20006770

University of Hawai'i Press
From Ancient Cham to Modern Dialects: Two Thousand Years of Language Contact and Change:
With an Appendix of Chamic Reconstructions and Loanwords
Author(s): Graham Thurgood
Source: Oceanic Linguistics Special Publications, No. 28, From Ancient Cham to Modern Dialects:
Two Thousand Years of Language Contact and Change (1999), pp. i, iii-vii, ix-xiii, xv-xvii, 1-259,
261-275, 277-397, 399-407
Published by: University of Hawai'i Press
Stable URL: http://www.jstor.org/stable/20006770
Accessed: 15-10-2015 08:16 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.
University of Hawai'i Press is collaborating with JSTOR to digitize, preserve and extend access to Oceanic Linguistics Special
Publications.
http://www.jstor.org
This content downloaded from 103.231.241.108 on Thu, 15 Oct 2015 08:16:36 UTC
All use subject to JSTOR Terms and Conditions
From Ancient Cham toModern Dialects
Oceanic
Special
Linguistics
No.
Publication
28
From Ancient Cham toModern Dialects:

Two Thousand
Language
With
Years
Contact
of
and Change
an appendix of Chamic
and
reconstructions
Graham
University
loanwords
Thurgood
of Hawai'i Press
Honolulu
1999 University of Hawai'i Press

All rights reserved
Printed in the United States of America
?
5 4 3 2 1
04 03 02 01 00 99
of Congress
Library
From
ancient
Cham
language
cm. ?
(Oceanic
and
Data
Graham.
Thurgood,
to modern
contact
p.
Cataloging-in-Publications
: two
dialects
/ Graham
change
linguistics
special
thousand
years
of
Thurgood.
publication
; no.
28)
Includes bibliographical references (p.

) and indexes.
ISBN 0-8248-2131-9
(pbk. :alk. paper)
2.
1. Cham
3.
language?Dialects.
in contact?Southeast
Language
499'.22?dc21
Cham
language?History.
I. Title.
II. Series.
Asia.
PL4491.94.T47
98-54334
1999
CIP
Camera-ready
University
and meet
of Hawai'i
the guidelines
Council
Printed
copy
prepared
by
the author.
are
on acid-free
books
paper
printed
and durability
for permanence
of the
on Library
Resources.
Press
by Cushing-Malloy,
Inc.
Contents
ix
Preface
xi
Acknowledgments
List
of Abbreviations
1 Introduction
and Conventions
xv.
4
Chamic convergence with Southeast Asia
The general tendencies: A broad overview
5
Previous work on Chamic reconstruction
6
The degree of adjustment
13
2 The Geographical
and Historical
The geographical setting

Chamic prehistory
15
17
Chamic history
The Chamic
3 Classification
and MK
Setting
14
inVietnam
27
14
languages
of the Chamic Languages
30
The place of Chamic within Austronesian

31
and broader affiliations
34
Malayo-Chamic
The Malayo-Chamic
subgroup 39
Within the Chamic languages: A sketch 40
a Chamic language 47
Acehnese,
A digression on Moken, a non-Chamic language
4 Altering
the Basic Word: From Disyllabic
58
toMonosyllabic
The iambic syllable: Early Mon-Khmer

influence 61
The loss of the vowel before medial -h- 63
and clusters 64
Disyllables with liquids > monosyllables
Loss of the unstressed initial syllable 65
vi Contents
5 Chamic Consonants
67
Presyliable consonants: The details 67

80
Main syllable onset consonants
Consonant clusters 93
Word-final
6 Chamic Vowels
consonants
99
104
The literature
105
The PC presy liable vowels
106
The inherited PC main syllable vowels
The borrowed PC main syllable vowels
PC vowel length
138
The main syllable vowels summarized
7 Nasals
and Nasalization
Nasalization
Nasalization
in Cham
Nasalization
Nasalization
Nasalization
in Tsat
160
inNorthern Roglai
inAcehnese
176
8 The Origins of Registers
151
152
in PC 152
in Chru
153
inHaroi
155
Nasalization
Nasalization
113
126
155
170
and Tones
178
179
Western Cham and the development of register
187
Phan Rang Cham, an incipient tone system
197
Haroi vowels and restructured register
tonal system 214
Tsat and its fully-developed
The
internal paths of change
9 PC Morphology:
Some Notes
232
237
A note on PC morphology
238
Verbs and verbal morphology
239
244
Nouns and nominal morphology
10 Contact, Multilingualism,
and Change
Non-linguistic
history 252
Chamic and the nature of language change
to a linguistic area 258
Accommodation
251
253
Contents
Vil
Appendix
I: Language Names
Acehnese
(and transliteration tables)
261
262
Bih 263
Chru
Haroi
Jarai
263
265
266
267
Krung
267
Noang
Phan Rang Cham
267
Rade 269
Rai 271
Northern Roglai
271
CacGiaRoglai
Southern Roglai
Tsat 272
272
Western Cham
Written Chamic
274
275
Appendix
272
II: The Chamic Lexicon
277
1.
to PC 280
Chamic Vocabulary Reconstructable
Words Borrowed After the Breakup of PC 346
3. English-Chamic
Index 365
2.
References
379
Author Index 395

Topic Index 399
Preface
The
title of this book From Ancient
toModern
Cham
Dialects:
Two Thousand
Years of Language Contact and Change, even more than showing my admiration
for J.Marvin Brown's work on Thai, reflects my belief that the term 'Cham' was
used at an earlier time as a general term for all the Chamic-speaking
peoples of
name
of the kingdom, Champa, it the etymological
Vietnam. It is reflected in the
source of the name Tsat used by the Utsat people of Hainan to describe their lan?
guage, and it occurs quite early in the Chinese dynastic records in their references
to Champa.
the title aside, two quite different views of research exist, each
Leaving
its own
with
one more
place,
the other
conservative,
more
One
exploratory.
view
is typified by the comment of the English poet Pope to lesser poets advising them
to keep their piece nine years. In this view, the endpoint is reached when every?
thing is completely clear and thoroughly documented. As will become evident to
exploratory
that
study
reconstruction,
reconstructs
first
on
focuses
in a quite different
has been written
readers, this monograph
2000
years
and
proto-Chamic
of
contact
language
tradition: it is an
based
then,
and
change.
on
that
Central
Asian
in this exploration
include the adaptation of Chamic to the Southeast
the
canonical
area,
linguistic
restructuring of the basic shape of the word,
major
changes
themes
ter,
tone,
in all
to the
and
these
and
vowel
and,
register,
of
inventories,
course,
the development
the role
played
by
of
regis?
bilingualism
developments.
from the outset
However,
offered
consonant
restructured
here
are,
even
more
than
itmust
usual,
be acknowledged
tentative,
preliminary,
that these analyses

and
undoubtedly
in
need of correction
In part, this reflects the fact that Chamic

and amplification.
studies are, in some senses, still in their infancy; in part, the numerous gaps inmy
background; and, in part, it reflects things that I have simply missed or misana
lyzed, and itwould be surprising if all my errors should prove minor. In a number
of ways, the conclusions presented here differ from conclusions reached in my
own earlier work,
ones.
In Chamic
be discovered
and where
studies,
only
they differ, these conclusions

the
seem
outlines
to be
clear;
supersede
far more
the earlier
remains
than has been found thus far.
IX
to
PREFACE
The preoccupation
throughout with borrowed forms directly reflects the
with
language contact and the relative chronology of language
preoccupation
contact. Thus, their presence in the tables and text is relevant, not only to the
involved, but also to the relative chronologies of the borrowings
If the loans participated in a change, they predate the change; if they
did not participate, then they were borrowed after the change.
reconstructions
themselves.
The original plan for this work did not include Acehnese.
It was only as
the first draft was nearing completion that a comparison with Acehnese made it
absolutely obvious thatAcehnese was simply a Chamic language whose speakers
lan?
had migrated to Sumatra, not a separate branch collateral with the mainland
guages. At that point the whole manuscript was reworked to incorporate Ace?
hnese, not with the intent of being definitive, but hoping to confirm beyond all
reasonable objections that Acehnese was Chamic. A myriad of details has been
left for another time when more
is known.
In addition to the main

ces:
one
is a set
of
texts, this work
transliteration
is accompanied
is an
another
tables,
by three appendi?
appendix
of
reconstructed
and borrowed forms, and the last is an appendix of forms arranged alphabetically
by English gloss. The transliteration of amultitude of differing orthographies was
this was done
necessary for clarity of exposition. Almost without exception,
aside from an error thatmay have crept in here and there.
mechanically,
The
borrowings
are
assessments
enough
other
lists
appendix
found
throughout
on.
based
both
the proto-Chamic
the manuscript,
Not
so that the reconstructions
all
the
reconstructions
along with
languages
are
the forms
included,
but
and
the
that these
there
are
can usually be done on the basis of the forms
given.
this appendix, both reconstructed forms and borrowings are

listed. Undoubtedly
there is some error in the identification of borrowings, but
none that invalidate any of the conclusions. A number of marginally
attested
Within
forms are included
in this appendix, along with the supporting data, with the

hope that other researchers will help clarify their etymologies. Where Chamic
forms are suspected of having as of yet unidentified Mon-Khmer
counterparts, it
is hoped that various scholars will help fill in these gaps.
Finally, it is expected that the forms reconstructed
Chamic
will
in the direction
be revised
Proto-Malayo-Chamic,
Chamic inscriptional
as better and more
in this work for Proto
of the forms found reconstructed

thorough use is made
data.
Graham Thurgood
California State University,
for
of theWritten
Chico
Acknowledgments
the single author, this work is a collaborative effort, only made possible
through the generous help of countless others. Much of my work on Chamic
would not have been possible without the help of Bob Blust, who provided me
Despite
with
insights, and copies of crucial papers. In particular, the

side of this work heavily reflects his help. He gave freely of
Malayo-Polynesian
his time, energy, and considerable expertise, making it possible not only to look
at the Chamic historical changes from the bottom-up?from
the modern Chamic
encouragement,
looking back to Proto-Chamic,

the perspective
of
but also to look at them from the top down?from
reconstructed
twin
These
Proto-Malayo-Polynesian.
tives produced amuch more sophisticated and insightful reconstruction

ysis than would have been otherwise possible.
perspec?
and anal?
The Acehnese portion of this work owes much toMark Durie, who con?
stantly supplemented my meager knowledge of Acehnese, patiently pointing out
over email what worked and what did not. He not only supplied most of the Ace?
a wealth of forms in an extremely valuable thesaurus
forms?including
(Daud Bukhari and Mark Durie, n.d.) and a hypercard program containing
Chamic-Acehnese
comparative material (Durie, 1990b), but he also looked them
over after the first draft of the manuscript was written, and supplied many that
hnese
were
still missing.
In a similar way, he critiqued and improved many of the anal?
not
of
Acehnese
but also those of Chamic. Inmany instances, he played
yses,
just
the devil's
advocate,
forcing
me
to rethink,
clarify,
and,
reformulate,
some
with
frequency, change positions taken in earlier drafts.

Whatever merits the analyses may have also owe a great deal to David
who
Solnit,
significantly
improved a number of the analyses in both major and
minor ways. The phonetic analysis of Tsat owes its essence to work done on it by
supplied a significant number of the forms. G?rard Diffloth,
and, most recently, David Tho?
Luangthongkum, Arthur Abramson,
IanMaddieson,
who
Theraphan
mas labored hard providing insights and straightening out my use of terminology
with regard to the terms register, voice quality, register complex, and so on and
clarified my discussion
sent me
invaluable
of the Chamic
materials
Phraya Prachakij-karacak's
interaction with
on Mon-Khmer
reconstruction
(1995) Some Languages
the MK.
as well
In addition, he
as a copy
of Siam, containing
of
a Jarai
XI
Acknowledgments
Xll
and a Rade wordlist
from the last century. At a much later point, Peter Ladefoged

the
suggested
explanation for the connection between breathiness and vowel rais?
ing and creakiness or tenseness and vowel lowering included here. Ouyang Jueya
kindly sent me additional forms, several of which turned out to be crucial for the
analysis of Tsat historical developments. Neil Baumgartner letme use the Cham
font he developed. A large number of other people made substantive suggestions
that are directly
reflected in the ideas presented here: Eric Oey, Patricia Donegan,

Jean
Stampe,
Tempeste, Zane Clark, Ibrahim b. Ismail, Ni Dabai, Joel
Jim Collins, Alan
Nevis,
Jerry Edmondson,
George Grace, Paul Benedict,
Stevens, David Thomas, Osh Larish, Keng-Fong Pang, Karen Mistry, and Elzbi
David
eta Thurgood. There

even recognize what
is no reason
to believe, however, that those thanked will

I have done with their suggestions,
let alone agree with
them.
In addition toMark Durie, who read the original "first" draft, a number
of other people kindly offered to read an earlier draft and provide me feedback:
David Thomas, Paul Benedict, John Wolff, Malcolm Ross, Martha Ratliff, Jim
and Jerry Edmondson,
and Sander Adelaar. Paul Benedict send me
Matisoff,
some
useful
notes
about
wider
Austronesian
as well
connections
as about
the ety?
mology of the u- prefix found inTsat. The feedback has been invaluable and has,
in some cases, substantially improved the description.
2 has been painstakingly
The Chamic lexicon contained in Appendix
gone through by a number of scholars, all of whom know more about Austrone?
and accuracy of the appendix owes its merits to
their help. Specifically, Bob Blust, Mark Durie, K. Alexander Adelaar, David
Thomas, and Paul Benedict have all contributed time and energy to the appendix.
In particular, Bob Blust, Mark Durie, and K. Alexander Adelaar commented on
sian than I do. The usefulness
the
forms,
one-by-one
where
necessary.
For
this
labor,
I cannot
thank
them
enough.
various people have knowingly or unknowingly helped me

with my understanding of the history and the historical documents pertaining to
Champa and the surrounding area. I wish to thank the following people for their
In addition,
and help: Bob Hsu, Barbara Andaya, Hilary Chappell, Pang Keng
Thanh
Nhan, Arun Sinha, John Wolff, Mike Feener (by way of John
Fong, Ngo
John Marston, Mackie Blanton, Chris Court, Mark Durie, Sander Ade?
Wolff),
laar, and the late Gwyn Williams.
suggestions
general way, I wish to acknowledge my intellectual debts to

my early teachers: my friend and mentor Jim Matisoff, who introduced me to
Southeast Asia, Mary Haas, who taught me much of what I know about the com?
In a more
parative method,
and Paul Benedict,
who never lost track of the big picture.
Acknowledgments
Xlll
Even more
to be modified
than usual, I fully expect some of the analyses presented here

on the basis of better and more detailed studies of the languages
cited, as well as on the basis of instrumental studies on these same languages. I

shall be astonished if all my errors should prove minor and grateful to readers for
their corrections. It goes without saying that, where these analyses differ, they
supersede my prior work on Chamic. This work is based upon research supported
by the National Science Foundation under Grant No. SBR-9512101.
List of Abbreviations
AC
Ay monier
Aceh.
Acehnese
Bahnar
(AC)
breathy register
C
and Conventions
and Cabaton
refers to the Bahnar forms cited inAy monier

Cabaton (1906)
the register complex that

includes a breathy-voiced component
refers to a form inAymonier and Cabaton (1906)
from Cabaton
Central-Eastern-Malayo-Polynesian
Coastal Chamic
Haroi, Western
Cham, and Phan Rang Cham
a stage not fully recoverable by reconstruction

hinted at by the transcriptions used in early
inscriptional Chamic
early-PC
and
a term used to designate
CEMP
but
1976; the numbers following

'Headley'
identify specific words discussed in this work
Headley
Highlands
(1906)
Headley
Chamic
Rade,
Jarai,
Chru,
Nb.
Nonthaburi Malay
NR
Northern Roglai
PAn
Proto-Austronesian
PC
Proto-Chamic;
N.
and Tsat
Roglai,
the earliest
stage
reliably
recover?
able by reconstruction
Phan Rang Cham
In this work, the Phan Rang Cham forms come

largely from Moussay's dictionary, with the conse?
quence that they are heavily influenced byWritten
Cham?and
thus
at times,
more
archaic
than
the
modern
spoken forms. Cited modern spoken forms

are specifically labeled as spoken forms.
PL
Pierre-Bernard
Lafont
xv
and Conventions
Abbreviations
XVI
PMK
PMP
Proto-Mon-Khmer
in a table, it refers to an
Proto-Malayo-Polynesian;
reconstruction
Austronesian
that
at
least
predates
Chamic; some of these forms do not reconstruct

all the way back to PMP
PNB
Proto-North-Bahnaric
PR Cham
Phan Rang Cham
PSB
Proto-South-Bahnaric
register complex
register
a complex of features thatmay include length,

pitch, and voice quality (phonation type)
the
same
register
complex
register
a term used to designate

includes
WMP
as
Roglai always refers toNorthern Roglai

specifically otherwise noted
Roglai
tense
(Smith 1972)
a tense-voiced
unless
the register complex
that
component
the
Western Malayo-Polynesian;
specifically,
PMP languages not in CEMP
reconstructed
borrowed
and
not
reconstructable
to PC
borrowed earlier than PC so reconstructable

x/*
0
borrowed but not clear if it reconstructs

indicates that the sound completely
indicates
that
no
relevant
examples
to PC
to PC
disappeared
have
been
found
(m)
(n)
-X
metathesis
vowel reflex the result of nasalization
are
Apparent irregularities in the correspondences
indicated by a hyphen followed a consonant indi?
cating precisely what is irregular:
-v
-c
-f
-VR
-t
irregular
vowel,
irregular
consonant,
irregular final,
irregular
vowel
irregular
tone,
register,
irregular nasalization, and so on. As the histori?

cal phonology is better understood, at least some
Abbreviations
and Conventions
XVll
of these apparent irregularities should disappear,

others will doubtless remain puzzles.
while
subscribed dot
The dot subscribed under various voiceless
stops
indicates that the following vowel is breathy, or
behaves as if itwere in the so-called breathy voice.
This orthographic convention makes the historical
developments far, farmore transparent, because
matches the transliteration found inMoussay
it
it allows for a straightfor?

ward, largely phonetically transparent translitera?
tion ofWestern Cham.
(1971), and because
Wr. Cham
Written
Cham from Ay monier
and Cabaton
(1906)
From Ancient Cham toModern Dialects
Introduction
Beyond the no longer fully-understood

deteriorating temples and an abundance
of beautiful statuary, little now remains of the Champa civilization
that once
flourished along the coastal plains of central and southern Vietnam (Ciochon and
James 1992:52-55). Almost a hundred temples along the coastline from Quang
Tri in the north to Phan Rang in the south give silent testimony to a mostly-for?
civilization of traders and artisans. Fragments of knowl?
gotten, once-formidable
edge of the past can still be retrieved, even from the crumpling remains of Tra
ki?u in the north, the first political capital of the Champa civilization. Today,
however, much of what is standing is no longer easily recognized, having been
buildings. In a valley seventeen miles to the
richer cluster of still upright buildings inMi-son, once
a center of Hindu worship where today twenty of the roughly seventy temples
built between the seventh and twelfth centuries AD still stand. Further south,
is the site of Vijaya, the new capital built after Indrapura
present-day Binh-dinh
absorbed
into the walls
west of Tra-ki?u
of modern
is amuch
fell in 982. Here
there is another group of temples

son, but on high places, possibly reflecting themore
were built. And, still further south, near the modern
Rang, are the important complex of temples called
built, not in a valley like Mi

perilous times inwhich they
cities of Nha Trang and Phan

Po Nagar, the spiritual center
in the south. These ruins are the physical remains of the Champa civilization,
which reached its zenith in the sixth or seventh century but has long since
returned to obscurity. Traces of its history remain in these fading temples and
record now inaccessible, as Ciochon and
sites, their legacy in an archaeological
James
wrote,
except
to trained
archaeologists
and
historians.
However, alongside the physical legacy embodied in these silent temples

from the distant past stands an enormously rich, still-living legacy, the modern
descendants of the Chamic language that was once the lingua franca of Champa.
This rich linguistic record lives on in the Tsat spoken on Hainan, the Rade, Jarai,
Haroi, Chru, and Roglai spoken in the southern Vietnam highlands, the Cham
Chapter
coast and the various Cham communities of Cam?

spoken along the Vietnamese
Acehnese
of
Sumatra?in
all the Chamic languages.
and
in
the
north
bodia,
an
The Chamic languages have
incredibly rich story to tell, one thatmay
ultimately prove more valuable to historians of language than the archaeological
records will prove to archaeologists. Captured in the Chamic linguistic record is a
luxurious complex of language changes, following a myriad of internal paths,
and responding to a diverse array of influences from other languages, that is
remarkable for both its richness and its clarity. The Chamic linguistic record has
much to tell us about the history of the Chamic languages and about the nature of
historical
change and the role of contact in that change.

The use of 'Cham' in title of the book From Ancient Cham
toModern
Two Thousand Years of Language Contact and Change reflects an inter?

between
its earlier usage and itsmodern usage. As back as a thousand years
play
ago and probably earlier, Cham was used as a general term for all the Chamic
Dialects:
of Vietnam. It is reflected, not just in the name of the kingdom,

+
pa), but also in the Chinese dynastic records. For example, the
Champa (Cham
History of the Song Dynasty (960-1279) records that, in 986, some Cham arrived
in Hainan from Zhancheng
(Zheng 1986:37). Zhancheng
(etymologically, Zhan
speaking peoples
'Cham' + cheng 'city') refers to Champa, with Zhan having subsequently under?
gone various sound changes within the history of Chinese. The form Zhan in the
dynastic
source
matches
the
phonetics
of
the
language
name
Tsat
(etymologi?
cally from Cham *cam) used by the Utsat people of Hainan (U 'people' + Tsat
'Cham', that is 'Cham people') perfectly. Other Chamic language names also
appear to have at one time included Cham as part of the name. The Phan Rang
Cham traditions refer to four different peoples with Cham as a component of
their name (Goschnick 1977:106): the Cham Raglai (the Roglai; from ra 'people'
+ glai 'forest'), the Cham Jarai (the Jarai), the Cham Kur (Cham + k?r 'Khmer',
theWestern Cham of Cambodia and Southern Vietnam), and the Cham Ro (from
Cham + ro 'remnant'). Goschnick
very tentatively associates the term Cham Ro
with the Haroi, but the match is far better with the Chru (from Cham + ro); it is
clear from its syllable structure that Chru is the reduction of two morphemes
and
Cham Ro is a perfect etymological fit. As for the Haroi, it is quite likely that their
name comes from the MK name Hr?, particularly in light of some of the other
variants of Haroi including Hroi and so on. However, an alternate designation of
the Haroi is the Bahnar Cham. That leaves only the Rade for which I have not yet
found a variant that includes Cham. To return to the main point, Cham was once
used widely to refer to the various groups of Chamic-speaking
people.
1.
It is likely that the designation

'forest people' was applied at different times
to more than one group of Chamic speakers. It cannot be automatically
assumed that all dialects designated Roglai are dialects of the same language.
Introduction
From Ancient Cham to modern dialects: Two Thousand Years of Lan?

and Change is a preliminary reading of that linguistic record. It
Contact
guage
reconstructs an early stage of Chamic as the foundation for tracing the develop?
ments of Chamic from itsAustronesian
forebears through early Chamic down to
itsmodern descendants, sketching the internal developments, noting the nature of
external contact, and examining the role played by each in the nature and direc?
tion of change. In the Chamic languages is hidden a rich storehouse of knowl?
edge, not just about the Chams and the history of the region, but also about the
interaction of language change and language contact: about external contact and
internal change, about the origins of register complexes from systems without
such complexes,
the convergence
about the origins of tones from nontonal languages, and about

of languages in a new linguistic area.
Not to be ignored is the potential Chamic contribution to the unraveling
of various puzzles about linguistic and non-linguistic history. Cham is the earliest
attested Austronesian
1975:53) dates
language. Coed?s (1939, cited inMarrison
the inscription found at Tra-ki?u near the old Cham capital of Indrapura (Amara
vati) from the middle of the fourth century, noting that this inscription is "le plus
ancien
texte,
"...the
oldest
actuelle
text,
connu,
presently
?crit
known,
dans
un
written
dialecte
that
malayo-polyn?sien",
in aMalayo-Polynesian
is,
dialect".
The Cham
inscription is older by three centuries than the "Old Malay"

of
inscriptions
Srivijaya in southeastern Sumatra. The text itself, associated with
a well near Indrapura, the old northern capital, is short but linguistically reveal?
ing,
as Marrison
notes.
yang naga puny a putauv. Ya ur?ng sep?y di ko, kurun

ko jem? labuh nari svarggah.
Ya ur?ng paribh? di ko, kurun saribu
thun davam di naraka, dengan tijuh kulo ko.
Siddham!
Ni
fortune!
this YANG serpent possess king. YA person respect DI him,

for him jewels fall from heaven. YA person insult DI him, for one-thou?
sand year remain DI hell, with seven family he.
Fortune! This
is the divine serpent of the king. Whoever

respects him,
for him jewels fall from heaven. Whoever
insults him, he will remain for
a thousand years in hell, with seven generations of his family. (Marrison
1975:53)
Marrison
(1975:53) observes that the language of the text is not that far from
in its grammar and its vocabulary. The similarities to
Cham or Malay
modern Malay and modern Cham grammar are evident in the yang and ya rela
modern
Chapter
tive markers,
of
syntax
both found in Cham,
the
sentence
equative
in the dengan
Ni
yang
'with' and di
punya
n?ga
'locative',
'This
putauv
the
one
in the
pos?
sessed by the king', in the use of punya as a genitive, and so on Hindu influence
a frequently used invocation of for?
is evident in the Sanskrit terms usiddham ?
?
tune;
n?ga
aka?
or dragon;
serpent
heaven,
svarggah
?
paribhu
to
insult,
nar
and kulo ?
is
of the remaining vocabulary
family". Most
as
the
Chamic.
from
earliest
the
oldest
And,
times,
transparently
Malayo-Polyne
sian text shows by the presence of Hindu terminology and its Indie script, contact
has played a significant role in Chamic.
hell,
the major focus of this study is on the history of language con?

Although
tact and change, the bulk of its work revolves around historical reconstruction of
(PC) and to the problems inherent in trying to reconstruct it, since
proto-Chamic
a historical
reconstruction
is necessary
for
recognizing,
and
unraveling,
interpret?
ing historical language contact. Without a reconstruction, it is often not possible

to differentiate between similarities that are genetically-inherited
and those that
are contact-induced.
driven by concern with understanding

the prior
history of Chamic language contact, the attempt has been made throughout this
work to identify and to determine, where possible, when individual loans were
incorporated into Chamic. There has been a consistent concern with differentiat?
However,
ing loans that predate the formation

PC.
These
sources
and
the
Even
intensity.
are
loans
chronology
more
about the historical
Chamic
an
source
of
important,
contact,
these
contact between
convergence
of PC and loans that postdate
important
with
but
loans
of
information,
not
also
sometimes
about
are
the peoples
Southeast
a crucial
source
the breakup of
about
just
its nature
of
the
and
information
involved.
Asia
languages have only begun to reveal what they can

about the history of the peoples of this area but certainly even this modest study
as valuable as the
reveals some important historical relationships. Nonetheless,
Without
doubt
the Chamic
considerations may be, the greatest value of the Chamic data may
non-linguistic
lie in the insights that Chamic gives us into the effects of language contact on lan?
guage change. As Eric Oey wrote (n.d.:l), the "closer study of the Chamic lan?
guages within their historical context offers an unparalleled opportunity
understand the nature of mainland
Southeast Asian areal influences
to better
that the
Chamic languages have undergone together with many of their non-PMP [Proto
Malayo-Polynesian]
neighbors." More recently, Solnit (1993:109) wrote express?
a
similar
view
the Chamic developments:
of
ing
"Clearly Chamic provides multi?
of
the
of
Southeast
Asian
areal features
ple examples
spread
including
tonality, and glottalized consonants, none of which are other?
monosyllabicity,
wise characteristic of Austronesian.
Its linguistic neighbors, from which the fea
Introduction
tures
are
spread,
are
all Mon-Khmer:
Bahnaric,
Katuic,
Khmer."
Viet-Muong,
insular PMP lan?

is available concerning
Indeed, because so much evidence
us
with
crucial informa?
guages, the Chamic languages of the mainland provide
tion about linguistic contact and adaptation to a new linguistic area.
2000 years ago, when theAustronesian-speaking
traders, artisans,
on
were
the
of Southeast
to
mainland
become Chamic arrived
and seafarers that
Asia, the language they spoke was disyllabic, nontonal, and nonregistral. Just as
importantly they had not had contact with the languages of themainland for some
About
four
thousand
years?recurrent,
sian speakers having
constantly
reached
self-renewing
about
myths
the islands through the Malay
Austrone?
notwith?
peninsula
standing.
linguistic record attests to the relatively "recent" arrival of Chamic

in
the immediate predecessor of all themodern
Vietnam:
proto-Chamic,
speakers
Chamic languages is a single unitary language, still quite close to its daughter
The
languages, with an obviously short time depth. If there is a single, linguistically

obvious, uncontroversial conclusion that can be drawn from the reconstruction of
it is that, despite the incessant uninformed statements to the con?
proto-Chamic,
that the Chamic
trary, the linguistic evidence alone establishes unequivocally
speakers from the
speakers of Vietnam represent an incursion of Austronesian
on
not
remnants
the
mainland
from the
the
of
Austronesian
left
islands,
speakers
some
initial expansion of Austronesian
six or eight
speakers out into the Pacific
thousand years ago. That the Chamic speakers are recent arrivals on themainland
is established beyond question by the linguistically-unitary,
quite tight-knit,
rather
easily-reconstructed
nature
of
PC.
Chamic
has
absolutely
none
of
the
diversity nor any of the time-depth associated with an older primary branch of a
language family and all of the characteristics of a recently-arrived, dialectally
unified immigrant group.
In the last 2000 years, the Chamic languages of Vietnam have undergone
in the canonical shape of their words, major changes in their
radical restructuring
consonantal
and
vocalic
phonological
basic structure of their phonological
inventories,
and,
in some
cases,
even
in the
systems.
THE GENERAL TENDENCIES:A BROAD OVERVIEW

return to the mainland by a pre-Chamic Austronesian
language that was
almost certainly disyllabic and atonal has provided us with an unusually clear
and external contact.
picture of the interaction between internal developments
The
Despite the essentially identical starting point provided by PC, the histories of the
in the subsequent paths
individual languages differed, although only moderately,
6 Chapter
of internal change and differed often radically in the nature of the resulting pho?
nological system. The crucial cause of these differences in the internal path fol?
lowed and in the ultimate phonological
system attained was the differences in the
contact situation, specifically, the differences in the phonological models encoun?
tered through contact. This exposure to new phonological
systems?through
to a significant degree which of the myriad of
available paths of change a given Chamic language actually took. External con?
tacts supplied the new models for phonological
restructuring and thus gave direc?
external
contact?determined
tionality
to the changes
not
others,
by selectively
some phonological
making
features, but
salient.
Among the general tendencies arising from contact with the languages
of the Southeast Asian mainland, beginning with the registral Mon-Khmer
(MK)
(Min), are the
languages and continuing with the tonal Vietnam and Hainanese
following:
that is,
> iambic > monosyllabic
increasing monophthongization,
the movement
from disyllabic
adjustments in vowel and consonant inventories,

the addition of glottalized consonants
the proliferation of vowel contrasts
the merger
and
ultimate
loss
that is,
of finals
the loss of voicing distinctions among the obstruents

radical changes in the phonological
systems, that is,
the development of register complexes and then tone systems
All of these represent areal tendencies in Southeast Asia and, in the Chamic data,
all of them represent directions of linguistic "drift" adapted under the influence
of contact with
Previous
languages already containing
work
on Chamic
In order to distinguish
larities, it isminimally
case, some of what
paucity of extensive
torical reconstruction
the features in question.
reconstruction
simi?
and historically-inherited
to
In
have
of
this
necessary
language history.
knowledge
is needed can be found in older texts, but given the relative
early texts this also means that it is imperative to have a his?
between
contact-induced
of PC. In the case of Chamic, we are dealing not with

contact
but
with a series of contact situations stretching over
situation
gle
period of time. And, given the limitations of the written records, it would
possible to discuss the contact in any meaningful way without a historical
struction: on a non-trivial
but simple
level, it would
neither be possible
a sin?
a long
not be
recon?
to accu
Introduction
alone begin to stratify them accurately as to when

rately identify the loans?let
and where they were borrowed, nor would it be possible to subgroup the dialects
and thus determine what changes correlate with changes in the contact patterns.
Thus, a reconstruction of Chamic is crucial to the unraveling the layers of histor?
ical
contact.
affil?
Although a number of earlier scholars recognized theAustronesian
iations of Chamic, or at least Cham, the earliest actual lexical reconstructions of
were those of Lee (1966) in his doctoral dissertation. A little earlier,
some of the sound correspondences between proto-Malayo-Polynesian
and Cham
had already been set out by Doris Blood (1962) in her concise and still useful
in Cham" Dyen (1971), often referring to
"Reflexes of Proto-Malayo-Polynesian
Chamic
(and to Thomas
Blood
the correspondences
between
(1963)) also discussed
and Chamic. Within Chamic itself, using data supplied
proto-Malayo-Polynesian
by Pastor Pham Xuan Tin, Dorothy Thomas (1963) laid out some of the internal
in her insightful "Proto-Malayo-Polynesian
reflexes in Rade,
correspondences
Jarai, and Chru". Lee's 1966 work followed. Then, Burnham (1976) incorporated
Haroi data into the reconstructions, while modifying
and improving a number of
in his work "The place of Haroi in the Chamic languages"
Lee's reconstructions
(sometimes citing personal communication with Lee).
In a general sense, Lee's 1966 dissertation, as modified by Burnham
1976, still remains the basic framework for the PC historical reconstructions.
Lee's dissertation contains 700 plus lexical reconstructions accompanied by pre?
inventories
liminary
occasional
of the PC consonant
or addition
modification
of
several
and vowel
forms
systems. Aside
here
or
there
from the
by Lee,
Burn?
ham, or others (cf. Durie 1990a), these reconstructions have remained essentially
unaltered until now, simply because so little has been done in the meantime on
Chamic
reconstruction.
Increases
in the database
since Lee 1966
1966, our knowledge of the mainland Chamic languages has increased

significantly, and much of this has been incorporated into these reconstructions.
Lee basically used four mainland
languages: N. Roglai, Rade, Jarai, and Phan
Since Lee
Rang Cham. The data base for this work not only includes additional sources for
Lee's four languages but has been expanded to include another four mainland lan?
guages: Haroi (Burnham 1976), Chru, Western Cham, and Written Chamic as
well
as Tsat,
spoken
on Hainan,
and Acehnese,
spoken
in northern
Sumatra.
The
addition of Haroi, Chru, and Western Cham resulted in minor adjustments here
and there, but little more. Further, although the additional use ofWritten Chamic
forms was sometimes valuable for its occasional preservation of an older form,
Chapter
aside from certain insights

in the
advances
into the older vowel
system,
it has led to only minor
reconstructions.
a language whose
significant was the inclusion of Acehnese,
as
as
was
seen
Niemann
1891.
with
Cham
early
Subsequent scholars
by
affinity
as
such
the
issue
affiliation
of
its
Cowan, Shorto, Durie and others
examining
Far more
similar conclusions. The argument for this position will be

for reconstruction comes
later in this work. The significance of Acehnese
from its early departure from the mainland, which sometimes enabled it to retain
archaic features that bear significantly upon PC reconstruction but which have
have also reached

made
inAustronesian
Increases
left behind on the mainland.
from the dialect continuum
disappeared
A second area where
material
comparative
our knowledge
has increased
is in the availability of reli?

and
PMP
reconstructions. Far more is
able, accessible proto-Austronesian
(PAn)
readily available to the comparativist now than was available in the late 1960s
and
and the early 1970s, leading to modifications
of our PC reconstructions
our
to
MK
between
PC
material
and
bor?
differentiate
native
ability
increasing
rowings, an area that is difficult but in which important progress has been made.
This work
PMP
has benefited
of Blust
reconstructions
to
laar. Access
clear,
internally
and
greatly
to
from having
the proto-Malayic
consistent
access
to the extensive
of Ade?
reconstructions
reconstructions
of
PMP
and
proto
fit
how prospective PC reconstructions
the higher order reconstructions of Austronesian
(An), specifically with
PMP (based on Blust's dictionary files (1990-1995) and personal communication
with Blust) and proto-Malayic
(1988, 1992) and personal
(based on Adelaar
Malayic
in with
has made
it possible
to examine
communication).
In particular, the presence of a large body of reliable PMP reconstruc?

to be done on both a bottom-up and a
tions allows the Chamic reconstructions
top-down
basis.
Here,
the
term
bottom-up
refers
in comparative work and requires no explication.

less common and does require comment. This
to the more
common
procedure
The term top-down, however, is

term was apparently introduced
into Austronesian
comparative literature by Blust (1972:1), who was discussing
from the top down". Certainly, this top-down
what he termed "reconstruction
refers to by the term
reconstruction
is what Anttila
(1972:346)
explicitly
"inverted reconstructions", which in reference to Chamic would simply mean
within Chamic are not
that the interpretations of the sound correspondences
based solely on the data within Chamic itself but are also based on our knowl?
is of
edge of higher-level PMP reconstructions. This use of an outside witness
particular value when
there is no obvious directionality
to a sound change,
that is,
Introduction
in those cases when a sound change could have gone either way, the ability to use
PMP reconstructions as a check often makes the direction of the change clear,
it obvious which of the alternatives within Chamic represents the
thus making
older
stage.
The availability of these MK and Austronesian

sources, when used along
our
the
PC
of
with
correspondence patterns, often makes it possi?
understanding
ble to not just identify loans but frequently to determine whether the loan pre?
the
dates or post-dates the breakup of PC. Beyond the expanded database?even
the major impetus behind the adjustments in
invaluable inclusion of Acehnese,
the reconstruction of PC comes from the greatly improved ability to identify
loanwords.
invaluable source that remains only

there is one potentially
earlier records of inscriptional Chamic. The more that we
partially tapped?the
know about the earlier inscriptions, the further back that we will be able to push
However,
The value of these older inscriptions is greatly increased

itself was a dialect continuum?a
the
fact
that
Chamic
early
string of related
by
a
coast
there is
As
and interacting dialects along the
of Vietnam.
consequence,
every reason to believe that there were Post-PC changes that affected all the
the PC reconstructions.
of the dialect continuum, with
members
stage
was
lost
everywhere.
Thus,
the result that the evidence
we
when
reconstruct
PC,
in reality,
of an earlier
what
we
are
is not the earliest stage of Chamic but a somewhat later stage.

reconstructing
What evidence we have of the earlier stages is, in part, provided by Acehnese,
which dropped out of the dialect chain, when it left themainland, thus saving evi?
dence
of
earlier
stages,
and,
in part,
by
the
inscriptional
evidence,
which
should
fill in some of the gaps.
Borrowings:
their identification
and interpretation
In order to analyze PC, it is essential to identify borrowings. The reason is that

the Chamic
languages have been in intimate contact with MK languages for
about two thousand years and during that time have borrowed extensively from
them. Without
to reconstruct
identifying the various layers of borrowings, it is neither possible

the history of the Chamic contact with MK nor to reconstruct
itself. Fortunately,
it is usually possible, not just to identify borrowed
items, but also to determine their relative chronology.
The problem of detecting borrowings is complicated by the existence of
Chamic
borrowing between Chamic speakers and theMK groups with which they
had contact. Thus, as David Blood points out (p.c.), there are numerous Cham
loans in Chrau, as the Chrau were dominated by the Chams, so the existence of a
mutual
form in Chrau does not authenticate
it as originally MK. He adds thatMnong
evi
10
Chapter
is also suspect, as the central and eastern Mnongs have been under heavy
Rade dominance and Bahnar evidence is also suspect because of heavy mutual
cultural interaction with the Jarai. Hr? evidence is similarly suspect because of
dence
heavy interaction with the Haroi, as is at least some Katu evidence, as the Katu
have clearly borrowed extensively from Haroi (G?rard Diffloth, p.c.).
it is likely that individual MK languages are largely free from
Although
Chamic
(David Thomas suggests Koho, Stieng, Rengao, Jeh, and

borrowings
this
work
has tended to identify loans, not on the basis of any
Bahnaric),
an
on
evaluation of the overall evidence: the degree of regularity
single factor, but
West
of the Chamic
Austronesian
able
sound correspondences,
etymology, the existence
reconstructions
of
various
MK
the existence
or non-existence
or non-existence
subgroups,
and
of a good
of the form in the avail?
so on.
In particular,
among
sources, the existence of a form in Bahnar ismuch weaker evidence of its

MK
in origin, than for example, its existence
in Thomas' (Prachacakij
being
karacak 1995) West Bahnaric reconstructions, H. Blood's (1967, 1968, 1974)
theMK
reconstructions
proto-Mnong
in many
(which,
cases,
are
often
proto
actually
as he draws heavily on Koho, Stieng, and Chrau (David Thomas,

Efimov's
reconstructions
(1987) proto-South-Bahnaric
p.c.)),
(from the Gage
translation of Efimov's wordlist), and Smith's (1972) proto-North Bahnaric. The
South-Bahnaric
of a form in several MK
existence
subgroups, of course, is still stronger evidence.

single work for the identification of MK borrowings
(1976), who focused on the potential MK borrowings in Lee's Chamic
The most
isHeadley
reconstructions.
He
valuable
estimated
that
10%
of Lee's
reconstructions
were
MK
bor?
a number
that now appears to be on the low side. Headley listed those

rowings,
forms he thought might be borrowings, either from MK or, in a small number of
cases, from elsewhere; inmost cases, based on his familiarity with MK, Headley
identified particular Chamic items as potentially borrowed from MK. In all but a
few cases, subsequent research has borne out his suspicions. Looking at his iden?
tification from a Chamic perspective, some of Headley's putative borrowings dis?
within Chamic,
play irregular correspondences
inherited An forms, and some exist both inMK
some have sounds not found
in
but have no history

one
or more branches
in
and Chamic
inAn. Further, some have been reconstructed

of MK (see sources listed in the paragraph above). In fact, further investigation
has fully confirmed most of the identifications. In addition to those, other forms
also reconstructed for PC by Lee (1966) and by Burnham (1976) have also turned
elsewhere
out to be MK
The
borrowings.
identification
of the extensive
layer of MK (and, occasionally,

has
had
several
other) borrowings
consequences. First, and in a sense
most important, is the realization of how extensive and how intimate theMK con?
tact was. The second consequence
is that the identification of these as borrowings
in Chamic
11
Introduction
has led to a reassessment
of the consonant
by Lee (and, by Burnham).

The bases for the identification
and vowel
inventories
reconstructed
of certain words as loans needs to be dis?
cussed, particularly as the strength of the supporting evidence may vary consider?
ably from case to case. Itwould be, of course, possible to simply assume that all
are native words unless there is overwhelming,
irrefutable evidence that
are
a
of the source. In the
identification
including
precise
they
borrowings,
stance
it
would
this
make
theoretical
Chamic context,
quite impossible to effec?
tively distinguish between inherited An forms and borrowed MK forms, leading
to the reconstruction of a PC that contained a large number of post-Chamic MK
words
loans. Thus, a less dogmatic, more pragmatic approach to loan identification has
been taken: if the sum of the evidence, taken as a whole, suggests that the word is
more likely a borrowing than an inherited form, then it is treated as such; if the
bulk of the evidence
In actual
suggests that it is native, then it is treated as native.

fact, few of the words are problematic. Certain words
are
loans. For example, if a form which occurs in Chamic is also

unquestionably
found reconstructed in one or more of theMK subgroups thus far reconstructed
for Vietnam, that is, if it occurs in one or more of Smith's proto-North-Bahnaric,
Blood's
Efimov's
proto-Mnong,
a preliminary
or Peiros'
proto-South-Bahnaric,
proto-Katuic?
of PKatuic, then from a Chamic viewpoint it is loan,

with the regularity of the intra-Chamic correspondence
patterns determining
whether it reconstructs to the PC stage or whether it was borrowed after the
reconstruction
breakup of PC. Further support is often provided by the failure of such forms to
occur in An languages outside of Chamic, although occasionally
such a form
may have a cognate inMalay.
an identified MK donor language itmay
Even without
reliably
which
designate
a form
as
lack AN etymologies
structions
available
only occur
but
which
loan.
For
and which
contain
in borrowed words. Aside
instance,
there
are
still be possible
a number
do not show up in any of theMK

"loan
phonemes",
that
is,
sounds
from the handful of An words
to
of words
recon?
which
that have
developed implosives under very specific conditions, forms with implosives can
usually be designated as loans. Similarly, as is established later in this work, cer?
tain PC words contain vowels which only occur inMK borrowings; these too are
as borrowings.
In other cases, it is not the presence of an unexpected sound that identi?
fies the word as a loan but instead the existence of marked irregularities in the
correspondence patterns; for instance, if the initial and the vowel are irregular in
marked
several languages and the word does not appear to exist inAn outside of Chamic,
are quite irregular within
it is a likely loan. Similarly, if the correspondences
mainland Chamic, the word does not seem to exist inAn outside of Chamic, and
12
Chapter
the form apparently does not occur inAcehnese,

it is likely a post- PC loan that
post-dates not just the breakup of Chamic but also themovement of theAcehnese
to Sumatra.
case
The
of
would,
course,
become
even
more
convincing
with
the
discovery of a likely MK source for the loan, but even without such a source, the
bulk of the available evidence suggests such forms are loans, not inherited forms.
Certain
are
words
native
unquestionably
An.
Certain
forms
reconstruct
back to PMP or even back to PAn. The PAn and PMP forms cited in this work
(unless otherwise noted) come exclusively from Blust's published work (or, per?
it is obvious even on the basis of my own limited personal
sonal communication);
the various PMP forms found floating around in the lit?
that
of
many
"expertise"
erature are unreliable. Not
surprisingly, the PMP forms with PC reflexes have PC

that show the forms are inherited, not borrowed.
correspondences
Other forms for which we have limited documentation
seem to be inher?
ited, at least at the PC level. Some forms occur outside Chamic, for instance, in
Malay, and pattern quite regularly within Chamic, but do not seem to reconstruct
to PMP. These are treated as inherited, at least from the Chamic perspective, a
for the analyses in this work, as
decision without any significant consequences
this
pattern regularly and there is no evidence of borrowing. Within
a
the
Acehnese
sometimes
into
context,
present
particular
Malay borrowings
problem, as not only was there a long period of mutual interaction between the
the words
Acehnese
and
the Malay,
but
also
some
of
the borrowings
may
not
be
readily
dis?
from inherited material.
tinguishable
In practice, of course, a specific word might be difficult to evaluate. Usu?
the
ally
problem revolves around the lack of information about its wider distribu?
tion, particularly its wider distribution in theMK languages. For AN, Blust's files
and his comparative dictionary (in progress) are extensive enough to provide a
reasonably accurate estimate about a particular word's distribution in AN as a
if a form does not occur in either Smith (1972), in Blood
(1967, 1968, 1974), in Efimov (1987), in Thomas (Prachacakij-karacak
1995) or
in Peiros (1996), it is difficult to tell if the gap is fortuitous or meaningful. That
whole;
however,
forMK
is, there is no obvious way for me to determine if the word is likely to be a native
MK etymon or not. A particular area of indeterminacy involves words found in
Bahnar and one or more Chamic languages but not in any of theMK reconstruc?
tions; since Bahnar has borrowed numerous Chamic words, without more evi?
dence it is unclear what the direction of the borrowing was.
Nonetheless,
despite the existence of countless potential areas of diffi?
culty in the identification of loans, the actual task is usually not problematic with
the overall level of reliability quite high. Discussions
of the status of particular
words
are scattered
throughout
this work,
as are discussions
of the evidence
the designations.
for
13
Introduction
Finally,
post-PC
it is important to distinguish
Thus,
borrowings.
if the
intra-Chamic
between
pre-PC borrowings
correspondence
patterns
are
and
regu?
lar, the loan predates the breakup of PC. In such cases, the word ismarked by the
to PC, while at the same time also
asterisk as reconstructing
accompanying
x to
indicate that, although it reconstructs to PC, it is nonetheless an
marked with
early pre-Chamic borrowing from MK, e.g., x*?arj 'hole; door' was borrowed
from MK before the PC stage. On the other hand, if the intra-Chamic correspon?
are irregular, the form was borrowed after the breakup of PC. In these
x
cases, the form is simply labelled with
indicating it is a borrowing, one that
'crossbow'
is a post-PC borrowing from
formation
of
the
PC, e.g. xsrap
postdates
dences
MK.
The degree
of adjustment
the late MK
loans, the
improved identification of loans?particularly
to
occasional
due
the inclusion of Acehnese,
expanded data base?particularly
reconstruc?
in
the
of
earlier
and
the
increased
adjustments
availability
analyses,
a clearer picture of the language that
tions of PMP and proto-Malayic?providing
The
PC evolved out of, all have resulted in the accumulative modification

of countless
details: roughly 60% of the specific lexical reconstructions differ in minor or
major ways from the reconstructions found in Lee (1966). This accretion of small
changes combined with the identification of the bulk of the post-PC MK loans
has
resulted
in revisions
in the PC
vowel
and
consonant
inventories
(with
the
spe?
cific details found scattered throughout this work). Nonetheless,

despite the large
number of smaller changes, Lee's pioneering work still remains the framework
for
these
reconstructions.
The
Geographical
and
Historical
Setting
The history of Chamic linguistic contact needs to be placed

cal and geographical context.
in a broader histori?
THE GEOGRAPHICAL SETTING

The geographical
setting has played its part in the history of the Chamic-speak
ing people. Vietnam is readily divided into the highlands formed by the Annamite
Cordillera and an often quite narrow strip of land that runs along the coastline,
with the north-south coastline itself being cut up into segments by the various riv?
ers that flow eastward into the South China Sea. Thus, these rivers, which pro?
vided good harbors, carved the geographical configuration around which various
riverine political entities developed. The rivers contained the harbors, with the
surrounding coastal strips integrated as rice growing areas, and the highland areas
were more loosely integrated into the coastal political entities as sources of high?
In the
lands forest products, often as important for trade as for consumption.
there were frequent power struggles between the political units that
in
the various river basins, first between Funan and Champa and later,
developed
after the decline of Funan, between the various riverine entities within Champa
itself, with frequent shifts in power as one or the other river-mouth harbors rose
early years,
to ascendancy (cf. K. R. Hall, 1985). Later, the Vietnamese would move south
along the coast, settling in the areas most suited to wet rice irrigation; this move?
ment naturally forced the Chams away from the coastal strip up into the high?
lands.
14
and Historical
Geographical
15
Setting
settlements existed in coastal areas because of the harbors,

Austronesian
the first major harbor being at Oc-eo, the port city of Funan (see "Chamic
settlement elsewhere, other
history" on page 17). As is typical of Austronesian
with
later settlements were
scattered along the narrow belt of land along the coast, typ?
to
mouths.
As part of the trading network, trading posts were
close
river
ically
also established in various places, including the southeast part of Hainan island,
where some of the northern Cham would flee in 986, with a second migration fol?
in 1486, and in parts of Cambodia. During most of its history and cer?
tainly during its ascendancy, Champa remained coastal. And, despite the fact that
the MK speakers are now largely found in the highland areas, it was in and
lowing
that the initial, intense contact between Aus?

tronesian speakers and MK speakers occurred. However, first after the loss of the
northern capital in 982 and then again after 1471 when the Vietnamese moved
around these coastal communities
down along the narrow coastal plain, Chamic

lands, while still remaining in contact with MK
Chamic
speakers retreated
into the high?
speakers.
prehistory
observes, Southeast Asian prehistory has often been

Key es (1995:182)
divided into periods named after sites found in northern Vietnam: the Hoa-binh,
As
the Bac-son,
and
the Dong-son
cultures.
However,
the Chamic
speakers
are
only
and instead are asso?
marginally associated with the Dong-son cultural complex,

ciated with a fourth complex, the Sa Huynh culture of southern Vietnam.
In Bellwood's
summary of the Sa Huynh culture (1985:275-279ff;
sites Bellwood
the
mentions
have been indicated in Figure 1), he writes that the sites

stretch from Hue and Da Nang in the north, to the type site Sa Huynh near the
central Vietnam coastline, and then southwards to the Mekong Delta. Bellwood
(1985:278) reports that the carbon dates from Phu Hoa and Hang Gon suggest an
overall date range for the Sa Huynh culture between 600 BC and 0, but he goes
on to say that the assemblages continue well into the middle of the first millen?
nium, making the first part contemporary with the Funan civilization and the later
with the early Champa civilization.
Bellwood
(1985:276) associates the Sa Huynh culture with the Chamic
settlers from the islands. The Sa Huynh burial jars have close parallels in the
Early Metal phase jars in the Philippines, northern Borneo, and the Celebes Sea
region of northern Indonesia. The dates at Niah and Tab?n forjar burial go back
as early as the end of the second millennium BC, suggesting that the practice in
predates the practice in Sa Huynh. Bellwood dismisses sugges?
tions that the stone burial jars of the Sa Huynh should instead be associated with
similar practices in northern Laos. Other parallels are found in the "almost identi
northern Borneo
16
Chapter
cal knobbed pennanular stone earrings (the so-called 'lingling-o'), and of a spe?
cial kind of earring or pendant with two animals heads (presumably deer), in a
number of sites inVietnam, Palawan and Sarawak".
Figure
1: The Sa Huynh
culture
Red River
+ Bac-son
Son
"JxDong
+1 Lang Vac
a
Key:
Site
Hainan
Hue
+
>+^>aNang
South
China
Sea
+ S Sa Huynh
+y^
Q vn+
Oc-eo
(Funan) I
PhuHoa
lang Gon
Mekong
The Sa Huynh sites reflect the Austronesian-speaking

group that settled
on the coast of Vietnam from an earlier homeland in perhaps Malaya or, more
likely, Borneo, sometime before 600 BC, although even earlier dates are now
It was some of these Austronesian
archaeologists.
being reported by Vietnamese
contact
with the MK people then living
after
extended
speaking people who,
along the coast, would become the speakers of Chamic.
and Historical
Geographical
17
Setting
records are supplemented

the archaeological
salvaged through historical reconstruction, which
Here
record,
by the linguistic
indicates that this
con?
period included intense, intimate, assimilative linguistic and non-linguistic
tact between the pre-Chamic Austronesian-speaking
and
MK-speaking
people
peoples, as evidenced from the massive
incorporation of loan words (including
pronouns,
basic
length
the
structure,
contrasts,
and so on), the restructuring
terms, basic vocabulary,
kinship
word
and
of
incorporation
new
consonants,
new
of the
vowel
vowels,
so on.
CHAMIC HISTORY
must have immediately
Hall
(1955; 1981) and Blust
brought
speakers.
coast
(1992a) argue that the first contact was probably not along the Vietnamese
at
the
South
China
but
rather
the
Gulf
of
Funan,
Thailand,
Sea,
facing
facing
around 500 BC, or perhaps even earlier?at
the site at Oc-eo, the port city of
Funan. By the middle of the sixth century, Funan had been conquered by the
The first Austronesian

them
incursions
onto
the mainland
into contact with MK
Khmers, but both Hall (1955:23; 1981:25) and Blust (1992a:50) suggest that ear?
lier Funan was Austronesian
speaking. In the first edition of his 1955 history of
Southeast Asia, Hall asserted that the Funanese were Malays
(1955:23, cited in
Blust (1992a:50)), a position that he repeated some twenty-five years later, when
he wrote writes
that the "Funanese were of Malay race, and still in the tribal state
(1981:25). Blust, citing Hall (1955) with approval, points
at the dawn of history"

out that (1992a:50):
The fall of Funan probably has a special significance for understanding

the history of AN [An] languages on the Southeast Asian mainland.
If
Funan
was
AN-speaking,
in the early
centuries
of
the Christian
era
a sin?
gle dialect chain would have extended almost unbroken from the south?
ern tip of theMalay Peninsula to Champa. The expansion of the Khmers
into the region of theMekong delta would then have divided an earlier
language continuum
into two separate and smaller dialect chains,...
question of whether or not Funan was originally Austronesian

speaking aside, it is likely that the long period of contact between Austronesian
and MK speakers began at Funan. If so, this would further suggest that the initial
language contact was not with the more northerly Bahnaric but rather more
groups; and, even ifHall and Blust are not correct, the first
southerly Mon-Khmer
The
intense contact also began at Funan. In any case, one would certainly still specu?
late that there was a string of trading posts down along the coast, stretching as far
Chapter
18
south as the eastern coast of modern Malaysia
which were dominated
by Austro?
traders.
nesian-speaking
Without records of the actual language spoken, the evidence for Funan
the available
Austronesian
being
speaking remains circumstantial. Nonetheless,
Chinese descriptions record at least an Austronesian presence in Funan and along
the coast to the south. K. R. Hall (1985:38) notes that, in 240, K'ang T'ai, aWu
envoy to Funan, reported to the emperor that Funan's authority reached from the
lower Mekong Delta to the upper Malay Peninsula, a stretch coterminous with
what was almost assuredly a string of Austronesian
speaking trading colonies. In
the Southern Ch'i history, the Funan of Jayavarman, the great king who died in
is described (D. G. E. Hall 1981:33) as a community of seafaring people,
"carrying on both trade and piracy, and constantly preying on their neighbours", a
514,
picture that would just as accurately describe the neighboring Chams. To extend
to the people of Funan, they were "a
Durie's characterization
of the Acehnese
people of the coastal margins, engaged in fishing the sea, in wet rice cultivation
of their preferred homelands,
in maritime
trade (and sometimes piracy)..."
are
a
of very Malay-like
(1996:114). That is, the descriptions
people.
The archaeological
record also suggests what Hall interprets as an Aus?
remains
tronesian connection. K. R. Hall (1985:40) describes the archaeological
at Oc-eo as showing that "the coast was occupied in the early first century AD by
Malay fishing and hunting groups". This interpretation is based on the conclusion
Sa Huynh culture both show the influence of Borneo
not
it
is
clear
that the influence is from the same region of
people, although
evidence from Borneo
Borneo. K. R. Hall (1985:40) writes that archaeological
that Funan and the Chamic
from
the presumed
on the west
known
as
site
of Yeh-po'-t'i,
coast of Borneo,
and which
y?pas"
which
an ancient
trade
notably
included
"substantiate
a Borneo
port
to have
believed
"carved sacrificial
cultural
link
to Funan".
than
logical evidence.
extension
of
one
the an earlier
exists
however,
explanation,
If, as speculated
by Hall
and Blust,
Austronesian-speaking
to account
for
the
the Cham were
Funanese
kingdom,
posts
Else?
Sa Huynh
(1985:276) has suggested that aspects of the Chamic

culture show links with northern Borneo culture.
where Bellwood
More
been
archaeo?
simply an
these
con?
nections receive a simple explanation. First, however, it is not clear that the
Borneo influences on Funan and on the Sa Huynh cultures were the same, leaving
the possibility we talking about more than one group of people. And, second, the
establishment
of a Borneo
does not automatically make it with Aus?

It seems evident to Adelaar and to myself that many
connection
people.
tronesian-speaking
of the now Austronesian
languages on Borneo
would
associate with a language shift from MK
substratum, one that I

toAustronesian.
show aMK
and Historical
Geographical
Setting
19
The historical
records provide several types of evidence that indicate

that the people of Funan were the same people found in neighboring Champa, but
this leaves open the possibility
that the people themselves were MK speaking.
For instance, there is a fifth-century Chinese report of a Funan prince who fled to
that is, Champa, and eventually became king of the Chams (K. R. Hall,
1985:71). However, it is easy to read too much into this since the later history of
Champa is dominated by intense contact and extensive cooperation between MK
Lin-yi,
and Austronesian
speakers. For instance, to paraphrase K. R. Hall (1985:178), the

Chinese envoys who visited Funan in the 240s reported that around 220 the
Chams and the Funanese were making joint naval raids and land attacks against
the Red River Delta
region. Thus, these frequent reports of joint Funanese and

ventures
Cham raiding
only argue for mutual cooperation.
It is important to point out both that there was also frequent tension not
just between Funan and Champa but also between the various coastal centers
within Champa itself. As K. R. Hall (1985:178) writes, "... the Cham realm's
early history was characterized by shifting alliances among regional centers that
were
at the river mouths of the Cham coast..." Changing power

were
at
the root of this factionalism. Funan's power base was mostly
relationships
solid when trade routes were either overland or at least followed the coastline
concentrated
allowed more
closely. As better sailing techniques and improved knowledge
direct routes to be taken to China, the geographical position of Funan ceased to
be an advantage and instead became a deterrent, with the bulk of the international
trade
to Champa
shifting
worked
themselves
out,
the
along
some
tension
eastern
coast
between
of Vietnam.
Funan
and
As
Champa
shifts
these
was
inevita?
ble. By the end of the sixth century, however, the changes in trading routes left
Funan a commercial backwater, as virtually all international trade shifted to the
ports
along
Funanese
the Vietnamese
coast.
In part
as
reaction
to
these
changes,
the
inland, refocusing their economy on developing rice

lands in the upper Mekong Delta (K. R. Hall, 1985:75), and by the end of the
sixth century Funan had ceased to be even partially Austronesian-speaking,
rulers withdrew
instead part of the Khmer world.

As for Champa itself, the first date of import for its future historical path
is probably 111 AD, before the first historical reference to the Chamic people
becoming
themselves.
In that year, northern Vietnam
became a province of the Han empire,

of north Vietnam. In particular, the
use
to
"Vietnamese peasantry began
Chinese methods of irrigation and agricul?
tural terracing" (Keyes 1995:182). In later centuries, these methods of irrigation
and terracing would be brought to the south, with consequences
for theMK and
an event that would
Chamic
peoples
lead to the Sinicization
then living on the coastal plains.
20
Chapter
term refer?
appearance of the Chams themselves?a
in a role that they would
ring at that time to all the Chamic speaking people?was
over
the
In
centuries.
the
137
AD, Coed?s
year
suggests
play repeatedly
a
"a
from
band
of
about
barbarians
thousand
(1968:43),
beyond the frontier of
The initial historical
Jih-nan" attempted to invade Hsiang-lin. Coed?s continues, "their name Ch'ii

lien, although written with different characters, can scarcely be dissociated from
that of the founder of Lin-yi."
the first recognition
of Champa
However,
itself, writes Coed?s
(1968:17), was by the Chinese, who began to speak of the kingdom of Champa
along the coast of Vietnam around 190 to 193 AD. Coed?s wrote (1968:42) that
The History
governor
includes, around the year 280, a report in which the Chinese

of Tongking
complains about the raids of the kingdom of Lin-yi
of Chin
(Champa). The governor, T'ao Huang, says of the kingdom that it "touches Funan
in the south. Their tribes are numerous; their friendly bands render mutual aid
taking advantage of the ruggedness of their region, they do not submit [to
China]."
By about the sixth century the Champa civilization reached its zenith,
but continued to flourish until the Vietnamese
"push to the South" began in the
a
tenth century. Meanwhile
writing system based on Indie models evolved, in
which the earliest Cham linguistic record is the inscription found at Tra-ki?u, dat?
ing from the middle of the fourth century (The inscription is cited in full on
page 3). Various temples were built in the north, particularly around the spiritual
another cluster to the south at the site of Vijaya, and still fur?
capital of Mi-son,
are
ther south
the important complex of temples called Po Nagar near the modern
cities of Nha Trang and Phan Rang. This whole
tact with MK speakers.
The Chamic
where
dialect
period
involved continued
con?
north at least as far as Quang-Tri,

Katuic speakers. The linguistic evidence
chain extended
they undoubtedly encountered

suggests that it was the Chamic speakers who left their linguistic imprint on the
Katuic languages in the form of borrowings and perhaps even some morphology,
if as Reid (1994) argues the morphology
found in some dialects of Katu isAus?
tronesian, particularly Katu proper. Under pressure from the north, these Chamic
speakers left, probably to become the modern Acehnese of northern Sumatra.
The beginning of the rapid decline started when the Vietnamese
sacked
at
the capital
Indrapura in 982. From this point on, the remaining history of
historians term the "push to the
Champa is dominated by what the Vietnamese
down along the coastal plains of Vietnam.

South", the Vietnamese movement
Within
after
the
twenty years
sacking of Indrapura, the Chams had effectively
abandoned their northern provinces and the capital was moved further south to
Vijaya.
Geographical
and Historical
Figure 2: Champa from
With
21
Setting
inception to absorption
(important sites)
the fall of Indrapura, the Chams inhabiting the northern provinces

the departure of the Acehnese,
the northern Cham con?
resettled elsewhere. With
sisted of two groups of still identifiable modern descendants: the Tsat speakers of
("The history of Tsat contact" on page 224), who are quite probably the
in the Chinese report of 986, and the Northern
refugees from Champa mentioned
Hainan
Roglai, who fled to the south with the fall of the capital, eventually coming to live
in the Vietnamese highlands. History also records a third group of refugees from
northern Champa, the group recorded by Guangzhou
(Canton City) as 310 refu?
from
cited
gees
Zhancheng. Although
by Zheng (1986:37) in connection with
providing
confirmation
of the dating the arrival of the Tsat
in Hainan,
the group
22
Chapter
to have fled to Guangzhou, not Hainan. Since the Chams more than
in Guangzhou
(Canton) at the time, Guangzhou
likely had trading connections
itself seems
sense
makes
as a destination
The
for
the emigrants.
shows that the Northern Roglai and the Utsat

linguistic
were once speakers of the same northern Cham dialect. Zheng (1986:37) notes
that the History of the Song Dynasty (960-1279) makes it clear that some of the
northern Cham went to Hainan. Specifically,
in 986, Pu-Luo-E and a hundred of
evidence
his clan arrived, having not just fled Zhancheng

(Champa) but having been
harassed by the people of Jiaozhi, the name given by the Northern Sung dynasty
to northern parts of Vietnam. This report, including the Pu used in the name of
the leader, affirms the arrival of the Cham from Champa, the group we now iden?
tify as the Utsat. Having arrived in Hainan, these Cham entered into the Hain
anese linguistic area, an area characterized by the richly tonal Be and Li of the
Tai-Kadai family and the Southern Min Hainanese dialect of Chinese.
The effective end of Champa as a regionally-dominant
political entity
came with the fall of the southern capital at Vijaya in 1471. Much of the remain?
der of Champa was incorporated
into Vietnam
and, although remnants of
now
some
was
to small territories situ?
for
it
reduced
existed
time,
Champa
quite
even
ated south of Cape Varella, where
today many Cham still live.
Just as happened after the fall of the northern capital, the fall of Vijaya
also
led
to
diaspora
of
Chamic
speakers,
some
to
the
highlands,
some
to
Hainan, Guangzhou
(Canton), Malaka, Aceh, Java, Thailand, and Cambodia (the
Western Cham). With reference to Hainan, again citing from Zheng (1986:37), in
1486 the True Records of the Emperor Xian Zong of the Ming Dynasty
(1368
1644) record over 1000 new refugees inHainan from Zhancheng (Champa). With
reference to Guangzhou
(Canton City), in 988 AD the Guangzhou records report
310 refugees from Zhancheng. The Sejarah Melayu records that after the fall of
Vijaya, the two sons of King Pau Kubah fled, with Syah Indera Berman going to
Malaka, and Syah Pau Ling going toAceh where he started the line of Aceh kings
(Abdul Rahman
arrival of Cham
ence of a Cham
1994 [1987]: 104). Other sources and texts record the

account notes the exist?
in Java. In Thailand, a 1662 missionary
at
colony
Ayudhya (Lafont 1994b:73) these appear to the same
al-Ahmadi
group now found in Bangkok, where they still exist as an identifiable ethnic com?
munity. At various times, Cham also fled to Cambodia. For instance, after the fall
of Vijaya in 1471, the Khmer Annals record thatmany Cham took refuge in Cam?
bodia; later, in 1692, the Khmer Royal Chronicles record another large migration
(MakPhoen
1994:76-77).
As Figure 3 shows, this last major ancient diaspora along with several
still later, more modern migrations correlates with the modern distribution of the
Chamic
languages.
Some of the Chamic
speakers retreated into the highlands
of
and Historical
Geographical
Setting
23
Vietnam, others fled to Hainan to become speakers of Tsat, some went to Cambo?
dia (apparently some as war captives, others willingly) where they speak what are
Western Cham dialects, and still others ultimately went to northern Sumatra,
eventually
to become
speakers of Acehnese.
Figure 3: The modern distribution
of Chamic
Once they were separated from the Chamic community, as Durie notes
the
Acehnese
re-entered the Austronesian world, greatly expanding their
(p.c.),
contacts with Malay while their language's typological convergence with MK
it remains unclear when and by what
languages was largely arrested. However,
route theAcehnese made their way to north Sumatra. As for dating their arrival in
Sumatra, Durie (1996:115), for example, assumes thatAcehnese may have been
there for a period of more than a thousand years, based on the striking dialect
diversity. Any real precision with respect to an arrival date is not possible without
more
information
The Acehnese
than we
now
have.
and Kelantan
took on their travels, there is evidence that

respect to the path the Acehnese
there was an important, Chamic presence in Kelantan, on the east coast of the
With
Chapter
24
Malaysian
peninsula but this influence looks to have been quite late. As Abdul
traditions
Rahman al-Ahmadi (1994 [1987]: 105-106) notes, local Kelantanese
an
area
some
authors believe
have the King of Kelantan coming from Kembayat,
to be Champa, although others maintain it is Cambodia.
of a Chamic presence is found in the numerous
Less
evi?
controversial
place names related to

Champa: Pengkalan Cepa, Kampong Cepa, and Gong Cepa, to cite but a few. The
Cepa of these names is obviously Champa, with the expected sound changes.
These place names and other influences were the result of an Acehnese presence
dence
not just a Chamic presence. G?rard Diffloth has suggested that Ace?
hnese contains loanwords restricted to Aslian; however, even if data showing
the
Aslian
loanwords from groups on the Malay peninsula can be assembled,
in Kelantan,
finding would
had
a presence
substantiation
only provide
for the contention
that the Acehnese
in Kelantan.
Table 1: The Malayic
dialects
PMP
Nb.
Malay
of Nonthaburi,
Pattani
Pattani, and Kelantan
Kelantan
*-i
-i
-i
-i
*laki
lakilaki
laki
llaki
*waRi
hari
*gigi
*beli
gigi
b?li
(I)
PC
*-sy
*laksy
'male;
person'
gigi
gig*
ayi
*hursy
'day;
gigi
bali
*gigsy
'tooth'
*p9-blsy
sun'
'sell'
-u
-u
-u
-u
*-ow
tahu
tahu
tahu
tahu
*thow
'know;
able'
can;
*kutu
kutu
gutu
kutu
*kutow
'louse,
head'
*kuku
kuku
kuku
kuku
*kukow
'claw;
*kayow
'tree;
-u
*taqu
-i?
fingernail'
*kahiw
kayu
kayu
kayu
kayu
wood'
*batu
batu
batu
batu
batu
*batow
'stone'
*baqeRU
baharu
bayu
bayu
bayu
*bahrow
'new; just
now'
*aku
aku
ku
*qabu
abu
?abu
abu
*kow
'I (familiar)'
*habow
'ashes'
Two things argue for this Acehnese presence in Kelantanese being fairly
recent. First, the records of Acehnese
presence in the area themselves suggest a
recent
Kelantanese
itself shows little other Acehnese
presence. Second,
fairly
influence. Despite the obvious Chamic influences in the place names, the modern
dialects
of the northeast
coast of Malaysia
and southern Thailand, Nonthaburi
Geographical
and Historical
25
Setting
Malay (a group relocated from Pattani to central Thailand),

Kelantanese Malay subgroup, not with PC, but with Malayic.
show, with reference to four obvious vocalic innovations
Pattani Malay and

As Tables 1 and 2
the
differentiating
from
of
Nonthaburi
(data
Malayic
Malayo-Chamic,
Tadmor 1995), Pattani (data from Tadmor 1995), and Kelantan (data from Abdul
Hamid Mahmood,
1994) consistently pattern with Malay, the representative of
branches
and the Chamic
not with PC.
Malayic,
Specifically, as Table 1 shows, in final position Chamic has innovated,

the two high PMP vowels *-i and *-u, giving *-ey and *-ow,
diphthongizing
one of these three dialects shares this Chamic innovation.
not
respectively;
Table 2: The Malayic
PMP
Malay
*-ay
*ma-atay
dialects
of Nonthaburi,
Pattani
Nb
Pattani, and Kelantan
Kelantan
-i
-i
-i
-i
mati
mati
mati
mati
(II)
PC
*-ay
*matay
'die'
*qatay
hati
ati
hati
ati
*hatay
'liver'
(*qaqay)
kaki
kaki
kaki
kaki
*kakay
'leg; foot'
*-uy
-i
-i
-i
-i
*hapuy
api
babi
?api
api
api
*babuy
*-uy
*?apuy
'fire'
*babuy
'wild pig'
As is obvious both from the PMP forms in Table 2 and from Adelaar's
1992 reconstruction of Malayic, Malayic has innovatively monophthongized
the
two PMP final diphthongs *-ay and *-uy, merging both as *-i; the three dialects
share this Malayic
innovation. Thus, although there have been some Acehnese
influence in Kelantan, it is obvious not just from this evidence but from all the
nor Chamic, but aMalayic
is not Acehnese,
linguistic evidence that Kelantanese
dialect.
Two
other
general
observations
can
also
be made
about
the
forms
of
Table 2, neither of which invalidates their subgrouping implications. First, it is

not the case that PMP *-ay forms all have -i reflexes inMalay (and in western
Indonesia as well as in a small set of languages historically heavily influenced by
Malay); some forms keep -ay reflexes inMalay (Blust, p.c.). Dyen, for example,
reconstructed *-ey (in addition to *-ay) for precisely those forms that have Malay
-i reflexes. However, whatever the ultimate solution to the problem posed by the the fact remains that the
i/-ay split in reflexes for higher level reconstruction,
not
Kelantanese
reflexes pattern with Malayic,
with PC. To the degree that the
-i reflexes are unexpected,
the argument that Kelantanese
Malayic
subgroups
with Malayic,
not Chamic,
is only strengthened.
Chapter
26
Note that the limited amount of data presented here makes the solution
to the problem seem obvious: simply reconstruct both *-ey and *-ay. However,
the apparent obviousness disappears once it is realized that this particular split in
reflexes is limited basically to PMalayic and, as noted already, to a number of
languages in western Indonesian known to have been strongly influenced by
Malay. Thus, I fully sympathize with Blust's reluctance to posit an additional
vowel phoneme solely on the basis of one segment of one subgroup of Austrone
sian.
in mind,
Bear
that Dyen's
however,
based on different
subgrouping
his methodology,
requires
of both
another
and
*-ey
is
*-ay
coupled with
and these assumptions,

proto-phoneme.
*hahi in
be the nonexistent
*qaqay
in
it
is
*kaki
in
but
and
PC,
Malayic
*kakay
*hahay
how both Chamic and Malay developed the same aberrant reflex
Malayic
in PC. And,
an
assumptions
reconstructing
the PMP
Second,
and the nonexistent
remains
reconstruction
unsolved
form
problem.
should
whatever
however,
Again,
the
ultimate
source
of
share this oddity supports

this irregularity, the fact the Chamic and Malayic
rather than refutes the arguments for the unity of Chamic and Malayic as a super
ordinate subgroup (both subgroups share reflexes of this unexpected lexical inno?
(as the
vation) and for the patterning of the Kelantanese
languages with Malayic
*-ay
The Cham
The MK
directly
the
remaining
and Chamic
significant,
area
of
component
long-term
becomes
form
-i, rather
for wet-rice
-ay).
remaining
on the mainland
speakers who
changes
in
stayed on the mainland

their
patterns
related to the Vietnamese-dominated
suitable
than
irrigation.
The
of
continued
language
contact,
to undergo
changes
inhabitation of the coastal plain, an
Vietnamese
movement
into
these
areas
was
abrupt in times of war and gradual in times of peace, but incessant. In

response to this incursion, some segments of the Chamic and MK groups inhabit?
ing the coastal plains adapted, assimilated, and in some cases even shifted to
culture in
although not without changing the incoming Vietnamese
while other segments of the Chamic and MK peoples reacted by
into
the hills, pushed not so much by the march of an army but more by
retreating
the almost imperceptible conversion of the coastal plains, by the Vietnamese
Vietnamese,
the process,
farmers, slowly but steadily, one field at a time.

Steadily over time, the once coastal-dwelling
speakers of Cham lan?

or
tomove further into the
and
have
assimilated
continued
either
guages
adapted
networks
between
extensive
Chamic speak?
The
interaction
highlands.
previous
ers were often weakened and sometimes even broken, with new patterns of affili?
ation evolving,
but now in some cases not with fellow Chamic
speakers, but with
and Historical
Geographical
27
Setting
become a Chamic-speaking
speakers. The Haroi, who have effectively
Hr? are an excellent example of this
branch of the otherwise Bahnaric-speaking
sort of cultural realignment. Eventually, as a consequence of the breaking of old
MK
alliances
and the substitution of new linguistic and cultural realignments,

among many of the Chamic languages became more marked,
ferentiation
the dif?
becom?
ing the Rade,
the Jarai, the Chru, the Haroi, and themodern Cham.

the Chamic and MK speakers that remained in the coastal
Typically
assimilated
into the Vietnamese
culture,
eventually
wrote:
Southeast
Asia.
As
throughout
Keyes (1995:19)
plain
a process
found
a growing distinction between the hill and lowland peoples,

did not develop between the two. Rather, throughout
boundaries
sharp
most of Southeast Asia, hill peoples were incorporated into social sys?
tems dominated by the lowland peoples.
Despite
And, although some of theMK and Chamic speakers retreated into the
hills, much of the disappearance of the Chamic MK speakers along the coastal
plain must be attributed not to their being killed or even displaced but to their
absorption into the emerging lowland civilization. Again, Keyes (1995:183-184)
writes:
Once
the various
territories had been conquered, Vietnamese migrants

into and settle these areas. Here, they often intermarried
with Chams and Khmer, and, even when they did not, they were exposed
to the different social and cultural patterns of these Indianized peoples.
would move
tended to result in some compromising of the dominant

tradition, at least among the peasantry. Many of the
cultural differences between northern and southern Vietnamese
can be
These
contacts
Chinese-derived
traced
to such
compromises.
In many cases, such assimilation

led to the total absorption of peoples into the
even
dominant Vietnamese
in cases where the group has maintained a
but
culture,
distinct linguistic identity, there have not only been enormous cultural changes
but the languages have been slowly restructured in the direction of Vietnamese,
as exemplified by the dramatic evolution of tones in Phan Rang Cham.
The Chamic
The modern
Chamic
MK
and MK
languages
in Vietnam
distribution
languages
of Chamic and MK languages is shown in Figure 4. The

are found in central and southern Vietnam interspersed with
languages.
Figure 4: The modem
X"
? Bru
distribution
and Mon-Khmer
ofChamic
luang Tri
South
China
Hue
\,
iJPacolN^
V^Phuqng\_
3
r
Katu
Jeh
Duan
Sea
*Da
Nang
Takua
^'"7'Cua
Quang
Ngai
Katua
x"
^f
im
2e-C /"
' Hre
Sedang V-'\
:
&
?%-vReneac?
^
D\.->-^
-v*
Kayong
Todrah
Renga?^-v
Halang
:
Kontum
>
tMonom
Qui
Nhon
Tuy
vHoa
N. Roglai
E. Mnong
*
f
..?/
_
? s'
Stieng
/
\
Nha
Mnong
^A-^C.
Trang
^
Dalat
"r
^^
Koho
?^/;\*TPhan
Y
Saigon
?
) Chrau
i Jro
Cham
Geographical
and Historical
Setting
29
languages are now highlands languages spoken by

from
the coastal plains. For those in the southern Viet?
living away
namese highlands, the major linguistic contact has been with MK
highlands lan?
Most
of the Chamic
hill tribesmen
guages, and this mutual contact has resulted in what Solnit termed "the southern
Vietnamese
highlands subregion", a linguistic area defined by certain common
characteristics
("Reflexes of PC glottalized obstruents" on page 91).
linguistic
What
is equally evident from Figure 4 (from Gregerson and Thomas
(1980:xi)) is that there are no longer any Chamic speakers living in the north, but
as already mentioned,
the Acehnese were at one point the most northerly Chamic
the descendants of then most
speakers, and upon the departure of the Acehnese,
northerly Chamic speakers live on in the modern Tsat speakers of Hainan and the
modern Northern Roglai speakers found in the south (page 224).
Finally, of course, there is the modern diaspora of Chamic and other
speakers from Vietnam, as a result of the conflicts of the Vietnam war, and from
Cambodia, as a result of the Khmer Rouge genocidal killings. These modern ref?
ugees are found scattered throughout the world, inMalaysia, Hong Kong, France,
?
even occasionally
Australia, and so on
showing up in places as distant as the
central valley of California and the large and growing Chamic (Dega) community
in central North Carolina (David Thomas, p.c.).
of
Classification
Chamic
the
Languages
and the determina?

languages as Austronesian
tion of their relationship to the rest of Austronesian
has clear implications for our
this part of Southeast Asia.
of
the
the
of
and
history
understanding
prehistory
The classification
John
of the Chamic
one
Crawfurd,
of
the earliest
to examine
scholars
Cham
its Aus?
recognized
tronesian character, describing Cham as the "Malay of Champa" as early as 1822.

However,
Chamic,
the
by
had
become
turn
of
the
century
controversial,
the
classification
a controversy
that
of
resulted
Cham,
from
and,
a failure
thus,
to
genetic and typological criteria for classification.

distinguish
The scholars following Crawfurd were inclined either to be confused by
theMK-like
typology of Chamic or to view it as an "intermediate link" between
between
languages of the islands.

languages of the mainland and theAustronesian
in 1889, Etienne Aymonier, who believed, along with many of his contem?
migrated to the islands from this part of the
poraries, that theMalayo-Polynesians
wrote
that Cham formed a kind of intermediary link
Southeast Asian mainland,
theMK
Thus,
between Khmer
and Malay
(Aymonier
1889:5-6
(translation mine)):
language as well as the majority of the tribes in

the forests of southern Indo-China, the Cham language represents a
mainland group related to the Polynesian family of languages found in
...that like the Khmer
and of the Indian Ocean; knowledge of it will

help balance the study of these languages; one can suggest that Cham
serves as an intermediate link between Khmer and Malay, for example.
all the islands of Pacific
30
of the Chamic
Classification
31
Languages
typological features in Chamic coupled with the presence of some

borrowed material also accounts for Schmidt's (1906) description of the Chamic
and for Thomas Sebeok's mis?
languages as "Austroasiatische Mischsprache"
The MK-like
this earlier
guided 1942 claim that these languages are Austroasiatic.
Despite
a modern
are
Chamic
Austronesian
from
the
confusion,
languages
indisputably
influ?
is
under
what
and
what
of
interest
is
how
and
perspective,
determining
ences these languages came to acquire the often typologically MK-like
systems which they possess today.
forms and
THE PLACE OF CHAMIC WITHIN AUSTRONESIAN

The
existence
family was certainly recognized early by

in 1708. In 1852 its geographical extent was sketched nicely
of the Austronesian
Hadrianus Relandus
by Crawfurd, who described what he called the "Malayan"

lowing terms (1852:cxxxiii):
languages
in the fol?
A certain connexion, of more or less extent, is well ascertained to exist

to Easter
between most of the languages which prevail from Madagascar
Island in the Pacific, and from Formosa, on the coast of China, to New
Zealand. It exists, then, over two hundred degrees of longitude and sev?
enty of latitude, or over a fifth part of the surface of this earth.
Crawfurd
continues
with
his
geographical
survey
of
languages,
adding
to his
Malayan
... the innumerable islands of the Indian

Archipelago,
New Guinea ?
of the great group of the Philippines ?
the North and South Pacific ?
and of Madagascar.
In 1852,
the Austronesian
language
in the world.
family was
the most
from Sumatra to
of the islands of
geographically
dis?
persed language family

Placing Chamic more precisely within the Austronesian
family requires
a family tree, along with its subgroups, but none of the subgrouping assumptions
about PAn are totally without their distractors. Nonetheless,
while there are dif?
ferences in details, except for Dyen (1965; 1995), who has his own family tree,
and two of Dyen's students (Wolff 1991, 1995; Tsuchida 1982), who regard For
1. The modern
recognized
perspective probably begins with Pittman (1959), who clearly

nature of Jarai, one of the Chamic languages.
the Austronesian
Chapter
32
languages, modern scholars

languages as branching off from Philippine
in
Formosa
in
homeland
the
Austronesian
and, in one sense or
agree
placing
another, in their recognition that the Austronesian
languages of Formosa repre?
sent a higher branch on the family tree than the remaining Malayo-Polynesian
mosan
in Tryon (1995)). In Blust's view (1977 and else?

languages (cf. the discussion
are
four primary subgroups of An, three of them For
where), for example, there
and one non-Formosan
and Paiwanic)
Tsouic,
(MP), the
Formosan
the
that
in
Others
5.
argue
languages
Figure
configuration presented
consist of a single branch; still others argue about the subgrouping of the Formo?
mosan
(Atayalic,
san languages
land
(cf. Starosta
scholars place the home?
however, most
1995). Again,
on Formosa.
such a homeland
also exist for a Formosan
arguments
Non-linguistic
insular Austronesian
homeland
for the
(1978, 1985, 1991) has argued for

languages. Bellwood
on archaeological
grounds. The dating of the archaeological
that Neolithic
sites on the east coast of Formosa are part of the
record suggests
Yuan-shan culture, which Bellwood
associates with
later Philippine
culture complex
and Indone?
on Formosa
Bellwood
places the Yuan-shan
or
some
so
in
the Philip?
dates
than
the
earlier
800
earliest
years
(4300 BC)
by
or
far
than
for
Indonesia
the
and
earlier
dates
(c. 2500
later)
pines (c. 3500 BC
(1870 BC) (dates from Blust 1995:592).
BC) and Melanesia
sian settlement.
to regard
it is misleading
However,
Austronesian
themselves
languages
as having originated in Formosa. The Formosan languages represent a starting

pointing for the migration into the Pacific, but there is also linguistic evidence for
an
earlier
Austronesian
has
language
Austronesian
on
presence
left modern
in
loans
the mainland.
the earlier
descendents,
the
While
non-Austronesian
no
early
of
the mainland.
Austronesian
languages
Austronesian
presence
has
left
For
example, in at least two branches of Tai-Kadai, there are early, readily identifi?
able Austronesian
borrowings (Thurgood, 1994). The unique correspondence pat?
in
terns of borrowed Austronesian
disyllabic roots are particularly conspicuous
forms of proto-Tai and proto-Kam-Sui.
among the overwhelmingly monosyllabic
For
Sui
example,
*?'a:n2'
proto-Kam-Sui
PAn
*polan
'moon';
*ft'am3
-t-i-f
*danum
*?bl/rien1,
to
corresponds
'(fresh)water';
sian (PWMP) *bujak corresponds

2.
to proto-Tai
corresponds
PAn
and
the
proto-Tai
and
proto-Kam
and
*nl/r9m4,
Proto-Western-Malayo-Polyne
to proto-Tai *7bl/ro:k7 'flower'.
These
are but
Thai scholars often attribute many of these words to Thai contact with Malay
in southern Thailand, but for at least these forms the reconstructions date
back to proto-Tai and the forms are found even in the Tai dialects spoken in
southwestern China, the area of the Tai homeland. The dating of proto-Tai
and the distribution of the forms makes it clear that these forms were bor?
rowed long before the earliest contact with Malay
in southern Thailand.
of the Chamic
Classification
33
Languages
a few of many
Kam-Sui
early Austronesian words borrowed early into proto-Tai and proto

(see Thurgood 1994 for details and more examples).
These forms are significant because the Tai-Kadai reconstructions and
show that these words represent, not recent borrowings
languages, but ancient borrowings into proto-Tai and
sources which make it possible to pro?
from early Austronesian
the subgrouping evidence

into the Tai and Kam-Sui
proto-Kam-Sui
vide the approximate
Kam-Sui
time and place of the borrowings.

reconstruct
borrowings
to proto-Tai
and
the Tai and the
Both
and
proto-Kam-Sui
thus
pre?
date the breakup of these proto-languages. An examination of the geographical

distribution of the Tai languages makes it clear that the area where the proto-Tai
began to break up must be somewhere in the Guizhou area (if not even further to
the north) and since the area where the proto-Kam-Sui began to break up must be
in the Guizhou and Guangxi area, it follows that the early Austronesian
contact,
which predates the breakup of proto-Tai and proto-Kam-Sui, must also have
as far north as Guizhou. The proto-Tai and the
at the very least?
occurred ?
homelands
proto-Kam-Sui
place these early Austronesian
speakers somewhere
south of the Yangtze
discussion).
What
some 8,000 years ago or so (see Thurgood 1994 for a fuller

this does is to establish the existence of early Austronesian
speakers presumably along the south bank of the upper reaches of theYangtze?
early Austronesian
speakers whose descendants later left the mainland to become
Austronesian
arrival in Indonesia and

speakers. Given that the Austronesian
was considerably later in prehistory than estimated dates for Formosa
and the northern Philippines,
it also makes a departure point somewhere across
from Formosa and the Philippines the most reasonable place to begin the journey
Melanesia
out into the Pacific.

In any case, despite differences of opinion about some of the higher
level subgrouping details, the place of Malayo-Chamic
within this configuration
does not itself seem controversial, nor do any of the disagreements
about the
higher-level
subgroupings
reached about Chamic.
seem
to have
any
for
consequences
It at least seems evident
that Chamic
the
conclusions
subgroups neither
with one the Formosan
branches nor with the Central-Eastern Malayo-Polynesian

a part of the languages usu?
subgroup, leaving Malayo-Chamic
uncontroversially
as Blust has pointed out
termed
Western
However,
ally
Malayo-Polynesian.
(p.c.), it is important to realize that PWMP is not a well-defined

subgroup defined
own
set
the
PWMP is
existence
of
its
of
innovations;
instead,
by
independent
not
those
in
Central-Eastern
included
basically
Malayo-Polynesian
languages
that is, PWMP is a cover term for those PMP languages that
Malayo-Polynesian,
fall outside of Central-Malayo-Polynesian.
Thus, the split of PMP is not binary;
a
it
is
is
that
CEMP
indeed,
quite probable
subgroup within aMP dialect contin?
uum, with Malayo-Chamic
simply being a subgroup within
the continuum.
Chapter
34
language subgroup
Following a convention found in Ross (1995b:263),
are
on
shared
innovations
established
the
bases
of
distinguished orthograph
ings
ically from collections of languages placed together for other reasons. In Figure
5, subgroupings established on the bases of shared innovations are in boldface,
shared ancestor (Ross
while language groupings that do not have an exclusively
are
"Formosan
in
italics.
Thus
the
italicized
label
1995b:263)
languages" indi?
cates a collection
of languages descended (along with PMP) from PAn. However,

is not being made that there was a single, unified "Proto-Formosan"
the claim
from which
Polynesian
the Formosan languages descended. Moving down the tree, Malayo

is itself defined by a set of shared innovations. In turn, it consists of
an
Central-Eastern-Malayo-Polynesian,
isWestern Malayo-Polynesian,
plement
part
the
of
innovation-defined
innovation-defined
subgroup;
its
com?
languages not
those Malayo-Polynesian
Central-Eastern-Malayo-Polynesian.
It is worth bearing inmind that the historical subgrouping of languages,

when properly done, as these subgroupings are, is done not on the basis of the
geographical distribution of the languages, but on the basis of shared historical
(Blust 1995b, Ross 1992). The Austronesian
family tree above is
based on such shared historical innovations; the fact that the family tree has such
striking correlations with geography is because, to a large degree, the current lin?
innovations
guistic distribution
MALAYO-CHAMIC
still reflects the older migration

AND BROADER
patterns fairly accurately.
AFFILIATIONS
Within Western Malayo-Polynesian,

the Chamic languages are part of what Blust
terms
the
(1992a)
Malayo-Chamic
subgroup, a subgroup that pairs the Chamic
Acehnese
(cf. "Acehnese, a Chamic language" on page 47)
languages including
with Adelaar's
(1988, 1992) Malayic
languages, a cluster of languages which
includes Malay. The existence of a special relationship between both the main?
land Chamic languages and Acehnese and, on a higher level, between the Chamic
and theMalayic
languages has long been argued for in the literature. The recon?
structions
found
in this work
serve to confirm what
earlier scholars have
long
maintained.
(1985), and James Collins

(1975), Blust (1981), Adelaar
(1991) have all argued for a special connection between the Chamic and the
Malayic
languages. Collins (1992:110), however, adds a word of caution, writing
Marrison
that
without
"comparisons
full
are
of
reconstructions
superficial".
Now,
PC,
with
Proto-Malayic,
this
reconstruction
and
Proto-Acehnese,
of
PC,
with
Ade?
it is
laar's proto-Malayic,
and Durie's partial reconstruction of proto-Acehnese,
are borrowed,
far more evident that the uniquely Malay elements in Acehnese
while
the uniquely Chamic
elements
are inherited.
Figure 5: P roto-Austronesian family tree

(modified from Blust 1977; Ross 1995b)
Austronesian
(PAn)
Formosan
Malayo
languages
Polynesian
(PMP)
Atayalic
Tsouic
i
Central
Western
Atayal
Eastern
Malayo
Polynesian
SHWNG
I
Sundanese,
Paiwan
Malayo
Malayo
Madurese,
Tsou
Eastern
Polynesian
Javanese,
Polynesian
Central
Chamic,
Malayo
Polynesian
Malayic,
Paiwanic
B im?nese,
ai,
Manggar
Ngadha,
Rotinese
Maya,
Tetum,
Ilocano,
Alune,
Tagalog
Asilulu,
Buli,
Gim?n,
I
Motu,
Lakalai,
Roviana,
Numfor,
Sa'a,
Dusner,
Trukese,
Serui-Laut,
Wandamen,
Buruese
Oceanic
Waropen
Iaai,
Fijian,
Hawaiian
Vietnam,
N.
Sumatra,
Malaysia,
Sundaland,
Sulawesi,
Lesser
Sundas,
southern
central
Moluccas
and
Southern
Halmahera,
Oceania
Western
New Guinea
Madagascar,
Borneo,
Philippines
36
Chapter
Figure 6: The Malayo-Chamic
languages
Malayo-Chamic
Proto-Chamic
Malayic
(Adelaar 1992)
Proto-
(PC)
Iban Other
/
Coastal
/Highlands
/
Chamic
Ma!ay
Chamic
Northern
Cham
. Western
/\
Northern
Acehnese
Rade
Jarai Chru Roglai
Tsat
Haroi
Phan Rang
Cham
Cham
language that Blust suggests might fall in a wider subgroup that includes
is Sundanese (1992a:44). However, Moken definitely falls out?
Malayo-Chamic
side of this group, as most likely do Rejang and Maloh.3
Another
The
innovated
numerals
One piece of evidence forMalayo-Chamic

and, potentially, for its broader affilia?
as Nothofer has quite cor?
tions are a series of innovated numerals. Although,
are
not
numerals
alone
sufficient
the
for subgrouping,
out,
rectly pointed
certainly
the
innovated
numerals
tionably not reconstructable
back
for
seven,
eight,
and
to PMP, do provide
nine,
which
highly
are
unques?
suggestive
sub
3. Collins (personal communication

cited in Blust 1992a:74, fn. 6) appears to be
are
immediate relatives of Acehnese
alone in his opinion that the most
not
Chamic.
Malayic,
Classification
grouping
of the Chamic
in Dyen
(cf. discussions
inNothofer (1985)).
evidence
(1975), as emended
37
Languages
(1965), Blust
(1981), and Nothofer
inMalayo-Chamic
Table 3: Innovated numerals
PMP
PC
Malay
*esa;
*sa
satu
satu
'one7
*duha
*dua
dua
dua
'two'
*telu
*klow
(tiga)
(tiga)
'three'
*epat
*pa:t
empat
'four7
Iban
*lima
*lima
?mpat
lima
*enem
*nam
?nam
*pitu
*tujuh
tujuh
tujoh
'seven'
*walu
*dua-lapan
> *lapan
d?lapan
delapan
'eight'
<
*siwa
lapan
'nine'
'five'
'six'
lapan
selapan
'nine'
*sa-lapan
s?mbilan
*samilan
<'eight'
*puluq
*pluh
*Ratus
*ratus
*ribu
*ribow
-f
puluh
'ten'
ratus
'hundred'
ribu
ribu
'thousand'
that theMalayic
languages, represented inTable 3 by Malay and Iban, share
the innovation of *tiga 'three', reflecting a genetic closeness not shared with the
Chamic languages.
Note
Table 3 compares the numerals of PMP with the numerals of PC, Malay,
and Iban, making the innovations apparent. The most widely distributed innova?
tion is in the number 'seven', which is not only distributed throughout Malayo
but also is found in languages beyond this group, and thus may correlate
with a broader subgroup. Although not as widely distributed as the innovations
Chamic
'eight' and 'nine' occur not just throughout

a
number of other languages, including Maloh
Malayo-Chamic
as
and Rejang. However,
Blust (1992a:44) cautions, emending Blust 1981, the
appearance of these innovated numerals in Rejang and Maloh may be due to bor?
for
'seven',
the innovations
for
but also within
rowing.
38
Chapter
Table 4: The Chamic
PC
numerals
Aceh.
Rade
Jarai
sa
sa;
NR
Tsat
sa
sa33
PR Cham
Haroi
tha
sa;
ha; ho
ha
'two'
*dua
duwa
dua
dua
dua
thua11
thua
twa
'three'
*klow
lhea
tl?u
klau
tlau
kiu33
tl?u
klow
pa?24
pa?
pa?
'four'
pa?
p??
'five'
*lima
limAg
ema
rama
pa:?
lum?
'six'
*nam
nam
n?m
n?m
*tujuh
tujoh
kjuh
tajuh
ptuat
*pa:t
'seven'
'eight'
ma1 33
lamia
lami
n?m
nain?3
n?m
n?m
tijuh
su55
cas?h
tac?h
capan
*dua-lapan
>
*lapan
lapan
<'nine'
'eight'
*dua-ambil-an
sap?n
talip?n
pain?42
?
lap?t
rapan
'eight'
lapan
sapan
t?mpan
?
'nine'
*sa-lapan
'nine'
*samilan
samilan
'nine'
<'eight'
'ten
^pluh
Notes:
For
*dua
ploh
'nine',
and
the
that
*dua-lapan,
'eight'
thalip?n
salapat
'nine',
the
duap?n
dua
dua
thua11
rapan;
capan
pain?42
lap?n
pluh
pluh
spluh
piu55
apoh
second
N.
Roglai
is, another
*dua
form
-pan
the Tsat
form
to the original
added
Jarai morpheme
Cham for 'nine' is from Aymonier
and
thua-
has
and Cabaton
unexpected
pluh
are
<
from
In
*dua-lapan.
length.
The
PR
(1906).
The
innovated forms for 'seven', 'eight', and 'nine' in Table 3 make it

clear
that PC subgroups with theMalayic
abundantly
languages.
Table 4 shows the distribution of these innovated numerals within the
Chamic
reflexes ("Acehnese, a
languages. Notice the regularity of the Acehnese
on
discusses
the
of
Acehnese
within PC).
page 47,
language"
membership
In Table 4, there is an interesting indeterminacy in the reconstruction of some of
Chamic
the forms listed in the table as descended

*lapan
cases,
could
it is not
be
the
possible
truncation
to tell. As
of
from PC *lapan. Forms with a reflex of
*dua-lapan
a consequence,
or
be
from
as many
*sa-lapan;
as four
of
in most
the Chamic
languages might have a word for 'eight' that was descended from 'nine'.
This widespread
confusion of the forms for 'eight' and 'nine' both
within
Chamic
double-prefixed
the
(Table 4) and outside of Chamic (Minangkabau, Kerinci),
forms for 'eight' in Northern Roglai and Tsat and for 'nine' in
of the Chamic
Classification
39
Languages
Phan Rang Cham, and the complete inversion of 'eight' and 'nine' in Rade and
Jarai, all indicate that the semantics of the original roots was anything but trans?
parent. It certainly appears as if the Rade and Jarai speakers ignored the subtrac
ordered the
tive meaning
of the root and instead, not totally unreasonably,
numeral beginning with *sa- 'one' (*sa-lapan) before the numeral beginning
with
*dua
'two'
(*dua-lapan).
of reflexes of PMP *telu
The existence
and the rest
'three' inAcehnese
rather than the laterMalay and Iban innovation tiga 'three', indicates
and Iban are closer to each other than either of them is to PC as well
of Chamic,
thatMalay
as being another piece of evidence thatAcehnese patterns more closely with the
Chamic than theMalayic
languages (page 47). Within Chamic, the fact that Rade
and Jarai have both reversed the reflexes of 'eight' and 'nine' (one suspects that
suggests that Rade and
they were no longer primarily traders and merchants)
Jarai have been in particularly close contact, a conclusion that is supported by all
sorts
of other
evidence.
innovations
intriguing semantics behind the Malayo-Chamic

'eight', and 'nine' are discussed by Blust (1981:467,
The
of
fn. 5). He writes

'seven',
that the forms for *tujuh 'seven' are usually thought to be from the word for
'index finger', which Blust notes is "the second finger of the second hand in fin?
ger-counting". The form *tujuh itself does not date back as far as PMP, but pre?
cisely where and when itwas innovated is not yet clear. The other innovations for
'eight' and 'nine' involve what Blust terms "subtractives". These forms he relates
to the roots *alap and *ambil
away
(from
ten)'
and
both
'fetch, take', with

and
*sa-alap-an
being
*dua-lap-an
*sa-ambil-an
'one
being
'two taken
taken
away
(from ten)', with an the etymology based on 'one less' also occurring in many
Borneo languages and in South Sulawesi languages (Adelaar, p.c.). In the PR
Cham form tampan 'eight' (from Blood 1962:11), the -p- is unexpected, but the
presence
of
the medial
-m-
away
(from ten)', which
above.
The
Acehnese
form
suggests
it comes
that
would make
sikuruiQi)
'nine'
from
the root *kura:rj 'less', appears to have originated

identical semantics.
The Malayo-Chamic
from
it semantically
taken
*dua-ambil-an
'two
parallel
to *dua-lap-an
*sa-kura:rj,
although
independently
involving
but from almost
subgroup
The Malayo-Chamic
subgroup is justified by a set of shared innovations. In addi?
tion to the innovated numerals, which may not be confined just to the Malayo
Chamic languages, there are several innovations largely but apparently not exclu?
sively restricted to theMalayo-Chamic
subgroup. Blust
(1992a:34-44,
emending
40
Chapter
1981) discusses the following: (1) the shift of PMP *q > *h, although the
PMP *q > h is somewhat wider than just Malayo-Chamic,
of
scope
including as it
does Balinese, Javanese, Sundanese, and the Batak languages, (2) the merger of
PMP *R and *r but with the merged reflex remaining at the same time distinct
Blust
from other proto-phonemes,

and (3) the loss of PMP *w in initial position. The
of
*R
*r
and
also
merger
happened in Batak and Balinese, and so did the loss of
*w (although some Batak languages have a residue of it). The lexically-specific
innovation Blust cites is the introduction of an initial *h in PMP *waRi (*waRi >
*ari > *hari), an innovation that follows the loss of PMP *w in initial position.
Two observations need to be made about these shared innovations. First,
the PC reconstructions reflect all of these changes; thus, there is no question that
the parameters of the subgroup. Second, while there is some
about precisely where to draw the lines at the periphery of theMalayo
PC falls within
question
Chamic subgroup,
define
Within
the linguistic distribution
of the innovations mentioned
here
a core.
the Chamic
languages:
A sketch
have noted with reference to other Austronesian

(1995:39-74)
are
two
that
there
distinct
patterns of shared innovations, each pointing
languages
to a different historical origin. In one pattern, which reflects genetic inheritance
Pawley
and Ross
(1995:50), "all member languages of a subgroup exclusively share a common set

of innovations", from which
it can be inferred that "the subgroup has been
a
formed because
community speaking a single language has become separated
geographically
4.
and/or socially
into two or more
communities
after separation,
The history of Chamic studies shows some promising early work but then
scholars appear to have turned to Thai and Khmer studies, leaving Chamic
studies largely dormant until a mild resurgence of interest in the last several
decades. The earliest wordlist is that in Crawfurd (1822, discussed by Craw?
furd himself 1852). The history of much of the early research is set out by
Lafont (1987b; 1994b), who notes that much of the early work on Chamic
was linguistic. In addition to Crawfurd's list, there was a two-page list pub?
lished in 1868 by A. Bastian, who two years later published a paper on the
language and the origins of the Cham people. Several years later in 1875,
Lafont continues, A. Morice published the first Cham glossary, containing
some 800 words, as part of a work on the Cham and Stieng (aMK language)
languages. Then, in 1880 K.F. Holle published a replica of the Cham alpha?
bet, following up this publication with a commentary on it in 1882.
Classification
of the Chamic
in the speech of each of the new communities
[and] changes have occurred

what
was
one
has
language
41
Languages
become
two
until
or more."
In the pattern reflecting areal diffusion of changes

uum (1995:50), "innovations form an overlapping pattern,
languages A, B, and C reflect one bunch of innovations,
another bunch, languages D, E, F, and G yet another, and
along a dialect contin?

such that, for example,
languages C, D, and E
languages G and H still

a different bunch of innovations", from which can be inferred that "the languages
of the subgroup once formed a network of related dialects. During this phase,
innovations occurred at various places in the network, spreading from their dia?
lect of origin into neighbouring dialects, but without affecting the whole network.
Over time the dialects have diverged until they have become mutually unintelligi?
ble for practical purposes, but they continue to reflect the innovation pattern of
the former" dialect chain.
In the case of the Chamic
languages both genetic inheritance and areal

diffusion have produced their patterns of shared innovations, but fortunately in
most although not all cases it is possible to distinguish one from the other and
thus to infer information about which dialects developed from an earlier common
source and also to infer much
about the earlier patterns of contact and bilingual
ism.
initial period of contact with MK speakers, undoubtedly very early

in the establishment of Champa. It involved intense contact including consider?
The
able bilingualism and led to the dramatic restructuring found in PC. After that ini?
seems to have
tial contact-induced
the language of Champa
restructuring,
continued
for
some
time
as a
largely
differentiated
dialect
chain
in coastal
com?
is no evidence
in the linguistic record

along the coast of Vietnam. There
of any marked dialect differentiation until the historical events associated with
the Vietnamese push to the south forced radically new patterns of language con?
tact as the result of the partial destruction of the old language contact patterns.
munities
the Chamic languages con?

Perhaps for as long as the first millennium,
a
stituted
dialect chain that certainly extended along the coast of Vietnam (Blust
1981:32; cf. also Blagden 1929 (summarized inDyen 1971a:202)) and may even
have extended as far south as the east coast of theMalaysian peninsula, although
it is more
features found along the east coast of

likely that the Chamic-like
date from a much later Acehnese
influence. The breakup of the dialect
Malaysia
chain along the coast of Vietnam
into clearly distinct languages occurred after the

the coastal communities was disrupted by the
a
Vietnamese push to the south, push that caused many Chamic speakers tomove
back into the highlands. This retreat into the highlands resulted in a new set of
language networks for almost all Chamic speakers, with some like the Haroi
eventually coming to be part of aMK social network, while others like the Phan
rich interactional network between
42
Chapter
Rang Cham eventually
came to be part of the lowlands society dominated
by the
Vietnamese.
Figure
7: The PC dialect
chain and itsmodern
reflexes
Proto-Chamic
(PC)
\.
Coastal
j Highlands
Chamic
/
Chamic
Northern
Cham
Acehnese
Rade
Northern
Western
Jarai Chru Roglai
Tsat Haroi
Cham
Phan Rang
Cham
The basic relationships among the Chamic languages are shown in Fig?
shows groupings due to common genetic inheritance and linguistic
to bilingualism. Often the data reflects both types of influence:
due
convergence
inmany cases, languages reflect genetically-inherited
similarities as well as simi?
larities due to areal contact. The interplay between these two factors, recognized
ure 7, which
from the beginning of the comparative method and reflected in the well-known
between the "family tree" model and the "wave" theory, is found
distinction
throughout the Chamic data.

The genetic and the areal relationships between the various modern lan?
are
illustrated in Figure 7. Although future research may show that Ace?
guages
is particularly close to one or another of the mainland Chamic languages,
the evidence thus far suggests that the Acehnese
left Champa before any strongly
marked distinction developed among the Chamic languages. It appears that at the
time of their departure, the Acehnese were the most northerly of the Chamic
hnese
groups,
covering
an
area
now
populated
by,
among
others,
the modern
Katuic
speakers. Only subsequent to the Acehnese

departure did the remaining
two
areal
into
the Highlands Chamic
split
major
configurations:
guages
lan?
Ian
of the Chamic
Classification
43
Languages
is shown
guages and the Coastal Chamic languages. Thus, in Figure 7 Acehnese
as separated from the other languages. Itwas undoubtedly only after 1000 AD or
so that a distinction of any significance existed between these two areas, with the
sharp separation of the Chamic speakers into Highlands Chamic and Coastal
Chamic developing
largely after the Vietnamese began moving south. In Figure
two
these
7,
linguistic areas are indicated by dotted lines. In the case of Jarai and
Rade, where
genetic
for mutual
the evidence
subgrouping
is clear but the evidence
influence
is not, only the influence of contact
is indicated
for
(by dotted
lines).
these areal configurations are subgrouping relationships (shown
by solid lines): specifically, the Coastal Chamic languages look to have once been
a single dialect before splitting up into Haroi, Western Cham, and Phan Rang
Cham. Similarly, within the Highlands Chamic area at least the Chru, Northern
Roglai, and Tsat were a single dialect before splitting up. However, the Rade and
Within
probably never a genetic subgroup but instead were simply the

of the original dialect chain, which were eventually forced up
into the highlands. Naturally, all these languages also participated in various pat?
the Jarai were
middle members
terns of areally-induced changes.

The subgrouping relationships merit a special comment. Tsat and North?
ern Roglai represent a Northern Cham dialect that split into two under the impe?
capture of the northern capital at Indrapura. The
and Phan Rang Cham represents another dialect
theVietnamese capture of the southern capital. Their
tus provided by the Vietnamese

division of Cham intoWestern
this time following
division,
early
relationship
with
Haroi
is also
apparently
subgrouping
one,
but
the
evi?
dence for this is not as clear because
the relationships are obscured by the fact

later go on to participate in two additional areal configurations:
first interacted with the Highlands Chamic group before becoming
that Haroi would

the Haroi
socially and linguistically
The relationship
realigned with
of Tsat and Northern
theMK-speaking
Hr?.
Roglai
and the syntax

Despite the radical differences in the phonology, the morphology,
of the modern languages, the subgrouping relationship between Tsat of Hainan
island and Northern Roglai of theVietnamese highlands is obvious from the com?
parative evidence. These two languages share two quite distinctive, quite unusual
innovations, either one of which would by itself suggest a subgrouping relation?
ship: Tsat and Northern Roglai share the innovation of preploded final nasals and
the shared loss of PC *-s in PC *-a:s. As
guages probably constituted
late as around 1000 AD,
these two lan?
a single Northern Cham dialect.
44
Chapter
Northern Roglai and Chru seem to have a special relationship. First,

there are various shared similarities among the nasalized vowels. Second, there is
the simple note inGrimes (1988:615), which states without further comment that
Southern Roglai is closely related to Chru and Northern Roglai.
Note that Tsat is subgrouped specifically with Northern Roglai, not sim?
ply with Roglai. Cac Gia Roglai (Cobbey 1977) and Cat Gia Roglai5 (Lee 1998)
both share reflexes of earlier preploded final nasals, with the denasalization
ing occurred under the same conditions. Several other dialects of Roglai
hav?
share
reflexes of earlier preploded final nasals as evidenced by the occasional final -p,
-t, or -k from a former nasal final, but Gia Le Roglai is almost entirely lacking in
evidence of preploded finals.6 However, a brief examination of Cobbey's
1977
fiche containing wordlists made it clear that neither his Cac Gia Roglai nor Gia
-a and Lee's Cat Gia
Le Roglai share the sound change PC *-a:s > N. Roglai/Tsat
Tsat
lacks
the
crucial
forms.
the
is
Thus,
Roglai
subgrouping
specifically with N.
to
the
other
dialects
for
I have seen data.
the
exclusion
of
which
Roglai
Roglai
in Tsat are quite unrelated to any of the
subsequent developments
the
mainland
Chamic languages. Instead, despite
among
remaining
The
developments
the obvious parallels with developments
elsewhere in Chamic, the Tsat develop?
ments reflect contact with the tonal, monosyllabic
languages of Hainan: the
movement
from iambic to monosyllabic,
the reduction of the *-l- and *-r- in
onset
to /-i-/,
clusters
appearance
the Tsat
the drastic
reduction
of
final
consonant
of tones are all the product of contact pressures
arrival
contrasts,
and
that developed
the
after
on Hainan.
The relationship
of Haroi with Western
and Phan Rang Cham
The other subgrouping relationship is between the Coastal Chamic languages.

is really no question about the relationship between Western and Phan
as they were the same language until the fall of the southern capital
Cham,
Rang
There
in 1471.
A more interesting question involves the nature of the Haroi relation?
no longer resembles its closest genetic relations as it has changed
Haroi
ships.
under
the influence of Hr?. Specifically, modern Haroi has become a
radically
restructured register system under the influence of Hr?. It has also come under
the influence of contact
evidence
5.
6.
indicates
from the Highlands

languages Rade and Jarai, but the
that, prior to this social and linguistic realignment with Hr?,
It is unclear tome whether we are talking about two languages or one here.
The one exception seems to be the reflex of *dun rnose' ddug?, with its glot
talized final nasal.
Classification
of the Chamic
Languages
45
Haroi's genetic affiliations were with Western Cham dialects. Although Haroi and
Western Cham are now no longer in contact, two interesting changes probably
induced by contact are shared with it. One is the lowering of centering diph?
thongs and triphthongs as part of a set of changes that also affected both Kvoeu
Hor
and Timothy Friberg's Cham and Headley's Kompong Thorn Cham (see
page 133 and page 135). The second change is that Haroi and Headley's
Kompong Thorn Cham (but not Kvoeu-Hor and Timothy Friberg's Western Cham
shared the change of the PC *-arj > *-irj (?8.3.2).

A third innovation, however, most likely reflects an early common inher?
itance from Haroi's origin as a Cham dialect. This change is the early develop?
ment of breathy-voiced
often termed "second register".
register complex,
(1978))
Although
aspects of second register development differ from language to lan?

the basic core of second register development
looks like it occurred early
guage,
in Cham, before Cham broke up into Haroi and Cham, with Cham subsequently
breaking up into Phan Rang andWestern Cham. Thus, the evidence seems to sup?
port what Lee (1966:2-3) noted in passing, "Hroy [Haroi] or Bahnar Cham may
be a dialect of Cham proper."
In a later 1974 paper about advanced-tongue-root
and register in Haroi,
Lee again considered the possibility that both Haroi and Cham developed register
together. However, his assumptions about Chamic subgrouping presented formi?
dable obstacles to such an analysis. Lee had apparently forgotten or abandoned
an earlier subgrouping suggestion made in his thesis where he
suggested that
Haroi and Cham
subgrouped together. Almost in indirect reference to his earlier

Lee
wrote
suggestion,
(1977b:97), "The difficulty in trying to make it a shared
innovation stemming from an era when they were still one language is that
Northern Roglai, which is evidently much closer to Cham than either Cham or
Northern Roglai is to Haroi, shows no evidence that I have been able to observe
of a register system" [italics mine]. However, this comment fails to distinguish
similarity due to inheritance from that due to contact.
In short, Haroi has undergone at least three distinguishable
layers of
areal influence beyond theMK influence common to all PC languages: the early
influence of Coastal Cham (or perhaps, justWestern Cham), the later influence of
and the restructuring influence of Hr?. It is not surprising that
and
(1976) ended up placing Haroi squarely between Highlands
Coastal Chamic, as the Haroi have related to and been influenced linguistically at
different times by both groups.
Highlands
Burnham
Chamic,
The relationship
of Rade
and Jarai
The
between
Rade
similarities
among other similarities
and Jarai appear primarily contact-induced:

the change of PC *-a:s > -aih, PC *-us > -uih
share
they
Chapter
46
and *rVl- > hi- as well
as a number of other minor
similarities, including the

and
interesting switching of 'eight' and 'nine', noted already in "Malayo-Chamic
broader affiliations" on page 34. However, it is also evident from the large num?
in the two thatmuch of the similarity between them, aside from
shared retentions may be contact induced.
Rade appears to constitute the end point in a Chamic dialect chain that
was only broken up after the Vietnamese
occupation of the south. More specifi?
not
the
PC
but
retains
Jarai,
cally, Rade,
*-sy < PMP *-i that became *-aj
throughout the rest of the mainland Chamic languages. Presumably, a dialect
ber of differences
and this change

after Acehnese
speakers left the mainland,
to
have
all
which
with
the
but
Rade,
appears
departure of theAce?
spread through
hnese became the most geographically
of
the
isolated
languages in the chain,
chain still existed
one
and,
was
assumes,
the most
likewise
when
isolated
the
change
spread
through the other remaining mainland languages. Rade has reduced its inventory
of presyllable onsets much more radically than Jarai or any of the other Chamic
also suggest that some the similarities between
languages. Other distinctions
these two languages are either independent developments or the result of contact.
A second change that appears to have swept through the dialect chain
that existed until roughly 1000 AD is the loss of the nasal component of medial
clusters consisting
still
clusters
of stops preceded
appeared
in Chamic
after
languages
still found
in modern Acehnese,
from all the mainland Chamic
nasals. Clearly
the breakup
of PC
such medial
as
these
are
long since totally disappeared

1822
For
instance, in Crawfurd's
languages.7
as
are
recent
the only forms with such medial clusters
forms identifiable
wordlist
loans.
by homorganic
The
native
The Highlands
clusters
had
versus Coastal
but these have
long
since
Chamic
lost
their
nasal
component.
distinction
The next major development

is the split of the mainland Chamic languages into a
Highlands Chamic region and a Coastal Chamic region. The distinction between
the Highlands Chamic area and the Coastal Chamic area could only have come
as a distinct
linguistic area after the fall of the northern capital of

Champa in 982, which drove Chamic speakers back from their coastal communi?
ties and into the highlands. And, in a sense, the changes characteristic of the
into existence
Chamic region probably emerged only much later after the fall of the
southern capital in 1471, a date thatmarks the beginning of the emergence of new
Coastal
7.
In light of this, it is quite surprising that there is an apparent lack of evidence

in early Cham written records for the nasal component of these clusters. This
lack, however, may merely reflect my own lack of expertise about inscrip
tional
Cham.
of the Chamic
Classification
patterns of Cham contact with Vietnamese,

structure.
phonological
Solnit
The Highlands
Chamic
to characterize
(1993:110)
In general
47
Languages
a language with a distinctly
different
that led
languages have certain similarities
it as the Southern Vietnam highlands sub-area.
languages have not undergone as radical a
the Highlands
restructuring process as the Coastal Chamic languages, in keeping with the simi?
larities between the structures of the Chamic and theMK languages of the high?
lands. Most of the forms are iambic in structure and have remained that way for
terms,
languages and Tsat, which under contact with more

monosyllabic
languages, have themselves become more monosyllabic.
More specifically
the Highlands Chamic languages have kept the PC
some time, unlike the Coastal
*b and *?. As a common
retention, of course, this does not pro?

vide evidence of any sort of subgrouping relationship, but it does provide an
interesting example of the influence of patterns of bilingualism on what does and
sub
does not change (cf. discussion
in "Solnit's Southern Vietnam Highlands
contrast between
area"
on page
91).
languages are marked by changes in a number of

salient features. Among
these changes is the development of a breathy-voiced
that happened in
register complex from the voiced obstruents, a development
somewhat parallel but somewhat distinct ways in Phan Rang Cham, Western
The Coastal Chamic
Cham, and Haroi, suggesting that the partial influence of patterns of areal contact
rather than just common inheritance is needed to account for the developments. A
second and obviously related change is the subsequent loss of the voicing of the
voiced
obstruents
and the co-occurring

loss of the distinction between voiced
and glottalized
obstruents
is
throughout these languages, which
clearly an areal feature, not a shared innovation, as it happened independently in
the different coastal languages. Not coincidentally,
these changes are in the direc?
obstruents
tion of the structure of Vietnamese,
the dominant
language in the Coastal Chamic
area.
ACEHNESE, A CHAMIC LANGUAGE

It is quite correctly simply assumed in the literature that the mainland Chamic
languages form a subgroup. The questions that exist revolve around the relation?
ship of Acehnese with the mainland Chamic languages. That there is a special
and the Chamic languages was recognized as
relationship between Acehnese
as
1891
by Niemann, who subgrouped Acehnese and Cham together, noting
early
similarities in their verb morphology,
in their treatment of inherited vowels, and
48
Chapter
in various
instances
(1981:523)
writes:
of apparent
lexical agreement. Along
these lines, Cowan
language belongs to theWest-Austronesian

languages
but, although situated in Sumatra, it has its nearest relatives in the
Chamic languages of the South-East Asian continent. This relationship is
The Achehnese
evident on all levels: phonology

(e.g. final stress, large-scale monosyl
of certain final vowels
labism, consonant clusters, diphthongization
deriving originally from long vowels), morphology
use of the substantival infix -?-), and lexicon.
(e.g. lack of suffixes,
in fact, is
Quite early, Niemann suggested amigration of Chams toAceh?which,
what happened.
Over the years various other scholars have agreed with Niemann and
Cowan at least to the extent of recognizing not only that a special subgrouping
and the mainland Chamic languages but
relationship exists between Acehnese
are
also that there
and mainland Chamic that need
similarities between Acehnese
these scholars are Blagden8 (1929), Cowan (1933,
1948, 1974, 1981, 1983, 1988), Shorto (1975, 1977), I.V Collins (1975), Blust
(1981), and Durie (1990a). The literature leaves no question that there are strik?
to be accounted
ing parallels
for. Among
between
Acehnese
and Chamic.
has a special subgrouping relationship

Thus, the claim that Acehnese
with Chamic is not original to this work, and the more specific claim that Ace?
hnese is simply a Chamic language has also already been made in the literature.
In fact, the belief that a special subgrouping relationship exists between Ace?
hnese and the mainland Chamic languages seems to be shared by all but one of
the scholars who have done extensive, detailed comparisons of the languages
involved.
8. Blagden's view, however, was quite distinct from the views of other scholars
cited here. He followed Kern's 1889 hypothesis inwhich Austronesian
speak?
ers originated on the mainland and the Cham represent the remaining rem?
nants of this original group of Austronesian
speakers on the mainland. Thus,
as related to the Chamic speakers, he also
although he viewed the Acehnese
as having originated on the main?
viewed the Cham including the Acehnese
land (specifically somewhere between Kra and Penang). Thus, Blagden's sce?
in
nario is quite at odds with the migrations
supported by the reconstructions
this work, which instead suggest the Chams originated through amigration to
coastal Vietnam
theAcehnese
Borneo and then only subsequent to that did

from there to northern Sumatra (conceivably with stops
from Southwest
migrate
along the way).
Classification
of the Chamic
49
Languages
Only J. Collins (1991), who has noted the parallels between Acehnese
seems to dissent. While
the existence of parallels
and Malay,
recognizing
between Acehnese
and mainland Chamic, Collins argued for the possibility that
Acehnese and Malay subgrouped more closely than either with mainland Chamic
and that the Acehnese
and mainland Chamic parallels were due to parallel but
combined with the mutual but independent influence
independent developments
on
contact
MK
and mainland Chamic. Collins has argued that
of
both Acehnese
are
needed to settle the question. It is easy to see his
detailed reconstructions
it would not be possible to determine
point, as without such reconstructions,
where
However,
reflect contact
the parallels
now
the
reconstructions
and where
and
the
they reflect genetic
accompanying
inheritance.
correspondence
pat?
terns exist: not only do they make it clear thatAcehnese

is a Chamic language but
establish that the very real special connection with Malay, aside from features
also found in Chamic, reflects the long history of contact between Acehnese and
Malay.
The precise nature of the special subgrouping relationship between Ace?

hnese and the mainland Chamic languages was impossible to determine without
to
The specific question
that required more reconstruction
reconstructions.
resolve was whether Acehnese
and mainland Chamic were sister languages or
whether Acehnese was simply another of the Chamic languages. Here, the recon?
structions provide substantiation for the claim that Acehnese
is simply another
was
correct
Chamic language. Thus, Niemann
when
he
(1891)
suggested that
is a Chamic language whose speakers migrated to northern Sumatra, a
Acehnese
position also taken by Cowan (n.d.:l), who specifically states that, before the
migration of the Acehnese
speakers to Sumatra, Acehnese was not distinct from
Chamic.
The evidence presented in various works of Cowan, in Shorto, and Durie

is certainly substantial enough to establish the special relationship of Acehnese
and Chamic. However, for those with doubts, the integration of theAcehnese data
into the reconstruction of Chamic, along with the numerous subgrouping argu?
ments, should provide a conclusive demonstration of the nature of the relation?
is not just
ship. These reconstructions fully substantiate the claim that Acehnese
a
with
Chamic
but
that
it
is
Chamic
subgrouped
language.
None of this, however, is to claim that modern Acehnese
looks like the
does not. In fact, Acehnese often looks more like
languages?it
PC than any of the modern mainland Chamic dialects do. This often striking sim?
ilarity to PC is not at all surprising, however. The Acehnese
speakers left for
Sumatra while all the mainland Chamic languages still looked a lot like PC, pre?
modern mainland
serving many of its PC-like features, while the remaining mainland Chamic lan?
guages were subsequently subjected to even more intense language contact over a
long period of time?at the very least for another six or seven hundred years.
Meanwhile
behind, have inmany
speakers, having left much of theMK contact

cases retained elements lost on the mainland. For example,
the Acehnese
50
Chapter
Acehnese
still preserves the four-way distinction in first syllable vowels inherited

from PMP.9 Elsewhere
in Chamic, this distinction has been reduced either par?
or
tially
drastically, although its earlier existence is hinted at in the data. On the
other hand, like the rest of Chamic, Acehnese has a rich array of second syllable
vowels, including those borrowed into pre-PC from MK sources. Thus, modern
to a stage after the incorporation of MK vowels into PC but
before the four-way distinction had been reduced to a limited three-way distinc?
tion. And, Acehnese's potential for providing a window into the paths of morpho?
Acehnese
is witness
logical and syntactic change is still largely unexplored.

The patterns of innovated numerals in "Malayo-Chamic
and broader
on page 34 provided some evidence

that Chamic was part of a
not
defined
termed
that included both
broader,
yet fully
group,
Malayo-Chamic
the mainland Chamic languages and Acehnese,
but at best this only establishes
affiliations"
that both the Chamic languages and Acehnese were all part of a larger subgroup.
It does not provide evidence that Acehnese
should be subgrouped together with
nor
does
it
is a Chamic lan?
mainland Chamic,
provide evidence that Acehnese
guage.
to provide evidence that establishes

it is now possible
the
However,
a
as
of
PMP.
The
evidence
distinct
of
PC, including Acehnese,
integrity
subgroup
consists of a cluster of shared innovations?shared
innovations found among the
consonants,
rowed
among
from
MK
the vowels,
sources,
and
in the
so on.
shared
Some
of pre-PC
borrowings
of
the
shared
innovations
material
are
bor?
innova?
tions not found anywhere else inAustronesian;

for others, individual parallels are
but certainly not in the particular configuration found here.
found elsewhere
Shared
innovations
among
the
consonants
and mainland Chamic presyllables have precisely the same

(1) Both Acehnese
consonantal onset inventories, including the identical change of the PMP first syl?
lable *n- toM
(Table 11; "Presyllable onsets: retention and reduction" on
page 75ff).
(2) As H. K. J. Cowan (1948), among others, noticed much earlier, the
as
restructuring of PMP disyllabic forms produced the same clusters inAcehnese
can be reconstructed for PC (see the discussion of primary clusters in "Primary
clusters" on page 93).
9.
In place of MK contact, however, there has been intense contact with Malay,
as J. Collins makes clear. One consequence of this Malay contact is that in a
number of ways Acehnese now looks somewhat more Malay-like
than PC or
the Chamic
languages
remaining on the mainland,
particularly
in the lexicon.
of the Chamic
Classification
In addition
51
Languages
in the primary clusters, there is also the

*-h- as the second element of the
to the reduction
of onset clusters with medial
shared presence
cluster (see "The loss of the vowel before medial -h-" on page 63 and "Voiced
and voiceless
'aspirated' consonants" on page 84).
reduce the
Not only do the mainland Chamic languages and Acehnese
same PMP disyllables
maintain disyllabicity
to monosyllables
with cluster onsets, both languages also
in the same forms (see the discussion of secondary reduc?
clusters" on page 94).
tion "Secondary
(3) Both Acehnese
*kl- and *tl- to *kl- and *gl

then goes a step further merging both to *kl
and mainland
and *dl- to *gl-, although Acehnese

> lh- inDurie's Acehnese.
PC merge
and mainland PC, there are a handful of

(4) Finally, in both Acehnese
native PMP words which developed glottalized obstruents (see "The origins of
forms with glottalized
glottalized obstruents" on page 86). The monosyllabic
the disyl?
obstruents on the mainland now have glottal stop reflexes inAcehnese;
labic forms have voiced obstruent reflexes inAcehnese. This shared innovation of
obstruents, otherwise
strong subgrouping evidence.
glottalized
Shared
innovations
among
unattested
inAustronesian,
constitutes
by itself
the vowels
of the high vowels interacts with the loss of final *-r

(1) The diphthongization
("The splits in the PMP high vowels *i and *u" on page 114). In syllable-final
position, PAn *-u (possibly early PC *-u) diphthongizes,
becoming PC *-ow.
However, with PAn *-ur forms, which have PC *-u reflex, the Chamic reflexes do
not reflect any diphthongization,
not yet lost the final *-r when
making it clear that the forms with final *-ur had

the PMP *-u in open syllables went to PC final
*-ow. Instead, the PMP *-ur lost the final -r only after this change, and thus did
not diphthongize. Acehnese
shares not only the diphthongization
of the *-u but
also shares the constraint
that the loss of final /-r/ must
of open final *-u, resulting
diphthongization
the Chamic languages.
be sequenced after the

in undiphthongized final open /-u/ in
The subgrouping value of the diphthongization, which occurs indepen?

in Austronesian,
elsewhere
is increased greatly by the presence of this
dently
shared sequencing constraint. It is worth noting that, while the typological paral?
lelism is interesting, it is the occurrence of these changes
thatmakes it of interest for subgrouping.
in precisely
the same
words
(2) As Durie (1990a) noticed, Acehnese and PC share regular reflexes of

borrowed from MK into PC. More specifically, the PC vowels borrowed
vowels
from MK,
*s,
*o,
*a,
*ia,
*ua,
and
*iaw,
all
turn up
in the
same
borrowed
words
Chapter
52
inAcehnese
and with regular reflexes (see the discussion of borrowed main sylla?
"The borrowed PC main syllable vowels" on page 126). In addition,
for most of these vowels, the vowel itself is not completely restricted to borrowed
ble vowels
words, but instead it also turns up in one or two native PMP words as well and, in
these cases, the Acehnese word also has the innovated vowel in precisely the
same
native
words.
Austronesian
that only a shared common inheritance from PC can account for

this sort of configuration. Aside from the question of how and where Acehnese
would have access to the borrowed MK vocabulary involved, it is unlikely in the
Note
extreme
that two languages would independently borrow essentially

also innovating new vowels in a handful of Austronesian
words, while
the same
inherited
in both Ace?
roots, resulting in the same adjustments to the vowel configurations
hnese and the mainland
languages and leaving both the mainland languages and
Acehnese
in virtually perfect correspondence patterns with each other.
The distribution
and patterning of the borrowed MK vowels by itself

constitutes strong evidence that Acehnese
is a Chamic language that migrated
to northern Sumatra, evidence only made stronger by the
from the mainland
handful of parallel innovations found in the inherited Austronesian
vocabulary.
(3) As discussed in considerable detail in "The borrowed PC main sylla?
ble vowels"
correlate
on page
vowel
length
126, Cowan
in mainland
(1948,
Chamic
1974), Shorto (1975), and this work all

vowels
with
the
subsequent
develop?
ment of diphthongization
inmodern Acehnese. Cowan (1948, 1974) wrote exten?
on
as
the
did
Shorto (1975); both noticed that the vowel length
sively
topic,
contrast between long and short /a/ in closed syllables correlated with vocalic dis?
including in PMP roots. Shorto concludes that the corre?
are
he
found
spondences
proof of a subgrouping relationship between Acehnese
and the rest of Chamic. Collins (1991:116), in contrast to Cowan, Shorto, and the
tinctions
inAcehnese,
inAcehnese and the

position taken in this work, argues that the diphthongization
vowel length in PC are historically distinct phenomena.
Both Cowan and Shorto are correct. The correlation in closed syllables
between PC length and modern Acehnese diphthongization
is well-attested
in the
literature; the existence of PC reconstructions only strengthens the case, estab?
lishing that the PC length contrast between *-a- and *-a:-, which was inherited
as a length distinction, subsequently developed into diphthongiza?
into Acehnese
tion. Note, as Cowan (n.d.:4) argued with reference to this development,
that the
of certain long vowels in closed syllables is of a
diphthongization
of PMP final *-i and *-u in open
younger date than the diphthongization
a
was
"confirmed
the
fact
it
fact
that
still active inNorth Sumatra at
syllables,
by
of
in
the time
Arabic loanwords with long /a/
Islamization; witness its occurring
Acehnese
much
in final closed
syllables."
of the Chamic
Classification
53
Languages
The reconstructed PC length distinctions correspond almost exception

closed syllable diphthongs and also with the
with
the modern Acehnese
lessly
skewed distributional patterns, making it quite impossible to attribute the corre?
spondences to parallel but independent developments. Specifically, PC length in /
not before all finals, but only before final *-?, *-rj, *-k, *-l, *
a/ is distinguished,
r, *-n, and *-t.Where the Acehnese data is available, with a handful of irregular
of /a/ in
forms, the same distributional constraints apply to the diphthongization
(see the examples
Acehnese
independent developments
PC also displays
*-o-
vs.
in *-u:-
and
length distinctions
in ?6.5). There is little likelihood

and the rest of Chamic.
and discussion
in both Acehnese
but
*-u-,
again
in very
only
these two vowels
with
in *-o:- vs.
for two other vowels,
length distinctions
will
contexts.
limited
of
Whether
also have Acehnese
the
correlates
seen.
remains
to be
Shared
innovations
in the lexicon
and the rest of PC share a great deal of vocabulary borrowed

(1) Both Acehnese
sources
Proto-South
from MK
(Proto-North-Bahnaric
(PNB), Proto-Mnong,
Bahnaric, Proto-Katuic), as many people have pointed out.
Table 5 shows a handful of MK borrowings that reconstruct to PC and
have regular reflexes inAcehnese;
others are found throughout this work and in
II.While
individual words may well have been borrowed indepen?
Appendix
into both Acehnese
dently
that
unlikely
all or even
most
and the mainland

of
the
shared
Chamic
MK
it is highly
languages,
can
borrowings
be
accounted
for
in this way.
The word 'cotton' reconstructs to PWMP, but not to PMP. The word
itself also appears to be a late loan intoWMP, but the variant inAcehnese
and
mainland Chamic comes from MK.
Table 5: MK borrowings found
PNB
PMnong
PC
inAcehnese
and mainland
Chamic
Acehnese
*kron
x*kro:rj
kruarj
'river'
*tap
x*klap
tob,
*cem
*sim
x*cim
cic?m
'bird'
*ha
*ha
top
'stab;
poke'
x*ha
hah -f
'open (mouth to say sthg.)'
*?c
x*?sh
??
'excrement;
*joh
x*coh
coh
*kan
*karj
*p?r
*p?r
*kapayh
-f
defecate'
'peck (of bird)'
x*ka:rj
kuiarj
x*p9r
phA
x*kapa:s
gapuiah
-i
'chin; jaw'
'to fly'
'cotton'
Chapter
54
(2) A second set of MK borrowings provides even more
and the mainland
languages?the
relationship between Acehnese
insight into the

MK borrow?
ings that do not reconstruct to PC but which nonetheless show up inAcehnese!

[Note that the forms listed in the post-PC column in Table 6 are explicitly marked
x
with the as not reconstructing to PC].
Some of these words show up elsewhere inAustronesian,
although usu?
ally in a slightly different form. However,
that were borrowed into various Chamic
theAcehnese
variants look to be forms
after the breakup of PC. The

that
shows
of
these
forms
in
Acehnese
Acehnese
presence
speakers left Champa
not only after the breakup of PC but also after these words were borrowed from
MK
dialects
sources.
Table 6: Post-PC MK borrowings
PNB
Acehnese
*groh
xgroh
kloh
'to bark'
*c9k?w
xcagau
cagea
'Malaysian
xdhual/r-f
dhoi
*bartoh
xpatuh
buiruitoh
*kamon
xkamuan
kuimuan
xcfsl
dm a
thai
*bad?k
xpruac (?) pruet
*orj
xho:n
*tach?m
*cum
xcum
*brak
x?9mra:k
(3) In addition
of the pronoun
-f
'dust;
-i
fog,
bear'
mist'
'to explode'
sister's
'nephew;
son'
'shallow'
'stomach;
h'uarj
'wasp'
com
'kiss;
muira?
Chamic
and mainland
post-PC
PMnong
"mon
inAcehnese
intestine'
smell'
'peacock'
to the borrowings, there are also various bits and pieces

that look to be related as well as some shared borrowed
system
terms
(cf. xk9muan 'nephew, sister's son' and x*?u:rj 'husband;
kinship
male'). However, among the singular pronouns, the only particularly noteworthy
MK
is the use of the innovated first person polite

form PC *hulun from 'slave', but this innovation is not restricted just to Chamic.
More interesting are the plural pronouns, which at the PC level have PC
innovation
shared with Acehnese
*g9p 'group; other', which

form is borrowed from MK
The evidence
Cowan
between
from Acehnese
shows up as a part of various plural pronouns. The

and, of relevance here, found inAcehnese.
epic poetry
epic poetry what he described as "the similarity

verse form sanja' and a related Cham verse used
the national Achehnese
observed
in Acehnese
in the song of the kadhar (amusician-officiant)
at the ceremony of the sacrifice of
of the Chamic
Classification
55
Languages
the buffalo" (in 1933, cited from Cowan 1982:156). Cowan observed the exist?
ence of a structure that consisted of lines of eight disyllabic metrical feet, alter?
nately describable as four pairs of such feet, with specific patterns of internal
rhyming.
However, in addition to what Cowan noticed, there are other indications
of the origins of the sanja' poetic tradition. Preserved in the rhyme schemes is
evidence of early vowel changes, changes so early that they predate the origin of
Acehnese
itself. As G. W. J. Drewes wrote (1979:4), Acehnese
epic poetry is
written
in what
the Acehnese called sanja' (from the Arabic saj), several features
are of particular interest. First, for the purposes of rhyming, long and
the significance of which will be made clear
short syllables are not distinguished,
in amoment. Second, two syllables make up a foot, with the accent on the second
of which
syllable. As Drewes notes the foot is iambic or, from the viewpoint of historical
comparison, not unlike the structure of Chamic. And, in this poetry, the most
common
rhyme scheme consists of the last vowel of the sixth foot rhyming with
the last syllable of the fourth foot. Thus, in the following example from Drewes
(1979:4), the two italicized words rhyme:
gah ban / gadjah / si? ban/ piti II
ph?t di / gaki II sa ngon / rusa,
'fame as an elephant, a body as a ricebird;
his sense of honour is in his legs, just as with
the deer'
if one examines, for example, Hikajat Potjut Muhamat, an epic poem

However,
from roughly the end of the 17th century, it becomes clear that, while most of
what counts as a rhyme actually does rhyme, not all of the so-called rhymes are
still phonetically
transparent. As Drewes notes,
count as rhyming (using Drewes' orthography):
counted as rhyming with
in this epic the following
eue
[ui9],
are eu [m] and

are ? [s] and
[e],
u [39]
are o
eu
[sa].
and
[0]
vowels
In the Hikajat
these rhymes are bizarre phonetically, but if the

Potjut Muhamat
are replaced with their corresponding
in the Hikajat Potjut Muhamat
forms from PC, the rhyme scheme, which ignores length differences, is perfect:
vowels
counted
as
rhyming
with

counted
as
rhyming
with
are
*-a:-
are *i
are
[09]
*-u-
(nasalized)
and
(nasalized)
and
and
*-09
<
*-?9
*-a:-,
*-i -,
<
*-sw
<
*-u.
Chapter
56
That is, all the forms that count as rhyming with /a/ were /a/ earlier in history, all
the forms that count as rhyming with HI were HI earlier, and all the forms that
count as rhyming with lui were lui earlier. The comparative evidence suggests
that the first two rhymes were phonetic rhymes in an earlier stage of Acehnese.
However, the third set of rhyming forms is astounding: the last time these rhymes
actually rhymed phonetically has to be at the very least contemporary with an
early stage in the history of Chamic! That is, by the most
least 800 years earlier!
Acehnese
as a Chamic
conservative
estimate at
language
though the linguistic evidence presented so far is more than adequate to

establish the relationship beyond question,
the countless other details given
remainder
this
work
full
substantiation to the long-held
the
of
throughout
give
Even
thatAcehnese
is a Chamic language. Certain of the shared innovations are
so specific and so striking in their uniqueness that even on their own they would
is a Chamic language.
constitute evidence thatAcehnese
belief
hnese
tion
(p.c.), since in subgrouping terms Ace?

Following Durie's suggestion
is simply another Chamic language, this work will simply use the designa?
"Chamic"
to
refer
suggestion Aceh-Chamic
Achino-Chamic
(1975).
The contribution
to
the
Chamic
family,
rather
(1990a), Cowan's Chamo-Achehic
of Acehnese
than
Durie's
earlier
(1988), or Shorto's
to PC reconstruction
The finding thatAcehnese

is a Chamic languages makes it invaluable as a source
on
of information
the reconstruction
of PC and on language contact and
that has only been modestly
change?one
tapped in this work. Before the Ace?
hnese
speakers
split
off,
it seems
reasonable
to assume
that Chamic
speakers
con?
coast?a
stituted a relatively homogenous
dialect chain along the Vietnamese
in part by the fact that the dialects
relative homogeneity which was maintained
were most likely subject to essentially the same type of pressures from MK lan?
guage contact and by the tendency for changes originating in one part of the dia?
of PC *-ay in
lect chain to spread throughout the whole (cf. "Diphthongization
Rade" on page 125), for a post-PC change found everywhere on the mainland,
except Rade). Even after the breakup of PC, convergence among the mainland
languages was promoted both by the similarities in theMK contact patterns and
by the tendency for innovations to spread throughout a dialect chain?tendencies
of the Chamic
Classification
which
the mainland
among
57
Languages
Chamic
languages
have
until modern
continued
times.
speakers left the mainland, they left behind the lev?

chain that
eling effects of the spread of changes through the dialect chain?a
existed at least well into the 18th century and the linguistic pressures arising out
of the continuing MK contact on the mainland. As a consequence, for possibly as
much as a thousand years, Acehnese has been outside the influence of many of
When
the Acehnese
the linguistic pressures that have influenced themainland Chamic languages. And
also as a consequence, Acehnese
preserves a living record of an intermediate
one
that has been largely lost through subsequently
stage in the history Chamic,
leveling in the languages remaining on the mainland, albeit a stage that must be
used with some caution, but which nonetheless is potentially invaluable.
The methodological
problem, of course, lies in figuring how to interpret
If a linguis?
tic feature occurs only inAcehnese, does it represent an independent innovation
inAcehnese or a retention preserved only inAcehnese and lost on the mainland?
Sometimes it is clear; for example, it is clear from comparisons with non-Chamic
differences
between Acehnese
Austronesian
languages
are
Acehnese
and the mainland
that the medial
Chamic
homorganic
languages.
nasal plus stop clusters
in
a retention.
In other
cases,
the answer
is not
as readily
apparent.
An
area
of potential
is interpreting the apparently large number of lexical innovations, par?

MK
ticularly
borrowings, shared by the mainland
languages but not Acehnese
and reconstructed by Lee (1966). Subsequent work has shown many of these to
difficulty
be
post-PC
borrowings,
whose
occurrence
only
in the mainland
simply
indicates
that the borrowings took place after theAcehnese

speakers had left themainland.
Their status as borrowings is sometimes evident from the presence of otherwise
unexpected clusters or from irregular correspondence patterns ("Post-PC clusters
Chamic languages" on page 96). However, in other cases, a post-PC
from
MK may not leave recognizable traces. Thus, even some of the
borrowing
MK forms reconstructed in this work at the PC level may in reality be post-PC
inmainland
borrowings,
not yet formally detectable
as such.
10. In this regard, Acehnese

is certainly not without parallels. Within Austrone?
sian, Malagasy
presents another example of a language being removed from
a linguistic area before the areal tendencies have an opportunity to fully
as Blust (p.c.) points out, has a Philippine-type morpho
affect it.Malagasy,
syntax, unlike any other language in its immediate subgroup (Barito), or any
language of southern or western Borneo. The detailed agreements of gram?
matical morphemes
show clearly that this is due to retention. Malagasy
left
southern Borneo before the widespread breakdown of this type of system
spread through much of western Indonesia.
Chapter
58
The areas of indeterminacy about when and from where forms are bor?
has
been significantly reduced, but there is still a small subset of forms
rowed
whose history is unclear. Certainly one problem is that the MK borrowings are
from
several
sources:
MK
some
are
from
some
Bahnaric,
are
from
Katuic,
and,
some are from theMK
languages spoken in Funan.

of the MK borrowings in the modern mainland
possibly,
Chamic
Comparisons
new
to the
MK
in
Acehnese
wrinkles
with
adds
the
several
borrowings
languages
as
at
that
the
Ace?
has
it
is
least
Durie
suggested,
problem: first,
possible,
(p.c.)
hnese picked up some MK words from the Aslian MK languages spoken in the
Kelantan
established
area, a possibility
suggested earlier by G. Diffloth (p.c.). It is well
that there was a Chamic influence in Kelantan and there is no reason
to assume
well-known
sions,
speakers. Second, as the

has said on a number of occa?
that this influence did not include Chamic

Mon-Khmer
coast
the northeast
specialist G?rard Diffloth

of
northern
Sumatra
a number
has
of MK
place
that some of the MK borrowings

certainly leaving open the possibility
in Sumatra.
hnese were borrowed even after the arrival of the Acehnese
names,
in Ace?
is complicated.
Thus, interpretation of the MK component inAcehnese
In some cases, our understanding
is improved. For example, much of the MK
material that Lee reconstructs for PC does not occur inAcehnese. The reworking
of Lee's PC reconstructions
many
these MK
of
forms
in light of additional
absent
from
Acehnese
evidence
do
not
has made
actually
it clear that
reconstruct
back
to PC; rather many of these forms are post-PC MK borrowings which presum?
departed from the
ably entered various mainland
languages after the Acehnese
In other cases, if as Diffloth notes (p.c.), there isMK material inAce?
does not appear in the mainland languages, itmust be accounted for,
but if Acehnese
speakers were the most northerly of the Chamic speakers, they
mainland.
hnese which
have picked up some of this vocabulary from Katuic speakers, also in the
north, before leaving for Sumatra. In any case, although there are obvious poten?
tial sources for these forms, the answers to these questions have not yet been
may
found.
A DIGRESSION ONMOKEN, A NON-CHAMIC LANGUAGE

appropriate place for a note on Moken, just to point out
is not Chamic and to give some of the obvious evidence. Moken, at
least the sixty or soMoken words recorded by Martha Blanche Lewis (1960) and
cited in Blust's dictionary, are enough to demonstrate thatMoken is not a Chamic
Here might
thatMoken
language.
be the most
not share the PAn *q > *h change found in the whole of

a lack which sets it, not just outside of Chamic, but outside
as well; instead, it has the change PAn *q > Moken /k/, a rather
It does
Malayo-Chamic,
Malayo-Chamic
Classification
of the Chamic
59
Languages
unusual reflex of PAn *q. Like Chamic, it does diphthongize both the high vow?
els in word-final position, but in this its similarity to Chamic is only typological.
The Chamic and Moken
reflexes of the word-final high vowels are split in
entirely different ways
*-ow
(< PAn
*-u),
Moken
in the two languages.

has
-uy,
-oi and
Instead of PC *-ey (< PAn *-i) and

-ui,
-oi,
respectively,
at
least
suggest?
ing that the PAn *-i and *-u have been merged inMoken. In any case, theMoken
developments make it clear thatMoken does not share in the Chamic diphthongi?
zation of word-final high vowels nor in the PAn *q > PC /h/ change. In fact, thus
far, although there are some typologically
similarities, there is no evidence that
Moken shares any of the more marked inherited innovations characteristic of the
Chamic
languages.
the Basic
Word:
Altering
toMonosyllabic
From Disyllabic
On the basis of the material
in Blood
(1962), Greenberg
(1970:139)
wrote,
con?
cerning Phan Rang Cham:

This
earlier,
thus
language,
tended
like the closely

toward
loss
extensive
producing
related Jarai, Rade, and Chru mentioned
or reduction
of
the vowel
of
the first
syllable,
monosyllabism.
Although not in precisely these terms, early writers commented on the

movement from canonically disyllabic forms to canonically iambic forms, that is,
to forms with an unstressed presyllable and a stressed main syllable. For instance,
(1963:61) notes that for Jarai and Chru, PMP disyllabic forms with
*r, *R, *q, or *h usually lose the first vowel, while inmost other instances
the first syllable vowel is reduced to shwa. Greenberg (1970:137), characterizing
Thomas
medial
the process
in more global terms, writes, "These languages have generally
norm by
reduced the typical Austronesian
canonical CVCVC to the monosyllabic
reduction or loss of the first vowel". That is, the PMP canonical CVCVC has
sesquisyllabic with an iambic stress pattern or else simply monosyllabic.
structure has its correlates
This adjustment of the canonical morpheme
in both the reduction in the membership
and the restrictions in the distribution of
became
the consonant
and
vowel
introduced from MK
inventories.
Various
new
in borrowed words. The most
consonants
(and
vowels)
were
salient of these borrowed con?
the series of voiced glottalized obstruents, since there were no glot?

talized consonants in the PMP predecessor, but there have also been a myriad of
sonants were
60
From Disyllabic
to Monosyllabic
61
less conspicuous changes, one of the most widespread being the introduction of
new clusters through the borrowing of words containing them. As for the vowels,
there has been proliferation through outright borrowing from MK sources as well
as through internal developments
under MK
influence.
THE IAMBIC SYLLABLE:EARLY MON-KHMER
INFLUENCE
the history of Chamic, there is a clearly-marked,

continual move?
Throughout
ment from disyllabic to iambic and from iambic to monosyllabic,
but the initial
was
a
to
stress under
for
the
PC
these
shift
for
final
impetus
changes
preference
the canonical preference was set for stress on the final sylla?
ble, the preconditions had been set up for the restructuring of the Chamic lexicon.
are either monosyllabic
or what Matisoff
In general, MK morphemes
MK
influence. Once
termed sesquisyllabic,
i.e., a syllable and a half, with a
(1973) picturesquely
stress pattern characterized by Donegan
(1993:5) as iambic, that is, "words in
which a light (open) syllable precedes a heavy (closed or long-voweled)
second
syllable", that is, words with an unstressed presyllable and a stressed main sylla?
ble. Both Matisoff
(1973) and Donegan
(1993) note in passing that words in
(of which MK is one of the two major branches) were either
proto-Austroasiatic
iambic or monosyllabic.
Of direct interest here is the fact that this characteriza?
tion fits the MK languages of Vietnam perfectly; for example, Chrau (Thomas
Vietnamese
is monosyl?
1971) is iambic, Mnong and Rolom are monosyllabic,
on.
so
As
stress
Lee (1974:645) mentions, "Ultimate
in languages with
labic, and
phonological words of more than one syllable is regular throughout much of
Southeast Asia." And then he adds, "The Chamic languages without exception
have
only
ultimate
stress."
The restructuring of Chamic lexicon and phonology both provide elo?

quent testimony to the intensity and the intimacy of theAustronesian contact with
MK. Lexically, of the roughly 700 forms Lee (1966) reconstructed for PC, Head
10% is a con?
ley (1976) identified roughly 10% of them as MK in origin?and,
servative figure. Included among the MK
incorporated early enough to be
incorporated into PC are basic vocabulary including pronouns, and a number of
kinship terms. The fact that these reconstruct to PC shows that the early contact
was intense and intimate, suggesting both considerable bilingualism and inter?
marriage.
the nature of the earlier system is not absolutely clear, the first
speakers to come into contact with theMK speakers along the coast
most likely had a system, not with exclusively penultimate stress, but a system
like that in modern Malay with penultimate stress in the vast majority of disyl
While
Austronesian
Chapter
62
labic words but ultimate stress in a small group of words with a reduced shwa as
the initial syllable, as Lee (1974:646) suggests. In this interpretation, contact with
MK led to a shift of preference from penultimate to ultimate stress, rather than
an
introducing
entirely
new
stress
pattern.
The shift in preference, however, set inmotion a chain of events that has
itself in different ways in different languages and, of course, under
manifested
different contact conditions with the consequences
already present in the conso?
inventories are reduced
inventories of PC: The PC presyllable
of the earlier PMP first syllable inventories, while the PC main syllable
inventories are larger than the inventories found in the earlier PMP second sylla?
nant and vowel

versions
bles.
The
internal paths of this historical developmental

continuum from
are
to
of
canonical final
transparent. The adoption
disyllabicity
monosyllabicity
stress by PC was enough to set up the internal preconditions for movement
in the
All that remained was for the changes to be trig?
of monosyllabicity.
gered by subsequent intense, extended contact with a monosyllabic
language. In
the
PC,
pretonic syllable still had four possible distinct vowels, but since then all
the mainland Chamic languages have shown a steady erosion of the pretonic syl?
direction
lable, beginning with the reduction of vowel distinctions. With the exception of
the descendant languages have reduced the vowel distinctions at least
Acehnese,
somewhat, although in a language such as Roglai, the presyllable still shows a
three-way distinction in certain environments. In some of the other languages, the
pretonic syllable still exists but in still others the presyllable has undergone a
complete loss of vowel distinctions; for example, in Chru and Rade (for the Rade,
see Table 9, page 66), only one vowel is found in the pretonic syllable?a
shwa.
The movement
toward monosyllabicity
In individual modern languages, the PC disyllabic, iambic forms have preceded

even further in the direction of monosyllabicity,
with the number of consonantal
even
more reduced. After the break?
contrasts
in
the presyllable being
and vocalic
up of PC, some languages, such as Roglai, Rade, and Jarai, apparently in contact
with languages typologically very much like PC and thus under minimal external
the basic word shape, have remained largely unchanged.
pressure to modify
However,
other Chamic
languages,
under the influence of different
patterns of
contact, have in varying degrees gone even more towards monosyllabicity.

One path that leads to a dramatic movement from disyllabic tomonosyl?
a generation or two is illustrated by Phan Rang Cham, where the pre
syllables were first omitted in informal, colloquial speech and now seem to have
been dropped entirely by some speakers. Doris Blood (1962:11) writes,
labic within
to Monosyllabic
From Disyllabic
63
Cham words may have two or rarely three syllables, but there is a strong
The final syllable of disyllabic words,
tendency toward monosyllabicity.
here
to as
referred
the main
carries
syllable,
stress.
the primary
The
con?
of the initial syllable... may vary considerably or

may be dropped entirely. Often in normal speech a word that is some?
times heard as a two-syllable words is fused into one, as in per?w >
sonants and vowels
pr?w new, and in kel?k > kl?k to tickle. The following

been
observed:
~
per?w
Scholars
pr?w
phir?w
tend to maintain
~
phr?w
fir?w
full forms of words
variations have
-
fr?w.
in their speech.
by the loss
As a general rule the speech of non-scholars is characterized

of preliminary syllable, reduction in vocalism or assimilation
lable with
the main
of that syl?
syllable.
In these examples, the reduced forms have come about by rather natural reduc?
tions and then loss of the presyllable vowels, but the complete loss of the presyl?
lable can come about quite dramatically.
as in PR Cham, this tendency to drop syllables
is
also
found throughout Chamic, even in Acehnese,
colloquial speech
long removed from the MK sphere of influence. Durie (p.c.) comments that it
occurs in personal names and in certain highly frequent terms. Thus instead of
Although
not as marked
in more
anwk manok, a villager might say nwk manok

'chicken' and instead of miyup
rumo h 'under the house', they might say y up moh.
Although usually not as dramatically captured inmid-change,
through?
out Chamic the reduction of disyllables, sometimes all the way tomonosyllables,
has occurred and almost all of it has in part been motivated by the tendency of the
presyllable vowel to reduce to shwa and then drop, as shown in colloquial Phan
forms.
Rang Cham variation, collapsing the disyllabic forms into monosyllabic
The bulk of such reductions follow one of three patterns of reduction.
The
loss
of
the
vowel
before
medial
-h
The first of the three patterns which led to monosyllables

may have been com?
the
PC
the
When
the
initial
of
main
stage.
pleted by
syllable was *h, as in Table
7, the initial of the pretonic syllable and the initial of themain syllables coalesced
into an initial cluster, a pattern noticed by Blood (1962) and commented on by
Greenberg (1970:139). One consequence of their origin from the coalescence of
forms.
disyllables is that such clusters only occur inmodern monosyllabic
Table 7 illustrates these developments nicely. InMalay, a closely-related
but non-Chamic language, the forms are still disyllabic with themedial -h- as the
onset of the second syllable. However, in PC, as reflected both inAcehnese and in
Chapter
64
Chamic languages, these forms have been reduced tomono?

with
various clusters in which the second consonant is -h-.
syllables beginning
the various mainland
Table 7: From disyllabic
PMP
*h to monosyllables
syllables with medial
Malay PC
Aceh.
Chru
NR
Tsat
*paqit
pahit
*phit
phet
phi:?
phi:?
phi?24
'bitter; bile'
*paqa
paha
*pha
pha
pha
pha
pha33
'thigh'
pohon
*phun
phon
phun
phut
phun33
'trunk;
pha:?24
'chisel;
?
*paqat
pahat
*pha:t
phuiat
pha:?
*daqan
dahan
*dha:n
dhuian
tha:n
tha:t
*taqu
tahu
*thow
thsa
thau
thau
tiau?42^
'know;
*taqun
tahun
*thun
thon
thun
thut
thun33
'year'
plus
(1977:78)
-h-
actual
'branch;
clusters,
not
aspirated
noted that in Chru such forms behave phonologically
unitary
phonemes,
an-h?
'a plane',
an
in which
tion of disyllables.
have
as evidence
citing
there is an established
of which
are
combinations
This
parallels,
infixed
to
plane'
bough'
able'
(p.c.) has pointed out, these conso?
In some instances, as G?rard Diffloth

nant
stem'
alternations
nominalizing
such
-an-
as
ph?
occurs.
stops.
Fuller
as clusters, not as
'to plane'
Certainly,
and
where
etymology, such clusters inevitably come from the reduc?

too is a point of convergence with MK languages, many
for
example,
Khmer
khaat
'lose'
and
k-om-haat
'loss'.
DISYLLABLES WITH LIQUIDS > MONOSYLLABLES AND CLUSTERS

a second pattern in which disyllables,
but with an originally
*-l- or *-r-, coalesced into monosyllables
with initial clusters. Notice that
the original disyllabic forms are still retained as disyllables outside Chamic in
Table
8 shows
medial
PMP and Malay

guages Acehnese
and within Chamic
in PC as well as in the modern Chamic
lan?
and Northern Roglai (not shown in Table 8), but the disyllables
been
have
reduced to clusters in Chru and Tsat. Within Chru and Tsat following
the initials of the pretonic syllable and the
In Tsat, the process of mor?
syllable coalesced, producing monosyllables.
one
structure
with the *-l- or the *-r
has
further
gone
step
pheme
simplification
the loss of the shwa in the presyllable,
main
of initial cluster becoming
an -i- glide.
1. The patterns of these secondarily-derived

detail in Chapter 5.
clusters
are discussed
in more
Table 8: From disyllabic

PMP
65
to Monosyllabic
From Disyllabic
Malay
darah
*daRaq
PC
Aceh.
Chru
*darah
darah
drah
*daRa
dara
*dara
dara
dra
*bulu
bulu
*bulow
bulsa
blau
pulau
*bulan
*baqeRU
*pulaw
pulo
pala:u
Tsat
sia55
'girl'
phia11
-1
bulan
*bula:n
buluian
*bahrow
baro-f
barhau
phia11
pla:?
pie?24 -v
pula
*pula
-phian11
pala
pia33
?
gulung
*gulung
?
*bara
*qabaRa
gulon
*gulun
karam
*karam
kura
*kura
?
?
?
parlarj-v
bra
kram
?
*palsy
phia11
?
?
kra
plai
-v
'body hair'
'island'
baharu
*pala:t -1 paluiat
'blood'
bla:n
*palaj
liquids to monosyllables
syllables with medial
piai33
'moon'
'new'
'palm; sole'
'to
plant'
'to roll'
'shoulder'
'sink;
sunk'
'turtle'
'village'
Quite parallel to these secondary clusters, which have developed within

the history of Chamic, are a set of earlier primary clusters, that is, clusters which
had already become clusters by the PC stage and which had developed in a simi?
lar way from the loss of the vowel preceding a medial *-l- or *-r- in an earlier
disyllabic
form.
LOSS OF THE UNSTRESSED INITIAL SYLLABLE

cases, however, the loss of the shwa of the presyllable resulting in the
initial and the main syllable initial would have
juxtaposition of the presyllable
resulted in a highly-marked cluster. Thus, it is not surprising that when the main
In most
syllable began with any other consonant

syllable was lost (see Table 9).
than *h- or a liquid, the whole
pretonic
from disyllabic to iambic to

As this table shows, throughout Chamic there has been a unidirec?
monosyllabic.
Outside of Chamic, PMP has
tional movement along the path tomonosyllabicity.
Table 9 shows the unidirectional
movement
forms as does Malay. Within Chamic, the disyllables are rapidly

through the reduction and loss of the presyHables: the
becoming monosyllables
but it
four-way vowel distinction of the PC presyllable still exists inAcehnese,
fully disyllabic
has been reduced to a limited three-way distinction inNorthern Roglai, while in

Rade the presyllable has been reduced even more, with the four-way vowel con?
trast reduced
to just a shwa and several of the initial consonants
having been
Chapter
66
dropped. In Tsat, the process has gone even further, with almost all the remaining
disyllables reduced tomonosyllables
through the total loss of the presyllable.
to monosyllables
Table 9: From disyllables

PMP
Malay
mamah
*mamaq
*qumah
*lima
*tarjan
*lapaR
PC
Aceh.
Rade
Tsat
*mamah
mamAh
mamah
huma
*huma
lima
*lima
tang?n
*tana:n
lapar
*lapa
padi
*paday
ma55
'chew'
umArj
hama
ma33
'dry field'
limAn
ema
ma33
'five'
kanan
rja:n33
'hand;
epa
pa33
'hungry'
pade
madie
thad?42
na55
'rice (paddy)'
'shoot (bow)'
*baseq
panah
basah
*panah
*basah
panah
basah
manah
masah
sa55
'wet;
*m-uda
muda
*muda
muda
mada
tha11
'young;
*panaq
arm'
damp'
tender'
Thus, although achieved through the interaction of various internal pres?

was originally set
sures and paths, the Chamic tendency toward monosyllabicity
toward monosyllabic?
into motion by language contact and, where the movement
it is due to continued
ity has continued,

have
was
even
more
restricted
set into motion
preferred
language contact with
morpheme
structures.
That
languages
is,
the
that
process
contact, resulting in iambic forms; the subsequent

seems to be due in large part to subsequent Phan
reduction to monosyllables
Rang Cham contact with the monosyllabic Vietnamese and Utsat contact with the
monosyllabic
languages
by MK
of Hainan.
Consonants
Chamic
for the develop?

shift to a preference for final stress has its consequences
ments among the consonants: not only do final consonants have their own unique
path of development, but so do the presyllable and themain syllable initial conso?
The
nants. For the presyllable initials, the tendency is for the contrasts to be steadily
reduced, beginning with the transition from PMP to PC. For themain syllable ini?
tials, the opposite was initially true; the initial contact with MK languages
to monosyllables,
expanded the inventory, as did the reduction of disyllables
which introduced new clusters. However,
this introduction of new contrasts is
as an incidental byproduct of the reduction of disyllables to mono?
as with all consonants, the central tendency among the main-syllable
best viewed
syllables;
initials
was
also
towards
sense
Making
-language
beyond
variation
the
out
of
contrasts.
reduction
of
these
tendencies
and,
in particular,
language-to
gets reduced and by how much, requires looking

to the differences
in patterns of
internal motivations
just the language

It becomes clear that direct correlations
bilingualism.
developments
the
in what
exist between
the internal
and the external bilingualism.
PRESYLLABLE CONSONANTS:THE DETAILS

is only a very limited number of consonants that can begin the presyllable
there are six voiceless obstruents: three stops, an affri?
(Table 10). Specifically,
are four voiced obstruents: three voiced stops and
two
and
fricatives.
There
cate,
There
a voiced affricate. Finally, there are three sonorants: one nasal (*m-), and two liq?
uids (*1- and *r-), with the further likelihood of an additional, marginal *ft-,
which
'breath;
seems
to
occur
only
in
two
words:
*?amuk
'mosquito'
and
*?awa
life'.
67
68
Chapter
Table 10: The presyllable
*t-
*p-
consonants
*c-
*k
*s-
*h
*h-
*d-
*i-
*m-
*1-
*fl-
*2
(rare)
*r
All of these presyllabic consonants existed as such in the PMP language

spoken by the PMP speakers who originally came toVietnam, although of course
various MK forms with similar structures have been added to the total number of
forms. Within
within
Chamic
itself, the full array of these forms is found unaltered

records of Cham, although many of these presyllable
have since undergone the changes reflected in the modern Phan Rang
the older written
consonants
Cham andWestern
all of the
Similarly, even inmodern Acehnese,
presyllable consonants except for *h- are still found as such.
Changes in the presyllable consonants have occurred in all the Chamic
Cham dialects.
languages, with the distribution of the changes suggesting that contact patterns
were a major determinant of the paths of change. Acehnese, which retains all the
consonants except *h-, left the mainland
before the bulk of the
presyllable
on
The
the
mainland,
changes began.
languages remaining
except for Rade,
underwent
similarities
patterns.
changes that were at least in part influenced by contact, as many of the

in the developments
cut across subgrouping lines but match contact
The
lone
exception
is Rade,
which,
as
the geography
suggests,
patterns
as if it were on the edge of the dialect chain existing among the Chamic dialects
remaining on the mainland. Certainly, the developments among the Rade presyl?
lable initial consonants are unique to Rade.
The
rather salient
lack of presyHables beginning with *n-, despite the

existence of *n- in the presyllables of PMP, seems to correlate directly with a
similar scarcity of such forms in, say, Malay, where, although a handful of appar?
ently disyllabic native words with n- seems to exist, the overwhelming majority
of disyllabic forms beginning with n- are either obviously borrowed or just as
obviously a secondary result of the effects of verbal prefixes. In this context, Lee
(1974:652) writes that, although only Manley
(1972:25) seems to specifically
note the absence, "apart from reduplicative patterns and some cases of syllabic
nasals homorganic with the first consonant of the following syllable, normally m
is the only nasal that can occur as the onset syllable inmany languages of Viet?
nam."
1.
Initial *h- in presyllables
does not occur in Durie's
reconstructions
of Proto
Acehnese.
Chamic
69
Consonants
Then, Lee suggests the thus-far-unexceptional

*n- to PC *1-, citing five of the examples in Table 11.
*n- > PC */
Table 11: PMP

PMP
Malay
*niyuR
nyor
*h-in-ipi
mimpi
W.Cham
Aceh.
*la?ur
boh u
la?u
lumpoa
lupai
*lumpsy
?
*lanah
nipis
*nipis
NR
PC
nanah
*nanaq
of PMP
correspondence
lipeh
*lipih
*lawa:n
*niwag
lan?h
lapay
?
'pus'
lupih
lapih
'thin (material)'
luwa:k
lawan
'thin'
nasi
'coconut palm'
'to dream'
la?u
'cooked
las?y
*lasey
rice'
In the case of 'dream', the PC forms and the Acehnese form in particular look to
be the product of the string PMP *h-in-ipi > *nipi > *lipi > *l-um-pi > *lumpey,
stuck into the ultimate Acehnese
with two layers of distinct verbal morphology
is restricted
form. This particular change

Chamic, including Acehnese.
to the Chamic
component
of Malayo
Chamic, xnagar 'country, city; area'

borrowing narjgroe represent distinct variants borrowed inde?
initial is not the expected l\-l and its vowel suggests an
The
Acehnese
pendently.
form is borrowed from
earlier *-i, both of which suggest that the Acehnese
into mainland
The Indie borrowing
and the Acehnese
Malay.
The voiceless
The
reflexes
Rade,
also
obstruents
of the voiceless
the *p- has merged
merged
*t-,
*k-,
a number of MK
sonant of a minor
as c-. Elsewhere,
and
*c-
as fc-,
a merger
also
in the minor
found
syllables
of
In fact, Chong (MK) allows only k- as the first con?

In Haroi, the PC presyllable *t- has merged with *c
languages.
syllable.
as Table
are, by and large, straightforward. In

the *b- and the *m-, all becoming m-; Rade has
obstruents
with
12 shows,
the various
are kept
series
distinct.
12 are of particular inter?

est and merit more discussion
than they will be given here. In addition to the
listed *p-, *t-, and *k- reflexes, there are also forms with what appear to be the
Certain other Chru reflexes not listed in Table
reflexes
par-,
tar-,
and
kar-,
respectively.
However,
these
forms
are most
likely
related to the reduction of earlier trisyllabic (or, polysyllablic)

forms, rather than
forms
of
the
conditioned.
Some
apparently were
trisyllabic
being phonologically
For example, in some
the residue of either borrowed or native morphology.
'RECIPROCAL'(Goschnick
the par- may be related to the Haroi pala1977:115). Thus, in Chru the form for 'divide; share' is parpha, while in Phan
Rang Cham there are three forms, one of which is parapha. In addition, there is
the form 'rabbit', which is tarpaii in Chru, but *tarapay in PC. Finally, there are
instances,
70
Chapter
such forms which
many
apparently do not occur elsewhere
in PC, and thus may
be borrowings.
Table 12: The voiceless

PC
Aceh.
Rade
*p
p-
m-
k-;
k-
*k
*c
g-
obstruents
Jarai Chru
Tsat
Haroi W.Cham
p-
p-
p-
0-
p-
p-
k-
k-
k-
0-
k-
kk-
t-
t-
0-
c-
c-
0-
c-
t-
k-
t-
c-;
k-
c-
j-
NR
xs-
c-;
*7_
PR Cham
p
tc c-
(unmarked)
In most
to tell from the material available

cases, it is not possible
or not forms written with a word-initial
vowel actually have glottal
whether
onsets.
at
However,
least
so far,
it really
does
seem
not
to have
any
consequences
for the analysis. InWestern Cham the pretonic syllable is lost more often than
not.With the la- pretonic syllable, for instance, it is lost in every case except atau
'ghost;
corpse'.
Acehnese
reflexes
initial *k~ and *c
of presyllable
The Acehnese
reflexes of PC presyllable *k- and presyllable *c- are voiced if and

if
the
main
only
syllable initial is also a voiceless stop (see Table 13, page 70),
but not if the main syllable initial is a sonorant (Mark Durie, p.c., informs me that
this process is productive inmodern Acehnese).
Were it not a borrowing, the word guda
'horse' would be an exception to

this pattern; that is, guda 'horse' has a presyllable voiced stop preceding a main
syllable voiced stop. However, there is good evidence it is borrowed: Aside from
not patterning as might be expected diachronically,
it is not widespread
in Aus?
tronesian and it is suspiciously close to the Gujarati ghoda 'horse' (Karen Mistry,
that Gujarati traders were early visitors to this region.
p.c.), and it is well-known
Table 13: PC *k- > Acehnese
Malay
kita
PC
g-; *o > Acehnese
Acehnese
*campa
juimpa
*k-ita
gui-ta-?oa
'Champa'
'we
(incL)'
kaki
*kakay
gaki
kapas
x*kapa:s
gapuiah
'foot;
leg'
'
cotton'
kuku
*kukow
guksa
'claw;
k?ntut
*kantut
toh
kutu
*kutow
kuda
x?
-v, gateh
guintat
-f
'fart;
gutsa
'louse,
guda
'horse'
fingernail'
flatus
ventrus'
head'
Chamic
71
Consonants
PC *s- and *h
14 and 15, the reflexes of presyllabic *s- and *h- are examined.
the *s- has been retained in presyllables, while the *h- has been lost.
In Table
InAcehnese
In Chru and N. Roglai, the two series are retained largely unchanged. In Chru,
there is a minor but interesting split in the reflexes of *s-. If the main syllable
stop, the vowel of the presyllable was completely
began with a voiceless
with an initial sp-, st-, or sk- cluster as
dropped, leaving the Chru monosyllabic
its onset; otherwise, the complete presyllable remains unaltered. This association
of consonant manner with a split in consonantal reflexes reflects voice quality
differences (Cf. Thurgood, 1980).
Table 14: Reflexes
PC
Aceh.
s-
*s-
Rade
Jarai
h-;
h-
t, k
NR
Chru
s-
sa-;
0-
/_p,
*s
of PC presyllabic
s-
Haroi
W. Cham
PR Cham
h-;
h-
sa
/_vd.
ob st.
elsewhere
sa
In the remaining five languages, the reflexes of *s- and *h- have merged
completely. In Jarai,Western Cham, and Phan Rang Cham, both *s- and *h- have
merged,
h-,
becoming
parent by the frequent

Phan Rang Cham.
Table 15: Reflexes
this
although
is made
development
total loss of the whole
of PC presyllabic
presyllable
Aceh.
Rade
Jarai
Chru
*h
0-
h-;
h-
h-
NR
Haroi
h-
h-;
0-
/_vd.
t, k
obst.
elsewhere
trans?
less
Cham
and
*h
PC
/_p,
somewhat
inWestern
W.Cham
PR Cham
hh0
0-
h-
h-
h-
0-
h-
h-
h-
h-
h-
0-
h-
h-
h-
h-
h-
h-
hh-
in Rade and Haroi,

the merged
reflexes are further split,
Finally,
case
on
in
each
is
it
the
basis
of
although
slightly different conditioning factors.
In both Rade and Haroi, if the main syllable began with a voiceless stop, the h
initial was simply lost (some transcriptions show a glottal onset being added sec?
ondarily to what would otherwise be a vowel-initial word). In Haroi, the process
Chapter
72
a step further with h- also being lost before main syllables with a voiced
obstruent initial. Cf. also "Secondary clusters" on page 94. Again, the association
of manner differences with splits in consonantal reflexes reflects voice quality
went
differences
(Cf. Thurgood,
The voiced
stops and breathiness
1980).
The
patterns for voiced stops

correspondence
breathiness that probably dates back to PC. All
an association
shows
of with
the modern Chamic languages

of
have
reflexes
Acehnese
and
Rade
except
presyllable *b-, *d-, and *g-,
split
with the splits conditioned by the manner of the initial consonant of themain syl?
lable. The reflexes of PC *j- have not undergone such splits.
Table 16: Examples
of reflexes of the PC presyllable
PC
Jarai
Chru
NR
*batow
patau
patau
patau
W. Cham
patau
voiced obstruents
PR Cham
patSw;
patow
'stone'
*bitu?
patu?
patu?-f
pitu?
patu?
pit??; pat??
'star'
*basah
pasah
pasah
pasah
pasah
path?h
'wet;
x*bato
pato
pato
pato
pato
pato
'teach'
*batuk
pat?k
patu?
pitu?
tiki:?
pat??
taki?
pat??
taki?
'cough'
*dikit
taki:?
tapa
damp'
little'
'few;
'arm
*dapa
tapa
tapa
tupa
*gatal
katal
katal
katan
katal
katal
'itchy'
*gatak
kat?k
kata?
kata?
kata?
kata?
'sap;
papah
papah
'mouth'
*babah
span'
resin'
*babuy
babui
pabui
babui
papui
papuy
'wild pig'
*dada
tada
tada
dada
tata
tata
'chest'
*bulow
blau
blau
bilau
plau
palow
'body hair'
*burja
ba?a
ba?a
buna
parjur -f
pirju
'flower'
*bula:n
blan
bla:n
bila:t
plan
pilan
'moon'
*darah
drah
drah
darah
taran
taran
'blood'
*dara
dra
dra
dara
tra
tara
'girl'
*dalam
dl?m
darlam
*gunam
gan?m
ganam
All
dalap
tal?m
tal?m
kan?m
'inside'
'cloud'
left the mainland before the

indicates that Acehnese
initials took place. Rade, apparently on the edge of the
its own series of developments
left behind, underwent
the evidence
loss of the presyllable

Chamic dialect chain
largely if not totally independent of the other mainland dialects. All the remain?
ing Chamic languages have undergone splits correlated with the manner of the
Chamic
73
Consonants
initial of the main syllable.2 As for Haroi, although subsequent changes have
resulted in the total devoicing of all PC presyllable voiced stops, traces of earlier
split patterns are reflected in the vowel reflexes of the presyllables.
Table
16 shows examples of the split in the reflexes of PC presyllable

which
stops,
depend upon the nature of the main syllable initial. In Jarai,
N.
W.
Chru,
Cham, and PR Cham, if the main syllable began with a
Roglai,
voiceless stop, the voiceless stop reflex is rarely but occasionally accompanied by
voiced
breathy voice on the vowel. If the main syllable began with a voiced obstruent,
the reflexes may be either voiced or voiceless in Jarai, are voiceless in Chru, are
voiced inN. Roglai, are voiceless inW. Cham but with variable residual breathy
on the vowel, and are voiceless
in PR Cham but without any residual
if
voice.
the
main
breathy
Finally,
syllable consonant is a sonorant, the reflex is
voiced in Jarai, Chru, and N. Roglai, and voiceless but with accompanying resid?
voice
ual breathy voice on the vowel inW. and PR Cham.

The same patterns (although based on all the data, not just the examples
in Table 16) are presented schematically in Table 17.
Table 17: The reflexes of the PC presyllable
PC
Aceh.
Rade
Jarai
Chru
*b-
b-
m-
p-;
p-;
p-;
b-
b-
b-
p-
p-
p-
p-pp-
/_vl.
/_vd.
m-
stops
stops
/_sonorants
*d-
/_vl.
/_vd.
d-
b-
b-
p-
p-
0-
t-;
t-;
t-;
c-
t-/;
d-
d-
d-
t-
t-
t-
t c-
t- t-
d-
c-
d-
d-
*g-
k-;
g-
g-
k-
k-
k-
k-
k-
/_sonorants
k-
g-
g-
g-
k-
*j-
0-
j"
j"
t-/;
c-
k-;
j"
p
p
kk-;
g-
j"
p-
p-/;
k-;
stops
p
p-/;
p-
d-
p-/;
p-
b-
k-
PR Cham
p-/;
m-
0-
g-
p-
b-
0-
stops
Haroi W.Cham
b-
0-
stops
NR
m-
/_sonorants
/_vl.
2.
voiced obstruents
k-;
k
c-
k-
kkk k-
c ?-
Tsat has essentially lost all its presyllables,

leaving too little trace of earlier
of
to
the
of
loss
determine
whether
Tsat also split the reflexes
stages
process
of the PC voiced stops.
74
Chapter
The patterns themselves are as described for the examples in Table 16,
although in general terms, the cross-linguistic
patterning of the changes illus?
trates a persistent drift towards voiceless obstruents. The pattern of the drift is
evident in the table: in terms of the manner of the main syllable consonant, the
loss of voicing (and breathiness) ismost favored before voiceless obstruents,
favored before voiced obstruents, and least favored before voiced sonorants.
sonorants
The
The
less
nasals
The reflexes of the PC nasals *m- and *?-, shown in Table 18, require two com?
First, in Rade, the *p- and the *b- have also merged with the *m- giving
ments.
m-.
Table 18: The reflexes of the PC presyllable

PC
Aceh.
Rade
Chru
NR
Haroi
PR Cham
m-
m-
m-
m-
Jarai
*/?-)
*m-
m-
*mata
mata
mata
mata
mata
mata
mita
*manu?
mano?
man??
man??
man??
man??
man??
min??
'chicken'
*muda
muda
mada
mada
mada
mida
mathia
mita
'young'
*?amuk
m-
*m- (and, possibly,
nasal
jamo?
?
*?awa
Second,
ewa
the
near
jawa
nonexistence
jam??
jamo?
jawa
lawa
of
PC
*?-
'eye'
camo?
cawia
indicates
yawa
the
'mosquito'
'breath'
phoneme
was
in various ways
marginal at the PC level, and the reflexes show its dissolution
throughout Chamic. In large part the reflexes of initial *?- have blended with
those of *j-, although there is also merger with the reflexes of *y-. The change of
PC first syllable *ft- to ill in N. Roglai is at least reminiscent of the change of
PMP first syllable *n- to ill in PC. It is worth, however, noting that 'breath,
breath, life, soul' is still manifested as /fiaw?/ 'soul' inHeadley's Western Cham,
suggesting
that thisWestern
Cham
initial still reflects the PMP initial quite faith?
fully.
The liquids
In terms of Chamic
reflexes,
the PC *m-. The distinction
the *r- and the *1- are a little more

between
the presyllable
interesting than
*r- and the *1- has been
Chamic
75
Consonants
totally lost in Rade, Haroi, and Jarai, with the Rade reflexes further merging with
*d- and *j-, and then ultimately disappearing completely.
of PC presyllable
Table 19: Reflexes

PC
Aceh.
Rade
Jarai
liquids
NR
Chru
Haroi
PR Cham
*rata:k
nutuia?
etak
rata?
rata:?
rata:?
lata? -1 rata?
*ribow
ribsa
eb?u
rabau
rabau
rubau
laph?au
ripow
rasa
rasa
rusa
lasa
ratha
rusa
*rusa
'bean; pea'
'thousand'
'Sambhur
deer'
*1
*lima
limAt)
lakoa
ema
rama
lama
lurna
lamia
limi
'five'
eksi
lakai
lakai
lak?i
lik?y
'male; person'
larjet
erjit
rakai
?
larji:?
larjl:?
lagi?
larji?
'sky'
luka
eka
laka
lika-
'wound,
*laksy
*lanit
*luka
onsets:
Presyllable
Within
retention
the history of Chamic,
includes
the
reduction
gradual
laka
scar'
and reduction
the persistent movement
the presyllable:
of
towards monosyllabicity
the
first,
consonant
and
vowel
inventories are progressively reduced and then eventually the entire presyllable is
lost. Acehnese
retains 12 of the 13 distinctions (marginally 14), undoubtedly in
it
because
left the mainland early, thus avoiding the subsequent bilingual
part
contact which would
lead the other languages
toward the reduction
of the initial
contrasts.
Rade displays the most drastic reductions in the presyllable consonantal

In Rade, the original 13 consonants have been reduced to just three. The
onsets.
voiced
apical
and
*b-,
*p-
consonants
have
*r-,
merged,
*1-,
*j-,
and
m-,
becoming
*dthe
have
*t-,
been
*c-,
lost,
*k-,
and
the bilabials
*g-
have
*m-,
merged,
becoming k-, and, the two voiceless fricatives have merged, becoming h- (with
the h- subsequently
in some contexts
(see "The voiceless
disappearing
obstruents" on page 69; Table 14)).
to just six, paralleling
Haroi has reduced the original 13 distinctions
Rade
*t-
in the reductions
and
*c-
coalesced,
itmakes. The *1- and *r- have coalesced,

becoming
(with the h- subsequently
c-;
the
disappearing
*s-
and
*h-
have
in some contexts
becoming
coalesced,
(see Table
becoming
/-; the
h
14, page 71)).
Chapter
76
And, as happens partially or completely throughout Chamic, the voicing distinc?

tion is lost with the voiced obstruents. The obvious parallels between Rade and
Haroi apparently reflect the influence of a period of mutual contact.
Jarai,
like
both
*h-, leaving Jarai with
Rade
Table 20: Retention

PC
Aceh.
*r-
and Haroi,
11 presyllable
Rade
NR
Jarai
Chru
0-
r-
r-
r-
1-
*1-0-
r-
1-
1-
J"
d-
*d-0-
*t-
t-
*c*k-
and
*1- as well
Haroi
r-
W.Cham
c-
1-
1
c
j"
t-; d-
t-; d-
c-
k-
t-
t-
t-
c-
c-
k-
c-
c-
c-
c-
c-
k-
k-
k-
k-
k-
k-
k-
*g-
g-
k-
k-;
k-;
k-
*m-
m-
m-
m-
*p-
p-
m-
p-
*b-
b-
m-
p-;
*s-
s-
h-; 0-
h-
sa-;
s-
h-; 0-
*h-
0-
h-
h-
h-
h-
h-;0-
13
13
g-
m-
m-
pb-
p-;
g-
pb-
s-
p-;
b-
PR Cham
c-
1-
j"
k-;
and
t-; d-
g-
*s-
1-
J"
*j-
as
consonants
in presyllable
and reduction
r-
0~
*r-
coalesces
onsets.
t-
t-/;
t-; t
t-
k-
k-;
m-
mm
p-
p-
p-
p-;
k-;
p
p-
p-;
h hh
h-
Totals:
13
12
11
12
12
In all the languages except Rade and Haroi, at least some of the voiced
coalescing with the reflexes of the voiceless stops in the pro?
stops have devoiced

cess.
Also,
in both
dialects
of Cham,
*s-
and
*h-
have
merged
as h-.
that the table does not include Tsat, which does not have presyl?
let
alone
lables,
presyllable onsets. Tsat has dropped all presyllable consonants
except those that have coalesced with the initial of the main syllable to become
Notice
part of the onset of amonosyllable.

Sporadic
As
"dissimilation"
is quite obvious
within
the mainland
(cf. text above and Table 20), throughout Chamic particularly

Chamic languages after the breakup there has been a ten
77
Consonants
Chamic
of the presyllables. This

dency toward the reduction of the contrasts in the initials
so
not
much
reduction has been accomplished,
through the dropping of initials,
the merger of var?
although this happens occasionally, but instead largely through
ious presyllable initials (cf. the situation in Rade in which some initials have dis?
appeared while others have undergone widespread merger).
in which there
Mark Durie (p.c.) pointed out another set of presyllables
Table
21). These
has been sporadic "dissimilation" of the presyllable initial (see
involve instances where both the presyllable and the main syllable
with /c-/, or, more
begin with the same initial: both begin with /k-/, both begin
the change occurs, the presyllable
both begin with /s-/. When
problematically,
initial usually becomes /t-/, but sometimes it becomes lc-1 and sometimes lk-1. In
each case, the change has occurred sporadically, sometimes happening, some?
times not. As with the changes discussed earlier, the result of these changes in the
reduction of number of patterns found involving the presyllable initials.
dissimilations
to discuss the developments of the presyllable initials in

*k-.
21
Table
of
the examples
one-by-one, beginning with the forms with initial
form
that
of
In those cases where there is a Malay form, the initial
faithfully
form. For the form 'claw; fingernail', most of the
reflects the older Austronesian
It is instructive
form guksa is quite regular

reflect the original PC *k-: The Acehnese
of
reflexes
presyllable initial *k- and
including the voiced initial (see "Acehnese
*c-" on page 70), and most of the remaining forms have also retained a velar
languages
two forms do reflect a change: the Jarai (PL) form takau has a
reflex. However,
lt-1 reflex, and both the Jarai (Lee) form cakau and the Phan Rang Cham (Lee)
cak?w
initials
have
reflecting
Table 21: Sporadic

Malay
kuku
change from
Jarai
Aceh.
PC
*kukow
>
kaki
neither
guksa
lt-1
*kakay
> lt-1
takau
gigi
lt-1
*gigey
> lt-1
gigoa
?
NR
PR Cham Wr. Cham
kukau
kakow
cak?w
?
takai
rak?h -i kaka
?
?
tSgai
*t-.
*?- to ft-/ in initials of the presyllable
takai
nor
?
gaki -v
*kaka:s ?
>
a *k-
kukau
'fingernail'
kakai
takay
kak?h
'claw;
'foot'
takai
kakai)
'fish scales'
takah
?
digai
tafc?y
'tooth'
tag?i
For 'foot', only theAcehnese gaki -v (unless, as the vowel suggests, this is a bor?
rowing from Malay), and one of the twoWritten Cham forms recorded by Aymo?
nier and Cabaton kakai reflects the original *k-. The remaining forms reflect an
initial lt-1. The pair ofWritten Chamic (AC) forms make it clear that both variants
have existed within the history of Cham, and suggest that the *k- developed into
Chapter
78
lt-1 during that time. For 'fish scale', the form in Jarai (PL) should be compared
with Khmer sraka (listed inAymonier and Cabaton). Of the remaining forms all
seem to have retained reflexes of the original *k- except for one of the two Phan
Rang Cham variants. The variant recorded by Lee takah reflects lt-1, not the *k-,
but the presence of both variants within Phan Rang Cham suggests that the *k
variant existed within the history of Cham. Finally, with the voiced velar of
form has retained evidence of the older velar; else?
'tooth', only the Acehnese
where
the reflexes have become
alveolar stops.
can be made about
these changes. First, the

of
Northern
PC:
the
Acehnese,
Jarai,
Roglai, Western
changes postdate
breakup
Cham, and Phan Rang Cham all retain evidence of the earlier *k-. Second, the
Several
are
changes
observations
sporadic
change
it sometimes
appears
to reflect
in nature:
the
becomes
an earlier
*k-,
*k-
only
sometimes,
changes
lc-1 and sometimes

sometimes
a *t-,
and
lt-1. The
sometimes
and when
reflex
it does
sometimes
a *c. Third,
two
words changed almost everywhere while two others only changed sporadically, a
pattern consistent with the existence of a tight-knit dialect chain in which
changes in one dialect often spread to contiguous dialects. Fourth, the apparent
to participate in this change (or, in the change of *c- to lt-1
discussed below (with the marginal exception of the word 'lizard')) suggests that
Acehnese had already the mainland when the bulk of these changes took place.
Consider the forms in Table 22 with initial *c-. With
'grandchild', the
failure of Acehnese
older *c- is still faithfully retained only in two Chamic sources, Acehnese and the
Cham (AC) cacauv (along with the variant with the initial lt-1). Similarly,
Written
with
'chop; strike', only theWritten
Fortunately,
the
Aymonier
and
Cham
Cabaton
(AC) cacauh
dictionary
retains the older initial.
contains
numerous
older
(AC) retains the

'great grandchild',
older initial (along with a form reflecting the more recent lt-1). Finally, the MK
borrowing 'lizard' is interesting primarily because it is a late borrowing but none?
theless it has sporadically undergone the change from *c- to lt-1 inAcehnese
(see
forms. With
again only
theWritten
Cham
in Acehnese)
suggesting that, although the change from *c- to lt-1
to have
have occurred long enough after the breakup of PC for Acehnese
escaped most of its influence, it is still reflected in this late borrowing.
the doublet
must
As with the *k- to lc-1 and lt-1 changes, several observations can be made
about this data. First, the change postdate the breakup of PC, as theAcehnese and
Written Cham doublets make clear, and, if 'lizard' is ignored, it is only theWrit?
to partic?
ten Cham doublets that do so. Second, the apparent failure of Acehnese
the exception of the word
'lizard', suggests that
of this occurred. Finally, the doublets inWritten
that the change occurred after the breakup of PC and spread
ipate in this change, with

Acehnese
left before much
Cham
indicate
throughout
the dialect chain.
Chamic
79
Consonants
Table 22: Sporadic
cucu
Aceh.
PC
Malay
> lt-1
x*cica?
> lt-1
tica?
tac??
?
c??
c?n
cica?;
cacai?;
'great
ticai?
grandchild'
caca?
'lizard;
kac??
tac??
k?rjk??
strike'
'chop,
ticauh
tice?
tace?
cacauh;
tacoh
ticoh
'grandchild'
ticauv
> lt-1
tacoh
CAt
x*cic?t
cacauv;
taco
tico
> lt-1
taco
x*cacoh
cicak
in initials of the presyllable

PR Cham Wr. Cham
NR
Jarai
cuco
x*cuco
cicit
*c- to/t-/
change from
gecko'
Finally, there is a third set of forms which behave in a parallel way but
are
which
only apparent when one looks outside of the Chamic languages, that is,
there is a pair of words in which lt-1 reflects what extra-Chamic evidence makes
clear was an earlier ls-1 (see Table 23). Within
that
attestation
I am
aware
of
for
an
earlier
Cham form turns up unexpectedly,
Written
Table 23: Sporadic

Malay
susu
PC
change from
Jarai
**susow
>
sisir
xxsisi(r)
> xtasi
and,
or
I do not expect to find any.
*s- to ft-/ in initials of the presyllable

W. Cham PR Cham Wr. Cham
?
?
tasau
tisau
kasi-i
there is no
an Acehnese
unless
NR
?
tasi
*s-;
itself, however,
?
tasau
*tasow
Chamic
'breast'
tasou
tathow
?
tasi
'a
tathi
comb'
tasi; tasi? -f
on the basis of the parallels with the forms inTables 16 and 17 com?
Nonetheless,
bined with the external evidence offered by theMalay forms, it is likely that these
forms
were
**susow
and
xxsisi
'comb;
hand
of
bananas'
in post-PC
and
have
simply dissimilated giving the forms found in the tables.

In any case, the "dissimilations" discussed above, despite their sporadic
nature, despite the fact that they postdate the breakup of PC, and despite their
sometimes
idiosyncratic patterning have in common that they are part of the
change toward the reduction of presyllables?a
change that has its impetus sup?
not
the
internal
of
these
by
plied
dynamics
languages but by the structures of the
are
in contact with.
languages that the Chamic languages
The
extra-Chamic
As
the patterns
correspondences
in Table 24 and the examples in Table 25 show, inMalay,

guage clearly subgrouped outside of Chamic, the first syllable consonants
spond precisely with the onsets reconstructed for Chamic.
a lan?
corre?
Chapter
80
Table 24: Reflexes
of PC presyllable
Aceh.
Rade
liquids
Jarai
PC
*1
1-
0-
*r
r-
0-
j
d
*j
j-
*d
d-
Chru, NR
W. & PR Cham
Malay
1
Haroi
c
c
t-;d
j
t-/;
The PC presyllable liquids are retained unaltered inAcehnese. The iden?

tical patterns of reflexes in Chru and Roglai as well as those inWestern and Phan
Rang Cham reflect the subgrouping.
Table 25: Reflexes
Malay
lima
of PC presyllable
Aceh.
PC
*lima
limArj
liquids (examples)
Rade
Jarai
Haroi
ema
rama
lamia
PR Cham
limi
'five'
'
lapar
*lapa
rusa
*rusa
epa
rusa
rapa
lapa
lapa
rasa
lasa
ritha
'
hungry'
Sambhur
deer'
*dara
dara
era
dra
caria
tara
darah
*darah
darah
erah
drah
cariah
tar?h
'blood'
danau
*danaw
dano
enau
danau
caniau
tanaw
'lake'
jarum
*jarum
jarom
er?m
jr?m-v
car?m
car?m
'needle'
jalan
*jala:n
-lu?an
elan
jalan
cal?an
?alan
'road;
dara
The Malay forms, the PC forms, the Acehnese,

all agree in their reflexes.
Main
syllable
onset
path'
and the Phan Rang Cham
forms
consonants
the borrowings have been culled out, the correspondences

among the
are com?
are
onsets
If
these
reflexes
main
straightforward.
remaining
syllable
Once
the reflexes of the presyllable onsets, it becomes apparent that certain

have different reflexes, depending upon whether they were presylla?
ble or main syllable onsets.
pared with
consonants
The
voiceless
obstruents
largely unaltered. Only the *s- in

even
Phan Rang Cham has changed, and,
in this case, it is obvious both from
Western Cham and from the older inscriptional citations that it was an s- at an
The voiceless
obstruents
have been retained
earlier point in Cham.
Chamic
81
Consonants
Table 26: Reflexes
of the voiceless
obstruents
PC
Aceh.
Rade
Jarai Chru
NR
Tsat
Haroi
*p-
p-
p-
p-
p-
p-
p-
p-
t- *t-t-
t-
t-
t-
t-
t-
W.Cham
t-
k-
k-
k-
k-
k-
k-
k-
k k-
*c-
c-
c-
c-
c-
c-
ts-
c-
c c-
*?-
0-
0-;
?-
0-;
0-;
.7.
s-
h-
*s-
s-
s-
s-
s-
*h-
h-
h-
h-
h-
?-
0-;
pt
*k-
.?.
.7.
PR Cham
0-;
.7.
.7.
s-
s-
s-
h-
h-
h-h
th
syllable PC *s- onset is retained as such in

Phan
languages except
Rang Cham, the presyllable PC *s- became h
in Chru and N. Roglai. It is worth noting thatWritten Cham has two dis?
Notice
that, while
the main
all modern
except
tinct representations
for /s-/, with the one Moussay

transliterates as lth-1 being the
older form, and the one transliterated as ls-1 being more recent. Thus, in some
sense, the forms with ls-1 are suspicious; that is, these may eventually turn out to
be borrowed but this remains to be seen.
The apparent two-way alternation in the reflexes of the PC glottal stop is
more a question of orthographic conventions than of reality. In several languages,
a clear reflex of the PC glottal stop occurs intervocalically
in word-medial posi?
tion, where the retained glottal stop is written as either a glottal stop or a hyphen.
In word-initial position before a vowel, nothing occurs in the orthographies but
from the various phonetic descriptions
it seems evident that such forms actually
a
with
not
onset?it
is
indicated in the orthographies.
begin
glottal
simply
To digress for a moment, Adelaar (1988:62) writes that, for his proto
the proto-phoneme
*c is not well-attested,
Malayic,
citing certain words,
a
form
with
the
PC
xcum
and another with
smell'3
'kiss;
ing
counterpart
counterpart
x*picah
'broken;
includ?
the PC
break':
the number of etyma with *c is restricted, and many of theMalayic

lex?
c
are
borrowed. But there is still a number of them
containing
emes
which
3.
hitherto
could not be explained
away as borrowings,
and the
As Blust pointed out (p.c.), if PC 'kiss; smell' is aMK borrowing, its distri?
bution in western Indonesia is interesting. Iban doesn't seem to have it, but as
Blust (p.c.) suggested this would be because Iban is part of the Malayo
Chamic population that never left the southwest Borneo homeland. He fur?
ther notes that Malay, Minangkabau
other languages which may have
Madurese;
Ngaju Dayak
and Banjarese all have it, as do some

borrowed it from Malay
(Sundanese,
presumably got it from Banjarese).
Chapter
82
ultimately depend on the reduction of their

includes e.g. PM *cari, *cu(rj)kup, *curi, *cium. It
number, which
appears that reflexes of *cium and *curi must have been borrowed from
Northern Indian languages, where we find Bengali cum(a) 'a kiss', curi,
of PM *c will
elimination
'theft', and cor 'thief, and Hindi

and chor 'thief. Since other Northern
chuma
cori
'a kiss',
chori
'a theft',
usually have
Indian languages
similar or related terms for 'kiss' and 'theft' or 'thief, and the regular
PMP etyma for these notions are *ajek and *tarjkaw respectively, *cium
and *curi must
be erroneous
reconstructions
built on Northern
Indian
*c, such as *kucirj 'cat', *pacah

'small'
*kacil
be
the result of secondary phonolog?
and
'scattered',
may
ical developments within Malayic
itself (cf. Zorc 1983:12-3). It is very
loanwords.
Other
of PM
instances
on the one hand, and bor?

likely that these phonological
developments
from
and Southeast-Asian
Northern
Dravidian,
Chinese,
Indian,
rowing
on
will
the
evidence for PM
eliminate
the
other,
ultimately
languages
*c. And
of PM *c has far-reaching
the elimination
PAn/PMP
consequences
for
*c.
Of the two PC counterparts he mentions specifically, the first is a borrowing, but

the second reconstructs at least within Chamic. However, as Adelaar observes for
the vast
Malayic,
majority
of
within
the words
with
Malayo-Chamic
are
*c-
bor?
rowed. Only one of the forms with a presyllable *c- looks to be PMP, the word
*caba:rj 'branch', but in PC it is x*ca?arj, with an unexpected glottalized Ibl. Of
the 30 or so PC forms whose main syllable begins with *c, although many recon?
struct
to the PC
level,
other
than
none
x*picah,
appear
to have
strong
PMP
recon?
structions !
The voiced
As
obstruents
is clear from the reflex patterns in Table 27, before it became differentiated
and Coastal Chamic, PC had four voiced obstruents.
into a Highlands
Table 27: Reflexes
of PC voiced obstruents
Chamic
Hainan
Coastal
Tsat
(Lowlands Chamic)
Haroi W. & PR Cham
PC
Aceh.
Chamic
Highlands
(Plateau Chamic)
Rade Jarai Chru NR
*b-
b-
b-
b-
b-
b-
*d*,
d-
d-
d-
d-
d-
phth-
pht
th-
g-
g-
g-
g-
g-
kh-
khk
*j-
j-
j-
j-
j-
j-
s-
?s-
Chamic
We
Tra-ki?u
initials were voiced
know that the Cham
(Coed?s
and
'with',
83
Consonants
(1939; cited inMarrison

'seven'.
tujoh
is no
There
(1975)));
reason
in the inscription found at

for example, di 'at', dengan
to trust
not
the words
that
written
with voiced stops in the script were voiced stops, at least originally.
The real questions revolve around when the various devoicings
took
place. Much of this must be surmised as all we know directly is that, on the basis
of Cabaton (1901), the voiced stops had become voiceless by 1901. The devoic
ing in Tsat and the Coastal Chamic are instructive for two reasons: first, the two
occurred
independently of each other and, second, the reflexes correlate with dif?
ferent linguistic sub-areas. In the southern Vietnamese
highlands, Rade, Jarai,
Chru, and N. Roglai have preserved the original PC voicing; in the Hainan lin?
guistic sub-area Tsat has devoiced, and in the linguistic sub-area along the coast?
line, the Haroi and Cham have undergone devoicing.
to N. Roglai, the devoicing
In Tsat, closely-related
only after the arrival in Hainan in 986, having fled Champa
occurred
probably
following the Viet?

at
northern
The change was
of
the
in
982.
sacking
capital
Indrapura
contact
in
influenced
the
situation
Hainan, as much of
undoubtedly
by
language
the phonological
of
Tsat has been restructured under the influence of the
system
namese
languages of Hainan. Notice that, in any case, the Tsat ismost closely related to
Northern Roglai, a language which still has voiced obstruents; thus, the devoic?
?
?
the separation of Cham and Roglai.
ing in Tsat happened after
long after
It is likely that the devoicing in Coastal Chamic, that is, in Haroi and in
Cham,
postdates
Highlands
of the distinction
the emergence
Chamic.
There
is also
reason
every
between Coastal Chamic
to
suspect
that Haroi
were by then quite separate and that the parallel developments

result
a similar
of
response
to
similar
situations
that
is,
and
in each were
the
type
of
and
Cham
the
conver?
gence typical of a linguistic area.

What must be resisted,
changes
tion must
in Coastal Chamic
be resisted
is the temptation to attribute the

however,
toVietnamese
influence. The reason that this tempta?
is that the same changes are found inWestern Cham, a

split off from Cham immediately after theVietnamese con?
group that apparently

quest of the southern capital atVijaya. Thus, if this historical scenario is accurate,
it is quite unlikely that the innovated changes in Coastal Chamic are primarily
due to Vietnamese
influence. It is possible, however, that the various changes
common
to Coastal Chamic might
occurred before Cham differentiated
be
the modern
reflexes
into Haroi, Western
Cham,
of changes that
and Phan Rang
Cham.
Even more
voiced obstruents
ple,
important to the history of these languages is the fact that the

are associated with breathiness throughout Chamic. For exam?
in the two Cham dialects,
the reflexes are voiceless
unaspirated
stops fol
84
Chapter
lowed by a breathy voiced vowel. Also, for Chru, Fuller (1977:85) writes that the
initial stops are often accompanied by a breathy quality on the vowel. In
Haroi, various vowel splitting patterns make it clear that the voiced obstruents are
voiced
to be associated with breathy voice. And, in the earlier discussion of the reflexes
voiced stops, the patterns of splitting in the reflexes suggest that
breathy phonation is associated with the voiced stops. In short, this and other evi?
of the presyllable
dence makes
it clear that throughout Chamic
voiced
stops are associated
with
breathiness.
The
sonorants
The PC sonorants are retained as such throughout Chamic. As noted elsewhere,

the sonorants behave as a natural class for the sake of certain register complex
and tonal developments.
Table 28: Reflexes of PC sonorants
PC
Aceh.
NR
Tsat
Haroi
m-
m-
m-
m-
m-
m-
m-
*n-
n-
n-
n-
n-
Jarai Chru
n-
n-
n-
W. Cham
PR Cham
m-
n-
*n-
0-
0-
rj-
rj-
0-
0-
rj-
rjrj
*ft-
ft-
?-
ft-
ft-
ft-
ft-
ft-
ft-
1-
ll-1- 1-
1-
1-
1-
11
*r-r-
r-
r-
r-
z-
r-r r-
z-
r*y.
y-
y-
y-
y-
y-
*w-
w-
w-
w-
w-
w-
Also,
*w-,
Rade
*m-
alternate
as might
be expected,
in a fully-predictable
v-
1-
yw-
in various languages
way
between
fi?
-y- and
yw
w-
the reflexes of *y- and

-i- and
between
-w-
and
-u-, respectively, depending on whether or not the reflex has been analyzed as
vocalic or consonantal in the grammar in question (cf. Lee 1974:662, fn. 5).
It is important to note that, while an initial *w- is reconstructable
in
some words for PC *w-, none of the words with an initial PC *w- reconstruct to
the PMP level. The secondary nature of such forms is fully consistent with Blust's
claim that word-initial *w- was lost in proto-Malayo-Chamic.
(1992a:34-44)
Voiced
and voiceless
'aspirated'
consonants
These voiced and voiceless
'aspirated' initials come from two sources. The inher?

ited initials result from the coalescence of the initial of the presyllable with the
Chamic
85
Consonants
syllable (see Table 29, below, for examples). In addition, there are
a number of examples that areMK borrowings, for example, PC *pha 'different'
and PNB *pha 'different' and PC x*khan 'blanket; cloth' and PNB *kh?n 'blan?
*h- of themain
ket'.
In Chamic, although there is some question about whether these are

units or clusters, they often behave as clusters, not just phonologically
but also
like
clusters
in
sound
and
they partake
phonetically. They frequently
morpholog?
ical alternations
in which
the vowel before medial
they are split as if they were clusters (see "The loss of

-h-" on page 63). Of course, such an analysis rests cru?
that alternating forms are synchronically related to one
cially on the assumption

another, an assumption that is clearly true diachronically but may be questionable
synchronically. With reference to Jarai, Blust (1980b: 133) records the alternation
[phi?] [pahi?] 'bitter' < PC *phi? (his own transcription), an alternation that sug?
a cluster. With reference to Acehnese,
as
gests that the initial is phonetically
Durie
(1985:19) makes clear, both the voiceless aspirated and the voiced aspi?
rated (his murmured) stops can be treated as phonemic sequences of C + h. Durie
(p.c.) argues that there are both phonetic (1985:19) and phonological grounds for
clusters. Ace?
treating Acehnese C + h sequences as demonstrably phonological
hnese (Durie 1985:19) nicely illustrates the evidence found in morphological
alternations; cf. /phet/ 'bitter', which, when it takes the infix l-mn-l, becomes
'bitterness'. And, in Acehnese,
this process of infixation is wide?
/puinuihet/
spread enough tomake it clear not only that the infixation itself is productive but
also that these onsets are not just historically but also synchronically clusters.
The diachronic source of such C + h clusters in all those instances where
there is a clear etymology and the forms have not been borrowed is from the
reduction of disyllables.
Table 29: Reflexes of voiceless
PC
Aceh.
Rade
Jarai
Chru NR
*ph-
ph-
ph-
ph-
ph-
*th-
th-
th-
th-
th-
*kh-
kh?
xch-
'aspirated' consonants
and voiced
kh-
kh-
kh-
ch-/;
s-
s-
Tsat
Haroi
W. Cham
PR Cham
ph-
ph-
ph-
ph-
ph
th-
th-
th-
th-
th
kh-
kh-
kh-
kh
ch-
s-/;
?
ch-/;
c-/;
c- s-
s-ch-/;
thy
*bh- ?
bh-
bh-/;
ph-
ph-
ph-/;
ph-
*dh-
dh-
*gh-
kh-?
ph
pah
dh-
th-
th-
th-
th-
th-
th-
th
gh-
kh-
kh-
kh-
kh-
kh-
kh-
kh
Chapter
86
'aspirates' are retained as such throughout Chamic, but

series has been largely lost. Thus, the voiced aspirates are only
retained as such in the orthography of Cham and in Rade, with the *bh- variably
The voiceless
the voiced
retained as /bh-/ in Jarai and the *dh- apparently retained as /dh-/ in Acehnese.
evidence for a former voiced series is found in both Western and
Nonetheless,
Phan Rang Cham, where the reflexes of these obstruents are in the breathy-voice
quality often found on the vowels after the former voiced obstruents; as Lee notes
this is a clear indication that these initials were once voiced
(1974:648-649),
obstruents.
these
In discussing
Lee
aspirates,
also
notes
that
there
are
some
also
in Roglai, but he assumes these are probably Rade borrowings.

It needs to be pointed out that xch- does not reconstruct to the PC level,
nor apparently do the forms containing it occur inAcehnese. Notice the consider?
voiced affricates
able, unconditioned variation in its reflexes.

The process of reducing the PMP
forms to aspirates in PC was not

as
PC
the
Lee
(1974:649) again notes, PMP *tuqah
stage;
totally completed by
'old' is PC *tuha which drops the vowel in Rade kha, sometimes in Jarai taha vs.
some variation
tha, but not in Roglai tuha and Cham taha. Undoubtedly,
forms has existed in Chamic for a considerable period of time.
in such
As pointed out earlier, the development of PC and post-Chamic clusters

consisting of stop plus -h- is an obvious adaptation to the MK languages with
which the PC speakers had come into contact.
Glottalized
The
voiced
obstruents
problem in talking about glottalized voiced obstruents is deter?

mining
they are. Greenberg (1970:134), paraphrasing Ladefoged, men?
tions at least three phonetically distinct possibilities, all referred to as implosives:
immediate
just what
"segments in which the larynx is actually lowered producing an ingressive air

stream upon the release of the oral stop, segments with laryngealized voicing,
lan?
and segments which are preglottalized."
In the literature on the Chamic
these segments are frequently termed "preglottalized". Certainly some
such as the glottalized obstruents of Jarai, are definitely phoneti?
manifestations,
(Bob Blust, p.c.); other manifestations may very well not be.
cally preglottalized
guages,
since
Nonetheless,
the
variants
never
actually
nically and all pattern together in Chamic,

nically
contrast
with
one
another
synchro?
the variants can be treated as diachro
non-distinctive.
The origins
of glottalized
obstruents
The overwhelming majority of PC glottalized obstruents are pre-Chamic borrow?

ings from MK sources; a large number of the remaining forms with glottalized
Chamic
Consonants
87
represent borrowings post-dating the breakup of PC. As an example,

Solnit (1993:109) cites theMK loan PC xtfirj -f 'tube; pipe (for smoking)'. How?
obstruents
ever, the earliest layer of glottalized obstruents was

developed in a small set of PMP forms.
not borrowed
but rather
Over the last forty years, linguists have largely focused their attention on
this small subset of inherited forms with glottalized obstruents. The first author to
comment on the crucial correspondences was Dorothy Thomas, who wrote about
the glottalized lb/ in the Rade, Jarai, and Chru forms for 'hair' (cf. Rade ??k
the Chamic forms to Dyen's (1953) PMP *buhuk 'hair', she
'hair'). Comparing
commented (1963:63), "The preglottalized b of *buhuk reflexes probably repre?
sents a metathesis of the *b and *h."
A little later, Greenberg
(1970:137), in his oft-cited paper "Some gener?

consonants,
concerning glottalic
especially implosives", noting that for
Chamic, Dyen's *h > pre-Chamic *? (and, as will be relevant for other examples,
Dyen's *q > pre-Chamic *h), recognized that the PMP *buhuk had been replaced
by the pre-PC level by *bu?uk. He then stated that when the first consonant in a
alizations
CVCV(C) structure is a voiced stop and the second is the laryngeal ?; the result is
a glottalized voiced obstruent, or, in his terms, a voiced injective.
To the word 'head hair', Greenberg added 'stench' and Lee (1974:649;
653-654) added 'climb', both of which work in a parallel way. Later writers, e.g.
Blust
(1980b) and Solnit (1993), have also cited

approvingly. For the forms, see Table 30.
Table 30: PC glottalized
PMP
obstruents
Malay
?
Pre-PC
*buhuk
*bu?uk
*nahik
naik
*da?ik
*bahu
bau
*ba?u
*?ow
PC
the forms and the analysis
in PMP words
Aceh.
Rade
PR Cham
*?uk-1
?o?
??k
?u?
'head hair'
*cfi?
?e?
dl?
cfi?
'climb;
??u
?ow
'stench'
ascend'
*daqan
dahan
*dahan
*daqis
dahi
*adahi
*bow
bsa
*dha:n
dhuian
adhan
*?adh?y
dhoa
adhei
'stench'
than
'branch'
they
'forehead'
Ignoring for a minute the last two forms on Table 30, several additional com?
ments might be made about the data and its analysis. First, the form for 'climb;
ascend' looks good, despite the PMP initial; if the PMP form is accurate, the *d
is simply a backformation from what was misanalyzed
as a nasal prefixed *dahik.
However, the form does occur in Bahnar (AC) dak; it conceivably could turn out
to simply be a lookalike borrowed from Bahnar. Second, although both Thomas
Chapter
88
(1963) and Lee (1974) mention metathesis,

involved deletion of the first syllable vowel,
in the set. Once
ascend'
seems
the two segments
it is more
likely that the change

particularly if we include 'climb;
are juxtaposed, metathesis
hardly
necessary.
seems to have a glottal stop

Finally, it is necessary to note thatAcehnese
for glottalized
obstruents
forms, but voiced
heading monosyllabic
obstruent reflexes for forms in disyllables or as part of an onset cluster. Further,
reflex
in the case of the Acehnese
rather than first

forms, it is likely that coalescence
was
As
Durie
involved.
out,
syllable dropping
(p.c.) points
although Acehnese
on
some occasions,
of
other
Chamic
does drop syllables independently
languages
is consistent with the dropping in
this is rare?mostly
the dropping inAcehnese
to have con?
the mainland
it
would
be
Thus,
languages.
surprising for Acehnese
in
CV?
this
environment
when other
sistently dropped syllables just
particular
Chamic
languages did not.

In addition, the word
'stench' requires some discussion, as its analysis is

a variant with a voiced rather than a glottalized
the
of
existence
complicated by
initial. The most wide spread Chamic reflexes of 'stench' point to a form with a
initial, but the Acehnese
example points instead to a *b-;
a
existence
of
*bvariant finds support inAymonier
the
patterning,
two
which
lists
variants, one with an initial voiced stop
dictionary,
initial.
Doris
Blood
(1962) also recorded a Phan Rang
glottalized
glottal
aside from the

and Cabaton's
and one with a
Cham variant
that reflects an initial voiced
stop.
turning to the interaction between the first three forms, which con?
tain the PMP *-h- and the two additional forms on Table 30, which contain PMP
Now
*-q-,
it becomes
clear
(1) PMP
(2) PMP
That
is,
the
change
that
*-h->
*-q->
of
the changes
were
chronologically
PMalayo-Chamic*-h->
Pre-PC*-?-
PMalayo-Chamic*-?->
Pre-PC*-h-.
PMP
*-h-
to pre-PC
*-?-
occurred
ordered:
/V_V
before
the
change
of
PMP *-q- to pre-PC *-h-. It is only after these ordered changes, that the loss of
the first syllable vowel led to the development of the Chamic glottalized initials.
Note that this change is shared throughout PC: these forms have the
expected glottal stops as their reflexes inAcehnese
obstruents as their reflexes inmainland Chamic.
and the expected
glottalized
sources have also been suggested for the origins of other Chamic
glottalized obstruents, but as of yet the data has yet to provide any clear substan?
tiation for any of these suggestions. For instance, Greenberg (1970), following
Other
(1963:60), proposed that some Chamic glottalized obstruents might have

developed from the loss of the first vowel in disyllabic forms with a glottal pre
Thomas
Chamic
89
Consonants
syllable onset and a voiced stop main syllable onset. The suggestion itself seems
plausible but the data cited in support of the notion is most likely spurious.
(1970:137) writes, "...from PMP ?ijut) 'nose' Thomas (1963) gives
Greenberg
Jarai, Chru ?dug, Rade ?durj or ?ac(u/;." However, contra Greenberg, neither the
Jarai nor the Chru forms actually occur in Thomas; instead, both Chru and Jarai
have ?ad?tj instead. And, while the Rade forms ?dug and ?adug do occur in Tho?
mas, the crucial Rade form ?dur?, likewise seems spurious as a form for 'nose'. It
does not occur with a glottalized consonant in Lee 1966, Tharp and Y-Bham
?
all of which are sources for Rade. It is,
1980, Y-Chang 1979, or Egerod 1978
of course, possible that the form in Thomas is not an error, but itwould take more
than this one form to substantiate the hypothesis. Another similar spurious pair of
forms also appears to exist in Thomas (1963:66), which lists the forms Idok, Idol
'stand' from PC *do:k 'sit', but which again do not occur in the Rade sources
consulted. Contrary to what has been indicated, the Chamic forms for both PC
*idun 'nose' and *do:k 'sit; stay' show completely regular reflexes of the original
stops, with no evidence of the secondarily-derived
implosion.
as various authors have noted, still other forms with unex?
Nonetheless,
voiced
glottalized obstruents do occur in the Chamic data (see Table 31). Lee
(1974:654) cites two forms with reservations, one for 'open the eyes' and another
for 'pain; ache'. The first, has a strong PMP etymology and patterns regularly in
pected
Chamic
*?la:?, despite
the potentially
irregular vowel
glottalized initial is unexpected and not possible

'pain; ache', it is regular within most of Chamic,
or
the Acehnese
these
forms
seem
to the extra-Chamic
*bulat
Malay
?
PC
glottalized
Acehnese
bhut
*?la:t
?
*lindiq
cabang
to account for. As
for x*pacfi?
but it is not directly related to
Thus,
Lee's
reservations
about
-v?
obstruents
Rade
?
x*ca?a:n
x*dih
cabman,
?
?
x*?arj
PR Cham
?la?
'open
wide'
eyes
ache'
pacfi?
'pain;
ka?an
ca?an
'branch;
cfih
d?h
'sleep; lie down'
??n
??n
'hole;
x*parfi?
*caban
the
However,
well-founded.
Table 31: Unexpectedly

PMP
forms.
Malay
inAcehnese.
tree
fork'
door'
However, other anomalous forms remain in need of explanations. For

instance, the glottalized medial consonant in x*ca?a:g 'branch of a tree' cannot
be explained, if the PMP etymology of the form is reliable. Further, themedial Ibl
form should be a /w/, but it isn't.Within Chamic at least, the form
looks as if always had a glottalized medial consonant: theAcehnese Ibl is the nor?
mal reflex of a glottalized consonant in a disyllabic
of theMalay
Chapter
90
In another form, suggested by Solnit, x*rfih 'sleep; lie down', if it is

to cited provenience PMP *lindiq, the glottalized obstruent needs an
explanation. Elsewhere Blust (1980b: 143) notes two words as possibly PMP in
related
initials. The first, meaning

'sticky, as glutinous
origin but with preglottalized
rice', looks to have been independently borrowed more than once into Chamic
and reflects something like x(ma)klit, x?lit 'sticky, as glutinous rice'; note the
PMP root reflected inMalay pulut. The other root
similarity to the well-attested
is *cfoh 'distant; far', which is regular in Chamic, which has apparent PMP corre?
spondences (cf. Malay jauti), and which has an inexplicably glottalized initial.
intriguing is the suggestion by Blust (1980b) that some of the Jarai
glottalized initials may derive from earlier clusters of homorganic nasal plus stop,
but thus far the crucial forms needed to confirm or disconfirm the suggestion are
More
missing.
there
Finally,
are
glottalized
consonants,
including
various
nasals,
scat?
tered seemingly
(mentioned by
randomly throughout various Jarai dialects
Haudricourt
1950) that require more careful examination. To take a single exam?
turns up in Jarai with a preglot?
ple, the PC form x*?an 'hole; door' unexpectedly
talized nasal, despite
the rest of the reflexes of PC *?- showing up as
at
in that dialect in this particular case, the presence of
least
stops,
preglottalized
Bahnar
and
mag
proto-North-Bahnaric
*qmarj
'door'
suggest
that
the
Jarai
form
is a late borrowing from either Bahnar or one of the Bahnaric languages in the
area. The Jarai data in Lafont's (1968) dictionary include a rich array of oddities
in which homorganic nasals appear sporadically
in lieu of anticipated stops,
sometimes
the
same
and sometimes
preglottalized
remains
way
to be
seen.
In
not. Whether
any
event,
the
this can be explained

alternation
of
in
glottalized
stops and homorganic nasals is found throughout at least the Tai-Kadai

family (cf. the Tai subgroup (Li 1977), the Kam-Sui subgroup (Thurgood 1988a),
and the Hlai (Li) subgroup (Thurgood 1991).
voiced
what is happening in Jarai will doubtless require more

Understanding
an
data. However,
examination of the Jarai dialect forms in Lafont's dictionary
suggests that in at least one of the Jarai dialects the voiced stops may have
become glottalized, as in certain Vietnamese
dialects, and that some dialect bor?
rowing
has
occurred.
Finally, it is likely that some of the forms with glottalized consonants

but nonetheless reconstructed in this work to PC are actually post-PC loans that
are formally
languages
guages
their regular correspondence

patterns in the
Despite
to
the
their
limitation
occur,
highlands Chamic lan?
they
suggests they were borrowed after the breakup of PC, but that
fit the correspondence patterns. See Appendix II for such cases.
undetectable.
in which
strongly
nonetheless
they
91
Chamic
Consonants
Reflexes
of PC glottalized
obstruents
In PC, the correspondences

and their reflexes are straightforward (see Table 32).
The PC correspondences
consist of two layers: an older layer limited to three or
four native PMP words with glottalized
number of MK
borrowings
Table 32: Reflexes
obstruents
inherited from PC and a large

obstruents.
incorporated along with their glottalized
of PC glottalized
obstruent
Aceh.
*?-
?-;b-
?-
?-
?-
?-
?-
?-
?-;d-
rf-
rf-
rf-
rf-
d-
rf-
rf
?j-
?j-
i-
%-
a-
iy
iy
*<f-
x*?j-
Jarai
Rade
-;j-
Chru
NR
W.&PR
Cham
PC
Tsat
Haroi
As Table 32 shows, PC has three glottalized obstruents, which have been retained
as such in the modern languages, except for the reflexes of *?j- in Tsat and Chru.
InAcehnese,
the reflexes of glottalized obstruents in disyllables or in onset clus?
ters seem to be the corresponding homorganic stops, but otherwise they seem to
be a simple glottal stop (see "The origins of glottalized obstruents" on page 86
for discussion of specific examples; this entire section owes much of itsmerit to
email discussions with Mark Durie).
Solnit's
Southern
Vietnam Highlands
sub-area
In the area of Southern Vietnam in which PC and the contemporaneous MK lan?

guages were spoken, there was a contrast between glottalized voiced obstruents4
and plain voiced obstruents. At some point probably postdating the Vietnamese
conquest
of
the
south,
southern
Vietnam
was
split
into
two
smaller
sub-areas.
sub
One,
by Solnit (1993:110) as the Southern Vietnam Highlands
area, is the home of the Highlands Chamic languages (Rade, Jarai, Chru, and N.
Roglai) and Bahnaric languages this area is characterized by contrastively glot?
talized voiced obstruents.
described
The other area, now far more
influenced by Vietnamese,
consists of the
Coastal Chamic languages (Haroi and Cham) and is characterized by the loss of
this contrast. It is that the Coastal Chamic languages share the range of innova?
tions that they do in part because
4.
at one time these were all the same dialect and
As Blust notes (p.c.), preglottalized

stops are also an areal feature in central
Taiwan and implosives are found in a number of languages of northern
Sarawak, and in a region extending from Bimanese
(eastern Sumbawa)
through much
of Flores
in the Lesser
Sunda Islands, and into southeastern
Sulawesi.
92
Chapter
that changes occurred before the original language differentiated

into Haroi,
Western Cham, and Phan Rang Cham and in part because of similar influences.
had most likely moved to Hainan before the distinc?
The Tsat-speakers
tion between
the Highlands
and the Coastal Chamic
came
area, but instead were
in neither
into being, and thus

influenced by the
they participated
linguistic
in Coastal
Hainan
sprachbund. Sometimes
change in Tsat parallel changes
characteristic of Hainan lan?
Chamic, but in each case these are developments
some
at
In
least
of
with
the
addition,
guages.
changes, the similarity is only typo?
as
the
details
differ
logical
significantly.
Table 33: The PC contrast between voiced and glottalized
PC
Rade
Jarai Chru
NR
*b-
b-
b-
b-
b-
*d-
d-
d-
d-
d-
j"
*j-
j"
j"
j"
Tsat
Haroi
ph-
ph-
p-
th-
th-
t-
s-
c-s-
W.Cham
?-
?-
?-
?-
?-
?-
??-
*rf-
rf-
rf-
rf-
rf-
rf-
rf-
rf-
x*?j- ?j-
?j-
i-
?j-
?i-
?j-
?j?j-
the
Coastal
Chamic
PR Cham
*?-
Among
obstruents
languages?Haroi,
rf
Western
Cham,
and
Phan Rang Cham, the voiced obstruents have devoiced, removing the contrast
between the glottalized and the plain series. As Lee points out (1974:653) for
Cham, now that the PC voiced obstruents have devoiced, the old glottalized series
tends to vary freely between voiced and glottalized variants.
This phonetic variation should not be surprising as voiced stops (plo?
and
voiced glottalized obstruents (implosives) are not all that unlike one
sives)
another. Ladefoged
(1971:26-27) writes:
between implosives and plosives is one of degree rather
than kind. In the formation of voiced plosives inmany languages (e.g.,
of the vibrating
there is often a small downward movement
English)
vocal cords. This allows a greater amount of air to pass up through the
The difference
glottis before the pressure of the air in the mouth has increased so much
that there is insufficient difference in pressure from below to above the
vocal cords to cause them to vibrate. An
which
this downward movement
implosive
is comparatively
is simply a sound in
large and rapid.
Beyond the phonetic plausibility, there is another factor that contributes

to the variation. As Lee (1974:653) notes, there is similar variation inVietnamese
dialects, where
the voiced alveolars and bilabial segments are voiced
in some dia
Chamic
93
Consonants
in others. As Henderson
lects but glottalized
variation is characteristic of Southeast Asia.
(1965:422-425)
has written,
such
CONSONANT CLUSTERS
The
languages can be
throughout the modern Chamic
at
the PC level, sec?
into three historical strata: primary clusters inherited
clusters
distributed
grouped
ondary clusters derived from PC disyllabic
Primary
forms, and MK borrowings.
clusters
at the PC level have either *1- or *r- as the sec?
All of the clusters reconstructed
these clusters were already clusters in the lan?

from, or these clusters developed
guage that PC descended
early in pre
Chamic?it
is not possible to tell from internal evidence.
ond member
of the cluster. Either
Table 34: Reflexes
of PC primary
clusters
PC
Aceh.
Rade
Jarai Chru
NR
*pl-
pi-
pi-
pi-
pi-
pi-
*bl-
bl-
bl-
bl-
bl-
bl-
*kl-
lh-
tl-
kl-
tl-
*gl-
lh-
dl-
gl-/;
Tsat
Haroi
pi-
pi-
phi-
pi-
kl-
ki-
kltl-
gl-
dl-
khi-
tl-
W. Cham
PR Cham
pl
pi-
pl
kl
kl-kl
dl
*pr-
pr-
pr-
pr-
pr-
pr-
pi-
pr-
prpr
*tr-
tr-
tr-
tr-
tr-
tr-
ts(i)-
tr-
trtr-
*kr-
kr-
kr-
kr-
kr-
kr-
ki-
kr-
kr-
kr
*br-
br-
br-
br-
br-
br-
phi-
pr-
pr-
pr
*dr-
dr-
dr-
dr-
dr-
dr-
si-
tr-
trtr
Note
that
in Tsat,
the medial
*1- or
*r- has
been
vocalized,
becoming
-i-.
The forms for *kl- and *gl- also merit comment. Lee (1966) reconstructs both
*tl- and *klas well as both *dl- and *gl-, but aside from one word in Rade?the
word for Roglai, the distinction between /gl-/ and /dl-/ seems to be restricted to
Jarai. In Jarai, the variation that exists looks cross-dialectal rather than genuinely
contrastive, including variation within dialects (cf. dleh, gleh 'tired').
it particularly obvious that C + h sequences derive
Acehnese makes
from the coalescence
of what were at an earlier point distinct segments. For
example, both *kl- and *gl- have merged to a single /lh-/ reflex, which Durie still
treats as a synchronie cluster (1985:20). As he notes with regard to not just /lh-/
but to all Acehnese clusters including an Ibl (1985:19) "The Ibl is psychologically
Chapter
94
real to native speakers, who intuitively use h to transcribe these consonants, and it
simplifies the description of the phonotactics."
Even more indicative of the diachronic origins are examples in which
the C + h are morphologically
separable. Durie notes that infixes often split up C
+ h clusters, as with /phet/
'bitterness'
/p-mn-het/
'bitter' which, when

"Voiced
(cf.
and
it takes the infix /-tun-/, becomes
voiceless
on
consonants"
'aspirated'
alternations often
page 84). As both Durie and Cowan note, the morphological
preserve both components of the earlier cluster intact. Durie (p.c.), for example,
notes that the combination of consonants that led to the initial /lh-/ in lhA 'thresh
in
is preserved
grain'
the
form
the same point was made by Cowan
Much
has
which
c-wm-wlA,
the
otherwise
lost
o.
(1981:539):
that, historically
speaking, the initial cluster lh
derives from *sl, *tl [*kl-, in this work], or *cl, clusters that are not
admitted in present-day Achehnese,
but apparently were in an earlier
...it should be noted
period of the language.

nal
consonants
the
that
suggests
language".
The
to which
question
to reappear;
tend
c-wm-wh;lhab
forcibly':
Cowan
In derivations with
the
crucial
I have
'paint,
*kl-,
*sl-,
question,
no
e.g.,
and
*cl-
of
lho
were
clusters
course,
tread
'stamp,
upon
is precisely
something
'swallow'....
s-wm-wlab;lhan
dye':
the origi?
the infix, however,
in "an
how
earlier
much
period
of
earlier?a
answer.
These PC primary clusters are inherited from PMP disyllables in which

the initial of the initial syllable coalesced with the *r- or *1- initial of the main
syllable to form amain syllable cluster onset (see Table 35, for examples).
Table 35: Sources of PC primary
PMP
clusters
PC
Aceh.
NR
blai
*belaq
*blsy
*blah
bloa
plah -i
blah
plah
'chop; split'
*tURun
*trun
tran
trut
tr?n
'descend'
*beRay
*brsy
bri-f
brai
Prgy
*beRas
*bra:s
bnuah
bra
*puluq
*pluh
siploh
sa pluh
*beli
In Table 35, the PMP and theMalay
PR Cham
'buy'
pl?h
counterparts
'give'
-1
prah
'rice (husked)'
'ten'
of the PC primary clusters are
disyllabic.
Secondary
clusters
Although no reduction occurred at either the PC stage, as the PC reconstruction

clusters are scattered throughout Chamic
indicates, various secondarily-derived
Chamic
95
Consonants
resulted from the collapse of an original disyllabic form when the onset of
syllable was *r- or *1- and the vowel of the presyllable was completely
lost. See also the discussion
in "Disyllables with liquids > monosyllables
and
which
the main
on page
clusters"
64.
Table 36: Clusters from
the reduction of PC disyllabic forms
PC
Aceh.
Rade
Jarai Chru
*kar-
kar-
kr-
kr-
kr-
NR
kar-
Tsat
?
*parear-
par-
pr-
pr-
pr-
par-
bar-
mr-
br-
br-
bar-
sar-
*sar-
ar-
tar*mar-
mar-
lar*dar-
hrhr??
hr-
sr-
sar-
Haroi W. Cham
PR Cham
kr-
kar-
kar
par-
par?
?
?
phi-
pr-;
pr-;
par-
par
har-
hr-
par
har?
hr-
har-
har-
0-
har-
hr-
har
mr-
mar-
mar-
0-
mar-
mar-
mir?
car-
?r-
?ar
tar-;
tar
jar-
0-
jr-
jr-
jar-
dar-
0-
dr-
dr-
dar-
s?
car-
tr
*sal-
hi-
*hal*ral-
al?
hi-
sal-
sal-
hi-
hi-
hal-
hal-
hi-
hi-
ral-
ral-
*mal-
mal-
*tal-
tal-
kl-
ml-
mltal-
?
?
?
?
nal-
hl-
hal
hal-
hi-
hal
hal-
ral-
ral
mal-
mal-
mil
tal?
mal-
mal-
tal-
tal-
0-
cal-
tal-
cal-
cal-
?al
cal-
tal-
tal
jal-
0-
jal-
jal-
jal-
0-
*dal-
dal-
0-
dl-
dal-
dal-
0-
*kal-
kal-
kl-
kl-
kal-
kal-
ki-
kal-
kl-
kal
*gal-
gal-
kl-
gl-
gal-
gal
khi-
kal-
kl-
kal
*bal-
bal-
mi-
bl-
bl-
bal-
phi-
pal-
pl-
pil
*pal-
pal-
pl-
pl-
pl-;
pal-
pi-
pal-
pl-
pal
tal-
pal
This widespread reduction largely represents what was historically inde?

In the Rade-Jarai cluster of languages, Rade
pendent but parallel developments.
has clusters in all cases, while Jarai has clusters in all but two cases. In the Chru
Roglai-Tsat cluster of languages, N. Roglai has not reduced any of the forms to
clusters, Chru has reduced some to clusters but not others, while Tsat has only
remnants of clusters and then only in those forms where the presyllable was not
lost. Finally, in the Haroi-Cham
completely
Cham have essentially retained the disyllabic
reduced
some
subgroup, Haroi and Phan Rang

forms, while Western Cham has
to clusters.
Chapter
96
The secondarily-derived
Chamic clusters are still disyllabic not only in
an
extra-Chamic
Malay,
language, and in PC, but also in, for example, modern
Northern Roglai and modern Phan Rang Cham, as the examples in Table 37
show.
Table 37: PC disyllabic forms

Malay
bulu
and Acehnese
inMalay
PC
Aceh.
NR
PR Cham
pilow; palow
tal?m
'body hair'
'inside; in'
hala?
'worm'
*bulow
bulsa
bilau
dalam
*dalam
dalam
ulat
*hulat
ulat
dalap
hu?a?
-lu?an
calan
'road;
jalan
kulit
*jala:n
*kulit
kulet
jalait
kuli:?
kali?
'skin'
malam
*malam
malam
malap
mil?m
'night; evening'
tali
*taley
talo a
talai
tal?y
'rope;
barah
*barah
barah
barah
darah
*darah
darah
darah
tar?h
'blood'
hari
*hursy
uroa
hurai
har?y
'day;
jurup
car?m
'needle'
kar?m
'to hatch'
miry?h
'red'
path'
string'
swollen'
'swell;
jarum
*jarum
jarom
k?ram
*karam
karom
karap
m eran
*mahirah
mirah
mariah
p?rak
xpirak -lf
pira?
surat
*surat
surat
paria? (m) pirak

sura?
har??
sun'
'silver;
money'
'write;
letter,
book'
tendency towards reduction of disyllables tomono?

treatment of disyllables
in
illustrated in the cross-linguistic
The Southeast Asian

is nicely
syllables
Chamic.
clusters
Post-PC
Lee
reconstructed
inmainland
Chamic
four clusters
languages
to the PC
level that subsequent research has

Chamic languages after the breakup of
shown to have entered various mainland

PC:
xcr-;
xjr-,
xsr-,
and
xgr-.
clusters
Table 38: Post-PC

post-PC Aceh.
Rade
xcr
cr-;
xjr
dr-
xsr
xgr
sr- ?
gr -9
hrgr
inmainland
Jarai Chru
tr-
cr
cr
jr
hr
Jr"
jr
gr
gr
srgr-
languages
Tsat Haroi W. Cham
NR
cr
sr
Chamic
si
?
khi-
PR Cham
cr
car
cr
car
?r-;
sr
sr
thr
kr
kr
kr
cr
cr
Chamic
97
Consonants
to the PC level. The first indication of their sta?
These clusters do not reconstruct
tus as post-PC borrowings

is their attestation pattern. Even in the patterns in
Table 38, representative forms are, if not lacking, at least marginally represented
in both Acehnese
and Tsat. The extensive gaps, however, are far more readily
apparent in Table 39, when unlike the typical Chamic cognate, items are fre?
quently unattested in the majority of the languages examined. Bear inmind these
sets are the strongest sets available; items excluded are weaker in one way or
another than those in Table 39.
Table 39: Post-PC
post-PC Aceh.
xcr-
aren
xcrih
xcroh
Rade
cr-;
?
xjr?
crok-f
croh
dr-
jr-
drau
si-
er?
cr?h
'strange'
'a stream'
croh
croh
car-
?r-
?r
'large
sr-
?
sroh
?
caro
sr-
sr??
xsr?p
xsr?h
cru
cru
siau11 cari?u
jro
hr-
-vr
ero
-i
jar'
'medicine'
'rainbow'
thrsr-
sr?u?
thru?
?
xsr? sr?
'to mark'
cr?h
caroh
jr?u
ero
sr-
xsr-
srs
'crossbow'
owe'
'debt,
rice'
'polish
hro?
hrok
sro:?
gr-
gr?
khi-
?
gri?rj
-v
kri?
-f
giarj
? ?
-vf
?
*groh groh
khim11
gram
gr?m
-f
-v
?
xgrak gra?
groh
'dirty'
kr?rj
kf
kr-
-v
kru?
xgrua?
xgram
kr-
xgrit
xgriarj
'subside'
'fishtrap'
?
xgr-
xgroh
PR Cham
cr-
??
xsro:k
W. Cham
jro
xjro
xsro?
(examples)
??
?
?
xjro
xjr?w
-n
ctih
languages
car-
trsh
Haroi
cr-
tr-
Chamic
Tsat
NR
inmainland
clusters
kr?g
?
tusk'
ffang;
'lie
prone'
kriam
kram
kr?m
'thunder'
kri?
kra?
kr3?
'vulture'
/kroh/
kr?h
'to bark'
-v
?
khia55 kr?h -v
98
Chapter
is
Typical of the sets for these clusters are the form 'dirty', which
attested in Jarai and Haroi as well as by a highly irregular form inWestern Cham,
and the form 'lie prone', which, although attested also has a variant completely
lacking the medial -r-, an alternation which is otherwise quite unexpected. Each
of the forms has something that makes it suspect. One of the forms with an Ace?
hnese counterpart is *groh 'to bark' (Durie's proto-Acehnese
reconstruction) but
'to bark' does not reconstruct to the PC level; the two forms in Durie's Acehnese
are dloh and kloh -i; aside from having two variants, the second form has an
irregular initial.
Stronger evidence of the non-PC nature of these forms, however, is
found in three other areas. First, all the more widely attested sets show irregulari?
ties; with the less attested patterns, there are often not enough forms to establish a
regular reflex, let alone an irregular one. Second, the vowel of many of the forms
is either the hi or the hi, both vowels only marginally attested in forms inherited
from PMP, with the overwhelming
attestation in MK borrowings. Third, and
none
of the forms with these clusters is attested in PMP.
quite telling,
The under-representation
inAcehnese
and Tsat along with the strongest
in
the
to suggest thatmany of these
combines
representation
highlands languages
forms were borrowed after theAcehnese had left themainland on the journey that
would
Roglai
eventually take them to Sumatra and after the Tsat had broken from the
and gone to Hainan, fleeing the Vietnamese push to the south.
The most probable sources of these forms are theMK languages, but if
so, the equivalent words in theMK languages have not yet been identified. Thus,
for example, with xsr- clusters, it is evident that both PMnong and PNB have an
*sr- cluster, but to confidently
identify MK as the source requires finding the
same
words.
Still other borrowed clusters exist. In PC, all of the forms beginning with
the clusters
x?l-,
xnr-,
xmr-
x?jr- and
are
see Table
borrowings,
40.
Table 40: Other clusters borrowed after the breakup of PC

W.
postPC
Aceh.
x?l-
bl_._
x?jr_
muir-
xmr-
Rade
Jarai
1-
61-; bl-
n-xnr-?
n-;
r-
Chru
NR
Tsat Haroi
? 61- 611161-;
61-;
n-
rfr_
?jr.
ir_
mr-
mr-
mr-
nr-
n-
ous ways.
In addition,
n-;
r-
Cham Cham
61
n-
n-;
...
?j.;?jr_
mr-
z<
An examination
PR
?jr.
7j_
?jr_
mr-
mr-
mr
*r
of the actual forms shows that they are irregular in vari?

themselves often indicate these are borrowed,
the clusters
Chamic
99
Consonants
not only from the viewpoint that they are otherwise not permitted PC clusters but
also because two of them include imploded consonants.
Table 41: Post-PC
borrowed
Rade
Aceh.
post-PC
clusters
x?jraw
x?jruah
ira:u
-1
'bamboo
?jrau
iruah
jluih;
PR Cham
NR
Chru
drau
?juah
?jrwah
deer'
giron
x61uar
-f
x6larj
blag
x?amra:k
muira?
-v
lar
6la:rj
6la:k
61an
'plains; yard'
amrak
amrai?
amra:?
amra?
'peacock'
amre?
murai
amr?? -f 'pepper; hot'

'thread'
mray
anroik
anon
-v
-v
anoin
-if
x?anrorj
-f
mara:i
enon
x?anro:rj
la
amrs?
mrai
xmray
-i
'to
luar-v
amrec
x?amr?c
xnran
sp.'
'barking
pran
-i
'carry
on
pole'
'numb'
-i
-f
lie'
anro?
-f
aro?
'toad'
surprisingly, not one of these forms has an established PMP etymology. It is

also evident from examination not just of the forms in Table 41 but the rest of the
data that these forms are concentrated among the Highlands Chamic languages,
Not
that is, the languages with
themost
intense recent contact with MK
languages.
CONSONANTS
WORD-FINAL
reflexes of the PC word-final consonants are straightforward. Most of the

and in the older
original finals are preserved in Rade, in Jarai, in Acehnese,
The
inscriptional Cham.
-t and
-p, all,
the mainland,
leaving
In some words,
within
Chamic
all the mainland
languages have lost the final
the Acehnese
only
evidence.
Everywhere
on
there has been a strong tendency for all the stop finals to reduce to
a glottal stop.
Final voiceless
stops and affricates
final stops and affricates are rapidly losing their contrastiveness

throughout Chamic. As Table 42 indicates, the Jarai final glottal stops tend to
have two reflexes: one reflecting the proto-final itself, the other a final glottal
The voiceless
stop.
For
instance,
there
are
two
reflexes
of
*-k
in Jarai:
the dominant
reflex
and a sizable number of glottal stops. Durie (p.c.) notes that in the Acehnese
ditional orthography the final -c is still preserved.
is -k,
tra?
100
Chapter
Table 42: Reflexes

PC
Aceh.
of PC final
Rade
Jarai
voiceless
Chru
stops and affricates
NR
Tsat
-?
-?
-?
Haroi
PR Cham
W.Cham
*-p
-p
-p
-p
-?
*-t
-t
-t;-?
-t;-?
-?
-?
-?
-?
*-k
*. ?
-?
-k
-k;-?
-?
-?
-?
-?
-?
-?
-?
-?
-?
-?
-?
-?
-?
-?-?
*_
-t
-?;
-i?;
-i?
-?
-i?
-i?
-?
-y?
-c
-c/
-?
-?
-?-?
u_
has written, there has been

Throughout Chamic, as Lee (1974:659-661)
a progressive reduction in the number of word-final consonants. Although PMP
had at least some voiced word-final obstruents, no word-final voiced consonants
are reconstructed
with
the mainland
as such for PC. This
is not surprising as no languages in contact

languages have final voiced stops. Similarly, PC *-s
reconstructed for PC, have subsequently merged to final -h
Chamic
and *-h, although

throughout Chamic, except where the *-s is simply lost. As for the liquids, the
final *-r is being lost throughout Chamic, while in Roglai final *-l has become -n,
merging with those examples of Roglai -n that have not denasalized.
The final *-h and *-s
Throughout Chamic, both final *-s and *-h have merged almost completely in the
modern languages, but in two cases they can still be separated by the differences
in the
vowel
accompanying
difference
between
final
reflexes.
*-s
and
*-h
Thus,
after
two
vowels
is reconstructable.
*-u-
and
*-a:-,
In addition,
the
sometimes
the final -s is still preserved in the Acehnese orthography, as Durie notes (p.c.).
After PC short and long *-a-, it is possible to sort out the difference
between final the final *-h and *-s, a gap reflecting an earlier merger within PMP.
Before final *-h the PMP distinction between PAn *-a- < e > and *a merged (Bob
Blust, p.c.). Since the major source of the length distinction in PC between long
and short -a- is the earlier distinction between PMP *-a- > PC short *-a- and
PMP *a > both PC short *-a- and long *-a:-, before final *-h PC did not develop a
distinction
between
*-a-
PC
and
*-a:-.
However, a limited PC length distinction does appear to have developed

before the final *-s. In inherited PC words the PAn *a > PC *-a:-, while it appears
that
the PAn
The
*-a-
> PC
*-a-,
correspondence
although
of
PAn
this
*-a-
is speculative
>
both
without
PC
*-a-
more
and
attestation.
*-a:-
is well
attested, but the number of PAn *-a- to PC *-a- forms is quite marginal with the
attestation possibly limited to two examples: *ma-alas
'lazy' and *dras 'fast'.
For the first, Blust (1992a) reconstructs *males (the -e- is a schwa) for iazy' in
Chamic
101
Consonants
his proto-Malayic;
for the second, he reconstructs PMP de Res, both with a schwa
before the final *-s.
In Table 43, the length contrast before PC final *-s is widely supported:
there are differences between *-as and *-a:s in all the languages in the table. In
contrast, the distinction between PC *-as and *-ah depends upon the two Ace?
hnese examples, which conceivably
Malay influence on Acehnese.
later
could be the result of widely-attested
However, even if the distinction between PC *-as and *-ah fails to hold
sets reconstructed as *-a:s still need to be reconstructed with length:
PC
the
up,
inAcehnese
the diphthongization
clearly reflects an earlier long vowel, and the
as
is
in
Chru.
still
retained
such
Thus, as Durie (p.c.) wrote me, the loss of
length
the final *-s inTsat reflects a two-stage process: first, the final *-s and *-h merged
to *-h, and then the final *-h was
lost after long *-a:-.
Table 43: PC short *-ah and *-as versus long *-a:s

PC
Rade
Aceh.
*-ah
-ah
*blah
plah -i
Jarai
-ah
-ah
blah
blah
egah
*lagah
NR
Tsat
-ah
-ah
-a55
blah
blah
phia*
'chop; split'
'tired'
Chru
ragah
lagan
lagah
khe55
panah
panah
pan?h
na55
'shoot (bow)'
sa55
'wet;
*panah
panah
manan
*basah
basah
masah
pasah
pasah
pasah
*-as
-aih
-ah
-ah
-ah
-ah
*ma-alas
malaih
*dras
draih
alah
?
,55
?alah
alah
alah
drah
drah
drah
damp'
55
'lazy'
sia 55
'fast'
Versus:
-a:s
*bra:s
x*kapa:s
*kaka:s
*?ata:s
-aih
-aih
-a:h
brxuah
braih
brain
brain
gapuiah
?
kapaih
kaih
atuiah
There
with
-tuah
taih
is one more
the discussion
kapaih
?
?ataih
-a33/11
bra
phia11
'husked
rice'
kapaih
kapa
pa33
'cotton'
karkaih
kaka
ka33
'fish
scales'
ata:h
ata
ta33
'far;
above'
set of forms that need to be mentioned
in connection
forms for 'gold'. Forms for 'gold'

are
etnas
the
found
resembling
Malay
throughout Southeast Asia. Within PC,
reconstruction suggests a borrowed form such as xama(:)s, with variable vowel
length. The Acehnese mwih, Chru mi:h, and Tsat ma33 suggest a long vowel,
while Western
of *-as and *-a:s?the
and Phan Rang Cham
suggest a short vowel.
102
Chapter
In a similar way, the preservation of the *-uh versus *-us distinction

illustrated by the examples in Table 44. Although
there is a complete merger
the Bireueng
is
in
dialect of Acehnese
(although not in all dialects), Northern Roglai,

reflexes retain traces of the former distinction between *-uh
and Tsat, the vowel

and *-us in Rade, Jarai, and Chru. The final *-s again produces a final -/- glide in
Rade and Jarai and vowel length in Chru, but is merged in N. Roglai and Tsat,
becoming *-h (the Tsat 55 tone is a reflex of an earlier *-h).
Table 44: PC *-uh versus *-us
PC
Aceh.
Rade
Jarai
Chru
-oh
-uh
-uh
-uh
-uh
-u55
?juh
?juh
iuh
?juh
?iu55
'firewood'
kjuh
tajuh
tajuh
tijuh
su55
'seven'
pluh
pluh
rabuh
spluh
labuh
pluh
labuh
piu55
'ten'
ebuh
phu55
'fall down;
-uih
-uih
-uih
-uh
*-uh
x*?juh
*tujuh
*pluh
tujoh
?
?
*labuh
NR
Tsat
drop
-oh
*tabus
phu5i
?
*fius
*tikus
tikoh
?
*marus
a?ih
?uih
ftiih
??h
?au55 -v
'blow nose'
takui
h
takuih
tukuh
ku55
'rat'
ruih
lak
Cowan
nants
maruh
as also
other
archaisms,
and Durie have noted, strongly suggesting
with
hnese. As
-h,
-y, and
Durie
'itch'
treatment of final *-s is of particular
-s, as well
preserves
'ransom;
save'
kakuih
The Acehnese
script
anchor'
-t, respectively,
has
taken
such
interest. The written
as final
-1, and
that the merger

place
within
the
-c, as both
of these conso?
history
of Ace?
notes
(p.c.), although final *-s has merged with *-h, the

is still preserved in some dialects (e.g. the dialect preserved in the
great Djajadiningrat
dictionary) where before front vowels the *-s has simply
with
but
after
/a/ in particular, the reflex shows up as l-ibl.
Ibl,
merged
distinction
The final sonorants
The PC final sonorants are retained essentially unchanged, with the exception
the loss of final *-l and *-r inAcehnese
and the N. Roglai alternations.
of
Chamic
103
Consonants
sonorants
Table 45: Reflexes of PC final

PC
Aceh.
Rade
Jarai Chru
NR
-n-n
-rj;-k
-n
-n
-rj
-1
-n
-1
-m
-m
-n
-n
-n
-n;
*-rj
-0
-0
-0
-0
*-l
-0
-1
-1
-n
-0
-0;
-0;
-1
-0;
-m
-m
-n
-m
-n
*-r
W.Cham
-n
-m
*-n
-m;
-n
Haroi
-n
*-m
-r
Tsat
-0;
-p
-t
-r
-0
-0;
PR Cham
-m
-rj
-1 -1
-r
-0;
-r
-0; -r
The splits in the N. Roglai final nasals are regular and are discussed in
some detail in the chapter on nasals and nasalization). Durie (p.c.) notes that in
traditional orthography the final -1, and -r are both preserved, e.g. *wil
Acehnese
'circle' is wil
it is
in the traditional writing. In some cases in spoken Acehnese,
and diphthongs but in other instances it shows up as the
lost after front vowels
glide -y; in other dialects a reflex is preserved even after the diphthongs.
Vowels
Chamic
speakers who arrived on the coast of the Southeast Asian main?

land spoke a basically disyllabic language with a relatively modest vowel inven?
The Austronesian
were
tory. There
four
basic
vowels:
*-a,
*-i,
*-u,
*-e
([-a])
as well
as
three
final
diphthongs: *-ay, *-uy, and *-aw; the four vowels occurred in both syllables
the disyllabic forms, while the diphthongs were restricted to the final syllable.
Under
PMP
Chamic
the
a change
languages,
influence
language adopted
that
of
intense
contact
the main
is reflected
in the
syllable
contrasts
with
MK
languages,
this
of
pre
stress of the neighboring MK

between
the vowel
invento?
ries of the pre- and the main syllable. Unlike the PMP disyllables where there
was a balanced four-way vowel contrast in both the syllables, in PC the vowel
inventories were anything but symmetrical: in the unstressed PC pretonic sylla?
ble, the four-way PMP vowel distinction has been retained in the presyllable, but
syllable the four-way distinction has been expanded to
so
or
distinct vowels, not counting length contrasts.
roughly 18
Some of these new main-syllable
vowels developed out of splits of
inherited PMP vowels, but the bulk of the forms with new vowels are found in
in the stressed main
borrowings from MK. Thus, the main vowels of PC include two

historical layers: those vowels inherited from PMP, which
readily-discernible
form the core of the basic vowel system, and those vowels which primarily reflect
pre-Chamic
occur in pre-Chamic MK borrowings.

influence and overwhelmingly
the two layers are not always completely
While often distinguishable,
distinct: sometimes the phonology of theMK borrowings matched the phonology
MK
lexicon, making the borrowed form indistinguishable on purely pho?

grounds from inherited PMP forms; undoubtedly, sometimes the pho
of the PMP
nological
104
Chamic
105
Vowels
nology of the MK borrowings was restructured by the pre-Chamic speakers to

match the phonology of the PMP lexicon, again making the forms blend phono
logically with the inherited PMP forms. In other instances, new phonological
contrasts entered pre-PC with theMK borrowings.
contrasts, the majority of the
Among forms carrying new phonological
are identifiable as MK
loans into pre-Chamic, while the bulk of the
remaining forms are potentially of MK origin as they lack etymologies, PMP or
words
of forms containing new vowels

in
among the MK forms are also one or
usually sprinkled
Two stages
two words with straightforward, well-attested
PMP etymologies.
appear to have occurred with such words: First, theMK contact led to the devel?
otherwise. However,
are MK borrowings,
although
the preponderance
of a vowel distinction already present in the phonet?

opment and phonemicization
ics of the PMP forms. Second, the development of the new sound in an PMP
form would have significantly lessened the need to restructure the incoming MK
loanwords containing the new vowel. Thus, when MK loanwords with new loan
phonemes were encountered, the new phonemes were borrowed along with the
loanword.
THE LITERATURE
between Cham
stopped at noting and commenting on the resemblances

and other languages?sometimes
showing great insight, the sci?
entific
comparative
While
earlier writers
study
of
Chamic
vowels
can
only
be
said
to have
begun
with
the first attempts to establish sound correspondences.

Sound correspondences
between PMP and PC were first given by Pittman (1959), followed by Blood
(1962), and Thomas (1963), and later by Dyen (1971a). In addition, the sound
correspondences
within
the modern
but also Burnham
Chamic
languages were discussed by Lee

and others. With these works, the
(1966), primarily,
(1976),
comparative study of Chamic began.
Now, in light of the work of Headley (1976) and of several writers who
have tentatively reconstructed subgroups of MK languages of Vietnam it is usu?
ally possible
to sort out theMK borrowings from the native PC material, and, as a

it is now also possible to revise these earlier analyses. For the iden?
consequence,
tification of MK words, in addition toHeadley, there are reconstructions available
of two branches of Mon-Khmer
found inVietnam. Within the Bahnaric branch,
reconstruction work on three subgroups have been carried out: South Bahnaric:
subset of South Bahnaric)

1967,
1987; H. Blood (on the Proto-Mnong
1974; North Bahnaric: Smith 1972; and West Bahnaric: Thomas
(in
Two
have
been
reconstructions
the
Katuic
branch
of
1995).
Prachacakij-karacak
Efimov
1968,
106
Chapter
done: Peiros
1996 and Thomas 1967. These works make possible the identifica?
features often
tion of a large number of MK loans while particular phonological
or
as
to
it
them
either
make
classify
post-PC loans.
possible
pre-Chamic
Further, the expansion of the data has led to numerous revisions of indi?
lexical
vidual
reconstructions
and minor
in the overall
modifications
vowel
inventory, although Lee's basic outline is still quite workable today. In addition,
of course, more recent studies of various Chamic
languages has also led to
increased understanding of PC vowels and their correspondences
(Edmondson
and Gregerson
1991; Friberg and Kvoeu-Hor

1977; Thurgood 1998).
As for PC vowel correspondences,
only Lee (1966) actually set out the
main vowel correspondences,
but he based his analysis on a corpus with numer?
ous unidentified MK words. As a result, over 10% of his reconstructed forms are
actually MK borrowings, many of them borrowed after the breakup of PC. Once
these are identified some of Lee's marginally-attested
vowel correspondences
either disappear or merge with other better-attested correspondences. Another
adjustment to Lee's scheme involves the reanalysis of his treatment of nasalized
vowels. Lee reconstructed a number of nasalized vowels in PC to account for the
failure of certain Roglai word-final nasals to denasalize. However, once the post
borrowings are removed from the data base, it becomes clear that the
PC MK
of the Roglai
majority
vowels
nasalized
changes
in PC.
Finally,
are internal to Roglai,

as
the
result
of
the
leaving only a handful of

identification
of
loans
and
in the analysis of countless individual words, Lee's *o and *o have

both been reconstructed as just a single *o, in part by assigning some *o reflexes
modifications
to *a,
but
occasionally
by
assigning
some
elsewhere.
This
accumulation
of minor
revisions has resulted in a modification
of the general scheme in Lee 1966.

As is true throughout this work, in the discussion
to follow four major
time divisions are distinguished:
(i) the PMP period predating contact with MK
languages the pre-Chamic period, (ii) the period in which early contact occurred
predates what we reconstruct as PC, (iii) the stretch of time during

which what we reconstruct as PC was spoken, and (iv) the period following the
but which
breakup of PC, including
the modern
period.
THE PC PRESYLLABLEVOWELS
The fully disyllabic forms of PMP were radically restructured in pre-Chamic. In
fact, by PC the reflexes of these once truly disyllabic forms had come to consist
of an unstressed atonic presyllable followed by an enriched, stressed tonic sylla?
ble. Under the influence of contact with neighboring MK languages, pre-Chamic
stress, leading to the proliferation
developed word-final
to the reduction
stressed final syllable and eventually
unstressed syllable.
of contrasts
in the
of contrasts
in the
107
Vowels
Chamic
This
of the first syllable of the PMP disyllable to an atonic

languages is reflected both in the fact that the
reduction
inmost
of the modern
presyllable
presyllable has become exclusively CV in structure and in the widespread reduc?
tion in vowel contrasts. Both the first syllable of the PMP ancestral disyllable and
the first syllable of PC had a four-way contrast, but this four-way distinction is
remaining Chamic languages have essentially
only preserved inAcehnese?the
lost most of the vowel contrasts, at most preserving a three-way distinction in
contexts.
specific
still preserves a four-way contrast, many of the

And, while Acehnese
other individual daughter languages have proceeded further along a path toward
with the directionality and the end point of these changes often
monosyllabicity,
so
not
much
by internal forces within the phonologies of the daughter
provided
structures of the languages
languages but by the phonological

were in contact with.
these languages
the four-way distinction, as is evident in the compari?

data with, for example, PMP orMalay, but, within mainland
PC still maintains
son of the Acehnese
a three-way,
Chamic,
than
rather
contrast
four-way,
can
only
be
partially
recon?
structed and then only in certain contexts. Some evidence is found in Northern
Roglai and some appears to exist in early inscriptional Cham, with additional bits
and pieces of evidence found scattered elsewhere. The evidence provided by the
existing forms in the various daughter languages must, however, be used with
some caution as there have been several internal realignments of the presyllable
vowels both in Roglai and in Cham, often giving the illusion that vowel contrasts
have been preserved but with the vowels
secondary
Thus,
developments.
as
in question actually being the product of
comparisons
among
PMP,
Acehnese,
and
clear, in the mainland Chamic languages, the presyllable

vowels have undergone secondary shifts and reductions that sometimes totally
obscure the nature of the original vowel.
The Written Cham data has undoubtedly preserved more information
mainland
Chamic make
than is contained
in this work, asWritten
Cham records contain the earliest Aus?
recorded,
require considerable philological work and
careful diachronic analysis. Perhaps some of this has been done already and I am
simply unaware of it. In any case, it is unclear how much would be learned from
also preserves the PMP four-way vowel
such philological work since Acehnese
tronesian
but this would
distinction, so such work would

what is already known.
Preservation
of the original
very likely do little more
PC four-way
than simply confirm
distinction
Chamic languages, the four-way contrast is still largely

a
as
preserved
three-way distinction after initial glottal stops, and remnants of the
In the modern mainland
108
Chapter
system are preserved inRoglai and Cham, but in these languages subsequent sec?
ondary vowel shifts have obscured what little remains of the original vowels.
glimpses of the original vowels have been preserved as the result of
inwhich the first syllable vowel shifts to the main syllable.
metathesis
sporadic
Occasional
Preservation
Inmainland
after initial glottal
stops
Chamic, there is a three-way distinction between *i, *u, and *a (with

of PMP *a and *a) still largely preserved in both Roglai and Cham
the merger
after an initial glottal stop.
Table 46: The preservation

Malay
PC
ikan
*?ika:n
hidung
ikat
*?idurj
*?ikat
urat
ular
akar
orang
Aceh.
of PC first syllable vowels after glottal

Chru
NR
PR Cham
akain
ika:t
ikan
'fish'
idorj
ikat
adurj
id?k
it?n
'nose'
aka?
ika?
ik??
'to tie'
*?urat
urat
ara?
ura?
ur??
'vein,
*?ular
uluia
ala
ula
ula
'snake'
*?ugha:r
ukhuia
akha
ukha
ukha
'root'
*?ura:rj
-1 unuarj
ara:rj
ura:k
urag
'person;
'child'
tendon'
someone'
anak
*?ana:k
anu??
ana:?
an?:?
a ni?
atas
*?ata:s
atuiah
ata:h
ata
at?h
'far;
asah
*?asah
asah
asah
asah
th?h
'sharpen'
asap
*?asap
asap
asa?
asa?
ath??
'smoke
apui
apuy
'fire'
'the wind'
'ghost;
api
apui
initials
*?apuy
apui
angin
*?arjin
arjsn
arjin
arjin
an?n
hantu
*?antow
uintsa
atau
atau
atow
above;
long'
(of a fire)'
corpse'
in Table 46 reflects the original PMP first syllable vowels

The Malay
well,
fairly
although the first vowel of orang comes from an earlier lui. It is also
worth noting that, while the Roglai data appears to reflect both the PMP and the
PC distinctions
rather faithfully, several of the Cham reflexes also have another

reflex, one reflecting the older vowel and the other reflecting the vowel's subse?
quent reduction to /a/.
Preservation
in other phonetic
environments
The PMP four-way distinction between *u, *i, *a, and *a has been preserved in
(three of these are reflected inTable 47), but is reduced to
Malay and inAcehnese
Chamic
109
Vowels
a three-way distinction
of *a and *a to *a,
probably phonetically a shwa in the onset syllable. This preservation is somewhat
sporadic, with much of the crucial evidence coming from much, much earlier
Written Cham records and inscriptions.
inmainland
Table 47: The preservation
PC
Malay
PR Cham
dikit
*dikit
dit -mv tiki:?
timun
*timun
timon
*gigey
gigoa
tulang
kuku
*tula:g
turnia
*kukow
rusa
*rusa
huma
bunga
kutu
the merger
of PC first syllable vowels
NR
Aceh.
Chamic with
turn?n
digai
in other environments
Wr. Cham
taki?
dikik
'few; little'
tarn ?n
tamun
'melon'
tak?y
tag?i
'tooth'
tulag; talag
kukau; kakou
'claw'
tulaik
talag
guksa
kukau
kakow
rusa
rusa
ritha;
*huma
umAg
huma
hamu -v
*buga
bugog
bug?
pigu
buga; bag!
'flower'
gutea
kutau
katow
katou
'head
*kutow
ratha
rusa;
'bone'
'
rasa
huma; ham?
deer'
Sambhur
'field'
louse'
Notice
that a three-way distinction is partially retained in Roglai, and
almost totally lost elsewhere inmainland Chamic. Despite the occasional illusion
of having kept an original PMP vowel in the first syllable, modern Phan Rang
Cham has virtually lost these distinctions. In fact, the vowels in all, not just some,
of the Phan Rang Cham forms are the result of reduction?in
most instances,
they
have
an
/a/
reflex.
In other
cases,
they
are
the
result
of
a secondary
develop?
these cases, they have an HI reflex. Modern Phan Rang Cham does have
its own distinctions among these vowels, but these distinctions are the result of
ment?in
various
secondary
developments.
In contrast, theWritten Cham forms, particularly the forms recorded in

the dictionary of Aymonier and Cabaton (1906) seem to preserve a great number
of older vowel distinctions. However, this material must be used with great cau?
tion. The entries reflect various
consequence,
of variation,
the Aymonier
some of which
time periods and sources blended together. As a

and Cabaton Written Cham material has a great deal
can be made
sense of and some of which
is simply
quite puzzling.
Sporadic preservation
of/u/through
metathesis
Aside
from what has been preserved inRoglai and Cham, there are also a handful
an original first vowel lui was preserved through sporadic met?
athesis (m).
of cases in which
110
Chapter
Table 48: Sporadic metathesis
*u
of first syllable
PC
Aceh.
Rade
Jarai
Chru
NR
*hursy
uroa
hrue (m)
hrai
harai
hurai
'day; sun'
hama
huma
'field'
habai
habuai
ara?
ura?
*huma
umAg
hama
hama
*hubsy
ubi -f
habsi
habai
*dursy
duroa
erue(m)
drai; trai
*?urat
urat
aru?t
?ara?
*hulat
ulat
hlu?t (m)
(m)
(m) 'taro; yam'

druai (m) daruai (m) 'thorn'
-v
hlu?t (m)
hala?
'vein'
hula?
'worm'
Note
that in Table 48, there is sporadic metathesis of first syllable *u, resulting in
a lui showing up in the second syllable, often along with the expected vowel
reflex. Although
they are not included in the table, in a number of cases, the lan?
in
variant. The directionality of
guages
question also have a non-metathesized
the forms are compared with
these changes is particularly clear when

metathesized Acehnese
forms.
in the three-way
shifts
Various
In addition
to the massive
distinction
reduction of the four-way distinction in first syllable

of Chamic, at times in Roglai and at times in Cham
throughout much
there have been secondary vowel
original PC vowel.
vowels
secondary
Independent
shifts further hindering
shifts in Roglai
of
an earlier
vowel
to lui
the identification
of the
and Cham
Two separate shifts are illustrated in Table 49. In Roglai,

the change
the non
after
*r-,
*1-,
*m-,
one well-attested
*t-
and
before
shift is
a bilabial.
Often Roglai
has variant forms, one with
the secondary
lui and one without;
occasionally,
one
data
by Awoi-hathe
variant
shows
up
in
the
recorded
et al.
while
another shows up in the material collected by Lee. An additional compli?

cating factor is the apparent sporadic vowel harmony in Roglai, in which the pre?
syllable vowel assimilates to the lui in the main syllable, as in 'needle', 'that',
and so on (see Appendix
II).
The other secondary development
illustrated by much of the same data
in Table 49 is the change in Phan Rang Cham of a presyllable vowel to HI after
either *r- or *1-. As with any vowel in the presyllable, there is variation between
HI and the reduced /a/. This Cham development, however, is only a part of amore
general change in Cham (see "Secondary shifts in Cham" on page 112 just
below).
ently,
Cabaton
at that
time,
(1901) noted
that the first vowel
in ribut
'storm' was,
appar?
a barred-i.
111
Vowels
Chamic
shift to /u/ and the Cham shift to /if
Table 49: The Roglai
Aceh.
PC
Malay
lapar
*lapa
l?mbu
x*lamo
labuh
PR Cham
NR
lapa
liiimo
lamo
*lama:n
lum?n
*labuh
labuh
l?mak
x*lama?
luim a?
l?bih
*labsh
ruibsh
-n
Wr. Cham
lipa; lapa
limo; lamo
lip?; lapa
'hungry'
lamau
'cow;
limin
limin; lam?n
libuh; labuh
'fall
down'
lima?; lama:?
'fat,
oil'
ox'
'elephant'
lum??
lipuh; lapuh
liml?; lam??
lubeh
lipsh; lap?h
lubaih; lab aih
'more'
-v
lipan
*limpa:n
limpui
an
lupa:t
lipan; lapan
lipan; lapan
'centipede'
mimpi
luka
*lumpsy
*luka
lumpoa
luka
lupai
lip?y; lap?y
lika-lik?h
lip?i; lap?i
lik?; luk?
'to dream'
lipag; lapag
li??ng; lubag;
'hole; pit'
ripug; rap?g
rabung
lubang
*luba:g
luba:k
r?bung
*rabug
rubuk
'wound'
labag
'bamboo
shoot'
nipis
lima
*lipih
*lima
lipeh
ribut
x*ribu:?
ribu
*ribow
mamah
*mamah
babah
*mabah
*mamih
limAg
lupih
lum?
lip?h; lap?h
limi; lami
rubu:?
ripu?; rapu?
ribea
rubau
mamAh mum?h
?
mubah
mamsh
mum?h
ripow;
rapow
mimih
'thin'
lapih
limi; lima
rabu?
ribuv;
ribou
babah, pabah
mimih,
'storm'
'thousand'
'chew'
mimih
papah
mimih
'five'
mamih
'mouth'
'sweet'
timun
*timun
tum?n
tamun
tamun
'melon'
tumbuh
*tamuh
tum?h
tam?h
timuh; tamuh
fto grow'
d?pa
*dapa
tupa
tapa
dap?
'armspan'
Roglai
of what
1. As
and
tim?n
In another subset of forms, various presyllable vowels have become ill in

but without more examples even a general idea
before certain alveolars,
to figure out.
is conditioning
this change
is well known,
in a large number of An
/n/are
is difficult
languages
It/ is dental while
alveolar.
Id/
112
Chapter
shift to /i/ before an alveolar
Table 50: The Roglai

PC
Malay
Aceh.
NR
PR Cham
Wr. Cham
darah
pat?y
tar?h
'banana'
*darah
pitai
darah
pat?i
darah
darah
'blood'
dada
*dada
dada
dada
tata
dada,
dalam
*dalam
dalam
batuk
*batuk
bato?
*gigsy
*bitu?
gigoa
*batsy
bini
tada
'chest'
t-al?m
dalam
'inside;
pitu?
pat??
batuk
digai
tak?y
tagsl
'cough'
'tooth'
pitu?
binai
pit??; pat??
batuk; butuk
'star'
dalap
in'
pinay
binai
bulu
*bulow
bulsa
bilau
pilow; palow
bulou; baluw
'virgin'
'body hair'
bulan
*bula:n
buluian
bila:t
'moon'
g?tai
pilan
kat?l
bulan; balan
katan
kata?
kat??
gatal; katal
gatak; k?tak
'itchy'
'sap; resin'
*binay
gatal
*gatal
*gatak
The
interaction
of
reconstruction
of presyllable
non-Acehnese
Chamic
Acehnese
data
and elsewhere
Secondary
these various
vowels
alone.
the
secondary developments makes
tenuous at least on the basis of
extremely
since
However,
outside of mainland
Chamic,
these
are
in
well-preserved
it is of little consequence.
shifts in Cham
In Phan Rang Cham, there is a somewhat similar shift from various presyllable
vowels to lil, but with several more conditions than in the Northern Roglai exam?
ples: the initial must be a bilabial stop and the following main syllable onset not
only must be an alveolar but it also must be either a sonorant or be a sonorant
descended
from a voiced
obstruent
(and, thus, the foUowing
vowel has breathy
voice).
Table 51: Shifts

Malay
PC
in Cham
Aceh.
to/i/before
Rade
*bara
mra
*banut
man?t
alveolar s
W.Cham
pra
bini
*binay
bulan
*bula:n
buluian
m?an
ea
bulu
*bulow
bulsa
ml?u
plau
manie
biji
x*bijsh
babi
*babuy
bijsh
bui
*babah
babah
mjsh
panai
plan
PR Cham
pira
'shoulder'
pinu?
'banyan,
pinay
'virgin;
woman'
pilan
'moon;
month'
balete'
pilow
'body hair'
pacsh
pa?sh
'seed'
papui
papuy
'wild pig'
papah
papah
'mouth'
Chamic
113
Vowels
This change, like many of the secondary shifts of presyllable vowels discussed in
this section, seems to be relatively recent as only a small number of these second?
arily-derived vowels are inAymonier and Cabaton's 1906 dictionary, but they are
rather common inMoussay's
1971 dictionary.
THE INHERITEDPC MAIN SYLLABLEVOWELS

system developed out of an PMP system with four basic vowels
and three diphthongs. The typical morpheme was disyllabic, with each of the four
vowels occurring in either syllable, but with the diphthongs restricted to the final
The PC vowel
syllable.
Figure 8: PMP second syllable vowels > PC main syllable vowels
PMP second syllable vowels
PC main
^U^j
-U
-i ?
*-ij
*-a
*-a-<e>
syllable vowels
>
*-uu
*-sy
>
*-ow
>
*a *a (short) / *-a:
*-ay
*-aw
*-uy
*-ay
*-uy
*-aw
The transition from PMP to PC vowels was not particularly complex.

Both PMP high vowels split, becoming diphthongs
in final position (at least
when lengthened by stress), but remaining unchanged in closed syllables. The
PMP shwa, which never occurred word-finally
in PMP, everywhere ceased to be
a shwa; in all contexts, it became PC *a. The old PMP *a reflexes split before
certain finals, with some reflexes of PMP *a becoming PC *-a- and others
becoming PC *-a:-, thus introducing in those positions a PC length distinction. In
PC, the old PMP shwa disappeared as such, although a new shwa appears in the
PC inventory brought in by pre-PC borrowings from MK. As for the old PMP
diphthongs, they entered PC unchanged.
The path from PMP to PC was illustrated in Figure 8. The next stage
the breakup of PC into its various daughter languages, including Ace?
a
hnese,
stage which ismore complicated than the transition from PMP to PC and
cannot be fully represented in Table 52 (cf. Figure 12 and accompanying text).
involves
this table introducing the vowels is somewhat simplified, it is a

good representation of the basic vowel reflex patterns, with several qualifications.
First, in two languages, the vowels correspondences are too complicated to repre?
sent clearly on the table. The vowels of Haroi, which have registrally-induced
Although
114
Chapter
splits, are discussed later (in "Haroi vowels and restructured register" on
page 197). The vowels of Acehnese, which often have multiple reflexes due to
vowel lowering induced by nasalization, are also treated later, as the correspon?
dences (Table 52), although regular, are somewhat over-simplified.
vowel
Table 52: The basic main syllable vowel reflexes

PMP
*i
PC
*-sy
Aceh.
-oa
*i-
*-i-
-e-;
*u
*-ow
-sa
K-u-
Rade
-s-
-o-
-o-;
*. a-
*-uy
tua
*a
a;
*-uy
-ui
*-ay
*-ay
-e;
*-aw
*-aw
-o;
Second,
?
-o
Jarai Chru NR
Tsat
W. Cham PR Cham
-si
-ai
-ai
-ai
-ai
-i-
-i-
-i-
-i-
-i-
-au
-au
-au
-au
-au;
-u-
-u-
-u-
-u-
-u-
-u-u
a -a-
-a-
-a-
-a-
-a-
-a
-a-
-ay
-i
-(i)a
-sy
-i-au
-ow
aa
-ui
-ui
-ui
-ui
-ui
-ui
-ie
-ai
-ai
-ai
-ad?42
-ai
-ay
-au
-au
-au
-au
-au
-au
-aw
the crucial
issue
of
vowel
length,
of
considerable
-uy
importance
to
both the analysis of borrowing and to PC subgrouping questions, is given a long

and detailed treatment later ("PC vowel length" on page 138). Third, various
minor subpatterns are treated elsewhere as they come up, including the sporadic
metathesis
found scattered throughout Chamic,
the Western
and Phan Rang
Cham reflexes of nasalized reflexes of PC *a found inWestern and Phan Rang
Cham, and the interesting reflexes of PC *-ay and *sy found in Rade.
The splits
in the PMP high vowels
*i and *u
(1963) laid out the basic patterns for the splitting of the PMP high vow?
els into two conditioned reflexes: For both high vowels, her tables make it clear
Thomas
that in open syllables (and before -h), the PMP *i and *u lengthen in open sylla?
bles and subsequently diphthongize, while in closed syllables (except before -h),
they remain high monophthongs.
An examination of Thomas' distributions, suggests that except for the
forms with final -h, the splits correlate with whether or not the forms occur word
finally. And, in fact, the final -h in these forms is spurious. Thomas's analysis of
the patterns was obscured by a spurious final -h inDempwolff
s PMP reconstruc?
tions (Dempwolff s "spiritus asper"); once Dempwolff
s spurious final -h is elim?
inated from Thomas' reconstructions,
her description of the split is accurate.
Chamic
115
Vowels
Thomas was
Despite using Dyen (1953) in which Dyen corrects Dempwolff,
s spurious finals since, for the forms not found
nonetheless misled by Dempwolff
s reconstructions.
inDyen, she reverted to Dempwolff
The diphthongization
patterns
cases will be discussed in this chapter, but the first to

Several diphthongization
occur historically
and thus the most central is the chain by which the high,
stressed *i and *u of PMP first lengthened under stress, becoming early-PC *-ij
is apparently preserved as such in the older Written Cham
and *-uu (the *-uu

as <
records
-uw
>),
on
going
<
to become
-sy
>
<
and
-ow
>
in later Written
(with the /-si/ also preserved as such in Rade), next becoming /-ai/ and
/-au/ (preserved as such in Jarai, Chru, and Northern Roglai), and then ultimately
and /-?u/ inmodern spoken Phan Rang Cham and Tsat.
ending up as l-?.%1
Cham
was restricted to the high, stressed vowels, and, as

the stressed vowels became long before diph?
emphasized,
Cowan
(1974:189)
thongizing. The unstressed high vowels did not lengthen, and thus did not partic?
ipate in the changes. Acehnese
provides some clear examples of the distinct
treatment of stressed and unstressed vowels. In particular, Acehnese has several
reflexes of PMP *aku T:

a diphthong, while
the vowel
the vowels
in the stressed independent morpheme kso is

affixes, the prefix ku- and the
in the two unstressed
suffix -ku(h), have remained undiphthongized.

The initial part of this diphthongization
(1985,
and,
p.c.)
as upgliding:
diphthongize
respectively.
vowels
*-ow,
color, that is, the *-sy

*-aj
and
*-au,
the
tendency
for
chain reflects what Donegan
tense
vowels
the PMP *-i and *-u become

both onsets underwent
Subsequently,
and
*-sy
ducing
as
describes
respectively.
Next,
the
loses its palatality, while
to lengthen
then
*-ij and *-uu,

leading to the PC
early-PC
lowering,
onsets
and
of
both
vowels
lose
their
the *-ow loses its labiality, pro?
respectively.
Figure 9: Diphthongization
PMP
>
early-PC >
*-i
>
*-u
>
*-i?
*-uu
>
>
chains for PC word-final
high vowels
PC
>
Jarai...
>
*-sy
*-ow
>
*-aj
*-au
>
l-?il
>
/-?u/
>
PR Cham, Tsat
Finally, the further lowering of *-aj and *-au in Tsat and in spoken Phan Rang
Cham to -?i and -?u was noted in Blood 1967 and later discussed in Donegan
(1985:133). Note that, except for the PC form *-ii, each of the posited steps is
116
attested either in one of inWritten
Cham or in one of the modern
Chapter
Chamic
lan?
guages.
chain posited in Figure 9 are

stages in the diphthongization
in the variation in Aymonier and Cabaton's 1906 dictionary of Cham, a
Most
attested
dictionary that is panchronic in the sense that the spellings of the words
ently come from texts spanning the whole history of Cham writing.
chain steps attested
10: Diphthongization
Figure
appar?
inWritten Cham
PMP
> Early
>
Later
>
...
> Modern
*-i
>
*-ij
>
*-sy
>
-aj
>
*-u
>
*-uu
>
*-ow
>
-au
*baru
>
baruw
>
barow
>
*tabu
>
t?buw
>
tabow
>
l-?il
>
/-?u/
...
>
barau
...
>
tabauw
'new'
'sugarcane'
Although the entries inAymonier and Cabaton are not dated, if the range ofWrit?
ten Cham variation is placed on the chain, three of the four stages posited for the
transition from early-PC *-uy to modern PR Cham /-?u/ are attested.
chain in Figure 9 and Figure 10 should be familiar
it is also well attested in the history of Germanic.
chain has
points out, the following diphthongization
to historical linguists because
As Donegan
(1985:214)
occurred repeatedly inGermanic,
Priebsch
and Collinson
(1966)
citing Stampe (1972) for the history of English,

for Standard German, and Sapir (1915) for Yid?
dish.
11: Germanic
Figure
Note
>
>
>
ii
>
uu
chains
diphthongization
sy
>
aw
>
>
4
au
>
ai
au
process has stressed vowels becoming long and only after

It is also worth noting that, if the shwa in
do
becoming long
they diphthongize.
PC were replaced by a and the -?- in Phan Rang Cham were replaced by a, the
PC chains and the Germanic chains in Figure 11 would be identical.
that the Germanic
Along
ants
of
with
proto-Malayic
the rest of Chamic, Acehnese

word-final
*-i
and
*-u,
where
vari?
inherited diphthongized
the
stressed
PMP
>
proto
Malayic high vowels *-i and *-u underwent lengthening and then diphthongiza?
tion.Within PC, the chain of Acehnese diphthongization was, of course, identical
with
the developments
in mainland
Chamic
up to the point where
the Acehnese
Chamic
117
Vowels
migrated to northern Sumatra, which seems to be while the PC diphthongs were

still preserved as *-sy and *-ow. As Durie notes (p.c.), the dialect data preserves
a complete record of the various Acehnese
stages in the development of PMP *-i
> proto-Malayic
*-i > PC *-sy to Acehnese
-oj. to Acehnese
*-sy > Acehnese
(Bireuen) -oa, that is, the dialect data contains the -sy posited for PC as well as
both the -oj, and the -oa posited as stages in the development of the Bireuen
forms. However, the dialect data has not preserved parallel forms for the stages in
the development
of
*-u
>
-sa.
12: Acehnese
Figure
diphthongization
chains
modern
PMP >
early-PC >
*-i
>
*-ij
>
*-u
>
*-uu
>
PC
> Acehnese
>
*-sy
>
-sy
>
>
*-su
>
*-ow
Acehnese
-oj
*-sy
>
>
>
Acehnese
-oa
-sa
After
took
that, the mainland Chamic languages and certain dialects of Acehnese
different paths of development. First, some Acehnese
dialects dissimilated
the
onsets
of
This
both
and
and
*-ow,
*-su,
syllabic
producing *-oj
respectively.
*-sy
dissimilative fortition effectively reversed the earlier vowel quality distinctions in
the onset of the diphthongs! Then, apparently reflecting the tendency for lax vow?
els to develop ingliding diphthongs
dialects
(Donegan 1985), these Acehnese
reduced both diphthong
to Acehnese
-oa
and
-sa,
upglides
into shwa, turning pre-Acehnese
*-oj and *-su
respectively.
despite the common starting point and the early shared paths of
the PC and some of the Acehnese
chains ultimately diverged
development,
but are
sharply, resulting in reflexes that correspond perfectly diachronically
Thus,
quite distinct phonetically.

Reflexes
of PMP
*-/- and *-?
In the transition from PMP to PC, the reflexes of the PMP high vowel *i split: in
syllables and in unstressed open syllables, it remained *-i-, while in open
stressed syllables, it became early-PC *-ij and then PC *-sy. It is important to
closed
note that, while most PMP forms do, not all the PMP open syllable *-i reflexes go
to PC *-sy. For a small number of grammatical forms, the PMP open-syllable
*i
a Chamic
pattern of reflexes thatmatches the reflex pattern for *-i-.

The last two of the three grammatical morphemes
in Table 53 sometimes
pattern with the *-sy forms and sometimes pattern with the *-i forms, suggesting
that there was originally an unstressed and a stressed variant, with the unstressed
displays
118
Chapter
variant
on
going
to pattern
with
*-i
and
the
one
stressed
with
Acehnese
*-sy.
sup?
plies several nice pairs of contrasting stressed versus unstressed reflexes, with
sod 'who' and the unstressed si 'title for
one of these being the stressed Acehnese
names'.2
people's
Table 53: PC open syllables with *-i in unstressed

PC
di
*d?
*-h?
*kam?
*kamsy
ta-
Haroi
?
tin?
ni
-nsi
PR Cham
ti
rat'
ni
'this'
kami
fwe
ni
-nai
kamoa
NR
?
Chru
ti
-noa
*-nsy
Jarai
Rade
Aceh.
syllables
hamsi
gamai
-iv
-iv
labu?kamin
(ex.)?
kamei;
-v
? k?me
It is important to note that the forms themselves are inherited, not borrowed.
ini and ni, and the
Malay, for example, has the preposition di, the demonstratives
reflex
Their
the
forms.
kami
for
pattern comes
pronoun
unique
corresponding
from
their
occurrence
cally
occur
stressed
in a unique
in some
but
these
environment:
phonological
contexts
unstressed
typi?
sources
Other
The majority of the remaining PC *-i finals are in early borrowings,

of these forms are found elsewhere inAustronesian.
Table 54: Two other early borrowings with PC open syllable

Rade
PC
xxsisi(r) kasi
*pagi
magi
Jarai
Chru
tasi
tasi
pagi
pagi
NR
kasi
pagi
Haroi
Tacomb'
pake
'tomorrow'
-v
suggests that 'comb' is a borrowing from MK,

Headley
in
presence
Malay as sisir requires that it also have been borrowed
The pattern reflects
between
content
two
*-/
tathi
casei
pakhi
although
PR Cham
-v
2.
forms
in others.
words
a restriction
and
on stress distribution,
grammatical
morphemes.
The
but, if so, its

intoMalay. In
not the difference

apparent
excep?
in practice usually encliticized

tions?the grammatical morphemes?are
a
a
are
of
and
thus
constructions
part
larger phonological word.
larger
into
Chamic
119
Vowels
any case, the actual PC reflex may simply reflect the final *-ir. As clearly
occurred with PMP forms ending in final *-ur, the vowel was not in an open syl?
lable when open syllable PMP *-i went to PC *sy. Rather, the *-ir lost the final r only later, and thus did not diphthongize. However, with the form pagi, which
shows up inMalay as pagi, the PC final *-i is both unexpected and unexplain
able, unless it is aMalay borrowing.
All the remaining forms that pattern like PC final *-i are loans, inmany
cases loans that postdate the breakup of PC. Headley has identified a number of
these
as
loans;
evidence
also
many
irregular
patterns.
correspondence
For
'scis?
that it is a MK
sors', Headley
loan, while another author has
(#1.54) suggests
suggested the less likely Tamil form k?l?kati 'areca-nut scissors'. For 'horn; ant?
ler', Headley (#1.34) identifies it isMK in origin, while it also looks like a PLB
form.
and probably all, of the forms below are loans. Headley (1976)
that three of these forms are borrowed on the basis of theMK data.
Most,
has suggested
His contention
receives support from Chamic: for all three forms, the correspon?
dence patterns within Chamic are irregular, indicating that the borrowing post?
dates the breakup of PC. With
'horn' even the representation of the borrowed
form as xtuki is quite problematic: first, it is quite possible itwas borrowed more
than once with different
second, there are at least three distinct vari?

ants in the data: Northern Roglai (Lee) tukii (with the barred-i being the expected
reflex of an earlier medial -r-), PNB *ake, and Proto-Mnong *nke. It is instructive
only has a form for one of these, unos for 'honey bee'.
thatAcehnese
55:
Table
post-PC
xkatri
forms. And,
Borrowings
Rade
katrsi
with
Jarai Chru
?
katrai
post-PC
NR
katri
-f
xtuki
ki
taki
open
Haroi
katroi
syllable
W. Cham
ei
ra?i
PR Cham
?
'scissors'
katray
-vf
taki
tuki
cake
take
-v-v-v
xla?i -f
*-i
la?i
la?i
take
-v
la?i
la?i
'horn;
antler'
lii
'winnowing
basket'
xtali
xhuni
hanue
kli-i
hani
?
hani
?
hun?
calei
hani
?
hani
tali
'flat (of rocks)'
hani
'honeybee'
Chamic, the reflex of PC *-i- was simply -i-, often marked

as
short or long by various authors but although length is
syllables
contrastive
in
individual languages, the length does not reconstruct to
apparently
PC. Not only does the length not agree across languages but within the individual
Throughout
in closed
languages
its occurrence
can be predicted on the basis of the environments.
120
Chapter
of PC *-/- < PMP
Table 56: Reflexes

Acehnese
Rade
Jarai
-e?;-s?
-im;
-??
-??
-et;
-st
-?t; -??
-en;
-sn
-in
-?n
-?p
NR
-ih -ih
-ih
-im
-im
W. Cham
PR Cham
-ih
-?n;
-im
-im
-ip
-i?
-i?
-i:?
-i:?
-it;
-i?
-i:?
-i:?
-il
-il
-il
-in
-in;
-in
-in
-in;
-?n
-i:n
-ig
-an
-?n
-ip
-iu?
-iu?
-iu?
?
-ep;
*-i
Chru
-ih-sh-ih
-eh;
?
-im -em;
-i?
-i?
-il -il
-it
-?n
-?n
-?w?
forms reflect two processes: the lowering of all high vowels and
on
in Acehnese"
lowering of all nasalized vowels ("Nasalization
seems
to
of
the
Jarai
reflexes
The
page 176).
represent something
variability
about Jarai rather than something about the source, as the variation shows up in
The Acehnese
the further
sources on Jarai.
both available
The upglide developing from PC *-ip in Chru, Northern Roglai, Western

and
Phan Rang Cham represents the fortition of the second formant transi?
Cham,
to
tion
the
final
-p,
rather
than
assimilation
to
the
"labiality"
of
the
*-p,
as
the
"labiality" of the *-p involves lip spreading while the "labiality" of the glide rep?
resents lip rounding (see Thurgood and Javkin (1975), for an acoustic explana?
tion of this class of sound changes). A parallel change occurs with PC *-ap.
The PC reflexes
developments
Written Cham.
Table 57: Reflexes

PMP
PC
*-i
*-sy
of the PC *-sy show more variation. The Acehnese

above. The PC *-sy is retained as such in Rade and
were discussed
Aceh.
-oa
of PC *-ey (< PMP

Rade
-si;
Jarai
-ai
-ue (m)
Chru
-ai
*-i)
NR
-ai;
Tsat W.Cham
-ai
-ay
PR Cham
-?y
-uai (m)
In Rade, Jarai, Chru, and Northern Roglai, the reflexes of PC *-sy become -ai. In
Tsat, Haroi, and Phan Rang Cham /-ai/ developed, through the lowering of -ai.
Reflexes
of PMP
*-w- and *-u
In the transition from PMP to PC, the reflexes of the PMP high vowel *u split: in
syllables and in unstressed open syllables, it remained *-u-, while in open
closed
Chamic
121
Vowels
stressed syllables,3 it became early-PC *-uu and then PC *-ow. PC also devel?
in the reflexes of *-u, but only before final
oped restricted length distinctions
glottal stops and velar nasals. (See "PC *-u- length distinctions" on page 141, for
further discussion).
The reflexes
of PC *-u are typically -u throughout Chamic, except in

and restructured register" on page 197) and Acehnese.
"Haroi
vowels
(see
In Acehnese,
the reflexes of PC high vowels are regularly lower, with reflexes
after a nasal lower still ("Nasalization inAcehnese" on page 176).
Haroi
The Chamic
reflexes of PC *-ow parallel the reflexes of PC *-sy. In

written Phan Rang Cham the *-ow is represented as -ow (the Phan Rang Cham in
Table 59); but in the modern spoken Phan Rang Cham, it first delabialized,
-?w, as it has in Rade, Tsat, Western
-ai, and then lowered, becoming
becoming
and Haroi.
Cham,
Table 58: Reflexes

PC
Aceh.
*-uh
-oh;
-oh;
*-un
-on;
*-um
-om
-on;
*-u?
*-u:?
-on
-om;
*-un
*-u:n
Rade
?
?
*-us
of PC *-u- < PMP
-o?;-o?
?
Jarai
*-u
NR
Chru
Tsat W. Cham
-uh
-uh
-uh
-uh
-u55
-uh
-?h
-uih
-uih
-u:h
-uh
-u55
-uh
-?h
-?n
-?n
-un
-ut;
-?n
-?n
-um
-um
-?m
-um;
um
-?n
-un;
-un
-on;
-??
-??
-??
-up
-un
-un
-uk;
-un
-o:n;
-o:n
-u?
-??
-un
-un
?
-um
-un
-u:k
-un
-?n
-un
-on
-on
-u?
-u:?
-un
-?? -??
-u:?
Although preserved in the orthography, in the modern

the onset of PC *-ow has been unrounded to /-?u/.
Table 59: Reflexes

PC
*-ow
3.
Aceh.
-sa
PR Cham
Rade
-?u
It is worth noting
cally long.
of PC *-ow < PMP

Jarai
Chru
-au
-au
that in all vowels
NR
-au
-u?;-o?
spoken Coastal
-u?;-o?
languages,
*-?
Tsat
Haroi
-au;
?u;
-(i)a
-?au
in word-final
W. Cham
PR Cham
-au
-ow
open syllables are phoneti?
122
Chapter
sources
Other
It is important to note that some forms that initially appear to come from a PC
*-u actually reflect an early PMP open syllable *-ur. Although the final -r itself is
no longer directly reflected anywhere in Chamic, PMP forms ending in final *-ur
have non-diphthongized
reflexes throughout Chamic.
(1) PMP
(2) PMP
>
PC
*-ur >
PC
*-u
*-3W
it clear that the forms with *-ur were not yet in

PMP
*-u
in open syllables went to PC *-ow. Instead, the
when
the
open syllables
PMP *-ur lost the final -r only after this change, and thus did not diphthongize.
The Chamic
lui reflexes make
Table 60: Reflexes

PMP
*-u vs. *-ur in Chamic
of PMP
PC
Malay
Aceh.
Rade
PR Cham
*bulu
bulu
*bulow
bulsa
ml?u
*aku
aku
*kow
ksa
k?u
palow
kow
'body hair'
'
F
*kutu
kutu
*kutow
gutsa
ka tau
katow
'head louse'
*ikur
?kor
*?iku
iku
ku
t?lur'egg'
*klu
klu
'testicles
*la?ur
liu
'coconut
*telur
"niyur
V nyor
*kajur ? kujur
*kaju
'tail'
iku
kaju
'spear;
(animal)'
palm'
lance'
the forms
forms have retained the final -r. In Acehnese,
final -r have a diphthong, while the form for 'tail', which had an earlier
PMP -r, has no such diphthongization.
Similarly, in the rest of the Chamic lan?
Notice
that the Malay
without
the forms with PMP -r did not diphthongize along with the forms in open
syllables. This striking parallelism in the changes and in their relative chronology
guages,
is just one of many pieces of evidence

Chamic dialect that left for Sumatra
that demonstrates
("Acehnese,
that Acehnese
a Chamic
was
language"
a
on
page 47).
Reflexes
of the PMP
shwa
In all contexts, the original PMP shwa disappeared, merging with the reflexes of
PC short *a. Meanwhile,
the PMP *a became PC long *-a:- before the finals * -?,
-rj, -k,
-1, -r, -n,
-p,
and,
more
marginally,
before
-t, producing
length contrast with the former PMP -a- reflexes, which
in these
words
always remained short in
Chamic
123
Vowels
introduction of various pre-PC MK borrowings strengthened the PC

contrast
between long and short *a (for a further discussion of the length
length
contrast, see "PC *-a- length distinctions" on page 146).
PC. The
Table 61: The reflexes of PMP shwa

PMP
PC
*gatel
*gatal
NR
Haroi
PR Cham
Aceh.
Rade
Chru
gatai
katal
katal
katan
kat?l
kat?l
kan??
tana?
tana?
cana?
tan??
'cook'
aka?
ika?
ak??
ik??
'to tie'
hadar)
hala?
hadak
*tanek
*tana?
*iket
*?ikat
ikat
*qajeng
*hadan
*qulej
*hulat
ararj -m had?g
ulat
hlu?t
hula?
'itchy'
'charcoal'
ath?arj hat?n
?
hal??
'worm'
(m)
In the examples in Table 61, as in all the examples found so far, PMP -a- has
become PC *-a- (The Chamic reflexes of PC *-a- are discussed on page 146ff.).
of PMP
Reflexes
*a
PC, a length contrast developed before final -?; -n; -k; -1; -r; -n; and mar?
ginally before -t. (See page 138ff. for further discussion of the length distinction.)
The Chamic reflexes of PC *a show a minimum of variation; the variation that
Within
does
occur
is usually
The Haroi
conditioned
reflexes,
by
the
syllable-final
consonant.
obscured
of regis
by the secondary development
are discussed
in "Haroi vowels and restructured
trally-induced vowel splitting,

register" on page 197. The Tsat correspondences
included in the table.
are too marginally
attested to be
The table is organized by the manner and place of the final consonant.
Among the final stops, the steady, incessant Chamic lenition of final *-p, *-t, *-k,
and *-c to a glottal stop is somewhat irregular in its reflexes. The upglide devel?
oping from PC *-ap inChru, Haroi, andWestern Cham is acoustically rather than
of *-ip in
articulatorily based. (See the brief discussion of the development
page 117.) The reflexes of PC final nasals are split in Roglai, with some forms
retaining the nasals and others developing
stops (see
homorganic voiceless
on
in Northern Roglai"
"Nasalization
page 170). The short vowel reflex of PC
*-a:n only occurs occasionally
in Phan Rang Cham. InWestern Cham and Phan
Cham
has
nasalization
sometimes
Rang
produced a high vowel reflex. Finally,
are
the Acehnese
reflexes
often split, with the most frequent conditioning factor
being whether
or not the vowel
is nasalized,
typically by the preceding
nasal con?
sonant.
124
Chapter
Table 62: The reflexes of PC *a

PC
Aceh.
*-a
-a;
Rade
Jarai
-An
-oh
-ah
Chru
W. Cham
NR
PR Cham
-a
-a
-a
-a
-ah
-ah
-ah
-ah;
-aih
-a:h
-a
-ah
-an
-a
-ih
*-ah
-ah;
*-a:s
-uiah
*-as
-aih
-ah
-ah
-ah
-ah
-ah
-ah
*-ap
-ap
-?p
-?p
-a?
-a?
-au?
-??
-ap
-ap
-a:u?
-a:?;
-au?
-aw?
-?t;-a?
-at;
-a?
-a?
-a?
-??
-??
-at;
-a?
-a:?
-a:?
-a?
-a?
*-a:p
-uiap
*-at
-at
-aih
-tut
*-a:t
-uiat;
*-ak
-a?
*-a:k
-uiak;
-uik
-at
-?k;-??
-?k;-??
-a?
-ak;-a?
-a?
-ak
-ak
-a:?
-a:?
-a?;
-??
-a:?
-a:?
-a?
*-a:?
-at;-a?
-??
x*-ac
-mat
-ac
-?i?
-a:i?
p-am
-am
-am;
-am
-am;
-am
-am
-a?;-ak
-i?
-ai?
-ay?
-ap;
-am
-am
-am
-an
-an
-an
-an
-ak
''-an
-an
-an
-an
-an
-an;
-man
-an
-an
-a:n
-a:n;
-an
-at
-a:t
-*g
-an;
-in
-an
-an;
-in;
-an;
-uin
-man;
*-an
-at)
*-ar
-uia;
-ia
-an
-an
-a:n
-a:k
-arj;
-tr\
-??
-??
-a?
-a?
-??;
-a?
-ar
-ar-;
-a:r
-a
-?r
-an;
-an
-??;
-??
-ar
-?r;
-ar
*-a:r
-uia
-a
-a
-a
-a
-a
-a
*-al
-ai
-?l
-al
-al
-an
-?l
-?l
*-a:l
-uia
-al
-al
-a:l
-an
-al
-al
of the PMP diphthongs
Reflexes
The
reflexes
of
PMP
*-ay,
*-aw,
and
*-uy
are
PC
*-ay,
tively. It is worth noticing that the first two diphthongs

counterparts to the word-final PC *-sy and *-ow, which
from
-in;
-in
-in
-a:n
-i?
-a?;
-a?
-a:i?
*-an
-a:n
-??
-ih
-ah;
PMP
*-i
and
*-aw,
and
*-uy,
respec?
are, in effect, the long

developed word finally
*-u.
Chamic
125
Vowels
The Chamic
Two
well-attested.
reflexes of PC *-ay, *-aw, and *-uy are straightforward and

interesting about these correspon?
things are particularly
dences. The split in the Acehnese reflexes of *-aw correlates with vowel nasaliza?
tion; the lower vowel reflex is the nasalized one. In Rade, Chru, and Tsat, loans
on the bases of their distinctive reflexes. Thus, the
are sometimes distinguishable
Rade reflex -ai indicates a loan; the actual reflex is the much more aberrant look?
ing -ie. In Tsat, the regular reflex of PC *-ay is -aii?42,with the *-y final strength?
ened to a glottal stop; the lack of a final glottal stop marks a form as a loan.
unexpected vowel length seems to occur with
more
to
be done on this.
work
needs
although
in Chru, otherwise
Finally,
loans,
Table 63: The reflexes of PC *-ay, *-aw, and *-wy

PC
Aceh.
-e;
*-ay
Rade
NR
Jarai Chru
-ai
-ie;
-ai
-ai
-o
-o;
-au
-au
-ui
-ui
-a:u;
*-uy
of PC *-ay
Diphthongization
In Rade,
the
accented
open
syllables,
Rade
PC
-ui
-ui
of PC
as PC
to Rade
(Lee's
-je
is -ie,
*-ay
was
*-ay
-ui
-ui
-uy
in final,
syllables.
-ay
(1974:655)
terms this change from PC
a characterization
-ye) metathesis,
occurred
PR Cham
-ai
Lee
only
to such
W.Cham
-ai
that
change
restricted
NR
Jarai
on the correspondence,
Commenting
-aw
in Rade
-ai
-?e
*-ay
*-ay
reflex
regular
-au
-au
x-a:u
-ui
-ay
x-a:i
-au
-au;
PR Cham
-ai
-a:i?42;
x-ai
*-aw
W. Cham
Tsat
which
is not
unreason?
able as a description of the diachronic correspondence with PC. However,

the
actual mechanism of change was undoubtedly not metathesis.
in the
Instead, amore likely chain involves a series of steps well-attested
literature: the change of PC *-ay to *-ee, dissimilative raising of the onset, and a
shift in syllabicity.
PC
*-ay
>
>
assimilation
*-ee
> onset raising >
syllabicity
>
-je
>
-ie
shift
126
Chapter
The first step involving assimilation

a monophthongization,
is, essentially
is a variant of the change of *-ay to -e, that

but accompanied
by the natural conse?
length. The onset raising of the next step is illustrated by

with examples from Finnish (*ee > ie in initial stressed
Donegan (1985:142-143)
from
the
transition
of Vulgar Latin to Old French (e, > ie, in accented
syllables),
and
the
from
Finca
open syllables),
Valparaiso dialect of Pokomchi, a Quichean
of retaining
quences
language,
examples
(ej.> ie, but apparently with no mention of stress or accent). Donegan's

parallel the Rade change. The final step, of course, involves a shift in
syllabicity, with the onset becoming a glide and the second part becoming the
syllable nucleus. Note that this last step changes a more closed syllable into a
more
open
syllable.
THE BORROWED PC MAIN SYLLABLEVOWELS

In addition to the vowel categories inherited from PMP, PC included a number of
that were borrowed into pre-Chamic from MK sources. Of course, many
vowels
MK borrowings fit nicely with the consonant and vowel categories inherited from
PMP. On the one hand, these are difficult to identify on internal grounds; on the
other hand, these had only a limited effect on the PC segmental inventory.
Table 64: PC main syllable vowels,
*-ia
*-i_?
*"i
inherited and borrowed
*-u *ua
*-u:-
[x-i?w] *-u-,
[x-iaw]
x*-uay
*uay
*-sy,
*-ow,
*-o
*-a:
*-ay
*-uy
*-aw
borrowings, however, had a significant effect on the vowel

categories of PC, an effect made salient by the fact that some vowel types are
but not exclusively, by MK borrowings. In effect,
exemplified overwhelmingly,
these new vowel types have been borrowed from MK. However, the mechanism
for developing
the new category always seems to involve more than simply bor?
Other MK
rowing words containing the vowel. Typically, but not invariably in this data, the
borrowing of a new vowel category also seems to be accompanied and, indeed,
Chamic
127
Vowels
facilitated by the innovative development of the new category out of native mate?
rial. The number of native tokens of the new type may be small ?
perhaps only
one
or
two
but
are
there
some.
usually
In the table, the vowel categories associated with MK borrowings are in

boldface to distinguish them from inherited vowel categories. As for the vowels
themselves, the *-i in open syllables, despite being in boldface, originates from
two sources, one involving borrowed MK forms, the other involving forms inher?
ited from PMP (page 117). And, finally, most of the vowel length distinctions
involve the interaction of influence of MK
borrowings
and inherited PMP mate?
rial.
PC*e
of words containing PC *s produces some fascinating asymme?

tries. Despite the fact that there are only a small number of PC forms that recon?
struct with *s and that their etymologies
are far from worked out, what we do
a
know provides
clear outline of the history. The distribution of reflexes in the
The examination
modern
InAcehnese
(Durie 1990a), before certain
languages is uncontroversial.
finals the *s is retained as such, and before others is backed to *a. And, on the
the Chamic
mainland,
close
mutual
in closed
languages Rade and Jarai, as is not unexpected

have
interaction,
the
same
set of
conditioned
the -e in open. The remaining
syllables,
reflexes:
the
given their
-s-
occurs
languages each have but a sin?
gle reflex.
Table 65: The reflexes of PC *?
PC
Aceh.
*s
s; a
Rade
-s-;
Jarai
-e
-s-;
-e
Chru NR
Tsat
The full set of reconstructed
W.Cham
PR Cham
forms is worth
listing. The first word in the

'more, surplus', related to Malay
list is the Acehnese
form lwbsh, PC *labsh

l?bih 'more' and a native PMP root. This word, however, is the only clearly PMP
word in the entire list. However, as Durie notes, theAcehnese l-z-l vowel is irregular.
Table 66: The reflexes of PC *lob?h
PNB
PMnong
Malay
l?bih
PC
Aceh.
*labsh
luibsh -v ebsh
Rade
PR Cham
lipsh
'more; surplus'
The next group of words all reconstruct fairly regularly to PC, but they have nei?
or other PMP counterpart, nor a MK counterpart in PNB or
ther a Malay
PMnong.
128
Chapter
containing PC *e
Table 67: Reflexes

PC
Rade
Aceh.
x*hagst
PR Cham
haket -f
-f
Is?
x*ls?
'why? what?'
'fall
into'
'to elbow'
x*nsh
'to drop'
kapl??
x*pals?
p??
p??
'pick, pluck'
x*racfs
erfe
rarfs
'Rhade'
x*ts?
t??-v
x*kls?
tie?
pAt
x*pst
'torn,
x*tulsh
x*tamsh
kl??
'steal'
tal?h
'untie'
kamsh
tamsh
worn'
'pillar; post'
x*?s?
6??
'elf.
x*g8
ke
'boat'
long,
thin
objects'
the next group of words, two pieces of evidence suggest that,

With
two which are post-PC borrowings, these forms are also early
for
last
the
except
MK incorporations but reconstructable to at least to PC. Note that the imperative
not only reconstructs to PC but also occurs inAcehnese. Each has a plausible MK
that
etymology,
is,
each
word
looks
a MK
like
(PNB), proto-Mnong,
proto-North-Bahnaric
have aMalay (or other PMP) counterpart.
Table 68: PC
PNB
PC
x*bs?
*beq
*randeh
Aceh.
Rade
bs?
edsh
x*radsh
x*glsh
*aseh
'IMPERATIVE'
rate h -v
'vehicle'
klsh
'tired'
vine'
'rope;
'cut'
?s?
'near(ly); about to'
x*?assh
assh
ath?h
'horse'
x*?sh
sh
?h
'excrement'
ale
*pale
*gle
x*?als
*babe
*be
xbubs
be
xwsh-vf
wsh
*weh
p??
x*rsh
x*js?
*?aseh
either
none of them seems
PR Cham
hru??
x*hurst
*ajeq
for
reconstructed
*e forms with MK etymologies
PMnong
xsreh
word
or PKatuic;
'medium
paps
w?h
'goat;
'turn
bamboo'
sheep'
aside'
to
Chamic
129
Vowels
The next group of words all reconstruct fairly regularly to PC, but they
have neither aMalay or other PMP counterpart, nor aMK counterpart in PNB or
PMnong.
Table 69: Reflexes

PC
x*hagst
containing PC *?
PR Cham
Rade
Aceh.
-f
Is?
x*ls?
hak?t -f
'why? what?'
le?
'fall
into'
'to elbow'
x*nsh
'to drop'
kapl??
x*pals?
p??
p??
'pick, pluck'
x*rarfs
erfe
rarfs
'Rhade'
x*ts?
t?? -v
x*kls?
tie?
pAt
x*pst
x*tulsh
kl??
'steal'
tal?h
'untie'
kamsh
tamsh
x*tamsh
'torn,
worn'
'pillar; post'
x*?s?
6??
'elf.
x*gs
ks
'boat'
long,
thin
objects'
The last set consists of the three words, aside from *labsh, which actu?
occur
both inMalay and Chamic. Neither the word for 'seed' nor the form
ally
for 'tickle' reconstruct to PC; in any case, the word for 'seed' is apparently a San?
skrit
borrowing.
More
of
interest,
however,
is the
fact
that both
appear
to recon?
struct to the PC level, with the Acehnese forms corresponding nicely.

As a possible native PMP root, only the form for 'great grandchild'
requires any careful examination. As for the initial c-, Adelaar (1988:62) writes
with respect to his proto-Malayic
that the proto-phoneme
*c is not well-attested,
at
words
it
least
In
any case, if the PC and
containing
making
mildly suspect.
are
are
not
related
and
if
both
MK
in
forms
Malay
origin, 'great grandchild'
might constitute a second native PMP root in Chamic containing *s, although I
suspect that the form is ultimately MK.
Table 70: PC *e forms without MK etymologies
Malay
PC
P-Aceh.
Aceh.
PR Cham
*(cs)cst
CAt
tac??
*bijeh
bijsh
pa?sh
'greatgrandchild'
'seed'
gli?-gli?-iv
kal?k
'tickle'
cicit
*cic?t
biji
x*bijsh
g?lak
xgilsk -lv *gle?
'laugh'
130
Chapter
In any case, theMK origins of PC *s are firmly established. Further, the

handful of irregular PC roots with PC *s ultimately from MK but with regular
roots are what one expects ifAcehnese
is a Chamic dialect. (The
proto-Acehnese
are
reconstructions
from Durie 1990a).
proto-Acehnese
PC shwa
The original PMP shwa merged with the PC reflexes of short *a, allowing a new
PC shwa to develop out of material incorporated from MK sources. The Chamic
reflexes of this new shwa are given in the table.
of PC *3
Table 71: Reflexes

PC
Aceh.
*-ah
*-at
Rade
-Ah
-ih
-ah;
-At
-??
*-ak
Chru
Jarai
NR
Tsat W.Cham
?
-ah
PR Cham
-ah
-ah
-ah
-??
-a?
-a?
-a?
-a?;-??
-a?;-??
-ak
-?k;-i?
-a?
-a?
-a?
-a?
-a?
-au?
-?w?
-ih
-ah;
*-ap
-op
-??
-a?
-au?
-a?
*-an
-An
-an
-an
-an
-ak
-an
-an;
*-am
-om
-am
-am
-am
-ap
-an
-om
-am
*-an
-An
-an
-on
-an
-at
-en
-an
*-ar
-a
-ar
-ar;
-ar
-a
-?r
-?r
-or
-an
-?n
-?n
-an;
if the final *-p were not retained inAcehnese,

itwould be obvious
that
reflexes in the mainland
languages
*-ap correspondences
reflect a final *-p. Parallel changes happen with *-ap and *-ip; see "Reflexes of
Even
from the vowel
PMP
change.
*-i- and *-i" on page

The
-i-
reflexes
117 for a brief discussion
in Western
and
Phan
Rang
Cham
of the acoustics
result
from
of the
nasaliza?
tion.
The sources
of shwa
Thus far, only a handful of forms reconstructed with shwa have potential Austro?
nesian etymologies. When
there is a readily-determinable
source, forms with
shwa are overwhelmingly
from MK, and even when the source is not fully identi?
fiable, it is often obvious from either the phonotactics or from the presence
non-PMP phoneme that the form was not PMP in origin.
of a
Chamic
131
Vowels
of PC forms containing shwa and potentially related to

forms are listed in Table 72. Of these only *bukan with its
The handful
(or PMP)
Malay
*bukan
proto-Malayic
is
reconstruction
*lan
although
convincing,
totally
also
looks plausible. The form *lan has a potential PMP source in *talun 'fallow
land; secondary forest'. In the case of theMalay word for 'hatch', the first sylla?
ble vowels inMalay and Acehnese
do not match, possibly indicating that the
word was
independently borrowed into Malay and PC. The last word, although
related to the Indonesian mythical bird the garuda, is a borrowing in Chamic.
and PC words with PC *a
Table 72: Malay

PNB
P-Malayic
PC
Malay
*bukan
bukan
Aceh.
Rade
PR Cham
bukAn
mak?n
kdw'-kan
'other;
*bukan
k?ram
*lan
lAn
l?n
Ian
'earth; soil'
*karam
karom
kr?m
kar?m
'to hatch'
kra?
'vulture;
All
tions toMK.
reconstructed
xgrak
garuda
different'
garuda'
in the following table have probable etymological connec?

are
Some
reconstructed for theMnong branch of MK and some are
the words
for the North Bahnaric

reconstructions
Proto-Mnong
naric reconstructions
branch. None
are from H. Blood
of them appear inMalay. The

(1968). The Proto-North-Bah
are from Smith (1972), occasionally
with minor modifica?
tions.
Table 73: MK words with PC *a
PNB
*c?n
PMnong
PC
*k?m
x*?ram
Aceh.
*kl3m
x*sam
*pan
x*pan
psg
*tap
x*klap
t?o?
*yah
x*yah
*dak
x*dan
*son
x*?an
*g?n
x*gan
*hag?r
x*sagar
*joq
**?jap
PR Cham
Rade
?r?m
som
?r?m
'to wrap'
'to nail;
'stab;
miy?h
dAn
dan
tag
??n
??n
*p?y
of
x*b-an-at
'pole;
'drum'
?jo?
?jow?
'correct;
a fence',
itself, although apparently originally
stop'
post'
right'
'to fan'
x*pa-pat
'a dam;
poke'
'to eat'
hak?r
The root x*bat 'to dam; to fence', which

tion
'stand;
k?g
hammer'
'PARTICLE'
hagar
?
-
arrow'
contains
borrowed
an
is the root used for the forma?

instrumental
from MK,
prefix.
The
prefix
is not restricted just to
132
Chapter
for instance, this infix not restricted to

forms. InAcehnese,
in disyllables
(cf. Chapter 9), a
specific forms but instead occurs productively
lan?
that says something about the intensity of the MK-Chamic
characteristic
a handful of fossilized
guage contact on the mainland that led to the borrowing.

Although not reconstructed for proto-North-Bahnaric
(with one excep?

Bahnar
in
74
in
and
Cabaton
the
words
Table
(1906)). It
appear
tion),
(Aymonier
to
to
not
fact
Smith
did not
if
attach
the
that
is
obvious what significance,
any,
reconstruct
On the one hand, a lot of the Bahnar

these for proto-North-Bahnaric.
from Cham; on the other, these forms do not appear to have
is borrowed
lexicon
An counterparts,
that they were not originally An.
suggesting
Table 74: Other PC *a forms with MK affiliations

PNB
Bahnar
(AC) PMnong
PC
banot
mram
bram;
Acehnese
x*b-an-at
x*?ram
*k?m
'a dam,
gar x*gar
'arrow'
'handle
gA
x*garj
?
t?l;
ap?r; p?r
fence'
(knife)'
g?n;g3n
*p?r
toi
'pole;
*p?r
until'
'arrive;
x*par
phA -i
'to fly'
x*gap
gop
'other;
?
gap
post'
x*tal
group'
The final set of forms have neither established MK nor Malay counter?
the first form looks unquestionably MK because of its glottal?
part. Nonetheless,
sources would
turn up
that careful search of MK
ized stop. It is possible
counterparts
for
at
Table 75: PC
PNB
PMnong
least
some
of
these.
*a words without PMP or MK etymologies

PC
Aceh.
?
Rade
PR Cham
x*?arj
gom
x*gam
xpayar
x*lap
x*madah
x*man
x*rak
x*sadar
-f
katar
'to burn'
kam
'to cover'
'corn;
?
myar
grain'
'to offer'
'fold'
mitah
'awaken'
mag
m?n
'from'
rak
r??
'grass;
hap?r
'remember'
-f
madih
hadar
tuwA
payai
low?
x*war
x*yar
g?m
x*katar
??rj
war
war
'forget'
yer
'lift,
weeds'
raise'
Chamic
133
Vowels
It is likely that these tables include at least some mistaken identifications

and misassociations,
but the overall configuration is unmistakable: the shwa has
entered PC through MK material incorporated into pre-Chamic.
PC
*ia,
x-i?u,
and
x-iaw
Three diphthongs were borrowed from MK sources: *ia was borrowed into pre
Chamic, becoming PC *ia, while x-i?w and x-iaw occur in post-PC borrowings.
For the diphthong *ia:, only the word 'water' appears to be a PMP root.
Table 76: Reflexes of PC Har
PC
*?iar
Aceh.
ia
Rade
sa
'water'
Jarai Chru NR
?ia
ia
Tsat
ia
?ia33
Haroi W. Cham
ea
?ea
-f
PR Cham
ya;
ier
'water
(fresh)'
The Chamic reflexes of *ia are regular, with the slight variation condi?
tioned by the finals. The two forms with nasalized vowels are marked with (n). In
working with Chamic forms containing the sequence -ia-, themain difficulty is in
forms that were originally disyllabic,
such as the disyllabicity
distinguishing
reflected inMalay tian 'stomach', from forms with an -ia- borrowed from MK.
This problem has not been completely
Table 77: Reflexes
PC
*-ia
Aceh.
?
Rade
solved here.
of PC Ha
Jarai Chru NR
Tsat
Haroi
W.Cham
-ea
-ya
-ea
-ya
-ia
-ia
-ia
-ia
*-iar
-ia
-ea
-ia
-ia
-ia
-ia
-ea;
*-ia?
-ia?;
-i??
-i??
-ia?
-ia?
-ia?
-ea?;
-ia?
-e? -ii? (n)

-ia? -?a?; -ia?
*-iat
-e?
-i??
-i??
-ia?
-ia
-ea;
-ia?
-ii? (n)
*-ian
-iag
-ien
-sg
PR Cham
-ea?;
-y??
-ea?;
-y??;
-?? (n)
?
-iag
-iag;
-iag
-eag;
-iag
-ieg
-iak;
(-yak)
In both Haroi
and Western
the onsets of these centering diph?

thongs have been lowered in every case, except for *-iag inWestern Cham. This
parallels but is more extensive than the lowering of the centering diphthongs
Cham,
134
Chapter
lui, which similarly occurs in Haroi and Western Cham (see

page 135). In the case of Haroi, the second reflex shows the result of subsequent
raising and backing of the onset to Ii/ under the influence of breathy voice
with
beginning
(page 197).
In Table 78, the PC *ia and *-iya- have clear examples inAcehnese, with
all of the examples except for 'dance' attested in Acehnese. The PC form for
'dance'
a counterpart
have
may
in Vietnamese
m?a
'dance'.
in origin, with a counterpart in

Malay tian. It is interesting that, within Malay, tian is disyllabic. The last two
forms, 'stomach' and 'wing', behave like disyllabic forms in some languages, in
which case the second vowel is long, and sometimes like monosyllabic
forms, in
which case the vowels behave as if they were *-ia-. The final *-iap produces
The word
upglides,
is PMP
'stomach'
*tiya:n
as does final *-p elsewhere.
Table 78: The reflexes of PC *-iya

PC
Aceh.
*liyah
Hah
*liya
Jarai
Rade
Chru
liah
haliya
eya
liah
layah
lia
raya
-i
riya
tamia
*timiya
Haroi
NR
PR Cham
leah
ly?h
'lick;
layia
liya
'ginger'
tamya;
'dance'
timiya
taste'
mya
tian
tiyuian
*tiya:n
*tiya:p
tio?-vf
tiyuiap
Aside
from the obviously

the
onomatopoetic,
ki?n
x-iaw
and
tiain
matiait
te an
tiaiu?
tia:?
te au?
onomatopoetic
x-iaw
only
occur
word
in forms
tyan
'stomach'
tyaw?
'chase'
'cat', which
postdating
is obviously
the
breakup
of PC.
Table 79: The forms with x-iaw and x-i?w
PC
Jarai
Rade
*miaw
xkri?w
xdriaw
x?i?w
mieo
?
NR
Tsat
miau
mi?u
miau33
krs?
ki?u
W.Cham
magiau
rfi?u
?au
egau
-v
PR Cham
miyaw
triew
?au
?au
cat
'castrate'
-i
dreu
-iv
xligiaw
Chru
ta:i?*2 iu
'exclaim'
iw
'left (side)'
ligiw;
'outside'
-iv
rag?au
lagiu
lagiw; giw
in one language or the other for all

Not only are there irregular correspondences
the forms, except for 'left (side)', but these forms are unattested in a number of
languages.
Chamic
135
Vowels
PC *ua and x*uay

no doubt exists that the majority of PC forms with *ua and x*uay were
incorporated into Chamic as the result of pre-Chamic contact with MK speakers,
two of the *ua forms are undoubtedly PMP: the PC *buat 'to do' and *dua 'two'
Although
are identical in form and meaning to theirMalay counterparts.

The rhymes marked with x*- only occur in PC words
ultimately from
sources. That is, even at this preliminary stage of investigation, it seems evi?
dent that six of the twelve rhymes are exclusively MK in origin.
MK
Table 80: Reflexes of PC *ua and **uay

PC
Aceh.
-uwa
*-ua
x*-uah
-uat
*-uat
*-ua?
x*-uam
?
?
-uan
x*-uan
x*-ual
x*-uay
x*-uac
Rade
Jarai Chru
NR
Tsat
Haroi
W. Cham
PR Cham
-ua
-ua
-ua
-ua
-ua
-oa;
-ua
-oa
-wa
-uah
-uah
-uah
-uah
-ua55
-oah;
-uh
-oah
-w?h
-?t;
-u??;
-ua?
-u??
-ua?
-oa?
-oa?
-oa?
-uot
-3t
-u??
-u??
-ua?
-ua?
-ua?
-??
-uom
-om
-o:m
-o:p
-uan
-om;
-urn
-uon
-on
-uan;
-uat;
-uan
-on;
-ul
-ul
-uan
-uan
-ual;
-uan;
-ual
-uan
-om
-w??
-om
-uan
-?n
-ual
-?l
-un
?
-ul
-ua
-ue
-uai
-uai
-uai
-ua:i
-oai;
-uai
-oy
-uat
-u??
-u?i?
-uai?
-ue?;
-oi?
-?ai?;
-uai?
-?y?
-oi
-ui
-uai
-oy
_?i?;
-?i?
_ui?
_?y?
-oa;
-ur
-ur
-uai?
?
-ui
x*-uay
___
x*.uc
_u?
-ui
-uai
_ui7;
_ui7
-uai
_ui?
___
-uc
x*-uar
-uor;
-ua
-ua
-u?
-ua
-ua
There are two major tendencies which are evident in the reflexes of these
a tendency that seems to
diphthongs. There is a tendency to monophthongize,
have existed
in Chamic
contact with MK. Notice
ever since it acquired diphthongs and triphthongs through

that the reduction of triphthongs to diphthongs and of
136
Chapter
is scattered throughout these languages, and is par?

diphthongs tomonophthongs
in
evident
Phan
Haroi,
ticularly
Rang Cham, and Jarai.
The second tendency is to lower the onset of centering diphthongs. For
the first five rhymes listed in Table 80, the onset is lowered inHaroi andWestern
Cham but without monophthongization.
In a number of other languages, there is
both lowering and monophthongization,
although in several cases length is pre?
served. PR Cham, for instance, frequently reduces /ua/ to loi.
Haroi goes further than any of the other languages in lowering the onsets
of these diphthongs. In fact, Haroi appears to have lowered the onsets of all these
for
except
diphthongs
three:
, *-uat,
x*uay
and
x*-ual.
In the
column
for Haroi,
the first reflex given shows the reflexes with the lowered onset. The second reflex
given shows the effects of breathy voice, which inmost cases raised the vowels to
/u7 but sometimes
backed
time (see "Haroi vowels
it to lui, typically monophthongizing

them at the same
and restructured register" on page 197, for further dis?
cussion).
PC*o
the exception of at most six words, when the reflexes of PC *-o have ety?
mologies,
they have MK etymologies. Two obviously PMP forms are *do:k 'sit;
stay' and *boh 'fruit; egg; elf. for small round objects', related toMalay duduk
With
and buah, both of which have unquestionable PMP etymologies,

duduk representing a reduplicated form of the root.
Table 81: PC *o words with Malay
PNB
Bahnar Malay
l?mbu
lamo;
PC
Aceh.
Jarai
x*lamo
humo
ramo
The
NR
lamo
PR Cham
limo;
t?rung
x*trog
truag
trog
buah
*boh
boh
boh
boh
duduk
*do:k
dua?
do:?
do:?
pon
to?
empu
*po
po
po-v
po
po-v
*lon
lAn
Ion
two
forms
are
the
far more
lat
'cow;
ox'
lamo
trok
next
theMalay
(or AN) counterparts
ramo -n
*trog
with
crog
trog,
'sit; live; stay'

'master;
l?n
problematic
'eggplant'
'fruit; egg; elf.'
lord'
'earth; soil'
x*lamo
'cow'
and
x*trog 'eggplant', which although found inMP (cf. Malay lembu and t?rung ), are
not widespread
inMP and do not reconstruct back to PMP. The evidence that
these two are borrowings includes the dual reflexes for 'eggplant' found in Phan
Rang Cham and the fact that both forms are found in MK: x*trog 'eggplant'
reconstructs
as *trog in proto-North
Bahnaric
and 'cow; ox' is found in Bahnar.
Chamic
137
Vowels
In both cases, these forms may have been late borrowings into Malay. The last
two words *po 'master; lord' and 'earth, soil' look to be old An roots, cf. Malay
empu
and
PAn
*Calun
and
PMP
'fallow
*talun
land,
secondary
forest'.
Table 82 illustrates reflexes of open syllable PC *-o without An counter?

parts. Among the open syllable examples, only 'cow; ox' and 'master; lord' seem
to have PMP counterparts; in contrast, at least three of these reconstruct to proto
North-Bahnaric.
The length distinction between

in the section on vowel length.
is discussed
the reflexes of PC *-o:- and *-o
Table 82: The reflexes of PC *o

PNB
Aceh.
PC
Rade
Jarai Chru
mato
x*bato
pato
pato
NR
W. Cham PR Cham
pato
pato
pato
dio
klo
kla
bio
x*blo
dio
x*glo
dio
glo
xjro
'split'
-n
jro
?
xjro
x*kro
kro
kro
emo
ramo
luimo
x*lamo
-i
'large
ero
jro
-i
kro
lamo
po
po
po
[hlo]
x*ralo
-v
'dry'
lamo
lamo
lamo;
'cow;
ox'
limo
po
ralo
jar'
'rainbow'
-n
po
'brain'
jro
ero
**po
'teach'
po
ralo
po
-v
ralo
ralo
'master'
'flesh;
meat'
ro
x*tamo
x*tuho
xkuho
tarmo
ka no
kaho
-iv
*c?w
co
ro
'cage'
tamo
'intact'
tuho
kaho
'snare'
kuho
kaho
'Koho'
-v
-iv
cuco
x*cuco
ro
?
taco
co
taco
taco
tico
'grand
-child'
*cano
x*tano
ka no
ta?o
ta?o
ta?o
-n
ta?o
tano
'male'
Two of the forms in Table 82 have Malay counterparts in l?mbu 'cow;

'grandchild', but l?mbu looks to have been independently borrowed
ox' and cucu

into Malay
and
PC,
as
the medial
correspondence
of Malay
-mb-
to Acehnese
-m
is unexpected.
The majority of the *o vowels lack any sort of PMP etymology. And,
although theMK etymologies have not yet been worked out, it appears thatmost
of these forms entered PC through MK sources. Nonetheless
there is a minority
that appear to have come from PMP forms with *u.
138
Chapter
PC VOWEL LENGTH
length distinctions are amajor source of controversy, with the sole
of
agreement being that a vowel length distinction existed in the mainland
point
Chamic languages. Both Shorto (1975) and Cowan (1974; n.d.) claim that the
long vowels inmainland Chamic correlate with diphthongization patterns inAce?
The PC vowel
hnese?a
position that Collins (1991:116) disputes, while asserting that the Ace?
and the mainland Chamic vowel length distinctions are
hnese diphthongization
distinct phenomena occurring independently within each language, presumably
also having developed independently within the various Chamic languages.
In contrast, Shorto (1975) and Cowan (1974; n.d.) not only correlate the
diphthongization
length distinction of mainland Chamic with Acehnese
but go much, much further, asserting the PC vowel length distinctions are not the
vowel
at the PC level but instead are themselves inherited from

results of developments
PAn. Shorto argues, apparently on the basis of general beliefs about what can be
and mainland
borrowed, that the vowel length distinction found in Acehnese
Chamic
does not result from borrowing

.. .the creation of so fundamental
(1975:90):
a distinction
as that of vowel
length by
borrowing seems intrinsically unlikely. On the other hand, contact with
MK languages, in many of which (as in PMK) a vowel-length distinc?
tion
operates,
ceivably
would
favour
the
retention
of
an original
distinction
con?
lost elsewhere.
That is, like Cowan, Shorto suggests that PAn had an original vowel
tinction which was retained in Chamic but lost elsewhere.
length dis?
reluctance to accept that a length distinction could have been

have more merit if the PC speakers were totally unfamiliar with
the population that spoke PC was not solely Austrone?
vowel length. However,
sian speaking. The massive incorporation of MK vocabulary and the restructuring
Shorto's
borrowed would
lexicon along MK lines strongly suggest that PC was spoken by popu?

that
lation
included a large number of originally MK speakers who had shifted to
PC along with an originally Austronesian
speaking population that was itself
the
vowel
For
such
length distinctions would have been
bilingual.
speakers,
of the An
brought into PC as part of the borrowed MK

inent part of the PC vocabulary.
lexicon that constitutes
such a prom?
As part of their proposed scenarios for the retention of an earlier PAn

in PC, both Cowan (n.d.:3) and Shorto (1975:100)
length distinction
was
that
PC
the
able to retain its distinctive length while itwas for
thesis
present
the most part lost in the rest of An, because PC remained on the mainland in con
vowel
Chamic
139
Vowels
languages with a length distinction, while the remainder of the An

the mainland, breaking off contact with MK languages, and thus
left
languages
lost vowel length. One of the many problems with this position is that the bulk of
modern scholarship views the Chamic speakers as having returned to mainland
tact with MK
from the islands within
the last two thousand years or so?a contention that is

of PC, which shows PC as a subgroup with
fully supported by the reconstructions

a relatively shallow time depth.
Collins (1991:116) rejects both the contention that vowel length in PC is

the retention of a PAn vowel length distinction and the claim thatAcehnese diph?
thongization patterns correlate with PC vowel length. Collins is certainly correct
in stating that PC vowel length was not inherited from PAn; in fact, much of this
section on PC vowel length is devoted to showing how PC length distinctions
evolved out of the interaction of PAn materials with MK contact and borrowings.
based in large part on the materials available to him at the
Undoubtedly
time, Collins' suspicion that the Acehnese diphthongs are not correlated with PC
vowel length has not held up. Once the various late MK borrowings have been
culled out, and the PC vowel length reconstructions established, it becomes clear
that the correlation between PC vowel length and Acehnese diphthongization
is
fully documented,
Cowan
and
Shorto
and, with many

now
accounted
of the potential
for.
Further,
exceptions
as Cowan
noted
mentioned
more
than
by
once
the transition from a

1974), for many of the forms inmodern Acehnese,
a
to
modern Acehnese diphthong is attested, that is, docu?
stressed, long vowel
mented within the history Acehnese
(1974:188):
(1948;
Now
it had been realized
long ago, on the evidence of the related lan?

and
the
loanwords, as well as the archaic spellings in the Arabic
guages
script and local dialect forms, that in very many cases the Achehnese
diphthongs derive from simple vowels in the final, i.e., stressed sylla?
bles. Thus it appeared that u/a derived from a in closed syllables, excep?
tionally (in PAn *apa if this reconstruction is correct) also in an open
syllable; oa and ia derived from /, and sd and ua from u, in the latter
two cases according to whether the syllable is open or closed (originally
closed).
Cowan
(1974:189)
continues,
noting
...that the original simple vowels must have been lengthened before
diphthongizing. This conclusion seems to be confirmed by the fact that,
with certain morphemes,
forms with the simple vowel still exist side by
side with forms with diphthongs.
In these cases the former are used
140
Chapter
without
and the latter with
first produced non-phonemic
stress and the implication is that this stress

lengthening and then diphthongization.
Thus, unlike Collins, Cowan and Shorto view the transition from long vowels to
diphthongs as straightforward with the correlation uncontroversial.
The PC vowel length distinctions are readily reconstructable, although
the earliest system that can be reconstructed looks to have been oddly configured:
In the reconstructed
system,
only
PC
*-u-,
*-o-,
and
*-a-
show
contrastive
length,
and then only before certain finals. As Lee (1966:117) noted, the "length contrast
seems to be fairly certain for *a, *u, and *o, but (as is true of the daughter lan?
guages) is limited to certain environments."
is a subpattern to the way in which vowels show length before
the
high vowel *-u- has a length contrast before glottal stops and
specific finals:
velar nasals; the mid vowel *-o- has a length contrast before glottal stops, velar
There
*-a- has a length contrast before glottal stops,
nasals, and *-k; and the low vowel

velar
ally,
and
nasals,
*-k
as well
as before
four
other
finals
(*-l,
*-r,
*-n,
and, margin?
*-t).
Figure
*-u-
13: Long versus short vowels and their environments
vs.
*-u:-
*-o-
vs.
*-o:-
/_.__?
/.___?
/....?
/.?n
/..__g
/____g
*-a-
vs.
*-a:
/__? k /??k
/_.._r
/....t
Perhaps, as suggested by length distinctions in Rade, a distinction may once have

it will never be possible to
existed between -a:m and -am as well. However,
reconstruct a full, balanced system of length contrasts as one most likely never
existed ?
length contrasts were probably always asymmetrical.
contrasts directly reflects
This highly-skewed
system of vowel-length
distribution matches the types of asym?
typologically-marked
in which the back vowels show
MK
of
in
the
Vietnam,
metry
languages
more distinctions
than in the front vowels (G?rard Diffloth, p.c.). The system
itself is not just found in the mainland languages, but is also directly reflected in
MK
contact. This
found
the inherited long vowels

the diphthongization
(inAcehnese
patterns of Acehnese
another
of PC subsequently diphthongized),
yet
piece of evidence that
providing
is simply another Chamic language.
Acehnese
Chamic
141
Vowels
It is reasonably clear just how the system of contrastive length came into
In
being.
general terms, the vowel length contrasts in PC reflect the interaction of
three major
factors
and
one:
one minor
the major
factors
are
contact
intense
with
languages with vowel length contrasts, the incorporation of numerous MK

loanwords containing such contrasts into pre-PC, and internal developments with
the material inherited from An.
MK
the long and short PC *-u- and *-o- the incorporated material bor?
seemed to play the central role in the phonemicization
of the
With
rowed from MK
change, especially in the case of *-o-, where all but a handful of the forms con?
taining the sound, regardless of length, areMK in origin.
With the length distinctions associated with PC *-a-, however, the much
came about through the internal-developments
in the inherited
as through as through contrasts with incorporated pre-PC MK loans.
in the reflexes of PMP *-a
Statistically, the prime catalyst was the developments
of the distinction
material
and
<
*-a-
-e- >, which
established
a distinction
between
PC
*-a-
and
*-a:-
in
inherited words. The contrasts developed through this internal mechanism were
augmented by the incorporation of countless MK -a- forms into Chamic, each
see "PC *-a
entering with either a long or short vowel. For further discussion,
on page 146.
factor has to do with required syllable weight. Most PC con?
are disyllabic or sesquisyllabic
(syllable and a half). However, some
length distinctions"
The minor
tent words
PC content words
are monosyllabic.
In PC, all monosyllabic

content words have
long vowels. Sometimes the long vowel is etymologically
expected; sometimes a
short vowel would otherwise be expected and the long vowel simply reflects the
requirement
that, for those vowels with
monosyllabic
content
word
be
long.
Thus,
a PC length contrast,
*do:k
'sit',
*ya:p
the vowel
'count',
and
of a
*pa:t
'four' all have

meet
long vowels precisely because otherwise these forms would not

the minimal
requirements for a content word. PC is not unique in this
the requirement that monosyllabic

content words be bimoraic is wide?
in
Austronesian
For
instance, in the Isbukun dialect of Bunan
(Blust, p.c.).
spread
a
Paul
from
class
Formosan
Li,
(data
notes),
language of eastern Taiwan, in the
respect;
and nuum 'six'/

pairs paat 'four' but saspat 'four (for counting people)'
lalabnum 'six (for counting people)',
roots are
the vowels in the monosyllabic
in
as
the
but
short
elsewhere.
roots,
secondarily long
monosyllabic
expected
PC *-u- length distinctions
PC *u has a limited and asymmetrically-distributed

vowel length distinction: *u
is found both short (Table 83) and long (Table 84) in two environments: before
142
Chapter
final -? and before final -g. Etymologically,

PMP sources and from MK sources.
the
Among
reflexes
of
the
the short vowels descend

*-u-
PC
short
there
vowel,
both from
are words
with
PMP etymologies
'flour', and 'nose') mixed in with
(apparently, for example,
words with established MK backgrounds (Table 83). No doubt exists that PC *-u
came
from
both
sources.
Table 83: Reflexes of PC *-u

PNB
Bahnar Malay
jalu
bak?
PC
Rade
NR
*?a?u?
aft??
a???
'beads'
*bitu?
mat??
'star'
*katug
'pull'
x*?u?
kat?g
???
pitu?
katuk
???
'dive;
*?adu?
ad??
adu?
'room'
el??
jalu?
man??
'bowl'
'chicken;
babu?
'drunk'
*jalu?
manuk
*manu?
mabuk
*mabu?
man??
*gulug
-guluk
'to roll'
hidung
*?idug
id?k
'nose'
jantung
*tug
tuk
'stomach;
gulung
ad?g
-i
l?sung
*lasug
rombong
*bug
b?g
*?anug
an?g
?
anug
?
tapug
*daqb?g
*gad?g
kadug
-i
risuk
es?g
'mortar'
'large
basket'
'package'
x*rfug
dug
duk
t?pung
*tapug
kap?g
r?bung
*rabug
eb?g
tupuk
rubuk
kandung
x*kadug
dug
fowl'
abdomen'
'heart'
submerge'
'wrap
up;
bundle'
'flour'
'bamboo
shoot'
'pocket; pouch'
*-u?
-??
-u?
*-ug
-?g
-uk;
-ug
intriguing and indicative of the amount of work that still remains

to be done are the words that appear in both MK and PMP, such as 'pocket;
pouch' and 'flour'. For these the direction of the borrowing remains to be deter?
More
mined.
Chamic
143
Vowels
In contrast to short PC *-u-, the reflexes of the long PC *-u:- all seem to
descend from MK sources (Table 84). The form ribut 'storm' is intriguing in that,
in addition to occurring in Bahnar (MK), it also appears inMalay. In wider Aus?
tronesian, though, ribut is restricted geographically
is a likely loan.
Table 84: Reflexes
PNB
Bahnar
to western
Indonesia and so
of PC *-u:
PC
Aceh.
Rade
PR Cham
?
x*?amu:g
amug
'bunch;
stalk'
amug
'snout;
muzzle'
?
x*?amu:g
*bru:g
x*ka?u:?
brug
?
*og
ug
x*?u:g
?akug
xcaku:g
*pu:g
'streaked;
kakug
juimpug
kag??
t?? a
x*tagu:?
x*tu:?
'worried;
sad'
'husband;
male'
apyag
'carry (several)'
'straw (rice)'
tako?
'get up; stand up'
to?
'to receive'
cakog
'descend'
xyu:?
d?g
striped'
'float'
xdu:g
dug
gut
x*gu:?
g??
l?u?
'below; bottom'
habut
*ribu:?
eb??
ripu?
'storm'
x*ju:?
j??
cu?
'black'
x*pu:?
p??
*-u:?
-??
-u:?
-ug
-u:k
p?k
*-u:g
-ug
'carry
in arms'
For PC *u, the vowel length distinction is the direct result of the incor?
the short vowels occur freely in both inherited
poration of MK material. While
PMP and incorporated MK words, the long vowels occur overwhelmingly
only in
words from MK sources.
PC *-a- length distinctions

the distribution
and the origins of PC *o are of interest. Like all Chamic

the
distribution
is anything but symmetrical. The PC *o occurs
length contrasts,
long and short before final -?, final -g, and final -k.
Both
144
Chapter
Table 85: Reflexes
PNB
PC
*kasuq
x*ho?
of PC *-o
Aceh.
NR
Rade
PR Cham
kaho?
east'
'upgrade;
xcho?
ko?
x*?ako?
cho?
?ho?
ako?
ako?
'scoop up; ladle'

'head'
x*?o?
'vomit'
x?anrog
-f
anro?-f
x*dhog
dhog
thok
x*prog
prog
prok
x*salog
hlog lar
x*yog
*(ka)r?g
xrog
rua g
*tr?g
x*trog
truag
x*pok
aro?
thog
prog
klog
'lift; take off
rog
rog-1
trog
trok
trog
'eggplant'
pok
po?
x*bok
-o?
-uag
*-ok
'forever'
'back
'to
-ok
(anat.)'
open'
'to cry'
"disk
shaped'
ho?
'pour
out;
bo?
po?-v
'to
-o?
-o?
hok
*-og
W
-i
-turok
x*tulok
*-o?
'knife'
yok
co?
x*hok
'toad'
y^g
x*cok
*?k
bleed'
'sweat;
x*go?
*h?k
-ok
-?g
-o?
-o?
spill'
swell'
The vowel itself was almost but not entirely borrowed. Undoubtedly,
it
does occur in one native PMP form, the form *do:k 'sit', found throughout PMP
and related toMalay duduk. Intriguingly, I suspect it is the word's monosyllabic?
ity that accounts for the vowel length in this form. The forms x*lamo 'cow; ox;
cattle' and x*trog 'eggplant, although found inMalay as l?mbu and t?rung, are
not widely attested in An and are most likely early borrowings. If not, there are
three PMP words with PC *o; if these two are borrowings, there is but one native
PMP form left.
mation
mentioned,
The remaining words are apparently borrowings, some predating the for?
of PC and some introduced later. Aside from the three words already
none
of
the words
four of the forms occur

and two more
in Aymonier
occurs
widely
attested
in Austronesian.
In contrast,
in Smith's
(see Table 85)

(1972) proto-North-Bahnaric
and Cabaton's (1906) dictionary of Cham (see Table
Chamic
145
Vowels
85). The -nr- cluster in 'toad', the glottalized stop in xcarfog -f 'flat basket', and
the initial ch- of 'scoop up; ladle' mark these as borrowed. Headley (1976) sug?
gests that x*yog 'to lift; take off and xrog 'back (anat.)' are MK in origin, while
Durie (1990a: 106), citing Cowan (1983:177), labels x*go? 'upgrade; above; east'
a MK
may
is, of 17 forms, at least 10 of them have MK

(ormay not) have an Austronesian heritage.
loan. That
of PC *-o:
Table 86: Reflexes
PNB
Bahnar
PC
Aceh.
NR
Rade
-so:?
tho?
x*ko:?
ko?
ko:?
ko? -1
x*mo:?
mo?
m?:?
s?ap
-f
b?k; b? x*?o:?
x*kho:g khuag
?
krog
x*kro:g
r?g
x*ro:g
bog
x*bo:g
xjo:g
*cug
?
*k?n
?
x*glo:g
*k?g
x*ko:g
-
*log
xho:g
*t?g
-f
?o:?
?o?
'face'cf.'nose'
tog
ato:k
ato g
'beat (gong)'
khog
kho:k
khog
'dry (weather)'
krog
kro:k
krog
'river'
x?anro:g
pr?k
l?k
x*pro:k
*do:k
With
the Bahnar
Chamic
forms
bog
pay
'coffin;
jog
cog
'axe'
dlog
jo:k
dlo:k
kog
ko:k
'tall; high; big'
kog
'bracelet'
hog -f
-v
test'
prove,
'wasp'
anro:k
anog
'carry (on pole)'
lok
lo:?
pro?
lo?
'to peel'
dok
do:?
to?
'sit;
-o:?
-o?;
enog
prok
plua?
dua?
casket'
klog
'try,
hog
x*-o:?
dence forMK
'raise;
x*lo:k
nourish'
rog
bo:k
?
anog
'wife'
log
x*lo:g
*og
rog
?
'lungs; placenta'
'white'
?o?
?
x*?ato:g
PR Cham
kaso?
x*so:?
*bog
one
affiliations;
-o?;
-o?
x*-o:g
-uag
-og
-og;
x*-o:k
-ua?
-ok
-o:?
-o:k
'squirrel'
live;
stay'
-5?
-og
-o?
the long vowels, aside from the one obviously PMP form, the evi?
origins is impressive. All but six of these forms occur in PNB or in
recorded
with
by Aymonier
long
-o:-
as MK
and Cabaton.
in origin:
Headley
'coffin;
(1976)
casket',
labels
'axe',
six
'river',
146
Chapter
'wasp', and 'bracelet'. The remaining six, while not yet established as
borrowings, certainly lack obvious PMP etymologies.
For PC *o, both the vowel and its length distinction are the direct result
'squirrel',
of contact with MK, as both the vowel and its length distinction
almost exclusively
inmaterial incorporated from MK.
are present
PC *-a- length distinctions

Although for PC *-u- and PC *-o- the PC vowel length distinction would at best
be marginal without the forms with long vowels incorporated into Chamic from
MK sources, the case is quite different with the length distinction associated with
the influence of MK contact played a role, the length contrast in
PC *-a- is fully and richly manifested
in inherited Austronesian
forms, largely
due to the manifestation
of PAn *e [a] as PC *-?-, while inmany cases PAn *-a
PC *-a-. While
as PC
is manifested
The
evolved
*-a:-.
non-MK
of
component
the PMP
from
vowels
*-a-
the
and
PC
PMP
*-a*-a-.
versus
with
And,
length
contrast
the possible
excep?
*-a:-
tion of several
reflexes
thus far unexplained

cases, these vowels have quite regular
in PC. Cowan (1974) noted that for PMP *-a- forms the PC reflexes are
always short; however, for PMP *-a- both short and long PC reflexes are found.
That is, while the PMP *-a- unexceptionally
merged with the PC short *-a
the
reflexes,
r, -n,
reflexes
-p, and marginally

These
the PMP
*a
before
are
patterns
is -a:-
with
but
syntactic
slot,
which,
-?, -g,
Before
exceptionless.
in the word
unless
final
-k,
-1,
long.4
totally
not
before
Specifically,
*a occurs
however,
is inexplicable
in an unstressed
ture associated
split.
-t, PMP
not,
in ten words,
reflex. This exception

frequency
*a are
of PMP
'name',
where
'name' occurred with

in turn,
prevented
it has
final
-n,
a short
considerable
length?a
fea?
stress?from
developing. Before ^t,PMP became long only

it
occurred
*-h- which
marginally:
specifically,
long in words with a medial
to
in
the
number
four *pa:t
'chisel, plane,
chisel, plane'),
dropped ('bad; wicked',
because it is a monosyllabic
content word, and, for no apparent reason, in *pala:t
-1 'palm; sole'. In the form 'palm; sole' the vowel pattern is irregular, with some
languages having a short vowel reflex, and some a long vowel reflex. However,
aside from this handful of cases, where PMP *a has long reflexes and where it
has short reflexes
4.
is totally predictable
from the final. [Note: The reconstructions
Before *-h, PAn had already merged PAn *-e- and *-a- (Blust 1995b:622
that the inherited Austronesian
forms in
623), with the ultimate consequence
PC did not develop a contrast between PC *-a and *-a:- before *-h. Nor did
such a contrast develop in PC from borrowed MK material.
Chamic
147
Vowels
used in this work are Blust's
for PMP (see References)
and Adelaar's
for Proto
Malayic]
the reflexes of PC *-a- (Table 87) and of PC *-a:- (Table 88),
Among
words both of PMP and of MK origin can be found easily. Numerous forms have
com?
Malay counterparts, many of which are common PMP forms. Others have
mon MK origins.
another
In Table 88, seven examples are reconstructed for proto-North-Bahnaric;

ten are found in Aymonier and Cabaton's Bahnar. Still others such as
x*ba:l 'mend; patch' are found in theMK

However, there are three Acehnese
subgroup PKatuic.
items that require discussion
because
they have a /-an/ final rather than the /-man/ final

*-a:n:
final
'fish', 'cup; bowl; dish', and 'road; path', and
expected from PC
to Durie (p.c.), ikan is aMalay borrowing. As he notes, the
'bowl'. According
usual Acehnese word for fish is wgkot. The second word cawan 'cup; bowl; dish'
is, according to Coope (1976) and Baxter (p.c.), a borrowing from Chinese. This
in one variant or another
analysis is consistent both with its failure to reconstruct within PMP and its irreg?
-lu?an 'yard; space in
ular final. The last word needing discussion is theAcehnese
reflex
of PC *jala:n 'road;
is
the
which
Durie
front of the house',
suggests
regular
path'. The difficulty involves the presence of another form in Acehnese jalan
'highway; main road' which at first glance looks like the appropriate cognate. On
closer inspection, however, -lwan looks to be the cognate form and jalan looks to
be a later borrowing from Malay. As Durie points out, -lvaon corresponds quite
regularly phonologically, while the final of jalan -v is irregular. Semantically, the
jalan is interesting: jalan, as Durie notes, is used only marginally to refer tomain
roads and highways, while ret/rot is the normal word for road. This suggests that
the semantic range of the original reflex of PC *jala:n > -lwon 'yard; space in
front of the house' was narrowed under the influence of the word ret/rot, with the
jalan borrowed only later from Malay.
The fact that a length distinction between PC *-a-
modern Acehnese
(Table 87) and PC

*-a:- is found quite readily both in the inherited PMP forms and in the incorpo?
rated MK material suggests that the length distinction comes from two sources:
one, it was developed
internally from the inherited PMP material (as described
above) and that it was
also present
in the material
incorporated
from MK
lan?
guages.5
5.
However, note that the contrast before PC *-r is found primarily in incorpo?
rated MK material, as most of the inherited PMP final -r'swere lost.
Table 87: Reflexes

PNB
of PC *-a
Bahnar Malay
PC
lama
l?mak
x*lama?
tak
t?tak
*ta?
masak
*tasa?
tanak
*tana?
?
tag?k
*tagak
*gatak
arag
*qm?g
arang-
Aceh.
Rade
luima?
?
-f
?
arag-m
*hadag
x*?ag
?
lag
pag
*glag
em??
lam??
kan??
tarn?
kag??
kat?k
kat??
had?g
hat?g
'sap; resin'
'charcoal'
??g
dl?g
??g
'hole; door'
klag
'look at; watch'
'cleaver'
h?g
x*hag
*kh?n
kl?n
kain
?
*k?t
k?t
'name'
an?n
x*klan
lhan
bantal
*bantal
bantai
g?tai
*gatal
gatai
s?ndal
x*dal
*baqar
*-d?r
nan
x*khan ?
x*ba?ar
a wall'
'hot; spicy'
'sickle'
h?g
w?g
w?g
*?anan
kh?n
'ripe; cooked'
'cook'
'to make
*wag
?
cut'
tath??
?
u?g
kh?n
'cloth; blanket'
tl?n
kl?n
'boa; python'
pat?r-f
kat?l
'pillow'
kat?l
rf?l
'to wedge'
maar
*dar
pi?ar
ikat
*?ikat
ikat
surat
*surat
surat
hr??
'encircle'
ik??
'to
har??
'write;
tap?t
t?pat
*tapat
tiupat
ara
urat
*?urat
urat
kawat
*kawat
kap??
tap??
aru?t (m) ur??
kaw?t
sukat
*sukat
sukat
mak??
pusat
*pusat
pusat
m a s?t
ulat
*hulat
ulat
*_a7
-at;
*-ak
-a?
*-ag
-ag
-?g
*-an
-an
-an;
*_al
-ai
-?l
-at
-?t;
letter'
honest'
'vein,
tendon'
'wire'
'to measure'
path??
hal??
'navel;
center'
'worm'
-??; 4?
-??;
-?k,
tie'
'straight;
-i
hlu?t (m)
-a?
'itchy'
'paper'
dar
oil'
grease,
kas??
*paO
*h?g
anan
'fat,
'chop;
?
mag
PR Cham
-??
-a?;
-ak
-?g
-an
-an
-?l
*-ar
*-at
-a?
-??
Table 88: Reflexes of PC *-a:

Bahnar
PC
Aceh.
Rade
PR Cham
?mpat
anak
*pa:t
*?ana:k
puiat
anu??
p??
pa?
ani?
'four'
minyak
*mif?a:k
mita?
mini?
'oil'
x*?awa:k
awuia?
awak
awa?
'spoon; ladle'
jak
?a?
'invite'
Malay
uak
j?k
lag
anak
'child'
ajak
*ja:k
pinang
*pina:g
pinuig
manag
panig
'betel(-nut)'
hudang
*huda:g
uduiag
hadan
hat?g
'shrimp'
x*ka:g
kuiag
yang
lag
x*ya:g
*la:g
orang
?ag
*?ura:g
uruiag
kag
kag
'chin; jaw'
yag
yag
'spirit; god'
lag
lag
'spread
urag
'person'
cog
'wait'
ar?g
*ca:g
tulang
*tula:g
k?rang
*kra:g
lubang
*luba:g
klag
tuluiag
-1
out'
'bone'
talag
'clam'
kruiag
dag
x*tfa:g
lipag
'hole; pit'
dag
'lie
suppine'
lipan
*limpa:n
limpuian
epan
lipan
'centipede'
hujan
bulan
*huja:n
ujuian
najan
ha?an
'rain'
*bula:n
buluian
m?an
pilan
'moon'
dahan
*dha:n
dhuian
adhan
than
'branch'
kagan
ta gin
'hand;
arm'
elan
calan
'road;
path'
tang?n
*taga:n
jalan
*jala:n
-lu?an
tampal
x*ba:l
pagar
*paga:r
p agu? a
akar
*?ugha:r
ukhuia
tawar
*taba:r
tabuia
jahat
*jaha:t
juihuiat
pahat
pal
'mend'
paka
'fence'
agha
ukha
'root'
kaba
tapa
'tasteless'
jhat
?ha?
'bad; wicked'
pha?
'to chisel'
*pha:t
phuiat
*-a:?
-uia?;
-uiat
-??
-a?
*-a:k
-uiak;
-uik
-ak
-a?;
*-a:g
-mag;
-uig
-ag
-ag;-?g
*-a:n
-uian
-an
-an;
*-a:l
-uia
-al
-al
*-a:r
-uia
-a
-a
*-a:t
-uiat;
-at
-a?
-uit
-i?
-in,
-in
150
of
Then,
there
course,
are
a number
these are transparently borrowings

length in the Chamic languages.
Table 89: Borrowings
amra
*brak
*relat
borrowings
*-a:-,
but
containing
so these forms do not obscure
the analysis of
with *-a:
Bahnar Malay
PMnong
of
Chapter
m?rak
?
PC
x?amra:k
Rade
PR Cham
muira?-v
amrak
amra?
lat
klst-ivf
xla:t
jaga
Acehnese
xja:g
jaga -f
'peacock'
'flat'
'to guard'
First, not one of the three forms in Table 89 reconstructs to the PC level. In addi?
tion, all three forms are attested inMK languages, although 'guard' is apparently
ultimately Sanskrit (Coope 1986). The word 'peacock' has a medial /-mr-/ only
(1976, #1.47) identifies it as a MK loan, it
borrowings. Headley
as
reconstructs
*brak in Proto-Mnong,
*bra:? in PSB (Efimov), and *ria? in PKa
In the case of 'flat' the
tuic, and it has an irregular vowel reflex in Acehnese.
forms were borrowed from MK. The word reconstructs as *relat in PMnong,
inMK
found
is the likely source of the forms scattered throughout the mainland Chamic
is a quite unre?
languages. The lookalike form rata found inMalay and Acehnese
which
lated borrowing from Javanese (see Appendix

II). As for 'guard', Coope (1986)
comes
which
in
that
it
from
Sanskrit,
suggests
light of the similarity between the
and Acehnese,
Malay
on
the
one
hand,
and
the PC,
on
the
other,
looks
at
least
and Malay cannot

possible. What looks definite in any case is that the Acehnese
be directly related to the PC forms; the Acehnese
looks like it came from Malay,
with
the question of the ultimate origin of theMalay
The Acehnese
correlates
being left open.
of PC vowel length
The Acehnese
data in the above tables is of particular interest. The claims in the

in PC correlates
literature by Cowan and by Shorto that the length distinction
with certain Acehnese
vowel distinctions are substantiated by a careful compari?
son of Acehnese
vowels with these PC length distinctions. The distinction
found in the tables
long and short *a is clearly and unambiguously
above for the other two long vowels PC *-o:- and PC *-u:-, more Acehnese data
will be needed to determine whether the distinction is preserved there too.
between
PAn stress and vowel length compared with PC vowel length

The data discussed
in this section have shown that PC vowel length distinctions

are not inherited from PAn but instead represent a secondary development, and,
as a secondary development,
appear to have nothing to say about whether or not
Chamic
151
Vowels
stress (with resultant vowel length distinctions in various languages as a reflex)

needs to be reconstructed for PAn. Certainly Zorc (1978, 1983) has shown that in
certain Philippine languages the long/short vowel length distinction would corre?
late with the earlier existence of stress, and, as Ross (1992:47) writes, "Recent
work suggests that Zorc's system is reconstructable for PAn, since there are relics
of it in a few Formosan languages." Zorc and Ross are not alone in their sugges?
tion that PAn had a stress distinction; Wolff (1991), Peiros (cited in Ross
and Benedict (p.c.) also suggest reconstructing stress for PAn.
1992),
However, PC vowel length and the proposed PAn stress system (which
correlates with vowel length in certain Philippine languages) have no correlation
whatsoever with one another. First, as Ross (1992:47) notes, in the Philippine
languages "the long/short distinction is limited to the vowel of the penultimate
syllable", while the PC distinctions are limited to the ultimate syllable. Second,
and more importantly, the PC vowel length distinctions between long and short
*a correlate directly with the earlier distinction between PAn *e [a] and *a rein?
forced by the borrowing of MK words with a length distinction, while the length
distinctions between long and short *u and *o depend almost on length distinc?
tions between originally inherited and originally borrowed material.
THE MAIN SYLLABLEVOWELS SUMMARIZED

It goes without
fied. For
dences
saying that everywhere the details remain to be filled in and clari?

instance, much more can be learned about the Acehnese
correspon?
from
the
incorporation
of
a more
extensive
number
of
forms.
Similarly,
better understanding of borrowings will contribute to a better understanding of

the systemic interactions between the PMP and theMK heritage.
the outlines of the history of PC vowels seem clear. The PC
Nonetheless,
vowel system consists of a core of elements inherited from PMP supplemented
and enriched by MK borrowings. In addition, the subsequent reflexes of PC vow?
els in the various daughter languages is also straightforward, with the improve?
ment upon the foundation laid by Lee 1966 and others made possible by an
of which forms were borrowings and by a greatly
expanded understanding
expanded data base.
From these patterns we can learn something both about the nature of the
earlier cultural contact and about the influence of language contact on vowel sys?
tems. The intensity of the early contact between MK speakers and the pre-PC
speakers is attested to by the richness of the borrowed component of PC. The
effects of language contact are attested to by the restructuring of the original
PMP disyllables
into the iambic morphemes of PC and by the incorporation of a
number of new vowel distinctions
into the linguistic system.
and Nasalization
Nasals
Both
consonants
nasal
and
nasalized
vowels
are
for PC.
reconstructed
For
nasal?
are straightforward and based on regular

correspondence patterns. In fact, only two developments are worthy of comment:
In Cham, word-final nasalization seems to trigger simplification of complex vow?
the reconstructions
ized consonants,
els. In Roglai
and Tsat,
consonants
there is widespread
that
change
has
two
denasalization
of word-final
it makes
consequences:
late
loans
nasal
with
nasals conspicuous as they fail to undergo the change and it provides

strong subgrouping evidence for the pairing of Tsat with Northern Roglai.
For nasalized vowels, the bulk of the evidence comes from just three lan?
word-final
Chru
closely-related
guages?the
and
Northern
and Phan Rang Cham. While

the data itself, sometimes about its interpretation,
are
cases,
there are sometimes
from Western
forms
in certain
and,
Roglai,
loans,
there
are,
nonetheless,
certain
patterns
questions about
and sometimes about which
which
seem
clear.
NASALIZATION INPC
There are PC nasalized
these cases,
vowels
the reconstruction
not associated with adjacent nasal consonants. In

of PC nasalized vowels is based on the nasaliza?
in Chru, Roglai, and Haroi forms, and, in those rare Phan Rang
a reflex that would indicate prior nasalization (see "Nasaliza?
tion in Cham" on page 155, for a discussion of the Cham reflexes).
tion recorded
Cham
forms with
vowels. In Chru, all

vowel; inNorthern Roglai all of the
inHaroi about half the forms have nasalized vow?
Table 90 includes words with reflexes of nasalized

but one of the forms has an actual nasalized
forms have nasalized
vowels;
els; and, in Phan Rang Cham,
every word
has evidence
of prior nasalization,
152
153
and Nasalization
Nasals
either a vowel reflex characteristic

thong or triphthong simplification
Table 90: Reflexes
PC
Aceh.
or in the pattern of diph?

nasalization.
earlier
that indicates
of earlier nasalization
of PC nasalized
Chru
vowel
NR
Haroi
PR Cham
ma?ia?
ma?i??
ma????
mii?-v
'urinate'
*mah?w
mah?u
mah?u
mah?u
mihu
'thirst'
*kr?h
kr?h
?ia?
*ma?iak
*h?
khr?h -i kr?h
h?
h?
h?i -v
krih
'middle'
hi
'you; thou'
'work; do'
x*brua?
brua?-n
bru??
x?i?w
i?u
i?u
pro?
eau
pr??
iw
sr?:u?-l
sr??
sr?u?
thru?
'crossbow'
jr?u
cari?u
cru
'medicine'
thr?h
'nest;
kath?r
'porcupine'
?
xsr?p
?
xjr?w
jr?:u
?
xsr?h
-f
xkasu?r
sr?h
sr?h
kasu?
-f
kasu?
srouh
-f
?
xji? ji?
Notice
kasoa
sia
ji?
-v
-v
'left (side)'
swarm'
'taxes'
thatmore
than half of these have been identified as post-PC bor?

of the nasalized vowels are found in borrowings, but not all. Cer?
rowings. Many
tainly in this group,
'thirst; desire' appears to be related to an PMP form meaning
'want'.
NASALIZATION INCHRU
The nasalization
it occurs,
these
in Chru appears to be, etymologically-speaking,

sporadic. When
in the vicinity of initial nasals, *h-, and *?- particularly when
it occurs
co-occur
with
final
*-?;
however,
the
actual
occurrence
is, nonetheless,
not
fully predictable on the basis of correspondence patterns.

In Table 91, there are Chru forms which have corresponding nasalized
vowels in Northern Roglai. With the exception of 'raw; green', the nasalized
either by a nasal consonant or by a glottal stop, suggesting the
In these cases, it is probably nothing
of
possibility
perseverative nasalization.
more than the variable marking perseverative vowel nasalization. The case of
'raw; green' is more surprising; however, itmay be that the nasalization spread?
vowel
is preceded
ing from the earlier homorganic
nasal is still found inAcehnese.
154
Chapter
Table 91: Chru and Northern
nasalization
Roglai
PC
Aceh.
Chru
NR
Haroi
PR Cham
*ku?it
ku?et
ka?i:?
ku?i:?
kaf?ei?
ka?i?
'yellow;
ban?:?
bin??
mano?
man??
man??
man??
pinu?
min??
'banyan, h?lete'
'chicken; fowl'
*banut
*manu?
tam?h
*tamuh
*tamut
x*k-am-lo?
klo
x?amra:k
muir??
??
??
?au
kamlo
kumlo
kamlo
kamlo
amr?:?
amra:?
amra?
amra?
-v
muintah
tam?h
m?:?
tumeric'
'grow; sprout'
'hammer'
gu?h
mugu?h
matan
matan
xam-agu?h
*mantah
cam oh
tum?h
m?:?
*?u
-v
?u
she;
'he,
they'
'dumb; mute'
'peacock'
dawn'
'morning;
matan
mit?h
'raw;
green'
A second interesting set of Chru correspondences

is found in Table 92.
as the tendency for
has
described
appear to exhibit what Matisoff
(1985)
vowels to nasalize spontaneously after glottal stops and Ibl.
These
Table 92: Matisoff
s vowel nasalization
PC
Aceh.
Chru
tuiot
ta??:?
ka?i?g
ka?iak
ka?eag
ka?g
'loins; waist'
e?-f
?h
eh
eh
?h
'excrement'
*tu?ut
x*ka?iag
x*?sh
?
x*ka?u:?
lah??
x*lahi??
Haroi
PR Cham
ca?u?-v
tau?
tu?u:?
ka??:?
x*halua?
NR
with glottal or/h/initials
lah?a?
ka?ou?
hlua?
'knee'
?
?
sad'
'worried;
halwa?-i
?
'to
'sharp'
lose'
The third set of forms (Table 93), however, is not explicable in phonetic
terms, that is, the nasalization on the vowel seems to have no obvious phonetic
cause. What is immediately obvious, however, is the fact that, at the very least,
four of the six forms are borrowed, not native.
Table 93: Unexpected
PC
Aceh.
vowel nasalization
glo
x*glo
?
*kaka:s
x*ruay
?
x*karah
-f
ruai
-n
-n
kr?h
-n
sr?p
-n -f
PR Cham
klo
'brain;
kaka
kak?h
kak?h
'fish scales'
pha:t
aphan
phan
ruai
roai
karah
kr?h
roy
kar?h
dlo
-n
pha:n
?
xsra:p
-n
kark?:h
*bha:n
Haroi
NR
Chru
?
sr?p
-fv
marrow'
'sneeze'
'fly; bug; insect'

'ring'
"tired
of
155
and Nasalization
Nasals
NASALIZATION INHAROI
rare. Tegenfeldt-Mundhenk
are extremely
and
vowels
are
note
vowels
and
that
there
10
nasalized
(1977)
rarely occurring
are
that there
only two examples of each. In addition, there is allophonic nasaliza?
tion: "preceding final -p, -t, and -k, all vowels are nasalized" (1977:2).
In Haroi,
Goschnick
nasalized
Table 94: Haroi

postPC
xsr?p
xtu?i
vowel nasalization
in borrowed words
Aceh.
Chru
Haroi
PR Cham
sr?:u?-l
sr??
sr?u?
thru?
'crossbow'
tuai
thu?i -i
t?ai
toy
'guest; stranger'
jr?u
cari?u
cru
'medicine'
NR
?
xjr?w
jr?:u
The existence
of nasalization
wise unaccountably nasalized words

tion inChru, above).
inHaroi
is interesting in that all the other?

are
inHaroi
borrowed (cf. the similar situa?
NASALIZATION INCHAM
One of the more
striking effects of nasalization occurs in Cham, where nasaliza?

tion appears to trigger vowel deletion. As Lee (1974:655) noticed, inWestern and
Phan Rang Cham PC *a vowels (and nasalized PC *a vowels), either with inher?
ited nasalized vowels or with vowels secondarily nasalized by a preceding nasal
have special Cham reflexes. Sometimes PC *a becomes HI (and,
on
additional conditioning factors, sometimes HI or lui) and, in certain
depending
the
diphthongs,
original /a/ is dropped.
In the first set of examples, an inherited nasalized *a becomes l-i-l. In
consonant,
the second set of examples, a preceding consonant secondarily nasalizes the

vowel, leading to an l-i-l reflex. In the third set of examples, the reflex is III, rather
than /i/; the conditioning factor involved is unclear. In the fourth set of examples,
three words
'striped' (which may be the same root as 'flower'), and

'soul, spirit; shadow') have the reflex l-u-l. This reflex appears to be conditioned
by a combination of their word-final position after a velar nasal and by the fact
('flower',
that the initial *b- makes
the whole word breathy-voiced.

In the next to last set of
it can be seen that some PC *-a forms have also undergone the change.
examples,
In the table in "Reflexes
of PMP *-i- and *-i" on page 117, it can be seen that

some PC *-a became ? in Cham; perhaps the change from *-a to ? occurred first,
with the ? > -i change following. The last set of examples constitutes a problem;
despite
the obviously
nasalized
vowel,
the vowel reflex remains unaffected.
156
Chapter
Table 95: PC *a > -i in PR Cham and Western Cham

PC
W. Cham
*kr?h
krih;kih
-lvn
xkr?m
krim
-v
mamih
*mamah
tami
*tama
krih
krim
'middle; half
-v
mimih
tanih
*tanah
PR Cham
'bamboo'
'chew'
tanih
'earth,
soil'
tami
'enter'
*?ama
mi
ami
'father'
*huma
hami
hamu
*lima
lami
lami; limi
'five'
*?ina
ni
ini
*panah
panih
panih
'mother; major; big'

'shoot (bow); a bow'
*pina:g
*?ana:k
panig
ni?
panig
ani?
'betel (-nut)'
'child'
*lama:n
lamin
limin
maf?i?
mifti?
lagi
ligi; lagi
'sesame'
ta?i
'ask'
tag?n
'hand'
tagi
'ear'
4
sky'
*mif?a:k
*laga
-v
*taga
tagan
*taga:n
-v
*ta(li)ga
*lagit
*ma?am
*buga
lag!:?
-1
-v
'field'
'elephant'
'oil'
ma?im
ligi?
minim
pagur -f
pigu
'flower'
pigu
'striped'
twill'
'weave;
?
*buga
*bugat
pag??
pig??
'soul,
x*b-an-at
pana?
pini?
'a dam;
x*mag
mag
m?g
'from'
kan?m
'cloud'
?
*gunam
*nam
n?m
n?m
'six'
*?anan
nan
n?n
'that
spirit;
shadow'
a fence'
(third
p.)'
The direction of the nasalization is also of interest, being perseverative

rather than anticipatory, a type of nasalization reminiscent of the nasalization in
Malay and in other Austronesian
languages of the region (cf. Court (1970)).
157
and Nasalization
Nasals
In addition, throughout the data, cases can be found inwhich themodern

Cham reflex is an /a/, phonetically
[a], rather than ill. These cases are examples
a
I'll was subsequently
in
which
of further reduction,
secondarily developed
reduced to a /a/. An examination of the older Written Cham forms inAymonier
and Cabaton makes clear, many of the forms with an orthographic <a> inModern
Cham had an ill in the earlier records.
Finally, after *m- in the presyllable, the vowel reflex is almost excep
tionlessly HI, rather than ill, effectively neutralizing the earlier four-way vowel
distinction. The patterns suggest that the various vowels were first reduced to
l-i-l under the influence of nasalization.
shwa and then later became

this change
described
in 1901.
Table 96: After Cham presyllable
PMP
Cabaton
Malay
PC
NR
PR Cham
mamAh
mum?h
mimih
'chew'
mate
matai
mitay
mi??m
'die'
Aceh.
*mamaq
mamah
*ma-atay
mati
*matay
*inum;
minum
*minum;
*inem
*mamah
minom
'to drink'
*minam
mata
*mata
mata
mata
mita
'eye'
*manuk
manuk
*manu?
mano?
man??
min??
'chicken;
*ma-qasin
masin
*masin
masen
masit
mithin
'salted;
*mata
fowl'
salty'
(?)
*ma-esem
masam
*masam
masam
m asap
mith?m
'sour;
*m-uda
muda
*muda
muda
mida
mita
'young;
that the original presyllable

attested in both Malay and PMP.
Note
Nasalization
and vowel deletion
In addition
to those environments
there are also

deleted.
several cases,
In Lee's
nasalization
terms,
"In
vowel
unripe'
as is
in Acehnese,
in diphthongs
in which
discussed
two
is still preserved
vinegar'
other
by Lee,
in which
are
there
environments
in Cham. One of these environments
reduced to a HI,
a nasalized *a was
*a was
a nasalized
is where
traces
of
earlier
*? is preceded by *i or
*u as part of a complex
peak" (1966:120).
Interpretation of the historical origins of these changes, however, is not
as clear as it was in the case of a nasalized *? being reduced to a HI. Lee pointed
out the connection with nasalization,
but other factors are also obviously
involved. One factor is the non-Chamic-like
syllable structure: all of these forms
involve originally
closed syllables were also originally
heavy.
158
Chapter
is abundantly clear is that all five forms cited in Table 97 are post
borrowings. None of these forms shows up inMalay or PMP, but the first
are
four
attested in either PSB (Efimov) or in PKatuic, and the one remaining
form is also apparently a borrowing. One suspects that, in addition to simple
What
Chamic
nasalization, there may also have been some adjustment

more restricted syllable canon of Cham.
Table 97: Cham forms with lost nasalized
P-Katuic
PSB
post-PC
NR
*kA:ji
xkhiag -fi
khiag -f
xkasu?r
kasu?
?
*agkAji
?
*kamhuar
xmu?r
*brua?
*br_q
-f
bru??
x*bru??
An examination
'want; desire'
kh?g
kasur
kath?r
mur
mu
pr??
'porcupine'
-f
'termite'
'work; do'
pr??
?
kr??
xgru??
in both Western
PR Cham
khin -f
mu?
-f
to fit the
vowels
W. Cham
-f
of borrowings
'lie
prone'
it clear that the vowel deletion occurs
of Table 97 makes
and Phan Rang Cham.

second environment for deletion
*? preceded a following
*u (Lee 1966:120). In this environment, Lee noted that the *? was lost except
when it followed a main syllable nasal, as in *naw 'go'. Rephrasing Lee's
Lee's
a nasalized
description,
immediately
followed
was
vowel
a nasalized
Table 98: Loss of nasalized

PC
Aceh.
?
*mahaw
lost when
Chru
NR
W.Cham
PR Cham
mah?u
mah?u
mahu
mihu
cru
cru
ira:u
x?jraw
da?o
*danaw
?
maw
*boh
*naw
but
not
when
it
/a/ before /u/ in Cham
jr?:u
?
a *u,
it preceded
consonant.
?
xjr?w
is where
jr?u
-1
<
?rau
?jrau
'thirst;
'medicine'
'bamboo
*jr-
-n ?
tanaw
danau
danau
bamau
bum?u
pon
?au
?au
?au
mau
desire'
sp.'
'lake'
'mushroom'
pimaw
naw
'go; walk'
In Table 98, the forms for 'medicine' and 'thirst; desire' reconstruct to PC and in
both cases the vowels in question are nasalized. The form for 'bamboo species' is
added for comparison;
it provides an example of a word parallel to 'medicine'
but without a nasalized vowel. The three remaining forms on the table confirm
that, if the immediately
did not occur.
Lee's observation
Cham deletion
An
examination
shows that nasalization
preceding
of the relevant vowels

correlates with
consonant was a nasal, the
in Chru and Northern
the Cham deletions,
Cham deletion does not occur after an immediately
preceding
with
Roglai
the caveat that
nasal consonant.
159
and Nasalization
Nasals
Lee (1966:120) notes that the deletions in the Cham forms in Table 99
could have been affected by either of the above conditioning environments, as the
nasalized vowel both follows an *i and precedes a *u. Like the other deletions,
these forms involve nasalized
Table 99: Other deletions

PC
Aceh.
?
x?i?w
vowels
NR
i?u
mia
miau
Lee viewed
iu
iw
lagiu
ligiw;
magiau
miyaw
mi?u
-v
'cat'
'cat', which he reconstructed with a PC nasalized

as it failed to undergo vowel deletion in Cham. However,
vowel
deletion
a nasal conso?
the vowel followed
nant, parallel to the four forms found Table 98.

Despite the apparent clarity of the examples
of Cham
(side)'
'outside'
lagiw
the word
as exceptional,
alternately, it is simply another form in which
vowel,
'left
xligiaw
*miaw
PR Cham
W.Cham
i?u
in Cham
vowels
of nasalized
Chru
in heavy syllables.
and nasalization
in the last three tables, the

tentative. Most
is somewhat
analysis
likely these partial patterns are not the product of internal developments within
that occurred before the
Chamic, but instead, the residue of MK developments
forms were borrowed into Cham. After all, only one of the forms involving a
has a good PMP etymology?the
form 'thirst; desire'. The form xkh?ag
is a late borrowing, as ismade clear by the extreme irregularity in
deletion
'want; desire'
its correspondences.
The
form
'termite'
is borrowed,
and
are
so
forms
the
for
'porcupine' and 'work; do'. The remaining forms could well be as of yet uniden?
tified loans, possibly even late loans. In such cases, the apparent correspondence
patterns would represent not what has happened in the history of Chamic, but
instead may have come from different Chamic
forms from different donor languages.
languages borrowing
the various
Other changes that at first appear to be connected are, on closer inspec?

tion, unrelated. The three forms below have only changed in Phan Rang Cham,
but not inWestern Cham, clearly indicating that the change is recent?post-dat?
ing the split of Western Cham and Phan Rang Cham.
Table 100: Vowel coalescence
Aceh.
post-PC
xhanua? ?
?
xtu?y
xkamuan
kuimuan
Chru
restricted
NR
to PR Cham
W. Cham PR Cham
hanua?
ha nu?? hanui?
tuai
thu?i
kamuan
kamu?n
-i
han??; n??
tuai
toy
kamuan
kamon
'right
-v (side)'
'guest'
-v
'nephew'
160
Chapter
Not only are these changes restricted to in Phan Rang Cham but also the
language contains other parallels. These changes seem to reflect the trend toward
diphthong
simplification
under the influence of Vietnamese
playing
itself out.
NASALIZATION INTSAT
Tsat
is an invaluable
source of valuable
information
on Chamic
nasalization.
consists of preliminary sketches in a

Unfortunately,
broad phonetic transcription. For the reflexes of PC finals, however, there is
in the recording of these finals, that is, variation beyond the
inconsistency
variation
found in a preliminary analysis. The rather wide-range of vari?
expected
the available Tsat material
ation in the transcription of these finals cannot be blamed on Ouyang, Zheng, and
three are experienced and skilled field workers whose transcriptions of
Ni. All
has proved completely

reliable in the past, so the variation
not
reflects something about the nature of the Tsat data itself. Two possibilities,
in
themselves:
variation
result
from
the
the
exclusive,
may
suggest
part
mutually
other
languages
fact that in certain sandhi environments
the older place of articulation has been

in
it
retained
from its retention in careful speech.
and
result
part may
faithfully
all the critical elements for us to understand the basic developments
Nonetheless,
the
glottalization of final nasals in Tsat are present in the transcriptions
involving
of one
or another
of
the
linguists
who
have
looked
at Tsat?
Ouyang,
Zheng,
and
who did
Ni, supplemented by the occasional word transcribed by Maddieson,
some instrumental work on the tones working with Pang's tapes. Thus, despite
considerable
seems
indeterminacy
in the details,
the overall
pattern of development
clear.
Two types of noticeable variation occur in the Tsat data: notational vari?
ation and actual variation. The notational variation is fully to be expected: the
data is not only preliminary but comes from multiple sources and authors, from
from published papers, and from conference papers.
personal communication,
The range of phonetic variation in the tonal transcription is clarified inMaddie?
son and Pang (1993), in which the various notational systems are standardized
explicitly that Tsat has a five-way tone system, with three
level tones, a rising tone, and a falling tone, an analysis implicit in Ouyang,
Zheng, and Ni but certainly not transparent. The notational variation can be
nicely illustrated by the single phonemic falling tone, which ismarked at various
and it is established
times as 53, 42, 32, and 21 (following Ni and Maddieson,

it is always marked in
this work as 42). Part of this variation is subphonemic, as the falling tone is allo
lower in the nasal finals than elsewhere,
phonically
something consistently
marked
in Ouyang and Zheng. Similarly, a notational omission,
rather than
is worth noting: the glottal stop which appears to always accom?
inconsistency,
both
the
pany
rising and the falling tone is not always marked in the Ouyang and
Nasals
161
and Nasalization
nor in the Ni notation, a reasonable enough omission given its

involved?thus
that obscures the diachronic processes
predictability,
glottal stops have been included in this transcription.
None of the notational variation, however, causes any problems. For all
Zheng
notation
but one
the Tsat forms cited in this work a standardized
transcription of tones has been

are
explicitly written; none of this involves sub?
adapted and all final glottal stops
stantive changes, but rather the standardization has been largely mechanical,
aside from the handful of instances in which the various descriptions are actually
different. Even
in these cases,
the cases of the few non-notational

likely reflect typographical
errors rather than actual differences.
the nasalized
Reconstituting
are trivial, as the correct choices in

differences are glaringly obvious, and most
the decisions
finals
is a second kind of "directional" variation, however, that is neither nota?

tional nor trivial, among the reflexes of PC final nasals. Fourteen of the forms in
the data base differ in their transcription of vowel length, sixteen of the 53 nasal
There
in their transcription of a nasal component, and six of the 53 nasal

in their transcription of the place of articulation. Without question
this fluctuation is in large part due to the fact that these distinctions are in the pro?
finals differ
finals differ
cess of disappearing and are evidently quite hard to hear and in part due to con
textually influenced variation, for example, in certain sandhi conditions the older
place of articulation is preserved.
the
However,
variation
obscures
both
the
synchronie
and
the diachronic
in effect, it is possible, to use a term I first heard used by

picture. Fortunately,
Marc Okrand, to "reconstitute" these Tsat finals by critically analyzing the data,
that is, by a judicious comparison of the different transcriptions, supplemented by
knowledge of the Tsat concurrence constraints, by taking into account Ouyang
and Zheng's observations on the directionality of the variation, and by utilizing
the insights provided by the occasional form thatMaddieson
transcribed. In this
way, it is possible
finals must be.
Reconstituting
to determine with considerable
accuracy what
themodern Tsat
the vowel length component
length is the easiest to reconstitute (Table 101). Fourteen of the forms in

the data base and nine of the 53 nasal final forms differ in recorded vowel length.
Vowel
IfMaddieson, Ouyang, Zheng, or Ni heard the vowel as long, it is long regardless

of what the others recorded; this assumption leads to patterns of vowel length
consistent with our diachronic and synchronie expectations.
162
Chapter
vowel
Table 101: Reconstituting

PNB
*ka
length
Tsat
Tsat
PC
Tsat
(reconstituted)
(O. & Z.)
*nam
na:n?33
na:t33
*huda:g
lantha:gn
*sa:g
sa:g33
sa:g33
sag33
'house'
*hadag
thaig?42
tha:k42
thag42
'charcoal'
*papa:n
pam11
pam11
pan11
*?ika:n
ka:n33
ka:n33
kan33
*padam
tham?42
tha:t42
than11
(Ni)
nan?33
'six'
-tha:gn
lanthagn
'shrimp; lobster'
-1
'board; plank'
'fish'
-i
'extinguish'
'extinguish' Ouyang and Zhang also have an intriguing phonetic

form in square brackets [thatn42] that foreshadows the discussion to come.
For
Reconstituting
the nasalization
The reconstitution
component
of the nasalization
component
can be done almost mechani?
cally (Table 102).

Tsat
PC
the nasalization
component
Tsat
Tsat
Tsat
(reconstituted)
(O. &Z.)
(Ni)
(Maddieson)
ha:?42
[hal?42]
nan?33
[na??33]
pa:n?n-t
[p ?42]
x*hag
haig?42
ha:k42
*nam
na:n?33
na:t
*lapan
pam?42
pa:t42
*?ariag
liag?33 -i
liak33
x*prog
piog?33
33
'hot; spicy'
'six'
'eight'
liag?33 -i
'crab'
pyok33
pio?33
'big'
-i
x*trog
tsiog?33
tsyok33
t?iog?33
'eggplant'
*?ura:g
za:g?33
za:k33
za:n?33
'person;
x*dhog
thog?33
thok33
tho?33
'knife'
*khag
khaig?42
[thokg33]
khak42
kha?42
'hard; stiff; strong'
tha:k42
thag42
'charcoal'
someone'
*hadag
*klam
thaig?42
kian?33
kiat33
*masam
sa:n?42
sa:t42
sa:n?42
'sour;
*hitam
tarn?42
ta:t42
ta:n?42
'black'
*dalam
lam?42
la:t42
lam?42
'deep; inside'
*gulam
khiag?42 -ft
khiak42
*padam
thain?42
tha:t42
'afternoon;
_._
'carry
than11
-t
night'
vinegar'
on
shoulder'
'extinguish'
163
and Nasalization
Nasals
For the nasal finals, fully sixteen of the 53 forms differ in their recording of a
records
if Ouyang, Zheng, Ni, or Maddieson
nasalization component. Again,
we
assume
to
If
in
the
be present
Tsat form.
this, fully
nasalization, it is assumed
48 of the 53 Tsat forms descended
from PC final nasals turn out to have a nasal?
ized reflex inmodern Tsat.

As for the remaining five forms, these are also assumed to have (or, at
least very recently had) a nasalized vowel. Not only are nasalized forms the
expected reflexes of the PC final nasals, but non-nasalized final alveolar or velar
consonants
do
not
seem
otherwise
to occur
in Tsat.
Further, while the nasals we are so carefully reconstituting still exist as

in specific sandhi contexts, they may only exist as vowel
nasals synchronically
nasalization in citation forms, except possibly for careful speech.
Reconstituting
the place
of articulation
the place of articulation can be restored in a similar manner. Ouyang and

Zheng note that, both with nasals and stops, older velars alternate with alveolars,
an observation which is fully supported by the diachronic facts. In addition, it is
Finally,
clear that both older alveolars
and velars sometimes
alternate with glottal stops.

Thus, if the "oldest" form found in Ouyang, Zheng, or Ni is taken as the basic
place of articulation, all but two of the forms can be assigned a place of articula?
tion that again fully accords with the diachronic facts.
The variation in the place of articulation for these reflexes is obvious in
comparisons of Ouyang and Zheng with Ni, and has been specifically com?
mented on by Ouyang and Zheng. The following patterns of variation are specifi?
cally noted by Ouyang and Zheng (1983:31); the directionality of the changes
reflects in part the difference between what is retained in sandhi forms and what
is found
in citation
forms,
a directionality
that mirrors
the historical
changes
involved:
variation with velars becoming
-k and -t
-g and -n
(with -k being the older form);

(with -g being the older form);
variation with nasals becoming

-n and -t
-g and -k
alveolars:
homorganic
stops:
(with -n being the older form);

(with -g being the older form);
variation with velar nasals disappearing
leaving behind nasalized vowels:
-ag and -a (with -ag being the older form).
164
Chapter
And, in addition, although not specifically noted by Ouyang and Zheng,

on the basis of comparing Ouyang and Zheng with Ni, there is variation between
-t or -k and glottal stop:
-t, -k and
(with -t or -k being the older form).
-?
Two of these patterns of variation are not just common inmodern Tsat but also
show up when Ouyang and Zheng's forms are compared with Ni's, that is, in two
cases a velar varies with an alveolar?for
which the original place of articulation
to be velar, and in four cases, a velar varies with a glottal stop?for
the original place of articulation is again assumed to velar.
In short, despite variation in the marking of length, nasality, and place of
articulation, it has been possible to determine with a fairly high degree of accu?
is assumed
which
racy the basic phonemic
shapes of the reflexes of the Tsat reflexes of PC final
nasals.
it is likely that the two places of articulation we are so carefully

"reconstituting" are consistently present in modern Tsat in certain sandhi posi?
tions (as will become clearer later), but it is equally likely that in citation forms
Further,
and may even have been largely

they have lost much of their distinctiveness,
reduced to nasalized vowels followed by a glottal stop, except perhaps in careful
speech.

PC
Tsat
the place
Tsat
of articulation
Tsat
(reconstituted)
(O. &Z.)
(Ni)
*?ura:g
za:g?33
za:k33
za:n?33
'person;
*tula:g
la:g33
la:g33
la:n33
'bone'
x*prog
piog?33
pyok33
pio?33
Tbig'
x*hag
haig?42
ha:k42
ha:?42
'hot; spicy'
*khag
khaig?42
khak42
kha?42
thog?33
thok33
tho?33
'knife'
x*dhog
someone'
[thokg33]
Internal
reconstruction
of preploded
nasals
This rather laborious reconstituting of the Tsat reflexes of the PC nasal finals was
a necessary prelude to the internal reconstruction of Tsat sandhi forms, which are
Nasals
165
and Nasalization
to
the final key to understanding
the fate of PC nasal finals in Tsat. However,
we
to
the
forms
determine
should examine the four attested sandhi
pattern?
begin
ing.
Table 104: The four attested

Sandhi form:
sandhi forms
form: PC
Reconstituted
tatn42(la:nn)
>
tarn?42
'section'
tsiakg42 (lai33)
>
tsiag?42
'where'
thokg33-
>
thog?33
'knife'
'
extinguish'
>
thatn42-
tham?42
x*dhog
*padam
Ouyang and Zheng (1983:31) list the first two forms of Table 104, noting that
certain sandhi forms have nasal finals preceded by homorganic stops. The other
two forms are also from Ouyang (p.c.). The first two forms are enough to make
two things clear: First, ignoring the arrows on the table for a minute, the modern
glottalized, nasalized finals have come from earlier final clusters consisting of a
stop followed by a homorganic nasal. Second, now that the pattern is clear, it is
possible to predict the sandhi forms from the reconstituted forms.
The
sandhi forms cited by Ouyang and Zheng bring to mind Court's

(1967:48) term preploded nasals, which he uses to describe the stop plus homor?
ganic nasal combinations which occur inM?ntu. To illustrate M?ntu preploded
nasals, he gives numerous M?ntu examples,
including the two forms cited in
Table 105. To Table 105 have been added Malay, PC, Northern Roglai, and the
four attested Tsat sandhi forms, along with the reconstituted forms.
Table 105: The preploded
sandhi forms
Tsat
M?ntu
/buratn/
/turakg/
Malay
bulan
tulang
padam
Tsat
>
PC
NR
*bula:n
bila:t
-phian11
'moon'
*tula:g
tula:k
-lain33
'bone'
x*dhog
thok
>
thog?33
tham?42
'knife'
'
extinguish'
>
tarn?42
'section'
tsiag?42
'at'
*padam
padap
sandhi
thokg33thatn42tatn42tsiakg42-
>
>
reconstituted
a so-called Land Dayak language of Kalimantan,

In M?ntu,
the pre?
nasals
out
have
of
final
nasals
also
the
in
discussion
ploded
(cf.
developed
"Reflexes from PC word-final nasals" on page 171). In the Northern Roglai data
listed here, PC final nasals presumably
became preploded
nasals before
loosing
166
Chapter
the nasal component and becoming homorganic voiceless stops. And, in Tsat, as
be shown, the PC final nasals became preploded nasals (at least some of
which are still preserved in the sandhi forms) and then in turn some of these
will
developed final glottalization.

The patterns in Tables
105 and 106 allow us to internally reconstruct the

forms. The vowel length is
sandhi forms on the basis of the reconstituted
Tsat
inherited from PC.

In Table 106 are the internally-reconstructed
sandhi forms. From these
forms, not only the modern Tsat reflexes can be derived, but so can the Northern
Roglai reflexes. Thus, there is no doubt that the modern Tsat and modern North?
ern Roglai forms descended from an earlier common source.
Table 106: The internally-reconstructed
PC
Tsat
Tsat
(reconstituted)
(sandhi)
*nam
>
*natn
*?ura:g
sandhi forms
>
*za:kg
*khag
*kh?kg
>
*hadag
*th?kg
>
na:n?33
'six'
someone'
za:g?33
'person;
khaig?42
thaig?42
'charcoal'
x*hag
*h?kg
>
haig?42
'hot;
*masam
*satn
>
sa:n?42
'sour;
*hitam
*tatn
>
tarn?42
'black'
*dalam
*tatn
<
tarn?42
'deep;
*gulam
*khiakg
>
khiag?42 -ft
'carry
*padam
*thatn
>
tham?42
'extinguish'
pam?42
'eight'
>
kian?33
'afternoon;
>
liag?33 -i
'crab'
-ft
>
*dua-lapan
*patn
*klam
*kiatn
*?ariag
*liakg
x*trog
*tsi?kg
>
tsiog?33
x*prog
*pi?kg
>
piog?33
x*dhog
*th?kg
>
thog?33
-i
spicy'
vinegar'
inside'
on
shoulder'
night'
'eggplant'
W
'knife'
that neither the Tsat preploded nasals nor the Northern Roglai pre?
final
later should be confused with the post-nasalized
ploded
of
discussed
Maddieson
and
Vietnamese,
(1996:129). Not
stops
by Ladefoged
only has there been no contact between the Vietnamese and the Tsat speakers but
the phenomena seem distinct; as Ladefoged and Maddieson write, inVietnamese
Note
nasals discussed
stops are usually released, but the release is by lowering the velum
so that a short voiceless nasal is produced".
the oral closure ismaintained,
"word-final
while
167
and Nasalization
Nasals
The origin of these Vietnamese post-nasalized

stops seems to have more
in common with the origins of echo vowels found throughout much of Austrone?
sian than with the Tsat, Roglai, and Land Dayak preploded nasals, which results
from a change in the timing of velar lowering for what was originally a final
nasal.
of modern
The distribution
glottalization
It is necessary to keep in mind that preploded final nasals are essentially final
clusters consisting of a stop followed by a homorganic nasal. The glottal stop
transcribed at various times by Ouyang, Zheng, Ni, and Maddieson
developed
when
of these stop plus nasal combinations
the stop component
articulation,
becoming
glottal
than
rather
an oral
closure
e.g.,
lost its place of

*-tn-
>
-?n-
and
*-kg- > -?g-. Itwas this process that produced what the various authors have writ?
ten in their phonemic transcription as either -n? or -g?, depending on the place of
articulation.
Given the assumption that the stop component of the preploded nasals
it is possible from the modern distribution of
the source of glottalization,
to
reconstruct
the earlier process. An examination of the modern
glottalization
distribution shows that glottalization occurs with a well-defined,
specific subset
was
of vowel plus final nasal combinations,

exceptions, just in the forms containing
in
cifically,
*-og,
but
the
reflexes
nowhere
else.
of
the
PC
that is, with two extremely interesting

the PC short vowels *-a- and *-o-, spe?
rhymes
*-ag,
*-am,
*-?n,
*-iag,
*-ian,
and
It is imperative not tomiss the importance of the fact that glottalization

was recorded in these forms. The only reconstructable source for the glottaliza?
tion in forms ending in PC final nasals is the earlier preploded nasals of Tsat (and
Thus,
Roglai).
'six',
form
just the occurrence

with
nasalized
of glottalization
vowel
(due
*?ura:g 'person', a form with a long vowel,

these
in the reflexes for PC *nam
to perseverative
and
nasalization),
argues for earlier preploded
nasals in
forms.
Thus, the real puzzle is not why some Tsat forms have glottalization?it
comes from Tsat forms descended from PC forms with nasal finals. The puzzle
why
some Tsat forms descended
is
from PC forms with nasal finals do not have
It is possible to be misled by a comparison with other languages

containing preploded final nasals into thinking that the presence of nasal onsets
prevents the development of preploded final nasals. However, at least in the case
glottalization.
of Tsat, this is the wrong answer. The glottalization

in the Tsat reflex of PC *nam
'six' makes that quite clear; the Tsat reflex of PC *nam 'six' obviously has a
nasal onset and just as obviously has glottalization
in its reflex. In fact, the pres?
ence of glottalization
in the Tsat reflex of PC *nam 'six' suggests the right
168
Chapter
answer: all final nasals became preploded but itwas only in certain highly favor?
that this preplosion was retained long enough for the stop
ably environments
component to develop
The solution
into glottal closure.

to understanding
the historical processes involved lies in
that all final nasals originally developed preplosion, which was then
realizing
environments. Thus it is necessary to characterize the

favored the retention of preplosion. Certainly, preplosion is
after nasal onsets, the retention of glottalization
in 'six' notwithstand?
subsequently
environments
disfavored
lost inmost
which
ing. Beyond this, keep inmind that final preplosion is essentially a syllable-final
cluster, so retention of the preploded nasals would have been least favored after
long vowels and most favored after short vowels. After long vowels, simplifica?
tion of the preploded cluster should be most likely; after short vowels, it should
be least likely. And, in fact, this matches
the modern distribution: except in the
word 'person', after long vowels, there is no evidence of glottalization, while
after
the PC
bution makes
everywhere
that
is, except
short
vowels
*-a-
and
*-o-,
glottalization
always
occurs.
This
distri?
perfect sense if the preploded nasals had simplified to plain nasals

except in the most favorable environment for retaining a final cluster,
before
the
two
short
*-a-
vowels
and
*-o-.
Two additional pieces of evidence argue for the analysis in which all PC
preploded before being simplified to simple nasals in most
final nasals became

environments
in Tsat,
one
internal
and
one
external.
The
internal
evidence
is the
on *?ura:g 'person' with its long vowel. While

it is
circumstances
led to this particular word keeping its
it nonetheless did and it
preploded nasal long enough to develop glottalization,
seems more likely that this is a specially-conditioned
retention than a specially
presence of glottalization
unclear just what special
conditioned
The
development.
second piece of evidence
is even more
The
language
to Tsat, Northern Roglai has a startlingly similar set of
it is necessary to
developments. Even without considering the Tsat developments,
conclude that Roglai also developed the highly-marked final preplosion in final
most-closely
compelling.
related
nasals, and then subsequently simplified the preploded nasals, but with the nasal
component, not the stop component being lost during the process. Instead of two
separate but independent
it ismuch more plausible
accounts
of the development of final preploded nasals,

that Tsat and Northern Roglai developed the
they became separate languages, with each lan?
to assume
preploded final nasals before

guage then simplifying the preploded nasals in its own way.
It is in this context that the glottalization on the Tsat reflex of *nam
'six'
extremely interesting. In closely-related Roglai, only PC final nasals in

forms like *nam with its nasal onset failed to ultimately have homorganic stop
reflexes in modern Roglai. Thus, it is clear on the basis of internal evidence
becomes
169
and Nasalization
Nasals
that all the forms without
within Roglai
syllable-initial
their accom?
nasals with
panying perseverative nasalization developed preploded final nasals on the path

to the modern homorganic
stop reflexes. What is left unclear from the Roglai
material alone is whether the forms with syllable-initial nasals also developed
preploded nasal finals and then subsequently lost them or whether the syllable
initial forms simply never developed preploded finals in the first place. The Tsat
form
*natn
<
na:n?33,
from
PC
*nam
with
'six',
its
internally-reconstructed
pre?
ploded nasal provides the answer; in the common ancestor of Northern Roglai
and Tsat, in which the preploded nasals were originally innovated, all PC final
nasals developed preplosion, with the subsequent simplification of these stop plus
nasal clusters only occurring after Tsat and Northern Roglai had separated, as is
evident from the often similar but clearly independent paths of cluster simplifica?
tion taken in each language.
The modern distribution of glottalized, nasalized finals in Tsat came
as
about
the result of a chronologically-ordered
sequence of changes, which is
represented in a somewhat simplified form in Figure 14:
Figure
14: Simplified derivation
of glottalized
nasal finals
'hot; spicy'
P-Roglai-Tsat preploded final nasals

coda simplification after long vowels
glottalization of prenasal stops h??g
'house'
PC
*hag
*sa:g
h?kg
?
sa:g
sa:kg
lengthening of /a/ before glottal stop

tone assignment
ha:?g
ha:?g42
modernTsat/ha:?g42/
sa:g33
/sa:g33/
is only a rough schematic representation of

Figure 14, it should be emphasized,
the basic outlines. For discussion of the details and for a justification of the chro?
nology, please
The Tsat nasal
see the preceding

reflexes with
discussion.
the 42 falling
the data base contains nine examples

developed a 42 falling tone. It is immediately
Finally,
(Table 105) that the presence
of a glottal
tone
of former preploded nasals having

from examining the forms
obvious
stop is one prerequisite
for the falling
170
Chapter
tone just as it is evident that the presence of a glottal stop by itself is not sufficient
to predict the occurrence of the tone. The second prerequisite seems to be the
presence of an inherited or a secondarily-derived
long vowel inmodern Tsat.
pair of conditions accounts for the data in Table 106, with two
Tsat /khiag?42 -ft/ from *tsiakg -ft 'carry on the shoulder' and /za:g?/
This
exceptions:
from
earlier
*za:kg
someone'.
'person;
The
tone
falling
on
'carry
on
the
shoul?
der' is unexpected, but so is the final; that is, the irregular tone is in a form that
also has other correspondence problems.
The lack of a falling tone in PC *?ura:g > Tsat *za:kg > za:g?, on the
other hand, could be attributed to any one of a number of unique characteristics
in this form was originally long; the PC vowels in the
other forms were originally short. As a member of the classifier and of the pro?
noun systems,
the word
itself more
than likely occurs in phonologically
unstressed environments, unlike other otherwise similar forms in the data base.
of this form: The PC vowel
of this, of course, tells us precisely why this particular form did not develop
a falling tone; its unique characteristics do, however, make it clear that the word
is not a counterexample
to the proposed analysis.
None
NASALIZATION INNORTHERN ROGLAI

The Roglai reflexes of PC final nasals are important for two things: Like Tsat, the
Northern Roglai reflexes of PC final nasals have passed through a stage in which
of preploded nasals, see
they were preploded final nasals (for a discussion
in Tsat" on page 160), providing compelling evidence that Tsat and
The
Roglai were once the same dialect (namely, proto-Northern Roglai/Tsat).
have
in
107
final
nasals
Northern Roglai voiceless
PC
Table
reflexes
of
stop
"Nasalization
from earlier preploded

indicated in the table.
evolved
Table 107: The presence

PC
>
*nam
>
nasals shared by Proto-Northern-Roglai/Tsat,
of preploded
P-NR/Tsat
*n?pm
NR
nam
nasals
inNorthern
Roglai and Tsat
Tsat (sandhi)
?
*bula:n
>
*bula:kg
ia bila:t
*tula:g
>
*tula:kg
tula:k
x*dhog
>
*thokg
thok
*padam
>
*pad?pm
padap
'six'
thokg33thatn42tatn42tsiakg42-
'moon'
'bone'
'knife'
'extinguish'
'section'
'at'
as
171
and Nasalization
Nasals
And, in addition to their importance for subgrouping, the unique treat?

ment of word-final nasals inNorthern Roglai is a rich source of information not
only about the history of Roglai nasals but also about the history of PC nasals and
about PC nasalized vowels. The modern Roglai reflexes often provide important
evidence about which forms are inherited and which are borrowed. And, further,
in the case of the borrowed
about the relative chronology
Reflexes
from PC word-final
forms, Roglai sometimes

of the borrowing.
also provides
information
nasals
nasals, there are two basic Roglai reflex patterns. In the most
of these two patterns, the word-final nasal has as its Roglai reflex a
For PC word-final
dominant
stop.
homorganic
Table 108: PC final nasals with Roglai

PC
NR
Chru
stop reflexes
PR Cham
ha:k
hag
dalap
talarn
kho:g
kho:k
khog
'dry (weather?)'
*khag
khag
khak
kh?g
*sa:g
sa:g
sa:k
thag
mil?m
'house'
x*ha:g
*dalam
dalam;
'bank (river); shore'

'deep; inside'
darlam
x*kho:g
*malam
malam
malap
x_na:n
hajam
*huja:n
*masin
masin
*?idug
adug
*phun
phun
'stem'
*pa-gha:g
-n la:i?
*bha:n
pha:n
*luba:g
*halim
laba:g
halim
x*fiag
?ag
hag
khan
x*hag
x*khan
xsua:n
xduan
-1
'door'
'night;
rina:t
ina:t;
evening'
'pineapple'
can
'rain'
huja:t
masit (?)
id?k
mithin
phut
ph?n
'trunk;
pakha:k
pakhag
'dry
pha:t
luba:k
phan
'sneeze'
ha?an;
it?g; tug
'salted;
salty'
'nose'
log;
over
plant'
fire'
lipag; lapag
halim
'hole; pit'
??g
h?g
'hole;
hak
khat
khan
'hot; spicy'
'cloth; blanket'
swan
'soul'
don
'bamboo
halip
?ak
sua:t
duat
-1
'rainy
season'
door'
hat'
172
Chapter
As Table 108 shows, Chru and Phan Rang Cham both retain the PC final
in Roglai, the PC final nasal consonants have changed into homorganic
but
nasal,
oral stops. This homorganic stop is the unmarked reflex of word-final nasals in
Roglai.
In Table
108, all the PC final nasals have denasalized becoming homor?

the
ganic stops,
expected reflex. Of particular interest for relative chronology are
the last two examples
'soul' and 'bamboo hat'. Neither form reconstructs to the
x
*
PC level (as the without an accompanying
indicates), however, both forms
have final stops, not nasals. Thus, these forms were borrowed into pre-Roglai
after the breakup of PC but in time to take part in the denasalization
of Roglai
final
nasals.
In the other basic pattern, the PC word-final

retention
correlates
that
with
the presence
of
a nasal
nasal has been retained, a

consonant
at the onset
of
the
final syllable, that is, the PC word-final nasal has been retained as a nasal when
the PC final syllable also began with a nasal. Notice that, in this data, there is a
correlation between Roglai nasalization and vowel length.
As was true for Table 108, in Table 109 the forms inherited from PC
include PC forms
structable back
inherited from PMP
to PC
(marked with
(marked with *) and borrowings recon

x*). Both sets of forms, of course, have
the change.
undergone
Historically,
once
the Roglai forms have descended from forms that were

to the "preploded" nasals of M?ntu Land Dayak
similar
phonetically
described by Court (1967) (discussed also on page 164). Blust
among others Court (1967), writes about preploded nasals:
(1991:148),
citing
In certain Land Dayak languages of southwest Borneo (Scott 1964:

1967), final nasals are often combined with a preceding obstruent,
in others: Mentu
voiced in some languages, voiceless
ciupm 'kiss'
Court
(Malay cium), Bukar-Sadong kaidn 'cloth' (Malay kain). Simple nasals

occur in final position if the final syllable begins with a nasal consonant:
Bukar-Sadong
Continuing,
he writes
terjan
'hand,
arm'.
that similar distributions
occur elsewhere:
...in Tunjung of southeast Borneo and in some dialects of Mentawai,

spoken in the Barrier Islands west of Sumatra (Bernd Nothofer, personal
a Chamic
Northern Roglai,
communication).
language of Vietnam,
final nasals as the homorganic voiceless
If
the
final
stop.
syllable begins with a nasal, however, the final nasal is
?
a clear indication of the former presence of preploded
preserved
nasals in that language as well.
reflects Proto-Austronesian
Table 109: PC final nasals with Roglai

PC
Chru
pana:g
*pina:g
nasal reflexes
NR
PRCham
pin?g
panig
'betel (areca palm)'
x*miag
mieg
'cheek; jaw'
*gunam
ganam
kan?m
'cloud'
*lama:n
lama:n
limin
lum?n
'elephant'
-i
cana:g
can?g
tanig
*taga:n
taga:n
tag?n
tag?n
'hand'
*timun
tamun
tum?n
tam?n
'melon;
*?agan
agan
in; g?n
*?anug
nam
*nam
*mam
-v
n?m
mem
mam
-v;
m?m
*?anan
nin
*binay
banai
'of animals'
-1
*ma?am
ma?a:m
-1
'package'
n?m
'six'
-v;
mum
-v
suckle'
'suck;
n?n
'that (third p.)'
binai
pinay
'virgin'
'of animals'
'woman'
maft?m
miflim
'the wind'
agin
agin
agin;
xkamuan
kamuan
kamu?n
kamon
'nephew'
ramo:g
lum?g -i
rimog;
'tiger'
ramo
kram
krim
-v
xkr?m
-lvn
xnran
-if;
xdran
-if
xcaguar
kra:m
drin
-f
canua
-1
-v
dr?n
-vf
pran
-i
gin
'bamboo'
tr?n
'numb'
?
-fg
'flat
cagu?
*miaw
miau
mi?u
miyaw
*naw
nau
n?u
naw
*samaw
sam?
*danaw
danau
danau
xhanu??
hanua?
ha nu??
*sana
sana
sana
-f
'fry'
-v
basket'
'cat'
'go; walk'
'prompt;
-n
-n
tanaw
twill'
'weave;
*?agin
xrimo:g
cucumber'
'name'
an?g
m um
bed'
'furniture;
xcana:g
on
time'
'lake'
han??
'right (side)'
hana
'roast;
parch'
174
Chapter
nizing
One key to understanding the phonetics of these changes lies in recog?

that, as in many of the Austronesian
languages of this area, in Northern
the salient pattern of nasalization of vowels by adjacent nasal consonants
Roglai
is perseverative,
Roglai
on
not anticipatory. That is,
the
vowels
rule: v > v / nasal_(C)#.
nasalization
not denying
While
also have been some anticipatory
that there may

word-final
preceding
nasal
it was
consonants,
the
nasalization
perseverative
that was crucial to blocking the denasalization of final nasals. Thus,

the perseverative nasalization of the vowel from the syllable-initial nasal conso?
assimilation
109, alone or in combination with whatever anticipa?

that may also have occurred, was salient enough to block the
tory assimilation
denasalization of the word-final nasals.
nant of the forms in Table
other
Forms with
reflex patterns
forms that fail to conform
Most
rowings,
a subset
although
of
to one of the two basic reflex patterns are bor?
such
aberrant
forms
may
instead
reflect
PC
nasal?
ized vowels. Although

the irregularities in their correspondence patterns make it
obvious that most such forms are recent borrowings, a small number of these
forms
are
in their
regular
In addition
assimilation
mary
from
nasalized
vowels
nant but which
Post-pre-Roglai
patterns.
correspondence
to the secondarily
nasal
preceding
with
no
nasalized
consonant,
obvious
reconstruct with nasalized
connection
vowels
there
vowels
are
to an
from perseverative
also
a handful
adjacent
nasal
of
pri?
conso?
at the PC stage.
borrowings
In Roglai, some the easiest post-PC borrowings to identify are those that have
of
been borrowed so recently that they have not undergone the denasalization
final
nasal
consonants.
In Table 110, the final word-final nasals of the Roglai forms have not
denasalized, despite the fact that there is no evidence of a preceding nasalized
vowel to stop the denasalization from taking place. On the basis of this (and vari?
ous other pieces of evidence) these forms have been identified as late borrowings
post-dating
the denasalization
of Roglai
final nasal consonants.
175
and Nasalization
Nasals
Table 110: Late borrowings

PC
Chru
xkhi:n
khin
xkhiag -fi
khiag
xgriag
griag
xdian
dian
-v
-v
khin -f
kh?n
'dare;
khiag -f
giag -f
kh?g
'want;
desire'
tusk'
kr?g
'fang;
di?n
-f
tien
'candle'
cum
-f
c?m
'kiss;
-f
phug
xphug
x(li)hug
x?ag
?a:g
lahog -f
?ag -f
gram
gram
-vf
xho:g
-f
ho:g
xham
-In
ha:m
'table'
-f
ham
smell'
'papaya'
kr?m
-v
-f
hog
-1 la:m
brave'
'leper'
lahog -v
xgram
Roglai
PR Cham
NR
xcum
xcum
into Northern
'thunder'
'wasp'
-f
'greed'
in Roglai
Secondary
final nasals
In addition
to inherited final nasals, Roglai also has a set of final nasals derived
from PC *-l. The original PC lateral final is still found in Cham and
secondarily
Chru.
*-/ > Roglai
Table 111: PC final
PR Cham
NR
Chru
PC
-n
*mal
mal
man
hap?l
mal
*gatal
katal
katan
kat?l
spal
x*sapal
sapan
sanan
*sanal
*wil
wil
'circle'
kapa:l
ba:l
*kapa:l
x*ba:l
'beam'
'itchy'
'pillow'
w?l
win
'round'
kapan
kapal
'thick'
ban
pal
'mend;
banan
x*b-an-a:l
'arm'
'rag'
xkarfual
kadual
kaduan
katfol
'heel'
x*ja:l
ja:l
hual
jan
?al-1
'casting
hol
'cloud,
x*hual
huan
xdhual/r
-f
thul-v
thun
x?abual
-v
bual
aban
del
den
xtfel
xsi?jual
-v
sa?ual
-v
th?r-vf
-v
si?juan
patch'
cf.
'mend'
net'
fog'
'dust; fog'
'blunt; dull'
'shallow'
?j?l;ha?j?l
'light (weight)'
176
Chapter
the change of final *-l (and *-r) to -n is an areal feature of

Although
Southeast Asia, within Chamic the change is apparently limited to Roglai. Note
that this change even applies to post-PC loans. The relative lateness of the change
within Roglai is apparent from relative chronology,
shift of Roglai final nasals to homorganic stops.
that is, the shift postdates
the
NASALIZATION INACEHNESE
as elsewhere
in Acehnese,
in Chamic,
is perseverative. The most
of this nasalization is on the main syllable vowels of Ace?
obvious manifestation
hnese, that is, the vowels in the second syllable, which it affects in several ways,
Nasalization
as Table
112 shows.
Table 112: The effect of perseverative

Malay
PC
Aceh.
on Acehnese
nasalization
NR
PR Cham
vowels
kulit
*kulit
kulet
kuli:?
kali?
'skin'
nipis
*lipih
lipeh
lupih
lip?h; lap?h
'thin'
langit
l?get
mamen
mum?h
mim?h
agen
ag?n
kunyit
*?agin
*ku?it
kuftst
ku?i:?
ag?n; g?n
ka?i?
mabuk
*mabu?
m ab o?
babu?
manuk
*manu?
mano?
man??
buta
*buta
buta
darah
*darah
darah
darah
tar?h
'blood'
ta?i
'ask'
angin
lag?:?
ligi?; lagi?
*lagit
*mamih
min??
mita
'yellow'
'blind'
tanya
*taga
ta?Ag
*huma
umAg
huma
hamu
lima
*lima
limAg
lum?
limi; lami
bunga
*buga
bugog
bug?
-v
'field'
'five'
'flower'
inAg
ini
'mother'
awuia?
awa?
'spoon;
*rata:k
ruituia?
rata:?
ripa?; rata?
'bean;
dahan
*dha:n
dhuian
tha:t
than
'branch;
hudang
*huda:g
uduiag
huda:k
'shrimp'
anak
*?ana:k
anu??
ana:?
hat?g; pag
ani?
minyak
*mi?a:k
mi?ui?
ma?a:?
mi?i?
'oil'
panig
'betel
pinang
*pina:g
pinuig
pin?g
fowl'
'chicken;
-i
huma
x*?awa:k
'the wind'
'drunk'
tina
*?ina
'sky'
*
sweet'
ladle'
pea'
bough'
'child'
(-nut)'
Nasals
177
and Nasalization
high vowels *-u- and *-i- are all lowered in

variants become loi and Id, while the nasalized
variants become hi and Id, respectively. The PC short *-a becomes /-a/, but the
nasalized variant becomes /-Ag/. The PC long *-a:- becomes /-uia-/, but the nasal?
The PC second-syllable
but the non-nasalized
Acehnese,
ized variant becomes
the shorter
l-m-l.
The
Origins
of
and
Registers
Tones
languages present case studies of a clarity unparalleled in the litera?

ture illustrating the radical typological restructuring of phonological
systems.
The language that the seafaring Austronesian-speaking
forerunners of the modern
The Chamic
Chamic
spoke when
and
tonal,
non-registral.
they arrived on themainland

From
this
starting
point
was essentially
have
descended
non
disyllabic,
a startling
array
of phonological
systems: the register system of Western Cham (Friberg and
and Gregerson,
Kvoeu-Hor,
1977; Edmondson
1993); the restructured register
of
Haroi
Burnham
1976; Thurgood, 1996); the quasi
system
(Lee, 1974, 1977b;
registral, incipiently tonal system of Phan Rang Cham (Thurgood, 1993; Han,
and Gregerson,
Edmondson,
1992); and the fully-tonal system of Tsat (Haudri
court, 1984; Benedict,
1984; Ni 1990ab; Thurgood 1992b, 1993). All this variety
has evolved from an essentially identical starting point, all of it has followed rel?
atively transparent paths of internal restructuring, and in each case the unique
endpoint has come about under the influence of contact with
ent
typologically
differ?
languages.
For each of these distinct developments,

it is possible to reconstruct a
clear outline of the entire internal path of development by which, from an essen?
tially atonal and disyllabic pre-Chamic origin, the modern register system (in
Western
Cham), the modern restructured register system (in Haroi), the incipient
tonal system (in Phan Rang Cham (Eastern Cham)), and the fully tonal system (in
Tsat) developed. Due to the relatively shallow time-depth and the richness of the
data, the Chamic languages provide outlines of the internal paths of development
involved in all four cases that are rather clear and relatively complete.
178
The Origins
179
and Tones
of Registers
For example, in the case of the Tsat data, we see a case in which we start
from a completely atonal (and, largely disyllabic) stage and we end with a fully
tonal (and, largely monosyllabic)
stage. The completeness of the data is unlike
much of the literature on tonogenesis, which typically involves just the later tone
splitting stage rather than true tonogenesis. And, further, in contrast to many of
the instances of tonogenesis reported in the literature on tonogenesis,
the end
a two-way
is not
product
tone
but
contrast,
a rich
five-way
tone
system.
There are, of course, several values to such case studies, among which is
their value as models for less transparent, more controversial changes that, often
because of the greater time-depth, require a high degree of extrapolation in the
analysis.
WESTERN CHAM AND THE DEVELOPMENTof REGISTER

a complex of features that tend to occur together: voice
quality (phonation type), vowel length, pitch, and voice quality induced vowel
as David Thomas has noted, individual languages may
gliding. Historically,
one
or
another
of those features, suppressing the other features. Thus,
emphasize
itself constitutes
Register
as he notes, ancient Khmer emphasized the voice quality feature, while Vietnam?
ese has progressively
emphasized pitch. Modern Central Khmer has now sub?
dued
the voice
the vowel ongliding feature

quality feature and emphasized
while
Khmer
has
Modern Northern
instead,
emphasized the vowel height feature.
And,
Modern
Vietnamese
still
has
voice
quality
features
in some
of
its tones.
of breathy voice with vowel raising and of creaky or

tense voice with vowel lowering has long been noted in the literature on register
1952). The explanation for this
complexes (noted at least as early as Henderson
correlation is to be found in the mechanics of the production of breathy and tense
The
correlation
voice: breathy voice characteristically

quent
enlongation
of
the vocal
tract,
involves a lowering of the larynx, a conse?

and
a resultant
lower
Fl?hence
vowel
rais?
tense or creaky voice characteristically

involves a raising of the larynx,
a consequent
a resultant higher Fl?hence
vocal
of
the
and
tract,
shortening
vowel lowering (Peter Ladefoged, p.c.).
ing, while
There is a second correlation, not as widely reported, between voice

vowel centralization. Exactly the same laryngeal gestures that pro?
and
quality
duce vowel height differences also affect the centralization of the vowels, except
that it is the effect on the F2 that determines
vowel centralization: the larynx is

lowered in producing breathy voice, the vocal tract is lengthened, the lengthened
vocal tract lengthens the wave lengths, lowering the formants and resulting in
vowels with a lower F2. The result, of course, is that these vowels are more cen?
tralized. Conversely,
the raising of the larynx in producing
creaky voice,
the con
180
Chapter
sequent shortening of the vocal tract, and the resulting shortened wave
produce higher formants. Vowels with a higher F2 are more peripheral.
lengths,
As forWestern Cham, its register system originated in two stages: first, a

distinct breathy phonation developed after the voiced obstruents (second register)
producing a phonation contrast with the vowel after the remaining consonants
versus
(first register); the two contrasting vowel quality differences?breathy
modal voice?have
resulted in two phonetically distinct sets of vowels, one asso?
ciated with breathy voice and the other with themodal (or, clear) voice. Later, the
register complex associated with breathy phonation was extended to the forms
with sonorant initials, see Figure
Figure
15: The development
PC initial classes:
15.
of Western Cham register
Distinct
voice
Vowel
registers:
quality:
PC initials (except
voiced obstruents +
modal
voice
==>
breathy-voiced
vowel set
sonorants)
PC voiced
obstruents
modal-voiced
breathy
+
==>
voice
vowel
set
sonorants
of the breathy register to include forms with initial sonorants

(shown in Figure 15) is atypical and restricted toWestern Cham. The develop?
ments in Haroi, Phan Rang Cham, and Tsat are more typical, with the sonorants
The
extension
not with
the voiced obstruents, but with the other reflexes.

that these two distinct phonation types affected the vowel
Again
two
distinct sets of vowels (see Table 113).
quality, resulting in
allophonically
Finally, the voiced and voiceless obstruents lost their voicing contrast, merging
in voice
into a single set of voiceless obstruents, phonemicizing
the differences
patterning,
notice
quality and vowel quality, and producing
the modern
register system of Western
Cham.
Inmodern Western Cham, the two vowel registers are distinct, although
the two registers for the vowel /a/ are distinguished, not by vowel quality, but by
other features. And, even now the two vowel sets are in part predictable from the
modern
initials. The original relationships are no longer transparent though,
because of the mergers
in the obstruents and because of the spreading, under
specified conditions,
main syllable.
of register from the pretonic
first syllable
to the stressed
inWestern Cham
Table 113: Vowel registers

Modal
Breathy
register
register
vowels:
vowels:
e1
ae
o e.
[From Edmondson
The
181
and Tones
of Registers
The Origins
and Gregerson
1993:67]
literature
The first synopsis of the diachronic origins ofWestern Cham register was Friberg
and Kvoeu-Hor's
(1977: 35-36, fn. 14) short footnote in their insightful paper on
Western Cham register patterns. It has since been expanded and developed by
and Gregerson
(1993), who supplement their analysis with instru?
data. The registers of Western Cham, like all registers, are clusters of co
occurring features: inmodern Western Cham (Friberg and Kvoeu-Hor (1977) and
Edmondson
mental
and Gregerson
first register vowels usually have
(1993:63-72)),
tenser voice quality, lower voice quality, and higher pitch, while the second regis?
ter vowels usually have laxer or breathy voice quality, high vowel quality, and
lower pitch. And, as these authors observed, the first register vowels derive from
the phonation type induced by proto-voiceless
initials, while the second register
Edmondson
vowels
derive
the phonation type induced by proto-voiced

initials. As
are
the
reminiscent
of Hend?
out,
point
developments
erson's (1952) description of the feature complexes associated with Khmer regis?
ter. See Table 114.
Edmondson
from
and Gregerson
Table 114: Contrasts
between modal and breathy register

Modal
original initials
voice quality
vowel quality
pitch distinctions
[Henderson,
voiced
Breathy voiced
First Register
Second Register
proto-voiceless
proto-voiced
tense,
clear
lower (open)
higher pitch
1952; Edmondson
lax, breathy
higher (closed)
lower pitch
and Gregerson,
1993:61-63]
Two things in Table 114 should be noted. First, although this is Hender?
son's (1952) description of the Khmer voice register distinctions, not of Western
182
Chapter
Cham, the complexes also characterize theWestern Cham distinctions (although

and Gregerson's
Edmondson
instrumental description did not find systematic
vowel quality differences between the registers). Second, the ordering of the fea?
tures parallels the three historical stages: the original initials > voice quality dis?
tinctions > vowel quality & pitch distinctions.
The PC voiced
consonants:
the two layers
The first stage in the history ofWestern

of
voice
breathy
after
the
the innovation
Cham register begins with
obstruents
voiced
*b-,
*d,
*g-,
and
*j-.
Later
another
layer was added when second register spread to the forms with sonorant initials.
Thus, the sonorant initial forms were added to the nucleus of Western Cham sec?
ond register forms from the voiced obstruents. Table
reflexes from originally voiced obstruents.
115 contains second register
115 shows the development of PC voiced obstruents intoWestern

second register. In Table 115 in both the monosyllables
and the disyllables,
it is the main syllable initial that developed into second register, regardless of the
Table
Cham
Notice
initial of the presyllable.

marked
with
a subscribed
els,
not
the
same
the consonants,
way
each
that
indicating
convention makes
orthographic
its presence
dot,
that all main
show
syllable
following
it easy to recognize
time
it appears,
register
second
This
vowel.
register
are
second register by marking
although
distinctions.
initial obstruents
it is of
In actuality,
course
the PC
the vow?
voiced
obstruents have long since lost their voicing and merged with the voiceless series.
InWestern Cham, it is possible for each syllable of a disyllabic word to
be in a different register. In Table 115, 'seven' and 'rice (paddy)' have first regis?
ter presyllable
less obstruent
vowels because
but
second
the presyllables originally began with a PC voice?

the main vowel
register main vowels, because
originally began with a PC voiced obstruent. In contrast, both syllables in the

forms for 'mouth' and 'tooth' are in the second register because both syllables
originally began with PC voiced obstruents. However, sometimes second register
has been lost in the presyllable, cf. 'chest'.
As for the other languages in the table, they display similar patterns: In
Tsat, both the 42 tone and the 11 tones show the effect of second register; only
for Phan Rang Cham, it is not a coinci?
dence that all the second register Western Cham forms have either a correspond?
low tone, as this tone developed from second
ing low tone or a glottal-final
the 55 tone from *-h fails to do so. As
register.
of Registers
The Origins
183
and Tones
Table 115: PC voiced obstruents > Western Cham breathy register

Tsat
Haroi
*habow
pha11
aph?au
pau
hapow
'ashes'
*babuy
phui11
paph?i
papui
papuy
'wild pig'
*hubsy
*ribow
phai11
aphui (m) pay
hap?y
'taro; yam'
'thousand'
pha11
W.Cham
PR Cham
PC
rapau
ripow
kabau
kapaw
'water
laph?au
x*kabaw
kaphiau
buffalo'
*dada
tha11
cathia
tata
tata
'chest'
*?idug
thug11
ath?g
it?g
'nose'
*huda:g
*dua
tha:gn
thua11
athiag
thua
pug
ni? tag
toa
hat?g
twa
'shrimp'
'two'
tlih
klsh
klsh
'tired'
x*glsh
*gigsy
khai11
cakhii
takay
tak?y
'tooth'
*huja:n
*do:k
sam11
asian
can
hagan
'rain'
tho?42
thu?
to?
to?
*?abih
phi55
aph?h
pih
apih
'sit; live; stay'

'all; finished'
*labuh
phu55
laph?h
lapuh
lapuh
'fall down'
papah
'mouth'
tap?h
'ransom'
'seven'
'rice (paddy)'
*babah
*tabus
pha55
paphlah
papah
?
phu55
*tujuh
su55
cas?h
*paday
tha??42
pathiai
ta?uh
tac?h
pafai
patay
Western Cham also has second register after certain sonorant initials that
are discussed in the section on phonation spreading ("Transparency and phona?
tion spreading" on page 183).
Transparency
and phonation
spreading
the situation is precisely as already described. However, for a

subset of the disyllabic words, the modern distribution of register has been com?
plicated by the spreading of the voice quality from the presyllable to the main
syllable (The discussion here disagrees, although only in minor ways, from the
For monosyllables,
Friberg and Kvoeu-Hor analysis of spreading found in Thurgood (1996)). Thus, it

may be the PC initial of the presyllable, not the PC initial of the main syllable,
that correlates with themodern register. Friberg and Kvoeu-Hor
(1977:36), in the
same footnote
referred to earlier, have explicitly
presented
the patterns.
184
Chapter
register after main
Breathy
initial sonorants
syllable
Throughout Chamic there is a hierarchical pattern in the spread of the breathy

voice associated with second register from the presyllable
through the medial
consonant to the main syllable: sonorants > voiceless fricatives > voiceless stops.
in Haroi, Phan Rang Cham, and Tsat, there is clear evidence that
Specifically,
voice
has spread through the medial sonorants to the main syllable. In
breathy
the same pattern of
there is every reason to assume that historically
in
Western
for
the second register
thus
occurred
Cham,
spreading
accounting
after the main-syllable
initial sonorants in Table 116.
addition,
Table 116: Breathy
voice spreading
NR
PC
W.Cham
*barah
barah
paran
*buga
*dalam
bug?
pagur-f
tal?m
*darah
darah
dalap
'swell;
pigu
tal?m
'flower'
'blood'
talan
'tongue'
daruai (m)
taruai (m)
taroy (m)
'thorn'
jala:t
?alan
?alan
<
*g-
bum?u
poh
mau
swollen'
'deep; inside'
tar?h
*dursy
maw
PR Cham
taran
gilah
*boh
sonorant
talan
*dilah
*jala:n
through a medial
'road; path'
'mushroom'
pimaw
In Table 116, theWestern Cham disyllables are in second register, as

indicated by the dot subscribed under the initial obstruents. The same patterns of
spreading are found in the Phan Rang Cham forms.
Western Cham has second register reflexes after initial sonorants in two
additional
environments
environments,
found elsewhere
disyllables
where
in Chamic:
both
where
in monosyllables
syllables
begin with
in these forms
second
register
reflexes
are
not
that begin with sonorants and in

sonorants. See Table 117. The
is unexpected. Perhaps it was

appearance of second register
extended to these forms due to some sort of acoustic similarity between nasalized
and breathy-voiced
vowels; perhaps itwas extended to these vowels as these are
the only other vowels following voiced consonants. In any case, the Friberg and
(1977:36) analysis implicitly assumes that at some point all vowels
sonorants
had developed second register, with it only to be lost later in
following
some cases. It would be interesting to find out if all dialects of Western Cham
Kvoeu-Hor
have second register in these forms.

In Table 117, second register has not been marked in any special way, as
it is fully predictable from the initials. The origin of second register with these
forms will
be
speculated
on after
the discussion
of
register
spreading
page 183, below.
on
The Origins
185
and Tones
of Registers
sets of disyllabic forms with sonorant main-syl?

lable initials. In these, the vowels are in first register, rather than second register.
In the Friberg and Kvoeu-Hor
(1977:36) analysis (followed in Thurgood 1996),
these are treated as the result of the spread of first register from the presyllable to
There are two additional
the main
vowels
after
syllable. Contra the Friberg and Kvoeu-Hor

after the sonorants in Table 117 are assumed
analysis, the first register

to be the expected reflex
a sonorant.
Table 117: Breathy-voiced

Tsat
PC
Haroi
x*ls?
le?24
*?ular
la33
*mamah
ma
x*lamo
mo
is extended to the sonorants
55
,33
11
m
la
mam?h
mamih
lamo
ula
'snake'
mimih
'chew'
lamo
'cow;
ami
'father'
'five'
ox;
cattle'
fowl'
*?ama
ma
*lima
ma33
lami
limi
*manu?
nu?24
manu?
man??
min??
'chicken;
*?ana:k
na?24
ana?
ni?
ani?
'child'
*?ini
ni33
ni
ni
ni
'this'
*lagit
x*war
lamia
rji?24
van
ama
'fall into'
alia
lamo
PR Cham
W. Cham
33
lagi?
wol
mi
lagi?
war
lagi?
'sky'
war
'forget'
In the disyllabic forms in Table 118, the vowel after the main-syllable
initial sonorant is in the first register, not in the second register. Thus, it is clear
that it is initial of the first register, not the initial of the second register, that has
determined
the register of the second register.

from
the first syllable to the second.
spread
Table 118: Main
In short, the first register has
syllable sonorants with first register reflexes
PC
Tsat
Haroi
*pina:g
na:g33
panag
panig
panig
'betel; betel-nut'
*tula:g
la:g33
calag
talag
talag
'bone'
kam?i
ka m ay
kam?y
'female,
*kumey
mai33
W. Cham
PR Cham
woman'
*pula
pia33
pala
p?a
pala
'to
*panah
na55
panih
*kulit
panih
kali?

'skin'
*hurey
li?24 -i
33
zai"5
pan?h
kalei?
*huma
ma 33
kli?
harii (m) ea hray

hami
hamia
'day; sun'
har?y
hamu
plant'
-v
'field'
186
Chapter
voice spreading
Breathy
through other main
syllable
initials
As Friberg and Kvoeu-Hor

spread through the medial
in the forms in Table
second register has also

observed, the breathy-voiced
voiceless fricatives, producing second register vowels
119. Note that in both 'new' and 'otter', the two words with
Ibl, the whole word
medial
Table 119: Spreading

PC
*bassy
x*buhay
*bahrow
is in second register, not just the second syllable.
of breathy voice through *s and *h
Tsat
Haroi
sai1
pas?i
W. Cham
pasay
pahas
phia11
priau
PR Cham
pith?y
-f
pahau
iron
phay
'otter'
pirow
'new; just now'
All
the Phan Rang Cham forms show evidence of the spread of second register in
their low tone reflexes, as do both the Tsat forms in their 11 tones. In Haroi, how?
ever,
'new'
only
shows
second
register
spreading.
The only Western Cham medi?is that block second register spreading are
the medial voiceless stops. As Table 120 shows, second register failed to spread
through the medial voiceless stops.
Table 120: Spread of breathy voice stopped by voiceless
PC
Tsat
*batsy
u11
Haroi
tai1
W.
Cham
stops
PR Cham
pat?i
patay
pat?y
'banana'
pato?
pat??
taki?
pat??
taki?
'cough'
*batuk
tu?42
*dikit
ki?*2
*batow
tau11
pat?u
patau
patow
'stone'
x*bato
to11
pato
pato
pato
'teach'
'few;
little'
as Friberg and Kvoeu-Hor

(1977:36) suggested, the voiced obstruents
before spreading could occur, perhaps the medial voiceless stops are
simply more resistant to spreading, or perhaps it is some combination of the two.
Perhaps,
devoiced
The history
of Western
Cham
contact
there are gaps in our knowledge of Western Cham, it has apparently

in contact with register languages. In the literature, the Western
always
Chams are viewed as a recently split off dialect of Cham, with theWestern Cham
Although
been
being a branch of Cham that migrated westward, particularly to Cambodia, after

Champa collapsed in the sixteenth century (Headley, 1991), with this migration
placing
the split with Phan Rang Cham around 1471. Undoubtedly,
many
of the
of Registers
The Origins
187
and Tones
Cham did migrate to Cambodia at that time, but it is also quite clear that
a significant number of them remained in southern Vietnam and doubtless these
Western
to have contact with Phan Rang speakers. Be that as itmay, judging

from a comparison of Western and Phan Rang Cham, both Phan Rang andWest?
ern Cham had already developed a common register system by the time of the
split, one probably shared by Haroi.
continued
The gaps in our knowledge revolve around identifying the specific con?
tact languages involved in the development of the original register system. With?
out more precise knowledge of the languages involved, it is impossible to do
much more
than speculate on the social mechanisms

involved. However, from the
nature and the number of MK loans incorporated into both Western and Phan
Rang Cham, it is clear that at very least the contact involved long-term bilingual?
ism and it also involved language shift from some MK languages to Cham, cer?
tainly at least in part from Bahnar.
It is not clear just how homogenous Western Cham is nor how similar
have been in different dialects. The two dialects used in
the paths of development

this work, one described
by Kvoeu-Hor and Friberg (1978) and one by Headley

(1991), both differ, although only inminor details. When some of the other dia?
lects of Western Cham are described, particularly those that have undergone dif?
ferent paths of contact since splitting off from Cham,

how different these are.
itwill be interesting to see
Very speculatively, itmay be that some of the Cham dialects in Cambo?

dia date back, not to the fall of the southern capital in 1471, but even earlier to the
fall of the northern capital roughly five hundred years earlier.
Phan Rang
Cham,
an incipient
tone
system
For a long period after the breakup of PC, Cham was in contact with atonal, pos?
sibly registral MK languages and developed a register system, before breaking up
into Phan Rang Cham, Western Cham, and Haroi. However, certainly since the
fall of the southern capital Vijaya, in Binh-dinh,
in 1471, a major influence on
Phan Rang
the fully tonal Vietnamese with many of the Phan Cham

in
In this setting, Phan Rang Cham has steadily
Vietnamese.
becoming bilingual
become less registral and more tonal.
The
has been
literature
The available descriptions of Phan Rang Cham tones are excellent. The earliest
description seems to be Doris Blood's in 1962, followed by David Blood's in
1967. These preliminary descriptions by the Bloods clearly distinguished a two
188
Chapter
way tonal distinction for Phan Rang Cham. Later, on the basis of the material in
Blood (1962), Greenberg
(1970:139) wrote that Cham has followed a regional
the
of lowered pitch after what were originally voiced
tendency,
development
the subsequent
obstruents, with
obstruents causing
loss of voicing
in the
distinctions
initial
...a hitherto atonal language to generate a tonal system of two levels,

rather than to double the number of tones as in other languages of
Southeast Asia.
Just four years later, Fr. G?rard Moussay
analyzed Phan Rang Cham in his
Chamic-French
that there were four tones in
dictionary writing (1971:xiii-xiv)
are
the Cham language, and, although they
noted in the transcription by any sign,
are
on
went
to describe them as: an even tone,
He
for
needed
they
speaking.
occurring when the vowels are used alone or else when preceded by a normal
consonant, a low tone when vowels are preceded by a long consonant (the exam?
ples indicate he meant the orthographically voiced obstruents), a departing ("ris?
ing") tone when the vowel is preceded by a normal consonant and followed by a
glottal stop, and a falling tone when the vowel is preceded by a long consonant
(again, what are orthographically voiced obstruents) and followed by glottal clo?
sure.
useful
say
In 1992, Moussay's
analysis was basically confirmed by the extremely
and Gregerson, who like Mous?
instrumental study by Han, Edmondson,
Phan
analyzed
depending
upon
The evolution
Rang
whether
Cham
as
having
or not Moussay's
of Phan Rang Cham
three-
"departing"
or
four-way
tone was
distinction,
phonemic.
tones
The path of Phan Rang Cham tonogenesis

is quite transparent. The first stage
probably dates back to before Phan Rang Cham, Western Cham, and Haroi split
up into separate dialects: after voiced obstruents, breathy phonation (second reg?
ister) developed while after the remaining obstruents there was the now contrast?
it is not clear when the
ing unmarked, modal phonation (first register). Although
low-pitch that accompanied the breathy phonation developed into low tone, con?
trasting with the higher pitched modal phonation which developed into high tone,
but from its inception that register system probably included redundant pitch dis?
tinctions.
Later, under the influence of bilingual contact with Vietnamese

speakers,
the pitch distinctions became more salient than the phonation differences and
when the voicing distinction between obstruents was lost, the result was a two
The Origins
of Registers
tonal distinction. With
way
originally
189
and Tones
the loss of the voicing distinction in the initials, the

the
low-pitched vowels became low-toned, while
breathy-voiced,
modal-voiced
vowels automatically constituted a higher-pitched
tone,
remaining
was
The
tone
default.
the
low
and
for
the
essentially by
only remaining step
high
tone to each be split further on the basis of the presence or absence of a final glot?
tal stop. With the final glottal stops, although it is obvious that Phan Rang Cham
final glottal stops affect pitch, it is not clear if the pitch difference is still allo
phonically predictable synchronically or if it is already fully phonemic. However,
the source of the pitch is known, the direction of the change is
diachronically,
clear, and even the inevitable outcome seems obvious. See Figure 16.
For monosyllables,
the developments are precisely as described, but for
disyllabic forms the situation ismore complicated. In some cases, the tone of the
main syllable is not predictable from the initial of the main syllable, but instead
must be predicted from the initial of the presyllable. Thus, if the PC presyllable
originally began with a voiced obstruent and themain syllable initial was a sono?
rant or *-h-, itwas the voiced obstruent of the presyllable that resulted in the low
toned reflex inmodern Phan Rang. In such cases, the breathiness of the presylla?
ble (originally, from its voiced obstruent), spread from the presyllable to themain
the main syllable began with other than a sonorant or *-h-, no
syllable. When
such spreading from the presyllable to the main syllable took place, at least in
Phan Rang Cham.
Remnants
of this earlier register system still exist in Phan Rang Cham.

and Gregerson point out (1992),
study, Han, Edmondson,
In their instrumental
for
instance,
that
breathiness
from PC voiced obstruents

labic
is regularly
found
in the monosyllables
descended
and it occurs, although only sporadically
in the disyl?
forms.
Figure
16: Phan Rang Cham
Initials
classes:
tonogenesis
Resulting
registers:
PC initials
(except voiced
obstruents)
modal voice
+ higher pitch
inmonosyllables
Resulting
tone
classes:
incipient high tone

with glottal finals
high tone with
non-glottal finals
low tone with
PC voiced
obstruents
breathy voice
+ lower pitch
glottal finals
low tone with
non-glottal
finals
190
Chapter
Now, of course, as Doris Blood

was confirmed by the Han, Edmondson,
forms are even more clearly distinguished
PC voiced
obstruents
> breathy
in passing and as
(1962:12) mentioned
and Gregerson studies, the Phan Rang
by low or low-rising pitch.
voice > low tone
It is important to realize that the path of development

is from PC voiced
obstruents to breathy phonation (second register) to low tone. If the middle step is
left out, it leaves the impression that low tone developed directly from voiced
obstruents?that
synopsis would be not just misleading but quite false.
Table 121: PC voiced obstruent > breathy voice > low tones
PNB
*pah
*qyurj
PC
W. Cham
*blsy
*blah
play
plah
plah
'chop; split'
*pa-bley
paplay
papl?y
'sell'
*brsy
pray
*bra:s
prah
'buy'
'give'
prah
-1
'rice (husked)'
f
x*darj
tan
t?n
*dua
toa
twa
'two'
*dha:n
than
than
'branch;
x*dhog
thon
thon
'knife'
*?adh?y
thay
they
x*glsh
klsh
klsh
'tired'
*glay
klai
r?m-klay
'forest;
kah
kah
*gah
gam
kom
X*,
gar
*guy
The
PR Cham
evidence
kui
for the earlier
stand;
stop'
bough'
'forehead'
wild,
savage'
'side, direction;
kam
'to cover'
k?r
k?r
'handle (knife)'
'carry on back'
kuy
bank'
stage with its contrast between

modal and breathy register exists everywhere: in Phan Rang Cham clear traces of
voice quality differences were found by Han, Edmondson, and Gregerson (1992)
and in the closely-related Western Cham dialect, there is a register system that is
register
very much like the earlier register system of Phan Rang. In addition, there is
other evidence: notice that it is not voicing per se that led to low tone, but that the
development of low tone was mediated through an intermediate stage of breathi?
ness. Among
other things, this apparently necessary
intermediate
stage accounts
The Origins
191
and Tones
of Registers
for the fact that it is only after the reflexes of the old PC voiced obstruents, which
led to breathiness on the following vowel, that low tone developed. After the
voiced sonorants, which are obviously voiced but which are far less likely to
following breathiness on the vowel, low tone never developed.
Both the correlations with PC and the subsequent chain of developments
are still evident from a comparison of PC, Western Cham, and the Phan Rang
As Table 121 shows, if PC initial was a voiced obstruent, the
monosyllables.
develop
Western
Cham form has a breathy vowel, and the Phan Rang Cham monosyllable
traces of residual breathiness.
has low-tone with
Table 122: Other PC initials > modal
PNB
*p?r
*tr?g
PC
PR Cham
*kra
kra
kra
x*po
po 'title'
po
'master;
'to open'
'monkey'
x*pah
pah
pah
x*par
par
'to fly'
*pliih
par
ha pluh
pl?h
'ten'
*tuh
tuh
t?h
'to pour'
tug
'stomach;
tuy
t?l
'to follow'
'intestines'
*tun
tug
*tuy
tui
x*tal
t?l
*kla:s
klah
x*trorj
trorj
trog
'eggplant'
*trey
tray
'full,
*klam
makl?m
tr?y
kl?m
*klaw
klau
klaw
'laugh'
*klow
klau
klow
'three'
*trun
tr?n
tr?n
'descend'
*wah
*roy
W. Cham
register > high ton
wan
x*ruay
ruai
x*war
war
x*mag
*naw
*wil
*nam
*?u
'to free'
rqag
nau
w?l 'circle'
kl?h
wan
roy
lord'
abdomen'
until'
'arrive;
'escape'
satiated'
'afternoon;
night'
'to fish'
'fly; bug; insect'
war
'forget'
mig
'from'
naw
'go; walk'
w?l
'round'
n?m
n?m
'six'
?u
?u
'he,
she;
they'
192
Chapter
121, the Western Cham forms are all in the breathy register
the
dot
under the initial obstruent). The Phan Rang Cham obstruents
(marked by
are marked in precisely the same way; in Phan Rang Cham the dot indicates not
register but that the form has a low tone (although usually with residual accompa?
In Table
nying breathiness).
Similarly, the origins of the contrasting Phan Rang tone are likewise evi?
dent from a comparison of the PC, Western Cham, and the Phan Rang monosylla?
bles. As Table 122 shows, if PC initial was not a voiced obstruent, theWestern
form has a first register vowel (except in the case of sonorants), and the
Phan Rang Cham monosyllable
has a high-tone without any traces of breathiness.
In Table 122, theWestern Cham forms containing initial obstruents are
Cham
register (indicated in the table by the lack of a dot under

theWestern Cham initial obstruent). However, there is a secondary development
restricted toWestern Cham (not found in either Phan Rang Cham or in closely
in the unmarked modal
related Haroi), where all sonorant initial forms are also in the breathy register
(marked in this table but not everywhere in this work by a dot under the initial
sonorant). In Phan Rang Cham, all of the forms are in the unmarked high tone
(indicated
Registers
in the table by the lack of a dot under the initial consonant).
split byfinal
glottal
stop
register split, Phan Rang Cham developed a low tone from the
a default high tone from the first register. Then, these two
and
register
tone classes were further split by the final glottal stop. The Phan Rang Cham low
After
the original
second
a final glottal stop became a low, glottal tone, indicated reasonably

in Table 123 by the combination of the dot under the obstruent and the
tone with
enough
final glottal stop.
Table 123: Breathy

PNB
PC
register + final glottal stop

W.Cham
*bap
pau?
p??
'fill; full'
*bru?-n-f
pr??
pr??
'rotten'
to?
to?
'sit; live; stay'
k3?
'kettle; glazed clay pot'
fc1^?
'vulture; garuda'
*do:k
k3?
xgrak
*ajeq
PR Cham
x*go?
fr9?
*jahit
?hi?
?hi?
'sew'
x*joh
?oh
x*ju:?
joh 'snap'
cu?
cu?
'broken; spoilt'
'black'
x*js?
???
?s?
'near; about to'
The Origins
of Registers
193
and Tones
Cham forms are all in the breathy register (marked by the

same dot under the obstruent that in Phan Rang Cham indicates low tone).
Notice
that theWestern
In a parallel way, the Phan Rang Cham high-tone with a final glottal stop
became a high, glottal tone, which Moussay describes as rising, indicated reason?
ably enough in the Table 124 by the absence of a dot under the initial consonant
and the presence of the final glottal stop.
An enlightening discussion of the tone splitting effect of final glottal
on
the Phan Rang tonal system is found inHan, Edmondson, and Gregerson
stop
inwhich they note that various excellent scholars have commented
(1992:41-42),
about the perceptible pitch difference between non-glottal final and glottal forms
(which descend from the PC voiceless stops: *-p, *-t, *-k, *-?). As Han, Edmond?
son, and Gregerson wrote, despite the fact that he considered the difference non
nonetheless wrote that such Cham forms had
(1967:29)
before
both
the final glottal stop and before final -h. And,
noticeably higher pitch
as discussed earlier, Moussay
further splits the low-toned and high-toned
into
phonemic,
forms with
Blood
and without
final glottal stops. Similarly, Han, Edmondson,

and
a
that
Thi
tone
also
Chau
four
with
(1987)
Gregerson report
system,
Hoang
posits
the glottal finals and the nonglottal finals distinguished.
Table 124: Modal
PC
register + final glottal
W. Cham
PR Cham
x*puac
puai?
poy?
*pa:t
pa?
pa? 'four'
x*pet
p??
p??
x*pro:k
stop
'scold; talk'
'pick, pluck'
pro?
pro?
'squirrel'
k?? x*ke?
k??
'bite; snap at; peck'
x*cat
ca?
ca?
'mountain
x*ko:?
ko?
ako?-l
'white'
x*rak
ra?
har??
'grass;
x*ls?
le?
le?
'fall into'
*ga?
x*lo:k
g??
lo?
ag??
lo?
'make, do'
mi?
'take;
ma?
*mat
range'
weeds'
'to peel'
fetch,
get'
Basing their analysis in part on instrumental findings, Han, Edmondson,

and Gregerson
(1992:41) differ from other investigators in distinguishing
only
three tones: the expected two low tones, one glottal-final and the other without a
glottal final, but for the high tones they only posit a single tone, preferring to ana?
lyze Phan Rang Cham as having a single high tone but with allophonic
ferences associated with the presence or absence of a glottal final.
pitch dif?
194
Chapter
From
a historical
the question of whether contemporary

perspective,
Phan Rang Cham has a four-tone system or a three-tone system is irrelevant. It is
clear what the current system developed out of and even what direction it seems
to be heading.
However, what is missing from these accounts is adequate information
on the Phan Rang reflexes of PC forms that ended in *-h. Only Blood (1967) has
commented on them specifically, and he treated them as patterning with the final
glottal stops, that is, he reported such forms as having allophonic but noticeable
extra high pitch. An instrumental examination of these forms would be quite
interesting.
Transparency
and phonation
spreading
in disyllabic forms is slightly more complicated. If the presyllable

a
with
PC
voiced obstruent and the initial onset of the main syllable was
began
either a sonorant or a medial voiceless fricative *s or *h, the breathy voice of the
The situation
second register spread from the presyllable

ing in a low-toned main syllable.
to the main
syllable ultimately
result?
125 shows, in disyllabic forms beginning with a PC voiced

sonorants
the breathy voice of the register spread through medial
As Table
obstruent,
resulting in the main syllable having second register inWestern Cham, indicated
by a dot under the obstruent, and in low tone in Phan Rang Cham, again indicated
by a dot under the obstruent.
PC
through sonorants
NR
W.Cham
PR Cham
*bara
bara
pra
*barah
barah
paran
*bulow
bilau
'body hair; feathers'
ia bila:t
plau
ea plan
pilow
*bula:n
pilan
'moon;
*buga
*dalam
bug?
pagur -f
tal?m
pigu
tal?m
'flower'
tanaw
'lake'
'blood'
dalap
-n
pir?
'shoulder'
danau
*darah
darah
tarah
tar?h
*dilah
gilah < *gdaruai (m)
talah
talan
t-aruai (m)
taroy (m)
'thorn'
kan?m
'cloud'
?alan
?alan
'road;
p?a
pila
'tusk;
pra
tara
'girl
*gunam
*jala:n
x*bala
jala:t
bala
*dara
dara
month'
'deep; inside'
*danaw
*dursy
swollen'
'swell;
'tongue'
path'
ivory'
(c.
teenage)'
The Origins
195
and Tones
of Registers
In addition to spreading through sonorants, the breathy voice of the sec?

ond register also spread through medial *s and *h. All theWestern Cham forms
except 'iron' are in second register, just as all the Phan Rang Cham forms are in
tone.
low

Malay
jahit
jahat
dahi
PC
*
PR Cham
W. Cham
?hi?
jahit
*s and *h
through medial
'sew'
?hi?
*jaha:t
?ha?
'bad;
wicked'
*?adh?y
thay
they
'forehead'
akar
*?ugha:r
kha
ukha
'root'
b?si
*basey
pasay
pith?y
'iron'
The form 'forehead' has been included in the table to show that voiced aspirated
stops also show low tone. The Malay forms in the table show that one source of
PC voiced aspirated
medial
initials is the reduction of earlier disyllabic forms.

However, the breathy voice of the second register did not spread through
voiceless stops either inWestern Cham or in Phan Rang Cham.
Table 127: Failure
PNB
Malay
batu
to spread through voiceless

PC
PR Cham
*batow
patau
patow,
patow
'stone7
*batsy
patay
pat?y
'banana'
batuk
*batuk
*dikit
pat??
taki?
pat??
taki?
'cough'
dikit
pato
'teach'
*pr?t
x*bato
None
W. Cham
stops
pato
'few;
little'
127 show second register and none of

forms show low tone. Particularly interesting is the second Phan
of theWestern
Cham forms in Table
the Phan Rang

Rang form listed for 'stone'. The initial /p/ has a dot under it, indicating a second
register vowel, but the It/ beginning the main syllable does not. In effect, the first
syllable of the form had second register, but the breathy voice did not spread to
the second syllable, and thus the main syllable does not have a low tone reflex.
sono?
The patterns of spreading are in themselves interesting. Medial
rants appear completely permeable to spreading, while medial obstruents are far
more
resistant, with only medial
The history
of Phan Rang Cham
*s and *h allowing
spreading.
contact
of
The Phan Rang data is interesting from the viewpoint of the mechanism
a
The
Phan
from
Western
involved.
Cham
Cham
transition
Rang
style
change
196
Chapter
toward
system
register
a Vietnamese
lution than a revolution ?

shifts in emphasis.
style
tone
system
represents
more
of
an evo?
the actual changes look to be little more than slight

To return to something Eug?nie Henderson
(1967:171) said
thirty years ago:

It is important to recognize that pitch is frequently only one of the pho?
netic components of "tone" as a phonological
category. A phonological
tone is in our area [South East Asia] very frequently a complex of other
as intensity, duration, voice quality, final
features besides pitch?such
and so on.
glottal constriction
are best described
The six tones of Vietnam
as complexes
of features with pitch

1984-5:16). And, among these features, it is
that are coming to the fore in the emergent
Phan Rang Cham tone system. The Vietnamese
system contains breathy versus
contrasts:
tone
the
which is "often accompanied
non-breathy
low-pitched huyen
contrasts
with
the
clear register found with the mid or
by breathy voice quality"
but one of these features (Thompson
those which are salient inVietnamese
high-mid pitched ngang tone. And, among the forms with final stops, the nang
tone with its the low-dropping-pitch
and which "ends in [a] stop or is cut off
contrasts
with
the sac tone with its high-rising pitch.
abruptly by [a] glottal stop"
That
is, the Vietnamese
tone
system
contains
the very
complexes
of phonological
oppositions that have formed the basis of the incipient tone system in Phan Rang.
The stages in Phan Rang Cham tonogenesis are still transparent. Distinct
layers of external contact have precipitated each of the stages of internal change.
contact with MK led to the restructuring of evenly-stressed
Austronesian
disylla?
bles into iambic syllables with final stress. Later contact with MK register lan?
to
led
guages
Vietnamese
successive
language adjusting
guage
system,
register
and
has produced an increasingly

Phan Rang Cham phonological
intense
contact
the
tone
system
of
tonal Phan Rang. The entire history of

restructurings has been the history of a
its internal paths of change
to follow paths illuminated by lan?
contact.
The
social contact
is bilingualism, with the Phan Rang Cham

with no significant language shifting,
speakers being bilingual
except of course, away from Phan Rang Cham, when Phan Rang speakers shift to
involved
in Vietnamese,
Vietnamese.
It seem appropriate
in the process of dismissing
adds an interesting twist:
...a few scholars
to close
this section with a quote from Lafont, who

of tones in Cham, inadvertently
the very existence
have described
never given evidence
Cham
for this assertion,
as a tone language but have

for there are no phonemic tones
The Origins
of Registers
in Cham
197
and Tones
It seems that this error originated

Cam-Vietnamien
as there are in Vietnamese.
of the Dictionnarie
from l'Introduction
(pp. xii-xiv)
Fran?is (1971) in which the author talked about four tones while he
referred to the intonation particular to the Phan Rang, Phan Ri region
where the Cham people were educated in Vietnamese
schools, intona?
tion which is not to be found elsewhere, neither in the Chir-D?c
region
nor among the Chams in Cambodia, except in interrogative sentences
which are always characterized by a higher register. (1994b: 12-13; note:
the grammar of the translation has been emended)
The explicit and somewhat puzzling claim that Cham has no tones at all can sim?
ply be dismissed, although there is no question that the Cham tones are not as
tones. More interesting, however, are the
perceptually salient as the Vietnamese
suggestion that not all dialects have tones and the almost inadvertent suggestion
that Cham
correlates directly with
tonogenesis
the degree of contact with Viet?
namese.
HAROI VOWELS AND RESTRUCTUREDREGISTER

comments on the Haroi vowels have been scattered throughout the
on
is too complex to be explained in passing
vowels, their development
chapter
and too intimately involved with register and Hr? contact to be discussed in isola?
Although
tion.
Thus,
it
is necessary
to bring
these
prior
expand upon them as, from a Chamic viewpoint,

a unique, fascinating path of development.
observations
together
here
and
the Haroi vowels have followed
The original PC vowel system underwent massive splitting and realign?

settling into the system of Modern Haroi. Prior to the realignments
within Haroi, there were a series of other earlier realignments often shared with
ment
before
Cham ("Pre-Haroi vowel changes" on page 199), which were not recog?
as
1996. However, the major vehicle for this complete
nized
such in Thurgood
vowel splitting. Under the influence of
restructuring was voice quality-induced
Western
tense voice
(induced by the proto-voiceless

obstruents) certain monophthongs
lowered. And, under the influence of breathy voice (induced by the proto
were raised, other monophthongs
voiced obstruents),
certain monophthongs
still
and
other
developed on-glides,
diphthongs had their onsets raised.
were
vowels
The first indication of Haroi's uniqueness is the extraordinary number of

and, if one is doing historical work, the existence of two or more reflexes
for each proto-vowel.

tems of most Chamic
In striking contrast to the typical nine- or ten-vowel sys?

languages, Haroi has a plethora of vowels: 11 simple vow
198
Chapter
els, each occurring both long and short (see Table 128; also cf. Tegenfeldt and
Goschnick,
1977:1), and 17 diphthongs and triphthongs. Beyond these, Haroi
also has some 10 rarely occurring nasalized vowels.
Table 128: Haroi
simple vowels
front
high:
closed
i u i
open
central
mid
back
s
low
are no vowel
contrasts in the Haroi presyllable. As one conse?

to
final syllable stress into pre-Chamic, all of the common
quence of the shift
or in the
vowel distinctions
in modern Haroi occur either in monosyllables
There
stressed main
syllable of disyllables.
Table 129: Haroi
diphthongs
and triphthongs
front
high
mid
central
ia
iau
ea
eau oi
back
ii
iai
ia
iau
ui
oa
eiou
low
ua
oai
ai
au
In addition, Haroi has a rich array of diphthongs and triphthongs (see Table 129),
the result of registrally-induced
changes.
it is not clear what significance to attach to the fact, it is worth
Although
it would be
also has an eleven-vowel
system. Minimally,
noting that Vietnamese
important to examine the vowel inventories of various other languages of the
many
region to see how common this particular configuration

ticular whether it occurs in Bahnar.
Haroi
restructured
is in the area and in par?
register
(1976) termed a restructured register system. In

the case of Haroi, between PC and modern Haroi the following chain of events
has occurred: (1) certain classes of initials led to distinctive phonation differences
Modern Haroi has what Huffman
of Registers
The Origins
199
and Tones
on the following vowels, that is, to voice quality distinctions; (2) the voice quality
that is, led to a register
differences on the vowels produced vowel distinctions,
vowel sets associated with different voice
system with vowel registers?distinct
quality differences; and, (3) the voice quality distinctions that originally condi?
tioned the vowel splits disappeared, leaving behind a large number of now uncon?
ditioned vowel distinctions; inHuffman's terms, it became a restructured register
system with the proliferation of vowels associated with restructured register sys?
tems.
vowel changes
Pre-Haroi
Prior to the alignments within Haroi, there were a series of other earlier realign?
ments, often shared with Western Cham, which were not recognized as pre-Haroi
in Thurgood 1966: the much earlier lowering of the onsets of certain diphthongs,
a change which is largely shared with Western Cham; the merger of PC *-ow
with the reflexes of *-?u, which is clear from the fact that the vowel of the rhyme
behaves like a low vowel during the registrally-induced
vowel splits; and, in a
number of cases, PC shwa was raised or backed before specific finals, again a
change that preceded
The early pre-Haroi
vowel
case,
reflexes of PC shwa
the changes
Among
system
the change
was
shifts due to the influence of voice quality.
the Haroi vowel
that occurred before
the development
was
determined
of
by
three
the
the registral realignment

distinct
syllable-final
reflexes
for PC
of the Haroi
shwa.
In each
consonant.
PC*-9f)>*-i()
Before final *-n, PC shwa became l-i-l. The most obvious thing about
130 is that three of the four Haroi reflexes of PC *-ag contain a barred-i.
The shwa in the remaining form is the expected reflex of an earlier barred-i which
Table
lowered due to the initial voiceless consonant (see "Registers and the
vowel splitting patterns" on page 201, for discussion of this vowel lowering).
Beyond this, however, there is something else quite interesting about the
has been
chart: the reflexes
inHeadley's Western Cham match the Haroi reflexes perfectly,

to
down
the
right
split conditioned by the voiceless obstruent initial. This sug?
that
this
gests
change predates from PC *-arj > *-in predates Haroi, and that at
this
least
dialect ofWestern Cham and Haroi were particularly close either genet?
ically, through contact, or both. Note that this change is not shared, however, with
the Kvoeu-Hor
and Friberg Western
Cham
forms or with
the Phan Rang Cham
forms.
200
Chapter
Table 130: The reflexes of PC *-dg

PNB
Rade
PC
Haroi
W.Cham
W.Cham
(Kvoeu-Hor)
(Headley)
x*?ag
*qy?n
??g
x*mag
man
x*dag
dan
x*pag
mig
thin
peg
?arj
??rj
'to eat'
man
mig
'from'
bit)
?ig
tag
tig
pag
pag
peg
PR Cham
c?g
'stand; stop'
pag
'to nail,
hammer'
PC
*-at,
*-a h remained-d?,
and
*-ak,
and
*-dk,
*-dh
as a shwa. Else?
Before final *-t, final *-k, and final *-h, PC shwa was preserved
where it became something else.
Table 131: The reflexes of PC *-a?, *-dk, and *-dh
PNB
PC
Chru
Rade
x*cat
Haroi
ca?
'mountain
papa?
'to fan'
ra?
'grass;
**&
'vulture;
pah
p?h
'to open'
jarlah -i
madah
tlih
'descend;
mathih
'awaken'
ci?-i
ca?
x*pa-pat
*p?y
x*rak
rak
ra?
"""
"""
xgrak
?
x*pah
x*glah
x*madah
?
madih
range'
weeds'
garuda'
collapse'
For these rhymes, the final shwa is preserved as such inHaroi, unless
began in PC with a voiced obstruent. In that case, the shwa was
barred-i (see page 201, for discussion
All
The
*-on,
other
examples
of PC
*-a-
>
of this vowel raising).
*-3
remaining words with PC shwa backed

*-ar
>,
and
*-am
>
*-om.
the syllable
raised to a
In these
cases,
to l-ol, that is, the rhyme *-an >

unless
the
initial
of
the main
sylla?
ble or of the presyllable was originally a voiced obstruent, the vowel remained an
open-o. However, when the initial of the presyllable was a voiced obstruent, this
open-o was raised to /u/ (see page 201, for discussion of this vowel raising).
201
and Tones
of Registers
The Origins
Table 132: The reflexes of the other PC forms with *-a
PNB
Rade
PC
l?n
*lan
*p?r
pakan
x*par
phiar
x*katar
katar
x*war
-v
drin
-if
-iv
war
'other; different'
tr?n
'numb'
wol
war
'earth'
pak?n
pol
katol
par
x*v
*yar
"hag?r
Haroi
15n
Ian
mak?n
*bukan
xdr?n
Chru
'to fly'
'corn;
grain'
'forget'
raise'
'lift,
dar
yol
thol
'handle (knife)'
x*dar
dar
x*gar
gr?n -ivf
gar
kh?l
x*sagar
hagar
sagar
akh?l
'drum'
x*sadar
hadar
sadar
ath?l
'remember'
x*?am
am
am
5m
'to roast'
pakhtim
x*pagam
It is obvious
that the changes
'bury'
'dove'
in the reflexes of the PC shwa forms took
the registrally-induced
(see page 201). Cer?
realignments
tainly, with the changes before the final velar nasal, this is obvious because the
change is shared with Western Cham, but, even with the forms that backed to an
place
vowel
before
open-o, the modern Haroi distribution requires that the change preceded the
in "Registers and the vowel splitting patterns" on
vowel alignments discussed
page 201.
Registers
and the vowel
splitting
patterns
The two specific marked registers are associated with the remaining changes that
produced the modern Haroi vowel splits: the first register, which consists of the
conditioned set of vowels associated with the tense pho?
allophonically-distinct
nation type that evolved after the PC voiceless obstruents and the second register,
set of vowels associated with the
which consists of the allophonically-distinct
these
that evolved after the PC voiced obstruents. When
breathy phonation
vowel-quality-conditioning
phonation differences were lost, allophonic vowel
differences became phonemic, and Haroi became a restructured register language
(see Figure
17).
202
Chapter
17: Restructured
Figure
register and Haroi
Vowel
Initial
classes:
high vowels
tense
obstruents
voice
lower
Other PC
modal
no
voice
PC voiced
breathy
obstruents
Restructured
register :
vowels:
PC voiceless
initials
vowel splitting
Effect on
quality:
low and mid,

vowels
voice
proliferation
of vowels
effect
raise
For monosyllabic
words, the various vowel types interacted with the
of articulation of the PC syllable-initial consonant to produce the modern
vowel splitting patterns. These vowel splitting patterns are summarized inTable 133.
manner
Table 133: Consonant
types, vowel classes, and vowel splitting
voiceless
obstruents
>
tense
voice
(= first register)
glottalized
voiced
aspirates,
voiced
obstruents
obstruents,
&
> breathy register

(= second register)
sonorants
high
(onset)
lowered;
vowels;
*-ag >
unchanged
raised and backed:
centering
diphthongs:
**-ia-
>
-ia-;
*ua
>
*oa
**-ua-
>
-ua;
*ia
>
*ea
unchanged
unchanged
-u- /___m,
mid
*s; *a; *o;
*-ey
>
unchanged
unchanged
unchanged
unchanged
*-oi
low
vowels
After
unchanged
>-ag
*-ig
voiceless
-?
raised:
t; i; u (u);
-ii (+ fronted)
developed
obstruents
-i- onset
(tense voice)
The tense voice quality of first register vowels, from the PC voiceless obstruents,
the voiceless obstruents did not affect
led to high vowel lowering. Otherwise,
The Origins
of Registers
203
and Tones
vowel quality. Note that the high vowels include not just the high vowel inherited
from PC high vowels, but also the secondarily-derived
high barred-i that devel?
before
final
velar
in
Western
Cham
both
and
Haroi (page 199).
nasals
oped
Table 134: Tense voice and vowel lowering after voiceless
PNB
PC
Rade
kasi
xxsisi(r)
x*dhog
*qb!c
Chru
tasi
obstruents
Haroi
casei
pi:?pit
pei?
phi:?
si:?
phei?
sei? 'sew'
'a comb'
'sleep; close eyes'

'bitter; bile'
*phit
phi?
*jahit
*kulit
jh?t
kl?t
kali:?
kalei?
'skin'
*ku?it
kaft??
kaft?:?
ka?ei?
'yellow;
*tasi?
kasi?
tasi?
casei?
'sea;
*thu
thu
thu
thou
'dry'
*?iku
ku
aku
akou
'tail'
*lukut
ek?t
laku:?
lakou?
'absent'
turn eric'
ocean'
'avoid'
*kantut
x*ka?u:?
katu:?
katou?; tout
'fart'
ka??:?
ka?ou?
'worried;
sad'
P9y
khin
psg
'to nail,
pound'
kh?n
'dare'
psg
xkhi:n
*tr?h
x*cuh
?uh
coh
'burn
*krih
kr??-f
kri:h -1
kreh
'to whittle'
*phun
ph?n
phun
phon
'trunk; log; plant'
x*truh
truh
truh
troh
'arrive'
toh
'change'
klum
tlom; kalom
'to cover'
*tuh
*klum
As
trns.'
*trun
tr?n
trun
tron
'descend'
*pluh
spluh
aploh
'ten'
*thun
pluh
th?n
thun
thon
'year'
*tuh
tuh
tuh
'to
*tuy
tui hlus
tui
catoh; toh
toi
pour'
'to follow'
is evident
from the examples in Table 134, both the lil and the lui
into the diphthongs /ei/ and /ou/, respectively, word-finally
or before a
develop
final glottal stop.
204
Chapter
obstruents
After glottalized
and sonorants
(modal voice)
vowel quality changes occurred after either the glottalized obstruents nor
after the sonorants. This is not surprising: both the glottalized obstruents and the
sonorants were associated with modal voice, a clear phonation type that would
No
not have been expected

There
to affect vowel quality.

a subset
is, however,
of
after
reflexes
sonorants
where
situa?
the
tion is complicated by an assimilative interaction involving a PC *i or *u in the

presyllable or a *y as the onset of the main syllable that caused vowel raising, a
phenomenon otherwise only associated with voiced obstruents (see Table 135).
The existence
by Burnham
of such irregular correspondences was noticed by Lee (1977b) and

(1976), who both commented on these unexpectedly
high vowel
reflexes.
assimilation
Table 135: Height
after high (semi-)vowels
PNB
PC
Rade
Chru
Haroi
Tag
x*ya:g
yag
ya:g
yiag
'spirit; god'
yiah
'destroy;
*yah
*raya
yap
ya:u?
yiau?
'count'
*buya
mya
bia
payia
'crocodile'
*?ular
ala
ala
alia
'snake'
*huma
hama
hama
hamia
'field'
*lima
ema
lama
lamia
'five'
?
ana ania
*?ina
eman
*lama:n
sam?
*yor
*kayua
x*hayua?
*yua
-v
yuan
kayua
-v
yuan
'mother;
lamian
lama:n
*samaw
*yuan
apart'
layia
*ya:p
*yun
take
-f
hamiau
'prompt;
yun
'Vietnamese'
kayua
kayua
yua?
yu?
major;
big'
'elephant'
ka-
on
time'
'because'
dah
yua
'harvest
(rice)'
'use'
There are three groups of sonorant-initial forms in Table 135 which have
unexpectedly high vowels. In each case, the presence of either *i, *u, or *y seems
to correlate with the otherwise unexpected reflex. The first group contains low
but with a barred-i onset; in each case, this onset follows *y. The second
group also contains low vowels, but in this case it appears that it is the high vowel
vowels,
205
and Tones
of Registers
The Origins
*i or *u in the immediately preceding syllable that conditioned the height. The

last group contains an l-u-l or /-ua/ which normally would have been lowered or
had its onset lowered much
earlier in the history of Haroi (page 135) but which is

in
modern
Haroi. This unexpected height correlates with the
unexpectedly high
initial *y. These changes
involve /a/ diphthongizing,
with the onset of /ia/
increasing its palatality as it assimilates to the preceding element, while
ond part of the diphthong decreases its sonority (Donegan 1985:145-46)
After
voiced aspirates
the sec?
(modal voice)
Contrary to what is implied in a table in Thurgood (1996), once recent borrow?

ings have been culled out, it becomes apparent that the so-called voiced aspirates
behave just like the glottalized obstruents and the sonorants?they
have no effect
on the vowel quality.
whatsoever
Table 136: Reflexes
Chru
Haroi
thog
thog
*?adh?y
dhog
adhsi
thai
thai-v
'forehead'
*dha:n
adhan
tha:n
than
'branch'
*jahit
jh?t
si:?
sei?
'sew'
'bad;
'knife'
*jaha:t
jhat
jaha:?
s?t
*?ugha:r
agha
akha
akha
'root'
*pa-gha:g
bhag
kha?
pakhag
kh??
'forbid'
*pa-gha?
The vowel
After
voiced aspirates
PC
Rade
x*dhog
stop
of the so-called
seen
-i
in /ei/ in 'sew' is the expected
diphthongization
'dry
wicked'
over
fire'
of HI before a glottal
earlier.
voiced obstruents
(breathy voice)
As both Lee
(1977b) and Burnham (1976) noticed, the breathy voice associated

with the second register (from PC voiced obstruents) caused various mid vowels
to raise and the low vowels to develop a barred-i onglide (see Table 137).
206
Chapter
Table 137: Breathy

PNB
PC
voice quality and raising after voiced obstruents

Rade
x*glsh
x*joh
joh
Chru
Haroi
gleh
tlth
joh
suh
khu?
'broken; spoilt'
'kettle; clay pot'
'fruit; egg; elf.'
go?
boh
go?
boh
*do:k
dok
do:?
phuh
thu?
x*gam
g?m
gam
khum
'to cover'
*bra:s
braih
bra:h
priah
'rice (husked)'
*ba
ba
ba
*bap
*pah
'tired'
*boh
x*go?
gah
blah
'sit; live; stay'
take,
phia
'bring,
phiau?
khiah
'fill; full'
carry'
'side, direction;
bank'
*gah
*blah
blah
pliah
'chop; split'
*glag
dl?g
tliag
'look
*ja:k
jak
dlie lui
sia?
'invite'
glai
tliai
'forest, jungle; wild'
x*druam
druom
dro:m
trum
'fell
*dua
dua
dua
thua
'two'
*glay
at; watch'
a tree'
In addition, certain diphthongs monophthongize

in specific environments.
tional examples of changes in PC shwa can be seen in page 199, above.
Transparency
and phonation
Addi?
spreading
For disyllabic words, the situation is complicated by the fact that some classes of
initial consonants allow the phonation generated by the initial of
main-syllable
the presyllable to spread to the vowel of the main syllable. In such cases it is ini?
tial of the presyllable, not the initial of the main
ister of the main vowel.
syllable,
that determines
the reg?
Except when the initial of the presyllable is *s or *h, the Haroi spreading
patterns are remarkably straightforward. If the main syllable begins with a sono?
rant, it is the initial of the presyllable, not the initial of the main syllable, that
the register of the main syllable vowel. In other words, except in the
case of presyllable *s or *h, initial sonorants are transparent to register spreading.
For example, as both Burnham (1976) and Lee (1977b:89) noticed, if the pretonic
determines
The Origins
of Registers
207
and Tones
the vowel split?
syllable begins with a voiced obstruent, the main syllable follows

ting patterns associated with breathy phonation (See Table 138).
Table 138: Breathy phonation
x*bala
through sonorants
Chru
Haroi
ewa
jawa
cawia
mla
bla
palia
'tusk; ivory'
Rade
PC
xjawa
spreading
soul,
'breath,
air'
*buga
maga
baga
pagia
'flower'
*dara
era
dra
caria
'girl (c. teenage)'
*bara
mra
bra
pria
'shoulder'
erah
*darah
*dras
*barah
drah
carian
'blood'
drah
carian
'fast'
brah
swollen'
elah
dalah
priah
caliah
'swell;
*dilah
*gunam
kanam
ganam
kaniam
'cloud'
*dalam
elam
dalam
caliam
'deep; inside'
*bula:n
mlan
ea
palian
'moon;
bla:n
'tongue'
month'
elan
jalam
calian
'road;
enau
danau
caniau
'lake'
*boh maw
mamau
bamau
pam?au
'mushroom'
*bulow
ml?u
blau
paliau
'body
druai (m)
car?i (m)
'thorn'
*jala:n
*danaw
erue (m)
*durey
path'
hair'
In a parallel way, if the pretonic syllable begins with a voiceless

obstruent (again, other than *s or *h) and the main syllable begins with a sono?
rant, the main syllable follows the vowel splitting patterns associated with voice?
less obstruent phonation. The examples in Table 139 show forms in which the
phonation induced by the initial voiceless obstruent of the pretonic syllable has
the reflexes of PC high vowels *-i
spread to the main syllable. As a consequence,
are the reflexes expected after voiceless stops, not the
reflexes expected after monosyllables
beginning with sonorants.
For the last two examples in Table 139, the final vowel can also be attrib?
uted to the otherwise well-attested process whereby a HI before a final glottal stop
and *-u after sonorants
becomes
/ei/.
In contrast
that allow a great dealing of spreading, the

initial
allow none: indeed, it appears thatmain-syllable
block spreading.
to the sonorants
obstruents
syllable-initial
obstruents completely
208
Chapter
Table 139: Tense voice spreading

PC
Rade
kam un
*timun
through sonorants
Chru
*tamuh
*kalih
Haroi
tam?h
cam oh
'grow; sprout'
tamun
camon
'melon;
karlih
kaleh
cucumber'
*kulit
kl?t
kali:?
kalei?
'miserly'
'skin'
*ku?it
kam?
ka??:?
ka?ei?
'yellow;
tumeric'
It is important to note that the vowel splits cannot be attributed directly

to the influence of PC initials; instead, the splits are due to the influence of an
intervening phonation type. These two patterns of phonation spreading constitute
strong evidence that itwas the phonations correlated with
consonants
sonants, not the syllable-initial
themselves,
the syllable-initial con?

that caused the vowel
In phonetic terms, what must have spread from the pretonic syllable
sonorant of the main syllable was a specific phonation
through the syllable-initial
not
manner
the
of
articulation or the voicing of the pretonic syllable-initial
type,
splitting.
consonant.
Voice quality
In disyllabic
and the reflexes of PC presyllable
*s- and *h
*s or *h is the initial of the presyllable, there is a voice

quality-based split in the reflexes. It ismore than likely that early inHaroi PC *s
and *h- merged as *h-. Then, before the before tense-voiced main syllables (< PC
voiceless
forms when
the initial *h- was
lost (the first set of examples

but
before
main
140)
breathy-voiced
syllables (< PC voiced obstruents
ond set of examples in Table 140), this presyllabic *h- was retained.
obstruents),
However
before
the main-syllable
initials associated
in Table
the sec?
with modal
voice,
that is, before glottalized obstruents or before sonorants (and, presumably, before
voiced obstruents, if there were any), the *h- (< PC *s- and *h-) remained Haroi
Ibl.
Parallel
in which the reflexes of initial consonants have

developments
differed depending upon voice quality, that is, phonation type, are attested in
in Chinese, and elsewhere (cf. Thur?
Thai, in the Yi languages (Lolo-Burmese),
good, 1980, and numerous other sources).]
Table 140: Register

PNB
and thefate of PC presyllable

NR
*s- and *h
PC
Rade
Haroi
*hatay
tie
hatai
at?i -i
'liver'
*sapuh
*satuk
t?k
stu? 'boil'
apoh
ato?
'boil; cook'
*habow
hab?u
habau
aphiau
'ashes'
*hudip
had?p
hadiu?
ath?p
'live,
*huda:g
hadag
hada:g
athiag
'shrimp;
'broom;
sweep'
alive'
lobster'
*hubsy
habei
habai
aphui (m)
'taro;
tuber;
yam'
*huja:n
asian
'rain'
athiam
'ant'
asiau
'hammock'
athuh
'look for;
hajan
had?m
haja:n
x*haduah
duah
duah
x*sadar
hadar
sadar
athul
'remember'
x*sagar
hagar
sagar
akh?l
'drum'
xsi?jual -v
ha?jul
sa?ual
ha?jul
'light (not heavy)'
*halow
hl?u
halau
halau
'pestle'
*halim
hl?m
halim
halim
'rainy
*hurey
hrue (m)
harai
*haway
*haluh
hawie
hawai
harii (m)
hawai
hluh
haluh
haluh
*hulun
hl?n
halun
halun
*sidam
adam
-i
*sijaw
search'
*hare
'day;
season'
sun'
'rattan'
'perforated;
pierce'
harip
*sarip
'sip; slurp;
suck
*hla
x*sula
hla
servant;
'slave;
r
in'
sala
hala
'leaf
*sarum
srum ?
har?m
'sheath-like'
*sana
sana
hana
'roast;
*sanig
sanag
'fry'
parch'
*cach?g
*samaw
han?g
sam?
han?g
-f
hamiau
'think'
'prompt;
on
time'
210
Chapter
Haroi
contact with Hr?
things are known about Haroi from the linguistic evidence. The sub
grouping evidence on page 40 argues that Haroi originated as a Cham dialect,
along with Phan Rang and Western Cham. Similarly, the patterns of apparently
Certain
convergences with Rade suggest that Haroi was influenced by

some contact with Rade. However,
the strongest influence on the structure of
modern Haroi was an extended period of intense contact with MK languages,
non-inherited
with Hr?. In fact, the very name Haroi might be an adaptation of Hr?.
in
The description
The evidence of Hr? contact is uncontroversial.
Schrock, et al. (1966:217) sums up the ethnographic situation nicely [although
Hr? and
their description uses a single term to designate both theMK-speaking
especially
Haroi, both names will be used here]. The authors note that,
although the Haroi and the Hr? speak two distinct languages, both linguistic
as parts of a single tribe, with the
groups are usually classified ethnographically
Haroi speaking a Chamic language and being influenced by the Cham and the
the Chamic-speaking
Rade and the Hr? speaking Hr?, a Northern Bahnaric language.

As an illustration of the linguistic differences, Schrock, et al. (1966:217)
provide a brief table of linguistic forms (see Table 141, for an updated, slightly
expanded, and slightly modified version of that table).
Table 141: The MK Hr? and the Chamic Haroi
Eastern
PNB
*qbok
What
Bahnar
Hr?
(MK)
Haroi PR
(Chamic) Rade
Cham
PC
sa
sa
I ?ig
?ag
??g
x*?ag
'to eat'
ec
ec
I ma?iam
man?m
mi??m
*min?m
'to drink'
hnan
hnamo
I sag
'house'
bok
I ka oi
sag
ae
*sa:g
bok
thag
?
'grandfather'
141 to show is the respective lin?

from a modern viewpoint, it is uncon?
Schrock et al. intended the forms in Table
guistic affiliations of Hr? and Haroi. And,

troversial that Hr? isMK, while Haroi is Chamic.
Here, however, it is not the genetic affiliations
but the interlanguage con?

from the Schrock et al. description that the
tact that is of interest. It is evident

Haroi are not only referred to as Bahnar Cham but are also from an ethnographic
viewpoint (1966:217), "considered a subgroup of the eastern division of the Bah?
nar tribe". Certainly, the ethnography documents

intense and extended contact,
contact which obviously included bilingualism.
Although they incorrectly guessed that the contact language was Bahnar,
rather than Hr?, Tegenfeldt
and Goschnick
(1977:1-2) correctly suggested a
211
and Tones
of Registers
The Origins
contact and Haroi restructured register. That the

was
not
Hr?,
Bahnar, is now clear as are the changes. Hr?
language in question
has undergone vowel realignments quite parallel to those undergone in Haroi.
The evidence can be culled, with some minor readjustments of the reconstruc?
between MK
causal connection
tions, from Smith's (1972) reconstruction

includes Hr?.
of Proto-North-Bahnaric
(PNB), which
However, before Haroi can be directly compared with Hr? (as seen in
1972 reconstructions),
certain of Smith's reconstructions require modifi?
cation. Typologically,
there is usually a rather straight correlation between the
Smith's
of register complexes and vowel height movement:

to lower, while breathy-register vowels tend to rise;
tend
tense-register
in
Smith's
reconstructions
there is no regular relationship between
nonetheless,
the voice quality differences and vowel lowering and raising. In Smith's PNB
voice
quality
differences
vowels
under
reconstructions,
tense
voice,
certain
PNB
reconstructed
are
vowels
some?
times raised and sometimes

other PNB vowels
lowered; similarly, under the breathy voice, certain

are similarly sometimes raised and sometimes lowered.
in the vowel reconstructions,

it is possible to
a system in which, when they changed height, the PNB vowels low?
ered under tense voice and raised under breathy voice. The solution to this was in
With minor modifications
reconstruct
the vowels
Inmany cases, unless one took the effects of voice quality

the actual vowel reconstruction was somewhat arbitrary: that
themselves.
into consideration,
is, in many cases,
it appears that Smith had no principled basis for deciding

should be reconstructed as, for example, mid in PNB and
then raised in certain languages or whether they should be reconstructed as high
in PNB and then lowered in certain languages. Working with the added constraint
whether
that
certain vowels
tense
Smith's
voice
causes
reconstructions
dences between
to lower
vowels
can
be
adjusted
and
breathy
rather
voice quality and vowel movement
voice
easily,
causes
making
vowels
the
to rise,
correspon?
both internally consistent
and
typologically
expected.
Smith's reconstruc?
Although more work needs to be done modifying
or
tions before it can be determined whether such changes are unexceptional
whether they are sometimes conditioned, the minor modifications
made
already
have major additional benefits: it is now obvious that Hr?, the language most inti?
in contact with Haroi, has undergone voice quality-related vowel height
changes that closely parallel those found in Haroi. As Table 142 shows, at least
some of the Hr? high vowels lowered under tense voice, the quality equivalent to
mately
tense voice under which Haroi high vowels lowered.

And, as Table 143 shows, the Hr? mid and low vowels have raised under
breathy voice, the voice quality equivalent to the Haroi breathy voice under
which Haroi mid and low vowels raised.
the Haroi
212
Chapter
In short, the Haroi and the Hr? changes are not just parallel,
Haroi changes took place while in intense contact with Hr?.
Table 142: Hr? high vowels
PNB
but the
lowered under tense voice
Hr?
(Smith, modified)
*tum
torn
'all'
*bic
bee
'fat (v)'
*c(h)uy
c?y
'plant
*pih
pen
'pound
*asih
asen
'horse'
*rih
r?h
'play (instrument)'
*taqnih
taneh
'earth,
*(q)bul
bo
'lizard;
*kaciyh
kac?h
'sneeze'
rice'
rice'
soil'
gecko;
salamander'
It is tempting, by the way, to attribute the vowel raising in at least the

*-?y vowels in Table 143 to the final off-glide. However, a comparison of
the Hr? reflexes of PNB *-ay under breathy voice with the reflexes of PNB *-ay
PNB
under
modal
(that
in the vowel
voice
is, clear)
makes
it clear
that
breathy
voice
played
a role
raising.
Table 143: Hr? mid and low vowels raised under breathy voice
PNB
Hr?
(Smith, modified)
*c?m
ci m
'bird'
*kl?c
kl?c
'deaf
*ph?w
'happy'
*adr?y
ph?w
ad?
*h?y
h?
'day;
*pl?y
pfi
'fruit; egg; elf. for round objects'
'pestle'
sun'
that when PNB *-ay has tense voice, the Hr? reflexes remain *-ay,
the PNB *-ay was under breathy voice, the high vowel reflex occurs.
Note
but when
1. Although
these particular patterns appear quite clear, much reworking of
Smith (1972) is needed to fully make sense of the interaction of vowel
changes and phonation types.
The Origins
of Registers
Table 144: PNB
PNB
213
and Tones
*-ay reflexes under tense versus breathy voice

Hr?
(Smith,
modified)
*bray
*katayh
bray
katayh
'thread'
'hip'
*may
may
'sister
*bagay
magay
'people'
*adr?y
ad?
'pestle'
*h?y
hi
'day;
*pl?y
p?
b?h
*qb?yh
in-law;
elder
sister'
sun'
'fruit; egg; clL for round objects'

'snake'
The conjectures about the earlier sociolinguistic

situation are somewhat
was
a
some
combination of
shift accompanied by long
likely
speculative, but it
term bilingualism. At the earliest stages, Haroi would still have been a dialect of
the then prestigious Cham language, and thus some speakers of Hr? may have
shifted to Haroi. Since then, however, Cham has suffered a considerable loss of
prestige, and at some point the roles became reversed with the Haroi assimilating
to the Hr?.
In any case, two things are well-attested: Ethnographically,
the Haroi
have been heavily influenced by Bahnar^ in particular by the Hr?. Under these
influences, Haroi has undergone a set of changes typologically parallel to the
changes in Hr?. Specifically, during this period of social and linguistic contact,
the Haroi vowel system has been realigned, coming to resemble the Hr? vowel
system.
Conclusions
The Haroi case nicely illustrates the respective roles played by external language
contact and by internal paths of change: the external contact has provided both
the impetus and the directionality
for the changes, while the language internal
structures have provided constraints on the potential paths for the changes.
As a corollary to the primary role played by contact in this and other
Haroi changes, the major Haroi changes have not come about gradually. Rather,
the opposite is true: since the major changes in Haroi came about with the onset
of intense contact, the major changes are characterized by short periods of rapid,
assimilative restructuring, beginning with the onset of intensive contact and fol?
lowed by periods of relative stasis and more minor
changes?continuing
until the
214
Chapter
next significant period of contact. The non-contact
induced changes are far more
minimal.
restructurings, powered by contact situations and presumably

abrupt rather than gradual, are found throughout Southeast Asia (for example,
1992a, 1996). In fact, it would be reasonable to conclude that, in the
Thurgood
Similar
Southeast Asian
area, it is contact
that has provided the impetus for

with
factors only influencing the
changes
language-internal
paths by which such changes take place.
Finally, the most interesting part of this particular change may be that it
seems to have been a change in part shared both by Haroi and Hr?. Given both
most
linguistic
of the major
the timing of the changes and their striking similarities, it looks like the best way
to view the changes is as being one that was shared by the Haroi and the Hr?
Hr? speakers the changes occurred in their Hr?, but
speakers. For monolingual
speakers bilingual in Hr?, the changes occurred not only in their
Hr? but were also extended to their Haroi. Thus, in this sense, the change was
shared between the two languages.
for those Haroi
TSAT AND ITS FULLY-DEVELOPEDTONAL SYSTEM

Tsat is a Chamic
near
Sanya
language spoken in theMuslim
on Hainan
island.
Although
there
was
villages
of Yanglan
undoubtedly
and Huixin
a trading
commu?
to this area was apparently

nity established earlier, the first major migration
at
in 982.
the
the
northern
after
fall
of
shortly
Indrapura to the Vietnamese
capital
to that,
Subsequent
there may
also
have
been
other
immigrations.
language itself is Tsat /tsa:n?42/ (transcribed from a Keng-Fong

Pang tape), a form that corresponds with absolutely complete regularity to the
Cham name for themselves /cam/. The initial, the final, the vowel length, the
The
glottalization, and even the tone are precisely what is expected (see Table 104 and
text for a discussion of the tonal developments). Although
the
the accompanying
language is Tsat, the people are called Utsat, an ethnonym consisting of the root
/tsam?42/ 'Cham', just discussed, with what is apparently the prefix lu-l also seen
in other ethnonyms
in the language, an etymology
suggested by Mark Durie
(p.c.).
Evolution
of the Tsat
tones
Over fifty years ago, Paul Benedict (1941) recognized that Tsat is Chamic, but
the Tsat tones were only reported more recently (Ouyang and Zheng, 1983; Ni,
1988ab, 1990ab). The comparative work has either accompanied the synchronie
work or quickly followed
Ni, 1988ab, 1990ab).
it (cf. Benedict,
1984; Haudricourt,
1984; Zheng
1986;
The Origins
of Registers
215
and Tones
The tone system itself is similar in its complexity to the tone systems of
its Chinese neighbors and more complex than that of Phan Rang Cham, and the
diachronic study of its tonal developments have been a major or minor focus of
1984; Haudri
papers by various authors (Maddieson and Pang 1993; Benedict
court 1984; Ouyang and Zheng 1983; Zheng 1986; Ni 1988ab, 1990ab; Thur?
good 1992b, 1993).
Tsat is of particular linguistic interest because in it, each stage in the
and fully-tonal is remark?
transition from disyllabic and atonal to monosyllabic
ably well-documented.
Perhaps it is the clearest such case in the linguistics liter?
ature. Part of the clarity comes from the relatively short time-depth; the rest of the
clarity is simply our luck in having most of the intermediate stages attested in
related languages.
Figure
18: Tsat tonogenesis
Initials
inmonosyllables
Resulting
classes:
registers:
Resulting
tone
Tones:
PC initials
(except voiced
obstruents)
PC voiced
obstruents
modal voiced,
high series
breathy voiced,
low
classes:
Finals:
55
<
24
< glottal
33
< voiced
55
<
42
< glottal
series
11
*-h
*-h
< voiced
Tones are marked with Chao
[Zhao] (1930) tone numbers. The numbers

indicate relative pitch height, with 5 being high, 3 in the middle, and 1
low. The first number indicates where the tone begins the second where
the tone ends. Thus, for example, 55 is a high, level tone.
a little more complicated,

like the evolution of Phan Rang
Although
Cham tones, the evolution of tones in Tsat is relatively straightforward. The mod?
ern Tsat tones are predictable from the voicing differences in the earlier initials
216
Chapter
from a final *-h; the relative chronology of this

to
with
other
the
respect
changes is unclear, something indicated by the
change
dotted line in Figure 18. Otherwise,
the earliest stage involved a splitting of the
?
lexicon into two groups
words with high-pitched, probably clear-voiced pho?
nation and words with a low-pitched, breathy-voiced phonation (Figure 18).
and finals. The 55 tone evolves
Next, both of these groups were split further by the final consonant. All
a mid-rising 24s (-s
words with a final glottal stop developed a contour tone?
42s
indicates a stopped tone) tone from the high-pitched series and a mid-falling
tone from the breathy-voiced
low-pitched series. All words with a nasal or a
a mid-level
a
vowel final developed
level tone?
33 tone from the high-pitched
series and a low-level 11 tone from the breathy-voiced
low-pitched series. Sev?
eral other subsets developed
in special ways,
are also
but these developments
transparent.
From final
*-h & *-s: (> tone 55)
among the Chamic languages, Tsat has developed a unique tone from the
reflexes of PC final *-h (< PMP *-q) and final *-s: the 55 high-level tone. Statisti?
cally, most such forms derive from final *-h, but a significant number of 55 tones
Alone
also come from PC final *-s (for examples,
nipis
habis
PC
Chru
*lipih
*?abih
lapin
abih
145).
tone
Table 145: Origin of the 55 high-level

Malay
see Table
NR
Tsat
lupih
abih
pi55
'thin (material)'
phi55 phi55
'all; finished,
done'
tikus
mamah
taku:h
*tikus
*mamah
tukuh
mum?h
sa?ioh
tanah
*tanah
tanah
si?joh
tan?h
deras
*dras
drah
drah
buah
*boh
boh
boh
nanah
*lanah
tujuh
*tujuh
x*(si)?joh
m?rah
s?-puluh
*mahirah
*pluh
lan?h
tajuh
mariah
spluh
(na11) ku55
ma55
2iu55
'drip; a drop'
na 55
'earth,
sia-'a55
'fast;
short
pho55
'fruit;
egg;
la11 na55
sa pluh
soil'
time'
elf.'
'pus'
'seven'
tijuh
mariah
'rat'
'chew'
za ,55
plU
55
'red'
'ten'
For PC forms ending in either *-h or *-s (except the *-a:s rhyme), the Tsat reflex
is invariably the 55 tone regardless of whether the syllable-initial consonant is a
voiced
obstruent,
a voiceless
obstruent,
a sonorant,
or whatever.
The Origins
The shared Roglai
217
and Tones
of Registers
and Tsat loss of PC *-s in reflexes of PC *-a:s
However, there is one fascinating and instructive exception to the generalization

that PC final *-h and *-s led to tone 55 in Tsat: the PC *-a:s rhyme. As is particu?
larly clear from the N. Roglai column in Table 146, the final *-s was simply lost
after the four words
ending
in PC *-a:s.
Table 146: The shared Roglai

Malay
PC
Chru
and Tsat loss of PC *-s in reflexes of PC *-a:s

NR
Tsat
-as
*-a:s
-a:h
-a
-a33/11
atas
*?ata:s
ata:h
ata
ta33
kapas
'far;
above;
long'
'cotton'
x*kapa:s
kapa:h
kapa
pa33
*kaka:s
karka:h
kaka
ka33
'fish scales'
beras
*bra:s
bra:h
bra
phia11
'rice (husked)'
-ah
*-ah
-ah
-ah
-a55
darah
*darah
drah
darah
sia55
'blood'
panah
*panah
panah
pan?h
na55
x*picah
pa can
tsa55
'broken;
b?lah
*blah
blah
phia55
'chop; split'
basah
*basah
sa55
'wet;
pecan
pasah
blah
pasah
break'
damp'
As a consequence
in Tsat, these forms have tone 11 or tone 33 reflexes,
not the 55 tone expected from words ending in a final *-s. Note first that this is a
very restricted, very specific loss: only the *-s is lost and then, as Durie suggested
(p.c.), only after long /a:/. In other environments, the *-s was not lost; that is, in
other environments,
the PC *-s remains /-s/: in Northern Roglai, it remains /-s/
and it shows up in Tsat as the 55 tone. Nor does the otherwise quite parallel PC
*-ah loose its final *-h; in Northern Roglai, *-h is retained as such everywhere,
and inTsat the final *-h is reflected everywhere
of examples in Table 146).
Given the highly specific environment
as the 55 tone (see the second set
for this change, the fact that it is

shared by Northern Roglai and Tsat is strong evidence for subgrouping the two
together. When combined with the fact that Northern Roglai and Tsat also both
share the innovation of final preploded nasals, no doubt exists that these two lan?
guages should be subgrouped together.
218
Chapter
From final
stops
All PC final strops (*-p, *-t, *-k, *-c, *-?) have as their Tsat reflex a final glottal
stop. The transcriptions of both Ouyang and Zheng and of Ni give the misleading
impression that even these final glottal stops have disappeared in some forms, but
the instrumental
studies
inMaddieson
and Pang
(1993) make
it clear the final
glottal stops are still present.

These forms ending
tinction correlated with
in final glottal stops have split into two tones, a dis?

themanner of the initial consonant: monosyllables
begin?
an originally voiced obstruent have a 42 falling tone; the remaining

have a contrasting 24 rising tone. The importance of initial voic?
monosyllables
ing in the evolution of the 42 tones was suggested by Benedict (1984); this pre?
ning with
cise configuration of features as the origin of the 42 tone was pointed out to me
by Eric Oey (p.c., 1992).
tone
Table 147: Origin of the 42 falling

PC
Malay
Chru
NR
Tsat
xtabiat
ta?ia?
tubia?
hidup
*hudip
hadiu?
hadiu?
phia?42
thiu?42
'go out; appear'

flive, alive'
x*js?
se?42
'near;
*do:k
je?
do:?
je?'near'
duduk
do:?
tho?42
'sit;
jua?
sua?42
'step
no?42
'upgrade;
phua?42
'do; work'
buat
'do'
x*jua?
jawa?
x*rp?
go?
*buat
bu??
about
live;
on;
to'
stay'
tread'
east'
The remaining monosyllables

ending in a glottal stop but not beginning with an
originally voiced obstruent have a rising 24 tone.
Just as with the 42 falling tone, the final glottal stop is also retained with
the 24 rising tone. What
is not yet obvious from these forms but what will
is that it was the
become obvious when voice quality spreading is discussed
not
the voicing per se, that led
breathiness associated with the voiced obstruents,
to the low tone reflexes ("Transparency and phonation spreading" on page 183).
for the final stops, not the
retains a number of the
which
language Malay,
finals as such. In this respect, however, the Malay orthography is somewhat
deceptive as the orthographic final <k> actually represents a glottal stop inmost
cases. Note thatAcehnese
(not included on the table) sometimes retains the final
Table
least of which
stops
as
148 shows several pieces
of evidence
is the extra-Chamic
such.
The Origins
219
and Tones
of Registers
Table 148: Origin of the 24 rising tone
PC
Chru
NR
Tsat
*phit
phi:?
phi:?
stu?
satu?
phi?24
tu?24
'bitter; bile'
*satuk
anak
*?ana:k
ana:?
ana:?
na?24
'child'
pahat
*pha:t
pha:?
Be?
?e?
Malay
pahit
x*?e?
x*rfa?
?mpat
'boil; cook'
pha:?24
'chisel,
?e?24
'elf. long, thin

objects'
?da?24
'crack
hua?24
'eat
plane'
open'
x*hua?
hua?
x*?o:?
?o:?
?o:? mata
?o?24
'face'cf.'nose'
x*le?
Is?
le?
le?24
'fall into'
*pa:t
pa:?
pa:?
pa?24
ra?
za?24
'grass;
?u:?
?u?24
'head hair'
ako?
ko?24
'head'
'hot; sunny'
ra?
x*rak
-1
*?uk
?u:?
-1
ako?
x*?ako?
rice'
'four'
x*parfia?
padia?
padia?
tfia?4
*ga?
t)a?
fj??
nau?24
'make,
x*cat
ca?
ca?
weeds'
do'
tsa?24
'mountain
f?a?24
'oil'
range'
minyak
*mifta:k
masak
*tasa?
tasa?
tisa?
sa?24
'ripe;
cooked'
sakit
*sakit
-saki:?
saki:?
ki?24
'sick,
painful'
ni?4
ma?a:?
langit
*lagit
lagi:?
lan?:?
asap
*?asap
asa?
asa?
sau?24
'sky'
'smoke
x*kle?
kls?
tie?
ke?24-m
'steal'
ma?
ma?24
'take;
ka?24
'to tie'
za?24
'vein,
za?24
sa:?24
'wave; ripple; surf'

'bad; wicked'
si?24
'sew'
*mat
ma?
ikat
*?ikat
aka?
ika?
urat
*?urat
ara?
ura?
riak
jahat
jahit
From final
Aside
xriya:?
*jaha:t
*
jahit
nasals
ria:?
jaha:?
si:?
chi:?
fetch,
get'
tendon'
or vowels
from a handful
with forms ending
-1
(of a fire)'
on page 160, the developments

in final nasals or final vowels parallel the developments with
of final nasals discussed
220
Chapter
final glottal stop: monosyllables

beginning with
have all
low-level tone; monosyllables
without
initial have a contrasting
Table 149: Origins
an originally
an originally
voiced
obstruent
voiced
obstruent
tone.
33 mid-level
of the 11 low-level
tone
PC
Chru
NR
Tsat
abu
*habow
habau
habau
pha11
'ashes'
ada
*?ada
tha11
'have,
adik
*?adsy
thai11
Malay
adai
adai
exist'
*babuy
pabui
ba bui
phui11
'younger sibling'
'wild pig'
dada
*dada
tada
dada
tha11
'chest'
dua
*dua
dua
dua
thua11
'two'
hidung
*?idurj
adug
id?k
thug11
'nose'
hada:rj
huda:k
la^haig11
'shrimp; lobster'
haja:n
huja:t
sam11
'rain'
tarbau
tub au
pha11
habai
habuai(m)
phai11
babi
hudang
*huda:rj
hujan
t?bu
*huja:n
ubi
*tabow
*hubsy
-v
'sugarcane'
'taro;
tuber;
ular
*?ular
ala
ula
(la11) la33
fsnake'
lipan
*limpa:n
lapa:n
lupa:t
a11 pain11
'centipede'
kandung
x*kadun
kadug
thug11
'pocket;
ribu
*ribow
rabau
rub au
pha11
yam'
pouch'
'thousand'
tone is essentially the residual tone class: it

The remaining 33 mid-level
involves no voiced obstruent initials, and, thus, no tone lowering; it involves no
Tsat final glottal stop, and, thus, no contour tones; and it involves no final *-h or
tone. In short, the 33 mid-level
tone is from the
*-s, and, thus, no high-level
or
a
vowel final
voiced obstruent onset.
nasal final forms lacking
the forms discussed thus far, the forms ending in PC *-h
Tsat 55 high-level tone reflexes. The tones of the remaining
forms are conditioned by two factors: whether the onset syllable was a voiced
To summarize
or *-s have modern
and whether the PC final was originally a stop. The presence of a

voiced obstruent onset resulted in a 42 falling tone, if the form ended in a glottal
stop, and in the 11 low-level tone, if it did not. The presence of a final glottal stop
obstruent
resulted in a contour tone: if the form had a voiced obstruent onset, the tone was
the 42 falling tone; if the onset was not a voiced obstruent, the tone was a 24 ris?
ing
tone.
The Origins
Table 150: Origins
Chru
pinang
*pina:g
pana:g
satu
*sa
sa
tahun
*thun
tali
tone
of the 33 mid-level
PC
Malay
221
and Tones
of Registers
Tsat
NR
pin?g
sa
na:g33
'betel;
sa33
'one'
thun
thut
thun33
'year'
tala i
lai33
*talsy
talai
'rope;
string'
tang?n
*taga:n
ta gain
tag?n
ga:n33
'hand;
arm'
tuha
*tuha
tha
tuha
ha33 xau33
'old (people)'
turun
*trun
trun
trut
tsun
ular
*?ular
ala
ula
(la11) la33
lapar
33
'descend'
'snake'
33
*lapa
lapa
x*?assh
assh
aseh
x*par
par
pa
pa
'hungry'
,33
'horse'
33
tulang
palai
piai
ruai
zuai
*?ika:n
akain
ika:t
ka:n33
'fish'
x*ka:g
talka:g
ka:k
ka:g33
'chin; jaw'
tula:k
la:*33
'bone'
tig33 ta33
'eye'
*tula:g
mata
*mata
Transparency
occurs
also
33
-n
x*ruay
ikan
'to fly'
pan
?-v
*palsy
As
betel-nut'
plai
ruai
tala:g
mata
and phonation
in Haroi,
mata
'village'
33
'fly; bug; insect'
spreading
Western
and
Cham,
Phan
there was
Cham,
Rang
spread?
ing of the breathiness generated by a voiced obstruent in the pretonic syllable

the main syllable, thereby resulting in a lowered tone in the main syllable.
And, as in the other languages in this study, the main-syllable-initial
are
sonorants
transparent
to spreading.
Notice
that
the
tone
class
151 is predictable not from the initial of the PC main

the initial of the PC presyllable.
in Table
to
the examples
of
syllable but from
through sonorants
PC
Chru
*dihlow
lahau
dihlau
lau11
'first (go); formerly'
jarum
*jarum
jrum ?
jurup
sun11
'needle'
bunga
*buga
baga
bug?
ga11
'flower'
baharu
*bahrow
barhau
bahrau
phia11
'new;
bulan
*bula:n
ea bla:n
ia bila:t
phian11
'moon;
Malay
dahulu
jalan
*jala:n
*bara
jala:n
bra
NR
jala:t
bara
Tsat
lam11
phia11
'road;
just
now'
month'
path'
'shoulder'
222
Chapter
Tsat is unique among the languages examined in also allowing

voiceless obstruents. In the examples of
spreading through main-syllable-initial
Table 152, it is the voiced obstruent of the presyllable, not the voiceless obstruent
of the main syllable, that correlates with the Tsat 11 low-level and the 42 falling
However,
tone. Thus, in these examples, in which the PC presyllable begins with a voiced
obstruent and the main syllable begins with a voiceless obstruent, the second reg?
ister phonation, undoubtedly, breathiness, has spread from the presyllable to the
main
syllable.

Malay
PC
d?pa
dikit
Chru
*dapa
*dikit
*batsy
b?si
*bassy
*buta
buta
through voiceless
tapa
taki:?
patai
pasai
?
x*bato
pato
batu
*batow
patau
batuk
*batuk
those
patu?
stops
NR
Tsat
tupa
tiki:?
pa11
ki7*2
'armspan'
'few; little'
pitai
u11 tai11
'banana'
sai11
'iron'
pisai
?
ta11
'blind'
pato
to11
'teach'
patau
tau11
'stone'
pitu?
tu?42
'cough'
There seem to be, however, a small set of examples, virtually identical to

in Table 152, in which the PC presyllable
also begins with a voiced
obstruent and the main
syllable also begins with the voiceless obstruent *-h-, but

it clear that the breathiness did not spread from the presyl?
lable to the main syllable. That is, it initially appears that, in Tsat, breathiness
spread through medial voiceless stops but not through voiceless medial l-b-l.
where
the tone makes
These patterns were reported as a puzzle in Thurgood (1993), a problem

heightened by the contrast with Phan Rang Cham, where just the opposite hap?
pened: breathiness spread through medial l-b-l, but did not spread through medial
stops (see Table 153). The essence of the puzzle lay in the fact that in
the medial voiceless l-b-l was apparently more transparent to spreading
than the medial voiceless stops (as might be expected), but in Tsat precisely the
voiceless
Cham
opposite
seemed
The
true.
resolution
to
the
puzzle,
however,
turns
out
to be
anticlimactic:
as
the Northern Roglai column in Table 153 indicates, in forms with medial l-b-l the
voicing of the voiced obstruents was lost before the breakup of Northern Roglai
and, more to the point, long before the spreading of breathiness would have taken
a result, by the time spreading occurred,
no
question
longer had voicing or breathiness.
place
in Tsat. As
the presyllables
in
The Origins
Table 153: The apparent failure

Malay
jahat
jahit
dahi
akar
PC
223
and Tones
of Registers
to spread through medial
NR
Tsat
PR Cham
sa:?24
?ha?
'bad; wicked;
*jahit
chi:?
si?24
?hi?
'sew'
*?adh?y
thai
thai33
they
'forehead'
*?ugha:r
ukha
kha33
ukha
'root'
*jaha:t
-h
badly'
Despite the obvious typological similarities, the Tsat tonal developments

and the spreading patterns are idiosyncratic to Tsat, making it evident that the
are quite independent of those in Haroi,
Tsat tonal and spreading developments
Western Cham, and Phan Rang Cham.
final constriction
Tones and secondary
This analysis describes what happens to the vast majority of the voiced syllables.
However, two cases remain which involve interesting twists: the first involves the
tonal development of PMP *-ay syllables acquiring final glottal stops; the second
involves final nasals acquiring glottal constriction. The development of the glot?
talization of final nasals was discussed on page 160ff.; the development of the
final glottal stops from PMP *-ay will be discussed below.
Secondarily-derived
falling
tones: PMP
*-ay > Tsat -aa?
Tsat has developed an epenthetic final glottal stop and a 42 falling tone on inher?
ited PMP *-ay forms. Although the original sources show considerable variation
in the subphonemic pitch of the tone, as Maddieson
and Pang (1993) correctly
observe Tsat has only one phonemic falling tone, and thus all the examples in
Table 154 are marked as 42.
Table 154: Tsat reflexes of PMP

PMP
*ma-atay
(*qaqay)
Malay
mati
PC
Tsat
*matay
taii?42
'die'
*kakay
kaii?42
'foot'
mari
*maray
zaii?42
'come'
padi
*paday
tha??42
'rice (paddy)'
kaki
*haway
*qatay
*-ay finals
hati
vari?42
'rattan'
*hatay
ta:!?42
'liver'
*glay
khiad?42
'forest; jungle'
224
Chapter
falling tone correlates with two features of the modern Tsat reflexes: the
presence of a glottal stop and the presence of a modern Tsat long vowel. These
same two features can also be seen in the other source of secondarily-derived
falling tones, the glottalized final nasals, which discussed on page 160ff.
The
The history
of Tsat
contact
The
linguistic evidence reveals the outlines of Tsat contact. More revealing than
any other finding is that, as is evident from irrefutable linguistic evidence
(page 40), Tsat subgroups with Northern Roglai. Along with various minor
pieces of evidence, there are two major shared innovations: Tsat and Northern
loss of final *-s just in the rhyme *-a:s

Roglai share the contextually-restricted
but nowhere else and they share the very, very unusual, and typologically highly
marked innovation of preploded final nasals. Either of these would normally be
considered
significant subgrouping evidence; the combination

For those without sufficient linguistic sophistication
strength of the linguistic evidence, it is possible to be mislead
/tsa:n?42/, which, as is shown in "Tsat and its fully-developed
is irrefutable.
to appreciate the
by the name Tsat
tonal system" on
page 214, is the modern Tsat reflex of the name Cham. However, it is only in con?
temporary times that the term Cham picks out, for example, just the Phan Rang
Cham
and
the Western
Cham.
thousand
years
ago,
all
the Chamic
speakers
the coast of Vietnam
along
presumably fell under the designation Cham so the
name alone tells us nothing other than what we already know?that
the Tsat are
Chamic.
The connection
of the N. Roglai
and northern
Cham
The modern N. Roglai are, as the apparent etymology of the name Roglai sug?
gests, 'forest people' (*ra < *?ura:g 'people' + *glay 'forest'), but it is unlikely
that it was a forest-dwelling
group that broke off from the Roglai and fled to
Hainan. Instead, they must have been living somewhere else with a different life?
style and had not yet come to live in the highlands. Further, both Roglai and Tsat
share two marked sound changes?the
loss of final PC *-s after *a and stop final
reflexes from PC final nasals as the result of the decomposition
of final nasals
forming preploded nasals. It is also unlikely that the speakers of N. Roglai/Tsat
were in the south; rather it is far more likely, particularly in light of the existence
of the preploded nasals, that these speakers were at the end of the Chamic dialect
all other things considered, the northern end of the chain.
chain?presumably,
it
is
Thus,
reasonably likely that the forerunners of the modern N. Roglai and
modern Utsat were then simply the northern Chams living in and around Indra
The Origins
of Registers
and Tones
225
pura before the fall of the northern capital Indrapura in 982, some of them traders
and some from other social groups. And, it was only after the fall of the northern
capital Indrapura that these northern Chams split up, with the forerunners of the
modern
Tsat
modern Roglai
/tsam?42/ (< PC *cam) fleeing to Hainan, the forerunners of the

fleeing to southern Vietnam, and, speculatively, others fleeing to
Cambodia.
It is also reasonable
to assume
that the composition of the emigrant

groups was different depending upon the destination. It is likely that a dispropor?
tionate number of the traders went to either the trading colony on Hainan or the
new capital in the south, that the majority of themembers of the ruling class most
likely favored southern Vietnam, and that the rest of southern Vietnam and per?
haps Cambodia attracted those with other skills and backgrounds. In any event by
around 1000, the northern Chams had abandoned the northern provinces.
This scenario, suggested by the linguistic subgrouping and by the subse?
quent migration patterns, also gives us an approximate date for the Tsat arrival in
is, sometime not too long after 982 when Indrapura fell to theViet?
and while the Chams were in the process of abandoning their northern
provinces. This date would fit nicely with the Tsat traditions (Sch?fer, 1967:75)
which maintain,
in one version, that they came to Hainan from the "Western
Hainan?that
namese
in Tang times [618-907], and another version that says that they
Regions"
reached Hainan in the Song period [960-1279]. This would account for the preva?
lence of stories about the times of the kings among the Roglai and for the stories
about the Roglai possessing some of the Cham royal treasures.
traditional dates match quite nicely with the Chinese dynastic
which
records,
Zheng (1986:37, footnote 4), in a footnote in her article on the
genetic relationship of Tsat to the Chamic languages, briefly discusses. The trans?
lation has been provided by Hilary Chappell; her comments are in square brack?
The
ets, while Mark Durie has added two notes in curly brackets.
to the Records ofQiong Prefecture (Qiongzhou Fuzhi), their
According
forebears "came in boats with their families as a result of troubled times
in the Song dynasty [960-1279 AD] and the Yuan dynasty" [1279-1368
to extracts
AD]. ['Qiong' is another name for Hainan Island]. According
from both the History of the Song Dynasty (Song Shi) and True Records
of the Emperor Xian Zong of theMing Dynasty (Ming Xian Zong Shi), it
can now be precisely determined that they came in several migrations
from Zhancheng
{Champa} to Hainan Island. The History of the Song
v.
489, p. 14 080 (revised edition Zhonghua Publishers) records
dynasty,
the
third year in the period of Yong Xi (986 AD)) Danzhou
"(During
Hainan
from
[modern Danxian,
Island] reported that a person
226 Chapter
led one hundred of his clan to the county

their
allegiance], having been harassed by the people of
capital [to give
Jiaozhi [northern parts of modern Vietnam and the name given to Viet?
nam by the Northern Song dynasty]".
Zhancheng
named Pu-Luo-E
these immigrants are Chamic speakers from Champa.

Unquestionably,
Two years later there is another report of immigrants from Champa, but these
(Canton City) on mainland
immigrants appear to have settled in Guangzhou
China.
In the first year of Duan Gong (988 AD), Guangzhou

[Canton City]
one
a
to
that
of the Yi
"Huxuan,
person
belong
again reported
others
assembled
310
tribes [Yi = non-Chinese
and
'barbarian']
people
and came to the country capital from Zhancheng
[Champa]".
once
However, this 988 AD record is difficult to interpret. On the one hand, it is clear
that Zheng is citing it in the context of discussing the arrival of the Tsat (Chams)
in Hainan. On the other hand, it is quite clear that the report is from Guangzhou
(Canton City), not Hainan. The most straightforward reading, however, is that
this is a report of Chams arriving presumably in the Guangzhou
(Canton) area.
It should be pointed out thatYi is being used here as a general term indi?
cating a non-Chinese
Yi people, a Loloish
barbarian, not as a specific term indicating a member

branch of Tibeto-Burman.
subgroup of Lolo-Burmese
of the
Then, again, not long after the fall of the southern capital Vijaya, in
in 1471, there is a another report of the immigration of refugees from
Binh-dinh,
Champa. The True Records of the Emperor Xian Zong of theMing Dynasty also
records (v.284)
(In the 22nd year of Cheng Hua (1486 AD)), the eleventh month, the
Xu Tong'ai, and others presented a
Imperial Censor to Guangdong,
to the throne, reporting: 'The prince of Zhancheng, Gulai,
memorial
the
leading
imperial concubines, the prince's descendants and his clans
1000 people?and
carrying local produce [probably as a
people?over
came
to Yazhou [modern day Yaxian] in Guangdong province,
tribute]
to
wishing
appeal to the throne'.
the first two immigrations, since it had been roughly 500 years since the
Cham had controlled the northern provinces, these immigrants would not have
been from the north, but instead must have been from the south?and,
thus, were
Unlike
not the first wave of immigrants,
the ones that originally
split off from the North
of Registers
The Origins
ern Roglai. And,

likely destination.
again,
227
and Tones
be a
for the traders the trading post at Hainan would
these records, Zheng (1986:37) concludes that the Utsat have

for over 1000 years, a conclusion that accords very nicely with
the subgrouping record and with the consequent conclusion that the first wave of
Tsat must have immigrated to Hainan from the northern part of the Champa and
Citing
been in Hainan
thus historically
rather early. The Chinese records, however, also suggest
there was a second wave of immigrants roughly five hundred years later.
that
with the date of the arrival of Islam
Correlations
several authors have tried to

Presumably because the modern Utsat are Muslim,
connect the date for the arrival of the Utsat on Hainan to the date for the arrival of
Islam in Champa. Benedict (1941:130), for example, suggests that, although the
"date of the introduction of Mohammedanism
into Annam remains uncertain",
since the earliest Islamic inscriptions are dated 1039 and from 1025 to 1035, "the
arrival date was somewhere around the middle of the elev?
earliest date potential

enth
century".
Similarly, in her dissertation, Pang begins an appendix entitled "Com?

ment on Utsat's Chamic origins" (1992:431-436) with the statement that she will
comment on the dating of the Utsat colony in Hainan and on whether or not the
Utsat were already Muslims when they arrived inHainan. However, much of the
for her dating starts with the assumption
that the Utsat were
argumentation
Muslim
already
when
they
arrived
in Hainan,
a position
that
she
provides
no
evi?
dence for, possibly following Benedict (1941:10) or possibly following Maspero

(1928:13-14) who states that the widespread conversion of Cham people to Islam
in part as a conse?
only occurred after the fall of Champa in 1470. Apparently
of this assumption about the Utsat being Muslim
upon arrival, Pang
argues for a date very roughly 625 years ago. Certainly, if we assume that the
Tsat were Muslim when they arrived, the temptation might exist to argue for a
quence
later
arrival.
It is, however, simply unclear when the Utsat became Muslim. On the
one hand, in the literature, it has been argued by some scholars that at least a sig?
nificant part of the merchant class was Muslim at a very early date. On the other
hand, it seems clear that at least the majority of the general population that con?
verted to Islam did so much later. In addition, even if the earlier date of 986 is
accepted as the arrival date for the Utsat on Hainan, the arrival of Islam might be
correlated,
including
not with
this early date, but with one of the later exoduses to Hainan,
those after the fall of Champa in 1470. Given the mixed evidence, it is
inappropriate to base any conclusions

Utsat to Islam.
on putative dates for the conversion
of the
228
Chapter
For example, a careful reading of Aymonier

1891, Ravaisse 1922, Mas?
1979
(translation 1985), and Chambert-Loir
pero 1928, Sch?fer 1967, Manguin
(1994:96), and Reid (1993, 1995) makes clear there is very early evidence of
Islam among the trading community and there is every reason to suspect that it
was disproportionately
the merchant community who went to Hainan. So if the
Tsat were partly Muslim when they left for Hainan, this is not surprising, nor
it call for a date as late as "the middle of the eleventh century". And, if the
it is not surprising that they became so on Hainan, as
Tsat were not yet Muslim,
we know that there was already aMuslim trading community on Hainan.
As for the significance of the Kufic inscriptions dated 1025-1035 and
would
In
1039, Ravaisse (1922), who reported them himself, drew limited conclusions.
two
noth?
"The
these
steles
indicates
of
the words of Manguin
presence
(1985:1):
ing more than the existence in the eleventh century of a community of Muslim
it need be added, Cham traders
merchants,
strangers to the country." However,
were undoubtedly part of the trading community as well, including quite proba?
adds, providing some perspective,
bly some who were Muslim. Then Manguin
"Nothing could be more normal than this settlement, when it is recalled thatMus?
had resided in such great number in South China since the eight
The
century.
trading station in Champa was only one of the links in a long chain,
which connected theMiddle East with Africa and China."
influence dates from a
Thus, while it is true that the bulk of theMuslim
lim merchants
influence in the trading communities dates from consider?

later time, theMuslim
earlier.
Various
contemporary records indicate that Islam existed among the
ably
merchants as early as 958, and probably earlier. Sch?fer (1967:75), for instance,
mentions
that the Chinese records show that in 958 the ambassador of the Cham
Indravarman III presented to the ruler of Chou "rose water, flasks of
Greek firs ("fierve fire oil"), and precious stones, was named (according to the
Abu Hasan". Three years later, Abu
Chinese P'u Ho-San (Pyu Ha-san)?plainly
was
Hasan
again sent to China by Jaya Indravarman I with "a letter on palmyra
monarch
leaves in an envelope of fragrant wood, and an offering of ivory, camphor, pea?

cocks, and twenty Tajik [Arab] vases' for the new Sung emperor."
Other early accounts also confirm the early existence of Islam, although
classes. As Sch?fer
this may have been largely restricted to the merchant
(1967:75) writes:
was right when, in 1891, he made the rash
statement that there were Muslim Chams as early as the ninth or tenth
century. ... This hypothesis, otherwise startling and incredible, would
[907-960]
explain the statement in the History of the Five Dynasties
It seems tome
thatAymonier
Wu tai shi, 74, 448od.] that the customs of the Chams

those of the Arabs" [lit. Tajik].,
are "the same as
of Registers
The Origins
229
and Tones
Thus, ifmerchants provided a significant portion of the immigration to Hainan,

then an Islam influence may have existed early enough to predate the earliest
probable migration.
But, if itwas not primarily amerchant class that fled toHainan, the dates
Islamization of Champa are probably too controversial to be use?
for widespread
ful for dating the migration. On the one hand, Coed?s concluded (1968:239):
"We have, however, no real proof that Islam had penetrated into Champa before
the Chams were driven out of Vijava in 1471." Henri Chambert-Loir
(1994:96)
provides an even later date for the Islamicization of the Cham court and part of
the end of 17th century. For instance, the Cham king, as
the population?around
Reid (1993:154) writes, citing Manguin, became Muslim only sometime between
1607 and 1676.
On the other hand, not all scholars agree with this position. Some have
that Islam had amuch earlier foothold in Champa, and not just among
the merchant class. Thus, after citing approvingly Aymonier's claim that Islam
concluded
was established
early in Champa,
Sch?fer
(1967:75) writes:
to the commercial ports

traders congregated is shown by a statement in the History
of Sung. After noting that among the animals used by the Chams, aside
from the water buffalo and the yellow ox (both familiar to the Chinese),
That
the influence of Islam was not restricted
where Arab
there was the 'mountain ox' (a banting or a kouprey?) which they did
not hitch to the plow but killed as a sacrifice to the ghosts. When about
to
kill
it,
their
'shamans'
invoked
it
thus:
"A-la-ghwa-gyep-b?t"",
can be translated (says our source) as "May it soon live in another

body!" Despite the pious Buddhist hope for the reincarnation of even a
lowly sacrificial victim, claimed as the meaning of this ritual formula, it
which
can hardly be other than butchered Arabic:
"Allah hu akbar!"?"God
is
great!"
In short, even if the dates for the arrival of Islam were known, it is not
clear that these dates would have any direct bearing on the question of when the
Tsat arrived in Hainan. However, the linguistic subgrouping evidence unequivo?
cally places the Tsat and the modern Roglai together as part of what could only
have been a northern Cham group, which dissolved into various subgroups with
the fall of Indrapura in 982. This scenario accords nicely with the arrival of a
group in 986, one or both of which could have been the Tsat. Thus, the
Chamic
linguistic and the historical evidence substantiate Zheng's conclusion

that the Tsat arrived inHainan roughly a thousand years ago.
(1986:37)
230
Chapter
Tsat contact on Hainan

the Utsat
arrival in Hainan, Tsat was strongly influenced apparently by

the
Min dialect that functions as the local lingua franca, but
Southern
Hainanese,
to a strong communal identity and sense of lan?
nonetheless, due undoubtedly
guage loyalty due at least partially to their being set apart by their Islamic beliefs
After
not
(Keng-Fong Pang, p.c.), the Utsat have retained their language. Nonetheless,
only is the lexicon rampant with Chinese loans, but the Tsat language has con?
verged dramatically with the language structures found in the languages of
Hainan.
arriving on Hainan, Tsat has gone from atonal to fully tonal, a

convergence with the surrounding languages.
striking example of phonological
Insofar as the contemporary situation is representative of prior history, the obvi?
Since
ous mechanism
for this change is bilingualism,

rather than shift. Not only are
many of the Utsat either partially or fully bilingual in the fully-tonal Hainanese,
but because the modern Tsat form a rather tight Muslim community, it is difficult
tomarry into it.Thus, the language shift that does occur is from Tsat
rather than the other way around.
for outsiders
to Hainanese
In earlier times, bilingualism and shift may also have contributed to its
that when the men arrived on Hainan
restructuring. One Tsat tradition maintains
they
took
women
wives
would
generation
tinct
What
from
among
the Li,
who
are
of
speakers
a tonal
The
language.
then have shifted
very well might
version
of
does
seem
the
tradition,
to Tsat, while the children of at least the first

have grown up bilingual in Tsat and Li. A quite dis?
however,
from
apparent
the
has
the men
various
stories,
and women
however,
arriving
is
that
together.
there
was
some early contact with the Li.

Precisely what sort of tone system did the Tsat come into contact with?
it is reported that the Tsat are bilingual in Hainanese,
it is not clear
Although
exactly
what
this means.
As
some
readers
are
aware,
the
term Hainanese
is often
used as a vague cover term for all the Southern Min dialects found on Hainan
island. Probably, the dialect in question is the Tan-chou dialect described by Ting
(1980), but it really does not make much difference. All the languages of Hainan
that Tsat speakers could have come into contact with have well-developed,
typo?
logically very similar tonal systems, any one of which would have provided the
typological model needed for the restructuring of the Tsat phonological
system.
In Table
155, the tone system of Tsat is compared with the tone system
dialect (Ting 1980), as it is probably the Tan-chou dialect of
the Tsat are bilingual in, and with two Li dialects (that is, the Hlai dia?
of the Tan-chou
Hainanese
lects), as the literature suggests early contact between the Tsat and the Li (Ouy?
ang and Zheng, 1980; cf. also Thurgood 1991 for a look at the reconstructed Hlai
tone
system).
of Registers
The Origins
231
and Tones
typological approximation of these languages to each other and to

Tsat is striking. All the languages have three level tones: Tsat has a high-level 55,
a mid-level
33, and a low-level 11; the Tan-chou dialect has a high-level 55, a
The
22, and a low-level 11; and, both Li dialects have three level tones. All
the languages have a rising tone, and all the languages except Tan-chou have a
falling tone.
mid-level
such as they are, are minor. Although

The differences,
the
the
and
tones,
falling tones, end in glottal stops, only
rising
lect of Li has final stops (indicated by the final -s) in the contour
the Tan-chou nor the Yu?nm?n dialect of Li are contour tones
the Tsat contour

the T?ngshi
dia?
tones; in neither
associated
with
final stops (although even in Tsat the final glottal stops for the rising and falling
were not always transcribed).
Table 155: The tones of Tsat, Tan-chou, and two Li languages
Austronesian:
Chamic
Chinese:
Tai-Kadai:
Hlai (LO
Southern
dialect
(Hainan)
Tsat
high level
falling tone
mid
level
rising tone
low level
55
(Hainan)
Tan-chou
55
42s
(Hainan)
T?nqsh?
Yu?nm?n
55
55s
43s
42
44
33
22
33
24s
35
13s
13
li?
li
11
11
Bear in mind, it is not being argued that the Tsat tonal system is bor?
rowed from one of these languages. Rather, the claim is that the patterns of Tsat
bilingualism with these languages influenced the internal paths of change within
Tsat itself, ultimately leading to the development of a typologically similar sys?
tem. Thus, an exact match is not likely, and would, in fact, be rather surprising.
at in this way, the interaction between the internal influences and the
external factors should typically result in a system distinct from the earlier Tsat
system as well as different from the system found in the contact language.
Looked
232
Chapter
Under extended contact with speakers of the tonal languages of Hainan,

Tsat speakers have radically restructured their phonological
system, innovating a
tonal system through the exploitation of internal paths of change that is extremely
similar to the tone systems of the other languages of Hainan.
THE INTERNALPATHS OF CHANGE

The Phan Rang data and the Tsat data show the chain of developments
that it is valuable to point out that this chain, leading to tonogenesis,
so clearly
may be a
if not all tonogenesis?as

from tone splitting.
distinguished
it is not all that obvious from the literature, in Southeast Asia true tono?
Although
genesis always seems to be mediated
through prior or at least coexistent phona?
There are several arguments
tion distinctions.
for this position. Voicing
model
for most
alone do not seem to be sufficient
for tonogenesis. Further, it is not

voicing per se that leads to tonogenesis, as vowels are voiced. Finally, it is fre?
quently not all the voiced consonants that develop the low tones, but rather just
the voiced obstruents, which are associated time and time again with breathy
voice, while the sonorants are ignored, which if anything are more voiced than
distinctions
the voiced
development
The origins
obstruents. Again,
of
low
tones,
of register
but
it is not the voicing

rather
complexes
some
other
per se that is crucial
quality
of
the voiced
in the
obstruents.
and tones
Gaps exist in our knowledge and will continue to do so for some time, but we
now know the basic stages in the development of Cham register, and its subse?
in theWestern Cham register system, into the incipient tones
quent development
of Phan Rang Cham, and into the restructured register system of Haroi. We also
of the full-blown
the stages in the essentially
independent development
tone system of Tsat on Hainan island.
know
Although Western Cham, Phan Rang Cham, and Haroi appear to have
developed a common register system before dividing into separate linguistic enti?
ties, the subsequent development were largely independent of one another, aside
from some obvious
later contact between the groups. As the prior discussion (and

156) makes clear, the developmental
paths often parallel one another,
common
to
the
of
convergence
type
reflecting
languages which share an identical
were
to
which
considerable
bilingual contact, and
starting point,
probably subject
Table
are guided by what are assumed to be universal phonetic tendencies. Fur?

ther, with very little exception, the internal paths of change are themselves pho?
which
netically plausible.
The Origins
of Registers
233
and Tones
initial stage in all cases was to develop a breathy-voiced

register
complex after the voiced obstruents, with Western Cham, Phan Rang Cham, and
Haroi probably evolving the same system before breaking up into dialects and
The
with Tsat doing the same thing but quite independently of the developments else?
where. Once breathy voice has developed after one set of initials, by default the
voice quality after the other consonants constitutes a contrasting register, and,
thus, we have register contrasts. Even from this point, although the developments
in individual languages may differ in detail, the changes remain similar typologi?
cally.
Cham developments
involved the fewest adjustments to the
one
The
already existing register system.
peculiarly Western Cham modification
was the extension of second register to a subset of the sonorant initial forms, an
The Western
not yet fully explicable.

Phan Rang Cham, of course, initially developed along identical lines
with Western Cham, with the differences coming after Phan Rang and Western
extension
split. The major difference between the two was that in Phan Rang Cham
the pitch differences came to the foreground with a corresponding backgrounding
of the phonation distinctions. The second register developed into the low tones
and the first register the high tones, with both tones being split again into those
forms ending in glottal stop .and those not.
Cham
The initial developments of the Haroi register also involve modifications

shared register system. However, unlike Western Cham or Phan
of the earlier
the non-breathy high vowels were further modified apparently by a

unique phonation associated with the voiceless stops and affricates, presumably a
tense phonation that has left its own special class of vowel reflexes in modern
Rang Cham,
Haroi. Elsewhere,
were
essentially
except
for some rather odd vowel
assimilation,
the vowels
unaffected.
register system was probably at this point only marginally

in
from the system inWestern Cham. Aside from differences
This Haroi
distinguishable
details and the possibility
it might have had more striking contrasts in vowel
Western
than
Cham, the two systems were very much alike typologically.
quality
most
And,
importantly, in both register systems, the vowel quality differences
were still fully predictable from the accompanying phonation distinctions. The
restructuring came with the loss of the phonation differences:

the voice quality differences disappeared, some of the accompanying vowel
quality differences were left behind, adding new phonemically distinct vowels to
dramatic Haroi
when
inventory and making Haroi into a restructured register system.

to Northern Roglai
and the Tsat
Tsat is closely-related
Although
in
the
the
other
three
changes
languages
changes occurred quite independently of
after the Tsat reached Hainan, the Tsat changes also tend to parallel the changes
the vowel
234
Chapter
in the other three languages. Tsat initially developed a two-way register contrast
a breathy register and, by default, a modal (clear) register. Thus,
distinguishing
the initial stage was the development of a register system.
Then, in the case of Tsat, the phonation distinctions were backgrounded
and the pitch distinctions were foregrounded,
of
leading to the development
tones. In a development unique to Tsat, a special tone class evolved from words
ending in PC *-s and *-h. Among the remaining forms, the former breathy regis?
ter developed into the 11 low tone in forms without a final glottal stop and into
the 42 falling tone in those forms with a final glottal stop; similarly, the modal
33 tone in words without a final glottal stop
register developed into the mid-level
and into the 24 rising tone in those words with a final glottal stop. Thus, aside
from some oddities involving glottalized final nasals as well as some epenthetic
final stops,
the Tsat developments,

despite the historical
followed
Rang Cham,
largely parallel lines of change.
Table 156: The paths of development
Haroi
Cham
restructured
register
voiceless
fully
tonal
tonal
33; 24s
vowel reflexes
high
tones
tones
(initial layer)
(default)
(default)
(default;
high
33; 24s
first register
(default)
also
some
occurs
unaffected)
after
(default;
unaffected)
sonorants
(second
voiced
incipiently
< first register
second register
obstruents
Tsat
(default)
obstruents
sonorants
Phan Rang
Cham
first register
obstruents
glottalized
from Phan
inmonosyllables
Western
register
separation
tones
tones
(default)
(default)
high
33; 24s
tones
tones
(default)
(default)
low
11,42s
layer)
second register
(initial layer)
vowel reflexes
<
second
tones
tones
register
of Registers
The Origins
235
and Tones
all the developmental

similarities, the resultant modern phono?
are
distinct
logical systems
radically
typologically. These striking differences in
outcomes have their origins inmodest adjustments made to the configurations of
features that comprise the register complexes. As Eug?nie Henderson noted some
Despite
second register is composed of a complex

thirty years ago, the breathy-voiced
cluster of features: a characteristically
lax or breathy voice quality, accompanied
by relatively lower pitch, but relatively higher vowel quality, while the contrast?
ing first register typically has tense, or modal phonation, accompanied by a rela?
tively higher pitch but relatively lower vowel quality. For reasons related to
external contact, not to internal pressures, Western Cham has essentially main?
tained a modified version of the earlier register system, while the pitch compo?
nent has risen to prominence in Phan Rang Cham and Tsat, and the vowel quality
differences have been focused on inHaroi.
Spreading
patterns
syllables are often not the reflection of the initial of themain

are
but
instead
the result of phonation spreading from the presyllable to
syllable
the main syllable. In this sort of spreading, there are two interacting patterns: one
is for certain classes of main-syllable
initials to be more transparent than others
The reflexes of main
to the spread of phonation, and the other is for the breathy-voiced

to be the most likely to spread (Table 157).
Table 157: The paths of breathy voice spreading
through
sonorants
in disyllables
through
*s,
second register
*h
through
voiceless stops
yes
yes
no
Phan Rang Cham
yes
yes
no
Haroi
yes
not clear
no
Tsat
yes
yes (*s)
yes
Western
Cham
Table 157 summarizes the spreading of the breathy-voiced

second regis?
ter from the presyllable
to the main syllable. Note that in all four languages
breathy voice has spread through the sonorants, resulting in the main syllable
having second register, not first register, reflexes. Notice thatmedial *s and *h are
almost as permeable,
with Western
Cham
and Phan Rang Cham
showing
clear
236
Chapter
evidence of spreading, with the Haroi data unclear, and with Tsat showing evi?
dence of spreading through *s. However, spreading of the breathy voice through
a medial voiceless stop is limited to Tsat.
The Chamic data tells us little about the spreading of tense voice, prima?
in every language except Haroi the vowel reflexes found after the
because
rily
voiceless obstruents and the vowel reflexes after the sonorants are identical?that
is, the same voice quality is found after both the voiceless obstruents and the
sonorants. Thus, the Chamic data on the spreading of tense voice is restricted to
one instance in Haroi, where tense voice (but only from voiceless stops and affri?
cates but not from *s and *h) spreads through medial sonorants lowering high
vowels
to mid
vowels.
PC Morphology:
Some Notes
Even a brief sketch of Chamic morphology

lies beyond the scope of this work, let
alone any sort of thorough survey. However, it is possible to discuss certain mor?
structures found in Chamic, particularly those structures that are
phological
to PC, with a special emphasis both on the MK influence on
Chamic morphology
and on the importance of the Acehnese data to the recon?
struction of PC morphology.
reconstructable
of aMK
The existence
from the beginning
core
of Chamic
early,
element
in Chamic
has been recognized
of the study of Cham. Crawfurd

terming
it the "Malay
of Champa",
(1822) recognized
but
subsequent
almost
the An
scholars
not only recognized theMK element inChamic but also were often mislead by it,
sometimes going as far as tomisclassify
it as aMK language. For Etienne Aymo?
nier (1889:5-6), who recognized both its MK and An elements, Chamic was
some kind of intermediary link between Khmer and Malay. Schmidt (1906) also
recognized the existence of both a MK element and an An element when he
described the Chamic languages as "Austroasiatische Mischsprache".
Thomas
Sebeok (1942) was led even further astray by theMK, misclassifying
the Chamic
as Austroasiatic. None of the investigators seemed to have trouble recognizing a
MK
element
Even
in Chamic.
It was only its interpretation that was controversial.

seems to have been
of a MK element in Acehnese
the existence
recognized early and with little enduring debate about its existence. Niemann
(1891), Cowan (1933, 1948, 1974, 1981, 1982, 1983, 1988, n.d.), Shorto (1975),
and Durie (1990a) all have not only recognized the existence of aMK element in
Acehnese
but also the last three have further recognized that, for the most part, it
is essentially the same MK element found in the mainland MK languages. The
237
238
Chapter
that exists again revolves around determining
controversy
the origins of this MK
element.
The debates about the genetic affiliations of Chamic have long since
laid to rest, but for those interested in a summary of the positions various
scholars have held, see Lafont (1987b; 1994b). However, all modern scholars
recognize the An character of Chamic, with the questions revolving not around
the classification
of Chamic but around how these An languages acquired the
been
MK
and the typologically MK-like

Both Cowan and Shorto
forms and systems which they now have.

that the MK features
rejected the possibility
been borrowed from MK?the

position argued for in this work?and
instead argue that in one sense or another the similarities are vestiges of a much,
much deeper genetic relationship between MK and An. In both cases, the crux of
their argument is essentially "belief-based";
that is, both scholars are convinced
that certain elements are almost in principle unborrowable! In particular, neither
have
length nor the shared grammatical morphemes
the vowel
are viewed
as borrow
able.
The assertion
that certain elements
are in principle
unborrowable
has
certainly been rejected by the preponderance of modern scholarship, leaving as

the real question to be answered whether the borrowing hypothesis is plausible in
this particular case?with
the borrowing hypothesis the only position supported
by much
of the evidence
in this work.1
interesting is the extent of morphological

borrowing. Shorto
Especially
one
notes
of
the
influences
of
MK
when
he
(1975:82)
suggests that the loss of
suffixes is an areal feature shared with the neighboring MK languages, a position
confirmed
by
an examination
of
these
languages.
A NOTE ON PC MORPHOLOGY
of the borrowing of MK material into An post-dates the breakup of PC.
Though especially notable in the highlands Chamic languages, where the contact
Much
with MK
has been more
intense and for a longer duration, post-PC borrowing is

the mainland Chamic languages. However, a small but signifi?
found throughout
seems to be reconstructable to PC. Both Cowan
cant amount of MK morphology
and
Shorto
(1975:81) recognize verbal and nominal morphology
(1948:432-431)
and the mainland Chamic lan?
of considerable antiquity found in both Acehnese
guages,
prefix
1.
specifically
*t9-,
a verbal
a causative
prefix
*ma-,
verbal prefix *pa-, a reflexive/reciprocal

and
nominalizing
instrumental
verbal
infix
*-an
The possibility of a much, much earlier genetic relationship between Austro?

or possibly with just theMon-Khmer
nesian and Austroasiatic,
component of
Austroasiatic,
notwithstanding.
239
Some Notes
PC Morphology:
are
in both Acehnese
and the mainland
languages (the reconstructions
those of this work). While the causative verbal prefix *pa- is native apparently to
both MK and PAn, the nominalizing
infix *-an- is definitely of MK origin.
In addition to these, there is a negative imperative *bs?, of obvious MK
infix *-am-, just as obviously of An origin, as well as
origin, and a nominalizing
found
a number of nonaffixial
grammatical morphemes
tions, and other connectives, which are sometimes
including pronouns, preposi?

from MK and sometimes from
An.
The borrowing did not all go one way, of course. Bahnar, a MK lan?
shows
the heavy borrowing of An lexical material as does Katu (Diffloth,
guage,
p.c.) and various other MK languages of the region.
Verbs
and verbal
morphology
The two inherited An prefixes discussed here, the *ta- and the *ma-, are not as
semantically transparent. Both are best retained inAcehnese, whose speakers not
only left the mainland quite some time ago?thus
leaving much of theMK influ?
ence behind?and
have since been in contact with speakers of other Malayic
lan?
guages?thus
helping
preserve
these prefixes, which Adelaar
reconstructs
for
proto-Malayic.
A glance at the lexicon makes it clear that not only were verbs borrowed
from MK but also so was some of the verbal morphology. The instrumental infix
*-an- is not just MK but may have been at least quasi-productive,
if one can
from
the
modern
judge
languages. Similarly, the negative imperative *bs? is
from MK. In both cases, the meaning
unquestionably
cal element is relatively transparent.
The
'inadvertent'
of the borrowed grammati?
prefix
The
'inadvertent' prefix, to use Durie's term, is reconstructed for PC as *ta-, with

the possibility that future research might allow the eventual addition of a final -r
to the prefix. Aymonier and Cabaton (1906:xxiii-xxiv)
describe it as having a
range of meanings for Cham including causativity, reciprocity, and stativity. This
prefix is reconstructed not just back to the proto-Malayo-Chamic

stage but to
PAn. Adelaar
reconstructs
as
it
*tAr- for proto-Malayic
but with the
(1992)
'inadvertent' including potentiality, non-controlledness,
and excessivity
meaning
(Adelaar, p.c.). Blust (p.c.) reconstructs it to PMP as *taR 'inadvertent; marker
of
involuntary
laar's PMalayic
and
action'
and
for PAn
as
*tar-
'inadvertent'.
The
match
with
Ade
and Blust's PMP and PAn looks perfect, both in terms of form
function.
240
Chapter
Table 158: The PC *td- 'inadvertent'

PMalayic
*tAr-
'inadvertent'
*ta-PC
'inadvertent'
tin-
'inadvertent'
Jarai
ta-
'reflexive'
Bahnar
ta-
Acehnese
'passive;
reciprocal;
causative'
tar-
Pacoh
'reciprocal'
t-
ta-
state'
'resultant
'involuntary'
look to be variants of *ta- seem to occur not just in the Chamic

the Bah?
of Vietnam but also in neighboring MK languages. Within
what
However,
languages
naric branch of MK,
for instance, Bahnar has a form Banker (1964) glosses as

causative'. Although Banker chosen to treat it as several
'passive; reciprocal;
homophonous prefixes rather than as a single unit, it looks to be the same element
as
historically. The *ta- also seems to be found in Chrau (Thomas 1971:152-153)
ta- 'causative, resultative'. And, within the Katuic branch, Pacoh (Watson 1966)
look as if they might
lists three prefixes which

ants
of
single
prefix:
tar-
'reciprocal',
t-
be historically
'resultant
state',
and
conditioned
ta-
vari?
'involuntary'.
The apparent relatedness of the MK to the PC forms is of considerable

and PMP *taR
interest. The PC prefix reconstructs to PAn *tar- 'inadvertent'
'inadvertent;
marker
of
involuntary
structs back not just to proto-Malayic

the Jarai form are also
connected,
action'
(Blust,
p.c.).
Thus,
the
form
recon?
but to PAn; certainly the Acehnese

and, if
inherited from PAn. How then does one
account for theMK
forms, that is, the Bahnar (MK: Bahnaric), Chrau (MK: Bah?
and
Pacoh
naric),
(MK: Katuic)? The phonetics and the semantics are similar. If
not simply lookalikes, which seems doubtful, are these forms the result of bor?
rowing from Chamic? If the forms are borrowed, the borrowings must have been
and Katuic, two distinct subgroups of MK. If the forms are
the
borrowed,
borrowing would not be too surprising as the evidence of intense
contact between Chamic and MK. Of particular interest are the Pacoh forms as
into both Bahnaric
they indicate intense contact between ancient Chams and the Katuic. Notice that
even this more northerly contact falls within the boundaries of Champa influence,
at least before the Vietnamese
began pushing southward. Other evidence of a
Chamic
contact with Katuic

into Acehnese,
particularly
Katu (Reid 1994), which
include apparently Katuic borrowings into Chamic,

strata in
and an apparent Austronesian morphological
one would presume were due to Chamic influences.
241
Some Notes
PC Morphology:
The obvious way to account for the Katuic strata found in Chamic is to
that Chamic influence extended up along the coast into Katuic territory.
assume
Certainly, an examination
there are a considerable
of the appendix of forms makes it abundantly clear that

number of MK forms, attested in the more northerly
Katuic but not in the more southerly Bahnaric. Further, many of these are attested
Thus, the most likely scenario is to assume that theAcehnese are the
first group
of the most northerly group of Chamic speakers?the
to
forced
leave themainland. This scenario is consistent with other evidence, spe?
cifically, with the fact that Acehnese
patterns neither with the remaining main?
inAcehnese.
descendants
languages nor the remaining highlands Chamic languages.

the
division of the original dialect chain into coastal and high?
Why
lands languages only became strongly marked after the Vietnamese
push to the
south had split up the chain?a division that occurred after the Acehnese had left
land coastal Chamic
not? Because
the mainland.
strata in
The other way to account for the Austronesian morphological
Katu is to invoke Austric, that is, to conclude that this strata is inherited from a
and Austronesian
(cf. Reid 1994). How?
proto-Austric consisting of Austroasiatic
ever, early contact between Chamic and Katuic speakers needs to be recognized
anyway to account for the presence of Katuic borrowings in Chamic and Chamic
borrowings in various Katuic languages (see Appendix II) so the opportunity for
the borrowing to take place has already been established. Further, the parallels
between the Cham prefix described by Aymonier and Cabaton (1906:xxiii-xxiv)
as expressing
described
causativity,
by Watson
as
reciprocity, and stativity and the Pacoh pr?fixai variants

expressing
'reciprocal'
(tar-),
'resultant
state'
(t-),
and
'involuntary' (ta-) are too close to be accounted for by mutual inheritance from
proto-Austric. The notion of mutual inheritance would require that, on the one
hand, the Katuic Pacoh inherited the prefix from proto-Austric by way of proto
Austroasiatic
with its phonetics and semantics largely unaltered, while on the
other hand the PC language Cham inherited the prefix also from proto-Austric
but by way of Austronesian,
again with the phonetics and semantics largely unal?
tered. Given that the time depth for PAn exceeds 6,000 years this is implausible.
The
*ma-
marker
The *ma- marker
(< proto-Malayo-Chamic
*man-) is a fully productive part of
of theMalayic branch of Malayo-Chamic
(cf. its central place in
the verbal morphology
of modern Malay).
themorphology
Aymonier and Cabaton (1906:364) describe the prefix md- in Cham as a

causative prefix, indicating a state, possession,
reciprocity, or mutuality, further
occurs
as
ma-.
also
that
it
in
Bahnar
noting
They specifically equate it with
242
Chapter
Malay m?-Zm?ng. An examination of the ma- section of their dictionary makes it

clear that the prefix is essentially the same prefix as inMalay and that its use,
although much diminished inmodern Cham, was once far more productive.
Table 159: The PC
Malay
rneng-
*/na- 'verb prefix1
PC
Acehnese
*ma-
mu?-
Jarai
Bahnar
ma
It clearly reconstructs to PC, as it exists inAcehnese,

Jarai, and Written Cham.
its occurrence in the modern mainland Chamic languages is now far
However,
more limited than it once was in PC.
The
'causative'
prefix
The PC
'causative' prefix *pa- is found throughout the Chamic languages except

for Tsat, where it got lost as part of the loss of almost all initial syllables. The pre?
on northern
fix is not just widespread
in Chamic languages including Acehnese
Sumatra but appears to either be productive or to have been productive in those
languages for which we have sufficient evidence and it occurs just as freely with
as MK
roots, except again for Tsat where sound change has led to the drop?
it
of
initial
ping
syllables. In addition to its appearance as part of the morphology,
also occurs widely in fossilized form as the initial syllable of a number of lexical
An
items in the appendixes.

Table 160: The PC */)5- 'causative7
PC
Malay
*pa-
early as 1906, Aymonier

as
borrowed
pa-
pui-
*pa-
As
Jarai
Acehnese
from
just inMK languages

tant branches of MK.
and Cabaton
Khmer,
contiguous
It certainly
a not
Bahnar
pa
(1906:xxiii)
unreasonable
described
assumption
as
the causative
it occurs
not
to the Chamic
occurs
languages but also in quite dis?

in the MK languages of Vietnam. For
in Bahnar, not just with the same form but also the same
It also must occur fossilized in Chrau (Thomas
meaning (Banker 1964:105-106).
'create'. Farther afield, it
1971:153), cf. the pair jeng 'become' and panhj?ng
instance, it occurs asps-
in Rischel's (1995) Mlabri of northern Indochina, as the only instance of

an entirely transparent prefix inMlabri. Rischel describes this prefix as a caus?
ative with two variants pa- and ba-, with the former occurring before stem-initial
voiced consonants, and the latter before voiceless consonants (Rischel 1995:90).
occurs
the
However,
occurs
243
Some Notes
PC Morphology:
causative
*pa-
prefix
also
to PAn.
reconstructs
It certainly
languages of Taiwan;
languages, that is, the Austronesian
(1995) includes both the PAn reconstruction *pa- and various
in the Formosan
for instance, Ross

instances of pa- occurring as a causative marker in Paiwan (cf. also Wolff
1995:566). Blust, in his dictionary (1990, 1991, 1992, 1993, 1994, 1995) recon?
s
structs it for PAn. He notes the causative *pain his reanalysis of Dempwolff
*patay and *matay, which he reanalyses as *pa-aCay 'kill' (with causative *pa-)
and *ma-aCay 'dead' (with stative *ma-). It also occurs in the Philippines and in
some languages of Borneo. It also seems to occur fossilized in some of the Oce?
anic languages (reported in Blust's dictionary).
Thus, the causative *pa- is found in both MK and PAn. If its presence in
MK had any effect on its presence in Chamic, it is probably nothing more than to
reinforce its use, making it less resistant to loss.
The negative
imperative
The negative imperative is without doubt borrowed from MK. Although

it is
found throughout the Chamic languages (Acehnese, Rade, Jarai, N. Roglai, Chru,
Jarai, PR Cham), it does not occur elsewhere inAustronesian.
Table 161: The PC *be?- rref?exivef reciprocal,
Acehnese
PC
Malay
x*be?
bs?
inadvertent'
Jarai
Bahnar
be?
be?
Within MK, Smith (1972) reconstructs it for PNB as *beq, noting its appearance
in Bahnar and Hr?, among other MK languages.
N. Roglai contains not only the reflex of the older PC x*bs? but also a
far more
recent, borrowed negative imperative represented

an
obvious
dar) [di],
borrowing of Vietnamese d?ng.
second,
The bipartite
by the form
negatives
Lee (1996) contains an excellent discussion of Chamic negatives, using N. Roglai

as his point of departure, but covering all of Chamic nicely. Lee reconstructs
three
markers:
negation
*?aka
*ka,
'not
yet'.
the
The
PC
pair
x*?uh...
and
x*?uh...x*?oh
*ka,
*?aka
'not,
'not
no;
yet'
negative'
typically
and
occur
PC
pre
verbally, often in conjunction with the post-verbal x*?oh 'not, no, negative'.
Both the simple negation 'not, no' and the 'not yet' usually occur in conjunction
with the final *?oh, but not always. The bipartite constructions containing these
negation markers
has
bipartite
are limited to the mainland
construction
for
'not yet'
languages,
although Acehnese
goh....lom.
also
244
Chapter
The Chamic negatives are difficult to date with any precision. In addi?
tion to the apparent absence of cognate negatives in Acehnese,
there are other
a
for
Not
ini?
reconstruction.
is
of
the
existence
very
problems
only
preglottalized
tial a problem (it suggests the form might be aMK borrowing), but the initial and
final of x*?uh both show irregularities. The origin of the structure is unclear. Cf.
1996 for further discussion.
Lee
and nominal
Nouns
morphology
The pieces of the nominal morphology

discussed here are the *-am- nominaliz?
1973, Dahl 1976,
ing infix inherited from An (*-um- 'Actor Focus', cf. Wolff
Starosta, Pawley, and Reid, 1982, Ross 1995) and the instrumental infix bor?
infix is thus far lim?
for the *-am- nominalizing
in
and in fossilized forms inWritten Cham; its occurrence
rowed from MK. The evidence

ited to Acehnese
ismainly important as it provides evidence that itwas once productive

in PC. The instrumental *-an- borrowed from MK is interesting because the pres?
ence of a borrowed infix?an
something
apparently productive infix?indicates
Acehnese
about the intensity of the contact.

to
In addition
these
and other grammatical

from An intoMK.
words
-dm- nominalizing
infix
The
were
there
affixes,
that were
various
pronouns,
prepositions,
both from MK
borrowed
into An
and
The PC *-am- nominalizing

infix is reconstructed for PC on the basis of Written
Cham and Acehnese. For Cham, Aymonier and Cabaton (1906:xxiv) do not really
define it but instead simply list -md- (-am-) as an infix and give a set of Written
Cham
examples:
pay?n
'equal',
and
< kararj
'deficit'
k-am-drag
'food,
b-am-dn-dtj
'less',
provisions'
<
p-am-ayau
bag
'to eat'
<
'comparison'
(the morpheme
is particularly
have been added). The last infixed word b-am-on-dg
-urn- and-a/7
a
because
has
both
the
it
double
that
both
infix,
is,
interesting
divisions
instrumental
infix and because
infix
not
is -am-,
-md-
as
it suggests
suggested
that the shape of the -urn- nominalizing

and Cabaton.
by Aymonier
For Acehnese, Durie (1985:32) discusses

there is both a pr?fixai variant and an infix. Cowan
root as being
termed
actor
as an example
of AN
focus.
of
origin
Durie
gives
and identified
the forms
it at some length, noting
that
(n.d.7) clearly recognized the

it with marking what Dahl (1976)
c-wn-arAtj
'cleverness'<
car\g
'clever'
its use.
245
Some Notes
PC Morphology:
Table 162: The PC *-am- nominalizing

PC
Malay
-urn-
infix
-ma
-uim-
*-am-
Bahnar
Wr. Cham
Acehnese
Although only occurring in fossilized forms inMalay, the infix is widely

attested in An. Of the fossilized forms of this infix that do occur inMalay, my
favorite is kancil 'the little one, the pelanduk' < apparently an infixed version of
offers a similar pair: gilang 'luster, glow' and g-em
ilang 'very bright, splendid'; the same pair is listed in Coope (1986) as gilang
'radiant, resplendent' in
'shining' when in isolation and as gilang-g-em-ilang
with
and
in
the
forms
meanings
reduplication,
Coope making it evident that in
Malay the semantics of the infix have been pretty well lost.
k?cil
The
'little'. Blust
(1988:14)
infix
instrumental
infix PC *-an- is borrowed from MK. This was recognized

and Cabaton 1906:xxiv), who note that theWritten Cham
The instrumental
ago by Aymonier
-n- comes from Khmer
nouns
instrumental
say from MK]

'a pile'
'to speak', and b-an-dk
p?9C
have
e.g.,
[I would
d-an-ak
been
and serves to form agentive
< dak
'a dam' < bdk
long
infix
'to pile
on',
p-an-uoc
'to dam' (the morpheme
or
<
'speech'
divisions
added).
It is found throughout Chamic as Table 162 indicates and as a glance

Jarai, Medcalf
through the glossary of reconstructed forms attests to. Within
discusses
the
instrumental
"occurs
when the
nominalization which
(1989:42)
infix -oN- is inserted between
Thus,
kih,
'to
sweep'
this with
the initial consonant and the initial vowel of a verb.
becomes
the widespread
'broom'."
k-on-ih,
However,
while
Medcalf
-in- infix inAustronesian,
equates
only the form sup?
seems
the
the
to
function
distinct
me, particularly in light of
ports
equation;
quite
amorpheme with identical shape and function in the neighboring MK languages.
The MK etymology of the form is widely attested. As Aymonier and
pointed out, a similar infix with a similar range of functions occurs in
It also occurs in Bahnar, a Bahnar language which has long been in con?
tact with Chamic languages; Banker (1964:100) writes that the instrumental infix
Cabaton
Khmer.
-an-
"functions
as a nominalizer,
the infix as sometimes

b?t
'to make
instrumental
a dam'
meaning
having
becomes
e.g.,
par
verbs
converting
b-sn-?t
'to fly'
to nouns."
Banker
describes
of the result of a verbal action, e.g.,
the meaning
'a dam'
becomes
and
p-dn-?r
at
other
times
having
'wing'.
an
246
Chapter
Table 163: The PC *-a/7- instrumental

PC
Aceh.
-uin-;
Jarai Chru
nui-
-an
-an
infix
PR Cham Bahnar
'
-an
instrumental
infix'
phuiat
*pha:t
ph??
pha:?
pahna:?
*p-an-ha:t
*thow
thsa;
tu-
thau
thau
'to chisel,
pha?
?
plane'
'chisel; plane'
thow
'know;
thunow
'magical
able'
knowledge'
x*bat
x*b-an-at
x*poh
poh
?
x*d-an-uc
?
?
ba?
b?t
'to dam; fence'
bani?
banu:?-f
pirn?
ban?t
'a dam;
poh
poh
du?? ?
?
'drumsticks'
t?y?
danui?
a fence'
'strikebound'
panoh
x*p-an-oh
x*duc
bi?
'to sting'
?
tan?y?
'stinger'
a disproportionate
number of forms with this instrumental
Although
infix seem to be MK in origin, the infix also shows up in roots inherited from
PMP. Thus, in Table 163, the first two roots cited are apparent PMP roots, cf.
to Blust (p.c.), the PMP root
Malay pahat and tahu, respectively. According
*paqet 'chisel' reconstructs to PMP based on a cognate distribution that includes
the Philippines and western Indonesia/Chamic.
The infix must have been rela?
based on the large number of varied roots inmainland Chamic
languages containing it and on the fact that it is still what Durie terms productive
tively productive,
in Acehnese.
Prepositions
can be reconstructed for PC. Table 164 shows three
A number of prepositions
of
The
first four show evidence of MK origin: The first has
groups
prepositions.
in various forms after the breakup of PC, with the variation in ini?
tials and finals obvious from the handful of forms given. The second reconstructs
to PC as x*js?. On the one hand, it is reconstructed for PNB as *ajeq 'near' and,
been borrowed
PC Morphology:
247
Some Notes
on the other, it does not appear to have any An etymology. The third evolved
from a verb meaning
'to follow', which may have a Vietnamese
origin, cf. Viet?
namese theo. The fourth reconstructs to PC but also occurs in Bahnar, making it
impossible to evaluate even tentatively without more data.
The next two prepositions
(x*man and *hayow) both reconstruct to PC,
but thus far neither has an obvious An orMK counterpart, although the vowel in
x*man certainly
suggests aMK
etymology.
Table 164: The Chamic prepositions

PC
Acehnese
bak
'at;
Jarai
on'
pa'at,
x*js?
*tuy
je?
*gah
gah
x*mat)
man
Bahnar
'at;
to';
b?t
?ai
'place'
on'
near
tui 'to follow'
'to follow;
gah
according
'side'
'side'
'from'
'from'
'like (prep.)'
*hayow
ta-
ta-
di
?
*danan
*dalam
*ka
to'
dalam;
lam
kui-
'to,
?
rjAn
dl?m (Lee)
l?m
ka-
ka-
'indirect
object'
relation
; kato'
'in
as a
as a
in, at'
'to, at'
'with'
'in; inside'
occurs
prefix
preposi-
both
and
'to;
indirect
object'
tion
last four prepositions are not only reconstructed for PC, but they obviously
have An etymologies. All four are found in closely-related Malay, as di, d?ngan,
dalam, and k?-, respectively, and the first two are found in the earliest Cham
The
inscriptions (Marrison 1975:54). What makes this set particularly interesting is

that three of the four have also been borrowed into Bahnar, with the forms for 'in;
inside' widely enough attested in Bahnaric languages for *l?m to be recon?
structed for PNB by Smith!
Pronouns
Among the personal pronouns readily reconstructable for PC, some are inherited
from PMP, while others are clearly either secondarily developed or borrowed
from MK. The first person singular pronoun PC *kau T
is inherited from PMP,
the first person polite form PC *hulun from 'slave' is innovated, but certainly the
innovation exists outside of PC; one suspects, however, that while the form
*hulun dates back to PC, only the meaning
'slave' existed at the time depth, with
248
Chapter
the first person meaning a later innovation not just in Chamic but throughout the
area. The other polite first person form *dahla? looks to be secondary; it does not
seem to be attested outside of Chamic, nor apparently is itwidely attested within
Chamic.
The
of one of the second person forms *ha is not firmly

etymology
it is not clear if it reconstructs to a pre-Chamic level or not. The other
established;
form in the table, however, xih is definitely a post-PC
Chamic. The third person *ftu is an An root.
from MK
borrowing
into
Table 165: The PC singular pronouns

PC
Aceh.
Rade
Jarai
Chru
Haroi W. Cham
PR Cham
*kau
ksa;
k?u
kau
kau;
k?u
kau
kow
halun
halun
hl?n-;
hal?n
T (polite);
tahla?
'I (polite)'
'you; thou'
hl?n
ulon
hlun
hul?n
*dahla?
I (familiar)'
kami
ku
*hulun
'
dalha?
h?(N)
ih
h?
ih
?u
?u
??
*ha
xih
*ftu
-?oa
-hl??
h?i-v
hi
slave'
hi
thou'
'you;
?au
?u
?u
'he,
she;
in
'we'
-v
-?u
they'
of the plural pronouns are inherited from An. The development of

from a form meaning
'body' is found here and is common through?
out Southeast Asia, with the parallel in Tibetan rang coming immediately
to
Most
the reflexive
mind. The plural form meaning

'group; other', which shows up in various plural
as
is
it is borrowed from MK and, not surprisingly,
pronouns,
particularly striking
does not apparently appear elsewhere in PMP. Thus, among
*gap 'other; group' is borrowed from MK at the PC level.
the pronouns,
Table 166: The PC plural pronouns

PC
Aceh.
*kamai
kamoa;
-v
mua
*ta
ta
*drai
droa
gap;
gui-;
Haroi
hamsi
-gamai
kamei
kami
(incl.)
-v?
-v
-ta
ata
ita
(excl
trai
-drai;
in 'we (inclusive)'
?in
?irj
x?in
g^p
Jarai
drsi
ga
g??
-v
g*7
W. Cham
PR Cham
Rade
tray
-trey
'we (ex.
'we (incl.)'
'we (incl.);
reflexive'
-iv
in
kau?
k?w'
(-k?n)
'we'
'other;
group'
the
PC Morphology:
249
Some Notes
Other pronouns of interest are the late post-PC pronominal pieces bor?
rowed into various Chamic subgroups from MK sources. Note that x?irj and xih
are found among the Highlands Chamic languages, the languages with the most
intense contact with MK.
Other
forms
In addition to the demonstratives which can be found in the glossary, there are
various other grammatical forms, most of which have An etymologies.
ya < *yarj 'relative clause marker': Marrison (1975:53) notes the exist?
ence of this relative clause marker in the older Cham documents.
attested
Malay
-kan: Aymonier and Cabaton (1906:xxiv) say about this suffix, widely
inMalay, "It may have existed in the older language as an analog to the
suffixes
-an,-
/,-
-kan.
It scarcely
now
exists
as
except
the
suffix
-kan
(-kand, -kanay) which, with maljdg

gave the expression mdljag-kan
(mQljdg
to
m?mulia-kan
the
'revere, honor'" [The Malay has
kand), comparable
Malay
been respelled here]. Thus, it is likely that the -kan once existed but has long
since been lost along with all the other suffixes.
sa-, siy: The older form is clearly PC *si-. Both Aymonier and Cabaton
(1906:xxiv) and Marrison (1975:57) note the existence of this honorific-marking
prefix. Marrison describes it "as a familiar element in titles, or as the first element
in siapa ['who']", while Aymonier and Cabaton simply list it as a prefix and then
illustrate it with the following examples: ai 'brother', sa-ai 'elder brother'; nai
'young
sa-nai
woman';
'woman'.
Cf.
also Acehnese
si
'title
for people's
names'.
It is also found inMalay.

dorn 'all': This quantifier, found in Khmer, is borrowed from MK (Mar?
rison 1975:56).
PC x*soh 'only; empty, free, leisure' :This quantifier is found in a full
of
Chamic
but it has an obvious MK
range
languages, including Acehnese,
source
in PNB
reasonable
'only' and in PKatuic *[s/c]anhah

'empty'. It is, thus,
that it is aMK borrowing.
last three connectives are found in Aymonier and Cabaton's dictio?
*sasoh
to assume
The
nary, and all also occur inMalay.

maka:
Aymonier
"well,
pun:
Aymonier
"then,
ampun:
and Cabaton
then,
because,
and,
and Cabaton
if, really;
(1906:386)
thus,
gloss
this connective
as
this connective
as
when".
(1906:288)gloss
also".
(1906:1) gloss this word, used in for?

as "forgiveness, forgive, ask forgiveness"
and Cabaton
Aymonier
mal contexts,
250
Chapter
that
Evaluating these forms is problematic. There was obviously a court Malay
existed in numerous courts in theMalay world, aMalay that occurs in the writ?
these specific forms occur in the stories
ings of these courts. As a consequence,
recorded in both theMalay of Pattani and theMalay of Malaka. Since the forms
do not occur elsewhere
in Chamic,
it is not possible
to draw conclusions
about
the antiquity of the forms.

Were this an attempt to provide a full grammatical sketch of PC, numer?
ous other grammatical elements would need to be treated. Additional grammati?
cal forms are found in the glossary, but a large amount of morphological
reconstruction
remains
vide a preliminary
to be done.
The
point
of
this
chapter,
sketch that gave at least a little of evidence
however,
of MK
was
to pro?
influence on
Chamic.
10
Contact,
and
Multilingualism,
Change
inmany ways
Although
tentative, unquestionably exploratory in nature, and, of

tedious, this study has nonetheless provided a detailed pre?
necessity,
frequently
examination
of the interaction between contact, multilingualism,
and
liminary
a
within
of
the
the
When
Chamic
within
change
history
languages.
placed
linguistic record reveals much not just
about the linguistic history of Champa but also about non-linguistic history.
Much of the history of the region is reflected in the patterns of geograph?
broader historical
ical distribution
setting,
the Chamic
and the deducible patterns of language contact. The early arriv?

peoples most likely landed south of Danang and thus probably
ing pre-Chamic
encountered Bahnarics.
Given the major restructuring of the arriving Austrone?

sian language that took place, these pre-Chamic people must have become
socially dominant, with this dominance leading many most probably Bahnaric
speaking people to shift to Cham, but bringing with them many MK characteris?
tics. The Vietnamese
"push to the south" around 1000 split the northern Cham

into a group that moved southward becoming the Roglai and a group that fled to
Hainan, becoming the Utsat. Upon reaching Hainan, the Utsat came into close
contact with Hainanese
the complete
and, perhaps, one or more Hlai

restructuring of Tsat.
(Li) languages,
leading to
Probably sometime around the fall of Indrapura in the north, although it

have
been as much as several centuries earlier or later, the Chamic speakers
may
were
to become theAcehnese
who
left themainland on a journey that would ulti?
mately
end in northern Sumatra.
The next major restructuring again correlates with a major historical

Vietnamese
conquest of the south, roughly 500 years ago. With the
final conquest of the south by the Vietnamese,
the Cham were not only pushed
event?the
251
252
Chapter
inward from the coast but lost much
10
of their close contact with each other. One
major group was pushed into the Central Highlands, becoming the Rade and the
Jarai in the process. Somewhat later, the Haroi followed, splitting off from the
Coastal Cham and coming under the influence not just of the Rade and Jarai, but
Hr?.
importantly, entering into a close relationship with the MK-speaking
Under this influence, Haroi was extensively restructured. The Coastal Cham in
more
the south came under increasingly strong Vietnamese

influence, particularly in
recent times with the spread of education and so on. The language of these
more
the Phan Rang Cham, has restructured
Cham,
in the direction of Vietnamese.
NON-LINGUISTIC HISTORY
Much
of the interaction between

is scattered
history
between
the linguistic developments

this work and the outlines
and non-linguistic
the interaction
of
throughout
linguistic and non-linguistic history are sketched in Chapter 2. However,
in addition to the insights the data gives us with regard to earlier contact situa?
tions, the linguistic record also clarifies our understanding of two important his?
torical events, one talked about at length in the prior literature and the other
that is, the relative positions
scarcely mentioned,
Sumatra
and
the Tsat
of Hainan
with
respect
of the Acehnese
to the rest
of
the Chamic
of northern
languages.
respect to both languages, the careful, detailed establishing and

of
patterns makes the relationships to the rest of
correspondence
interpretation
the Chamic languages clear. In the case of the relationship between Chamic and
Acehnese,
long a source of speculation in the literature, the examination of the
With
correspondence patterns between Acehnese and the rest of the Chamic languages
provide, at least to a historical linguist, unambiguous and convincing evidence
thatAcehnese was originally simply another Chamic language, which some time
after the breakup of PC, left for northern Sumatra. The evidence for this position
is found throughout this work and is summarized in "Acehnese, a Chamic lan?
guage" on page 47.
In fact, it is also fairly likely that, until they departed for Sumatra, the
were the more northerly end of the Chamic dialect chain, occupying
territory in the regions now occupied by Katuic speaking people. The
Acehnese
coastal
of a clearly Katuic element in Chamic and, significantly, in Acehnese

to be so. Being at that point the most northerly group of Chamic
this
requires
speakers, the Acehnese were probably the first to leave as the Vietnamese moved
left before any significant breakup of the Chamic
south. In effect, the Acehnese
existence
dialect continuum
evidence
makes
into coastal and highlands
languages occurred, as the linguistic
clear.
253
and Change
Multilingualism,
Contact,
left the ancestors of the Northern Roglai

This departure of the Acehnese
as the most northerly group of Chamic speakers and as such these
new northern Chamic speakers came under increasingly pressure from the north.
and the Utsat
in'The history of Tsat contact" on page 224 makes it clear that,

after the fall of the northern capital, some of what were now the new northern
Cham fled to Hainan, where their language developed
into the modern Tsat,
while other northern Cham fled to the south, eventually becoming speakers of
The discussion
Northern
in the southern Vietnamese
Roglai
dence
highlands. The subgrouping

the connection between Northern Roglai
establishes
unequivocally
Tsat from the present locations of the speakers;
Northern Roglai and the Tsat must have been
evi?
and
it has to be inferred that the
in the north when the split

occurred, which, in turn, necessitates a date of roughly 1000 AD, a date very
much in accord with the Chinese records noting the arrival of groups fleeing from
Champa. The rough date may itself be of as much interest to linguists as to histo?
rians as it basically marks the point at which Tsat began its transition from a typ?
ical post-PC Chamic language to the fully-tonal, monosyllabic
today. This date is, of course, of additional interest as theAcehnese
the mainland by that date.
Chamic
Despite
work
and the nature

being rather distinctive
nonetheless
agree
on
the
of language
change
and quite varied, the examples

respective
roles
language it is
must have left
played
by
external
discussed
in this
language
con?
tact and by internal paths of change: the external contact has provided the impe?
tus and determined the directionality and rate of the changes, while the language
internal structures have been largely restricted to providing constraints on the
potential paths for the changes.
Certainly, the major changes in these languages have not come about
gradually. Rather, the exact opposite is true: the major changes are characterized
by short periods of rapid, assimilative restructuring beginning with the onset of
intensive contact and are followed by periods of relative stasis characterized only
by more minor
The
changes?continuing
non-contact-induced
changes
until the next significant period of contact.

are minimal.
the non-linguistic historical findings are significant, the great?

Although
est value of the Chamic data may lie in the insights into the nature of historical
change that Chamic gives us. At the most general level, the Chamic languages
provide insights into the nature of linguistic change itself. In effect, the Chamic
data challenges certain assumptions, some implicit and some explicit, about the
nature of historical
change?not
that these assumptions
have not been challenged
254
Chapter
10
before and not that they will not be made and then challenged again. Specifically,
the following assumptions about the nature of linguistic change are called into
question: the belief that internal factors are more important than external factors
in implementing change, and the belief that rates of change are typically gradual
under
except
The
influence
circumstances.
catastrophic
of internal and external
factors
linguists have recently commented on the preference given to internal

over external explanations of language change. As Grace remarked while talking
about subgrouping and the nature of language change (1985:6), "our professional
Several
paradigm seems to assume that most

internal to the language itself." More
change is caused by conditions

Thomason
and Kaufman in their
recently
and Genetic Linguistics
139ff)
(1988:57-64;
linguistic
Language Contact, Creolization,

discussed what they saw as inappropriate prejudice in favor of internal explana?
tions, resulting in what they saw as a tendency to accept contact explanations
only in lieu of internal explanations, and then, only if the contact explanations
can be completely and fully documented. In fact, inmany instances in the litera?
ture the suggestion of external causes is rejected outright simply because a plausi?
ble internal path of development
is present.
The preference for internal explanations of change is intimately tied to a
that normally language change comes about from some sort of dis?
in
the language undergoing the changes. Since this sort of systemic
equilibrium
is assumed to be the typical cause of change within languages,
disequilibrium
once a plausible internal path of change is found, it is unfortunate but not surpris?
related belief
is often automatically
taken as the cause of
ing that this internal development
as
such
those that lead to the
of
Under
course,
change.
special circumstances,
that external pressures
development of pidgins and cre?les, it is acknowledged
can
be more
influential
than
internal
pressures,
but
such
cases
are
seen
as
atypical
changes. As a corollary, one piece of evidence that supports

the atypical nature of such changes is that they cannot be given a plausible inter?
nal accounting. Of course, this characterization
is in part a straw man, but in part
of normal historical
it does characterize many
of the implicit
if not explicit
assumptions
about the
nature of historical
change.
The Chamic changes, which fall well within the range of so-called nor?
mal historical change, show an unambiguous relationship between internal and
but
the existence of not only internally-plausible
external causation. Despite
often attested paths of internal development for the changes in question, it is the
internal systemic
outside, external models provided by bilingual contact?not
set
the
in
motion
and
determined
their direc
has
changes
disequilibrium?that
255
and Change
Multilingualism,
Contact,
tionality. The internal factors do not play a causative role. Rather, the major role
played by the internal structures of the languages is the limitations these struc?
tures place on the potentially-exploitable
paths of change.
Despite the existence of phonetically-plausible
paths of change in virtu?
case
and
the
fact
that
the
ally every
despite
preponderance of the evidence is cir?
case
for the dominance of the role of external influences in the
cumstantial, the
Even a cursory correlation of the Chamic
is compelling.
changes
contact
with
in
patterns makes it obvious that bilingualism with
changes
changes
speakers of non-Chamic
languages set the Chamic internal changes in motion
Chamic
and gave them their internal direction: in each and every case involving a major
restructuring of the phonology of a Chamic language, the timing and direction of
the change can be correlated perfectly with a new pattern of bilingual contact.
in fact, in some cases, such as the development of register complexes, tonal?
and
restructured register the direction of the change induced by the contact
ity,
not only matches the language patterns of the contact language but is largely oth?
erwise unattested in the 900 or so other Austronesian
languages. The failure to
And,
the centrality of external patterns of bilingualism

in these changes
a
to
to
account
considerable
role
coincidence
for countless
requires ascribing
recognize
repeated
correlations.
The data leads to the conclusion that for the Chamic changes it is con?
that both sets the
tact, not the existence of a language-internal disequilibrium,
in
motion
and
the
determines
of
The
direction
changes
change.
corollary, of
is
that
the
role
of
to
course,
major
language internal patterns is
impose constraints
on
the possible
straints
paths
to have
appear
of
change
frequently
in the
been
literature
mistaken
on
historical
for
causes.
change,
Certainly,
con?
these
this
is par?
of tone, register, and

ticularly obvious with respect to the development
restructured register. The development of phonation distinctions after certain ini?
tials left a system with the potential to develop in any one of the these direc?
tions?the
actual
This
choice
conclusion
was
determined
is strikingly
by
contact.
of the conclusions drawn by

and
Weinreich, Labov,
Herzog's (1968) "Empirical foundations for a theory of
in
which the authors correlate language change with language
language change"
variation and ascribe change not to gradual, incremental changes in one register
reminiscent
in the dialect of one group of speakers but instead to the spreading of a varying
this is not to say that the variants that spread might not be in
rather minor with only the accumulation of such variants
feature. Of course,
and of themselves
the variant, the change itself

resulting in dramatic change, but however minuscule
one
is abrupt?the
to
at in another way, the
from
variant
another.
Looked
change
cause always seems to be adjustment to a new model of language, whether it is a
variant within
the same language or amodel
outside the language.
256
Chapter
10
The fact that the significant changes within the Chamic languages corre?
late with new patterns of contact accounts for the abrupt, non-gradual nature of
the changes. For instance, even today the phonologies of Roglai, Jarai, and Chru
among the Highlands Chamic languages look very much like the system recon?
structed for PC despite the passage of well over a thousand years. In contrast,
have undergone intense bilingual contact and have been totally
restructured. The pattern that suggests itself is a fairly short period of restructur?
Tsat and Haroi
ing followed by a period of relative stasis until new patterns of linguistic contact
come into existence (cf. similar conclusions
in Grace (1985)).
to
of this is to argue that external explanations are automatically
be preferred to internal explanations. That would be nothing more than replacing
one prejudice with another. Rather, the suggestion is more balanced. First, the
clarity of the Chamic data forces recognition that the role played by multilingual?
None
ism in areas such as Southeast Asia
is far, far more central to language change

than previously thought, that the interplay between internal and external factors is
just that?an
interplay, and that the types of relatively sudden, abrupt changes
often found throughout Southeast Asia are more than likely the result of external
factors manipulating
internal factors?not
internal to the language itself.
As Thomason
preferred
when
As a corollary,
nal
and Kaufman
the balance
of
the result of marked
disequilibriums
have argued external explanations
the evidence,
such external explanations
or else
almost
mere
simplicity,
are to be
favors
inevitably have language
them.
inter?
correlates.
As an addendum, itmust also be recognized that, as Grace (1985) has

written, there is another external factor: in times of significant cultural change,
there is a tendency for languages to undergo more rapid change than in times of
relative stability. Some of this may simply reflect the mixing of various dialects
and so on, but some of itmay reflect adaptation to a new range of uses that a lan?
some cases the change may correlate with
guage is being required to perform?in
the transition from exclusively phatic and social uses to a system in which infor?
mation
transfer
plays
a much
more
central
These
considerations,
however,
it obvious
that theMK
substratum
role.
lie outside of the scope of this work.
'Inheritance'
accounts
The evidence
assembled
of the MK
substratum
in this work makes
in
is due to contact, but there are several older, alternate accounts which
attempted to invoke common genetic inheritance as a major determinant of the
Chamic
similarities.
and Chamic
attempts to account for the massive similarities between MK

run into two related, insurmountable obstacles. Reconstructed
PC
These
Multilingualism,
Contact,
and Change
257
not like a remnant of Austronesian

that
looks remarkably like proto-Malayic,
remained behind on the mainland, while the rest of Austronesian went out into
the Pacific. In fact, the comparative evidence makes it clear that PC and proto
Malayic were either sister languages or something very close to that. Thus, PC is
a close-knit, rather shallow subgroup with a time depth of roughly two thousand
years. However, the last time the Austronesian
languages were on the mainland
was at least 6,000 years ago, as the archaeological
and linguistic evidence both
make clear. Further, it is quite clear that the features shared between Chamic and
far too
theMK languages of Vietnam are strikingly similar, almost identical?far,
close to identical inmost cases to reflect common retention from a common rela?
tionship some five or six thousand years earlier. At that time depth, even if there
were an ancient genetic relationship between Austronesian
it
and Astronautic,
would be too far in the past to account for the striking similarities between PC
and MK.
Thus, the 'inheritance' accounts of both Cowan and Shorto are interest?
to
not for their merits as possible accounts of similarities, but
examine,
ing
instead to examine why both Cowan and Shorto found it necessary to invoke such
in the first place. In both cases, their preference for an inheritance rather
than a contact account of the similarities rests on beliefs that certain things are
not borrowable in principle and thus must be inherited. Cowan (n.d.:2) simply
accounts
states that, given

mainland Chamic
and
length agreement is shared by Acehnese
languages on the one hand and by the neighboring MK lan?
on
the
and given that the length distinction cannot have been bor?
other,
guages
rowed ready-made from MK into Chamic, the only other possible explanation
must be that the agreements point to a common genetic origin of vowel length in
MK
that the vowel
and PC where vowel
length was a distinctive feature. If these principles were

follow that the length distinction must be the result of a
to be accepted, it would
common genetic MK/An
origin, with the vowel length feature having been lost
else
in
PAn, "except for sporadic remnants reminiscent of it in certain
everywhere
languages" (Cowan 1983:179). In support of his position, Cowan (n.d.:3) argues
that Chamic represents a dialect of PAn "which not only remained on the conti?
nent but also stayed in the immediate vicinity of the Austro-Asiatic
mother
a
contention
that
is
false.
tongue",
demonstrably
For similar reasons, Shorto (1975:90) also concludes
that the vowel
and mainland Chamic did not result from
length distinction found inAcehnese
a distinction as that of vowel
so
"...the
creation
of
fundamental
borrowing:
length by borrowing seems intrinsically unlikely. On the other hand, contact with
MK languages, inmany of which (as in PMK) a vowel-length distinction oper?
lost else?
ates, would favour the retention of an original distinction conceivably
where." Like Cowan,
Shorto also argues that PAn had an original vowel
length
258
Chapter
10
which was retained in Chamic but lost elsewhere. However, not only
is there no apparent evidence elsewhere for a PAn or PMP vowel length distinc?
tion but the Chamic data can be accounted for without positing any length dis?
tinctions above the PC level ("PC vowel length" on page 138ff.).
distinction
Cowan (n.d.:9) draws a parallel conclusion about the shared morpholog?

ical elements, declaring that "the agreement which exists between the affixal sys?
tems of the Austronesian
languages is a
languages and those of the Mon-Khmer
rather close one" and concludes that the similarities cannot be explained by loans
due to contact
since grammatical morphemes,
unlike vocabulary,
are not easily
borrowed.
is interesting about both Cowan's and Shorto's positions is not the

conclusions
about the existence of a genetic relationship between MK
What
ultimate
(a position that ultimately may turn out to be right but which is certainly
not relevant to this data) but the fact that their conclusions came out of two false
and An
of
length could not have developed as the consequence
that grammatical
and a similar belief about the unlikelihood
could have been borrowed.
one that vowel
beliefs,
contact
language
morphemes
ACCOMMODATION TO A LINGUISTIC AREA

languages is as an example unparalleled in its clar?
on language change. As Eric Oey noted ear?
of
the
effects
of
areal
influences
ity
lier, the study of the Chamic languages provide an example par excellence of the
The other value of the Chamic
effects of the three distinct waves

guages,
of contact?with
the Bahnaric
and Mnong
lan?
then Vietnamese
and the languages of Hainan.

interwoven with the various waves
of contact are multiple

Intimately
Asian
areal
features
otherwise largely
examples of the development of Southeast
the development of rich registral systems, the devel?
unattested inAustronesian:
opment of restructured register, the development of full sets of tonal contrasts,
the internal development of glottalized obstruents, the drive toward monosyllab?
icity, and so on. Chamic provides example after example of internal "drifts"
through external contact. (Cf. "Chamic convergence with
given directionality
Southeast Asia" on page 4ff. and "The general tendencies: A broad overview" on
page 5ff., for a general discussion, with details and examples found throughout).
Particularly notable among the changes is the complete documentation
of register, restructured register, and tonal systems, from a
lan?
atonal, disyllabic
fully-tonal modern
origin to monosyllabic,
completely
guage (in the case of Tsat), cf. Chapter 8, in particular. Unless the Chamic devel?
opments are somehow
inexplicably viewed as an aberration, they provide an
of the development
Contact,
Multilingualism,
and Change
259
intriguing model for the much more inextricably interwoven layers of historical
that characterize the histories of many of the languages that have
development
been intermingling on the mainland for a much longer period of time.
lan?
In fact, because so much is known about the insular Austronesian
so
what
Austronesian
and
because
the
it
is
evident
guages
languages
pre-Chamic
looked like, the changes found in the Chamic languages of the mainland provide
us with invaluable insights into linguistic contact and adaptation to a new linguis?
tic area.
This outline of the history Chamic is a first approximation presented in

the hope that it will prove insightful and useful to others interested in these and
similar
questions.
Appendix I: Language Names

(and transliteration tables)
This appendix contains a list of Chamic language names, information about
which names correspond to which, and transliteration tables.
The transliteration tables are just that?
transliteration tables. Except by
sketches of the languages involved.
chance, they are not complete phonological
Thus, for example, various diphthongs are not listed separately as their transliter?
ation follows naturally from the transliteration of the individual components.
The transcription systems have been largely standardized. The standard?
ization was necessary for at least three reasons: First, many of these languages
have been transcribed in a modified version of the Vietnamese
alphabet Quoc
the
to
notation
familiar
accessible
those
with
Vietnamese or
ngu, making
readily
with
other
numerous
languages
instances,
of Vietnam
data
from
but somewhat
the
same
language
to others. Second,
opaque
has
been
transcribed
in
some?
times even by the same linguist on different occasions for different purposes.
Third, not just intra-language but also cross-language comparability was desired,
both for readers and for myself. Sometimes the same symbols are used for differ?
ent
languages
by-language
with
different
values,
a practice
that makes
sense
on
language
basis but makes
cross-linguistic
comparison difficult.
of the linguists working on Chamic languages are affiliated with
Many
the Summer Institute of Linguistics
its interest in translating the Bible
in part by
(SIL), an organization motivated
into as many languages as possible. In the
this involves a linguistic analysis followed by the
early stages of the process

development of an orthography. The language data being used in this work comes
from all different stages in this process. As a consequence,
the earliest work on
the language is often in a rather phonetically-transparent
while later work is sometimes appears in an orthography
transcription system,
that differs in certain
from the earlier phonemic transcription. Even the system of phonemic tran?
scription may be changed somewhat over time. In almost all cases, however, the
differences in transcription systems have been purely notational and the systems
ways
were
standardized by simply substituting one symbol for another.

The varied orthographies
used in the transcription of the different
Chamic languages have been transliterated with the intention of making the nota?
tion maximally
transparent to the linguistically
sophisticated
reader. First, where
261
262
Appendix
I: Language
(and transliteration
Names
tables)
the same sounds have been
transcribed with one symbol in one language and

a
another in another,
single transliteration was chosen in order tomake the sym?
bols as comparable as possible across languages. Second, in some cases, where a
sound has been transcribed more
or less uniformly in the sources, but where the

value of the symbol is not obvious to readers not familiar with
actual phonetic
an alternate transliteration was
Vietnamese
orthographic conventions,
which makes the phonetics more obvious to the non-specialist.
None
chosen
as criticisms of the
should be misinterpreted
in the works cited. The contrary is true: given
of these adjustments
orthographies originally employed

the intended purpose and primary audience, these orthographies are admirably
well-crafted. However, these orthographies were designed not to provide a maxi?
mally transparent transcription system for linguists, but, much more importantly,
to provide a maximally
readable orthographic system for speakers of the lan?
guages in question. And, where the potential users are literate in another lan?
it is likely in the majority of cases to be in Vietnamese. Thus, in many
these
cases,
orthographic systems often follow the orthographic conventions of
in Roglai /k/ is represented mainly
Vietnamese. For instance, as in Vietnamese,
<
<
>
<
< e >, and < h >, a system some?
i
k
before
with
>,
c>,
by
occurring only
guage,
what reminiscent
of < c> and < k > in English. Similarly, as in

< p >, < t >, < ch >, and < c/k > represent unaspirated
of the distribution
the Roglai
Vietnamese,
voiceless
stops,
aspirated
stops.
<
while
>,
ph
<
th >,
<
chh
>,
<
and
kh
>
represent
voiceless
Acehnese
a dialect
Bireuen,
in the North Aceh
region
Consonants:
Durie
(1985):
k-/-k/'
Ph
b
th
ch
kh
bh
dh ih
s
sy
1
y
Transliteration:
k-/-?/?
ch
kh
ph
b
th
d
9h
bh
dh
jh
gh
r
m
ny
ng
mb
nd
nj
ngg
m
'm
n
'n
?
'?
og
Names
I: Language
Appendix
263
tables)
Notes:
The last row represents Durie's funny nasals, rather than clusters consist?
ing of nasals followed by homorganic voiced stops.
Vowels:
Durie
Transliteration:
(1985):
eu
?, ?
?, ?
e,
ui
9, a
i
? e
Nasalization
sao
and diphthongs:
ismarked
Nasalization
'
preceding
by
the vowel, except after nasal con?
it is fully predictable.
The shwa at the end of various diphthongs
as -e, but transcribed here as /a/.
sonants, where
is written
in the orthography
BlH
The Bih may have been originally a Rade group but have diverged due to long
contact with the Mnong,
although it remains mutually
intelligible with Rade.
Maitre (1912:400) reported that the Pih (undoubtedly the Bih) spoke a corrupt
Rade dialect.
Chru
Chru, Churu, Chrau Hma, Cadoe Loang,
Seyu, Cru, Kru, Rai
Consonants:
Jr?ng et al. (1977:viii-x):

p
ph
b
ch
Transliteration:
th kh
g
th
d
ph
jh
h
-
d
1
?
w
n
n-h
kh
g
jh
s
b
w
nh
ng
ft
n-h
Appendix
264
I: Language
Names
tables)
Notes:
the glottal stop is marked in the original as < - >, but
Intervocalically,
(see "Length,
finally as part of the marking of length, pitch, and nasalization
<
>
and
nasalization"
n-h
The
below).
pitch,
represents an aspi?
peculiar symbol
rated
nasal.
Vowels:
Jr?ng et al. (1977:viii-x):
Transliteration:
ir
<y
Length, pitch, and nasalization:

In the orthography used by Jr?ng et al. (1977:viii-x) for Chru, a complex
of
diacritics has been used to mark three separate features: (a) vowel
system
that
length co-occurring with the "lowering of voice" (Jr?ng et al. (1977:viii-x),
is, most
likely,
the
of
lowering
pitch,
voice quality or phonation change

ence of a final glottal stop.
it may
although
(b) nasalization
be
actually
some
sort
of
of the vowel and (c) the pres?
In the Jr?ng et al. system, individual diacritics have different values

than they have word-finally.
In addition, these diacritics may
word-internally
mark
not
single
feature,
the diacritic
internally
but
complex
of
features.
Thus,
on the vowel < a > in < ? > marks
for
example,
word
two features: vowel
itmarks three features:

length plus voice lowering and nasalization; word-finally,
vowel length plus voice lowering, nasalization, and a final glottal stop.
Despite a reluctance to normalize an orthography, the diacritics in this
transcription system have been rewritten to make the phonetics more transparent
to those not familiar with Chru orthographic conventions:
(a) length plus voice
<
a
a:
is always
>, (b) nasalization
colon, i.e.
lowering is always marked by
marked by a tilde over the vowel, i.e. < ? > and (c) glottal stops are always
marked
by a glottal stop, i.e. < ? >.

Jr?ng et al.
< ? >
<
-a:-
-?- >
< -? >
-a:?
Transliteration
a tilde indicating nasalization
medial:
a colon
indicating
voice
lowering
and
length
final: a colon
indicating both length, and, apparently,

lowered pitch, and a final glottal stop
I: Language
Appendix
<
Names
a tilde for nasalization
-?:
medial:
-a?
final glottal stop

final: a final glottal stop
-?- >
and a colon
indicating
voice lowering and length
final: a colon indicating both length, and, apparently,
lowered pitch, a tilde indicating nasalization, and a
<-a
< ?>
265
tables)
HAROI
Haroi, Hrway, Hroi, Hroy, Bahnar Cham
Consonants:
Mundhenk
Other: Mundhenk,
Goschnick Lee,
Burnham:
and Goschnick
(dictionary):
ph
th
ch
P
ph
kh
d -/q
c/c
kh
ph
th
kh
h
?
b;7b d;?d
1/r
m
t
th
s
Transliteration:
1/r
nh
ng
d
1/r
ng;n
ft O
w
w
-uq
?w
-iq
-?w
-uh
7y
?w
-ly
-u?
-wh
-i?
-uh
Vowels:
tir
ie
uo
o*
As Mundhenk
one
of five
onsets:
e;e
o; o
e: e
o; o
and Goschnick
(1977) note, all the diphthongs
begin with
i, i, u, o, or e.
266
Appendix
I: Language
Names
tables)
In their dictionary, Mundhenk

and Goschnick apparently indicate short
and long l\J by < ie > and < i? >, respectively. However, the marking of length
involved in their use of their orthographic representation of hi is not as clear. In
this work, on the basis of indications found in various articles, the < u? > (with an
additional breve over the already marked <o>, < ?o >, and < uo > are treated as
and < u? > is treated as long hi.
short hi,
and nasalization:
Length
In addition, Haroi has both length and nasalization distinctions. Nasal?

ization is marked with a tilde, i.e., < ? >. In the dictionary, long vowels are
unmarked, while the short vowels are marked as short, i.e. < ? >. In the other
source and in Lee, it is the long vowels that are marked;
are indicated by a grave accent, i.e. < ? >, while the short vowels are
Mundhenk-Goschnick
long vowels
unmarked. In Burnham,
both long and short vowels are marked: the long vowels
are indicated by a macron, i.e. < ? >, while the short vowels are indicated by a
breve, i.e. < ? >, with the unmarked vowels presumably indicating vowels occur?
ring in contexts where no vowel length contrasts occur (ignoring, of course, the
possibilities of errors and forms where the length is unknown).
In this work, Haroi length is unmarked, shortness is indicated by a breve,
and nasalization is indicated by a tilde.
Jarai
Two
sources
have been used for Jarai: Pham Xu?n Tin's Lexique Polygone
is the source of the forms in Lee (1966), who uses a modification
which
(1955),
of Pham Xu?n Tin's orthography and Pierre Lafont's (PL) Lexique Jarai (1968),
the dialect of Plei Ku. However, it is clear that the two descriptions are of differ?
ent dialects.
Consonants:
Pham Xu?n Tin (1955):
Transliteration:
ptck
ptck
ph
b
b
th
d
4
kh
ph
b
g
j
dj
?
mn?g
?m
mn?ng
w
1, r
w
sh
th
d
d
j
?j
?n
?ft
?rj
1, r
71, s
kh
?
Names
I: Language
Appendix
Lee
Lafont (1968):
(1966):
tck
ptchk
ph th
b
d
b
4
j
dj
mnftng
**
m
?h
267
tables)
th
khph
g bd
2b
?d
kh
j
?j
g
?
mnftrj
1, r
r?g
w
1,
4,sh s
y
h
Vowels:
Pham Xu?n Tin (1955):
i
u*
a,
ir
Lee
Lafont (1968):
i
Transliteration:
(1966):
o% ?
(1) The -o, when it used as an off-glide, is written as -u. (2) Final
vowels marked with a breve have had final glottal stops added, where this is sup?
ported by the evidence.
Notes:
Krung
Krung is the only name reported for this group. They are found along the upper
La Liau, Ya Hiau, and Kra Bou north of Ban Me Thuot. Beyond the fact that they
are alleged to be Chamic speaking, there is no information on further affiliations.
Noang
Also
called La-Dang.
The Noang
are located southeast of Dalat
in South Viet?
nam.
Phan Rang
Cham
Cham has its own literary tradition, one that dates back some 1500 years involv?
ing an Indie script. There are two modern descendants of this tradition. One is
268
Appendix
I:Language
Names
tables)
Phan Rang Cham dictionary, which lists Cham forms in this Indie
Moussay's
script and then provides a two transliterations for each item, one representing a
variant of the modern pronunciation and the other providing a more historically
accurate representation of the writing system. The other is the excellent Cham
Vietnamese dictionary by B?i Kh?nh The (1995); this dictionary uses the same
and Cabaton (1906), but instead of their transliteration,
script as Aymonier
the author uses one that is useful and quite
transparent.
Consonants:
Moussay
P
ph
th
*
P
Ph th
s
b d
w
1
w r
m n
Moussay
(1971):
c
ch
?
ch
y
y
?
(1971):
(spoken)
Transliteration:
k
kh
k
kh
P
ph
p
Ph
ng
?
w
w
m
(written)
Transliteration:
p/p
ph
b
bh
p/p
ph
b
th
ch
kh
bh
dh
gh
jh
s
/n
?
w
r
m
?-/-?
t
th
t
th
s
d
1
r
n
t
th
d
dh
s
d
1
r
n
c
ch
?
?h
k
kh
k
kh
?j
y
y
?
c
ch
j
jh
s
k
kh
g
gh
?j
y
?
Note:
InWritten Cham, postvocalic k is almost always (but not always) a glot?
tal stop, as indicated by Moussay's
transcription of the spoken language, which
almost always shows a glottal stop.
The spoken Phan Rang Cham used throughout this work is from Mous?
say at only one important point does his representation
differ significantly
from
Appendix
Names
I: Language
269
tables)
pronunciation and that is with the diphthongs written < -sy >
in modern Phan Rang as i-?ii and /-au/,
and < -oy >, which are pronounced
the actual modern
respectively.
Vowels:
Moussay
i
Transliteration:
(1971):
ir
?, ?
?, ?
o, 3
In addition, for theMoussay and the Aymonier and Cabaton dictionaries

two other more specific transliterations are necessary: must be substituted for
< -?y > and < -?y > to bring their transliteration system in line with the other
transcriptions and the script.
Rade
Rade, Rhad?, Raday, Rde, Ede
Consonants:
Transliteration:
(1978:49):
Egerod
P
ph
b
th
ch
kh
t
P
ph th
b d
xf
s, xz
?b
?d
f
?
w
ph
b
th
ch
kh
1
n
h
?
s, z
d
1
?j
y
ph
b
th
ch
kh
s
?j
r
m
s
4
kh
Y-Chang(1979):
ch
Tharp (1980:vi):
4
1
r
ng
?g
270
Appendix
I: Language
Names
tables)
Notes:
Egerod (1978:49) notes that III and izi only occur in loans. He also notes
that initial glottal stop is left unwritten, as it is considered automatic before all
word-initial
vowels.
Vowels:
Egerod
Transliteration:
(1978:50):
u
Tharp(1980:vi):
Y-Chang(1979):
Note:
Egerod
unmarked.
The
marks
the
remaining
long vowels
unmarked. Note
In addition,
-ao
Lee
nant
except
has
been
mark
that short a
long,
short
the short vowels
leaving
as
vowels
as a
is written
as
transliterated
as
the
short,
is
the
-au.
structurally
leaving
in Tharp and Y-Chang.
notes that "Rade initial m, k, h followed
(1974:644)
/, r, h, w,
long vowels
authors
separate
syllable
with
by any conso?
noncontrastive
vowel. Occasionally
there is a contrast of consonant cluster and disyllabic word
which is handled in the traditional orthography by writing u for w and / for y in
a cluster (cf. hwie
'rattan' which is disyllabic and hut 'to fear' which ismono?
syllabic) and by inserting an a between k and h to indicate a disyllabic word (cf.
kh?n 'cloth' and kah?n
'soldier')." For the sake of making the syllable structure
more
the vowels
transparent,
in such
words
In all the Rade

with
the intervening
kp
orthographies,
shwa omitted.
have
been
the following
re-inserted.
initial clusters are written
>
k9p-
m9p?
>
k9t-
m9t
>
k9k-
m9k-
>
kgb-
m9b-h9b
kgd-
m9d-h9d
k9g-
m9g-
mp-
kt
mt-
kk
mk-
kb
mb-
hb
kd
md-
hd
>
kg
mg-
hg
>
hgg
Appendix
Names
I: Language
k?-
m?
h?
>
kg?-
m9?-
kd-
mrf
hd
>
k9d-
m9d-
km
mm-
hm
>
k9m-m9m-
kn
mn-
hn
>
k9n-
mn-
hn
>
kgg-
m9n-h9i]
ms-
hs
>
k9s-
m9s-
mh
hw
>
m9h-h9W
hj
h?j
ks
>
271
tables)
h9?
h9d
h9m
m9n-h9n
h9s
h9j- h9?j
in this work, the shwas have been reinserted

readily analyzable and more comparable with non-Rade
However,
to make
the forms more
forms.
Rai
The term Rai, according to Grimes (1988), is sometimes used to refer to a dialect
of southern Roglai, a usage that coincides with a comment by Lee (1966:3), but
also notes that the term is also used to refer to a Chru group. As
guages are extremely close together genetically, it is still unclear
these two lan?

to me whether
distinct groups are being referred to or not.

Grimes (1988) lists Rai as related not just to Chru, but specifically
dialect.
Seyu
to the
NORTHERN ROGLAI
Consonants:
et al. (1977:vi):
Awai-hathe
ch
c/k
ph
th
chh
kh
jh
gh
vh dh
s
b a
w
Transliteration:
P
ph
b
bh
h
dj
nh
-/q
ng
?
w
m
t
th
d
dh
s
d
1
r
ch
kh
jh gh
?j
y
Notes:
The
dialects
Ixl \About
it is pronounced
et al. (1977: vi) write that, in some

the IrI, Awai-hathe
as trilled, while in other dialects it has, in effect, disap
272
Appendix
Names
I: Language
tables)
peared, that is, initially, it has disappeared, leaving the onset vowel long medially,
it has disappeared,
leaving the adjoining vowels as a single long vowel and,
finally, it has become a semivowel, somewhat close to the Vietnamese < ir > [i].
(See also the introductory discussion of this section on transliteration.)
The glottal stop < -/q > ismarked in the original as < - > between
els, but as < q > finally.
vow?
Vowels:
et al. (1977:vi):
Awai-hathe
Transliteration:
i
e
Length
and nasalization:
In addition, N. Roglai has both
in isolation ismarked with a colon,
Length
tion is marked with
length and nasalization distinctions.

i.e. < a: >, while nasalization in isola?
a tilde, i.e., < ? >. However, the combination of length and

has been written with < ? > borrowed from Vietnamese, which coa?
lesces the marking of length and nasalization into a unitary symbol. In this work,
nasalization
length
and
nasalized
are
nasalization
vowel followed
separate
kept
and
are written
by a colon symbolizing
as <
?: >,
that
is, as
length.
Cac Gia Roglai

Cac Gia Roglai (Cobbey 1977) or Cat Gia Roglai
ferent from other Roglai dialects (Grimes 1988).
(Lee 1998) is considerably
dif?
Roglai
Southern
Like Chru, Rai is listed by Grimes
(1988) as a dialect.
Tsat
and Pang (1993) note, there are two descriptions and two tran?
scriptions of Tsat, one by Ouyang and Zheng (1983), with subsequent work by
Zheng (1986) and the other by Ni (1988ab; 1990ab). Both transcriptions essen?
As Maddieson
tially agree, with

However,
even
the only apparent differences
with
the
tones,
upon
ent that the two systems essentially
more
being in the transcription of tones.
careful
examination
agree (Maddieson
and Pang,
it becomes
appar?
1993).
Appendix
I: Language
Names
273
tables)
Consonants:
Zheng, Ni:
Ouyang,
p
ph
th
ts
Transliteration:
k
kh
ph
th
ts
k
kh
7b/b ?d/d ? d
Note:
ji
The /ph/ is phonetically
[$] and /kh/ is phonetically
[x].
Vowels:
Except where noted otherwise, the vowels are as in the original sources.
is
not yet completely understood; thus, the tables do not always indicate
Length
apparent irregularities in length.
Ouyang,
Zheng, Ni:
i
e
9
a
Transliteration:
u
u
9
Tones:
Maddieson
and Pang (1993) significantly refines the tonal system of

and Ni and it is this refined analysis that is used in this work.
Ouyang,
Zheng,
Certainly the historical data fully supports Maddieson's
adjustments. Thus, based
on Maddieson
and Pang, Tsat is analyzed in this work as having five etymologi?
cal tones: three level tones in non-checked
syllables and two contour tones in
one rising and one falling. In addition to these five etymologi
tones, there appear to be a handful of additional tonal configura?
cally-predictable
tions occurring only in recent loanwords.
and Pang (1993) argue that, despite what might be suggested
Maddieson
checked
syllables,
by the transcriptions used in the various papers on Tsat, the contour tones are
always associated with checked syllables. With reference to the various falling
tones transcribed
in the sources on Tsat, on the basis the absence of instrumental

data to support the existence of more than one falling tone, Maddie?
son and Pang suggest that there is a single falling tone, which only occurs in
or historical
checked
occurs
syllables. Similarly,
in checked syllables.
there is only a single rising tone, which
again only
274
Appendix
I: Language
Names
tables)
between the Ouyang and Zheng

the correspondences
forms, the Ni forms, and the standardized Tsat cited in this paper. Note that the
final glottal stop, found inMaddieson's
examination of all contour tones has been
167 shows
Table
in the transcription. Further, the pitch values used by Ni are followed,

instrumental work suggests that the pitch range of the contour
tones more closely approximates the Ni analysis.
indicated
as Maddieson's
Table 167: Comparison
of tone transcriptions
Ni
and Zheng
Ouyang
Standardized
55
55
55
high-level
42
53
42?
high-falling
mid-level
33
33
24
33
24?
35
11
11
low-rising
low-level
11
It is important to emphasize that theMaddieson

and Pang (1993) refinements of
the Tsat tonal system are completely substantiated by the historical comparisons.
Western
Cham
In the orthography used forWestern Cham by Kvoeu-Hor and Friberg (1978), the
voiced stops are used for voiceless stops and affricates followed by second regis?
ter vowels,
while
voiceless
are
stops
used
for,
not
voiceless
surprisingly,
stops.
this work uses

despite this treatment being etymologically-correct,
a
vowel
for
with
second
both
series,
stops
being indi?
register
following
cated by a subscribed dot under the immediately preceding p, t, k, or c.
However,
voiceless
Consonants:
Kvoeu-Hor
and Friberg (1978):
ptchk
Transliteration:
ptck
ph
th
chh
kh
ph
dj
kh
?
y
?j
ch
pt?k
sh
th
nh
ng
ft
Appendix
Names
I: Language
275
tables)
Vowels:
Kvoeu-Hor
ir
Note:
Kvoeu-Hor
Written
Transliteration:
between
and Friberg (1978):
-oa-
is no
There
and
and
-oa-.
contrast
Thus,
-ao
Friberg
inWestern
has
Cham
are written
these
been
as
transliterated
between
-ea-
and
as
-au.
-oa-,
-ea-
and
-ea-,
respectively.
nor
The
Chamic
data shows up from two sources in this work: in the transcription

of Phan Rang Cham used throughout (see Phan Rang Cham section above) and in
the citations from Aymonier and Cabaton also scattered throughout this work.
Written Chamic
As Eric Oey notes about the 587-page Aymonier and Cabaton Cham dic?
tionary, many of the cognates are incorrect and there are apparent problems with
the transliteration. Nonetheless,
when used with appropriate caution it constitutes
an invaluable source. The older orthography provides numerous insights into ear?
lier stages of the language, almost all of which can be substantiated by other evi?
dence. As Oey
dictionary
contains
further notes
variant
many
(n.d., p. 2), although Aymonier

forms
and
has
inadequacies
and Cabaton's
in the
script,
"the
variants and apparent quirks of the script may provide value clues to the histori?
cal development of the language."
The modifications
to the transliteration
forMoussay's Phan Rang Cham.

An additional source forWritten
The Cham-Vietnamese
dictionary Tir Dien
Chamic
are identical to those employed

forms is the 1995 B?i Kh?nh
Ch?m-Viet.
Appendix II: The Chamic Lexicon
The organization of Appendix 2 involves an initial division of the lexicon into

those forms that reconstruct to PC and those that do not. Thus, the first major set
to PC are fur?
to PC. The forms that reconstruct
of forms is those that reconstruct
into those with Austronesian

those with MK etymolo?
etymologies,
The
and
those
PC
reconstructions with
established
without
gies,
etymologies.
MK etymologies,
of course, represent the oldest layer of Chamic and MK con?
ther divided
tact.
The second major grouping, post-PC borrowings, represent post-PC lan?

contact.
guage
Language contact is amajor focus of this work so the inclusion of
a large section on post-PC loanwords is central to the investigation.
In addition,
throughout
also noted.
the lexicon, the presence of Chamic loanwords inMK languages is

Loan words both from MK into Chamic and from Chamic into MK
help document
where
and when
the extent of language contact while

the contact
providing
evidence
about
occurred.
The post-PC borrowings have been divided up according to origin, spe?

cifically, into words of Indie origin, of Arabic origin, and of MK origin.
Other loans also occur in Chamic, but it is these three groups that repre?
sent the dominant cultural contact, at least until more recent times. Now of
course the dominance
of the Vietnamese
require the analysis of the count?

The depth of the historically recent
would
less, almost daily loans from Vietnamese.

Vietnamese
influence is seen in the borrowing
'IMPERATIVE,
mon,
which
for
instance,
negative,
xvai,
cf.
don't',
'cotton
cloth'
of the post-PC
Vietnamese
d?ng.
is a post-PC
shows up in Chru as ba:i, cf. Vietnamese
Other
borrowing
xd9g
borrowing
loans
from
are
com?
Vietnamese,
v?i.
xd9g 'IMPERATIVE, negative, don't', N. Roglai d9g; d9g

namese dung. Looks to be borrowed directly from Vietnamese.
di; Viet?
xvai 'cotton cloth', Chru ba:i, Vietnamese

v?i. Note that the Chru is
an
borrowed from the Vietnamese.
related doublet also
However,
etymologically
exists in Chru in the form k9pa:h
PC level (cf. 'cotton' below).
'cotton';
this second form reconstructs
277
to the
278
Appendix
II: The Chamic
Lexicon
loans, some of considerable

time depth like PC x*cawa:n 'cup; bowl, dish', which reconstructs to PC. Others
are more recent like the post-PC borrowing xdaw 'sword; knife'.
time there have been various Chinese
Over
x*cawa:n
cawait,
Haroi
bowl,
'cup;
W.
C9wan,
Cham
cawan
Acehnese
dish',
cawan,
PR
Cham
-v, Chru
cawan
'petite
cuain,
tasse
N.
Roglai
? alcool',
Wr. Cham cawan, Malay cawan; MK: PMnong *ci?n 'bowl'. Coope (1976) lists
this as a borrowing from Chinese.
If so, it looks like the coalescence of a two
to
construction
the
the most likely candidates being 'tea' +
with
one,
syllable
'bowl'
[cf. Mandarin
chawan
'bowl'
'tea'
cha
+ wan
'bowl'
(Baxter, p.c.)] producing Malay
'tea-bowl;
tea-cup'
; Japanese
cawan 'cup; bowl; dish'.
xdaw 'sword; knife', Rade dau, Jarai (PL) dau, Jarai (Lee) dau, Chru
da:u -1,N. Roglai dau, Haroi thiau < *dau; dau 'machete's, like a knife but
longer', W. Cham dau, PR Cham daw, Wr. Cham daw; MK: Bahnar (AC) dao;
(DT) *daaw 'knife'. This word is probably a Chinese borrowing,
but the intermediate paths are anything but obvious. Cf. Baxter (1992) Old Chi?
nese *taw 'knife', Middle Chinese taw,Mandarin d?o.
MK: PKatuic
the handful of Chinese borrowings represent a relatively indirect con?

However,
tact without a significant impact on Chamic languages. Thus, these are left for
another
study
with
another
The post-PC
cussed
below
Comments
in their
focus.
borrowings
own
from Indie, Arabic,
and MK
sources are dis?
sections.
on methodology
and the lists
The problem of sorting out which borrowed forms date back to PC and which
were borrowed more recently is often present. Thus, comments on the distribu?
the sets are sometimes
tion are included with some of the sets. Specifically,
labelled as either restricted
to Highlands Chamic (Rade, Jarai, Chru, N. Roglai,

(which was a part of N. Rogali until itmoved to Hainan)), or as High?
lands Chamic plus Haroi andWestern Cham, the two originally Coastal Chamic
languages that have had an extended period of contact under the influence of
and Tsat
Thus, words restricted to Highlands Chamic plus

highlands MK
languages.
Haroi andWestern Cham are likely to have been borrowed from MK. Even when
the
existence
of
regular
correspondence
patterns
suggests
a form
reconstructs
to
is quite real that some of these regularly corresponding forms

PC, the possibility
are actually later borrowings, particularly in the case of forms that are not only
attestation but also restricted toHighlands Chamic.
without wider Austronesian
information is also evident in
Other potentially valuable distributional
the lexicon. Thus, the presence of a Phan Rang Cham form or aWritten Cham
279
Lexicon
II: The Chamic
Appendix
such a form establishes that a set of words is not just restricted

now
to the languages
found in the highlands, but instead is distributed throughout
mainland Chamic and thus reduces the possibility that the form was borrowed
form is obvious:
from MK
only after many of the Chamic languages had moved

lands. More difficult to evaluate is the importance of Acehnese
up into the high?

attestation. Cer?
tainly the lack of an Acehnese form is particularly difficult to evaluate; it is likely

that some forms have simply been lost, while others have been replaced by Malay
forms. However, even if there is a related form inAcehnese,
it is possible that the
word may have entered Acehnese after Acehnese
left the mainland. It has, in fact,
been suggested that Acehnese has a number of MK words not normally found in
Bahnaric
and Katuic,
but this remains to be established.
from the possibility of inadvertent omissions, the following lists

include not only all the reconstructions referred to in the text
of reconstructions
Aside
but also all the reconstructions
retained from Lee (1966). However, the list does

not begin to include each and every form in each and every language examined.
For example, four sources of Rade were examined and analyzed, but only one is
included in the list. Similarly, the work on Phan Rang Cham by the David and
Thomas, by Ernest Lee, and by Gerard Moussay were
and included in the analysis, but only the forms from
Doris Blood, by Dorothy

all thoroughly examined
Moussay are systematically
ple sources were examined
included in this appendix. In each case where multi?

and used, the source chosen for inclusion in the lexi?
con is the one that provided the longest list of cognate forms?no
other criterion
It needs to be emphasized that the analysis itself used all the available
was used.
forms,
and
occasionally
in the
text
an alternate
source
has
been
used
to
illustrate
a point, usually because that source happened to have fewer gaps than the others.
In the list of forms given here, except for Jarai, only one source for each language
has
been
included.
As
a consequence,
a form
occasionally
appears
more
margin?
ally attested than it is; inmost cases this simply means that the attestation was
found, not in the languages given below, but in several of the other languages
used for this study. Thus, while in general Appendix 2 includes the data used for
the analysis, it does not begin to include the whole data base; that would have led
to an appendix extending to hundreds of pages with little corresponding increase
in our knowledge.
The decision
has also been made to include various problematic sets,

as
labeled
such, in the sets below. Further work and feedback from vari?
clearly
ous readers should clarify much of this sort of indeterminancy ; the hope is to see
updated at a later time.
If there is a discrepancy between the reconstructions
in the body of the
text and the reconstructions
in the appendix of forms, the appendix of forms is
almost inevitably correct. While
the various parts of the body of the text were
this collection
280
II: The Chamic
Appendix
final
times, the appendix was always updated and represents

exist,
hopes, of course, that no such discrepancies
at different
written
version.
One
undoubtedly more
than one has slipped by.
1.
Vocabulary
Chamic
Lexicon
the
but
to PC
Reconstructable
As expected, a large number of these forms reconstructed for PC have obvious

An etymologies
It needs to
and these constitute the first set of forms presented.
be noted that many of these Austronesian
forms have been borrowed into MK
languages of Vietnam and are thus found in both the Bahnaric and in the Katuic
branches of MK. As a consequence,
it is often only the existence of fairly well
established Austronesian
that allow us to establish their ultimate
etymologies
Austronesian
1.1
affiliation.
of Austronesian
PC words
*-9n-9n-,
(Banker)
origin
'INSTRUMENTAL
Chru
-9n-,
N.
Acehnese
infix',
-an-, W.
Roglai
Cham
-9n-,
Rade
nui-,
-tun-,
PR
Cham
-9n-,
Jarai
-9n-.
*?abih
'all; finished, done', Jarai (PL) abih, Jarai (Lee) ?abih, Chru
abih 'completely', N. Roglai abih, Tsat phi55 phi55, Haroi aph?h; ph?h,W. Cham
pih, PR Cham apih, pin, Wr. Cham abih; bih, Malay habis 'done, finished,
entirely', PMalayic *habis, PWMP *qabiq, *qabis.
*?ada
PMP
ada,
*adaq;
'have,
there
is,
there
are',
Tsat
tha11, W.
mata
Cham
i?, Malay
*wada.
ado9, Rade adsi, Jarai (PL)

*?adsy 'younger sibling; cadet', Acehnese
ou
ad9i 'soeur
fr?re cadet', Jarai (Lee) ?ad9i, Chru ad9i, N. Roglai ad9i, Tsat
thai11, Haroi athii, W. Cham tay, PR Cham atsy; t?y, Wr. Cham ad?i, Malay
adik (Blust (p.c.) notes that the -k is the retention of the PWMP vocative marker
*-q), PMP *huaji-q (Blust (p.c.) notes that the *-q is a vocative marker).
dho9, Rade adhsi, Jarai (PL) thai, Jarai
*?adh?y 'forehead', Acehnese
N.
Tsat
Chru
th9i,
th9i,
thai,
thai33, Haroi th9i -v, W. Cham thay,
(Lee)
Roglai
PR Cham they, Wr. Cham dh?i, Malay dahi, PMP *daqih.
(also
*?adu? 'room', Rade ad??, Jarai (Lee) ?ad??, Chru adu?, N. Roglai adu?
a elf.), Haroi ath??, PR Cham at??; tu?, Wr. Cham aduk, Malay
cf.
p?raduan.
*?alih 'move residence',

alih, PMP *aliq.
Malay
Jarai (PL) ? ?, PR Cham al?h,Wr. Cham alih,
*?ama 'father', Rade ama, Jarai (PL) ama, Jarai (Lee) ?ama, Chru ama,
N.
Roglai
Cham
am?,
ami, Wr.
Tsat
Cham
ma11,
Haroi
ami,
PMP
ama
*ama,
'father;
PAn
address
term',
W.
Cham
mi,
*ama.
PR
Appendix
II: The Chamic
281
Lexicon
*?ana:k 'child', Acehnese anu??, Rade anak, Jarai (PL) an??, Jarai (Lee)
?an?? -1, Chru ana:?, N. Roglai ana:?, Tsat na?24, Haroi ana?, W. Cham ni?, PR
Cham ani?, Wr. Cham anik, Malay anak, PMalayic *anak, PMP *anak.
*?anan
Chru
'name',
anan,
W.
Cham
PR
n?n,
Cham
g?n, Wr.
an?n;
Cham anan; nan, PMP *najan; MK: Bahnar (AC) anan ?.

*?anan 'name', Acehnese
nan, Rade an?n, Jarai (PL) an?n, Jarai (Lee)
?an?n,
N.
Tsat
an?n,
Roglai
nan33,
anSn
Haroi
-v, PMP
*najan.
*?anan 'that (third p.)', Acehnese f?an, -nan, Rade n?n, Jarai (PL) an?n,
Jarai (Lee) ?anun < *u, Chru nin, Tsat nan33, Haroi n?n -v,W. Cham ngn 'there',
PR Cham n?n, Wr. Cham nan, PMalayic
*(a)na(?), PMalayic
*(i)na(n),
*(i)na(n),
PWMP
*(a)na(?),
*-nan
(note that this is reconstructed
as a PWMP
suffix).
*?anap 'front (of)', Rade ti an?p, Jarai (PL) an?p, Jarai (Lee) ?an?p,
Chru anau?, N. Roglai pa? an??, Haroi ari?au?,W. Cham kah nau?, PR Cham
an??, Wr. Cham anak, Malay hadap, PMP *qa(n)dep.
*?anin 'the wind', Acehnese
agen, Rade ag?n, Jarai (PL) agin, Jarai
(Lee) ?agin -1,Chru agin, N. Roglai agin, Tsat gin33,Haroi ag?n,W. Cham gin, PR
Cham agin; gin,Wr. Cham agin; gin, Malay angin, PMalayic *agin, PMP *hagin.
*?antow
'ghost;
Acehnese
corpse',
corpse',
W.
atau
Cham
PR
-presyllable,
at?u
Rade
uints9,
(Lee) ?at9u, Chru atgu 'corpse', N. Roglai
Jarai
'corpse',
at9u 'underworld', Haroi at?u 'ghost,

Cham
at?w,
Wr.
Cham
atuw,
Malay
hantu, PMP *qanitu, PAn *qaNiCu.

*?apuy 'fire', Acehnese apui, Rade pui, Jarai (PL) pui, Jarai (Lee) ?apui,
Chru apui, N. Roglai apui, Tsat pui33, Haroi apoi, W. Cham pui, PR Cham apuy;
puy, Wr. Cham apu?i, Malay api, PMalayic *api, PMP *hapuy.
*?asah
asah
Acehnese
'sharpen',
in
Rade
'whetstone',
sah
'whet',
Jarai (PL) ?sah, Jarai (Lee) ?asah, Chru asah, N. Roglai asah, Haroi as?h, W.
Cham sah, PR Cham than, Wr. Cham thah, Malay asah, PMalayic *asah 'grind;
sharpen',
PMP
*hasaq.
asap, Rade s?p pui, Jarai (PL)

*?asap 'smoke (of a fire)', Acehnese
Jarai
Chru
N.
Tsat sau?24, Haroi as?u?, W.
asa?,
asa?,
(Lee)
?s?p,
s?p,
Roglai
Cham sau?, PR Cham th??; ath??, Wr. Cham s?k; athak, Malay asap, PMalayic
(Adelaar) *as9p, PWMP *qasep.
*?asey 'flesh; meat; body; contents', Acehnese aso9, Rade assi 'body',
Jarai (PL) ?sar -vf, Jarai (Lee) ?as9i, Haroi as?i 'body', PMalayic *isi?, PMP
*hesi; MK: PKatuic (DT) *s?j 'meat'. Note that this form has been borrowed
from Chamic into Katuic, but apparently not into Bahnaric.
*?asow
?as9u,
Cham
Chru
athow;
'dog', Acehnese
as9u,
thow,
N.
Roglai
Wr.
Cham
as89, Rade as?u, Jarai (PL) asgu, Jarai (Lee)
asgu,
Tsat
suw;
a11
asug,
sau33,
Malay
Haroi
gigi
as?u,
asu
W.
'canine
Cham
sau,
tooth',
PR
PMa
282
Appendix
layic *asu?, PMP *asu, Pan asu; MK:

structs both in PMK and PAn.
II: The Chamic
Lexicon
*so. This root obviously
PMnong
recon?
*?ata:s 'far; above; long', Acehnese atui9h, Rade tain, Jarai (PL) ataih,
Jarai (Lee) ?ataih, Chru ata:h, N. Roglai ata, Tsat ta33,Haroi atah, W. Cham tah,
PR
Cham
PMP
Cham
atah,
tab, Malay
atas
'above',
PMalayic
*atas,
*atas.
*?awa
wa
t?h, Wr.
at?h;
'paternal',
'uncle,
aunt',
abuwa
Acehnese
PMP *ua? 'uncle, aunt' MK:

;
?, Rade
PMnong
awa
'uncle,
*wa. This
aunt',
Chru
is a variant of
*?awa.
*?ayup 'blow e.g. the wind; whistle, instrument', Acehnese yop 'blow
Rade
flute',
ay??, Jarai (PL) ay?p, Jarai (Lee) ay?p, Chru iu:?, N. Roglai ayu:?,
W.
Haroi ay??,
Cham y??, PR Cham ay??; y??, Wr. Cham ayuk; yuk, Malay tiup,
PMalayic *iup, PMP *heyup, PAn *Seyup.
*?iar -f 'water (fresh)', Acehnese
?9, Rade ea, Jarai (PL) ia, Jarai (Lee)
N.
Tsat
Chru
Haroi
ea, W. Cham ea, PR Cham ya; ier -f,
?ia,
ia,
?ia33,
Roglai ia,
Wr. Cham ia, Malay air; ayer, PMalayic *air, PMP *wahiR. There are problems
with
the PC correspondences
for this form. It is, perhaps, unrelated.
Acehnese
*?idug 'nose',
idog, Rade ad?g, Jarai (PL) adug; dug (Ss);
Jarai
Chru
(Lee) ?ad?g,
?dug (S),
adug, N. Roglai id?k, Tsat thug11, Haroi ath?g,
W. Cham tug, PR Cham it?g; at?g; t?g, Wr. Cham idug; adug; dug, Malay
hidung, PMalayic *hidug, PMP *ijug.
ikan < Malay,
*?ika:n 'fish', Acehnese
(Lee)
?akan,
Chru
aka:n,
N.
Roglai
ika:t,
Tsat
Rade kan, Jarai (PL) akan, Jarai

ka:n33,
Haroi
akan,
W.
Cham
kan,
ikan; kan, Wr. Cham ikan; kan, Malay ikan, PMP *hikan; MK: PNB
*ka, PMnong *ka, Bahnar (AC) ka, PSB (Efimov) *ka:, PKatuic *?9ka:. The
MK and the PC forms both date back to their respective proto-languages.
PR Cham
*?ikat 'to tie', Acehnese

ikat, Jarai (PL) aka?, Jarai (Lee) ?ak??; k??,
Chru aka?, N. Roglai ika?, Tsat ka?24, Haroi ak??, W. Cham k?k, PR Cham ik??;
k??, Wr. Cham ikak; kak, Malay ikat, PMalayic *ik9t, PMP *hiket; MK: PNB
*k?t, Bahnar (AC) k?t, PSB (Efimov) *k9t. [The PSB must be borrowed from
PC]
iku < *-r, Rade ku, Jarai (PL) aku, Jarai (Lee)
*?iku 'tail', Acehnese
?aku, Chru aku, N. Roglai iku, Tsat ku33, Haroi akou, W. Cham hla ku, PR Cham
iku,Wr. Cham ik?, Malay ?kor, PMalayic *ikur, PMP *ikuR.
*?ina 'mother (animal); big', Acehnese
inAg, Rade ana 'female', Jarai
(Lee) ?ana, Haroi ania, W. Cham ni in ni tagin 'thumb', PR Cham ini,Wr. Cham
im, Malay betina [be/t/ina] 'female', PMalayic *ina, PMP *(t)-ina.
*?in?; *iney 'this', Acehnese f?o9, -no9, Rade tinsi, Jarai (PL) anai, Jarai
(Lee) ?anai, Chru ni 'here, this', N. Roglai tin?, Tsat ni33, Haroi ni,W. Cham ni,
PR Cham ni,Wr. Cham ni, Malay ini; ni, PMalayic *(i)ni(?).
Appendix
II: The Chamic
Acehnese
'root',
*?ugha:r
283
Lexicon
<
ukhui9
*-r,
Rade
Jarai
agha,
akha,
(PL)
Jarai (Lee) ?akha, Chru akha, N. Roglai ukha, Tsat (kai33) kha33; ha33; ha33 eee,
Haroi akha, W. Cham kha, PR Cham ukha; akha; kha, Wr. Cham ukh?; agh?,
Malay akar, PMalayic *akar, PMP *wakaR.
*?ular
<
ului9
Acehnese
'snake',
*-r,
Rade
ala,
Jarai
(PL)
Jarai
al?,
(Lee) ?ala, Chru ala, N. Roglai ula, Tsat (la11) la33, Haroi alia 'snake (poison?
ous)', W. Cham la, PR Cham ula; ala; la, Wr. Cham ul?; al?; l?, Malay ular,
PMalayic *ul9r, PMP *hulaR, PAn *SulaR.
-1
*?ura:g
someone',
'person;
Acehnese
ar?g, Jarai (Lee) ?ar?g -1, Chru araig 'people,

za:g?33,
Haroi
-1 'elf.',
ar?g
W.
Cham
rag,
Rade
urui9g,
ar?g
elf.', N. Roglai
PR Cham
urag, Wr.
-1, Jarai
(PL)
ura:k; ra:k, Tsat
Cham
urag,
Malay
orang, PMalayic *urag 'person; someone'. Blust notes that this is restricted to
languges inWestern Borneo which are either closely related toMalay, or which
have been in a borrowing relationship with Malay for many centuries. It is listed
under the appendix
entitled
length irregularity in Chamic

*?urat
'vein,
vessel,
'loans'
in his comparative
difficult to evaluate.
tendon'
cf.
'root',
dictionary, making
Acehnese
urat;
urAt
-v
the
'vine',
Rade aru?t (m), Jarai (PL) ar?t, Jarai (Lee) ?ara? -v, Chru ara?, N. Roglai ura?,
Tsat za?24, Haroi ar?? -v, W. Cham r?? 'thread', PR Cham r??; ar??; ur??, Wr.
Cham arak, rak, Malay urat, PMP *uRat, PAn *uRaC; MK: Bahnar (AC) ara,
The antiquity of
(Efimov) *?urat cf. 'root', PKatuic *?9ra:? 'vein, leaf.
the PAn and PMP forms suggests that the borrowing is into PSB and PKatuic, not
PSB
the other
way
*babah;
around.
*mabah
'mouth',
Acehnese
babah,
Jarai
(PL)
bab?h,
mab?h,
Jarai (Lee) bah; mgbah -i, N. Roglai mubah (also a classifier); bah, Tsat pha55,
Haroi p9ph?ah, W. Cham papah, PR Cham papah, Wr. Cham pabatj, PMP
*baqbaq; *beqbaq.
bui, Jarai (PL) b?bui, Jarai (Lee) b9bui,
*babuy 'wild pig', Acehnese
Chru p9bui, N. Roglai ba bui, Tsat phui11, Haroi p9ph?i, W. Cham papui, PR
Cham papuy, Wr. Cham pabu?i, Malay babi 'pig', PMP *babuy, PAn *babuy.
baro -f, Rade mr?u, Jarai (PL)
*bahrow 'new; just now', Acehnese
p9hr9u, phrgu, Jarai (Lee) phrgu -i, Chru b9rh9u, N. Roglai bahr9u, Tsat phi9n,
Haroi ptiau, W. Cham pahau, PR Cham pirow, Wr. Cham biruw, Malay baharu;
baru, PMalayic *baharu?, PMP *baqeRU.
*balow 'widowed', Acehnese
bals9, Jarai (PL) bl9U, Jarai (Lee) bl9u,
Chru bl9u, N. Roglai bal9u, PR Cham pilow, Wr. Cham biluw, Malay balu, PMP
*balu.
*banut 'banyan, balete', Rade m9n?t, Jarai (PL) mgn?t, bgn??, Jarai
(Lee) b9n??, Chru b9n?:?, N. Roglai bin?? 'banyan, balete', PR Cham pinu?, Wr.
Cham binuk, PWMP *bunut 'Ficus species'.
284
Appendix
II: The Chamic
Lexicon
*bara 'shoulder', Rade mra, Jarai (PL) bra (Pk), Jarai (Lee) bra, Chru
N.
bra,
Roglai bara, Tsat phia11, Haroi pria, W. Cham pra, PR Cham pira, Wr.
Cham bir?, PMP *qabaRa.
*barah
Haroi
priah,
*baReq
'swell;
W.
Cham
swollen',
paran,
Acehnese
Malay
barah, Chru brah, N. Roglai
barah
'tumor,
abscess',
PMP
barah,
*baReq;
'abscess'.
*basah 'wet; damp', Acehnese

basah, Rade m9sah, Jarai (PL) p?s?h,
Jarai (Lee) p9sah, Chru p9sah, N. Roglai pasah, Tsat sa55,Haroi pgs?h, W. Cham
*basah,
pasah, PR Cham path?h, Wr. Cham pathah, Malay basah, PMalayic
PMP *baseq.
*batsy 'banana', Rade m9tei, Jarai (PL) p9t9i, Chru p9t9i, N. Roglai
Tsat
u11 tai11, Haroi p9t?i, W. Cham patay, PR Cham patsy, Wr. Cham
pit9i,
PMP
pat?i,
*punti; MK: PNB *p?t, PMnong *prit, PSB (Efimov) *p?:t, PKa?
tuic *p9riet, *?9tiet. Despite some similarities, neither theMK nor the PC words
for banana are borrowed from the other language.
'stone', Acehnese
bat89, Rade boh tau, Jarai (PL) p9t9u, Jarai
(Lee) p9t9u, Chru p9t9u, N. Roglai pat9u, Tsat tau11, Haroi p9t?u, W. Cham
patau, PR Cham pitow; patow; patow, Wr. Cham bituw; batuw, Malay batu,
PMalayic *batu, PMP *batu.
*batow
*batuk
bato?, Rade m9t?k, Jarai (PL) p?t?k, mat?k

'cough', Acehnese
(Pk), Jarai (Lee) p9t?k, Chru p9tu?, N. Roglai pitu?, Tsat tu?42,Haroi pato?, W.
Cham pat??, PR Cham pat??, Wr. Cham patuk, Malay batuk; MK: PMnong
*b9sy?k, PSB
are
only
(Efimov) *b9shi9?. Despite
the apparent similarity,
theMK
forms
lookalikes.
*bilit
Roglai
Cham
-i 'twist', Rade bl?t -i, Jarai (Lee) bl??, Chru bli:? 'roll up', N.
bili:?, Haroi pli? -first vowel, W. Cham pli? 'wrap up, roll, package', PR
li? -i, Wr. Cham lik, Malay b?lit 'twining round, coiling round', PMP
*belit.
'daughter', Jarai (PL) b?nai,

*binay 'virgin', Rade mgnie
bgnai, Chru bgnai 'female (animal)', N. Roglai binai 'of animals',
panai 'female (animal)', PR Cham pinay 'woman', Wr. Cham binai,
The
'wife', PMP *binay 'virgin', PAn *binay 'female; woman'.
Jarai (Lee)
W. Cham
Malay bini
PAn is an
infixed form (*b-in-ahi); the infixed variant is not attested anywhere in Taiwan.
*blah 'chop; split', Acehnese plah -i, Rade blah, Jarai (PL) bl?h, Jarai
(Lee) blah, Chru blah, N. Roglai blah, Tsat phia55, Haroi pliah, W. Cham plah,
PR Cham plah, Wr. Cham blah, PMP *belaq, Malay b?lah, PMalayic *bglah,
PNB *pah, PSB (Efimov) ^lah,
PMP *belaq 'crack, split open' MK:
;
PMnong
*blah 'split'. Note that this root also appears to reconstruct in three branches of
Bahnaric.
II: The Chamic
Appendix
285
Lexicon
*bley 'buy', Acehnese blo9, Rade blsi, Jarai (PL) bl9i, Jarai (Lee) bl9i,
Chru bl9i, N. Roglai bl9i, Haroi plii, W. Cham play, PR Cham pl?y, Wr. Cham
bl?i, Malay b?li, PMalayic *bali, PMP *beli, PAn *beli.
*boh 'fruit; egg; elf. for round objects', Acehnese boh, Rade boh, Jarai
(PL) boh, Jarai (Lee) boh, Chru boh, N. Roglai boh, Tsat pho55, Haroi phtm
buah elf.,
<phu?h>, W. Cham poh -v, PR Cham poh, Wr. Cham buah, Malay
PMalayic
*buah, PMP *buaq, PAn *buaq.

(*boh) *b9tih 'calf of leg', Acehnese
buiteh, Rade boh tih, Jarai (PL)

N.
boh
Chru
Roglai boh patih, Haroi p9teh, W.
p9tih,
p?tih, Jarai (Lee) p9tih,
Cham poh patih, PR Cham patih, Wr. Cham patih, Malay b?tis, PMP *betis.
*?
*bow 'stench', (the initial is a voiced variant of the more frequent
, but is also found
'smell',
Roglai
Malay
PMP
in Acehnese),
Acehnese
bs9, Wr. Cham
bau, Malay
bau
*bahu.
*?ow 'stench', Rade ??u, Jarai (PL) ?au, Jarai (Lee) ?9u, Chru ?gu, N.
?gu, Tsat ?933, Haroi ??u, W. Cham ?au, PR Cham ?ow, Wr. Cham ?uw,
bau 'smell', PMP *bahu.
*bra:s 'rice (husked)', Acehnese
brui9h, Rade braih, Jarai (PL) braih,

Jarai (Lee) braih, Chru bra:h 'pounded rice', N. Roglai bra, Tsat phia11, Haroi
priah, W. Cham prah 'pounded rice', PR Cham prah -1,Wr. Cham brah, Malay
b?ras,
PMP
*beRas.
*brsy 'give', Acehnese bri -f < Malay ?, Rade brsi, Jarai (PL) brgi, Jarai
(Lee) br9i, Chru br9i, N. Roglai br9i, Tsat ?, Haroi pr?i,W. Cham pray, PR Cham
pr?y, Wr. Cham br?i, Malay b?ri, PMalayic *b9ti?, PMP *beRay, PAn *beRay.
one would expect PMalayo
[The PC and PMalayic vowels are unexpected;
Chamic
*-ay,
*brsy
not
'to
the
of what
reflexes
permit'
<
'give'
was
apparently
cf.*brsy
'give',
*-i]
PMalayo-Chamic
Rade
brsi,
Chru
br9i,
N.
Roglai br9i, W. Cham pray.

*bru? -n -f 'rotten', Acehnese
bro?, Rade br??, Jarai (PL) br??, Jarai
-n
-f, Tsat zo?24 ?, Haroi pr?? -i,W. Cham
(Lee) br??, Chru bru?, N. Roglai br?k
pr??, PR Cham pr??, Wr. Cham bruk, Malay buruk 'worn, decayed (of vegeta?
bles)',
PMP
*buRuk 'rotten meat;
addled
eggs; bad character';
MK:
PKatuic
(DT) *qab/?_k.
*bru?? 'do; work', Rade bru??, Jarai (PL) m?? bru??, Jarai (Lee) bru??,
Chru brua? -n, N. Roglai bru??, Haroi pro?,W. Cham pr??, PR Cham pr??, Wr.
*bru9? 'work, industrious'; PKatuic (Dorothy
Thomas)
*br_q [Pacoh proaq, Katu briaq]. This set of forms plus the set for
*buat 'do, work' is inordinately interesting. First, it looks like *buat 'do work',
which has an excellent PAn genealogy
[PMP *buhat, PAn *buhat]. However, it
Cham bruk; MK:
PSB
(Efimov)
If it is a borrowing from MK,

then the PAn forms must be counted as chance look-alikes. If, as ismore likely, it
also occurs
inMK
in both PSB and in PKatuic.
286
Appendix
was borrowed
II: The Chamic
Lexicon
and PSB from An, then this is another An form bor?

(or, perhaps, just Katu and Pacoh) from a presumably Chamic
the coast, a little north of the northern-most capital of Champa.
into Katuic
rowed into Katuic
source along
Both the PSB and the PKatuic
ogy not expected
into Post-Chamic
from one of these sources but with
MK; perhaps the -r- was

*buat
N. Roglai
ferent',
bu9t
Acehnese
used in the phrase

N.
tuk9t
Roglai
Acehnese
different',
'the other day',

-i, Haroi
from
speaking Chamic.
bu9t
'action';
'action';
'do',
pubu9t
buat 'do', PMP *buhat, PAn *buhat.
bu??, Tsat phua?42, Malay

'other;
the -r- infix borrowed
introduced by MK bilinguals
'do; work',
*buk9n
m9k?n
forms contain an infixed -r-, a piece of morphol?

inMK. Perhaps itwas then reborrowed
but common
in Chamic
bukAn;
'on
the
Rade
contrary',
'dif?
Jarai (PL) pakon, Chru p9k9n

'other',
pakon
kAn
PR
Cham
(kow')-k?n
-i, Wr.
*bukan 'other', PWMP

(ggp)-g9n -i, Malay bukan 'not', PMalayic
(Blust) *beken 'other, different'. As Blust has pointed out, the first vowel of the
word *buk9n is aMalayo-Chamic
lexical innovation.
Cham
*bula:n 'moon; month', Acehnese buluian, Rade m?an, Jarai (PL) blan,
Jarai (Lee) blan, Chru ea bla:n, N. Roglai ia bila:t, Tsat lug11 phian11, Haroi
p9lian, W. Cham ea plan, PR Cham pilan, Wr. Cham bilan, Malay bulan, PMa?
layic *bulan, PMP *bulan, PAn *bulaN.
*bulow
'hair, body; downy feathers; plant floss', Acehnese

buls9,
Rade ml?u, Jarai (PL) blgu 'plume', Jarai (Lee) bl9u, Chru bl9u, N. Roglai bil9u,
Tsat phi9n, Haroi pallau, W. Cham plau, PR Cham pilow; palow, Wr. Cham
buluw, biluw, Malay bulu, PMalayic *bulu, PMP *bulu.
bugog, Rade maga, Jarai (PL) baga, Jarai
*buga 'flower', Acehnese
Chru
N.
(Lee) baga,
baga,
Roglai bug?, Tsat ga11, Haroi p9gia, W. Cham pagur
f, PR Cham pigu, Wr. Cham big?, Malay bunga, PMalayic *buga(?), PMP *buga.
*buga
cf.
'striped'
'flower',
Rade
maga
'checkered,
Chru
spotted',
bgga, N. Roglai bug?, PR Cham pigu, Wr. Cham big?. The 'striped, checkered'
looks to have evolved from 'flowery' or some such, as the shape of the
meaning
etymon
is the
same
as
that
for
'flower'.
*bugat
spirit; shadow, shade', Rade mag?t, Jarai (PL) bag?t,
Jarai (Lee) bag??, bag?t, Chru baga?, N. Roglai bug??, Haroi pag?a?, W. Cham
pag??, PR Cham pig??, Wr. Cham biguk.
'soul,
*buta 'blind', Acehnese buta, Tsat ta11, PR Cham mita, Wr. Cham mita
as
(same
eye?), Malay buta.
buya, Rade mya, Jarai (PL) b?a, Jarai
*buya 'crocodile', Acehnese
as
same
the
(Lee) bia Note:
bya, Chru bia Note: the same as bya, N. Roglai
biya, Haroi payia, W. Cham paya, PR Cham piya, Wr. Cham biy?, Malay
PMP *buqaya, PAn *buqaya.
buaya,
Appendix
II: The Chamic
287
Lexicon
buisoa, Rade masei, Jarai (PL) pasai, Jarai

*bassy 'iron', Acehnese
N.
Chru
(Lee) pasai,
Roglai pisai, Tsat sai11, Haroi pas?i, W. Cham pasay,
pasai,
PR Cham pith?y; path?y, Wr. Cham bith?i; pas?i, Malay b?si.
Cham
x*?la:t 'open eyes wide', Acehnese bluit, Chru ?la:?, N. Roglai ?la:?, PR
?la?, Wr. Cham ?lak, PMP bulat 'open eyes wide'; MK: Bahnar (AC)
bl?k.
If as it appears, the form is descended
from PMP *bulat, the initial is quite

a PC *?l- cluster in the data. Thus,
The
has
form
the
of
puzzling.
only example
the
obvious
the most plausible account is that
Austronesian
despite
etymology,
was
not
the PC
inherited directly but secondarily borrowed from an Austronesian
source.
*?ow see immediately below *bow

*?uk -1 'head hair', Acehnese
o?, Rade ??k, Jarai (PL) ??k; ?m?k
(Pkly), Jarai (Lee) ??k; ???, Chru ?u:? -1,N. Roglai ?u:?, Tsat ?u?24,Haroi ?u?,W.
Cham ?u?, PR Cham ?u?,Wr. Cham ?uk, PMalayic *bu?uk, PMP *buhuk.
*dada 'chest', Acehnese dada, Rade dah da 'breast' -i (the first initial
is irregular), Jarai (PL) tada (Ouest), Jarai (Lee) tada, Chru har tada 'breast', N.
Roglai dada, Tsat tha11, Haroi cathia, W. Cham tata; ?ata, PR Cham tata, Wr.
Cham tad?, Malay dada, PMP dahdah; MK: PKatuic (DT) *tath_. Note that it is
only inKatuic apparently that the Chamic form has been borrowed, another small
piece of evidence
that there was extensive Chamic
contact in the Katuic
speaking
area.
*dalam
(PL) hal?m
dalap,
Tsat
'inside; in', Acehnese dalam; lam, Rade hl?m lam; elam, Jarai
-i, dalam, Jarai (Lee) dl?m, Chru darlam; lam, dalam, N. Roglai
lam?42,
Haroi
Cham dalam, Malay

Bahnar (AC) lam.
*danaw
(Lee)
danau,
tanaw,
Wr.
PMP
Chru
Cham
l?m;
lim,
danau,
danaw,
Acehnese
N.
Rade
dano,
Roglai
Malay
Cham
tal?m,
PR Cham
pal?m, Wr.
*(d-)alam, PMP *dalem; MK: PNB *l?m,
dalam, PMalayic
'lake',
W.
caliam,
danau
danau
enau,
-n, Haroi
'mere,
pool,
Jarai
caniau
lake',
(PL)
'pond',
PMalayic
Jarai
danau,
PR
Cham
*danaw,
*danaw.
*dara 'girl (c. teenage)', Acehnese dara, Rade era, Jarai (PL) dra, Jarai
(Lee) dra, Chru dra 'teenage girl', N. Roglai dara, Haroi caria, W. Cham tra, PR
Cham tara,Wr. Cham dar?, Malay dara, PMP *daRa 'girl'.
*darah 'blood', Acehnese darah, Rade erah, Jarai (PL) dr?h, Jarai (Lee)
drah, Chru drah, N. Roglai darah, Tsat sia55, Haroi carian, W. Cham tarah, PR
Cham tar?h, Wr. Cham darah, Malay darah, PMalayic *darah, PMP *daRaq.
*dha:n 'branch; bough', Acehnese dhuian, Rade adhan, Jarai (PL) than;
dhan (Ouest), Jarai (Lee) than; dhan, Chru tha:n, N. Roglai tha:t, Haroi than, W.
than, PR Cham than, Wr. Cham dhan, Malay dahan, PMalayic *dahan,
PMP *daqan.
Cham
288
Appendix
*d? 'at', Acehnese

because as a function word
ta-
le
marque
'pour
but
II: The Chamic
Lexicon
di, Rade ti -v (The vowel difference here is probably

itwas subject to a different stress pattern.), Jarai (PL)
pour
Chru
suivi',
ta-,
N.
ta, W.
Roglai
ti
Cham
sa
'first', PR Cham ti,Wr. Cham d?,Malay di, PMalayic *di, PMP *di.
*dihlow
*hlow 'first (go); formerly; before', Acehnese
'formerly';
dilea, Rade el?u, Jarai (PL) hlau, Jarai (Lee) hlau; hlau, Chru la hau, N. Roglai
dihlau, Tsat lau11, Haroi hl?u, W. Cham talau, PR Cham tahlow, an??, Wr. Cham
dahluw, Malay dahulu; dulu, PMalayic *di hulu(?).
*dikit 'few; little', Acehnese
dit -mv, Chru taki:?, N. Roglai tiki:?, Tsat
W.
Cham
PR
Cham
Wr.
Cham takik, Wr. Cham (AC) dikik,
taki?,
ki?42,
taki?,
Malay dikit, PMP *dikit.
*dilah 'tongue', Acehnese dilah, lidah, Rade elah, Jarai (PL) jal?h; gl?h
Jarai
(Lee) jalah -i < *g-, -v, Chru dalah, N. Roglai gilah < *g-, Tsat la55,
(Hd.),
Haroi caliah, W. Cham talah, PR Cham talah, Wr. Cham dalah, Malay lidah;
dilah 'wick, lamp', PMalayic *dilah, PMP *dilaq.
*do:k 'sit; live; stay', Acehnese
dua?, Rade dok, Jarai (PL) do?, Jarai
do:?
Chru
N.
'remain, sit',
(Lee) do?,
Roglai do:?, Tsat tho?42 'sit', Haroi thu?,W.
Cham to?, PR Cham to?,Wr. Cham dauk, Malay duduk, PMalayic *duduk, PMP
*dukduk.
*do:k 'still' cf. 'sit', Rade adok, Chru do:?,W. Cham

po?, Wr.
Cham
dauk,
to? kug, PR Cham
duduk.
Malay
*do:k 'at; in' cf. 'sit', Haroi thu?. This "prepositional" use has devel?
oped from the main verb use, apparently through a verb serialization process.
*dras 'fast; short time', Acehnese draih -v, Jarai (Lee) drah cf. ?eh drah
'dysentery',
'fast
Cham
(coming
drah,
Chru
W.
back)',
Malay
N.
drah,
d?ras,
drah
Roglai
Cham
PMP
prah
*deRes
'short
'short
'fast,
time',
time;
Tsat
early',
sia55,
PR
Haroi
Cham
carian
trah, Wr.
rapid'.
droa 'self; elf.

(-self); body; living beings', Acehnese
*drey 'reflexive
for people', Rade drei in 'we (inclusive)',
Jarai (PL) drai 'corps'; ha 'm?me';
?u 'lui m?me', Jarai (Lee) drai, N. Roglai drai elf. living things, Tsat pi11 se n -v
[(trai -iv)], W. Cham tray 'living beings, animal, elf.', PR Cham tr?y,Wr. Cham
drei; min ?, Malay diri, PMP *diri.
*drsy 'we (incl.)' cf. 'body,
self, Rade drei, N. Roglai labu? drai;

labu? f??,W. Cham tray, PR Cham kh?l tr?y 'we (excl.), Malay diri, Salako
(Adelaar) diri? 'we (inclusive)'.
*dua 'two', Acehnese duwa, Rade dua, Jarai (PL) dua, Jarai (Lee) dua,
Chru dua, N. Roglai dua, Tsat thua11, Haroi thua, W. Cham coa, PR Cham twa,
Wr. Cham du?, Malay
dua, PMalayic
*dua(?), PMP *duha.
Appendix
II: The Chamic
289
Lexicon
duwa ploh, Rade dua pluh, Chru dua

*dua pluh 'twenty', Acehnese
pluh, N. Roglai dua pluh, W. Cham doa pluh, PR Cham twa pl?h, Malay dua.
puluh, PMP *duha ga puluq.
lapan, Jarai (PL) dua rapan -1 (two taken
*dua-lapan 'eight', Acehnese
Chru
Jarai
away from ten),
(Lee) capan -1,
talpan, Tsat pam?42, Tsat (St?bel)
/bad/ i.e., /p?t/, W. Cham
tap?n, PR Cham
talip?n, Wr. Cham
dalipan, Malay
d?lapan.
duroa, Rade erue (m), Jarai (PL) drai, Jarai

*dursy 'thorn', Acehnese
druai
(Lee) drai; trai, Chru
(m), N. Roglai daruai (m), Haroi car?i (m),W. Cham
taruai (m), PR Cham taroy (m),Wr. Cham daruai, Malay duri, PMP *duRi.
gAn,Malay d?ngan, Inscriptional Cham
*dagan 'with; and', Acehnese
PWMP
dengan,
*degen.
*dapa 'armspan', Rade pa, Jarai (PL) t?pa, Jarai (Lee) tapa, Chru tapa,
tupa, Tsat pa11, Haroi capa, W. Cham tapa, PR Cham tapa, Wr. Cham
tapa, Malay d?pa, PMP *depa.
N. Roglai
gatai, Rade kat?l, Jarai (PL) ga?? (Est), Jarai

*gatal 'itchy', Acehnese
(Lee) katal, Chru katal, N. Roglai katan, Haroi kat?l 'to itch', W. Cham kat?l,
PR Cham kat?l, Wr. Cham katal, Malay gatal, PMP *gatel.
gigoa, Rade egei, Jarai (PL) tagai, Jarai (Lee)
*gigsy 'tooth', Acehnese
tagai -i, Chru tagai, N. Roglai digai, Tsat (hu11) khai11, Haroi cakhii, W. Cham
takay, PR Cham takfy, Wr. Cham tag?i, Wr. Cham (AC) tag?i ,Malay gigi,
PMalayic *gigi. This root is confined to languages inwestern Indonesia.
*gila
Cham
kla
'foolish',
'crazy;
mad',
Chru gala
PR Cham
'foolish', N. Roglai
kila, Wr.
Cham
gila pito:? 'foolish', W.
gil?, Malay
gila,
cf. PMP
*ila,
*kila 'restless; wild' MK:

PSB (Efimov) *gala:.
;
'to
Acehnese
roll',
gulog, Jarai (PL) kal?g; (Hd.) ba-galug, Jarai
*gulug
N.
Chru
-v,
(Lee) glug,
Roglai pa-guluk; ta-guluk, Haroi ta-kal?g; kal?g,
pa-rlag
PR Cham kal?g, Wr. Cham galug, Malay gulung, PMP *gulug. With prefixation:
In Jarai, Chru and N. Roglai, the form seems to occur with a causative prefix, not
tomention also with a ta- prefixed version inN. Roglai and Haroi.
*gatak 'sap; resin', Rade kat?k, Jarai (Lee) kat?k, Chru kata?, N.
Roglai kata?, Haroi kat??, W. Cham kat??, PR Cham kat??, Wr. Cham katak, cf.
Achnese
guitah, Malay
g?tah 'sap; latex', PWMP *geteq 'tree sap'. The
PWMP
PWMP
and the Malay all reflect an earlier final *-h (<

(Blust), the Acehnese,
while
all
the
reflect an unexpected
Chamic
forms (except Acehnese)
*-q),
final *-k; theAcehnese

ismost likely borrowed from theMalay.
*habow 'ashes', Acehnese
absa, Rade habau, Jarai (PL) habau, Jarai
habau
N.
Chru
habau,
-vl,
(Lee)
Roglai habau, Tsat pha11, Haroi aphiau, W.
Cham pau, PR Cham hap3w; pow, Wr. Cham habuw, Malay abu, PMalayic
290
II: The Chamic
Appendix
*habu, Malay
abu, PMP *qabu; MK:
PSB
(Efimov)
*bu:h, PKatuic
Lexicon
*[s/h]a?bah,
*ha?bah.
arag -m, Rade had?g, Jarai (PL) had?g,

*hadag 'charcoal', Acehnese
Jarai (Lee) had?g, Chru hadag, N. Roglai hadak, Tsat thaig?42, Haroi athiag, W.
Cham tag, PR Cham hat?g; pag,Wr. Cham hadag; dag, PMP qajeg; MK: Bahnar
(AC) arag.
*halow
alea, Rade hl?u, Jarai (PL) hlau, Jarai (Lee)

'pestle', Acehnese
N.
Haroi
Chru
hlau,
halau, W. Cham hlau, PR Cham hlow,
halau,
Roglai halau,
Wr. Cham hluw, Malay alu, PMalayic *halu, PMP *lalu.
ate, Rade tie, Jarai (PL) ha tai (Ouest), Jarai
*hatay 'liver', Acehnese
(Lee) hatai, Chru hatai, N. Roglai hatai, Tsat tad?42, Haroi at?i -i,W. Cham tai,
PR Cham hatay, Wr. Cham hatai, Malay hati, PMalayic *hati, PMP *qatay, PAn
*qaCay.
'rattan (generic?)', Acehnese

awe, Rade hawie, Jarai (PL)
*haway
huai
huai
Chru
Jarai
'rod, rush,
hawai, N. Roglai hawai,
hawei;
(Lee)
(Hd.),
Tsat va??42, Haroi hawai, W. Cham haway -f, PR Cham haw?y, Wr. Cham
haw?i, Iban ui?, PMalayic *(h)ui, PMP *quay, PAn *quay.
*hijaw 'green; blue', Acehnese
Wr.
Cham hajaw; caw, Malay hijau,
?aw,
*hitam 'black', Acehnese
itam,
tain?42, Tsat (St?bel) /d?n/ i.e., /t?n/, PR
Malay
ijo, W. Cham
?au, PR Cham
ha?aw;
PMalayic *hijaw.
Jarai (PL) hatam j?? 'bleu-noir', Tsat
Cham hat?m; t?m, Wr. Cham hatam,
hitarri, PMalayic *hitam, PMP *qitem.

*hlow see *dihlow
ubi -f, Rade habsi, Jarai (PL)

*hubey 'taro; tuber; yam', Acehnese
N.
habai, Jarai (Lee) habai, Chru habai,
Roglai habuai (m), Tsat phai11, Haroi
PR
W.
Cham
Cham
pay,
hap?y; p?y, Wr. Cham hab?i, Malay ubi,
aphui (m),
PMP
*qubi.
Also
cf.
'a
plant'.
uduiag, Rade hadag, Jarai (PL)

*huda:g 'shrimp; lobster', Acehnese
N.
Jarai
Chru
(Lee) hadag,
h3da:g,
hada:g,
Roglai huda:k, Tsat la11 tha:gn, Haroi
PR
Cham
ni?
Cham
W.
tag,
hat?g; tag, Wr. Cham hudag; dag, Malay
athiag,
PMP
hudang; udang,
*qudag.
*hudsy 'after; behind', Chru hadai 'after', W. Cham hatay 'then,
PR Cham hat?y; t?y, Wr. Cham had?i, cf. Malay kemudian, PMalayic
PMP
*ma-udehi; MK: Bahnar (AC) hadoi.
*hudi?,
*hudip 'live, alive', Acehnese udep, Rade had?p, Jarai (PL) hadip, Jarai
(Lee) hadip, Chru hadiu?, N. Roglai hadiu?, Tsat thiu?42, Haroi ath?p, W. Cham
already',
tiu?, PR Cham hat?w?; tlw?, Wr. Cham hadiap, Malay hidup, PMalayic *hidup,
PMP *qudip, *quDip, PAn *qudip.
ujuian, Rade ha jan, Jarai (PL) haja:n, Jarai
*huja:n 'rain', Acehnese
(Lee) najan, Chru hajam, N. Roglai huja:t, Tsat sain11, Haroi asian, W. Cham
II: The Chamic
Appendix
291
Lexicon
can, PR Cham
ha?an; can, Wr. Cham hajan; jan, Malay

hujan, PMalayic
PMP
*hujan,
*quzan, PAn *quzaN.
*hulat 'worm', Acehnese
ulat, Rade hlu?t (m) 'caterpillar', Jarai (PL)
hl?t, Jarai (Lee) hl?t; hlu?t (m), Chru hala?, N. Roglai hula?, W. Cham hl?? 'cat?
erpillar', PR Cham hal??, Wr. Cham halak, Malay
*qulej, PAn *qulej.
*hulun
I (polite)', Acehnese
servant;
'slave;
lontuwan
wan,
ulon,
(polite);
(sg./pl.)
Ion-,
ulog-,
*hulat, PMP
ulat, PMalayic
ulon,
Ion, ul?g, ulogtu
ulontuwan-,
lontuwan-;
-Ion,
-log, Rade hl?n, Jarai (PL) hlun, Jarai (Lee) hl?n, Chru halun, N. Roglai hulut,
Haroi halun, W. Cham hl?n hl??; hul?n T,
PR Cham hal?n, Wr. Cham halun,
Malay ulun, PMP *qulun.
*huma 'cultivated field', Acehnese umAg, Rade hama 'swidden field',
Jarai (PL) hamua (m), hama, Jarai (Lee) hama; hamua, Chru hama 'paddy
field', N. Roglai hum? 'wet field', Tsat ma33, Haroi hamia, W. Cham hami, PR
Cham hamu -v, Wr. Cham ham?, Malay huma 'dry field', 'swidden', PMP
*quma,
PAn
*qumah.
uroa; mata uroa, Rade hrue (m); yag

*hursy 'day; sun', Acehnese
hrue, Jarai (PL) hrai, Jarai (Lee) hrai, Chru harai, N. Roglai hurai; ia hurai, Tsat
zai33, Haroi harii (m),W. Cham hray; ea hray, PR Cham har?y, Wr. Cham har?i,
Malay hari, PMalayic *hari, PMP *waRi.
*huta:n
utuian, PR Cham hatan; tan, Wr.

jungle', Acehnese
hatan; tan, Malay hutan; utan, PMalayic *hutan, PMP *qutan 'small,
wild herbaceous plants; scrub-land, bush'. As Blust (p.c.), points out the mean?
'forest;
Cham
ing
'forest'
is secondary,
*doh
'far,
but
shared
Chru
distant',
in both Malay
don
and Chamic.
PR
'near',
'loin, ?loign?', Wr. Cham hadauh; dauh, Malay

*zauq. The PC initial is quite unexpected.
*ita
'we
Acehnese
(incl.)',
gui-ta-?oa,
*guita < *kita plus ftoa 'this'; ta-; -tui(h),

tan-phi55;
ta33
'we
ta33
(excl.)';
za:k33
'we
Cham
hatfoh
don
'?loigne';
jauh, PMalayic
*jauh, PMP
ta-?oa
(neutral)
(sg./pl.)
Jarai (PL) Big ta 'we (excl.)',

(excl.)',
Haroi
ata
'we
(incl.)',
<
Tsat
PR
Cham khol ita,Wr. Cham it?, Malay kita, PMalayic *kita?, PMP *kita.
*ja:k 'invite', Rade jak, Jarai (Lee) jak, N. Roglai ja:?, Haroi sia?, PR
Cham ?a?,Wr. Cham jak, Malay ajak; MK: Bahnar (AC) j?k.
Roglai
Malay
*jahit 'sew', Rade jhit, Jarai (PL) (est), Jarai (Lee) si?; sit, Chru si:?, N.
chi:?, Tsat si?24,Haroi sei?,W. Cham ?hi?, PR Cham ?hi?,Wr. Cham jhik,
jahit, PMalayic *jahit, PMP *zaqit; MK: Bahnar (AC) cit.
*jala:n
house';
'road;
path',
-v < Malay,
Acehnese
-lu?an
'yard;
space
in
front
of
jalan
jalam, N. Roglai jala:t, Tsat
?alan, Wr. Cham jalan, Malay
the
elan, Jarai (PL) jalan, Jarai (Lee) jalan, Chru

lam11, Haroi calian, W. Cham ?alan, PR Cham
jalan, PMalayic *jalan, PMP *zalan.
Rade
292
Appendix
II: The Chamic
Lexicon
*jarum 'needle', Acehnese

jarom, Rade er?m, Jarai (PL) ?j?rum; jr?m
(Pk), Jarai (Lee) jr?m -a, Chru jrum ?, N. Roglai jump, Tsat sun11, Haroi car?m,
W. Cham ?rum -v, PR Cham car?m, Wr. Cham jarurri,Malay jarum, PMalayic
*jarum, PMP *zaRum; MK: PNB *jar?m, PKatuic (DT) *jar_m.
*judsy 'after,
to Highlands
Restricted
later; last', Rade edsi, N. Roglai judai, Haroi cath?i.

Chamic plus Haroi. [This form looks to be a variant of
but with an unexplained initial]

*jaha:t 'bad; wicked', Acehnese
juihuit; jhuit -v, Rade jhat, Jarai (PL)
s?t, Jarai (Lee) sat, Chru ja ha:?, Tsat sa:?24,Haroi s?t, PR Cham ?ha?,Wr. Cham
jhak, Malay jahat, PMalayic *jah(aa)t, PMP *zaqat; *zaqet.
*hudsy
'after; behind',
*kaju 'spear; lance', Rade kaju, Jarai (Lee) taju -i, Haroi kasu, Malay
kujur.
?), Jarai (PL) takai, Jarai

gaki -v (< Malay
*kakay 'foot', Acehnese
Chru
N.
Tsat
takai,
(Lee) takai,
kad?42, Haroi cak?i, W. Cham
Roglai takai,
takai, PR Cham takay, Wr. Cham takai, Wr. Cham (AC) kakai, takai, Malay
kaki, PMalayic *kaki, PMP *qaqay. There is a problem with the relationship
the PMP form, already noted by Blust in his dictionary. Note that the *k
the expected reflex is
reflex of PMP *q- in both PC and PMalayic is unexpected:
/h-/, not /k-/. It could be the case that, despite the similarities, the two forms are
with
it could be that the PMalayo-Chamic
simply unrelated,
sound,
or
it could
be
something
else?perhaps
the MK
an irregular
represents
velar
prefix,
which
some?
body parts. In any case, the form is valuable for subgrouping; this
that is, in both
unexpected form is shared in both branches of PMalayo-Chamic,
and
PC.
PMalayic
times marks
see
*kala:g
*rala:g
*kalsy 'dig', Rade klsi, Jarai (PL) kl?i, Jarai (Lee) klai, Chru klai -i, N.
Roglai kalai, Haroi kal?i, PR Cham kaley, Wr. Cham kal?i, Malay kali 'ditch';
gali -i, PMalayic *kali, PMP *kali.
*kam?;
'we
*kamsy
(ex.)',
Acehnese
kamoa;
mu?-;
-mru(h),
Rade
hamsi, Jarai (PL) ?ig g?mai 'we (incl.)'; Big mai (Hd.) 'we (incl.)'; Big hamai
(S) 'we (incl.)', Jarai (Lee) gamai -iv, N. Roglai labu? kam?n -v?, Tsat mi33,
Haroi kamei -v h?i -v (grammatical particle); kamen (-h?i); k?me h?i -v
(grammatical particle),
*kami, PMP *kami.
*kanam
'very
dark',
N.
PR Cham kami, Wr. Cham kami, Malay
-f 'dark', Jarai (PL) kanam,
Roglai
kanap
-f, Haroi
kan?m,
kami, PMalayic
Jarai (Lee) kanam, Chru kanam

W.
Cham
kan?m
'dark
of moon',
PR Cham kan?m, Wr. Cham kanam, Malay k?lam, PMP *kelem; MK: PKatuic
*kanham 'dark'. [The Katuic form is apparently borrowed].
Appendix
II: The Chamic
293
Lexicon
*kapa:l 'thick', Rade kapal, Jarai (PL) kSpal, Jarai (Lee) kapal, Chru
N.
Roglai kapan, Tsat pa:nn -t, Haroi kapal, W. Cham kapal, PR Cham
kapa:l,
kapal, Wr. Cham kapal, PMP *kapal.
*kapit
'squeeze;
pinch,
press'
maybe
the
same
as
*kapit
'close',
Jarai
(Lee) kap?? ?, Haroi cap?t -vf, PR Cham tapi?,Wr. Cham tapik, PMP *ka(m)pit
'hold together, squeeze',
*ga(m)pit 'to hold together'.
*karam
krom, Chru kram
Wr.
kar?m,
Cham
'to hatch', Acehnese karom, Rade kr?m 'sitting hen', Jarai (PL)
'sitting hen', N. Roglai karap, W. Cham kar?m, PR Cham
kararri,
Malay
ram.
k?ram,
*kawat, *kuat 'wire', Rade kaw?t, Jarai (Lee) ku?t, Chru kuat -f, Haroi
kaw?t, W. Cham kaw?t -vf, Malay kawat, PWMP *kawad.
*kayow 'tree; wood', Acehnese kayea, Rade kay?u, Jarai (PL) kayau,
kayau, Jarai (Lee) kayau, Chru kayau, N. Roglai kayau, Tsat (phun33) zau33,
Haroi kay?u -vr, W. Cham kayau, PR Cham kaySw, Wr. Cham kayuw, Malay
kayu 'stick; wood', PMalayic *kayu? 'stick, wood', PMP *kahiw.
*klow
'three', Acehnese
lhea, Rade tl?u, Jarai (PL) klau, Jarai (Lee)
klau, Chru klau, N. Roglai tlau, Tsat kiu33, Haroi tl?u, W. Cham klau, PR Cham
klow, PMalayic *talu, PMP *telu.
*klu 'testicles
Jarai (PL) tiflu, Chru klu 'testicles', PR
(of animal)',
Cham
t?lur 'egg', PMalayic *talur 'egg', PMP *qateluR, *qiteluR

klu, Malay
'egg', PAn *qiCeluR 'egg'.
*kow 'I (familiar)', Acehnese
kea; ku-; -ku(h), Rade k?u, Jarai (PL)
kau 'de inf?rieur ? son chef direct', Jarai (Lee) kau, Chru kau, kami, N. Roglai
kau,
Tsat
*aku,
kau33,
PMP
Haroi
k?u, W.
Cham
kau,
PR
Cham
k?w,
Malay
aku,
PMalayic
*aku.
*kra 'monkey', Rade kra, Jarai (PL) kra, Jarai (Lee) kra, Chru kra, N.
Tsat kia33, Haroi kra, W. Cham kra, PR Cham kra, Wr. Cham kr?,
kra,
Roglai
The loss
Malay k?ra, PWMP (Blust, p.c.) *keraq 'the chattering of monkeys'.
of final -q is irregular, but shared with both Malay and Chamic.
*kukow
'claw; fingernail', Acehnese

guksa, Rade ka k?u, Jarai (Lee)
nor
reflects
neither
*kJarai
(initial
(PL) takau < *t-, Chru karkau,
*t-),
N. Roglai kukau, Haroi kaic?u, W. Cham kakau, PR Cham kak?w, PR Cham
cakau
cak?w
(initial reflects neither *k- nor *t-), Wr. Cham kakuw, Wr. Cham (AC)
kukau, Malay kuku, PMP *kuhkuh. Note the sporadic nature of the changes of
the presyllable-initial
*k-. One contributing factor is interaction with the *k
body part prefix found throughout MK (cf. Smith 1975).
*kulit 'skin', Acehnese kulet, Rade kl?t, Jarai (PL) kalit, Jarai (Lee) klit,
Chru kali:?, N. Roglai kuli:?, Tsat li?24 -i; lo?24 -i, Haroi kalei?, W. Cham kli?, PR
Cham kali?, Wr. Cham kalik, Malay kulit, PMalayic *kulit, PMP *kulit 'skin;
bark'.
294
Appendix
*ku?it
kuflst
Acehnese
tumeric',
'yellow;
II: The Chamic
Rade
'tumeric',
Lexicon
Jarai
kan??,
Tsat
(PL) kan??, Jarai (Lee) kaf?i?, Chru kafi?:?, N. Roglai ku?i:? 'yellow-orange',
?i?24; gi55, Haroi ka?ei?, W. Cham ka?i? 'orange, safron', PR Cham ka?i?, Wr.
Cham kaftik, Malay kunyit 'tumeric', PMalayic *kunit 'yellow, tumeric', PMP
'tumeric'.
*kunij
*kura 'tortoise; turtle', Rade krua (m) 'turtle', Jarai (PL) kr?a (Ouest),
Jarai (Lee) krua (m), Chru kra, N. Roglai kura, Haroi kroa (m), PR Cham kara,
Wr.
Cham
kar?,
Malay
kura.
*kura:g 'less; insufficient',
W. Cham kurag, Acehnese
kuruiag, Malay
kurang.
'head louse', Acehnese

gutsa, Rade kat?u, Jarai (PL) katau,
Jarai (Lee) katau, Chru katau, N. Roglai kutau, Haroi kat?u, W. Cham katau, PR
Cham katow, Wr. Cham katuw, Malay kutu, PMalayic *kutu, PMP *kutu.
*kutow
*ka-
'to,
for
at',
(goal),
Acehnese
kui-,
Rade
ka-,
Jarai
(PL)
ka-
'pour
Jarai (Lee) [ka], Chru ka- 'to, concerning', N. Roglai hia ga;
marque le motif,
ga, Haroi ka-, PR Cham ka, Wr. Cham ka, Malay k?-, kepada, PMP *ki- 'to, for
(goal)'.
*kantut
'fart;
flatus
Acehnese
ventrus',
toh
guintat
-v,
Chru
katu:?,
Haroi katou?; tout, Malay k?ntut, PWMP *ka-qetut 'the noise of flatuence',
PMP *qe(n)tut.
If the Malay and the PC were derived directly from the PMP
have /h-/ as their initial; the actually occurring /k-/ reflects the
would
*qetut, they
*ka-
prefix,
with
*la?ur
the nasal
'coconut
an
expected
palm',
secondary
Acehnese
boh
development.
u, Chru
la?u,
N.
Roglai
la?u,
Haroi
la?u, W. Cham la?u, PR Cham liu, Wr. Cham li?, Malay nyor, PMalayic
PMP
*niur,
*niuR, PAn niyuR.
*laba:t 'walk, go', Rade ebat, Jarai (PL) rabat, N. Roglai luba:?, Tsat
pha:?42,W. Cham kapa? ?-i, PR Cham lipa?; lapa?, Wr. Cham libak; labak, PMP
*lampaq?.
'fall down; drop anchor', Rade ebuh, Jarai (PL) rab?h, Jarai
Chru
labuh, N. Roglai labuh, Tsat phu55, Haroi laph?h, W. Cham
(Lee) rabuh,
PR
labuh 'drop anchor',
Cham
lapuh,
lipuh; lapuh, Wr. Cham libuh, Malay
PMP
PMalayic *labuh,
*labuq.
*labuh
lakoa, Rade eksi, Jarai (PL) rakai,

*lakey 'male; person', Acehnese
Jarai (Lee) rakai -i; cakai -i, Chru lakai 'male', N. Roglai lakai, Tsat kai33,
Haroi lak?i, W. Cham lakay, PR Cham lik?y; lak?y, Wr. Cham lak?i, Malay lak
ilaki, PMalayic *laki
(-laki), PMP *laki.
-v 'sesame', Rade egu, Jarai (PL) raga; baga (Hd.), Jarai (Lee)
raga, N. Roglai lag?, Haroi laga; lag?u -v,W. Cham lagi, PR Cham ligi; lagi,Wr.
Cham ligi; lagi, PMP *lega; MK: PSB (Efimov) *laga:.
*lanah 'pus', Jarai (PL) ran?h, Jarai (Lee) ranah, N. Roglai lan?h, Tsat
*laga
la11 na55, Malay
nanah,
PMP
*nanaq.
II: The Chamic
Appendix
295
Lexicon
l?get, Rade eg?t, Jarai (PL) git (Pk), Chru tasi:?

*lagit 'sky', Acehnese
Tsat
N.
lagi:?,
gi?24, Haroi lagi?, W. Cham lagi?, PR Cham ligi?;
Roglai lag!:?,
lagi?,Wr. Cham ligik; lagik, Malay langit, PMalayic *lagit, PMP *lagit.
*lapa 'hungry', Rade epa, Jarai (PL) r?pa, Jarai (Lee) rapa, Chru lapa,
N. Roglai lapa, Tsat pa33, Haroi lapa, W. Cham lapa, PR Cham lipa; lapa, Wr.
Cham
lipa; lapa, Malay lapar, PMP *lapaR.

*lasey 'rice (cooked)', Rade esei, Jarai (PL) asai -i, Jarai (Lee) ?asai -i,
Chru la sai, Haroi la soi,W. Cham las?y, PR Cham lith?y, Wr. Cham lis?y, Malay
nasi.
'withered; faded', Acehnese

layea 'faded', Chru layau, W.
*layow
PR
Wr.
Cham
Cham
layu
layuw; liyuw, Malay
layau,
liy3w, layow,
'faded; withered', PMalayic *layu, PWMP *laqeyu.
*lima 'five', Acehnese
limAg, Rade em a, Jarai (PL) rama, Jarai (Lee)
Cham
rama, Chru lama, N. Roglai lum?, Tsat ma33, Haroi lamia, W. Cham lami, PR
Cham limi; lami, Wr. Cham limi; lami, Malay
lima, PMalayic *lima, PMP
*lima.
*limpa:n 'centipede', Acehnese

limprxian, Rade epan, Jarai (PL) rapan,
Jarai (Lee) rapan, Chru lapa:n, N. Roglai lupa:t, Tsat a11 pam11, Haroi lapan, W.
Cham lapan, PR Cham lipan; lapan, Wr. Cham lipan; lapan, Malay
(ha)lipan,
PMP
*qalu-hipan,
*lintah
PAn
*qalu-Sipan.
-i 'water
leech', Acehnese
lintah, Rade katah -i, Jarai (PL)
Jarai
Chru
N.
retan,
latah,
(Lee) ratah,
Roglai ritah -i, Haroi lat?h, W. Cham
latah, PR Cham lit?h; lat?h, Wr. Cham latah, Malay lintah 'horse leech', PMP
*qali-metaq,
PAn
*qali-meCaq;
MK:
Bahnar
(AC)
latah,
ratah.
*lipih 'thin (material)', Acehnese

lipeh, Rade epih, Jarai (Lee) rapih,
Chru lapih, N. Roglai lupih, Tsat pi55, Haroi lapeh, W. Cham lapih, PR Cham
lip?h; lap?h,Wr. Cham lipih; lapih, Malay nipis, PMalayic *nipis, *m/ipis, PMP
*nipis.
*liya 'ginger', Acehnese

haliya, Rade eya, Jarai (PL) raya, rala, Jarai
(Lee) raya, Chru lia (note: the same as lya), N. Roglai riya -i, Tsat za33, Haroi
layia, PR Cham liya; laya; kanrog-riya -i,Wr. Cham liy?; laya; ganraug riy?,
Malay halia, PMP *laqia; MK: Bahnar (AC) lia.
Roglai
Malay
laka
*luba:g 'hole; pit', Acehnese kubag -i 'wallowing hole', Chru laba:g, N.

luba:k, W. Cham lapag, PR Cham lipag; lapag, Wr. Cham libag; labag,
lubang.
*luka
'bite,
lik?, Malay
'wound, scar', Acehnese

luka, Rade eka, Jarai (Lee) raka, Chru
W.
Haroi
Cham
sore',
laka,
laka, PR Cham lika-lik?h, Wr. Cham
PMP
*luka.
luka,
*lumpey 'to dream', Acehnese

Jarai (Lee) rapai, Chru lapai, N. Roglai
lumpoa, Rade epei, Jarai (PL) rapai,

lupai, Tsat pai33, Haroi lap?i, W. Cham
296
Appendix
lapay, PR Cham lip?y; lap?y, Wr. Cham

*m-impi, PMP *h-in-ipi; *hipi; *h-um-ipi.
*labsh 'more; surplus', Acehnese
lip?i; lap?i, Malay
rab?h, N. Roglai
lubeh, Haroi
labaih, Malay l?bih, PMalayic

*lama? 'fat, grease,
II: The Chamic
mimpi,
Lexicon
PMalayic
ruibsh -v, Rade ebsh dlai, Jarai (Lee)

PR
Cham
laphth,
lipsh; lap?h, Wr. Cham libaih;
PMP
*lebih,
*lebiq.
luim a?, Rade em??, Jarai (PL)

oil', Acehnese
rama?, Jarai (Lee) rama? -1, Chru lama?, N. Roglai lum??, Tsat ma?42, Haroi
lamia?, W. Cham lama?, PR Cham limi?; lami?, Wr. Cham lamik; limik, Malay
l?mak, PMalayic *lamak, PMP *lemak; MK: PMnong *ram?q, *ramaq, Bahnar
(AC) rama; lama.
*lan 'earth', Acehnese
Lvn 'silt, Rade l?n, Jarai (PL) Ion, Jarai (Lee)

13n, Chru Ian, N. Roglai lat, Haroi Ion 'ground, floor; country', PR Cham l?n,
Wr. Cham Ian, PMP *talun 'fallow land, secondary forest', PAn *Calun.
lui sog, Rade es?g, Jarai (PL) ras?g, Jarai
*lasug -i 'mortar', Acehnese
risuk
Chru
N.
-i, Haroi lasog, W. Cham lasug, PR
(Lee) ras?g,
lasug,
Roglai
Cham lith?g; rath?g -i; lath?g, Wr. Cham lisug; rasug; lathug, Malay
l?sung,
PMP
PMalayic *lasug,
*lesug.
*ma-alas
'lazy', Acehnese malaih, Rade alah, Jarai (Lee) ?alah, Chru

alah, N. Roglai alah, Haroi al?h, W. Cham Iah, PR Cham alah mit?h, Wr. Cham
alah mitah, Malay malas, PMalayic (Blust) *males.
*ma-boh 'to lay egg', Jarai (PL) maboh, Jarai (Lee) maboh, Chru boh,
Haroi paph?h, PR Cham mipoh, Wr. Cham mibuah. Related to 'fruit; egg; small
round
Tsat
main,
below.
object'
*ma?in
'play',
?in33, Haroi
ma?en,
PMP
Acehnese
W.
Cham
rmu??n,
maTin,
Chru
PR
ma?in
Cham
'fun',
miin,
Wr.
N.
Cham
Roglai
miin,
ma?in,
Malay
*ayam.
*mabah
see
*babah
*mabu? 'drunk, intoxicated', Acehnese mabo?, Chru mabu?, N. Roglai

babu?, W. Cham map??, Malay mabuk, PMP *ma-buhek, PAn *(ma-)buSek.
*mah?w 'thirst; desire', Rade mahau, Jarai (PL) m?ha? -f, Jarai (Lee)
mahau, Chru mah?u, N. Roglai mah?u, Tsat hau?24 -t, Haroi mah?u, W. Cham
mahu 'desire; thirst', PR Cham mihu, Wr. Cham mih?, Malay mahu 'want'.
*mahirah 'red', Acehnese mirah, Rade hrah, Jarai (PL) mri?h (Est),
Jarai (Lee) mriah -m, Chru mariah, N. Roglai mariah, Tsat za55, Haroi mareah,
W. Cham mareah, PR Cham miry?h, Wr. Cham miriah, Malay m?rah, PMalayic
*(ma-)irah,
PMP
*malam
*ma-qiRaq.
'night;
evening',
Acehnese
malam
'night',
Rade
mlam
-1, Jarai
(PL) ml?m, Jarai (Lee) ml?m, Chru malam, N. Roglai malap, Haroi malam -1,
W. Cham mal?m, PR Cham mil?m, Wr. Cham milarri, Malay malam
'night',
dark'.
*halem
PMP
PWMP
*m-alem
*ma-lam,
(Blust)
'night,
'night',
PMalayic
Appendix
II: The Chamic
297
Lexicon
*malow
Jarai (PL) malau,
'blush; ashamed;
shy', Acehnese malsa,
Jarai (Lee) miau, Chru malau alah, N. Roglai malau, Haroi ml?u, W. Cham
malau, PR Cham milSw, Wr. Cham miluw, Malay malu, PMalayic *malu.
*mamah 'chew', Acehnese mamAh, Rade mamah, Jarai (PL) m?h, Jarai
(Lee) mamah, Chru bamah ?*mu-, N. Roglai mum?h, Tsat ma55, Haroi mamah,
W. Cham mamih, PR Cham mimih, Wr. Cham mimih, Malay mamah, PMalayic
PMP
*mamah,
*mamaq.
*mamih
'sweet', Acehnese mamen, Rade mamih, Jarai (PL) mih, Jarai

(Lee) mamih, N. Roglai mum?h, Tsat mi55, Haroi mam?h, PR Cham mim?h 'ai?
grelet', Wr. Cham mimih 'aigrelet', PMP *mamis.
*ma?am -1 'weave; twill', Acehnese maftum, Rade ma?am, Jarai (PL)
m?ft?m, Chru mafta:m -1 'knit', N. Roglai ma?am, Haroi ma?iam, W. Cham
PR Cham minim, Wr. Cham mi?im,
ma?im,
weave
(mats,
PMP
*a?em,
*ma-aftam
'plait,
baskets)'.
*maney 'bathe', Acehnese manoa, Rade manei, Jarai (PL) menai, Jarai
(Lee) manai, Chru manai, N. Roglai man? -v, Haroi mn?i < *-ai, W. Cham
manay, PR Cham miney, Wr. Cham min?y, Malay mandi.
*manu? 'chicken; fowl', Acehnese
mano?, Rade man??, Jarai (PL)
man??, Jarai (Lee) man??, Chru man??, N. Roglai man??, Tsat nu?24, Haroi
man??, W. Cham man??, PR Cham min??, Wr. Cham minuk, Malay manuk
'bird', PMalayic *manuk 'chicken', PMP *manuk.
mai,
r?i
-m, W.
mari-lah
Malay
masam,
'sour;
(Lee)
mai
Cham
'come
*masam
Jarai
Jarai (PL) rai, Chru marai, N. Roglai
'come',
*maray
Haroi
zaii?42,
PMP
here',
Cham
mai
Wr.
-r-, Tsat
Cham
N.
ea masam,
Rade
masam,
ia masam,
Chru
mai,
miray,
mirai;
*maRi.
*um-aRi;
Acehnese
vinegar',
masam,
-r-, PR
Jarai
ia masap,
Roglai
Tsat
(PL)
sain?42,
Haroi masam, W. Cham mas?m, PR Cham mith?m, Wr. Cham mitharri, Malay
masam, PMalayic
*asam, PMP *ma-esem. The word 'vinegar' is simply the
word
'sour'
the word
plus
*masin
'salted;
for water.
salty',
Acehnese
masen,
Rade
mas?n,
Jarai
(PL)
m as?n
(Hd.), Jarai (Lee) masin, Chru masin in 'pickled fish', N. Roglai masit (?), Tsat
sen?42 -vft, Haroi masen, W. Cham mas?n in 'fish juice', PR Cham mithin, Wr.
Cham misin, Malay masin, PMP *ma-qasin.
*masuh
-1 'tease
Malay
each
musuh
'fight (war)', Chru masuh, N. Roglai masuh, Haroi phia-masoh
other',
W.
Cham
masruh
-r, PR
Cham
mithuh,
Wr.
Cham
misuh,
'enemy'.
*mata
'eye', Acehnese mata, Jarai (PL) mata, Jarai (Lee) mata, Chru
mata, N. Roglai mata, Tsat tig33 ta33, Haroi mata, W. Cham mata, PR Cham
mita, Wr. Cham mita, Malay mata, PMalayic *mata, PMP *mata; MK: PNB
*m?t, PMnong *m?t, PSB (Efimov) *mat, PKatuic *mat. The PMP and theMK
sets of forms date back to their respective proto-languages.
298
II: The Chamic
Appendix
Acehnese
'die',
*matay
Chru
mate,
N.
matai,
Tsat
matai,
Roglai
Lexicon
taii?42,
Haroi pathia ?, W. Cham matai, PR Cham mitay, Wr. Cham mitai, Malay
PMalayic *mati, PMP *ma-atay.
*matow
<
patau
'son-in-law';
Chru
*p-,
martau,
PR
layic *b-in-antu
'child-in-law'.
(Blust,
Cham
Wr.
mitow,
Cham
W.
mat?u,
mituw,
(Lee)
'son/
PMa?
m?nantu,
Malay
in western
to languages
Confined
Jarai
matau
Cham
Indonesia
p.c.).
*miaw
Tsat
matau,
Roglai
daughter-in-law',
Haroi
m at?u,
Rade
'child-in-law',
*patow
N.
mati,
Acehnese
'cat',
m eau,
Haroi
miau33,
Rade
mia,
W.
Cham
miyaw -v; MK: PMnong *mi?w, PSB

this word requires no further comment.
Chru
mieo,
PR
magiau,
miyaw
Roglai
mi?u,
-v, Wr.
Cham
The imitative nature of
*ms:w.
(Efimov)
N.
miau,
Cham
*mi?a:k 'oil', Acehnese mi?ui?, N. Roglai ma?a:? h?, Tsat f?a?24,W.

Cham maf?i?, PR Cham mini?, Wr. Cham mif?ik, Malay minyak, PMP *mef?ak
is another shared irregularity in
'fat, grease, oil'. The -i- inMalayo-Chamic
Malayo-Chamic.
*minum; *minam 'to drink', Acehnese minom, Rade man?m, Jarai (PL)
maf??m; f??m (Hd.), Jarai (Lee) maf?um, Chru ma?um, Haroi maft?am, W. Cham
ma?um, PR Cham mi??m, Wr. Cham mi?urri, Malay minum, PMalayic *inum,
PMP
*um-inum.
Acehnese
'yesterday',
*mubursy
brai, Chru kabruai
baroa,
(m), W. Cham maproi
mabrue
Rade
(m),
(m), Malay
Jarai
(Lee)
'the good old
bahari
days'.
*muda
Cham
mida,
Malay
m?ng-,
tha11,
prefix',
PMalayic
*mantah
Jarai (Lee) mada,

Haroi
PMP
muda,
'verb
*maMalay
Tsat
intestine',
'raw;
Acehnese
tender',
unripe;
Jarai (PL) mada,
intestine',
'small
'young;
mathia,
*uda,
W.
Rade
muda,
mada
PR
mata,
Cham
Tsat
ta55, Haroi
Acehnese
Jarai
mui-/mu-,
PWMP
*mag-.
green;
unripe',
Acehnese
W.
in
Wr.
mita,
*m-uda.
*mag-,
m atah,
'small
mida
Chru mada, N. Roglai

Cham
ma-,
muintah,
Wr.
Rade
Cham
Cham
mata
-m,
PR
Cham
mit?h,
ma-,
m atah,
(PL) m atah, Jarai (Lee) m atah, Chru m atah -n 'unripe', N. Roglai

m atah,
in
Jarai
mat?h
Wr.
-n;
Cham
mitah, Malay rnentah, PMP *mataq; *mentaq; *a(n)taq.

*f?amuk 'mosquito', Acehnese
jamo?; ?amo?, Chru jam??, N. Roglai
earn
PR
Cham ?am o? -v, Wr. Cham jam auk,
o?; cam??,
jamo? -v, W. Cham
In large
Malay nyamuk, PMalayic *?amuk, PMP *nyamuk; *lamak; *lamuk.
part the reflexes of initial *?- have blended with those of *j-.
*?awa 'breath, soul, air', Rade ewa < *j-, Jarai (Lee) jawa, Chru jawa,
N. Roglai lawa, Haroi cawia 'breath, blow' < *j-; macawia, W. Cham (yawa),
W. Cham (Headley) /f?awa/ 'soul', PR Cham yawa 'blow out', Wr. Cham yaw?,
II: The Chamic
Appendix
Malay
jiwa, nyawa,
299
Lexicon
*f?awa, PMP *jawa
PMalayic
'air, soul, breath',
(*ma)
nihawa.
*?u 'he, she; they', Rade ?u, Jarai (PL) ?u, Jarai (Lee) ?u 'he', Chru
??, N. Roglai ??, Tsat ?au33 -vi, Haroi ?au -v,W. Cham ?u, PR Cham ?u, Wr.
Cham g?; MK: Bahnar (AC) ?i. [This form has cognates throughout An but the
are somewhat irregular]
vowel correspondences
*?us 'blow nose; sniffle; nasal mucus', Rade a?ih, Jarai (PL) ?Sh?h ?,
Jarai (Lee) ?uih, Chru ?i:h adung, N. Roglai ??h, Tsat ?au55 -v, W. Cham nun,
PR Cham ?uh; ha?uh, Wr. Cham ?uh; haftuh, Malay ingus, PMP *higus 'nasal
mucus'.
*nam 'six', Acehnese nam, Rade n?m, Jarai (PL) n?m, Jarai (Lee) n?m,
Chru nam, N. Roglai n?m, Tsat na:n?33, Haroi n?m, W. Cham n?m, PR Cham
n?m, Wr. Cham nani, Malay ?nam, PMalayic *anam, PMP *enem.
*naw
N. Roglai
'go; walk', Rade nau, Jarai (PL) nau, Jarai (Lee) nau, Chru nau,
n?u, Tsat na:u?33/42, Haroi nau, W. Cham nau, PR Cham naw, Wr.
Cham
PMP
nau,
*panaw
'go; walk'.
*pa:t 'four', Acehnese pui at, Rade p??, Jarai (PL) p??, Jarai (Lee) p??,
Chru pa:?, N. Roglai pa:?, Tsat pa?24, Haroi pa?,W. Cham pa?, PR Cham pa?,Wr.
Cham pak, Malay ?mpat, PMalayic *ampat, PMP *epat; MK: PMnong *pu?n.
[TheMK forms are not directly related to the PC forms]
*padam 'extinguish', N. Roglai padap, Tsat tham?42,W. Cham pap?m,
PR Cham patam, Wr. Cham padani, Malay padam, PMP *padem; MK: PNB
*p?t.
*paday
'rice,
paddy',
Acehnese
Jarai (Lee) padai, Chru padai, N. Roglai

Cham
pat-ai,
PR Cham
patay,
Wr.
Cham
pade,
Rade
madie,
Jarai
(PL)
padai,
padai, Tsat thau?42, Haroi pathiai, W.

padai,
Malay
padi.
pagina < *-r, Jarai (PL) paga, Jarai (Lee)

*pagar 'fence', Acehnese
Chru
N.
Tsat
paga,
kai33, Haroi pakhia, W. Cham paka, PR
paga,
Roglai paga,
Cham paka, Wr. Cham paga, Malay pagar, PMP *pager; MK: PMnong *p?r.
that PMP has a schwa in the second syllable.
*pagi 'tomorrow', Rade magi -v, Jarai (Lee) pagi, Chru pagi, N. Roglai
Haroi
pagi,
pakht, W. Cham pake -v, PR Cham pake -v,Wr. Cham page, Malay
pagi 'tomorrow'. This form does not look to be of any great antiquity.
Note
*pala:t -1 'palm; sole', Acehnese paluiat, Rade pl?? -1, Jarai (Lee) pl??
1, Chru p?a:?, N. Roglai pala:?, Tsat pie?24 -v, Haroi pala?, W. Cham pla?, PR
Cham pala?, Wr. Cham palak, PMP *palaj.
*palsy 'village', Jarai (PL) pla?, Jarai (Lee) plai, Chru plai -v, N. Roglai
Tsat
palai,
piai33, Haroi pal?i, W. Cham play 'region', PR Cham pal?y, Wr.
Cham pal?i, Malay balai, PMP *balay 'public building, community house'; MK:
PNB *pal?y, Bahnar (AC) paley.
[If theMalay and PMP are to be connected to
300
II: The Chamic
Appendix
Lexicon
the PC, two problems needed to be addressed: the initial is unexpectedly voice?
less in PC and the vowel seems to reflect an earlier *-i, not *-ay. Borrowing
is
most likely involved, probably into Chamic and intoMK]
*panah
'shoot
a bow',
(bow);
Acehnese
'arrow'
panah
notes
(Durie
variation
in nasalization),
Rade manah, Jarai (PL) panah, Jarai (Lee) panah,
Chru panah, N. Roglai pan?h, Tsat na55, Haroi pan?h, W. Cham panih, PR Cham
panih, Wr. Cham panih, Malay panah 'bow', PMalayic *panah, PMP *panaq,
PAn *panaq 'shoot an arrow; the flight of an arrow' MK:
PNB *p?n, PMnong
;
*p?n, PSB (Efimov) *paji, PKatuic *yapeji. The AN forms reconstruct to PAn.
the MK forms look related, they are not obviously
Although
have good etymologies
in theMK languages.
*papa:n
'board;
Acehnese
plank',
Cham papan; panah, PR Cham papan

PMP *papan.
Chru
papujan,
borrowings
Tsat
papa:n,
as they
pa:nn,
'table', Wr. Cham paban, Malay
W.
papan,
muntah
-i, Jarai (PL) patah, Jarai (Lee)
*patah 'vomit', Acehnese
*m/u(n)tah, PMP
patah, Tsat ta55, Haroi mat?h -i, Malay muntah, PMalayic
a
causative
prefix).
*pa-utaq (with
*pha 'thigh', Acehnese pha, Rade pha, Jarai (PL) pha, Jarai (Lee) pha,
Chru pha, N. Roglai pha, Tsat pha33, Haroi pha, W. Cham pha, PR Cham pha,
Wr. Cham ph?, Malay paha, PMP *paqa.
phuiat, Jarai (PL) ph?? (Pk), Jarai
*pha:t 'chisel, to plane', Acehnese
Chru
Tsat
W.
Cham
(Lee) ph??,
pha:?,
pha?, PR Cham pha?, Wr. Cham
pha:?24,
phak, Malay
pahat, PMP *paqet; MK: PSB (Efimov) *pha:t, *pha:?. The PSB
borrowed.
apparently
Acehnese
With
Rade
puinuihuiat,
*-an-
instrumental
m anhat,
Chru
infix:
'a
pahna:?
'chisel;
is
plane'*p-an-ha:t,
N.
plane',
Roglai
pahn?:?.
*phit 'bitter; bile', Acehnese

phet, Rade ph??, Jarai (PL) ph??, Jarai
N.
Chru
Tsat
(Lee) ph??,
Roglai phi:?,
phi:?,
phi?24, Haroi phei?, W. Cham phi?, PR
Cham phi?, Wr. Cham phik, Malay pahit, PMP *paqit.
phon, Rade ph?n 'stump,
*phun 'trunk; log; stem; plant', Acehnese
Jarai (PL) ph?n, Jarai (Lee) phun -1, Chru phun 'stem', N. Roglai phut
'stump', Tsat phun33, Haroi phon, W. Cham ph?n, PR Cham ph?n, Wr. Cham
stubble',
phun, Malay
pohon.
*pina:g
'betel
(areca
palm);
betel-nut',
Acehnese
pinuig,
Rade
manag,
Jarai (PL) panag, Jarai (Lee) panag, Chru pana:g, N. Roglai pin?g, Tsat na:g33,
Haroi panag, W. Cham panig, PR Cham panig, Wr. Cham panig, Malay pinang,
lateWMP *pinag ?; MK: PSB (Efimov) *pina:g; MK: PKatuic (DT) *pano?ng,
PKatuic *psnha:n.
Ultimately,
despite
its distribution
inWMP,
this might
be bor?
rowed.
*pluh 'ten', Acehnese

siploh, Rade pluh, Jarai (PL) pluh, Jarai (Lee)
pluh, Chru spluh, N. Roglai sa pluh, Tsat piu55, Haroi aploh, W. Cham ha pluh,
Appendix
II: The Chamic
301
Lexicon
PR Cham pl?h, Wr. Cham pluh, Malay
*puluh, PMP *sa-ga
s?-puluh, PMalayic
*sa-puluq.
puluq;
*po 'master; lord', Acehnese po, Rade po wide range of meanings and
uses, Jarai (PL) po, Jarai (Lee) po 'self, Chru po -v, N. Roglai po, Haroi po -v,
W. Cham po 'title', PR Cham po -v,Wr. Cham p?, Malay empu 'master', PMP
*empu
'grandparent'.
*pulaw 'island', Acehnese pulo, Rade plau ea, Chru bu:t pala:u -1, PR
Cham palaw, Wr. Cham palau, Malay pulau.
*pusat 'navel; center', Acehnese pusat, Rade mas?t, Jarai (PL) ph?n
Chru
s?t,
pasa?, N. Roglai pisa?, Haroi pasa?, W. Cham pas??, PR Cham path??,
Wr. Cham bathak, Malay pusat, PMP *pusej.
*putih 'white', Acehnese puteh, Tsat ti55, Haroi pateh 'light, fair', W.
Cham patih, PR Cham patih, Wr. Cham patih, Malay putih, PMalayic *putih,
PMP *putiq.
Rade m-, Jarai (PL) pa-, Jarai (Lee) pa-, Chru pa-;
*pa- 'CAUSATIVE',
per-,
N.
Roglai
Haroi
pa-,
pa-, W.
Cham
pa-,
*pa-; MK: Bahnar (AC) ba-. This prefix

and within MK most likely to PMK.
*pa-blsy
(*pa-
'sell',
'causative')
PR
Cham
pa-,
is reconstructable
Acehnese
Paiwan
to PAn
blsi
Rade
publoa,
PAn
pa-,
within AN
'buy',
Jarai (Lee) pablai; bablai, Chru pablai, N. Roglai pablai, Tsat li55 -ivt, Haroi
papl?i -m, W. Cham paplay, PR Cham papl?y, Wr. Cham pabl?y, Malay b?li
'buy',PMP*beli.
*ra 'person', N. Roglai
shortened form of PC *urag]
*ra-glay
mv,
N.
Roglai
*rabuh
(*raman?ih
'person'
radiai
'collapse,
rain Radglai, PR Cham ra,Wr. Cham r?.

'forest,
*glay
'Montagnard',
destroy',
Rade
jungle')
PR Cham
ebuh,
Malay
'
Roglai',
Raklay,
Wr.
roboh,
PMP
Rade
[A
-
raglai
Cham
raglai.
*Rebaq.
*rabug 'bamboo shoot', Rade eb?g, Jarai (PL) rabig -v, Jarai (Lee)
Chru
re??g -i,
rabug, N. Roglai rubuk, Tsat phug11; phog11, Haroi ph?g, PR Cham
ripug; rap?g, Wr. Cham ribug, Malay r?bung, PMP *rebug; MK: PNB *daqb?g
'bamboo sprouts', PSB (Efimov) *da'b1ag, PKatuic *?a?bag.
*rakit 'tie together; assemble', Acehnese
raket, Jarai (Lee) raid?, PR
Cham raki?,Wr. Cham rakik, Malay rakit, PMP *Rakit 'tie together; raft'.
*rala:g;
-i, Rade
Malay
*kala:g
thatch
'grass,
or congo;
roof
thatch',
Acehnese
hlag, Jarai (PL) hal?g, Jarai (Lee) hlag, Chru rala:g, N. Roglai
lalang
'k. of
long
grass'
narra
ag
rala:k,
*ralin 'candle wax', Rade hlin, Jarai (PL) halin, hlin, Jarai (Lee) hlin,
Chru ralin, N. Roglai ralit (hal?n 'candle wax'), PR Cham ral?n,Wr. Cham ralin,
cf. Malay Hlin, PMP *lilin 'beeswax'.
302
Appendix
II: The Chamic
Lexicon
*rag 'dry', Rade (Lee, Tharpe) rag, Haroi rag, PMP *keRag; (*ma
PSB (Efimov) *raji.
)Ragaw; *keRag; (*kaRig); MK: PNB *r?g 'dry wood',
Note that, although it is restricted to Highlands Chamic plus Haroi, it does seem
to reconstruct to PMP.
*ratus
-f
Acehnese
'hundred',
etuh
Rade
sirratoh,
-v,
Jarai
rat?h,
(PL)
Jarai (Lee) ratuh -v, Chru ratu:h, N. Roglai sa ratuh, Tsat ta11 tu33 -tf, Haroi
latouh -v,W. Cham ratuh, PR Cham rit?h; rat?h, Wr. Cham rituh; ratuh, Malay
PMP
ratus,
*sa-ratus.
*sa-ga-Ratus;
*raya ? 'big', Acehnese
raya, Haroi layia, PR Cham

*raya, PMP *Raya.
raya, Wr. Cham
raya, Malay raya, PMalayic

*ribow 'thousand', Acehnese
ribsa, Rade eb?u, Jarai (PL) rabau, Jarai

(Lee) rabau, Chru rabau, N. Roglai sa ng?n; rubau, Tsat pha n, Haroi laph?au, W.
Cham rapau, PR Cham ripow; rapow, Wr. Cham ribuw; rabuw, Malay ribu,
PWMP *ribu, no PMP; MK: PNB *rab?w, PSB (Efimov) *rabu:.
'
*rusa Sambhur deer', Acehnese
rusa, Jarai (PL) rasa, Jarai (Lee) rasa,
Chru
N.
rasa,
Cham
rasa,
rith?;
rusa,
Roglai
Haroi
rusa,
Malay
lasa, W.
PMP
Cham
Cham
PR
rasa,
ritha;
Wr.
ratha,
*Rusa.
*rusuk 'ribs', Acehnese

ruso?, Jarai (Lee) ras??, Chru rasu?, N. Roglai
tula:k ris?k -f, Haroi laso?, W. Cham talag las?? -i, PR Cham rath??, Wr. Cham
rusuk, PMP *Rusuk.
rasuk, Malay
*ra?a:n
Jarai
ra?an,
Cham
-ifl
(Lee)
kaftan
-i, PR
ra?an,
Chru
layam
li?an
Cham
ruinuin
Acehnese
'ladder,
-i;
-i, N.
la?an
-vm,
Cham
-i, Wr.
Rade
ri?a:t
Roglai
Jarai
ef?an,
-fl, Haroi
li?an
-i;
(PL)
W.
la?ian,
la?an
-i, PMP
'notched log ladder'. This form, despite being an apparent PMP ety?
is quite irregular in Chamic.
*sa 'one', Acehnese
sa, Rade sa, Jarai (PL) sa; ha (article), Jarai (Lee)
*haRezan
mon,
sa,
Chru
ha-,
sa, N.
sa, Tsat
Roglai
sa33, Haroi
sa;
ha;
ho, W.
Cham
sa;
ha,
PR
Cham tha,Wr. Cham s?, Malay satu, PMalayic *asa?, PMP *esa; *isa.
*sa pluh sa 'eleven', Acehnese
siblah, Rade pluh sa, Jarai (PL) pluh sa,
sa
Chru spluh sa, N. Roglai
pluh sa, W. Cham ha pluh sa, PR Cham tha pl?h
tha, Malay
s?b?las.
*sa-lapan
Wr.
Cham
'nine', Chru salpan, N. Roglai
salapat, PR Cham
thalip?n,
salipan.
*sakit 'sick, painful', Acehnese

saket, Jarai (Lee) ha kit, Chru -saki:?, N.
Tsat
PR
Cham
Wr. Cham hakik, Malay sakit, PMa?
saki:?,
haki?,
ki?24,
Roglai
PMP
*sakit.
layic *sakit,
*samilan
ten'
se
'nine',
W.
Cham
saml?n,
Malay
s?mbilan
('one
taken
ambilan).
from
303
Lexicon
II: The Chamic
Appendix
'broom; sweep', N. Roglai sapuh, Haroi apoh, W. Cham kai

sapu -f, PMP *sapu; MK: Bahnar (AC) co puih. [The Bahnar form
*sapuh
puh, Malay
is borrowed]
*sarip 'sip; slurp; suck in', Jarai (PL) hr?p (Pk), Jarai (Lee) hrip, Haroi
[The Malay and PMP second vowels do not
harip, Malay irup, PMP *hiRup.
match with the PC]
-i
*saya:p
Acehnese
'wing',
Rade
sayraap,
-n, W.
sia:??
Chru
siap,
che au? -i, PR Cham thyaw?, Wr. Cham thiap, Malay sayap, PMalayic
*sayap. Blust (p.c.) notes that there are no reflexes of *sayap outside of Malayo
Cham
but forms reflecting
Chamic,
'to fly' are found in the northern Phil?
*s-um-ayap
ippines.
'who;
*sey
be
the
unstressed
variant
'who', PR Cham
*siof
variant
this
word'
question
of
this
cf.
also
same
'honorific
same
prefix?',
cf.
also
'who',
si
Acehnese
morpheme,
which
might
Cham
hay
*sia, PMP *i-sai.
*sa-apa,
'title
which
soa, W.
Acehnese
morpheme,
they, Wr. Cham th?i, PMalayic
'honorific
*si-,
prefix'
the
be
might
for people's
stressed
Chru
names',
sa?
in sa?ai 'older sibling', Wr. Cham siy (Marrison 1975); sa- (AC), Malay si (No
in Malay). Blust (p.c.) notes that its recon?
longer has an honorific meaning
In a number
structed function is as an actor focus marker for personal nomin?is.
it has lost all grammatical
of languages, including Thao in Taiwan, and Malay,
function and is retained simply as a redundant marker of personal names. That is,
the reconstructed function in PAn and PMP was much like the sychronic function
of /si/ in Tagalog, marking the actor focus of personal nomin?is (proper names,
plus
a few
others).
'ant', Acehnese
sidom, Rade had?m, Jarai (PL) had?m, Jarai
of
adam
-i
first
Chru
hadSm,
(Lee)
syllable, N. Roglai sidap; sudap, Tsat a11
Haroi
W.
Cham
torn, PR Cham hat?m; t?m, Wr. Cham hadam;
ath?am,
than11,
dani, PMP (Blust; Adelaar) *sejem, *sijem; MK: Bahnar (AC) hudurn; hadani.
*sidam
[It is the Bahnar forms that are borrowed]

*sira -i 'salt', Acehnese
sira, Rade hra, Jarai (PL) hra, Jarai (Lee) hra,
Chru
sra, N.
Roglai
Cham
sar?, PMalayic
Malay
sukat.
*sukat
'to
sara,
Haroi
sra
-i, W.
Cham
sra
-i, PR
sara
Cham
-i, Wr.
*sira, PMP *qasiRa.

Acehnese
measure',
sukat
'of
volume',
Rade
mak??
-i,
*surat 'write; letter, book', Acehnese

surat, Rade hr??, Jarai (PL) hr??,
Jarai (Lee) hr??, Chru sra? in the word for paper, N. Roglai sura?, Haroi har??,
W. Cham hr??, PR Cham har??, Wr. Cham harak, Malay surat; MK: PSB (Efi?
mov)
sraq,
which,
*sara? 'book, written

Pacoh
uraq].
The
if the reconstruction
document',
*surat
appears
PKatuic
to also
is correct, suggests
(DT) *-uraaq
reconstruct
'writing'
in Efimov's
[Br?u
PSB,
the time depth of the borrowing
304
Appendix
II: The Chamic
Lexicon
even should neither the PSB or PKatuic recon?

into PSB as well as in Katuic;
struction hold up, this is evidence of reasonably early borrowing into PSB (or, at
least Bahnaric languages) and into PKatuic (or at least Katuic languages). The
root itself looks like it reconstructs at least toMalayo-Chamic,
but it is unclear
was.
what the original Malayo-Chamic
gloss
**susow > *tasow 'breast', Rade kas?u, Jarai (PL) tasau, Jarai (Lee)
tasau, N. Roglai tisau, Haroi cas?u, W. Cham tasau, PR Cham tathSw, Wr.
tathuw, Written Cham (AC) tasou, Malay susu, PMalayic *susu(?), PMP
*susu. The reconstruction with the double asterisk is what I assume, on the basis
Cham
and internal reconstruction, the earlier PC form would look

like, if it were not for gaps in the data base; the form with the single asterisk is
the earliest form reconstructable on the basis of just the evidence now available.
of external evidence
*ta? 'chop; cut', Acehnese

ta?, Jarai (PL) t??, Jarai (Lee) t??, Chru ta?,
N. Roglai ta?, Haroi t??,W. Cham t??, Malay
*tatak 'chop,
t?tak, PMalayic
hack', PMP *tektek 'chop; cut'; MK: Bahnar (AC) tak.
*taba:r 'tasteless; flat', Acehnese
tabraa < *-r, Rade kaba, Jarai (PL)
taba, Chru taba, N. Roglai taba, Haroi caphia, W. Cham tapa, PR Cham tapa,
Wr.
Cham
taba,
PWMP
tawar,
Malay
*tabaR.
*tabus
Tsat phu55, PR Cham
'ransom; save; rescue',

t?bus, PMP *tebus.
tabuh, Malay
*talsy
'rope;
string',
talo a, Rade
Acehnese
klsi
tap?h, Wr. Cham

Jarai
brag,
talai,
(PL)
Jarai (Lee) talai, Chru talai, N. Roglai talai, Tsat lai33, Haroi cal?i, W. Cham
talay, PR Cham tal?y, Wr. Cham tal?i, Malay tali, PMalayic *tali, PMP *talih,
PAn *CaliS.
*tama
tami,
PR Cham
'enter',
tami, Wr.
Acehnese
Cham
'to winnow',
Chru
tamAg,
tami,
tama,
*tama?
PMalayic
Haroi
'go
cama,
W.
PMP
inside',
Cham
*tama.
tampoa, Rade kapsi, hadsi, Jarai (PL)

*tampey
N.
Haroi
Roglai tupai,
tapai (Pk),
cap?i; cap?ai, W. Cham tapay, PR Cham
Wr.
Cham
tap?y; p?y,
tap?i, Malay tampi, PMP *tahepi; MK: Bahnar (AC)
tabeh;
Acehnese
tapai.
*tamuh
'grow;
sprout',
Chru
Cham
tamuh, PR Cham tam?h, Wr.

*tumbuh, PMP *tu(m)buq.
tam?h,
N.
Cham
Roglai
tum?h,
tamuh, Malay
Haroi
camoh,
W.
tumbuh, PMalayic
*taga 'ask', Acehnese

ta?Ag, Jarai (PL) ta?a, Jarai (Lee) ta?a, Chru
ta nia, N. Roglai tina, Haroi cafta, PR Cham tafti < *-g-, Wr. Cham ta??, Malay
tanya. [There are minor problems with this root: the medial nasal and subsequent
vowels pattern unexpectedly]
*taga:n
Chru
tagam,
N.
'hand; arm', Rade kagan,

Roglai
tag?n,
Tsat
ga:n33,
Jarai (PL) tagan, Jarai (Lee) tagan,
Haroi
cagan
'hand',
W.
Cham
tagan
Appendix
v, PR Cham
*tagan
tag?n, Wr.
305
Lexicon
II: The Chamic
Cham
tangin, Malay
tang?n, PMalayic
*tagan, PMP
'thumb'.
*tana? 'to cook', Rade kan??, Jarai (PL) tana?, Jarai (Lee) tana?, Chru
tana?, N. Roglai tan??, Haroi cana? 'cook rice', W. Cham tana?, PR Cham tan??,
Wr. Cham tanik, Malay tanak, PMalayic *tanak, *tanak, PMP *tanek.
'earth, soil', Acehnese
tanoh, Jarai (PL) tan?h (Pk), Jarai (Lee)
Tsat
tanah, Chru tanah, N. Roglai tan?h,
na55,W. Cham tanih, PR Cham tanih,
Wr. Cham tanih, Malay
tanah, PMalayic
*tanah, PMP *taneq; MK: PNB
*nteh
PSB
'earth',
(Efimov) *tne:h. The spread of MP
*taqneh 'dirt', PMnong
*tanah
forms indicates
the word
thus, at least the PNB form
is native to Austronesian;
looks to be borrowed!
*tapay
Rade
'rice
kapie, Chru
Cham tapai, Malay
wine;
Acehnese
alcohol',
taps
tapai, N. Roglai tapai, Haroi

tapai, PAn *tapay 'ferment'.
-v
'fermented
cap?i, PR Cham
rice
cake',
tapay, Wr.
Rade kapsi, Jarai (PL) tapai, Jarai (Lee) tapai, Chru

'rice wine'
Cf.
[This particular word seems to be related
tupai.
the making of rice wine, and yeast]
*tapug 'flour', Rade kap?g 'bread', Jarai (PL) tap?g, Jarai (Lee) tap?g,
Chru tapug, N. Roglai tupuk, PR Cham tap?g 'farine', Wr. Cham tapug, Malay
*tapsy
N.
Roglai
tapai,
to fermentation,
'yeast',
t?pung; MK: Bahnar (AC) tapug.

*tasi? 'sea; ocean', Rade kas??, Jarai (PL) ras?? -i, Jarai (Lee) ras?? -i,
Chru tasi:?, N. Roglai tasi:?, Haroi casei?, W. Cham tasi?, PR Cham tathi?, Wr.
Cham tathik, Malay tasik 'lake', PMalayic *tasik 'lake', PMP *tasik 'sea, salt?
water'.
Note
that
the Malay
refers only to lakes;

meaning
'sea;
tasik
the Chamic
has
undergone
forms, however,
a semantic
change
and
now
appear to still retain the older
saltwater'.
*thow 'know; can; able', Acehnese

thea; tu-, Rade th?u, Jarai (PL)
thau, Jarai (Lee) thau, Chru thau 'understand', N. Roglai thau, Tsat tiau?42 -i,
Haroi th?u, W. Cham thau, PR Cham th5w, Wr. Cham thSu, Malay tahu, PMa?
layic *tahu, PMP *taqu.
*thu 'year', Acehnese
thon, Rade th?n, Jarai (PL) thun, Jarai (Lee) th?n,
Chru thun, N. Roglai thut, Tsat thun33, Haroi thon, W. Cham th?n, PR Cham
th?n,Wr. Cham thun, Malay tahun, PMalayic *tahun, PMP *taqun.
*tikus 'rat', Acehnese
tikoh, Rade kakuih, Jarai (PL) takuih, Jarai (Lee)
takuih, Chru taku:h, N. Roglai tukuh, Tsat (na11) ku55, Haroi cakoh, W. Cham
takuh, PR Cham tak?h, Wr. Cham takuh, Malay tikus, PMalayic *tikus. [This
root is restricted to parts ofWMP]
*timun 'melon; cucumber', Acehnese
timon, Rade kam?n, Jarai (PL)
tamun, Jarai (Lee) tam?n 'a citrus', Chru tamun, N. Roglai tum?n, Haroi cam on,
306
W.
II: The Chamic
Appendix
tam?n
Cham
'cucumber';
tam?n
(AC) tamun, Malay

*tiney see *tiri?
*tin?;
*tinsy
'here',
cf.
PR
'melon',
padai
tamun, Wr. Cham
Cham
tam?n,
Lexicon
Wr.
Cham
timun, PMP *timun.
*ti, Acehnese
sinoa,
hinoa,
Rade
Jarai
tinsi,
(PL) anai, Chru ni 'here; this', N. Roglai tin!, Haroi ni 'here, this', W. Cham ni,
PR Cham ni,Wr. Cham ni, Malay sini, di sini.
*tiya:n
'pregnant', Rade
N.
matia:t
Roglai
tyan, Wr.
Cham
'stomach;
abdomen;
belly',
'pregnant',
tian, Malay
Tsat
tian
ten?33,
'uterus
Haroi
Cf.
-tiyraan
tean, W.
of a pregnant
*ti(?) 'PARTICLE'. Rade ti? -v, Rade

ti?-v.
Acehnese
in mratiyraan
tian, Jarai (PL) ki?n, Jarai (Lee) kian < *k- prefix, Chru tia:n,
Cham
woman',
(Lee) ti, N. Roglai
tean,
PMP
PR
Cham
*tian.
(Lee) ti, Haroi
'at'.
*trun 'descend', Acehnese

tr?n, Rade tr?n, Jarai (PL) tr?n, Jarai (Lee)
tr?n, Chru trun 'descend; southerly', N. Roglai trut, Tsat tsun33, Haroi tron, W.
Cham tr?n, PR Cham tr?n,Wr. Cham trun,Malay turun, PMP *tuRun.
*tu?ut 'knee', Acehnese
tuot; traot, Rade ka?t, Jarai (PL) k?ta??t, Jarai
(Lee) te??t, Chru ta??:?, N. Roglai tu?u:?, Haroi ca?u? -v, W. Cham ta?u?, PR
Cham tau?,Wr. Cham tauk, Malay lu/tut, PMalayic *tu?(ua)t, PMP *tuhud, PAn
*tuSud.
baroa, Rade mabrue

*tubursy 'yesterday', Acehnese
kabruai
N.
brai, Chru
(m) -i,
Roglai tubrai, Haroi macaprui
kaproy; paproy, Wr. Cham kabruai.
(m), Jarai (Lee)

(m), PR Cham
*tuha 'old (people)', Acehnese

tuha, Rade khua (m), Jarai (PL) tha,
Jarai (Lee) taha; tha, Chru tha, N. Roglai tuha, Tsat ha33 xau33, Haroi cana, W.
Cham taha, PR Cham taha, Wr. Cham taha, Malay tua; tuha, PMalayic *tuha(?),
PMP *tuqah.
tujoh, Rade kjuh, Jarai (PL) tajuh, Jarai (Lee)
*tujuh 'seven', Acehnese
N.
Chru
Roglai tijuh, Tsat su55, Haroi cas?h, W. Cham ta?uh, PR
tajuh,
tajuh,
Cham tac?h, Wr. Cham tajuh, Malay tujuh, but PMP *pitu.
[The form *tujuh
was
innovated is not
postdates the breakup of PMP, but precisely at what level it
clear yet]
tulraag, Rade klag, Jarai (PL) talog, Jarai
*tula:g 'bone', Acehnese
N.
Chru
(Lee) talag,
tala:g,
Roglai tula:k, Tsat la:g33, Haroi calag, W. Cham
talag, PR Cham talag, Wr. Cham talag, Malay tulang, PMalayic *tulag, PMP
*tuqelan; *tuqelag; MK: PMnong *katig. The similarities between theMK and
for 'bone' are intriguing.
Jarai (PL) tug (Pk), Chru tug, N. Roglai tuk,
*tug 'stomach; abdomen',
Haroi tog,W. Cham tug 'intestines', PR Cham t?g,Wr. Cham tug, Malay jantung
'heart'. Confined to languages inwestern Indonesia (Blust, p.c.).
the PMP words
*taive',
of
ter,
PMP
The
'inadvertent'.
*taR
'inadvertent;
PAn,
PMP,
PC forms reflect the same etymon; theMK forms (discussed

to be more than simply look-alikes. They appear borrowed.
*tabow
-v
Acehnese
'sugarcane',
Rade
trabea,
ta-
Jarai
'inadvertent',
'inadvertent',
*tar-
PAn
action',
involuntary
*tAr-
PMalayic
tra-
Acehnese
'INADVERTENT',
Malay
307
Lexicon
II: The Chamic
Appendix
'reflex?
marker
and
PMalayic,
in Chapter 9) look
Jarai
kab?u,
(PL)
tabau,
Jarai (Lee) tabau, Chru tarbau, N. Roglai tubau, Tsat pha11, Haroi caph?au, W.
Cham tapau, PR Cham tapow, Wr. Cham tabuw, Malay t?bu, PMP *tebuh.
*taliga 'ear', Acehnese
grali?uag -iv; puftuag, Rade kaga, Jarai (PL)
tagia (m), Jarai (Lee) tagia (m), Chru tania (m), N. Roglai ligi? (m); rigi? -i (m),
PR
Tsat ga33, Haroi cage a (m) 'ear, brain (place of thinking and memory)',
Cham tagi (m), Wr. Cham tang?, Malay
*taliga(?), PMP
t?linga, PMalayic
*taliga.
trapat, Rade kap??, Jarai (PL) tapa?,

*tapat 'straight; honest', Acehnese
Jarai (Lee) tapa? -1, Chru tapa?, N. Roglai tupa?, Haroi capa?, PR Cham tap??,
Wr. Cham tapak, Malay t?pat 'exactly', PWMP *tepet 'exact; precise'; MK:
Bahnar (AC) tap?t.
*wa
N.
'paternal',
Malay
'uncle,
ua
'old;
Haroi
wa,
Roglai
aunt
Jarai
aunt',
or uncle
(PL) wa,
wa,
older
Jarai
W.
Cham
than
one's
wa
(Lee)
wa,
PR
Cham
PMP
parents',
wa,
Chru
aunt',
'uncle,
Wr.
*ua?
Cham
'uncle,
wa
w?,
aunt'
MK: PMnong *wa 'uncle, aunt', PSB (Efimov) *wa:. Note the PMnong form,
but the antiquity of the PMP form suggests that this was an early borrowing into
PMnong and PSB, not the other way around.
*ya:p
Rade
'count',
yap,
Jarai
(Lee)
yap,
Chru
ya:u?,
N.
Roglai
ya:?,
Haroi yiau?, W. Cham yau?, PR Cham yaw?, Wr. Cham yap -f, PMP *ihap. The
vowel length results from a requirement found not just inAustronesian where it is
widespread
be
1.2
at least
but inmany
two moras
PC words
languages of the world
that vowels
inmonosyllable
roots
long.
of MK
origin
there is a second, large set of

items are those borrowings from MK sources
that were borrowed so early as to be reconstructable to PC. These forms include
numerous forms from two distinct MK subgroups, the more southerly Bahnaric
In addition
to the form inherited from Austronesian,
forms reconstructed
for PC. These
In many cases, of course, the forms are found

northerly Katuic.
so
it is not possible to tell which branch of MK the
both in Bahnaric and Katuic
and the more
borrowing came from, but in a significant subset of cases, the form is attested in
It is inevitable of course that there are accidental
only one of these branches.
gaps in the data, but the large number of forms attested in only one of these two
branches of MK makes it clear that both branches contributed material to pre-PC
308
Appendix
II: The Chamic
establishing that there was early, intense contact not just with Bahnaric
but also with Katuic speakers.
Lexicon
speakers
that a MK form reconstructs to PC is done almost exclu?

Establishing
on
sively
correspondence patterns within PC. In this regard, the existence of an
extra-Chamic Malay form does not mean that a form is necessarily Austronesian
as Malay
a number of MK
forms, for example, ketam 'crab'

'ant' quite independently of Chamic. Thus the occurrence of a form
both inMalay and Chamic does little to establish either that form must inevitably
nor does it contribute to determining a time depth for aMK bor?
be Austronesian
itself has borrowed
and semut
rowing.
It is important to realize that historically there was contact between Aus?

tronesian and MK languages before Austronesians
speakers reached the coast of
On the Malaysian
Vietnam.
peninsula, there are still Orang Asli speaking MK
languages, and in Borneo, Adelaar (1995) argues that there is linguistic evidence
of aMK influence on the Land Dayak. Adelaar presents two pieces of evidence
forMK influence on Land Dayak: one is the presence of preploded nasals, found
inMK languages and inAustronesian
languages in contact with MK languages,
and the other is the presence of MK vocabulary in Land Dayak. To this evidence
I would add another piece of evidence that suggests the shifting of MK speakers
to the Austronesian
Specifically,
languages of Borneo, including proto-Malayic.
forms found in the lexicon below (from Adelaar
throughout the proto-Malayic
a
are
there
should be vowel final but
number
of
forms
that, etymologically,
1992)
instead end in a final glottal stop (Blust, p.c., mentioned
the difficulty of account?
finals
On
the
for
these
ing
Malay peninsula, much the same
etymologically).
process can be seen with borrowings into Aslian languages from Malay, as is evi?
dent in Benjamin's
in his work the subgrouping of Orang Asli
(1976) wordlists
of
the
in
those lists, as in Land Dayak, as in proto
Malay peninsula;
languages
(cf. the forms in this lexicon), final glottal stops are often present on
Malay words which should end in open syllables. The significance of all this, of
course, is that some MK influence may have already been present in their lan?
Malayic
guage when the pre-Chamic speakers arrived inVietnam.

The evidence that a form is borrowed from a MK
source varies from
form to form, but two general types of evidence are widely represented in the fol?
lowing list. One type of evidence is provided by reconstructions of Austronesian
subgroups, on the one hand, and MK subgroups on the other. Certain forms are
found
in reconstructed
nowhere
in one or more
apparently
such forms have been borrowed
the phonology
subgroups of MK but, aside from PC,

Here the unmarked hypothesis is that
from MK. A second type of evidence is found in
else inAustronesian.
a number of specific vowels, certain diph?

(e.g. the implosives (with a small number of well
of the reconstruction:
thongs, particular consonants
309
Lexicon
II: The Chamic
Appendix
know exceptions), and otherwise unique clusters only occur inMK borrowings
(discussed in detail elsewhere in this work). The fact that a word contains one of
these elements is by itself evidence that it is a MK borrowing;
however, it is
almost
never
to depend
necessary
this
upon
evidence
as when
alone,
a word
con?
tains one of these elements, it is inevitably the case that the word also lacks an
Austronesian
etymology. Here again the unmarked hypothesis is that the word is
aMK borrowing.
x*?a:k
'a crow',
Acehnese
a?-a?,
Rade
ak,
Jarai
?a?, Chru
(Lee)
cim
a:?,
N. Roglai a:?, Haroi a? 'raven', W. Cham a?, PR Cham a?,Wr. Cham ak; MK:
Bahnar (AC) ?k, PKatuic *[k/g]a?a:?, *?a?a:?. One has to be careful with such
onomatopoetic
forms.
x*?ada 'duck', Chru ada, N. Roglai ada, Tsat tha11,W. Cham pa, PR
PSB (Efimov) *da:, PKatuic *yadia,
Cham ata; pa,Wr. Cham ad?; da; MK:
*?adia.
x*?adoh 'sing', Jarai (PL) ?doh, Jarai (Lee) ?adoh, Chru adoh, N. Roglai
toh, PR Cham atoh; toh, Wr. Cham adauh; MK: PSB (Efimov)
*?adoh. [The vowel also suggests aMK origin]
adoh, W. Cham
x*?aduan
cf.
'old',
x*maduan,
Rade
Jarai
aduon,
(PL)
Jarai
adon,
(Lee)
form obviously
is obviously related to x*maduan; both forms are
restricted to two Highlands Chamic languages.
[The vowel also suggests aMK
?adon.
This
origin]
x*?aha;
x*ha
'open
(mouth
to say
sthg.)',
cf.
Vietnamese
h?, Acehnese
hah -i, Rade
ha, Jarai (PL) ha, Jarai (Lee) ha, Chru ha 'slightly opened', N.
Roglai aha, Tsat ha33, Haroi ha 'wide open', W. Cham ha, PR Cham ha, Wr.
Cham h?; MK: PNB *ha, PMnong *ha, PSB (Efimov) *ha:?, *ha:, PKatuic
*taha,
*?ahah.
x*?ako? 'head', Rade k??, Jarai (PL) ak?, Jarai (Lee) ?ak5?, Chru ako?,
N. Roglai ako?, Tsat ko?24, Haroi ak3?, W. Cham k5?, PR Cham ak5?; ko?, Wr.
Cham akauk; kauk. [The vowel suggests aMK origin]
ale, Jarai (PL) als -v, ale -v, Jarai
(Lee) ?ale, Chru ale, N. Roglai ale, Haroi ale, W. Cham cut Is 'small bamboo';
MK: PNB *pale 'bamboo', PMnong *gle, PSB (Efimov) *gls:, PKatuic *?alae.
Restricted to Highlands Chamic plus Haroi andW. Cham.
[The vowel also sug?
a
gests MK origin]
x*?ale
'medium
x*?amu:g
Note
the restriction
'snout;
bamboo',
muzzle',
to Highlands
Rade
Rade
Chamic.
amug,
Chru
[The vowel
mo:g,
N.
Roglai
m?:k
-f.
length also suggests aMK
origin]
310
Appendix
II: The Chamic
Lexicon
x*?amu:g (?) 'bunch; ear of grain; stalk', Rade amug 'stalk (bananas)',
Chru
amu:g
-v
'stalk
of
bananas';
tarmung,
N.
amu:k
Roglai
'stalk',
Haroi
PR Cham (Blood) mog, Wr. Cham

amug, W. Cham mog 'stalk of bananas',
(AC) amug. Cf. MK: PKatuic (DT) *-phoong. [The vowel length suggests aMK
origin]
x*?ato:g 'hit with implement; beat (gong)', Rade tog, Jarai (PL) ?tog,
Chru ato:g, N. Roglai ato:k 'beat (gong)', Haroi atog, W. Cham tog, PR Cham
[The vowel also suggests a
atog, Wr. Cham ataug; MK: PKatuic (DT) *thiig.
MK origin]
x*?aw 'clothing; shirt', Viet, ?o, Rade au, Jarai (PL) au, Jarai (Lee) ?au,
Chru a:u -1 'shirt', N. Roglai au, Tsat ?a:un, Haroi au, W. Cham au, PR Cham
aw, Wr. Cham aw; MK: PNB *aw, PSB (Efimov) *?a:w.
x*?awa:k 'spoon; ladle', Acehnese awuia?, Rade awak, Jarai (PL) aw??,
awa?, PR Cham awa?, Wr. Cham awak; MK: Bahnar (AC)
Chru awa:?, Haroi

uak,
PKatuic
*hawa:?'ladle'.
e? -f, Rade eh, Jarai (PL) eh,

x*?eh 'excrement; defecate', Acehnese
Jarai (Lee) ??h, Chru eh, N. Roglai eh, Haroi eh,W. Cham eh, PR Cham ?h,Wr.
Cham aih; MK: PNB *?c, PSB (Efimov) *?ac, PKatuic *?eh. [The vowel also
suggests aMK origin]
x*?iaw
'call;
cry;
yell',
Rade
ieu
-v, Jarai
(PL)
iau,
iau,
Jarai
Chru iau, N. Roglai iau, Tsat ?ia33,Haroi ?ju,W. Cham iau, PR Cham
[The triphthong suggests aMK origin]
x*?ja:? 'hold; carry', Rade
Roglai ?ja:?,Haroi ja? 'carry in hand'
[The
Highlands Chamic plus Haroi.
(Lee)
?iau,
(Lee) ??w.
?j??, Jarai (PL) ?j??, Jarai (Lee) ?j??, N.

MK:
;
PMnong *yak 'carry'. Restricted to
initial also suggests aMK origin]
x*?jam 'soup, thin', Rade ?jam, Chru i?m, N. Roglai ?ja:p -1. Restricted
to Highlands Chamic.
[The initial also suggests aMK origin]
x*?juh 'firewood', Rade ?juh, Jarai (PL) ?j?h, Jarai (Lee) ?juh; ?j?h,
Chru iuh, N. Roglai ?juh, Tsat ?iu55, Haroi ?j?h, W. Cham ?juh, PR Cham ?j?h,
Wr. Cham ?juh. [The initial suggests aMK origin]
x*?jup 'smoke tobacco', Rade ?jup hat, Jarai (PL) ??j?p, Jarai (Lee)
?j?p, Chru iu? 'smoke', N. Roglai ?ju:?, Haroi ?ju?,W. Cham ?ju? -f, PR Cham
[The
?ju?,Wr. Cham ?juk; MK: PKatuic *ha?je:p, *kaje:p, *?ju:p, *?jo? 'suck'.
initial also suggests aMK origin]
x*?jap 'correct; right', Rade ?jo?, Jarai (PL) ?ja?, Jarai (Lee) ?ja?, Chru
N.
iau?,
Roglai ?ja?; dadi, Haroi j?u?,W. Cham ?jau?, PR Cham ?jow?,Wr. Cham
?jap;MK: PNB *joq. [The initial and the vowel also suggest aMK origin]
Appendix
II: The Chamic
311
Lexicon
Jarai (PL) ?ft?m, Jarai (Lee)

food; legumes',
'vegetables;
?a?am, Chru i?m, Tsat ?ia?33, PR Cham ?j?m, Wr. Cham ?jani. [The initial sug?
gests aMK origin]
x*?ftam
x*?o? 'vomit', Jarai (Lee) ?o?; ha???; ???, Chru o:? 'choke', PR Cham
(Lee) ?o?,Wr. Cham (AC) auk; MK: PNB *h?k, PMnong *hok, PSB (Efimov)
*h??, PKatuic (DT) *-hook 'choke on food'.
origin]
x*?u:g 'husband; male', Rade ug, Jarai (PL) ?g (Est), Jarai (Lee) ?og, PR
(Lee) ?og 'you (to a man), Wr. Cham (AC) og 'grandfather', PNB *?g
'son-in-law'.
[The vowel length also suggests aMK origin]
Cham
Roglai
x*?u?? -n 'to saw', Rade u??, Jarai (PL) (u?) ?nu??, Jarai (Lee) ?u??, N.
u??, PR Cham (Lee) ?ua?. [The vowel suggests aMK origin]
x*?am 'to roast', Rade am, Jarai (Lee) ?om, Chru am, N. Roglai ap,
Haroi om, W. Cham am -v, PR Cham (Lee) am. [The vowel suggests aMK ori?
gin]
x*ba:l 'mend; to patch',
Jarai (Lee) bal, Chru bail, N. Roglai ban,
Haroi phial, W. Cham pal, PR Cham pal, Wr. Cham bal; MK: PKatuic *-p_l 'to
patch'. With the instrumental infix: x*b-an-a:l 'rag', Rade manai, Jarai (PL)
banal, Jarai (Lee) banal, N. Roglai ban?n. Note: The connection between the
infixed and the non-infixed forms depends upon acceptance of the semantics of
the equation of 'rag' with 'mend; path'.
x*boh 'wash; launder', Rade boh, Jarai (PL) boh (Ouest), Jarai (Lee)
Chru
boh,
boh, N. Roglai boh, Haroi phun, PR Cham poh, Wr. Cham bauh. [The
vowel suggests aMK origin]
x*b-an-at
'a dam;
a fence',
Rade
man??,
Jarai
(PL)
bani?,
Jarai
(Lee)
bani?, Chru ban?:? -f, N. Roglai bana?, W. Cham pana? 'a dam', PR Cham pirii?,
Wr. Cham binik; MK: Bahnar (AC) banot. [The vowel and the infix also suggest
aMK origin]
x*ba?ar 'paper; book', Rade maar, Jarai (PL) ma?ar (Mod.), Chru
N.
ba?a:r,
Roglai ba?a, Haroi pa?ial, W. Cham pa?ar -il, PR Cham pi?ar, Wr.
Cham biar; MK: PNB *baqar 'paper; book', PSB (Efimov) *pa?a:r.
x*bala 'tusk; ivory', Rade mla, Jarai (PL) bia, Jarai (Lee) bia, Chru bia,
N. Roglai bala, Haroi palia, W. Cham pla, PR Cham pila, Wr. Cham bil?; MK:
PMnong *bla, PSB (Efimov) *bla:, PKatuic (DT) pal.k 'tusk' (the final -k is
unexpected).
x*bato 'teach', Rade mato, Jarai (PL) p?to, Jarai (Lee) pato, Chru pato,
N. Roglai pato, Tsat to11, Haroi pato, W. Cham pato, PR Cham pato, Wr. Cham
312
Appendix
II: The Chamic
Lexicon
patauw; MK: PSB (Efimov) *pa'diaw 'teach, dictate'. [The vowel also suggests
aMK origin]
'
x*be? IMPERATIVE, negative, don't', Acehnese be? 'don't', Chru be?,
N. Roglai be?, PR Cham p??, Wr. Cham baik; MK: PNB *beq. [The vowel also
x*biluay 'gourd', N. Roglai biluai, Wr.
*?alhuaj. [The vowel also suggests aMK origin]
x*blah
'skirt',
'square of cloth',
cf.
'flat
PR Cham plah
object',
Chru
Cham
blah
pluai; MK:
'sheet',
W.
PKatuic
Cham
'un habit; sheet', Wr. Cham blah; MK:
plah
PNB
*blah.
x*blo 'split; pierce; earring', Jarai (Lee) bio, Chru bio 'earring', Haroi
?lo 'earring'. Restricted to Highlands Chamic plus Haroi.
a
MK
gests
origin]
x*blak
'sound
of
turbulent
water
in
stream
rapids;
turn
over',
(Lee) bli? 'rolled up', Haroi pli? -i,W. Cham pli? -v 'spin', PR Cham
[The vowel suggests aMK origin]
Jarai
(Lee) pl??.
x*bo:g 'coffin; casket', Rade bog, Jarai (PL) ?og, Jarai (Lee) bog, Chru
bo:g, N. Roglai bo:k, Haroi phug, PR Cham pog, Wr. Cham baug; MK: PNB
*bog, Bahnar (AC) bog. [The vowel also suggests aMK origin]
x*bok 'to swell', Jarai (PL) tSb?k (Ouest), N. Roglai bo?, Haroi phu?,
W. Cham po? -v takuai 'goiter', PR Cham po? -v, Wr. Cham bok; MK: Bahnar
(AC) bop ?. [The vowel also suggests aMK origin]
x*buc 'uproot, pull up; scratch', Rade bue, Jarai (PL) b?c, Jarai (Lee)
Chru
buc,
bui?, N. Roglai pui?, Haroi ph?i?; p?i? -v; p?i? -i, W. Cham pui? 'pick,
pluck4 ?, PR Cham p?y?, Wr. Cham buc; MK: PKatuic *hapoj?, *yapuaj? 'pull
out'. [The final also suggests aMK origin]
x*buhay 'otter', Rade kamhe, Jarai (PL) pahai, N. Roglai buhai, W.
Cham pahas -f, PR Cham phay, Wr. Cham bhai; MK: Bahnar (AC) ph?y, PKa?
tuic (DT) *pih_, cf. Pacoh pih?y.
x*bat 'to dam; to fence', Rade bi? ea, Jarai (PL) bi?, Jarai (Lee) bi?,
Chru ba?,W. Cham pa? 'to dam up'. [The vowel suggests aMK origin]
x*?ah 'chipped; broken', Rade ?ah 'dented', Jarai (Lee) ?ah, Chru ?eh
[The initial sug?
v, N. Roglai ?ah, Haroi ??h. Restricted to Highlands Chamic.
gests aMK
origin]
x*?ag 'hole; door', Rade ??g, Jarai (PL) ?mag, Jarai (Lee) ??g, Chru ?ag
N.
'door',
Roglai ?ak, Haroi ??g, W. Cham pa??g, PR Cham ??g, Wr. Cham ?ag,
[The initial also sug?
Malay pintu; MK: PNB *qm?g 'door', Bahnar (AC) mag.
gests aMK origin]
313
Lexicon
II: The Chamic
Appendix
x*?ar 'coiled; wrapped around', Rade -ar, N. Roglai ?a 'coil (snake)',

PR Cham ??r,Wr. Cham Bar. [The initial suggests aMK origin]
x*?e? 'elf. long, thin objects', Jarai (Lee) ?e?, Chru ?e?, N. Roglai ?e?
objects', Tsat ?e?24, Haroi ?e?, W. Cham ??k -f, PR Cham ?e?, Wr.
Cham baik. [The initial and the vowel suggest aMK origin]
'stick-like
x*?ia? 'little', Rade ?i??, Jarai (PL) ?i??, Jarai (Lee) ?i??, Chru lu bia? -i,
PR Cham (Lee) ?i??. [The initial and the vowel suggest
Haroi ?ea? 'a moment',
aMK
origin]
x*?o:? 'face', cf. 'nose' Rade ?o? mata, Jarai (PL) ?o?, Jarai (Lee) ???;
?o:? mata 'cheek' and 'eye' ?, N. Roglai ?o:? mata, Tsat ?o?24,Haroi
Chru
???,
?o?,W. Cham ?o?, PR Cham ?o?, Wr. Cham ?auk; MK: Bahnar (AC) b?; buk;
bok. [The vowel also suggests aMK origin]
x*?ram 'arrow', Rade ?r?m, Jarai (PL) ?r?m, ?mr?m, Jarai (Lee) ?r?m,
Chru ?ram, N. Roglai ?rap, Haroi ?r?m, W. Cham ?r?m, PR Cham ?r?m, Wr.
Cham ?ram; MK: PMnong *k?m, Bahnar (AC) bram; mram.
[The initial conso?
nant and the vowel also suggest aMK
origin]
x*?u:n 'poor, cheap; easy', Rade ?un, Jarai (PL) ?m?nh, munh (CT),
Chru fiuan 'easy', Haroi ?un 'poor, cheap; easy', PR Cham ??n 'cheap; easy',
Wr. Cham ?uan; MK: Bahnar (AC) bon 'cheap, easy'.
[The initial consonant
a
and the vowel length also suggest MK origin]
x*?uah 'scold', Rade ?uah, Chru ?uah, PR Cham ?w?h, Wr. Cham ?yah.
[The initial and the vowel suggest aMK origin]
x*?uh...x*?oh
man
??h; W.
'non
plus';
Cham
'not,
?u...oh;
o; o
'yes/no
no;
Chru
Rade
negative',
?uh...au;
question
N.
marker';
?u;
Roglai
?u...?h;
?uh...oh;
PR Cham
ka...o
?u...?ja...?h;
Tsat
'pas
pu33
Jarai
?u
-i; Haroi
encore',
PSB
[The initial suggests aMK origin] Note: Not

(Efimov) *?o:h 'denial particle'.
do
the
initial
and
final
of
x*?uh
show irregularities, but the initial itself is
only
totally unexpected if the form is inherited from PAn. Also, while the negative is
often bipartite, it is not always so. The origin of this structure is still unclear. Cf.
Lee 1996 for further discussion. Acehnese
also has a bipartite construction for
'not yet' goh....lom, but the components do not seem related to the constructions
in the mainland Chamic languages. Cf. also 'not yet' immediately below.
x*?ag 'eat', Rade ?ag, Jarai (PL) ??g, Jarai (Lee) ??g -v, Chru ?ag, N.
Tsat ?ag33, Haroi ?ig, W. Cham ?ag, PR Cham ??g, Wr. Cham ?ag;
?ak,
Roglai
MK: PNB *c?g 'eat rice', PMnong *s?g, PSB (Efimov) *so:g. The initial of the
forms is quite unexpected, so it is unclear if these forms are related to the PC
form. [The initial and the vowel also suggest aMK origin]
MK
314
Appendix
II: The Chamic
Lexicon
x*?ag 'to burn', Jarai (Lee) ?ag, Chru ?ag, W. Cham Bag, PR Cham ??g,
Wr. Cham Bag. [The initial and the vowel suggest aMK origin]
'chop, strike', Rade kacoh, Jarai (PL) tacoh, Chru tacoh, N.
W.
Cham
tacoh, PR Cham tacan, Wr. Cham tacauh; Wr. Cham
Roglai ticoh,
[The vowel also suggests
(AC) cacauh, ticauh; MK: PSB (Efimov) *coh 'chop'.
x*cacoh
aMK origin]
x*cadiag 'finger', Rade kadieg, Jarai (PL) cadeg, Jarai (Lee) cadeg,
Haroi cadeag, W. Cham cadiag -i;MK: PNB *d?g 'little finger', PKatuic *?dag.
gests aMK
origin]
x*canah
tary;
crossroads',
'cross
N.
over;
Roglai
crossroads',
canah,
W.
Jarai
Cham
(Lee)
k?n
canah,
Chru
'crossroad'
canah
-i,
PR
'tribu?
Cham
can?h, Wr. Cham canah. If theW. Cham form is related, that would indicate that
the rest of these forms have an instrumental infix -an-, an indication of a likely
MK origin.
x*ca:g 'wait; await', Jarai (PL) pacag, Jarai (Lee) c?g -1,Chru ca:g, N.
Haroi pacag, W. Cham cag, PR Cham cog, Wr. Cham cog; MK:
ca:k,
Roglai
Bahnar (AC) ?ag. [The initial at least suggests aMK origin]
x*caramah
'generous', Rade kamah, Jarai (PL) camah, Jarai (Lee)
cam ah, N. Roglai caramah -n, PR Cham (Lee) kamah -v; MK: PSB (Efimov)
*sramah.
se? -i, Chru cia?, N. Roglai

x*cia? 'peel (with a knife)', Acehnese
a
cleave'.
vowel
MK
[The
suggests
origin]
'split,
x*cica?
'lizard;
gecko;
salamander',
Acehnese
cica?,
tica?,
cia?
Rade
t?klelo, Jarai (PL) k?g k?? (Ouest); c?g c?? (Est), W. Cham tac?? -i, PR Cham
kac??, Wr. Cham kacak, Wr. Cham (AC) caca?, Malay cicak 'garden lizard'.
x*cic?t 'great grandchild', P-Acehnese
cAt,
(Durie) *(ce)cet, Acehnese
Rade c?? 'great grandchild', Jarai (PL) tac?? (Ss), Jarai (Lee) tace? -1, Chru ce?
-n, N. Roglai tice?, Haroi cac??, PR Cham tac??, Wr. Cham tacaik, Wr. Cham
(AC) cacai?, ticai?, Malay cicit. [The vowel suggests aMK origin]
x*cih 'write; draw', Rade ein 'to write', Jarai (PL) cih rup, Jarai (Lee)
cih, Chru cih, N. Roglai chih 'to draw', Haroi ceh, W. Cham cih, PR Cham cih,
Wr. Cham cih; MK: PNB *ach?h, PMnong *cih, PSB (Efimov) *ci:h 'write,
draw'.
x*cim 'bird', Acehnese

cicem, Rade cim, Jarai (PL) cim, Jarai (Lee)
Tsat
N.
-tsun33 -v, Haroi cem, W. Cham cim, PR
Chru
cim,
cim,
Roglai cip,
Cham cim, Wr. Cham cirri;MK: PNB *c?m, PMnong *sim, Bahnar (AC) ?em,
PSB (Efimov) *shim, PKatuic *hace:m; *?ace:m.
Appendix
315
Lexicon
II: The Chamic
x*cig 'small gong', Rade cig 'small gong', Jarai (PL) cig, Chru ci:g 'gong
(with hump)', W. Cham ceg -v;MK: PNB *ch?g, PSB (Efimov) *cig. Restricted
to Highlands Chamic plus W. Cham.
c?h, PR Cham
x*coh 'dig', Jarai (PL) c?h, Jarai (Lee) coh, Haroi

Wr. Cham cauh. [The vowel suggests aMK origin]
c?h,
x*coh 'peck (of bird); strike (snake)', Acehnese

coh, Rade ?oh, Jarai
(Lee) coh, Chru coh, N. Roglai coh, Haroi c?h 'strike (snake)', W. Cham coh;
MK: PNB *joh. [The vowel also suggests aMK origin]
'to
x*cok
he a
c??
Haroi
cry',
'to wail,
cry
W.
mourning',
co?
Cham
'weep, mourn', PR Cham c??,Wr. Cham cauk; MK: PNB *kro. [The vowel also
x*cuah 'sand', Rade cuan, Jarai (PL) cuah, Jarai (Lee) cuah, Chru cuah,
N. Roglai cuah, Tsat tsua55, Haroi coah, W. Cham coah, PR Cham cw?h, Wr.
Cham cuah; MK: Bahnar (AC) cuah; coah, PKatuic *sa3uah, *ya3uah.
[The
vowel also suggests aMK origin]
x*cuan
x*cuco
N.
seize',
'grab,
tion to a Highlands
MK origin]
Chamic
language plus W. Cham.

Rade
'grandchild',
Chru co, N. Roglai

tacauw, Wr. Cham
tico, Haroi
Jarai
co,
caco, W. Cham
cuan.
Cham
cuat, W.
Roglai
taco;
(PL)
Note
the
restric?

taco,
Jarai
taco,
(Lee)
taco, PR Cham
taco, Wr. Cham

cucu, PMalayic *cucu?; MK:
(AC) cacauv, ticauv, Malay

PNB *c?w, PMnong *sau, PSB (Efimov) *saw. [Adelaar (p.c.) suggests that this
might be a very early Dravidian borrowing; if so, the presence of the form in
three branches of Bahnaric suggests either very early borrowing into Bahnaric
from Chamic or the possibility
it was originally MK, not Dravidian. However,
remember
that Austroasiatic
suggests aMK
are
speakers
found
in India.
In any
barbecue
whole
case,
the vowel
origin for the PC variant.]

'burn
x*cuh
Rade
trns.',
cuh
'kindle;
Jarai
animal',
(PL) c?h, Jarai (Lee) cuh, N. Roglai chuh 'kindle', Haroi coh, W. Cham cuh
'fry', PR Cham c?h, Wr. Cham cuh; MK: PMnong *chu 'burn', PSB (Efimov)
*jhu:
'burn,
on
set
x*cat
fire'.
'mountain
range',
Acehnese
cAt
'hill,
steep,
vertical',
Rade
ci?
-i,
Jarai (PL) ci?, Jarai (Lee) ci?, Chru ca?, N. Roglai ca?, Tsat tsa?24,Haroi ca?,W.
Cham ca?, PR Cham ca?,Wr. Cham cak. [The vowel suggests aMK origin]
x*d-an-uc
'stinger',
Chru
danui?,
PR Cham
tan?y?,
Wr.
Cham
danuc.
x*dar 'encircle', Rade dar 'go around', Jarai (PL) dar; dar, dar (Ouest)
d'un lieu', Jarai (Lee) dar; MK: PNB *t?p dar.
'circonf?rence
316
Appendix
IL The Chamic
Lexicon
'fell a tree', Rade druom, Jarai (PL) drom; grom (CT), Jarai
Chru
(Lee) drom,
dro:m, N. Roglai dro:p, Haroi trum dial.; carum, PR Cham
(Lee) ?rom, trom. [The vowel suggests aMK origin]
x*druam
x*duc 'to sting', Rade duc, Jarai (PL) du??, N. Roglai (ta?), W. Cham
t?y?,Wr. Cham due. [The final suggests aMK origin]
tui?, PR Cham
dAg, Rade dag, Jarai (PL) da? d?g, Jarai

x*dag 'stand; stop', Acehnese
N.
Chru
Tsat
(Lee) dag,
dag,
Roglai dak,
thag11, Haroi thig 'stand', W. Cham tag,
PR Cham tag, Wr. Cham dag; MK: PNB *qy?g, PMnong *dak, PKatuic *?tajig,
*?a?jig 'stand'. [The vowel also suggests aMK origin]
Roglai
x*dar 'bury', Rade dar, Jarai (PL) dar, Jarai (Lee) dar, Chru dar, N.
da, Haroi thul, PR Cham t?r,Wr. Cham dar. [The vowel suggests aMK
origin]
x*duh
'poor;
unlucky'
probably
connected
with
duh-dak; MK:
tuh 'sad', PR Cham puh-ta? '?tre de corv?e', Wr. Cham

PNB *d-an-?h. Note: The PNB contains an infix.
Cham
x*duh
honor',
'respect,
'serve',
W.
cf.
Cham
x*duh
'poor,
Chru
unlucky',
tuh, PR Cham
tuh, Wr.
duh,
Cham
N.
thu55, W.
Tsat
'serve',
duh
Roglai
hue?
duh.
x*da:g 'lie suppine; be on back', Rade dag, Jarai (Lee) dag, Chru da:g,
N. Roglai da:k, W. Cham dag, PR Cham dag, Wr. Cham dag, Malay t?r?ntang;
MK: PNB *qdlag 'lie on back', PKatuic *la-lieg 'face-up position'.
[If the
Malay form is related, it is through borrowing
also suggests aMK origin]
from aMK
x*da? 'crack open', Tsat ?da?24, PR Cham

gests aMK
source]
(Blood) da?.
[The initial
[The initial sug?
origin]
'to wedge',
Jarai (PL) d?l (Pk), Jarai (Lee) d?l, PR Cham d?l,
Wr. Cham dal, Malay s?ndal; MK: Bahnar (AC) stg. d?l. [The initial also sug?
gests aMK origin]
x*dal
x*day 'very', Chru dai, Haroi d?i, PR Cham

tial suggests aMK origin]
x*deh 'fry', Rade deh, PR Cham
(Blood) day
(Lee) padeh
-y.
(with causative
[The ini?
prefix).
[The vowel
x*di? 'climb; ascend', Acehnese e?, Rade d??, Jarai (PL) dl?, Jarai (Lee)
di:? 'climb; northerly, easterly', N. Roglai di:?, Haroi di?, W. Cham
Chru
di?,
di?, PR Cham di?, Wr. Cham dik, Malay naik, PMP *pa-nahik; MK: Bahnar
either the root
Two distinct possibilities
exist with this root:
(AC) dak.
from PMP and was borrowed into Bahnar, or the word was borrowed
into PC from Bahnar and the PMP forms are simply chance lookalikes. [The ini?
tial suggests aMK origin]
descended
Appendix
II: The Chamic
317
Lexicon
x*dih 'sleep; lie down', Acehnese

eh, Rade din, Jarai (Lee) dih, Chru
dih, Tsat di55, Haroi dih, W. Cham dih, PR Cham dih, Wr. Cham dih, Malay
tidur u?, PMP *tiduR; *tuduR. It is not likely that theMalay and PMP are related
to the PC.
[The initial suggests aMK
origin for the PC]
'pull', Rade (Tharpe) doh, Jarai (PL) d?h (Ouest et S), Jarai
(Lee) doh, Haroi had?h, PR Cham (Lee) doh. [The initial and the vowel suggest
aMK origin]
x*doh
x*dua 'carry (wear) on head', Rade dua, Jarai (PL) dua, Jarai (Lee)
Chru
dua,
dua, N. Roglai dua, Haroi doa, W. Cham doa, PR Cham dwa, Wr.
Cham du?. [The initial and the diphthong suggest aMK origin]
x*duac 'run; run off; flow', Rade due?, Jarai (PL) duai?, Jarai (Lee)
du?i?, Chru duai?, N. Roglai due?; duai?; dlai, Tsat ?doi?42, Haroi d?ai?, W.
Cham duai?, PR Cham doy?, Wr. Cham duac; MK: PNB *gad?w ?. [The initial
and the vowel also suggest aMK origin]
x*dug 'wrap up; bundle', Rade dug, Jarai (PL) pa an?g, N. Roglai duk,
Haroi d?g, W. Cham dug, PR Cham dug, Wr. Cham dug; MK: Bahnar (AC) dug.
[The initial also suggests aMK origin]
x*gan 'cross; pass over; go past', Rade g?n 'go past', Jarai (PL) g?n,
Jarai (Lee) g?n, N. Roglai gat, PR Cham k?n, Wr. Cham gan; MK: PMnong *g?n,
Bahnar (AC) pagan, PSB (Efimov) *gan.
x*ganiat 'narrow', Rade kani??, Jarai (PL) gani??, Chru gania? 'narrow;
crowded',
(N. Roglai caki??; paki??), Haroi kan?a?, W. Cham kanii? (n), PR
Cham kanl? (n),Wr. Cham ganik. [The vowel suggests aMK origin]
x*gay 'stick; wood', Rade gie, Jarai (PL) gai, Jarai (Lee) gai, Chru gai
'in various compounds', N. Roglai gai, PR Cham kay, Wr. Cham gai; MK: PKa?
tuic
*gir.
x*ge 'boat', N. Roglai ge, W. Cham ke, PR Cham ks, Wr. Cham gaiy.
vowel
[The
x*glac 'error; sin', Jarai (PL) gla??, Jarai (Lee) glai?, Chru gla:i?, N.
dla:i?, Haroi tliai?, PR Cham klay?, Wr. Cham glac. [The final suggests a
Roglai
MK origin]
x*gleh 'tired', Jarai (PL) gleh; dleh (S), Jarai (Lee) dleh; gleh, Chru
N.
gleh,
Roglai dleh, Haroi tlth, W. Cham klsh, PR Cham kleh, Wr. Cham glaih,
l?tih.
Malay
x*glo 'brain; marrow', Rade dio, Jarai (PL) dlo ak??, Jarai (Lee) dlo,
Chru glo -n, N. Roglai dlo, W. Cham I?lo,PR Cham I?lo,Wr. Cham glauw. [The
318
Appendix
II: The Chamic
Lexicon
x*glo:g 'tall; high; big', Rade dlog, Jarai (PL) dlog, Jarai (Lee) dlog
cf.
'long;
paglog 'to lengthen', Chru glo:g, N. Roglai dlo:k, Tsat khyog11, Haroi
catlug; tlug, W. Cham klag, PR Cham klag, Wr. Cham glaug. [The vowel sug?
gests aMK origin]
lhAh,
x*glah 'descend; sink; knock down; collapse, destroy', Acehnese
Chru jarlah -i, N. Roglai tadlah 'collapse', Haroi tlih, PR Cham klah, Wr. Cham
glah. [The vowel suggests aMK origin]
x*go? 'kettle; pot, glazed clay', Rade g?? 'cooking pot'; go, Jarai (PL)
g??, Jarai (Lee) g??, Chru go?, N. Roglai go?, Haroi khti?,W. Cham k??, PR Cham
k??, Wr. Cham gauk; MK: Bahnar (AC) g?. [The vowel also suggests aMK ori?
gin]
x*gu:? 'below, lower part', Rade g??, PR Cham ku? 'tirer du haut vers
le bas', Wr. Cham guk; MK: Bahnar (AC) gut. [The vowel length also suggests a
MK origin]
'return; go home; again', Jarai (PL) galal?, Jarai (Lee) gl?i?,
PR
Cham kalay?, Wr. Cham galac; MK: PKatuic *galhiAh,
kaliai?,
*yalhiAh, *ka[l/lh]a:jh. [The final also suggests aMK origin]
x*gulac
Haroi
x*guy 'carry on back', Rade gui, Jarai (PL) gui, Jarai (Lee) gui, Chru
gui, N. Roglai gui, Tsat khui11, Haroi kh?i, W. Cham kui, PR Cham kuy, Wr.
Cham gu?l; MK: PKatuic *gij, *ku:j.
x*gam
'to cover',
Acehnese
gom
'cover
with
cuplike
object;
invert
container', Rade g?m, Jarai (PL) gom, Jarai (Lee) g?m, Chru gam, N. Roglai gap,
Haroi khum; khiam, W. Cham kam, PR Cham kam, Wr. Cham gam. [The vowel
x*gag 'pole; post', Jarai (Lee) gag, Chru gag, N. Roglai gak, W. Cham
kag 'main beam', PR Cham k?g, Wr. Cham gag; MK: PNB *gag 'pole, spirit',
Bahnar (AC) g?g; g?g. [The vowel also suggests aMK origin]
gop, Rade ga? -v, Jarai (PL) ga? 'he,
x*gap 'other; group', Acehnese
N.
Chru
Jarai
she; they',
(Lee) ga-,
gau?,
Roglai -ga?,W. Cham kau?, PR Cham
MK:
Wr.
Bahnar
Cham
(AC) gap. [The vowel also sug?
k?w'(-k?n),
gap(-gan);
gests aMK
origin]
x*gar '(knife) handle', Acehnese gA, Rade gr?n -ivf, Jarai (PL) gar, Jarai
(Lee) g?r, Chru gar, N. Roglai ga, Tsat khan11, Haroi kh?l, W. Cham k?r, PR
Cham k?r, Wr. Cham gar; MK: Bahnar (AC) gar. [The vowel also suggests a
MK origin]
x*ha
see
x*?aha
Appendix
II: The Chamic
Lexicon
319
x*ha:k 'split', Rade hak 'whittle bark off, Jarai (Lee) hak, N. Roglai
ha:? 'ripped', Haroi ha?,Wr. Cham (AC) hak; MK: Bahnar (AC) hak; hek, PSB
(Efimov) *[ ]ha:? 'split, tear, divide'.
x*ha:g 'bank (river); shore', Rade hag ea, Jarai (PL) h?g, Jarai (Lee)
N.
hag,
Roglai ha:k, W. Cham hag, PR Cham hag, Wr. Cham hag; MK: PMnong
*hag, Bahnar (AC) ?ag. [<MK]
x*haduah 'look for; search', Rade duah, Jarai (Lee) haduah, Chru duah,
N. Roglai duah, Haroi athuh, W. Cham toah, PR Cham twah, Wr. Cham duah.
Jarai (PL) haget, Jarai (Lee) haget, Chru
x*haget -f 'why? what?',
PR
Cham
Wr.
Cham hagait; gait. [The vowel suggests
haket -f; k?t -f,
kayua ge,
aMK origin]
x*halua? 'sharp', Chru lah??, N. Roglai hlua?, PR Cham halwa? -i,Wr.
Cham haluak.
x*hala? 'chaff; rice dust', Chru lo:? 'remove husk', W. Cham (Headley)
to Highlands Chamic plus W. Cham. [The vowel suggests aMK
/la?/. Restricted
origin]
x*hag 'hot; spicy', Rade h?g, Jarai (PL) h?g, Jarai (Lee) h?g, Chru hag,
hak, Tsat haig?42, Haroi h?g, W. Cham h?g, PR Cham h?g, Wr. Cham
PNB *h?g, PMnong *h?g 'peppery', PSB (Efimov) *hag, PKatuic
*ha:g, *he:g 'hot'. Cf. Malay sahang 'black pepper' ?
N. Roglai
hag; MK:
PKatuic
MK
x*hago 'pine', Rade hago, Chru hago, N. Roglai hago; MK: PNB *hago,
*sagho:. Restricted to Highlands Chamic.
origin]
x*hia 'cry; weep', Rade hia, Jarai (PL) hla, Chru hia, N. Roglai hia,
Tsat hia33, Haroi he a, W. Cham hea, PR Cham hya, Wr. Cham hi?. [The vowel
x*ho? 'sweat; bleed; leak', Rade kah??, Jarai (Lee) h?? 'to bleed', W.
Cham hau? -v, PR Cham (Lee) h??, Wr. Cham (AC) huak; MK: PSB (Efimov)
*hu:c 'drink; leak'. [The vowel also suggests aMK origin]
x*hok 'pour out; spill', Rade h?k, Jarai (Lee) hok 'to abort', Chru ho?
h?? 'to spill, pour', PR Cham h??,Wr. Cham hauk; MK: PNB *?k,
Haroi
'spill',
PKatuic *[t/d]aho?, *haho?. [The vowel also suggests aMK origin]
x*hua 'pull', Jarai (PL) h?a (Ouest), Jarai (Lee) hua, Chru hua 'pull
along' hwa -vr, W. Cham hoa, PR Cham hwa, Wr. Cham hu?. [The vowel sug?
gests this is of MK origin]
320
Appendix
II: The Chamic
Lexicon
x*hua? 'eat rice', Rade hu??, Jarai (PL) hu??, Jarai (Lee) hua? -?, Chru
Tsat
hua?,
hua?24,W. Cham hoa?, PR Cham hw??, Wr. Cham huak. [The vowel
suggests aMK origin.]
x*hual 'cloud, fog, mist', Chru huai, N. Roglai huan, PR Cham hoi, Wr.
'smoke'. Cf. xdhual/r -f; here the same
huai; MK: PKatuic *[h/y]ahuol
a
but
with
different
etymon,
probably
pr?fixai element, has been borrowed more
Cham
than once into Chamic.

x*hayua? 'harvest (rice)', Chru yua?, N. Roglai riyua?, Haroi y??, W.
Restricted toHighlands Chamic plus Haroi andW. Cham. [Both the
Chamyoa?.
vowel and the restriction suggest this isMK in origin.]
x*huret 'rope; vine', Acehnese
uret, Rade hru?? (m), Jarai (PL) hr??
hru??
N.
Jarai
(Lee)
(Hd.),
(m),
Roglai hure?, Haroi har??, W. Cham hr??. [The
Acehnese
form
suimuiguip [Note: This Acehnese
x*ha?a:p'to yawn',
looks like it is from *sama + *gap], Rade haap, Jarai (Phraya Prachakij-karacak)
the first morpheme],
?aap, Chru sama?:?, N. Roglai sama-??:? [Note: Notice
Haroi ha-?au? [Notice that the word is treated as having two morphemes],
PMa?
POc
PMP
PSB
*huab;
(Efimov)
layic *uap,
(*ma)huab,
gga:p, PNB
*mawap,
Vietnamese
PKatuic
ng?p,
*hag?ua:p.
*kaqa55p,
'
x*jaray
jarai, W.
Cham
Jorai',
Rade
jarai
-v-i,
Chru
manih
jarai
-i, N.
Roglai
man?ih
?arai.
x*jaw 'to deliver', Jarai (PL) ja?, Jarai (Lee) jau 'to trust', PR Cham
Wr.
Cham jaw. With the *pa- 'causative' prefix: x*pajaw, Jarai (Lee)
?aw,
N.
Roglai pajau.
pajau,
x*je? 'near; about to' Rade je?, Jarai (PL) je?, Jarai (Lee) je? -1, Chru
je?; maje? 'near', N. Roglai je? si 'about to'; je? 'near', Tsat se?42, Haroi st?,W.
Cham ???; ma??? 'nearly', PR Cham ce?, Wr. Cham jaik; MK: PNB *ajeq
'near'. [The vowel also suggests aMK origin]
x*joh 'broken; spoilt', Rade joh, Jarai (PL) j?h, Jarai (Lee) joh, Chru
N.
joh,
Roglai joh, Haroi suh, W. Cham joh 'snap', PR Cham ?oh, Wr. Cham
[The
jauh.
x*ju:? 'black', Rade j??, Jarai (PL) j??, Jarai (Lee) j??, Chru ju:?, N.
Roglai ju:?, Haroi su?,W. Cham cu?, PR Cham cu?,Wr. Cham juk; MK: Bahnar
[The vowel length also suggests a
(AC) j?, PSB (Efimov) *ju:? 'black, dark'.
MK origin]
x*jua? 'step on; tread', Rade ju??, Jarai (PL) ju??, Jarai (Lee) ju??, Chru
jawa?, N. Roglai jua?, Tsat sua?42, Haroi so? -i,W. Cham ?oa?, PR Cham cw??,
Wr. Cham juak; MK: PMnong *jot 'trample', PSB (Efimov) *jo:t. [The vowel
also suggests aMK
origin]
Appendix
II: The Chamic
321
Lexicon
x*jum 'around' cf. x*pajum 'meet together', Rade j?m, Jarai (PL) jum,
Jarai (Lee) j?m 'around', N. Roglai pajup 'assemble, gather', PR Cham yom;
MK: PMnong *j?m.
x*juay 'don't', Chru juai, W. Cham ?uai, PR Cham ?oy, Wr. Cham juai.
x*jah 'PARTICLE', Rade jih 'completely'
'vraiment'.
PR Cham ?ah,Wr. Cham jah
'firstborn', Jarai (Lee) kacua, N. Roglai kacua, Haroi kacoa

'oldest child', W. Cham kacoa, PR Cham kacwa, Wr. Cham kacu?. [The vowel
x*kacua
x*kacua 'to spit', Jarai (Lee) kacua, PR Cham

lable of the original has dropped).
x*k-am-lo
'dumb;
mute',
cf. Mnong,
Acehnese
(Lee) cuwa
klo,
Rade
(the first syl?
Jarai
kamlo,
(PL) kamlo, Chru kamlo, N. Roglai kumlo, Haroi kamlo, W. Cham kam 1?, PR
Cham kamlo, Wr. Cham kamlauw; MK: PMnong
*kamlo, PSB (Efimov)
*kamlo:. Notice particularly the lack of the infix in the Acehnese
form, and the
treatment of the word as two separate morphemes
inW. Cham.
[The vowel also
x*ka:g 'chin; jaw', Acehnese

kuiag, Rade kag, Jarai (PL) kag, Jarai
(Lee) kag, Chru talka:g; ka:g, N. Roglai ka:k, Tsat ka:g33, Haroi kag, W. Cham
kag, PR Cham kag, Wr. Cham kag, Malay rahang ?; MK: PNB *kag, PMnong
*kag 'chin', PSB (Efimov) *ka:g, PKatuic *ta?ba:g 'jaw'.
x*kadug 'pocket; pouch', Jarai (Lee) kadug, Chru kadug 'bag', Tsat
Haroi
kath?g, W. Cham katug 'bag', PR Cham (Blood, Lee) kat?g, Malay
thug11,
MK:
PNB *gad?g, Bahnar (AC) kadug, PKatuic *kando:g, *kadug,
kandung;
'basket,
*3adug
[Note: The
bowl',
Katu
AD
3adug
'men's
basket
worn
around
waist'
form occurs
'pouch'. Despite
inMK, PC, and Malay, with the core meaning being

its occurrence inMalay, this form looks ultimately to be MK in
origin.].
x*ka?u:? 'barkcloth; tree species', Rade (Lee) ka???, Jarai (Lee) ka???,
N. Roglai (Lee) ka?u:?. Note the restriction to Highlands Chamic.
[The long
Cham
x*ka?u:? 'worried; sad', Jarai (Lee) ka???, Chru ka??:?, Haroi ka?ou?, PR
(Lee) ka?u?. [The vowel length suggests aMK origin]
x*kabac
Restricted
(an itch)', Chru kaba:i?, Haroi kaphiai? 'scratch'.

Chamic plus Haroi.
[The final also suggests aMK ori?
'scratch
to Highlands
gin]
322
II: The Chamic
Appendix
Lexicon
x*kabaw 'water buffalo', Rade kabau, Jarai (PL) kabau, Jarai (Lee)
kabau, Chru kaba:u -1, N. Roglai kabau, Haroi kaphiau, W. Cham kabau, PR
Cham kapaw, Wr. Cham kabaw, Malay k?rbau; MK: Bahnar (AC) kap?; kap?,
PSB (Efimov) *'grapu:. [Despite its occurrence in various Malayic dialects, this
word
looks to be aMK
borrowing]
x*kacuh 'to spit', Rade kacuh, Jarai (Lee) kacuh, Chru kacuh, Haroi
-f,W. Cham tacuh -i,Wr. Cham (AC) kacuh, PNB *kacuh, PSB (Efimov)
*choh, *jhoh, PKatuic *kacoh, *kaco:h.
kacoh
'hawk; bird of prey', Acehnese

kluiag 'kite' -i, Rade tlag -i
'bird of prey; hawk', Jarai (PL) klag, Jarai (Lee) klag 'eagle, kite', Chru kala:g
'bird of prey', N. Roglai kala:k 'bird of prey', Haroi kalag, PR Cham kalag, Wr.
x*kala:g
Cham
kalag, Malay
h?lang; MK: Bahnar

(1976) also suggested
*kalha:g. Headley
later borrowed into a handful ofWMP
(AC) klag-an, PSB *tsalag, PKatuic

that this was originally aMK etymon,
languages.
gapuiah, Rade kapaih, Jarai (PL) k?paih,

x*kapa:s 'cotton', Acehnese
Jarai (Lee) kapaih, Chru kapa:h, N. Roglai kapa, Tsat pa33, Haroi kapah, W.
PNB
Cham kapah, Malay kapas, PWMP *kapes 'cotton, Gossypium
;
spp.' MK:
*kapayh, PKatuic *kapa:jh, PKatuic *kapa:jh; *hapa:jh. Note that the PC vowel
does not match the PWMP, but does match the PNB and the PKatuic. Inciden?
tally,
Lee's
apparent
reconstruction
of
length
in the onset
syllable
is simply
a mis?
take, probably a typing error.

[The origin of this word is a subject of
an
like
This
word
looks
commentary.
early loan intoWMP, possibly from Indie.
Headley (1976:#2.1) notes possible Sanskrit and Hindi sources, cf. Sanskrit kar
(1989) analyzes it as an Austroasiatic
that
includes MK.]
larger family
paasa. Baxter
borrowing,
that is, from the
'to close', Rade (Lee) kapuat, kap?t, Jarai (Lee) kap?t, N.

Roglai (Lee) kapi:?, Haroi kap?a?. Note the restriction to Highlands Chamic,
plus Haroi. [The diphthong suggests aMK origin]
x*kapuat
x*kapuat ? 'handful', Rade kap?t, Jarai

[The vowel
kapoa?. Restricted to Highlands Chamic.
x*karah 'ring', Rade krah, Chru kr?h -n, N.
W. Cham karah, PR Cham kar?h, Wr. Cham karahMK origin]
x*katar
'corn;
grain',
(Lee) kate, Haroi katol. Note

Chamic languages plus Haroi.
Rade
ka tar,
Jarai
(Lee) kap?t, W. Cham ha

Roglai karah, Haroi kr?h,
[The nasalization suggests a
(Lee)
katar
'corn',
N.
Roglai
the restriction of the distribution to the Highlands

[The vowel also suggests aMK origin]
x*ke? 'bite; snap at; peck', Rade k??, Jarai (PL) k??, Jarai (Lee) k??,
Chru ke? 'bite', N. Roglai ke?, Haroi k??,W. Cham k??, PR Cham k??,Wr. Cham
kaik. [The vowel suggests aMK origin]
II: The Chamic
Appendix
323
Lexicon
x*khan 'cloth; blanket', Jarai (PL) kh?n, Jarai (Lee) kh?n, Chru khan,
N. Roglai khat 'cotton cloth', W. Cham kh?n, PR Cham kh?n, Wr. Cham khan,
Malay kain; MK: PNB *kh?n 'blanket', Bahnar (AC) kh?n.
x*khia? 'burnt smell', Rade khi?? 'crust at bottom of pot', Jarai (Lee)
khi??, Chru khia?, N. Roglai khia? 'scarred, scorched', Haroi kh?a? 'burnt
(rice)', PR Cham khy??, Wr. Cham khiak. [The diphthong suggests aMK origin]
x*kho:g 'dry (weather)', Acehnese khuag 'drought' (Durie notes varia?
tion in the nasalization), Rade khog 'end of rain', Jarai (Lee) khog, Chru kho:g
'fair', N. Roglai kho:k 'dry, sunny', Haroi khog 'refers to when it has finished
raining', PR Cham khog, Wr. Cham khaug. [The vowel suggests aMK origin]
x*kla:s
cf.
'escape',
Acehnese
'lose',
lhuiah,
Rade
Jarai
tlaih,
(PL)
klaih, Jarai (Lee) klaih, Chru kla:h, N. Roglai tla, Haroi tlah, W. Cham klah 'to
free', PR Cham kl?h, Wr. Cham klah-clm ti pay; MK: Bahnar (AC) klah, PKa?
tuic
*-lah.
*-klah,
(DT)
With
*pa-
causative
prefix:
'escape,
cause
to;
save;
'free', Rade tlaih, Jarai (PL) ta klaih

x*pa-kla:s, Acehnesepui-lhuiah
dual? (Hd.), Jarai (Lee) paklaih, Chru pakla:h, N. Roglai tatlah, PR Cham
pakl?h, Wr. Cham paklah; MK: PMnong *klas 'release'. With *pa- causative
release'
'take
prefix:
apart,
dismantle'
Rade
*pa-kla:s,
matlaih,
N.
Roglai
Cf.
patla.
'escape'.
x*klah
'to
lose',
cf.
Jarai
'escape',
(Lee)
klah,
Haroi
tlah
'to
slip
away
or escape, e.g. chicken from being tied', W. Cham lah, PR Cham (Lee) klah;
MK: PKatuic (DT) *-klah, *-lah.
[Note: This looks to be a doublet of the ety?
mon
for
'escape'].
uluia lhan, Rade tl?n, Jarai (PL) tian,

x*klan 'boa; python', Acehnese
kl?n, Jarai (Lee) kl?n, Chru klan 'python', N. Roglai tlat, Haroi tl?n 'boa', W.
Cham kl?n, PR Cham kl?n, Wr. Cham klan; MK: Bahnar (AC) kl?n, PSB (Efi?
mov)
*klan.
x*klay
'penis',
Chru
klai,
N.
Roglai
pitlai
'testicles',
Haroi
tlai, W.
Cham klai, PR Cham klay, Wr. Cham klai, PSB (Efimov) *klAw 'man, male,
penis', PKatuic *[k/g]alh[e/a]j.
x*kle? 'steal', Rade tie?; kan??, Jarai (PL) kl??, Jarai (Lee) kl??, Chru
kle?, N. Roglai tie?, Tsat ke?24 -m, W. Cham kl??, PR Cham kl??, Wr. Cham
klaik; MK: Bahnar (AC) kl?. [The vowel also suggests aMK origin]
x*kloh 'to break', Jarai (PL) kl?h, Jarai (Lee) kl?h -1, Chru kloh 'cut
(skin)', N. Roglai tloh, Haroi tl?h, W. Cham kloh 'cut (skin)', PR Cham kl?h,
Wr. Cham klauh. [The vowel suggests aMK origin]
x*kluan
klon, N. Roglai
'behind; buttocks', Rade tluon, Jarai (PL) klon, Jarai (Lee)

tluat 'buttocks', Haroi tl?n, PR Cham kl?n, Wr. Cham klaug;
324
Appendix
(DT) *tantuun 'behind'. Note

cal. [The vowel also suggests aMK origin]
MK: PKatuic
x*klun
-f 'tadpole',
II: The Chamic
that theMK
is spatial, not anatomi?
tl?n, Chru klu:t -f -1; MK:
Rade
Lexicon
PSB
(Efimov)
*kltun.
x*klap 'stab; poke', Rade tl??, Chru klau?, N. Roglai tla?, Haroi tl?u?,
Wr. Cham (AC) klap; klap; MK: PMnong *t?p, Bahnar (AC) klak, kl?k, kl?p,
kl?k. [The vowel also suggests aMK origin]
x*ko:? 'white', Rade ko?, Jarai (PL) ko, Jarai (Lee) ko? -1,Chru ko:?, N.
PR Cham k??;
Haroi ko?, W. Cham ko? in patih ko? 'very white',
ko:?,
Roglai
ak?? -1,Wr. Cham kauk; MK: PKatuic (DT) *klook. [The vowel also suggests a
MK origin]
x*ko:g 'bracelet', Rade kog, Jarai (PL) kog, Jarai (Lee) kog 'brass',
Chru ko:g, N. Roglai ko:k, Haroi kog, PR Cham kog, Wr. Cham kaug; MK: PNB
PKatuic (DT)
*kog, PMnong *kog, PSB (Efimov) *kAi^ 'copper bracelet',
*k?ng, PKatuic
*hakag; *hagkag.
origin]
x*koh 'cut off; shorten', Acehnese koh, Rade koh, Jarai (PL) koh, Jarai
(Lee) koh, Chru koh, N. Roglai koh, Haroi koh -1,W. Cham koh 'to lop off.
x*kro 'dry', Rade kro, Jarai (PL) kro rara, Jarai (Lee) kro, N. Roglai kro,
kro; MK: PMnong *kro, PSB (Efimov) *ro. Note that this is limited to
Highlands Chamic and Haroi among the Chamic languages, but reconstructs in
Haroi
two of the Bahnaric
branches.
kruag, Rade krog, Jarai (PL) krog, Jarai (Lee)

x*kro:g 'river', Acehnese
krog, Chru ia kro:g 'stream', N. Roglai kro:k, W. Cham karog, PR Cham krog,
Wr. Cham kraug; MK: PMnong *krog, Bahnar (AC) krog, PKatuic *karhuag.
x*kruac 'a citrus', Rade kru??, Jarai (PL) boh kra?? e, boh kr?al? (E)
Jarai
(Lee) kru?i?, PR Cham kroy?, Wr. Cham kruac, Proto-Hr?-Sedang
(Hd.),
*kruc; MK: PMnong *kroc, PSB (Efimov) *kruac. [The vowel and the final also
suggest aMK
origin]
x*kuac 'gather, amass',

to Highlands Chamic
Restricted
suggest aMK origin]
x*kuah
'shave,
scrape',
Chru kuai?, N. Roglai kuai?, W. Cham kuai?.

plus W. Cham. [The vowel and the final also
Rade
kueh
-v,
Jarai
(Lee)
kuah,
Chru
kuah,
N.
Roglai kuah, Haroi koah, W. Cham koah; MK: PNB *akoyh 'shave', PMnong
*kos 'shave off, PSB (Efimov) *ko:s, PKatuic *kuah, *kuajh 'scrape, shave'.
a
gests MK origin]
Appendix
II: The Chamic
325
Lexicon
'gibbon', Rade kue? -v, Jarai (PL) kra ku?ft, Chru kuan, W.
Cham kuan, PR Cham kon, Wr. Cham kuan; MK: PSB (Efimov) *kua1ji. [The
x*kuan

is 'lungs; placenta', Acehnese
*kulit 'skin' + x*so:? 'lungs; placenta'
Jarai
Rade
Jarai
kas??,
(PL) kas??,
(Lee) kas??, Chru kalso:?, N. Roglai
s?ap -fn,
kuli:? so:?, Haroi kalaso?, W. Cham kaso?, PR Cham tho? 'placenta', Wr. Cham
isAustrone?
thauk 'placenta; lungs' MK:

PKatuic *sah ?. [The first morpheme
;
sian; in the second, the vowels suggest aMK origin]
x*la:g 'spread out (a mat)', Rade lag, Jarai (PL) lag, Jarai (Lee) lag,
Chru la:g 'unroll', N. Roglai la:k, W. Cham lag, PR Cham lag, Wr. Cham lag;
MK: PMnong *l?k -f 'spread', Bahnar (AC) lag, PSB (Efimov) *la? -f 'spread,
unroll'.
x*la?i 'basket, winnowing',

Rade ei, Jarai (PL) ra?i, Jarai (Lee) ra?i,
la?i 'basket (shallow)', N. Roglai la?i, Haroi la?i, W. Cham la?i 'large
round', PR Cham lii,Wr. Cham lit; MK: PSB (Efimov) *la?i:, PKatuic *?arie,
Chru
*karie.
x*lac 'say', Rade lac, Jarai (PL) lal?, Jarai (Lee) l?i?, Chru la:i?, N.
Haroi lai?, W. Cham lai? cf. 'if, PR Cham lay?, Wr. Cham lac,
la:i?,
Roglai
[The final also suggests a MK
PMnong *lah, PSB (Efimov) *lah 'say, scold'.
origin]
x*lahia? 'to lose', Rade lue, Chru lah?a?,W. Cham lahii? (n). Restricted
toHighlands Chamic plus W. Cham. [The vowel and its nasalization also suggest
aMK origin]
x*laman
lam?n
-v
'weak,
MK: PKatuic
-v
soft,
'tired;
Rade
weak',
bloated',
PR
(DT) *qadl_h 'tired'.
em?n,
Cham
liman
Jarai
-v
(PL)
'feeble',
r?m?an,
Wr.
W.
Cham
Cham
liman;
x*le? 'fall into', Rade le? buh, Jarai (PL) le?, Jarai (Lee) le?, Chru le?, N.
le?, Tsat le?24,Haroi le?,W. Cham le?, PR Cham le?,Wr. Cham laik. [The
Roglai
vowel suggests aMK
origin]
x*lo:k 'to peel', Acehnese plua?, Rade lok 'take bark off tree', Jarai
(Lee) lok -v, Chru lo:?; lo:h -f, N. Roglai lo:?; calo:?, Haroi lo?,W. Cham lo?, PR
Cham lo?,Wr. Cham lauk, PMP *bulut 'coconut husk'; MK: Bahnar (AC) l?k,
PSB (Efimov) *plo:? 'peel, strip off skin, hull', PKatuic
x*lo:g 'try, prove,
Cham
test', Rade
(Blood) log, Wr. Cham laug; MK:

*liet, *luot.
[The vowel
log, Chru parlo:g 'try, prove, test', PR

PNB *log, PMnong *ralog 'try, test'.
326
Appendix
II: The Chamic
Lexicon
x*luay 'swim', Rade lue, Jarai (PL) loi -v, lua? (Hd.), Jarai (Lee) luai,
luai ia, N. Roglai luai, Haroi luai -v, PR Cham loy, Wr. Cham luai,
[The vowel also
PMnong *re ?, PSB (Efimov) *re:, PKatuic *[b/?b]aluo[j]h.
Chru
suggests aMK
origin]
x*luc 'complete; die, end; perished', Rade lue 'lose'; lue lie 'get lost',
Haroi lui? -v,W. Cham luai -vf 'stop, quit', PR Cham l?y?,Wr. Cham lue. [The
final suggests aMK origin]
Haroi
x*luay 'put, place, discard; allow', Rade lui, Chru luai, N. Roglai
lui,Wr. Cham buh. [The vowel suggests aMK origin]
luai,
x*luay 'reject; quit', Rade lui, Jarai (PL) lui, Jarai (Lee) lui, Chru luai,
W. Cham luai, PR Cham loy, Wr. Cham luai. [The vowel suggests aMK origin]
x*lap 'fold', Jarai (PL) lap (Ouest); lap (S), Chru lau?, PR Cham law?,
Wr. Cham lap; PMP *le(m)pit -f; *lepet -f 'wrap; fold'. Despite the similarities,
the PMP forms are not related. [The vowel suggests aMK origin]
x*maduan
see
x*?aduan
'awaken', Rade madih, Jarai (PL) padih, madih (Hd.), Jarai

(Lee) madih, Chru m a da h 'stay awake', N. Roglai madah, Haroi mathih, W.
Cham matan, PR Cham mitah, Wr. Cham midah- [The vowel suggests aMK ori?
x*madah
gin]
x*miag 'cheek; jaw', Acehnese miag, Rade mieg, Jarai (PL) meg, Jarai
(Lee) m?g; meg, W. Cham miag, PR Cham mieg, Wr. Cham mieg. [The diph?
thong suggests aMK origin]
x*mo:? 'wife', Rade mo?, Jarai (PL) m?? (Est), N. Roglai m?:? 'mid?
wife', Haroi mo?, W. Cham m?? -v, PR Cham (Lee) mu? -v. [The vowel suggests
aMK origin]
x*mag 'from', Rade mag, Jarai (PL) mag, Jarai (Lee) mag, Chru mig -v
-n (prep.), N. Roglai muni -vf, Haroi mig (grammatical particle), W. Cham mag,
PR Cham mig, Wr. Cham mig. [The vowel suggests aMK origin]
x*ftu? 'dive; submerge', Rade ???, Jarai (PL) paf???; ???, Jarai (Lee)
???, Chru ?u? ia, N. Roglai ???, Haroi ??k -f,W. Cham ??k -f 'submerge', PR
Cham ??k -f,Wr. Cham ?uk; MK: PSB (Efimov) *jiap 'dive, submerge, set' ? .
Cham
x*neh 'to elbow', Rade (Tharpe) ?eh -n, N. Roglai

(Lee) ?eh. [The vowel suggests aMK origin]
?eh -n julukiac,
PR
x*go? 'upgrade; above; east', Jarai (PL) g?? 'east', gah yang hrai ?le?,
Jarai (Lee) g?? 'east', Chru gah go? 'east', Tsat go?42/24 -t?, Haroi g??, PR Cham
(Lee) g??,Wr. Cham (AC) gauk. [The vowel suggests aMK origin]
327
Lexicon
II: The Chamic
Appendix
x*padiag -v 'carry; transport', Rade madi?g -v, Jarai (PL) bSdiag -i,
Chru pardiag, PR Cham pa?y?g, Wr. Cham pajiag; MK: PMnong *padi?g, PKa?
tuic (DT) *pataeng 'transport'.
[The diphthong also suggests aMK origin]
x*pagam 'dove', Jarai (PL) baragom ?, Haroi pakhum. Note the restric?
tion to the Highlands Chamic languages, plus Haroi. [The vowel suggests aMK
origin]
x*pah 'to slap', Rade pah, Jarai (Lee) pah, Chru pah, N. Roglai pah,
Haroi p?h, W. Cham pah, PR Cham p?h, Wr. Cham pah; MK: PKatuic *[h/
*habah.
s]am[p/b]ah, *[h/y]apah,
-f 'tell,
x*padar
send',
Haroi
demand,
-f
padai
summon'.
send;
Chru
command',
'command,
[The medial
to
tell
consonant
to
do
-1, N.
padar
sth',
W.
suggests aMK
Roglai
Cham
'tell,
pada
pad?r
'order,
origin]
x*padaw 'warm, hot', Rade madau, Jarai (PL) padau, Chru padau,
Tsat dam?42, Haroi padau 'be warm', W. Cham padau, PR Cham padaw, Wr.
Cham padaw; MK: PKatuic *[k/g]ata:w, *[k/g]ataw 'hot'. [The medial conso?
aMK
nant also suggests
origin]
x*padia? 'hot (weather); sunshine', Jarai (Lee) pa?i??, Chru padia?, N.

Roglai padia?, Tsat dia?24, Haroi pad?a?, W. Cham padea?, PR Cham pady??,
Wr. Cham padiak.
[The medial consonant and the vowel suggest aMK origin]
x*pale?
Restricted
pale?.
MK origin]
'to drop', Rade kapl??, Chru pale?, N. Roglai pale?, Haroi

to Highlands Chamic plus Haroi.
x*pet 'pick, pluck', Acehnese pAt, Rade p??, Jarai (PL) p??, Jarai (Lee)
Chru
p??,
pe?, N. Roglai pe?, Tsat pi55 -ft, Haroi p??, W. Cham p??, PR Cham
p??,Wr. Cham paik. [The vowel suggests aMK origin]
x*pha 'different', Jarai (Lee) pha, N. Roglai
Restricted
*pha.
pha, Haroi pha; MK: PNB
x*picah 'broken; break', Acehnese picah, Rade mcah, Jarai (PL) p?cah,
Jarai (Lee) pacah, Chru pacah, Tsat tsa55, Haroi pac?h, W. Cham pacah, PR
Cham pac?h, Wr. Cham pac?h, Malay p?cah; MK: PMnong *bacah 'break',
PSB (Efimov) *'pacah 'split, smash, break', PKatuic *pac[a/a]h,
*kac[a/a]h
'crack'. Note
me
that this word
like it originated
is found both inAustronesian
and MK,
but looks to
inMK.
x*pioh 'put, place', Rade pion, Chru pion, W. Cham mapiah. Restricted
to Highlands Chamic plus W. Cham. The vowel suggests that it is either bor?
rowed or the result of the coalescence of two morphemes.
x*pit 'sleep; close eyes', Rade pit, Jarai (PL) pit, Jarai (Lee) pit, Chru
'shut
pi:?
eyes', N. Roglai pi:?, Haroi pei?, PR Cham pi?,Wr. Cham pik 'mourir;
328
II: The Chamic
Appendix
fermer
l'oeil'; MK:
PKatuic
PNB
*qb?c 'lie down',
*bic, PSB
PMnong
Lexicon
*bik,
(Efimov)
*?bvj?, *?bAJ?.
-1 'land leech', Rade plum -1, Jarai (PL) plum, Chru plo:m -vl,
W.
Cham plom -v, PR Cham plom, Wr. Cham plorri; MK: PNB
plu:p,
*pl?m, PMnong *plom, PSB (Efimov) *pU:m, PKatuic *balhA:m, *palhA:m.
7*plum
N. Roglai
poh, Chru poh, N. Roglai p?h 'to

x*poh 'strike; pound', Acehnese
[The vowel suggests aMK origin].
pound', PR Cham pap?h, Wr. Cham papauh.
With
*-an-
instrumental
infix:
Chru panoh, N. Roglai
'drumstick'
pan?h, Haroi
cf.
x*p-an-oh
panah
'to drum'.
x*poh,
hanoh
Rade
[The vowel
-i,
suggests
MK origin]
x*pok 'to open',
po?; MK:
PMnong
Rade p?k, Jarai (PL) p?k, Jarai (Lee) p?k, N. Roglai

*pak, PSB (Efimov) *pa:?. [The vowel also suggests a MK
origin]
x*pro:k 'squirrel', Rade prok, Jarai (PL) pro, Jarai (Lee) pr??, Chru
Haroi
pro:?,
pro?, W. Cham pro?, PR Cham pro?,Wr. Cham prauk; MK: PMnong
*prok, Bahnar (AC) pr?k, PSB (Efimov) *pr?:?> PKatuic *[b/?b]a(r/rh]og, *ta[r/
rh]og. [The vowel also suggests aMK origin]
x*prog 'big', Rade pr?g, Jarai (PL) pr?g, Jarai (Lee) pr?g, Chru prog, N.
Roglai prok, Tsat piog?33, Haroi pr?g, W. Cham prug -v, PR Cham pr?g, Wr.
Cham praug. [The vowel suggests aMK origin]
x*pu:? 'carry in arms', Rade p??, Chru po:?, N. Roglai pu:?, Haroi po?,
W. Cham po?, Wr. Cham (AC) pu?; MK: Bahnar (AC) p?k. [The vowel length
'blow away; chase',
Jarai (PL) p?p?h,
(*pa- 'causative')
Jarai (Lee) papuh 'chase', Chru pun 'blow', N. Roglai papuh; MK: PMnong
*p?h 'blow'. Restricted to Highlands Chamic.
x*pa-puh
x*puley 'gourd; squash', cf. *biluay, Rade plei, Jarai (PL) plui (m),
Jarai (Lee) plai, plui, ploi, N. Roglai pluai (m), W. Cham ploi (m), PR Cham
ploy (m), Wr. Cham pluai -v;MK: PNB *pl?y.
x*pu:g
'straw
(rice)',
Acehnese
Chru
juimpug,
W. Cham pog, PR Cham apy?g, Wr. Cham apiag.

MK origin]
Tsat
apo:g,
[The vowel
'rice',
pug33
length suggests
x*puac 'scold; talk', Jarai (PL) p?al?, Jarai (Lee) pu?i?, Chru puai?, N.
Roglai puai? 'strong feelings; impatient', Haroi p?ai?, W. Cham puai?, PR Cham
poy?, Wr. Cham puac. [The final suggests aMK origin]
x*pa-pat
Roglai
pa?, Haroi
'causative')
'to
papa?; MK:
PNB
(*pa-
fan',
Jarai
*p?y,
(Lee)
PKatuic
papi?
N.
'to vibrate',
*[h/y]api?;
*?api?.
Appendix
II: The Chamic
Restricted
to Highlands
329
Lexicon
Chamic
plus Haroi.
ori?
gin]
W. Cham pah, PR
x*pah 'to open', Chru pah 'open up', Haroi p?h,
Cham p?h, Wr. Cham pah; MK: Bahnar (AC) paha, PKatuic (DT) *paah. [The
Chru pag; pa:g
x*pag 'to nail; to hammer', Rade p?g, Jarai (Lee) p?g,
'to pound', N. Roglai pak, Haroi pag, W. Cham pag, PR Cham pag 'frapper sur
la tete', Wr. Cham pag, PMP *paku; MK: PMnong *p3g. [The vowel also sug?
gests aMK origin]
It is only the Rade
x*par 'to fly', Acehnese phA -i, Rade phiar -iv Note:
that looks to be borrowed after PC, Jarai (Lee) par, Chru par, N. Roglai pa, Tsat
Wr. Cham par; MK: PNB *p?r,
pan33, Haroi pol, W. Cham par, PR Cham par,
PMnong *p?r, Bahnar (AC) ap?r; par, PSB (Efimov) *par, PKatuic (DT) *p?r,
PKatuic
*par, *pa:r.
'Rhade', Rade ede, Chru manih rade, N. Roglai man?ih rade,

PR
Cham rade, Wr. Cham rade. [The vowel and the meaning
rade,
a
suggest MK origin]
x*ralo 'flesh; meat', Rade hlo, Jarai (Lee) hlo, Chru ralo, N. Roglai ralo
x*rade
W. Cham
elf., W. Cham ralo, PR Cham ralo, Wr. Cham ralauw.
[The vowel suggests aMK
origin]
rhah -iv, Rade rau, Jarai (PL) rau, Jarai (Lee)
x*raw 'wash', Acehnese
ra:u
Tsat
N.
Chru
rau,
za:u33, Haroi rau,W. Cham rau, PR Cham
-1,
Roglai rau,
raw, Wr.
Cham
rau; MK:
(Efimov) *raw, PKatuic
PNB
*-raw,
PMnong
*raw,
Bahnar
(AC)
rao,
PSB
*?ariaw.
x*reh 'cut', Jarai (Lee) r?h, Chru reh 'clean fish', N. Roglai reh 'oper?
ate, dissect', Haroi r?h, PR Cham (Lee) reh, PMnong *sreh, PSB (Efimov)
*sre:h
'chop',
suggests aMK
PKatuic
*hareh,
*tareh,
*[s/c]arhe:?,
*tarhe:?.
[The
vowel
also
origin]
x*ribui? 'storm', Rade eb??, Jarai (Lee) rab??, Chru rabu:?, N. Roglai
Haroi
rubu:?,
laphu?, W. Cham gin rapu?, PR Cham ripu?; rapu?,Wr. Cham rib?k,
Malay ribut; MK: Bahnar (AC) habut, PKatuic (DT) *rapuuq. Within wider
this form is restricted to languages in western Indonesia and so is
Austronesian,
most
likely also a loan into these An
languages.
x*ro 'cage', Jarai (PL) ro (Pk), Jarai (Lee) ro, Chru ro, N. Roglai
Haroi ro,W. Cham ro. [The vowel suggests aMK origin]
ro,
x*ro:g 'raise; nourish', Rade rog 'take care of livestock', Jarai (PL) rog,
Jarai (Lee) rog, Haroi rog 'raise, feed', PR Cham rog,Wr. Cham raug; MK: Bah?
nar (AC) r?g. [The vowel also suggests aMK origin]
330
Appendix
II: The Chamic
Lexicon
x*ruah 'choose', Rade ruah 'election', Jarai (PL) ru?h, Jarai (Lee) ruah,
Chru rawah, Haroi roah, PR Cham rw?h, Wr. Cham ruah; MK: PNB *r?yh, PKa?
tuic (DT) *r_s (vowel unclear), PKatuic *hariah. [The diphthong also suggests a
MK origin]
x*ruay
Chru
ruai
-n, N.
ruai, PR Cham
'fly; bug; insect',

ruai,
Roglai
Tsat
Rade
rue, Jarai (PL) rual, Jarai (Lee) ruai,

a11
za:i33;
zuai33;
roy, Wr. Cham
roai, W.
la?24, Haroi
Cham
PNB
ruai; MK:
*roy, PMnong *rahway, PSB
(Efimov) *rahwa:y, PKatuic (DT) *r??y, PKatuic *[h/y]aruaj, *ra-ruaj, *?aruaj.
x*ruay 'to crawl', Rade rui, Jarai (PL) rui, Jarai (Lee) rui, Tsat zoi33,
Haroi rui,W. Cham ruai. [The triphthong suggests aMK origin]
x*rak
Rade
weeds',
'grass;
rak,
Jarai
(PL)
rak,
Chru
ra?
weeds',
'grass;
N. Roglai ra?, Tsat za?24, Haroi ra?,W. Cham ra?, PR Cham ra?; hara?, Wr. Cham
rak; harak. [The vowel suggests aMK origin]
suiat
'bail (water to catch fish); splash; shake out', Acehnese
sac ea 'bail water to catch fish', Jarai (Lee) s?i?, Chru sa:i?, N.
sa:i?, Tsat sai?24, Haroi s?? -v,W. Cham sai?, PR Cham thay?, Wr. Cham
x*sac
'bail', Rade
Roglai
thac. [The final consonant
x*sadar
hadar,
Chru
'remind;
'remember',
sa dar,
cause
suggests aMK
N.
Roglai
to remember',
vowel suggests aMK
Rade
sida,
PR
hadar,
Haroi
Cham
origin]
Jarai
athul
hat?r;
(PL)
'feel,
t?r, Wr.
hadar,
sense',
Cham
Jarai
W.
(Lee)
Cham
hadar;
hadar;
pa-d?r
dar.
[The
origin]
x*sagar 'drum', Rade hagar, Jarai (PL) hag?r, Jarai (Lee) hag?r, Chru
N.
sagar,
Roglai saga, Haroi akhul, W. Cham k?r, PR Cham hak?r, k?r, Wr.
Cham hagar, hagar; gar; MK: PNB *hag?r, Bahnar (AC) hagar; car, PSB (Efi?
[The vowel also suggests a MK ori?
mov) *sagghar, PKatuic *sagir; *sagka:r.
gin]
x*sadap 'old (things)', Rade had?p, Jarai (Lee) had?p ??, N. Roglai
The restriction of the distribution to Highlands Chamic suggests that,
this may be a post-PC borrowing.
[The
despite the regular correspondences,
medial consonant also suggests aMK origin]
sada?.
x*salog 'forever; eternally', Rade hl?g lar, Jarai (Lee) hlog, PR Cham
Wr.
Cham klaug. [The vowel suggests aMK origin]
klog -i,
x*sapal
'arm',
Acehnese
sapai,
Rade
p?l
'forearm',
Jarai
(PL)
hapal,
Jarai (Lee) hapal, Chru spal 'forearm', N. Roglai sapan, W. Cham pha pal 'fore?
arm muscle',
PR Cham hap?l, Wr. Cham hapal; MK: PKatuic (DT) *qapaal
'shoulder'.
II: The Chamic
Appendix
331
Lexicon
x*sapuat 'to harvest', Rade puot -v?, Jarai (PL) pu??, Jarai (Lee)
Chru
hapu??,
spua?, N. Roglai sapu??; sapua?. Restricted to Highlands Chamic.
x*(si)?joh 'drip; a drop', Jarai (PL) cf. t??joh, Jarai (Lee) ?joh; cf. ta?joh
'to leak', Chru sa?ioh, N. Roglai si?joh, Tsat ?iu55,W. Cham ta?joh, PR Cham
ta?j?h, Wr. Cham ta/?auh; MK: PNB *katoh 'drip; drop'. [The initial and the
vowel also suggest aMK origin]
x*siya:m 'good; nice; pretty', Rade siam 'beautiful', Jarai (PL) hi?m,
Jarai (Lee) hiam, Chru sia:m, N. Roglai sia:p, Haroi seam, W. Cham seam, PR
Cham thyam, Wr. Cham siani; MK: PKatuic *[l/lh][e?e]:m, *la-[l/lh][e/e]:m.
x*soh 'only; empty; free, leisure', Acehnese
soh, Jarai (PL) sah (Hd.),
Jarai (Lee) soh 'empty', N. Roglai (Lee) soh;W. Cham soh, PR Cham than, Wr.
Cham thauh; MK: PNB *sasoh, PKatuic *[s/c]anhah 'empty'.
[The vowel also
suggests aMK
origin]
x*sua 'pull out; seize', cf. xsuac 'pull out', Rade (Tharpe) sua, Jarai
sua
'seize', Haroi soa 'pull'. Note the distribution is restricted to High?
(Lee)
lands Chamic plus Haroi.
x*sua 'skin, dead', Jarai (Lee) sua, N. Roglai
sua, PR Cham
(Lee) sua.
[The vowel
x*suay 'fish trap', Rade sue 'long fishtrap, N. Roglai

suai. Restricted
x*sula 'leaf, Rade hia, Jarai (PL) hia, Jarai (Lee) hia, Chru sala, N.
hia:?
-h-f, Haroi hala, W. Cham hia, PR Cham hala, Wr. Cham hal?; MK:
Roglai
PNB *hla, PSB (Efimov) *_lha:, PWB (Thomas) *hlaa, PKatuic *[p/b]alhah,
*halhah.
x*suay 'slow; long time', Rade sui, Jarai (Lee) sui, Chru suai, N. Roglai
suai, Haroi sui, W. Cham suai, PR Cham (Lee) soy; sroy -m. [The vowel sug?
gests aMK origin]
x*sam
Chru masam
'to
wrap',
som
Acehnese
'cover'; MK: PMnong
'hide,
*klam.
put
?,
away'
Jarai
(Lee)
som
-1,
ori?
gin]
x*sag ? 'with; and', Rade mab?t h?g, Jarai (PL) h?g, Chru sag; sa- 'neg?
ative
particle',
[The vowel
W.
Cham
suggests aMK
h?g
-v
'with;
and',
PR
Cham
th?g
-v, Wr.
Cham
saug.
origin]
x*sana 'crossbow', Rade hana, Jarai (PL) hna (Pk), Chru sana, W.
Cham hanig -f tapog, PR Cham hani, Wr. Cham hani; MK: PMnong *sana, PSB
(Efimov) *sana: 'bow', PKatuic *sanha:.
332
II: The Chamic
Appendix
Lexicon
x*tadrua? 'lid', Rade kadru??, Chru tadrua?, PSB (Efimov) *khr?:p 'to
cover; a lid'. Restricted to Highlands Chamic.
[The vowel also suggests this is
of MK origin]
x*tagu:? 'get up; to stand up', Rade kag?? pit, Jarai (PL) dag?? -i, Jarai
(Lee) tagu?, Chru tagu:?, N. Roglai tagu:k -f, Haroi cakhu?, W. Cham ta ko?, PR
Cham tako?, Wr. Cham tagok. [The vowel length suggests aMK origin]
takua, Rade kakue, Jarai (PL) takuai, Jarai
x*takuay 'neck', Acehnese
Chru
N.
takuai,
takuai,
(Lee)
Roglai takuai, Tsat kua:i33, Haroi cak?ai, W. Cham
PR
takuai. [The vowel suggests aMK origin]
Wr.
Cham
Cham
takuai,
takoy,
x*taliat f? 'flute (front flute)', Rade dig kli?? -v 'side flute', Jarai (PL)
Chru
talia?, N. Roglai talia?. [The vowel suggests aMK origin]
talia?,
x*taluc
'last,
cf.
lastborn',
(Lee) talui?, N. Roglai talui?, Haroi

[The final suggests aMK origin]
kluc
Rade
x*luc,
cal?i?, PR Cham
Jarai
sibling',
'youngest
taluc.
tal?y?, Wr. Cham
x*taluc plus reflex of *apui 'fire' is 'firebrand', Rade (Tharpe) kluic,

Jarai (Lee) talui?, N. Roglai (Lee) talui?. Limited to Highlands Chamic.
[The
a
final also suggests MK origin]
x*tameh
tameh, Rade kameh, Jarai (PL) tameh,

'pillar; post', Acehnese
Jarai (Lee) tameh. [The vowel suggests aMK origin]
x*tano
N.
Roglai
son',
Wr.
'male',
Rade
tano
-n, Haroi
cano;
Cham
tanauw
'male;
ka no,
Jarai
(PL)
ta?o
dial.,
W.
son';
MK:
ta no,
Cham
Jarai
PR
tano,
*cano
PNB
(Lee)
Cham
'male;
ta no,
ta no, Chru
tano
'male;
husband'.
[The

x*tarapay 'rabbit', Rade pai, Jarai (PL) pai, Jarai (Lee) pai, Chru
N.
Roglai tarapai, Haroi kapai -i,W. Cham tapai, PR Cham tapay, Wr.
tarpa:i,
Cham tapay, Malay
tapai Treng.; MK: PMnong
*tarpay, PSB (Efimov)
*tsrapa:y.
x*tatuan
aMK
Chru
'wobbly',
tat?an,
W.
Cham
tat?an.
vowel
[The
suggests
origin]
'hang up', Rade (Tharpe) ka?ak, N. Roglai (Lee) ta?a?, PR

(Lee) ka???; MK: PKatuic (DT) *_bak 'hang up'. [The medial consonant
x*ta?ak
Cham
x*te?
'torn,
tion to Highlands
x*toh
worn',
Rade
te?, N.
Roglai
te?.
Note
the
restric?
Chamic.
'undress;
tl? -v, Chru
take
off,
Jarai
(Lee)
Roglai toh, Haroi t?h,W. Cham toh. Restricted

t?h
-1, Chru
to Highlands
toh
'dismantle',
Chamic
N.
plus Haroi.
333
Lexicon
II: The Chamic
Appendix
x*tram 'soak', Rade tram, Jarai (PL) tr?m (Pk et N), Jarai (Lee) tram,
Chru tram, N. Roglai trap, Haroi tr?m; dial, car?m, PR Cham tr?m, Wr. Cham
tram, PMP (*r)endem; MK: PNB *tr?m, PSB (Efimov) *tram, PKatuic *tarh[a/
a]m.
truag, Rade trag, Jarai (PL) trag, Jarai (Lee)

x*trog 'eggplant', Acehnese
trog, Chru trog, N. Roglai trok, Tsat (hu11) tsiog?33 -medial, Haroi trag,W. Cham
trag, PR Cham trag; cr?g, Wr. Cham traug; craug, Malay t?rung; MK: PNB
*haggig; *sakig ?.
*tr?g, PKatuic
x*truh 'arrive', Acehnese

troh, Rade truh, Jarai (PL) truh, Jarai (Lee)
PNB *tr?h.
;
truh, Chru truh, N. Roglai truh, Haroi troh 'escape' MK:
x*tu:? 'to receive', Rade tu? a 'accept; consent, Jarai (Lee) tu?, Chru
[du:?ma?], N. Roglai tu:?,W. Cham to?, PR Cham to?,Wr. Cham tok. [The vowel
length suggests aMK origin]
'snare', Rade kaho, N. Roglai tuho, Haroi cano. Note the
to Highlands Chamic, plus Haroi.
[The vowel also suggests aMK ori?
x*tuho
restriction
gin]
x*tuleh
cal?h, W. Cham
MK origin]
'untie', Jarai (PL) tal?h, Jarai (Lee) taleh, Chru taleh, Haroi
taleh, PR Cham tal?h, Wr. Cham talaih. [The vowel suggests a
x*tul?k 'disk shape; spherical shape', Jarai (Lee) tala?, N. Roglai (Lee)
tul?k -nf,W. Cham (Headley) /tal?k/, PR Cham (Lee) kal??.
[The vowel sug?
gests aMK origin]
x*tu?uac
cf.
'beak',
'lips',
Rade
ka?oc
-f, Jarai
(PL)
t??uk,
Jarai
(Lee)
ta?u?i?, Chru ta?uai?, N. Roglai tu?ue?, Haroi ca??ai?, W. Cham ca?uai? 'lips,
bill', PR Cham caboy?, Wr. Cham cabuac; MK: PKatuic *ca?bah, *tar?bah ?.
[The vowel and the final consonant also suggest aMK origin]
x*tal
(PL)
tal,
Jarai
'arrive;
(Lee)
Rade
tal ka
'sufficient',
Tsat
until',
tal
'to
the
tan33,
point
that,
Haroi
t?l
so much
-v
'come'
Jarai
that',
dir.
v., W.
Cham t?l, PR Cham t?l,Wr. Cham tal; MK: Bahnar (AC) t?l; toi, PSB (Efimov)
*tat 'arrive; reach'?.
x*wil 'round', Rade wll 'circle', Jarai (PL) wll, Jarai (Lee) wll, Chru wil
N.
'circle',
Roglai win, Haroi wll, W. Cham wll 'circle', PR Cham wll, Wr.
Cham wil; MK: PMnong *wll, PSB (Efimov) *wil.
x*wir 'turn around; dizzy; churning of rapids', Rade wir, Jarai (PL) wir,
Jarai (Lee) wir 'dizzy'; MK: PSB (Efimov) *wi:, *wa: 'dizziness'. Restricted to
Highlands Chamic.
334
II: The Chamic
Appendix
Lexicon
x*war 'forget', Acehnese

tuwA, Rade war, Jarai (PL) war, Jarai (Lee)
Chru
N.
war,
(rabit),
Roglai wabi?, Tsat van33, Haroi wol, W. Cham war, PR
Cham war, Wr. Cham war. [The vowel suggests aMK origin]
x*ya:g 'spirit; god', Jarai (PL) yag hrai, Rade yag, Jarai (Lee) yag, Chru
ya:g, N. Roglai ya:k, Haroi yiag, PR Cham yag, Wr. Cham yag, Malay yang;
MK: PNB *yag 'spirit', PMnong *yag, Bahnar (AC) i?g, PSB (Efimov) *ya:g,
PKatuic *?aje:g, *?aje:g.
x*yog 'to lift; take off, Rade y?g, Jarai (Lee) yog, N. Roglai yok 'carry
[The
by hand', Haroi y?g; yig ?. Restricted to Highlands Chamic plus Haroi.
x*yuam
Rade
'expensive',
yom 'price, cost'. Restricted

x*yun 'hammock',
yuom,
the Highlands
Chru
Tsat
yo:m,
Chamic
zuan33,
W.
Cham
[The vowel
andW. Cham.
Chru ayun, PR Cham ay?n; MK: Bahnar
(AC) ay?n.
x*yah 'PARTICLE', Jarai (Lee) yah 'why', N. Roglai yah, PR Cham

Wr.
Cham miyah; MK: PMnong *yah 'emphatic'.
miy?h,
a
MK
gests
origin]
x*yar 'lift, raise', Haroi yol, PR Cham yer, Wr. Cham yer.
1.3
PC words
of uncertain
[The vowel
(or other) origin
This
third list of forms that reconstruct to PC consists of words that, given the
current state of knowledge, seem to lack an etymology.
Some of these will inev?
itably turn out to be MK borrowings, but simply have not yet been identified as
such. For instance, many of the forms that are listed as restricted to Highlands
are likely to turn out to be borrowings, perhaps even post-PC Chamic

borrowings but ones that cannot yet be confidently labelled as such yet.
Chamic
Note that many of the forms occur both in Chamic and in Bahnaric, but
this by itself is not enough to establish that the form is ultimately MK as many
forms are found widespread
in Bahnaric languages.
demonstrably Austronesian
*?abaw
'large
ocean
Acehnese
'snail',
snail',
W.
Cham
ubo,
pau,
PR
Rade
Cham
abau,
apaw,
Chru
Wr.
N.
abau,
Cham
Roglai
abau
ab aw.
*?agam 'incest; desire, lust', Jarai (PL) ?g?m, Jarai (Lee) ?ag?m, Chru
agam, PR Cham ak?m, Wr. Cham agarri.
*?aka
see
*ka
*?ala 'below; beneath',

alia, W.
Cham
la tog
'under
the
Jarai (Lee) ?ala, Chru ala, N. Roglai

house',
PR
Cham
ala, Wr.
Cham
ala, Haroi
ala.
335
Lexicon
II: The Chamic
Appendix
*?alak 'yolk', Rade al?? 'eye' ?, Jarai (PL) ?l?k, Jarai (Lee) ?al??.
*?a?u? 'beads', Rade a??? 'necklace', Jarai (PL) a???, Jarai (Lee) ?a?u?
N.
-vl,
Roglai a??? 'seed bead necklace', Haroi a???, PR Cham ?u?,Wr. Cham
ftuk.
*?anug 'package', Rade an?g, Jarai (PL) an?g, Jarai (Lee) ?anug 'to
Haroi
an?g -v 'bundle', PR Cham an?g, Wr. Cham anug; MK: Bahnar
wrap',
(AC) anug.
gui, Chru agui, W. Cham gui 'wear' ?, PR
*?aguy 'to use', Acehnese
Cham aguy; guy, Wr. Cham agu?i; gy?i.
W.
*?apan 'hold; take', Jarai (PL) ?pan, Chru apan, N. Roglai apat,
Cham pan, PR Cham ap?n; pan, Wr. Cham apan.
*?ariag 'crab', Rade arieg, Jarai (PL) areg, Jarai (Lee) ?areg, Chru
N.
ra?iag,
Roglai ayak; ariak, Tsat liag?33 -i, Haroi areag, W. Cham riag, PR
Cham arieg; ryag, Wr. Cham arieg; riag.
*?asuk 'shavings', Jarai (Lee) ?as?k, Haroi asok. Note the restriction to
*?ay 'elder sibling', Rade ie 'older sister's husband, older brother's
wife', N. Roglai ai, Haroi ai 'brother-in-law
(of elder sibling); sister-in-law',
PR Cham ay, Wr. Cham ai.
*?ura? 'now', Rade ar??, Jarai (PL) r??, Jarai (Lee) ?ar??, Chru ara? ni,
PR Cham ur??; ar??, Wr. Cham urak.
*?urah 'bedbug', Rade areh -v, Jarai (PL) arah, Jarai (Lee) ?ar?h, Chru
ura? -f, Tsat zua55 (m),W. Cham rah, PR Cham ar?h, Wr. Cham
arah, N. Roglai
arah
*?usar
'soup
solids',
'flesh,
N.
Roglai
meat',
usa,
Acehnese
Haroi
as?l,
sia,
Jarai
PR Cham
(Lee)
?asar;
ath?r, Wr.
?as?r,
Cham
Chru
asa:r
asar.
*ba 'bring, take, carry', Acehnese ba, Rade ba, Jarai (Lee) ba, Chru ba,
N. Roglai ba, Haroi phia, PR Cham pa, Wr. Cham b?.
*ba 'to lead', Rade at?t ba, Jarai (Lee) ba, Chru ba in ba jalam 'ad?
vise', N. Roglai ba, W. Cham pa, PR Cham pa, Wr. Cham b?.
Chru pabug 'peak of roof,
*babug 'roof; ridge of (house, mountain)',
N. Roglai babuk 'tall center pole of house', Haroi paph?g 'roof, W. Cham
papug, PR Cham papug, Wr. Cham pabug, Malay bumbung; bubung, PMP
*bubug; MK: PKatuic (DT) *-phuung 'roof.
bantai, Jarai (PL) hanal ak??, Chru patal,
'pillow', Acehnese
Wr.
Cham patar -f,Malay bantai.
W. Cham pat?l, PR Cham pat?r -f,
*bap 'fill; full', Jarai (PL) bu??; b??, Jarai (Lee) b??, N. Roglai paba?,
Haroi phlau?, W. Cham pau?, PR Cham p??,Wr. Cham bak.
*bantal
*bapha
see
*mabha
336
Appendix
II: The Chamic
Lexicon
*bha:n 'sneeze', Jarai (PL) ph?n (Pk), Jarai (Lee) phan, Chru pham -n
pha:t, Haroi aphan, W. Cham pahan -v, PR Cham phan, Wr.
la:i?, N. Roglai
Cham bh?n.
*bhu 'dry' cf 'dry in the sun',
Rade
(Lee) bhu, Jarai (Lee) bhu, PR
Cham
(Lee) phu.
*bia? 'true, right; good', Chru bia?, Haroi phia?, W. Cham pea? 'indeed;
true; very', PR Cham py??, Wr. Cham biak, Malay baik 'good'.
*bijow 'shaman', Rade mj?u 'diviner', Chru pajau, N. Roglai bijau,
Haroi paslau, PR Cham pac?w, Wr. Cham pajuw; MK: PNB *paj?w 'sorcerer'.
'to butt', Rade manuh, Jarai (PL) b?banuh, Jarai (Lee) banuh,
N. Roglai bin?h, W. Cham panuh, PR Cham pinuh, Wr. Cham binuh.
*bit 'forget', Rade war bit, Jarai (PL) rablt, Jarai (Lee) rabit, N. Roglai
wabi? -1,Haroi phi?, PR Cham piw?l.
*binuh
*bitu? 'star', Rade mat??, Jarai (PL) p?t??, Jarai (Lee) patu?, Chru patu?
-f, N. Roglai pitu?, Haroi pato?, W. Cham pat??, PR Cham pit??; pat??; pat??,
Wr. Cham bituk; batuk.
*blah flat object', Jarai (Lee) blah, Chru blah elf., PR Cham plah, Wr.
Cham blah.
*blus -f 'to blow', Jarai (Lee) bluh -v, Chru blu:h, N. Roglai bluh, Haroi
pl?h, W. Cham pluh 'breathe, puff, PR Cham pluh, Wr. Cham bluh.
*boh
maw
'mushroom',
Rade
mamau,
Jarai
boh
(PL)
m?u,
Jarai
(Lee)
bam au, Chru bam au, N. Roglai bum au, Haroi pa miau, W. Cham poh m au, PR
Cham pimaw, Wr. Cham bimaw.
[The first element is the widespread An round
classifier
it
is
the
PC
second
element that has the unknown origin]
*boh;
object
*bru:g 'streaked; colorful, striped', Rade brug 'striped', Jarai (Lee)
brog, Haroi prug. Restricted to Highlands Chamic plus Haroi.
*buh 'wear ornaments',
Jarai (PL) buh, Jarai (Lee) buh, Chru buh, N.
to Highlands Chamic plus Haroi.
Haroi
Restricted
Roglai buh,
ph?n.
*bug 'large basket', Rade bug 'back basket', Jarai (PL) ??g -i, Jarai
(Lee) bug, Chru bug 'storage', Haroi ph?g, PR Cham pug, Wr. Cham bug, Malay
rombong.
'Cham', Rade cam, Jarai (PL) cam, Chru manih cam, N. Roglai
cap, Tsat tsam?42, Haroi cam 'Montagnards, highlanders', W. Cham
*cam
man?ih
cam,
PR
Cham
x*ca?a:g
cam, Wr.
'branch;
Cham
fork
cam.
of
tree',
Acehnese
cabuiag,
Rade
Jarai
ka?ag,
(PL) ta?ag i?, Chru ca?a:g, W. Cham ca?ag, PR Cham ca?ag, Wr. Cham ca?ag,
The
PMP *cabag 'bifurcation', Malay cabang; MK: PKatuic (DT) *-beeng.
glottalized ?- is quite unexpected, as is the medial Ibl inMalay, which probably
this forms seems to be attested
should be /w/. Although
PC it certainly patterns as a borrowing.
inAustronesian,
within
Appendix
II: The Chamic
337
Lexicon
*campa < *cam + pa 'Champa', Acehnese

Wr. Cham campa, Wr. Cham (AC) campa.
*cata
Restricted
'parrot',
to Highlands
N.
Roglai
'pierce, prick', Rade

Chamic.
*cut 'to dress; wear',

cu?, Wr.
Cham
Cham
cata,
W.
Cham
(Headley)
cata.
??t 'wear (ring)', Chru cu:?. Restricted
to
Rade cut, Jarai (Lee) cut,W. Cham cu?, PR Cham
cuk.
*dadit
Wr.
Cham
campa,
Chamic.
*cut
Highlands
cata, W.
juimpa, PR Cham
'a fan', N. Roglai
dadi:?, Haroi
cathi? 'to fan', PR Cham
tati?,
tadik.
*dahla?
'I (polite)', Jarai (PL) darn 'de sup?rieur ? inf?rieur', Chru

'
dalha?, W. Cham hl?n hl?? P, PR Cham tahla?, Wr. Cham dahlak.
*dap 'line up; straighten', Jarai (PL) d?p, Chru da? 'put in, place', N.
da?, PR Cham t??,Wr. Cham dak.
*dih 'that; there', Acehnese hideh, sideh 'there (far)', Rade adih, Jarai
(PL) ?dih, Jarai (Lee) dih, Chru dih 'there (far)', N. Roglai udih, Haroi thth, PR
Cham te h -v,Wr. Cham de h -v.
Roglai
*dra:g 'hornbill rhinoceros', Chru dra:g, N. Roglai dra:k, cf. Malay

?nggang. Restricted to Highlands Chamic.
*dra? -n 'hands on hips', Jarai (Lee) dr??, N. Roglai dr?? -n. Restricted
*duy 'to guide', Jarai (PL) dui, Jarai (Lee) dui, Chru dui, du:i, N. Roglai
dui, Haroi thui, PR Cham puy,Wr. Cham du?i.
*gah 'side, direction; bank', Jarai (PL) gah, Jarai (Lee) gah, Chru gah,
N. Roglai gah, Haroi khiah, W. Cham kah, PR Cham kah, Wr. Cham gah; MK:
Bahnar (AC) gah.
Jarai (Lee) ganam -f, Chru ga?ap -v?, N. Roglai
*gahnap 'wealth(y)',
gahn?p, PR Cham kan?p-mipa, Wr. Cham ganup.
*gaw 'rim', Rade gau 'back of knife', Jarai (Lee) gau 'helix', Chru gau,
PR Cham kaw 'dos'; hakaw, Wr. Cham gaw; hagaw.
*gig 'stove', Chru gi:g apui 'cooking fire', W. Cham kig 'cooking fire',
PR Cham kig 'cusine', Wr. Cham gig.
*glag 'look at; watch' Rade dl?g, Jarai (PL) lag -i; dl?g (Est), Haroi
tllag, PR Cham klag, Wr. Cham glag; MK: Bahnar (AC) lag.
*glay 'forest, jungle; wild, savage', Rade dlie lui, Jarai (PL) glai, dlai
et
(N
S), Jarai (Lee) glai, Chru glai 'jungle; forest; wild'; kaih, N. Roglai dlai;
kaih, Tsat khiaii?42, Haroi
Cham
tllai [cf. tlua], W. Cham klai, PR Cham r?m-klay, Wr.
rarri-glai.
*gulam 'carry on shoulder', Acehnese

gulam, Rade klam, Jarai (PL)
Jarai
Chru
N.
(Lee) gl?m,
gal?m,
galam,
Roglai gulap, Tsat khiag?42 -ft, Haroi
W.
PR
Cham
Cham
kaliam,
kil?m; kal?m, Wr. Cham gila ni; gala ni.
kl?m,
338
Lexicon
II: The Chamic
Appendix
*gunam 'cloud', Rade kanam, Jarai (PL) gan?m, Jarai (Lee) gan?m,
Chru ganam 'rain cloud', Haroi kanlam, PR Cham kan?m, Wr. Cham ganani.
*gut 'cave', Jarai (Lee) g??, Chru gu:? 'lair', N. Roglai gu:? 'cave', PR
Cham ko?,Wr. Cham gok.
*ha 'you; thou', Acehnese gata (sg./pl.) (neutral); ta-; -tui(h), Jarai (PL)
h? (N), Jarai (Lee) ha, Chru h?, N. Roglai h?, Tsat ha33, Haroi h?i -v,W. Cham
hi, PR Cham hi,Wr. Cham hi.
*hadum -1 'how much, many?', Rade d?m, Jarai (PL) had?m, Jarai
dorn -v, Chru hadu:m
(Lee) hadom;
Cham
hatom
-v; hat?m
-v, PR
hadu:p -1, Haroi
-vl, N. Roglai
hatom
Cham
-v;
torn
-v, Wr.
Cham
ath?m, W.
hadom.
*haley 'who; question word', Rade hlei, Chru halai, N. Roglai

Tsat ?a33 za:g33 -i v?, Haroi halai -v (grammatical particle), W. Cham
'whatever', PR Cham hal?y; ley, Wr. Cham hal?i.
alai,
hlay
*halim 'rainy season', Rade hllm 'rain for two or three days'; lip
Jarai
'flood',
(Lee) hlim, Chru halim 'wet weather', N. Roglai halip 'continual
halim 'flood', PR Cham halim 'rainy season', Wr. Cham
Haroi
rain; flood',
halim.
'perforated; pierce', Rade hluh, Jarai (Lee) hluh, Chru haluh, N.

haluh 'to have a hole', PR Cham hal?h, Wr. Cham haluh.
Haroi
haluh,
*hayow 'like (prep.)', Chru yau, W. Cham yau, PR Cham y?w, Wr.
*haluh
Roglai
Cham
yug
?.
*huna?
'asthma',
Chru
N.
nana?,
Roglai
hun??,
W.
Cham
-f, PR
hanic
Cham hani?, PR Cham hani?, Wr. Cham hanik.

*jah 'weed, clear brush', Rade jah, Jarai (PL) jah, Jarai (Lee) jah, Chru
jah 'chop small wood', N. Roglai jah, W. Cham ?ah 'cut with a knife', PR
Cham ?ah, Wr. Cham jah.
*jalu? 'bowl', Rade el??, Chru jalu?, N. Roglai jalu? (also a elf.), PR
Cham
(Lee) pagln-cal??; MK: Bahnar

*ka,
*?aka
'not
yet',
Rade
(AC) jalu.
ka...oh;
Jarai
(Lee)
ka;
?aka;
Chru
ka...au;
ka; ka Buh; Haroi ka?; W. Cham ka...o; PR Cham ka...o 'pas encore'.
*ka:l 'to lock; bolt', Rade kal, Jarai (Lee) kal, Chru kal -1; k-an-al 'a
lock', PR Cham (Lee) kian -vf.
*ka?iag 'loins; waist', Rade kaieg 'rib (back)', Jarai (Lee) ka?iag, Chru
N. Roglai
ka?iag 'waist', N. Roglai ka?iak 'lower back', Haroi ka?eag, W. Cham ka??g, PR
Cham ka?g,Wr. Cham kaig.
*kacaw
'scratch,
scrape',
Rade
kacau,
Chru
kacau,
N.
*kadow 'jump', Rade kad?u, Jarai (Lee) kadau, N. Roglai
-d, PR Cham kat?w, Wr. Cham kaduw.
Roglai
kacau.
Restricted
kad?u
kadau, Haroi
Appendix
II: The Chamic
339
Lexicon
*kaka:s 'fish scales', Rade kaih, Jarai (PL) rak?h -i, Chru kark?:h -n;
N.
karka:h,
Roglai kaka, Tsat ka33, Haroi kak?h 'fin of fish, shell of anteater',
W. Cham kakah, PR Cham kak?h, Wr. Cham kakah, Wr. Cham (AC) kakah.
*kala 'bald; bare', Rade k?? kla, Jarai (PL) kla, Chru kala ak?, N.
Tsat kiu33 -f?, Haroi kala; kalo, W. Cham kla, PR Cham kala, Wr.
kahlo,
Roglai
Cham kal?.
*kalih 'miserly', Jarai (PL) kaTih, Jarai (Lee) klih, Chru karlih, Haroi
kaleh, PR Cham kallh, Wr. Cham kalih.
*kapit 'to close', Rade (Lee) kapl?, Jarai (Lee) kapit, PR Cham (Lee)
kapi?. Restricted to Highlands Chamic.
*karam
'sink; sunk', Jarai (PL) kr?m, Jarai (Lee) kr?m, Chru kram, N.
Haroi
kr?m, PR Cham kar?m, Wr. Cham kararri, Malay karam;
Roglai karap,
MK: PNB *kr?m.
*kata:l 'thunder (-bolt); lightning', Jarai (PL) katal, Chru kata:l, N.
katan, PR Cham katal, Wr. Cham katal.
*kata:g 'strong; well', Rade katag, Jarai (Lee) katag, Chru kadag -1,
Haroi katag, PR Cham (Lee, Blood) kat?g -1.
Roglai
*katit 'to crush', Rade ka tit -1 'wring, twist', Jarai (Lee) katit, N. Roglai
kati:?, Haroi kat?t -f 'fall on something', W. Cham kat?? 'pinch'. Restricted to
Highlands Chamic, Haroi, andW. Cham.
*katrow
'pigeon', Rade katr?u, Jarai (Lee) katrau, Chru katrau, N.

Roglai katrau, Haroi katrau, W. Cham katrau play, PR Cham katr?w-catoy;
MK: PNB *catr?w, PMnong *katap.
*katug
Tpuir,
Rade
kat?g,
Jarai
(Lee)
katug,
Chru
katug,
N.
Roglai
katuk, Haroi katog 'tear (thread)', PR Cham (Lee) katug.

*kayua 'because', Rade kayua dah, Jarai (PL) yua (ka), Jarai (Lee) yua
ka, Chru kayua, N. Roglai kayua, Haroi kayua ka-, W. Cham kayoa, PR Cham
kaywa,
Wr.
Cham
kayu?.
*khag 'hard; stiff; strong', Acehnese

kAg 'strong' -v, Rade kh?g, Jarai
Jarai
(PL) kh?g,
(Lee) kh?g, Chru khag, N. Roglai khak 'solid, dense', Tsat
khaig?42, Haroi kh?g, W. Cham kh?g, PR Cham kh?g, Wr. Cham khag.
*klam 'afternoon; night', Rade tlam, Jarai (PL) kl?m (Hd.), Jarai (Lee)
Chru
kl?m,
klam, N. Roglai tlap, Tsat kian?33, Haroi tl?m 'night', W. Cham
makl?m, PR Cham kl?m; makl?m, Wr. Cham klam; maklani; MK: Bahnar (AC)
klam.
*klap 'old', Chru kla?, W. Cham klau?, PR Cham kl??, Wr. Cham klak;
MK: PNB *kr?q.
*klaw 'laugh', Rade tlau, Jarai (PL) kl?u, Jarai (Lee) klau, Chru klau,
N. Roglai tlau, Tsat kiau33, Haroi tlau, W. Cham klau, PR Cham klaw, Wr. Cham
340
Appendix
II: The Chamic
Lexicon
tawa, PMalayic
*tawa?, and PMP *tawa do not
[The forms in Malay
to
related
the
PC]
appear
*klum 'to cover', Jarai (Lee) kl?m, Chru klum 'cover up', N. Roglai
tlup, Haroi tl?m; kalom.
klaw.
*krah 'middle; half, Rade ti krah, Jarai (Lee) kr?h, Chru kr?h, N.
khr?h -i, Tsat kia55, Haroi kr?h, W. Cham krih; kih, PR Cham krih, Wr.
Roglai
Cham krih
*kra:g 'clam;
kra:k
'clam',
Acehnese
shellfish',
knuag, Chru kara:g -v, N. Roglai
k?rang.
Malay
*krih 'whittle', Rade kri? -f 'whittle, sharpen', Jarai (PL) krih, Jarai
(Lee) krih, Chru kri:h -1, N. Roglai krih, Haroi kreh 'sharpen'; car?h -v, PR
Cham krih, Wr. Cham krih.
Jarai (PL) kamai, Jarai (Lee) kamai, Chru
*kumey 'female, woman',
kamai, N. Roglai kumai -n, Tsat mai33, Haroi kam?i, W. Cham kamay, PR Cham
kam?y, Wr. Cham kam?i.
Cham
*la 'spleen; pancreas',

Wr.
Cham l?.
la,
*labua
PR Cham
Tsat
khe55
taro',
plant;
(Lee) kap?a
*lagah
lagan,
'a
Jarai (PL) la, Jarai (Lee) la, N. Roglai

labua
Chru
'spinach',
Haroi
laphua
di:?la, PR
'taro
root',
-i.
'tired', Rade
-v, Haroi
egah, Jarai (Lee) ragah, Chru

W.
lakhlah,
Cham
lakah
'ache',
lagah, N. Roglai
PR
Cham
likah;
likoy; likah, Wr. Cham ligah; liguai; lagah; MK: Bahnar (AC) ragah.
Jarai (PL) rajau, Jarai (Lee)
*lajaw or *rajaw 'hammock',
Restricted to Highlands Chamic.
*lakow
Roglai
rajau.
'ask for', Rade ak?u, Jarai (PL) rakau, Jarai (Lee) rakau, N.
lakaw, PR Cham lik?w; lak?w, Wr. Cham lakaw.
lakau, W. Cham
*lama:n
'elephant',
Chru lama:n, N. Roglai

Wr. Cham limin.
Rade
eman,
*lanag 'earthworm',
Chru la nag, N. Roglai lanak
Wr. Cham lanig.
Rade
Jarai
(PL)
raman,
lamian, W. Cham
lum?n, Haroi
Jarai
(Lee)
raman,
lam?n, PR Cham
limin,
en?g, Jarai (PL) r?n?g, Jarai (Lee) ran?g,

lan?g, W. Cham lanag, PR Cham lanlg,
-f, Haroi
*lasun
'onion', Rade es?n, Jarai (PL) ras?n, Jarai (Lee) ras?n, W. Cham
las?n, PR Cham lith?n; lath?n, Wr. Cham lisun; lith?n; lathun.
Rade ewag, Jarai (PL) rSwag, Jarai
*lawa:g 'thin; lean; emaciated',
(Lee) rawag, Chru lawa:g, N. Roglai luwa:k 'lean; thin', Tsat va:n33 -f, Haroi
lawag, W. Cham lawag, PR Cham liwag; lawag, Wr. Cham liwag; lawag, PMP
*niwag.
x*lamo 'cow; ox; cattle', Acehnese humo, Rade emo, Jarai (PL) ramo,
Jarai (Lee) ramo, Chru lamo, N. Roglai lamo -n, Tsat mo33, Haroi lamo, W.
II: The Chamic
Appendix
Cham
MK:
341
Lexicon
lamo, PR Cham limo; lamo, Wr. Cham limauw; lamauw, Malay l?mbu;
Bahnar (AC) ramo; lamo. [This word has a limited distribution inAN.]
liah, Jarai (PL) liah, Jarai (Lee) li?h-i,
*liyah 'lick; taste', Acehnese
Chru layah 'taste', N. Roglai liah, Tsat lia55, Haroi leah, PR Cham ly?h, Wr.
Cham liah; MK: cf. PEastern MK *liat 'lick'.
*lu 'much, many', Acehnese
la, Rade lu, Jarai (PL) lu, Jarai (Lee) lu
'very', Chru lu la; lu bi?, Tsat lu33,Haroi lu; lo -v,W. Cham lo pay, PR Cham lo
-v,Wr. Cham lo;MK: Bahnar (AC) la.
*lukut 'absent', Rade ek?t, Jarai (PL) r?kut, Chru laku:?, N. Roglai
liku:? -v, Haroi lakou? 'avoid; escape'?, PR Cham (Lee) liku?.
*ma?i?k 'urinate', Acehnese ?ia?, Rade ma?iek, Jarai (PL) ma?a?
?a?,
Chru maTia?, N. Roglai ma?i??, Haroi ma????, W. Cham ma?ii? (n), PR Cham
mii? -v (n),Wr. Cham miik; MK: Bahnar (AC) ik.
'naked', Acehnese
Ion, Rade ml?n, Jarai (PL) hlun,
*(ma)(sa)lun
mahlun, Jarai (Lee) mahl?n, Chru sarlun, N. Roglai salut, Haroi mahal?n, PR
Cham mil?n, Wr. Cham milun.
*ma?ih 'soured', Jarai (Lee) ma?ih, W. Cham ma?ih
mimih, Wr. Cham mimih.
'spoiled', PR Cham
*mabha, *bapha 'divide; share', Rade mabha, Jarai (Lee) papha < *p-,
Chru parpha, N. Roglai mupha, Tsat pha33, W. Cham papha, PR Cham pha;
parapha; rapha, Wr. Cham bh?; p?rabh?; rabh?; MK: PMnong *pa?.
'move,
*magey
agitate;
wobbly,
Rade
loose',
shaky,
Chru
magei,
magai, N. Roglai magai, W. Cham makay.

Rade
'fox',
*maja
Jarai
'weasel',
mja
(PL)
m?ja,
Chru
maja
'weasel',
N. Roglai maja, PR Cham mica 'civette', Wr. Cham mij?.

*makrah 'middle; half, cf. *khrah, Rade ti krah; m a kr?h, Jarai (PL)
makr?h, Jarai (Lee) makrah, Haroi makr?h, W. Cham krih.
*mal 'beam',
Cham
mal,
Wr.
Cham
-v 'suck; suckle',
*mam
m?m
-v, PR
-v, Chru
Cham
Jarai (Lee) mal, Chru mal, N. Roglai man, Haroi mal, PR

mal.
mem
m?m
-v; mum
-v; m?m,
-v, N.
Wr.
Rade mam,
Roglai
Cham
mam,
muni;
Haroi
mem
-v, W.
Jarai (Lee)
m?m
Cham
mani.
'itch', Rade ruih lak, N. Roglai maruh

Chamic.
*marus
to Highlands
Jarai (PL) pam?m,
'itchy, sores'. Restricted
*mat 'take; fetch, get', Acehnese mat, Rade m??, Jarai (PL) m??, Jarai
(Lee) m??, Chru ma?, N. Roglai ma?, Tsat ma?24, Haroi m?? 'to take, get, catch,
seize', W. Cham ma?, PR Cham ml?, Wr. Cham mik.
*mit 'always', Chru mit, W. Cham
Highlands Chamic language plus W. Cham.
/mit/ (Headley). Restricted
to one
342
II: The Chamic
Appendix
Lexicon
*ga? 'make, do', Rade g??, Jarai (PL) g??, Jarai (Lee) g??, Chru ga?, N.
g??, Tsat gau?24, Haroi g??, W. Cham g??, PR Cham ag?? cf. g??, Wr.
Roglai
Cham agap; gap.
*pa?
-i 'to,
distance)',
Chru
towards',
PR Cham
pa?
'at',
(Lee) pa?, Wr. Cham
N.
pa?, Haroi
Roglai
pa-
(AC) phak -i;MK: Bahnar
'at,
in (far
(AC) ph?
*pa?a:k 'armpit', Rade p?l-ak, Jarai (PL) pa???, Jarai (Lee) pa???, Chru
N.
Roglai ala pa?a:?, Haroi pa?a?, W. Cham pa?a?, PR Cham paa?, Wr.
pa?a:?,
Cham
paak.
*pabah 'spittle, slaver, drool', Rade

bah, Chru ia pabah, Tsat pha55.
*padar 'spin; turn', Jarai (PL) padar; padar (Ouest), Jarai (Lee) pad?r;
pa dar, Chru padar -1, N. Roglai pada 'turn upside down', W. Cham t?r, Wr.
Cham (AC) padar.
*padey 'rest', Rade madei, Jarai (Lee) pedai, Chru padai, N. Roglai
Haroi
pathli, PR Cham pat?y, Wr. Cham pad?y; MK: PNB *bad?y.
padai,
*paga:g
'protect',
Rade
magag
used
in
phrase
meaning
'to
get
shot', Jarai (PL) pagag, p?gig -v, Jarai (Lee) pagag, Chru pagan -fl 'to barri?
cade'. Restricted to Highlands Chamic.
*padi? 'pain; ache', Jarai (Lee) pa?ia? -m, Chru padi:?, N. Roglai
padi:?, Haroi padi?, W. Cham padi?, PR Cham padi?, Wr. Cham padik. Note:
the Acehnese may,
Acehnese pui de h, Malay p?dih represent a distinct etymon;
in fact, be borrowed from Malay.
*pajum 'meet together', cf *jum, Jarai (Lee) bajan?m, N. Roglai pajup,
Haroi pas?m, PR Cham pac?m, Wr. Cham pajum.
*pak-ke 'lizard; gecko', Acehnese pa?e, Rade p?k ke, Jarai (PL) p?k
ke, Jarai (Lee) pakake, Chru pak ke, N. Roglai pake, PR Cham pake, Wr. Cham
pakaiy.
*pag 'to make a wall', Jarai (Lee) p?g cf. kh?n p?g 'curtain', W. Cham
(AC) pag; MK: Bahnar (AC) p?g; pag. Restricted to Highlands Chamic andW.
Cham.
*pataw 'master; lord', Rade matau, Jarai (PL) patau, Jarai (Lee) patau,
Chru patau, N. Roglai pitau, Haroi patau, W. Cham patau, PR Cham pataw, Wr.
Cham patau; MK: PNB *pat?w, PSB (Efimov) *'pataw 'king, state'. Marrison
(1975:53) follows Aymonier and Cabaton (1906) in suggesting that this word
might
be
composed
of
*po
plus
tau
'person'
as
in Tagalog
tao
'man'.
Rade magan 'plate'; big 'bowl or

x*pigan 'bowl; dish', Acehnesepigan,
or
gourd', W. Cham paf?in , PR Cham pag?n-?alu?,
plate made from squash
Malay pinggan, PMP *piggan; MK: PMnong *tiggan 'bowl'; *bag 'bowl', PSB
(Efimov) *[ba]ggan, PKatuic (DT) *p/t-in9an> PKatuic *pagha:n 'bowl'. This
Appendix
II: The Chamic
343
Lexicon
word, according to Coope, is a Persian borrowing by way of Hindi; certainly, the

vowel length of all the Chamic forms suggests a short vowel, rather than the long
vowel that would be expected from the proposed PMP reconstruction. Note not
only that the form also occurs in PMnong, PSB, and PKatuic but that it shows
considerable variation in these MK languages, strongly suggesting it was bor?
This form is a borrowing into all the languages of the area.
*pioh 'keep, store; conserve', Rade pion 'to put, to place', Jarai (PL)
pioh wai, Jarai (Lee) pioh, PR Cham pyah, Wr. Cham piah
*plag 'citronella grass', Rade plag, Chru pla:g -1,N. Roglai plak, Haroi
rowed intoMK.
apl?g, W. Cham plag, Wr. Cham (AC) plag.

*pras 'scratch (of chicken)', Rade (Tharpe) praih, Chru pra:h, N. Roglai
Restricted
pipra.
'to plant', Acehnese pula, Rade pla, Jarai (PL) pl?, Jarai (Lee)
p?a, Chru pala, N. Roglai pila, Tsat pia33, Haroi pala, W. Cham pla, PR Cham
*pula
Wr.
pala,
Cham
pal?.
*pusa:g < (? *po + *sa:g 'master' + 'house' (AC)) 'husband', Chru

N.
pasa:g,
Roglai pisa:k, W. Cham pasag, PR Cham pathag, Wr. Cham pathag.
[The word for 'house' looks borrowed but the word *po 'master' may be inher?
ited]
(causative
prefix
*pa-gha:g
*pa-) 'dry over fire', Rade bhag -i
to
have coalesced),
Jarai (PL) kh?g, Jarai (Lee) pa khag, N. Roglai
(appears
Haroi
'roast,
broil',
pakha:k
pakhag, PR Cham pakhag. The Malay form pang
a tempting comparison,
gang 'to roast' makes

would
require
would
be
but the *g- to > gh- change that it
unique.
'forbid', Jarai (Lee) pakh??, Chru kha?,

*pa-gha? (*pa- 'causative')
Haroi kh??, PR Cham kh?? trlh,Wr. Cham ghak drih.
*pa-klah 'divide; separate', cf. *klah 'to lose', Rade kah, Jarai (PL)
p?klah; pacalah (Pk), Jarai (Lee) peklah, Chru klah, N. Roglai patla; pak, PR
Cham kl?h-nlh, Wr. Cham klah-nih.
*pa-pah
papah
'causative')
(*pa-
'clap hands', W. Cham pah.
'clap,
rub',
slap;
Restricted
Rade
pah,
to Highlands
N.
Chamic
Roglai
plus W.
Cham.
*rah/s
Cham
ran, Wr.
-f 'separate
Cham
a fire', Rade
raih -f pui, PR
raih pui, N. Roglai
rah.
*ram 'dead fire', Rade pui ram, W. Cham pui r?m.

*ranam
Wr.
Cham
rana
*rawag
Cham
'love',
Chru
ranam,
'visit',
Chru
rawag,
Haroi
W.
Cham
ran?m,
PR
Cham
ran?m,
ni.
N.
Roglai
lawak
-i, PR
Cham
raw?g,
rawag.
Wr.
344
Appendix
'land, interior; earth', N. Roglai
*riya
II: The Chamic
riya, PR Cham
Lexicon
(Lee) riya, Wr.
Cham
(AC) riy?.
*rata:k 'bean; pea', Acehnese nutuia?, Rade etak, Jarai (PL) rata?, Jarai
(Lee) rata?, Chru rata:?, N. Roglai rata:?, Haroi lata? -1,W. Cham rata?, PR Cham
rita?;
rata?, Wr.
Cham
rit?k;
sui a g 'hut,
Acehnese
'house',
*sa:g
rat?k.
tent',
Rade
Jarai
sag,
(PL)
sag,
Jarai
(Lee) sag also 'family', Chru sa:g, N. Roglai sa:k, Tsat sa:g33, Haroi sag, W.
Cham sag, PR Cham thag, Wr. Cham sag.
*sa?ay 'elder sibling', Chru sa?ai, N. Roglai sa?ai, Haroi ca??i -i (Note:
looks like a compound, not a disyllabic root), PR Cham ay, Wr. Cham ai; MK:
PKatuic *sa:j; *sa-sa:j 'older sibling'. The existence of this root in both PC and
PKatuic
some
needs
*salih
Restricted
hamiau.
explanation.
m alih
<
*pa-, Chru salih, W. Cham halih.

W.
and
Cham.
(Chru)
*samaw 'prompt; on time', Jarai (Lee) hamau, Chru sam? -f, Haroi
Restricted to Highlands Chamic plus Haroi.
Rade
'trade',
to Highlands
Chamic
*samu
Rade (Tharpe) hamo, Rade (Lee)
'compare;
comparable',
N.
PR
Cham
hamo,
(Lee) hmu.
Roglai (Lee) sam?,
*sana 'roast; parch', Jarai (PL) han?, Jarai (Lee) ha na, Chru sana 'fry',
sana -n, Haroi nana, W. Cham hana (first syllable
N. Roglai
atypically
PR
retained),
Cham
Wr.
hana,
Cham
han?.
anal k?? -i, Jarai (PL) ha nal ak??, Jarai (Lee)

'pillow',
sanan.
N.
Restricted
hanal,
Roglai
*sanig 'think', Rade han?g; m?n, Chru sa na g, N. Roglai sin?g, Haroi
*sanal
hanlg,
W.
Cham
*saraw
Rade
sana
'sting,
to Highlands
Restricted
*sarum
g -i, PR
Cham
hurt',
Chamic
'sheath-like',
Chru
sahig
sr?u,
-i, Wr.
N.
Cham
sanig.
Roglai
sarau,
W.
hrau.
Cham
andW. Cham.
Jarai
(Lee)
hrum
'scabbard',
Chru
srum
?, Haroi
har?m, PR Cham har?m, Wr. Cham harum; MK: Bahnar (AC) co rum. Although
the final nasal is a complete mismatch, Malay sarung comes tomind.
*satuk 'boil; cook', Rade t?k, Jarai (PL) hatuk, Jarai (Lee) hat?k, Chru
stu? 'boil', N. Roglai satu? 'boil', Tsat tu?24,Haroi ato?, W. Cham t??, PR Cham
tuk; hatuk; MK: Bahnar (AC) hat?k.
*sijaw 'hammock', N. Roglai sijau, Haroi asiau.
lands Chamic plus Haroi.
hat??; t??,Wr. Cham
Restricted
to High?
*sukat 'cork; stopper', Rade k?t, Jarai (PL) hak?t, Jarai (Lee) hak?t -v,
Chru ska?, N. Roglai suka? -v 'poke in; insert'. Restricted toHighlands Chamic.
*tagak -f 'cleaver', Rade kag?? 'small bush knife', Jarai (Lee) tag??,
Chru tagak -f 'bushknife (long curved handle)', N. Roglai tagak -f 'bush knife',
Wr. Cham
(AC) taga?; MK: Bahnar
(AC) tagak.
*tamaha
345
Lexicon
II: The Chamic
Appendix
tuwan
Acehnese
in-law',
'parent
kamha,
Jarai
'also an address
term',
-mf,
tarn aha, Haroi cahma
(Lee) tahma; tahmua, N. Roglai

W. Cham tarn aha, PR Cham (Lee) hma.
Rade
*tamut 'hammer', Rade kam?t, Jarai (PL) mut, t?mut, Jarai (Lee)
tam?t, Chru m?:?, N. Roglai m?:?, W. Cham mu?. Restricted
to Highlands
Chamic plus W. Cham.
*tana:? 'fagot; bamboo
strip', Jarai (PL) cana?, Jarai (Lee) tana?
'fagot',
PR
Cham
can?
tan?;
?, Wr.
tanak.
Cham
'deaf, Rade kag?l, Jarai (PL) t?gll; t?gal (N), N. Roglai (dan),
Tsat (ga33) gin33 v?, W. Cham tagoh -f, PR Cham tag?h -fv,Wr. Cham tagaurj.
*tapa? 'lie full length', Rade kap??, Chru tapa? 'standing upright', W.
*tagal
Cham tap?? also

*tarah
to Highlands
'sit up'. Restricted

cut
'trim;
up;
to
Jarai
sculpture',
'to
trah
(Lee)
N.
sculpture',
Roglai tarah, PR Cham tar?h, Wr. Cham tarah.

*tasa? 'ripe; cooked', Rade kas??, Jarai (PL) tasa?, Jarai (Lee) tasa?,
Chru tasa?, N. Roglai tisa?, Tsat sa?24, Haroi casa?, W. Cham tas??, PR Cham
tath??,
Wr.
tathak.
Cham
*thu 'dry', Acehnese

tho, Rade thu, Jarai (PL) thu, Jarai (Lee) thu, Chru
'to
thu, N. Roglai thu; pathu
dry', Tsat thu33,Haroi thou,W. Cham thu, PR Cham
thu,Wr. Cham th?.
Chru
'dance',
*timiya
N.
tamia,
Roglai
-v, W.
timiya
Cham
tamanea,
PR Cham tamya; mya, Wr. Cham tami?; mi?.

*tiya:p
N. Roglai
run
'chase,
after',
Acehnese
tia:?, Haroi teau?, PR Cham

'full,
*trey
satiated',
Rade
ti??
Rade
tiyuiap,
-fv, Chru
tia:u?,
tyaw?, Wr. Cham tiap.
trei,
Chru
trai, N.
Roglai
trai, Tsat
tsia33
v?,
troi -1,W. Cham tray, PR Cham tr?y,Wr. Cham tr?y.

*tuh 'change; metamorphose',
Jarai (PL) tuh, Jarai (Lee) tuh, N. Roglai
tuh, Haroi toh 'give birth (animals)'. Restricted to Highlands Chamic and, pos?
Haroi
sibly, Haroi.
*tuh
Roglai
'pour', Rade tuh, Jarai (PL) tuh, Jarai (Lee) tuh, Chru tuh, N.
Haroi
tuh,
catoh; toh, W. Cham tuh, PR Cham t?h,Wr. Cham tuh.
>
'follow'
*tuy
to',
'according
Rade
tui
hlue
hlue
'according to', Chru tui 'according to', N. Roglai

namese, W. Cham tui 'according to, with', PR Cham
'imitate,
copy,
follow',
theu -v 'follow'
< Viet?
tuy 'suivre', Wr. Cham
tu?i.
*tuy 'to follow',

dorn,
tuy,Wr.
Haroi
Cham
toi
dir.
v.
only
Rade tui hlue, Jarai (PL) tui, Chru tui, N. Roglai

occurs
as
an
auxiliary
verb,
W.
Cham
tui, PR
theu;
Cham
tu?i.
*tuy 'to mate',

Chru tui; tu:?,W. Cham
tui.
Restricted
to Highlands
346
Appendix
II: The Chamic
Lexicon
tuirho? -m, Rade kadl?k, Jarai (Lee)

*taglak 'choke; cough', Acehnese
tagl?k, Chru targla?, Haroi catlla?.
*wah 'to fish', Rade wah, Jarai (PL) wah, Jarai (Lee) wah, Chru wah
'fish with pole', N. Roglai wah, Tsat va55, Haroi w?? -f 'hang on a hook', W.
Cham wah, PR Cham w?h, Wr. Cham wah
*wah 'weave; twist', Rade
(Lee) wah.
(Lee) wah, Jarai (Lee) wah, N. Roglai
(Lee)
wah, PR Cham
*wag 'sickle', Rade w?g kuok 'hoe', Chru wag, N. Roglai wak 'blade
for weeding', Haroi w?g, W. Cham w?g, PR Cham w?g, Wr. Cham wag; MK:
Bahnar (AC) u?g.
*yah
Wr.
Cham
take apart', Haroi yiah, W. Cham yah, PR Cham y?h,
'destroy;
yah.
'yoke', Chru yau, W. Cham yau, PR Cham y?w, Wr. Cham yuw.
'to
to
use', Rade (Tharpe, Egerod) yua, Haroi yua. Restricted
*yua
Highlands Chamic.
*yow
2.
Words
Borrowed
After
the Breakup
of PC
contains lists of forms borrowed into Chamic languages after the

In addition to the continued borrowing of MK forms throughout
of
PC.
breakup
the history of the Chamic languages, there were also contributions of loanwords
This
from
2.1
section
two main
sources
of
cultural
influence:
India
and Arabia.
Of Indie origin
layer of Indie borrowings in insular Southeast Asia seems to date back

a little over two thousand years. As discussed earlier in this work, Indie loan?
The oldest
words are found in the earliest Indie inscriptions and six forms that reconstruct
the PC level have proposed Indie etymologies.
x*?asar
'seed',
Rade
as?r
-1, Jarai
(PL)
asar,
Tsat
sa?42
-f, Haroi
to
as?l
'grain', W. Cham
s?r, PR Cham ath?r, Wr. Cham asar; MK: PMnong *gg?r ?,

*ggar. [? < Indie, cf. Sanskrit saara. It is unclear if theMK forms
PSB (Efimov)
listed are related.]
x*?aseh 'horse', Rade aseh, Jarai (PL) aseh, Jarai (Lee) ?aseh, Chru
N.
aseh,
Roglai aseh, Tsat se33, Haroi as?h, W. Cham seh, PR Cham ath?h, Wr.
Cham asaih; MK: PNB *aseh; MK: PMnong *?aseh, PSB (Efimov) *?ase:h,
PKatuic *?aseh. [< ? Indie; Headley, #2.2; cf. Sanskrit ashva]
Appendix
II: The Chamic
347
Lexicon
x*bih 'poison; venom', Jarai (Lee) bih, Chru bih, Haroi phlh, PR Cham
Wr.
Cham
bih. [< Indie; Headley, #2.9; cf. PIE *visa]; MK: PKatuic (DT)
pih,
*piih.
(Durie) *bijeh, Acehnese bijeh, Rade mjeh,
x*bijeh 'seed', P-Acehnese
Jarai
Jarai (PL) pajeh; bajeh (S),
(Lee) paj?h -? -i, Chru pajeh, N. Roglai bijeh,
Tsat se55, Haroi pasih, W. Cham pa?eh 'seed for planting', PR Cham pa?eh, Wr.
Cham pajaih -i, Malay
form is from Sanskrit]
to Coope,
[? < Indie cf. Sanskrit; according
biji.
this
x*kapa:s 'cotton'. MK < Indie ?; Headley, #2.1; Hindi; Sanskrit kar

Note:
this as a
There is, so far, no PC internal grounds for designating
paasa;
borrowing.
Incidentally, Lee's apparent reconstruction of length in the onset syl?
lable is simply a mistake, more than likely simply a typing mistake. Much ink
has been spilled discussing
this form and its origins.
x*radeh 'vehicle', Rade edeh, Jarai (PL) radeh, Jarai (Lee) radeh, Chru
radeh, N. Roglai radeh, Tsat the11, Haroi lathih, W. Cham rateh, PR Cham rite h
-v; rateh -v, Wr. Cham rideh, radeh; MK: PMnong *randeh, PSB (Efimov)
[? < Indie; Headley,
*rande:h.
#2.8; Sanskrit
ratha]
For these six, it is likely that they were already in the Austronesian
developed into Chamic.
language that
The remaining Indie loans clearly postdate that period, as may one or
two of the forms above now reconstructed to PC. For some of the words listed
below,
Indie
are
languages
as
suggested
the
source.
In most
cases,
it is clear
that
the words are borrowings and in some but not all cases it appears that the ultimate
source was an Indie language. However,
it is not likely that the Indie language
itself was always the immediate donor. Instead, it is likely that some of these
forms
were
borrowed
from
other
Austronesian
languages
of
the
area,
such
as
Malay.
x?ama(:)s 'gold' ,Acehnese muih, muiih (long), Rade mah (short), Jarai
(PL) mah, Jarai (Lee) mah, Chru mi:h (long), N. Roglai m?h, Tsat ma33 (long),
Haroi mah, W. Cham mih, PR Cham mih (short), Wr. Cham mih (short), Malay
?mas, mas; MK: PNB *mah, PMnong *mah, Bahnar (AC) mah- [< ?]. In addi?
tion to its apparent resemblance to Sanskrit hema-; heman (Headley, #4.5), it
looks remarkably like Chinese,
Chinese kim, Mandarin jin.
x?amrec
'pepper;
hot',
cf. Baxter
Rade
(1992:768) Old Chinese
amrec
-v, Chru
amre?
-f, N.
*krjin, Middle
Roglai
amre?
'red pepper', Haroi amr?k 'red pepper shaker (bamboo tube)', W. Cham mr?? f, PR Cham amr?? -f; mr?? -f, Wr. Cham amraik; mraik; MK: Bahnar (AC)
348
Appendix
amre, PSB (Efimov) *mre?, PKatuic

#2.7; Sanskrit marica]
II: The Chamic
Lexicon
[? < Indie; Headley,
*pa[r/rh]i?, *pa[r/rh]i?.
xja:l 'casting net', Acehnese

juia, Rade jal, Jarai (Lee) jal, Chru ja:l
'conical net', N. Roglai jan,W. Cham cal, PR Cham c?l -1,Wr. Cham jal, Malay
jala; MK: Bahnar (AC) j?l; jol. [< ? Indie; Headley, #2.6; Sanskrit jaala]
xja:g 'to guard; gate (way)', Acehnese
jaga < Malay, Rade gak -i, Chru
<
N.
cf.
Sanskrit
Roglai ja:k, Malay jaga. [? Indie,
ja:g,
(according to Coope)]
-f
xmana:s
anu? h
boh
Acehnese
'pineapple',
xmanus -f 'man; person',

manih
Chru
Rade manuih,
N.
'person',
Roglai
teinan
Rade
'pineapple',
-i, Jarai
f; rina:t -f, Malay nanas. [ < Portuguese

were introduced from South America]
>
xnagar
(independently
Cham
nak?r
-i, Malay
make
spondences
xlagar
'country;
borrowed
[< Indie
this
that
form
Acehnese
area',
cf.
Sanskrit
a
was
(Lee)
rup,
Chru
-f, N.
ru:p
The
nagara].
post-PC
<
ina:t
Pineapples
*-i
'country'
laga, W.
corre?
irregular
borrowing.
rupa, Rade
-f 'body',
ru:p
Roglai
naggroa
[? <
Roglai
ia lagar, N. Roglai
Chru
xrupa 'form; image; body', Acehnese

Jarai
N.
panan,
*-s, W.
manusia.
apparently by way of Hindi.
into Acehnese),
n?geri.
it clear
city;
(Lee)
Jarai (Lee)
-v <
manlh
-f, Haroi
Cham m anus -f, PR Cham minuyh -f,Wr. Cham minuis -f,Malay

Indie; Headley, #2.4; Sanskrit manus, manushya, manusha]
x~na:n
[ < Portu?
Jarai (PL) manuih,
man?ih
-f, W.
nanas.
-f, PR Cham minan, Wr. Cham minah, Malay

guese, apparently by way of Hindi]
manuih,
manaih
-f, Chru
Cham manas
rup, Jarai (PL) rup,
Haroi
rup
-fl
W.
'picture',
Cham
rup -f, PR Cham rup -f,Wr. Cham r?p, Malay rupa; MK: PSB (Efimov)
Note
that this is an Indie loanword; the PSB is borrowed from Chamic.
[<
*ru:p.
Indie; Headley, #2.3; cf. Sanskrit r?pa]
'sound;
xsap
voice;
Rade
language',
s?p
'hear;
obey;
tape-recorder',
Jarai (PL) s?p; hi?p, Jarai (Lee) s?p, Chru sap -f 'voice, sound', N. Roglai
?p
-f), Tsat
Wr. Cham
se?24,
Haroi
s?p
-f, W.
sap. [? < Indie; Headley,
Cham
s?p
#2.n;
-vf,
Sanskrit
PR
Cham
-f, cf.
s?p
(s-an
ch?p
-f,
shabda]
xtara 'sky', PR Cham tara,Wr. Cham tara. This form may (ormay not)
be related to PSB *tru:? and PMnong *tr?k. If so, the phonetics of the Chamic
forms may eventually provide clues as to which language was the donor. [? <
Sanskrit
tara;
a Sanskrit
source
exclusive with
sarily mutually
is clear is that it is a borrowing.]
x*yuan
man?ih
-v
'Vietnamese',
has
also
theMK
Rade
been
proposed,
suggestion.
yuan
suggestion
not
neces?
In all this, the one thing that
-v, Chru
manih
Yuan,
N.
Roglai
yuat, Haroi yun, W. Cham Yuan, PR Cham yon, Wr. Cham yuan; MK:
II: The Chamic
Appendix
349
Lexicon
PNB *yun, PSB (Efimov) *'yyuan. [? < Indic; Headley, #2.10; Sanskrit yavana
This last proposed Indic etymology
'barbarian, Greek'].
is, at the very least,
speculative.
In the literature Indic etymologies

that can now be shown to have Austronesian
have been proposed
for two forms
origins.
*phit 'bitter; bile', cf. Sanskrit pitta., see Acehnese

phet, Rade phi?,
Jarai (PL) phi?, Jarai (Lee) phi?, Chru phi:?, N. Roglai phi:?, Tsat phi?24, Haroi
phei?, W. Cham phi?, PR Cham phi?, Wr. Cham phik, Malay pahit, PMP *paqit.
it has been suggested that this form is Indic in origin, the suggestion
Although
lacks any merit as the form is reconstructable back to PAn.
*sira -i 'salt', Acehnese
sira, Rade hra, Jarai (PL) hra, Jarai (Lee) hra,
Chru
sra, N.
Roglai
Haroi
sara,
sra
-i, W.
sra
Cham
-i, PR
Cham
sara
-i, Wr.
Cham
sar?, PMalayic *sira, PMP *qasiRa. Aymonier and Cabaton suggested

Sanskrit source for this form but it now can be established as Austronesian.
2.2
Even
Of Arabic
origin
the earliest Arabic
butions. None
borrowings
contributions
are much
later than the earlier Indic contri?
of them reconstruct
as widely
to PC, nor are the relatively few post-Chamic

distributed as the Indic forms.
x?alak 'rice wine; liquor; alcohol', Acehnese

ara? -i, Jarai (PL) ?lak,
N.
W.
l?c
Cham
PR
Cham al?k; l?k, Wr.
-f,
Roglai alak,
patih -f,
Cham alak, lak, Malay arak 'rice wine; liquor' MK:
PSB (Efimov) *?alak. As
;
Coope correctly notes, the form is a borrowing from Arabic, so the Efimov recon?
struction bears closer examination.
Either the PKatuic time depth is rather shal?
or
some
the
reconstruction
has
low,
problems. [<Arabic]
Chru alak
xtalabat
'worship', Chru talbat

Wr.
talabat -f. [<Arabic]
Cham
talap?t -f,
xumur
late borrowing
'age,
e.g.
the modern
from Arabic
age',
-f, N. Roglai
talabat
Acehnese
Malay
umu,
-f, PR Cham
umur.
[< Very
umuru].
This handful of forms underplays theArabic contribution. Even a cursory exam?

ination of Aymonier and Cabaton's dictionary of Cham would show a large num?
ber of Arabic loanwords, but a survey of the rest of the Chamic languages would
indicate that the Arabic loans were not widely disseminated.
350
II: The Chamic
Appendix
Of MK
2.3
Lexicon
(and other) origin
The overwhelming majority of the words in this section are of MK origin. Many
have been provided with partial MK etymologies
and many more are found
restricted to the Highlands Chamic area, where the only neighbors areMK speak?
ing. Only a handful of these words have been borrowed from another source,
such as Malay or French.
x?abual -v 'blunt; dull', Rade b?l -v, Chru bual, N. Roglai
Cham bul; MK: PKatuic (DT) *p_l 'dull'.
x?amiat
'fr?re
cadet
du
'uncle,
mia?
p?re',
Cham mii? (n)'uncle,
-v
amiet
Rade
aunt',
Chru
N.
'maternal',
'uncle,
Jarai
aunt',
mi??,
Roglai
aban -v, PR
Haroi
met
(PL)
met
-v, W.
aunt', PR Cham ml? (n) 'paternal', Wr. Cham mik.
x?amuan
see
x?anro:g
'carry
xkamuan
(two
objects
from
suspended
Rade
stick)',
enog
-v,
Jarai (PL) anog; ?nog (N), Jarai (Lee) ?anog, Chru ano:g, N. Roglai anro:k, Tsat
nog33, Haroi anog, W. Cham nog 'carry with pole (shoulders)', PR Cham anog,
Wr. Cham anaug; MK: PNB *t?g 'carry on pole', Bahnar (AC) anog, PSB (Efi?
mov)
*tu:g
on
'carry
-f
x?anrog
PR
Cham
aro?, Wr.
pole'.
'toad',
Jarai
Cham
arauk.
(PL)
N.
ajl-?ro?,
anro?
Roglai
aro?
-f, Haroi
-f,
x?i?w 'left (side)', Rade di?u -iv, Jarai (PL) lau, Jarai (Lee) ?ieau -v,
Chru iau, N. Roglai iau, Tsat taii?42 -iv, Haroi eau, W. Cham iu, PR Cham iw,
Wr. Cham iw; MK: PNB *haqew, PMnong *ki?w, PSB (Efimov) *'giaw.
'bamboo
x?jraw
Chru
ira:u
sp.',
-1 'thornless,
?rau < *jr- 'thornless,

*ka[l/lh]a:w,
Rade
drau
'thornless
long-sectioned',
N.
long-sectioned',
PR Cham
Roglai
bamboo',
long-sectioned
?jrau,
Haroi
?jrau, W.
Cham
PKatuic
(Lee) ?jaw, draw,
*3a[l/lh]a:w.
gluih; jluih, Jarai (PL) ?jru?h, Jarai

x?jruah 'barking deer', Achenese
(Lee) ?jruah, Chru iruah 'barking deer', N. Roglai ?juah, Haroi ?jroah 'barking
deer', W. Cham ?joah 'mouse de er', PR Cham ?jrw?h, Wr. Cham ?jruah.
'morning;
x?am-xagu?h
dawn',
Rade
im
'early',
aguah
'morning';
magac 'light, morning', Jarai (PL) maguah, guah, Jarai (Lee) maguah, Chru a:m
-1, gu?h, N. Roglai mugu?h; gu?h, Haroi am. Both forms are restricted in their
distribution
to Highlands
x?amra:k
'peacock',
Chamic
plus Haroi;
Acehnese
muira?
both forms contain loan phonemes.

-v, Rade
amrak,
Jarai
(PL)
amr??,
Chru amr?:?, N. Roglai amra:?, Haroi amra?, W. Cham mra?, PR Cham mra?;
amra?, Wr. Cham amrak; mrak, Malay m?rak; MK: PMnong *brak, Bahnar
(AC) amra, PSB (Efimov) *bra:?, PKatuic *ria?.
-vf
xaguat
'jew's
aguat -f 'Jew's harp'.

-f
xbamo:g
351
Lexicon
II: The Chamic
Appendix
harp',
Restricted
'banana
(Tharpe) mamog, N. Roglai
Rade
aguat
to Highlands
blossom',
Rade
-v
'one
string
mog
'stalk
of
r?m, Chru param
Rade
bananas',
bumo:k -f, Haroi pa mug, PR Cham
xbaram -i 'light (fire)', Rade

found in the highlands languages.
Chru
instrument',
Chamic.
(Lee) mo? -f.
-i, Tsat zan11. Only
xbay 'basket, kind of, Rade bai, Jarai (PL) bai, Jarai (Lee) bai, Chru
ba:i (large, round), N. Roglai bai, Haroi bai; phiai, W. Cham pai (kep), PR Cham
pay; hapay, Wr. Cham bai; MK: Bahnar (AC) bai.
xbiam -lv 'crop (of bird)', Rade bi?m 'goiter', Jarai (PL) bem (Ouest);
(DT) *phl_m 'crop, craw', PKatuic *biam 'craw of fowl'.
MK:
PKatuic
Chamic.
xbo 'empty', Rade (Tharpe) bo, Jarai (Lee) bo. Restricted

[The vowel indicates aMK origin]
to Highlands
xbu:r -f 'soup', Acehnese bu 'rice, Rade bur -f, Jarai (Lee) bur, Chru bu
'porridge', N. Roglai bu, Haroi phu -f 'cooked rice', W. Cham pu, PR Cham pu
f,Wr. Cham b?, Malay bubur 'broth', PMP *buRbuR;MK: PSB (Efimov) *po:r
'gruel, soup'. Although
quite irregular reflexes.
this form is also found in PMP, within Chamic
it has
xbube 'goat; sheep', Rade be, Jarai (PL) be, babe (S), Jarai (Lee) babe,
Chru pabe, N. Roglai mube, Tsat phe11, Haroi phi, W. Cham pape, PR Cham
pape
'goat';
PMnong
papo
'sheep',
Wr.
Cham
pabaiy
*be, PSB (Efimov) *babe:, PKatuic
'goat';
*?b[e/e]:,
pabauh;
MK:
PNB
*babe,
*?b[e/e]:?.
xbum -vl 'blind', Rade bum -1, Jarai (PL) bom; bum, Jarai (Lee) bum;
bom
Blood)
-v, Chru
bo:m
mata
-vl, N.
Roglai
bu:p
-1, Haroi
phum
-1, PR
Cham
(Lee,
bom 'night blind'.
x?a? -i 'carry on back', Rade ???, Jarai (PL) ???, Jarai (Lee) b?? -i, Chru
ba? -i, N. Roglai ba? -i, Haroi ??? 'carry in carrying cloth', W. Cham p?? -i, PR
Cham p??, Wr. Cham bak -i; MK: PNB *p?q ?, PMnong *ba?, PMalayic *ba?
'carry'. The irregularities in the Chamic forms plus existence of the two Bah?
naric forms mark the word as MK;
thus, the PMalayic form looks to be bor?
rowed.
x?ag 'table', Chru ?a:g, N. Roglai ?ag -f, Haroi ba:g -?-vr; MK:
(Efimov) *b?(o):g. Restricted to Highlands Chamic plus Haroi.
x?ig 'we',
'(incl.)', PMnong
PSB
Jarai (PL) ?ig-, Haroi big sia 'other'; MK: PNB *(q)b?n
*b?n, PSB (Efimov) *ba:tn 'we (incl.)'. Restricted to High?
lands Chamic.
352
Lexicon
II: The Chamic
Appendix
blag 'rice field, rice growing

x?lag 'plains, delta; yard', Acehnese
cultivated
Chru
flood
?la:g 'yard', N. Roglai ?la:k, W. Cham ?lag
plain',
plain,
Wr.
PR
Cham
Cham
?lag; MK: Bahnar (AC) blag; balag,
?lag,
'plains, delta',
PKatuic
*traaji 'field, plain'.

x?lit see x(ma)klit
x?luar -f 'to tell a lie', Rade luar -v, Jarai (PL) ?lar, Jarai (Lee) ?lor -v;
blor -v, Chru lar -i, N. Roglai la -i, PR Cham (Blood) ??r -f 'slander'.
x?uan -ivf 'island', Jarai (PL) ?ul (Ss), Chru bu:t -ivf pala:u, N. Roglai
PR
?on 'island; small hill', Wr. Cham ?uan.
Cham
?uat,
x?uh 'see', Rade ?uh, Jarai (PL) ?uh, Jarai (Lee) ?uh, Chru ?uh, N.
Roglai ?uh, Tsat ?u55, Haroi ?5h -vr,W. Cham ?oh -v, PR Cham ?oh; ?oh -v,Wr.
Cham ?oh; MK: Bahnar (AC) b?h.
Cham
cay, Wr.
trai? -f, PR
Jarai (PL) cai, Chru ca:i -f, N. Roglai
xca:i -f 'sap; resin',

cai.
Cham
xcabi? 'sack', Jarai (PL) cabl? (Ouest), Jarai (Lee) cabi?, Chru sabi:?, PR
Cham capi?; kapi?, Wr. Cham cabik; kabik.
xcagam 'handspan', Rade kagam, Jarai (PL) cagam, Jarai (Lee) cagam
1,Chru sagam, N. Roglai cagap 'thumb to middle finger', Tsat kha:gn ?-l, Haroi
'measure
cakhiam
-1,Wr.
cakam
Cham
xcagow
jagau,
Chru
to middle
thumb
from
W.
finger',
Cham
-1, PR
cakam
Cham
cagam.
bear',
'Malaysian
'black
sagau
bear',
N.
Acehnese
Roglai
cagea, Rade
Haroi
cagau,
Jarai (PL)
kag?u,
cakhiau,
W.
Cham
du? 'black bear', PR Cham cak3w, Wr. Cham caguw; MK: PMnong
*cak?w, Bahnar (AC) ?ag?u, PSB (Efimov) *'jrakaw. This word for 'bear' is
found throughout Southeast Asia. Probably originally aMK word.
cakau
xcada:g; xrada:g 'crack open', Rade kadag 'split, divide', Jarai (Lee)
cadag, Chru cada:g; sada:g, N. Roglai cada:k, Haroi cadag, PR Cham rad?g
mi, Wr.
Cham
radag.
xcadog
N.
Roglai
Cham
*kando:g,
cadok
-f
'flat
-f
cadaug; MK:
basket',
'large,
round
PMnong
Chru
basket',
*d?g
cadog
round
'large,
Haroi
'winnowing
cadog,
PR
basket',
basket',
winnowing
Cham
cadog
PKatuic
-1,Wr.
*kadug,
*3adug.
xcaku:g 'carry (several)', Rade kakug, Jarai (PL) cSkug (Pk), Jarai
(Lee) cakug, Chru sako:g, N. Roglai caku:k, Haroi cakog 'two carry something
heavy on a stick over the shoulders', W. Cham cakog, PR Cham cakog; takog,
Wr. Cham cakog; MK: Bahnar (AC) cakug; cokag ?, PKatuic *caka:g 'carry on
pole'.
II: The Chamic
Appendix
353
Lexicon
xcana:g 'furniture; bed', Jarai (PL) canag, Jarai (Lee) canag, Chru
cana:g, N. Roglai can?g, W. Cham canig 'wooden bed', PR Cham tanig -i,Wr.
Cham tanig 'bed'; MK: PSB (Efimov) *'cA:g 'bed'.
cante?, Wr. Cham (AC) cantik, Malay
'beautiful', Acehnese
is a late borrowing from Malay into both Cham and Acehnese.
xcantik
cantik. This
xcaguar -f 'flat basket',

canua
-f, Chru
cagua
Haroi
basket',
nowing
-gf
basket
'winnowing
W.
cagoa,
-f, Jarai (Lee)
Rade kaguor, Jarai (PL) cSgua

Cham
N.
(pointed)',
-vf
cagoa
'win?
cagu?
Roglai
'winnowing'.
xcap 'bundle, tie', Rade cap, Jarai (Lee) c?p, c?? -f ca? -f 'of chignon',
Chru ciau? ?i, Haroi c?u? 'to tie', W. Cham cau?, PR Cham (Lee) c??,Wr. Cham
(AC) cap.
xcata?see
xtatuh
xch-an-ar,
'dibble
cf.
stick',
xchar
cf. cha; chana? kalai

Roglai
these terms to Highlands Chamic.
chana
Haroi
'plant
with
digger'.
'posthole
Chru
stick',
Note
N.
ca?ar,
the restriction
xchar -f 'gong', Rade char, Jarai (PL) car, Chru sa:r, N. Roglai
sal, PR Cham char, Wr. Cham char; MK: Bahnar (AC) car.
of
cha,
xcho? 'scoop up; ladle out', Jarai (PL) so? ia, Jarai (Lee) s5?, Chru so?
ia 'dip water', N. Roglai cho? (ia), W. Cham ch??, PR Cham ch5?, Wr. Cham
jhauk; MK: Bahnar (AC) kh. chak.
xchum
-1 'pants;
chum
Rade
clothes',
-1 cf.
chi?m
Jarai
'cloth',
sum glog (Mod.), Jarai (Lee) som -v, Chru sum pha -1, N. Roglai cupha
Cham capa -v 'trousers'. Restricted to Highlands Chamic plus W. Cham.
xchap 'dirty', Rade cho?, Chru sau?, N. Roglai
soc
-fv groe,
PR Cham
xchar
'plant
tasow?,
with
Wr.
Cham
Chru
stick',
Chru
ciam
-1, N.
'feed;
Roglai
Rade
nourish',
ciap,
N.
clem
cha
Roglai
Chamic
ce am, W.
Haroi
-f,W. Cham
tica?; magap
tasap.
car,
restriction of these items to the Highlands

xciam
(PL)
-1,W.
Wr. Cham ciem, PSB (Efimov) *siam, PKatuic
chana.
Note
the
languages.
Jarai
-v; meiern,
Cham
cf.
kiam
*hace:m,
(Lee)
-iv, PR
[pa]
Cham
cem
-v,
ciem
-v,
*?ace:m.
xciag 'carry on side', Rade cleg, Jarai (PL) cieg -v, Chru ce:g -v, Haroi
ce g -v,Wr. Cham (AC) qag.
xcreh
'mark;
draw
line',
cf.
'write;
draw',
Rade
treh
'make
lines',
Jarai
(Lee) creh 'to strike', PR Cham cr?h, Wr. Cham craih.

xcrih 'strange; unusual',
crlh, Wr.
Cham
Chru crih crag, N. Roglai
ctih -n, PR Cham
crih.
354
Appendix
II: The Chamic
Lexicon
xcroh 'stream, creek', Rade crok -f, Jarai (PL) croh, Jarai (Lee) croh,
Chru croh, N. Roglai croh, Haroi caroh, W. Cham croh, PR Cham croh, Wr.
Cham
crauh.
xcu:r 'lime (for betel)', Rade c?r -f, Jarai (PL) c?r, Jarai (Lee) cur, Chru
cu:r, N. Roglai cu, Haroi col, W. Cham cu -f, PR Cham cur,Wr. Cham cur; MK:
PSB (Efimov) *cu:r.
xcum 'kiss; smell', Acehnese
com, Rade (bi) ??m, Jarai (PL) cum,
cum
'nose
Jarai
Chru
bacum,
kiss', N. Roglai cum -f, Haroi com, W.
(Lee) cum,
Cham cum, PR Cham c?m, Wr. Cham cum, Malay cium 'kiss, smell' MK:
PNB
;
*tach?m, PMnong *c?m, PSB (Efimov) *cum 'nuzzle'.
-vf 'squeeze; grasp', Acehnese
(Pk), Jarai (Lee) capa? 'to knead',
jupat -i, Rade kap?t, Jarai (PL)

Chru capa?, N. Roglai capa?,
xcupa/et
k?pit,
Haroi
capit
capa?,
W.
Cham
-f.
cap?t
xcu?uay; xtu?uay 'lips; gums', cf. 'beak', Rade ka?ue, Jarai (Lee)
ca?uai; ta?uai; tabuai; sa?uai, Chru sa?uai, N. Roglai cu ?uai; cubuai, Haroi
ca??ai, W. Cham ca?uai, PR Cham caboy, Wr. Cham ca?uai; MK: PKatuic
*ha?bar, *tam?bar 'lips' ? [cf.Malay bibir].
dhoi < *-l, Rade dhul 'fog, mist,
xdhual/r -f 'dust; fog, mist', Acehnese

vapor',
Chru
thul
-v, N.
Roglai
Cham dhur -v. [The vowel

xdian
-v, W.
indicates aMK
Rade
'candle',
thun
pui
di?n,
th?l
Cham
-v, PR
th?r
Cham
-vf, Wr.
origin]
dian
Chru
-v, N.
Roglai
then?42,W. Cham tian, PR Cham tien, Wr. Cham dien, Malay

(DT) *taen.
di?n
-f, Tsat
dian; MK: PKatuic
xdriaw 'exclaim; acclaim', Jarai (PL) dreu, Haroi triau, PR Cham triew,
[The vowel indicates aMK origin]
Wr. Cham driew.

xdran
-if
trtin 'benumbed;
'numb',
asleep
cf.
xnran
-if, Chru
drin
-v, N.
Roglai
diin
-vf, Haroi
(foot)', PR Cham tr?n,Wr. Cham dran.
xdap -ifv 'hide', Rade due? d?p -fv, Jarai (PL) kad?p, pad?p; bad?p
'quelque chose', Jarai (Lee) (pa)dop -fv, Chru padau?, N. Roglai da?; pada? -i;
kada:? -v, W. Cham patau?; tau?, PR Cham thow?; pakadow? cacher dans un
endroit secret; pa?ow?; kadow?; kawa?, Wr. Cham thap; p?gadap; gadap ?tre
cacher; kawak, Malay
?ndap.
xdaw 'all', Chru dau

so far, both Highlands Chamic
xdel
'shallow',
den; MK: PMnong
'at all', Tsat ?dau33. Only found in Chru and Tsat

languages;
Acehnese
*thal, PKatuic
dui a
[<MK]
-i, Rade
*[h/s]andial,
edal
-v, Chru
del,
N.
Roglai
*[p/b]a?di:l.
xdig -f 'tube; pipe (for smoking)', Rade dig, Jarai (PL) dig, Jarai (Lee)
dig; dig, Chru dig -1,N. Roglai dit -f, Haroi dig, W. Cham dig v? ?ju? 'pipe'; dig
Appendix
II: The Chamic
355
Lexicon
*qd?g 'bamboo
tju?, PR Cham dig, Wr. Cham dig; MK: PNB

*dig, Bahnar (AC) dig, PSB (Efimov) *dig.
PMnong
pipe',
'glutinous rice', Rade dio?, Jarai (PL) dia?? ?, Haroi diu?, W.

to Highlands Chamic plus Haroi and W.
(Headley) /diau?/. Restricted
xdiap
Cham
Cham.
xdoh 'flow; run off, Rade doh, N. Roglai doh 'subside'.

Note
run
flow'
to two Highlands
x*duac.
Cf.
off;
'run;
languages.
the
restriction
Roglai
xdu:g 'float', Rade dug, Jarai (PL) padug; dug hie g hlo g, Chru do:g, N.
du:k, Haroi dog -vr, W. Cham dog, Wr. Cham (AC) do:g; MK: Bahnar
(AC) d?g.
xduan 'bamboo hat', Rade duon, Jarai (PL) hado?n, Jarai (Lee) duan v, N. Roglai duat, Tsat duat33 -f, PR Cham don, Wr. Cham don; MK: PMnong
*?du?n, Bahnar (AC) duan, PSB (Efimov) *'duan, PKatuic *?duan.
xdvt -v 'small', Rade diet, Haroi d?t.
Restricted
to Highlands
Chamic
plus Haroi.
-lv 'tickle', P-Acehnese
(Durie) *gle?, Acehnese
gli?-gli? -vi,
Jarai (PL) glek, Jarai (Lee) gl?k ??, Chru galek -f; glek -if, N. Roglai gilet -f,
Haroi kal?k -vf, PR Cham kal?k, Wr. Cham galaik, Malay g?le te k; g?li; g?lak
xgalek
'laugh'; MK: PKatuic
(DT) *-lek.
xgiam/p -ivf 'carry under arm', Jarai (PL) g?p, Chru kiap -1,N. Roglai
'clutch,
kiap
Chamic,
carry
Haroi,
under
and W.
arm',
W.
Cham
kiam
-v.
Cham.
xgoh 'clean', Chru goh, N. Roglai goh; MK:

restriction to closely-related Chru and N. Roglai.
xgriag
'fang;
tusk',
Rade
gri?g
-v, Chru
griag,
N.
W. Cham kr?g -v, PR Cham kr?g, Wr. Cham gr?ln, PSB

eye
to Highlands
Restricted
PNB
Roglai
*ragoh.
Note
giag
-f 'incisor',
(Efimov)
the
*gani:g 'fang;
tooth'.
xgrit 'dirty', Jarai (PL) grl?, Jarai (Lee) grl?, Haroi khri?; kri?,W. Cham
kroc -vf, PR Cham (Lee) kri?.
dloh; kloh -i,
(Durie) *groh, Acehnese
xgroh 'to bark', P-Acehnese
N.
Jarai
Jarai
Chru
(PL) groh,
(Lee) groh,
groh,
groh,
Roglai groh, Tsat
khia55, Haroi kr?h -v; krouh -v, PR Cham kro h,Wr. Cham grauh; MK: PMnong
Rade
*groh,
PKatuic
xgrua?'lie
*garuah.
prone'.
xgrak 'vulture; garuda', Jarai (PL) gri?, Jarai (Lee) gri?, N. Roglai gra?,
Haroi kri?,W. Cham kra?? PR Cham kra?,Wr. Cham grak, Malay Garuda 'eagle
of Vishnu'.
356
Appendix
II: The Chamic
Lexicon
xgram -vf 'thunder', Rade gr?m, Jarai (PL) grom, Jarai (Lee) grom, Chru
gram, N. Roglai gram -f, Tsat khim11 or khien11, Haroi kriam, W. Cham kram -v,
PR Cham kr?m -v,Wr. Cham grurri;MK: Bahnar (AC) gram, PKatuic *garim.
xgua?'lie
prone'.
xhadruam
'book', Rade hadruom hr??, Haroi catr?m; hatrum -i. Thus

seems
to two languages, one a Highlands
this
restricted
far,
language and the
now
other
found in the highlands.
xham -1 'greed', Jarai (Lee) ham, Chru ha:m -1 la:m, N. Roglai
MK: Vietnamese
th?m lam, PKatuic *ha:m 'greedy'.
ham -f;
xhanu?? -v 'right (side)', Rade hanu??, Jarai (PL) hanu??, Jarai (Lee)
hanu??, Chru hanua?, N. Roglai ha nu?? (this form either indicates the original
or that it was reanalyzed in Northern Roglai), Haroi
form was bimorphemic,
hand?,W. Cham hanui?, PR Cham han??; n??, Wr. Cham hanuk; nuk, PMP *ka
wanan, *qaqay; *taqu 'right hand', Malay kanan. Note that it is possible that
this form may ultimately turn out to be related to the Austronesian
forms given
here, and the irregularities may simply reflect the reduction of a multi-syllabic,
multi-morphemic
entity to a single disyllabic
form.
xho:g -f 'wasp', Acehnese

h?ag, Rade hog, Jarai (PL) hog, Jarai (Lee)
hog, Chru ho:g, N. Roglai hog -f, Haroi hog, W. Cham hog; MK: PNB *og, PSB
(Efimov) *?o:g, PKatuic *ha:g.
xhuac 'fear', Rade hul? -f, Jarai (PL) hul? -f, Jarai (Lee) hu?? -f, Chru
huai? -f, N. Roglai hue?, Haroi hw?i? -vr, W. Cham huai?, PR Cham (Blood)
hoc; MK: PSB (Efimov) *rahyu:.
xhuc 'whistle', Rade hoc, Jarai (PL) hoac, Chru srui:t -if, N. Roglai
huai? -v, W. Cham hui?, PR Cham huyt -f, Wr. Cham huit -f; MK: PKatuic
*gaho:j? *haho:j?.
xhuni 'bee; honey', Acehnese
hani, Jarai (Lee) hani, Chru hani; hani:g
hani, W. Cham ea hani; hog, PR Cham
pected final ill that primarily marks this
some related forms inMalayic
(Adelaar).
xhama?
-ivf
Chru mahu?, N. Roglai

Cham
ham?t
-f, Wr.
'hear', Rade hami?, Jarai (PL) ham!?, Jarai (Lee) hml?,

hma? -v, Haroi hmak -f,W. Cham hamlt -f metial?, PR
Cham
hamit
xih 'you; thou', Rade

(polite)'; MK: PNB *?h.
Cham
unoa, Rade ea hanue (m), Jarai (PL)

'honey', N. Roglai hun?, Tsat ni33, Haroi
hani, Wr. Cham hani. It is the unex?
as a borrowing;
itmay, however, have
-f.
ih 'you (sg.)', Jarai (PL) ih 'to a single person
xjam 'plate; dish', Rade jam 'dish', Jarai (PL) jam, N. Roglai (jia), W.
cam -1, PR Cham cam -1,Wr. Cham jam; MK: PMnong *jam, Bahnar
(AC) jam.
II: The Chamic
Appendix
357
Lexicon
xja?u 'dry in sun', Rade ?hu madie, Jarai (PL) c???? -f, Chru sa?u -i
to
'dry rice', N. Roglai ja?u padai, Haroi ca?u, W. Cham ca?u -i 'spread rice
ba?u.
The
varies
Cham
Wr.
first
PR
Cham pabu -i,
inexplicably.
syllable
dry',
Jarai (Lee) sig 'stilts', PR Cham
xjhuag -n (?) 'long-legged',
MK: PSB (Efimov) *jo:g 'long', PKatuic *(ha)?jog, *ga?Jog 'long'.
xjia 'taxes', Rade jia, Jarai (Lee) jia, Chru ji?, N. Roglai
v, PR Cham (Lee) ci.
(Lee) ch?g;
ji?, Haroi
sia -
xjo:g 'axe', Rade jog, Jarai (PL) jog, Jarai (Lee) jog, Chru jo:g, N. Roglai
Tsat
suog11; sog11, Haroi sug; sug, PR Cham ?og; a?og, Wr. Cham jaug;
jo:k,
ajaug; MK: PNB *c?g, PMnong *sug, PSB (Efimov) *su:g, PKatuic *cu[a/a]g.
xjr?w 'medicine', Rade drau, Jarai (PL) jrau, Jarai (Lee) jrau, Chru
jra:u, N. Roglai jrau, Tsat sia(:)un, Haroi cari?u, W. Cham cru, PR Cham cru,
Wr. Cham jr?; MK: Bahnar (AC) jar?u, PSB (Efimov) *jrA:w?, PKatuic
*taha:w.
*harhaw,
xjro 'large jar', Chru jro, N. Roglai

jro 'water crock'
(also a elf.).
xjro 'rainbow', Rade ke? ero; keft kro, Jarai (PL) ero (Se), Jarai (Lee)
cro ?, N. Roglai tagalo jro, Haroi caro -vr, W. Cham cro -i. Restricted to High?
lands Chamic plus Haroi andW. Cham.
xjut -fl 'small bamboo', Rade j?t, Chru
Restricted to Highlands Chamic plus W. Cham.
jut -il, W.
Cham
cut
-f.
xjag -vf 'become', cf. *pajag, Rade jig -v, Jarai (PL) p?gi?g, Jarai (Lee)
Chru
jig -v,
jiag -v, N. Roglai jak, Haroi slg,W. Cham ?iag -v, PR Cham ?yag
v, Wr. Cham jiag; MK: PMnong *j?g, PSB (Efimov) *jeg 'be born; become'.
xkadual
'heel', Rade kadul, Jarai (PL) kadul, Jarai (Lee) kadul, Chru
'elbow', N. Roglai kaduan, Haroi kadul, W. Cham kadual, PR Cham
kadol, Wr. Cham kadual; MK: PNB *kaqn?l, PMnong *kandal, PSB (Efimov)
kadual
PKatuic
*gandA:l,
*gan?dA:l,
*sandual.
-f 'porcupine', Rade kasua -f, Jarai (PL) kats?a, Jarai (Lee)

kasu?
Chru
-f, N. Roglai kasu? -f, Haroi kasoa, W. Cham kasur, PR
-f,
Cham kath?r, Wr. Cham kathur; MK: Bahnar (AC) gacor, PKatuic *[h/?]agkAjh,
xkasu?r
kasua
*yagkA:jh.
xkatri 'scissors', Rade katrei, Chru katrai -f, N. Roglai katri, Haroi
W.
Cham katray, Malay k?l?kati 'areca-nut scissors' from Tamil?; MK:
katr?i,
PKatuic *-darh[e/a/a]j, Bru kantraj.L.
xkhi:n 'dare; brave', Jarai (PL) khin, Jarai (Lee) khin, Chru khin, N.
Roglai khin -f, Haroi kh?n, W. Cham kh?n 'covet; desire', PR Cham khin, Wr.
Cham khin, Proto-Hr?-Sedang
*khln.
358
II: The Chamic
Appendix
Lexicon
xkhiag -fi 'want; desire; marry', Rade cl?g -v, Jarai (PL) ki?g, Jarai
(Lee) ki?g -iv, Chru khiag -v, N. Roglai khiag -f also verbal auxiliary, Haroi
ceag -i,W. Cham khin -f, PR Cham khlg, Wr. Cham khiri-kh?g; MK: PSB (Efi?
mov)
desire'.
'want,
*kA:ji
xkr?m -lvn 'bamboo',

Chru
-1 'large,
kra:m
Rade
N.
thornless',
kram, Jarai (PL) kram, Jarai (Lee) kram,

kram
Roglai
kram
-f, Haroi
Cham
-1, W.
krim -v, PR Cham krim -v,Wr. Cham krirri.
Haroi
'to cut
pa-kreau
off
xkuac
dead
-1 'pick up (handful)'
Chamic plus Haroi.
xkuho
man?ih
-iv
'Koho',
kuac
Rade
claw',
'scratch,
koai?
-f
PNB
;MK:
cit.
Jarai
'scrape,
*kuc. Again,
kaho
Rade
kaho
-f, Haroi
to Highlands
restricted
manih
Chru
-iv,
kuac
(Lee)
-v, N.
Roglai
cr?u 'Tring, Koho', W. Cham kaho.
kuho; man?ih
xkun -f 'fold; bundle; curled', Rade k?n, Jarai (Lee) kun 'curled', N.
kun -f 'lie in a heap'. Restricted to Highlands Chamic.
Roglai
-f, Tsat
Rade
'grasshopper',
xkuto:k/p
kutoik
to?42 ?, Haroi
katok
-f, W.
-vf, Chru kato:k
katuop
-f MK:
;
katok
Cham
-f, N. Roglai
PKatuic
xkajap 'firm; solid', Rade kaj?p, Chru kajap -f,W. Cham

MK: PMnong *kalj?p. Restricted to Highlands Chamic.
x?amuan
xkamuan,
amuon
niece',
-v, Wr.
(Efimov)
'nephew,
N.
Cham
'nephew;
Jarai
niece',
'nephew,
'nephew,
kamon
PSB
-i,
-v
branches'
Rade ku?t, Chru kuit, N. Roglai
xkuit 'tangerine',
ki?u
Jarai (PL) kre?, Jarai (Lee) kreu, N. Roglai
xkri?w 'castrate',
Roglai
kamuan;
*kamo:n
(PL)
kamu?n,
MK:
'nephew,
sister's
aman,
Haroi
amon,
PNB
*mon
niece',
?amon,
(Lee)
W.
Cham
'nephew',
PKatuic
kac?p
km m uan,
Acehnese
son',
Jarai
*gu[a/o]?.
Chru
kamuan,
PMnong
*[s/?]amha:n,
-vf;
Rade
kamuan
PR
Cham
*kamon,
*karmha:n
niece'.
xla:t 'flat', Rade lat, Jarai (PL) lat, Jarai (Lee) lat; la?, N. Roglai la:?,
Haroi la?, PR Cham klet -ivf, PMnong *ralat. Headley (1976) identified the PC
form as borrowed from MK, an analysis supported both by the internal irregular?
ities and by the presence of the form in various MK languages. Two lookalikes,
rata and Malay rata 'smooth; level' are unrelated to the PC. As Ade?
Acehnese
is PMP
laar (p.c.) noted, the ultimate source of the Malay and the Acehnese
*datar. From PMP *datar, Malay has, not unexpectedly, datar, but in addition
Malay has rata, borrowed from Javanese (also descended from PMP *datar). The
Malay form rata was then borrowed into Acehnese.
xla?an 'cold', Rade e?t -f, Jarai (PL) ra?ot -fv, Jarai (Lee) ra?ot -fv, N.
Roglai la?at, Tsat ?an33, Haroi la?a? -vf, W. Cham la??n -v, PR Cham li?n; la?n,
Wr. Cham lian; laan; MK: PMnong *?lik, PSB (Efimov) *la?i:? 'cool, chill'.
II: The Chamic
Appendix
359
Lexicon
*lan 'earth' + xkliat -v '?' > 'clay', Rade l?n tliet, Jarai (PL) l?n, Chru
lan klia?, N. Roglai lat matli:?. The xkliat -v '?' is restricted to Highlands
Chamic.
x(li)hug 'papaya', Rade tei hug -1, Jarai (PL) hog (Pk), Jarai (Lee) hug,
Chru lahog -v, N. Roglai lahog -f,W. Cham lahSg -v, PR Cham (Lee) lihog; MK:
PNB *rah?g.
xligiaw 'outside', Rade egau -v, Jarai (PL) r?g?a?, Jarai (Lee) ragiau,
lagiau -v,W. Cham lagiu, PR Cham ligiw; lagiw; giw, Wr. Cham ligiw;
lagiw; giw. [The vowel suggests aMK origin.]
Haroi
xluan -v 'to swallow', Rade lun, Jarai (PL) l?n, Chru lu?n, N. Roglai
luat -v, Tsat luan33 -vf, Haroi lun,W. Cham luan -v, Malay telan, PMP *telen;
*tilen; MK: PKatuic *hali:n, *halA:n.
xlayuh 'shake', Rade eyuh, Jarai (PL) ray?h, Jarai (Lee) rayuh, Chru
N.
Roglai yuh 'wiggle', Haroi lay?h, W. Cham yuh 'chill, shivering', PR
layuh,
Cham y?h; r?h -i, Wr. Cham yuh; ruh; MK: PMnong *ragu 'shake', Bahnar
(AC) ruh.
kliat -iv, Rade ?llt 'stuck;
x(ma)klit; x?lit 'sticky; pasty', Acehnese
caught', Jarai (PL) ?lit, Jarai (Lee) tli? -i, N. Roglai tli:? 'in sticky rice', Haroi
?li?;MK: Bahnar (AC) klep.
-f 'rich',
xmadar
mada,
PR
ami?,
Chru
Cham
xme?
mipa
-vf
Jarai
-f, Wr.
-f, N.
(awai),
Roglai
'to cheat',
Chru
mada
-f, N.
Roglai
mid?.
Acehnese
'mother',
ame
m a dar
(Lee)
Cham
ma,
mak,
Haroi
umi,
ami?
Rade
mi,
'mother;
ami?,
address
Jarai
(PL)
term',
W.
Cham me?; m??, PR Cham ame?; me?, Wr. Cham maik, Malay (?)ma(?); MK:
PMnong *me; *me?, PSB (Efimov) *me: 'mother, female', PKatuic *hambe:?,
*?amh[e/e]?.
xmray
mara:i,
N.
'thread', Rade mrai,

mrai;
Roglai
murai,
Tsat
Jarai (PL) mrai,

za:i33,
Haroi
marai;
Jarai (Lee) mrai, Chru

camroi
'string',
W.
Cham mrai, PR Cham mray, Wr. Cham mrai; MK: PNB *bray, PMnong *bray,
PSB (Efimov) *bra:y 'web, thread', PKatuic *[h/s]amriaj, *pariaj, *pej.
xmuar -f 'termite', Rade muor -v, Jarai (PL) mua, Jarai (Lee) mua, Chru
mua, N. Roglai mua, Tsat mua33 'termite'?, Haroi moa, W. Cham mur, PR Cham
mu -f,Wr. Cham m?; MK: PKatuic *kamhuar.
xnran
-if
'numb',
Chru pran -i;MK: PKatuic

'silver'.
cf.
xdran
-if,
Jarai
(PL)
kron,
Jarai
(Lee)
nron
-v?,
*sabi:n.
x*gan 'money' from 'silver', Rade g?n, Jarai (Lee) g?n, cf. Tibetan dgul
[This form is from the Tibeto-Burman word for 'silver'.]
360
Appendix
II: The Chamic
Lexicon
xpajag -vf 'create', cf. xjag, Jarai (PL) pajig -v, Jarai (Lee) pajig
Chru pajiag -v, N. Roglai pajak, Haroi pasig, W. Cham pa?iag -v.
-v,
buiruitoh -i, Rade matuh, Jarai (Lee)

xpatuh 'to explode', Acehnese
N.
Haroi
Chru
partuh -r,
Roglai patuh,
patuh,
patouh -v, PR Cham pat?h, Wr.
Cham patuh; MK: PSB *bratoh, PMnong *bartoh, PSB (Efimov) *bratoh.
Roglai
xpayar -f 'to offer', Rade myar, Jarai (Lee) payar, Chru paya? -f, N.
paya ga, Haroi payol, PR Cham (Lee) payai, Wr. Cham (AC) payar, cf.
Malay
bayar
'to pay'.
-v
xpetrug
Rade
madrog,
to Highlands
Chamic
'rich',
*padr?g. Restricted
Haroi
-vr;
petrug
MK:
PMnong
plus Haroi.
xphaw 'gun', Rade phau, Jarai (PL) phau (modern), Jarai (Lee) phau,
Chru pha:u -1 'rifle', N. Roglai phau, Haroi phau, W. Cham phau, PR Cham
Wr.
phaw,
Cham
phaw.
xphug 'leper; leprosy', Rade ph?g, Jarai (PL) ph?g, Jarai (Lee) phug, N.
Roglai phug -f, Haroi p?g -i.
xpirak -If 'silver; money', Acehnese pira?, Rade pr?k, Jarai (PL) prak,
Jarai (Lee) pr?k, Chru pria? j?n -1 (m), N. Roglai paria? (m), W. Cham parea?
(m), PR Cham pary?? (m); pirak 'white', Wr. Cham pariak; birak 'white',
Blust
Malay p?rak; MK: PKatuic *[p/b]arha?, Khmer prak 'silver; money'.
(p.c.) notes that this form is a loan inmany An
MK
languages and is found in some
as well.
languages
xpliar
-f
'hail',
Rade
'snow,
plier
Chru
plia
-f, Haroi
paleal,
W.
Cham
(Headley) /plia/; MK: PNB *pr?l, PMnong *pli?r, PSB (Efimov) *pliar, PWB
(Thomas) *priaw, PKatuic *parhiel. Restricted to Highlands Chamic plus Haroi.
xpra:l -vf 'strong; well',
(Lee)
pr?n
-f, Chru
pram
Rade pral 'well, healthy',
-f 'strength',
N.
Roglai
pran
Jarai (PL) pran, Jarai
-f, Haroi
pral;
-f, W.
pran
Cham prin -f, PR Cham prin -f,Wr. Cham prin.

(?)
xpruac
'stomach;
intestine,
large',
Acehnese
pruat,
Rade
proc
-v,
Jarai (PL) proal?, prua??, Jarai (Lee) pru?i?; pre? -f, Chru pruai?, N. Roglai para?
-f; puai?, Haroi pr?ai?, W. Cham proi? -f, PR Cham proy?, Wr. Cham pruac; MK:
PKatuic
*ruaj?
'instestine'.
xprus/h -vf 'to squirt', Rade pruih, Chru kapru:h, N. Roglai kapru -vf,
Haroi prouh -v 'spit out', W. Cham pruh. Restricted to Highlands Chamic plus
Haroi andW. Cham.
xpak -fvl 'granary', Rade (Lee) pi? -v, Jarai (PL) pik -f, Jarai (Lee) pik
f, N. Roglai pa:? -1. Restricted to Highlands Chamic.
xraba:g
raba:k, W. Cham
-if
'bridge,
suspension',
rapo g 'canal'
Rade
-v, PR Cham
kaban
-if, Chru
raba:g,
N.
Roglai
ripag; rapag, Wr. Cham
rabag.
ribag;
Appendix
II: The Chamic
xrada:g
see
xramah
361
Lexicon
xcada:g
-v
emeh
Rade
'rhinoceros',
'rhinoceros
horn',
Jarai
(Lee)
ramah, Chru ramah, N. Roglai rum?h, Haroi lamlah, W. Cham ramih, PR Cham
ramih, Wr. Cham ramih; MK: Bahnar (AC) ramai, PKatuic (DT) *ramaas.
xramiat 'prepare; put away', Rade emiet -v, Jarai (PL) r?met pioh,
Chru ramia?, N. Roglai lumia? -i, Haroi lamia?, W. Cham ramii? (n), PR Cham
raml? (n),Wr. Cham ramik; MK: Bahnar (AC) ram?t, PKatuic *mian 'prepare'.
xraguat 'sad', Rade eguot -v, Jarai (PL) r?got, ragua?, Jarai (Lee) ragua?
-v, Chru raget-f.
xrimo:g 'tiger', Acehnese

rimuag, Rade emog, Jarai (PL) ramog, Jarai
N.
Chru
ramo:g,
(Lee) ramog,
Roglai lum?g -i, Haroi lamug -vr, W. Cham
Wr.
Cham rimaug; ramaug, Malay rimau; hari
ramog, PR Cham rimog; ramog,
mau. This history of this word is quite unclear; Hudson reports the same word in
theWest
Barito branch of his Barito
but with
the meaning
'leopard':
languages
Kapuas
of southeast Borneo,
(1967:14)
harimau, Ba'mang harimau, Ketingan

and Siang horomaug. Note
haramaug, Dohoi haramaug, Murung (II) horomaug,

the final velar nasal in several of these.
xriya:?
-1 'wave;
Acehnese
surf',
ripple;
riyuia?,
Chru
ria:?
Tsat
'wave',
(pho11) za?24, W. Cham raya?, PR Cham raya?, Wr. Cham ray?k, Malay riak.
Blust (p.c.) notes that xriak is confined to a few languages in Borneo, plus Malay
and Malagasy
lan?
(which, of course, is a Southeast Barito subgroup of Malayic
form riaka is a borrowing
guages of Borneo); the Malagasy
denced by it Ixl to Ixl correspondence with Malay (Adelaar).
as evi?
from Malay
ruag, Rade r3g 'upper back', Jarai (PL)

xrog 'back (anat.)', Acehnese
<
Jarai
*kChru
(Lee) rog,
grog
r?g,
'upper back', N. Roglai tulaik turok 'back
bone', Haroi r?g -vr, W. Cham rag, PR Cham rog -1,Wr. Cham raug; MK: PNB
*(ka)r?g, PKatuic *[k/g]arhag; *[k/g]alhog; *[k/g]alho:g.
xrua? -v 'painful',
rawa?
-v
saki:?,
'sick',
PR
Cham
N.
Rade
rua?
Roglai
rw??, Wr.
Cham
ru??, Jarai (PL) ru??, Jarai (Lee) ru??, Chru

'ache',
Haroi
r??;
ru??
-v
-vr, W.
Cham
roa?
ruak.
xruc 'pull', Jarai (PL) rot; ruh, Jarai (Lee) rue, Chru grui?; prui?, N.
rui? 'set trap or bow', Haroi rui? -v, PR Cham r?y?,Wr. Cham rue; MK:
Roglai
PKatuic
(DT) *r.q 'pull up'.
xraga 'shake', Rade ega; MK: PMnong *ragu 'shake', PSB (Efimov)
*ra'ggu:. Restricted to Rade.
*sa xsit 'little; a few', Chru sit -If, Haroi aset -vf 'a little', W. Cham
haslt
-f;
sit
-f, PR
petit'; MK: Bahnar
Cham
asit
-i, sit;
sit
(AC) to?iet, PKatuic
'petit,
peu',
Wr.
Cham
asit,
sit;
sit
*[h/s]ag?i:t.
'peu
362
ea
Rade
'gasoline',
xsag
ma?i?
II: The Chamic
Appendix
s?g,
sag, N.
Chru
Roglai
ia
Lexicon
Cham
-f, W.
sag
s?g.
xsagat 'asthma; rhinitis', Jarai (PL) bagat -i, Jarai (Lee) hagot -vi
'asthma', N. Roglai saga? 'stuffed-up nose'. Restricted to Highlands Chamic.
MK]
[Probably
xsi?jual -v 'light (not heavy)', Rade ha?jul, Jarai (PL) ?jul; ?jhul, Jarai
(Lee) ?jhul -v, Chru sa?ual, N. Roglai si?juan, Haroi ha?jul, W. Cham ?jual, PR
Cham ha?jol; ?jol,Wr. Cham ?jual; ha?jaul; MK: Bahnar (AC) hajac.
-iv
xsimag
Cham
g -v
samo
Rade
'cement',
-i, Chru
sim?g
samag,
N.
sim?g,
Roglai
W.
[< French].
xxsisi(r) > xtasi 'a comb; hand of bananas', Rade kasi; kasir matei,
Jarai (PL) tasi; tasi, Jarai (Lee) tasi; sir; si -if, Chru tasi; tasi, N. Roglai kasi;
tasi, Tsat si33, Haroi casei; casei, W. Cham tasi; tasi, PR Cham tathi; tathi, Wr.
Cham tathi, Malay sisir; MK: PKatuic *kaci:?; *[h/s]anci:?. The post-PC bor?
Malay
xxsisi(r) form is what I assume the earliest Chamic

sisir) before it underwent dissimilation).
Roglai
xsr?p 'crossbow', Jarai (PL) harau, Jarai (Lee) hra? -f, Chru sr?:u? -1,N.
sr??, Haroi sr?u?, PR Cham thru?,Wr. Cham thruk.
rowing
form looked like (cf.
xsra:p -f 'tired of, Jarai (Lee) hr?p, Chru sr?p -f 'fed up with',
. Restricted to
sr?p -fv 'tired of
xsr?
'debt,
Chamic
Highlands
xsroh
Cham
-n?
sroh; MK:
Chru
owe',
sr?, N.
sr?, W.
Roglai
sre.
Cham
Haroi
to
Restricted
plus W. Cham.
rice',
'polish
PKatuic
Acehnese
*saruah
sroh;
rhoh,
Chru
Rade
'pound
rice',
W.
'pound'.
xsro:k 'fishtrap', Rade hrok, Chru sro:?, N. Roglai

Highlands Chamic.
xsro? -n 'subside',
sroh
hro?, Chru
sro? -n.
sro:?. Restricted
Restricted
to
to Highlands
Chamic.
xsr?h 'nest; swarm', Rade hruh, Jarai (PL) hr?h, Jarai (Lee) hr?h, Chru
N.
sr?h,
Roglai sr?h, Haroi srouh -v; cahrouh -v, W. Cham sruh, PR Cham thr?h,
Wr. Cham thruh; MK: PKatuic (DT) *sr.h, PKatuic *soh, *so:h. Note that this
form borrowed into PC is only attested in Katuic thus far.
xsuac
N.
extract',
Highlands
'pull out', Rade
Roglai
Chamic
suai?
-f, Haroi
kasu??,
Jarai (PL) so??, Chru
s?ai?, W.
Cham
soa?
-f 'extract'.
sua? -f 'pull,
Restricted
to
plus Haroi andW. Cham.
xtali -if 'flat (of large rocks)',

PR Cham tali, Wr. Cham tali.
Jarai (Lee) kli -i (borrowed), Haroi calei,
-v 'corn; grain',
xtagay
363
Lexicon
II: The Chamic
Appendix
tagai,W. Cham
Chru
PSB
tagai,
(Efimov)
*tagA:y.
xtatuh; xcata? 'shake (blanket); tremble', Chru tartuh, N. Roglai tatah,

cata?, tata?. The first variant seems restricted to Highlands Chamic;
the second occurs in coastal Chamic dialects.
W. Cham
xthaum 'visit', Rade bi tuom 'visit', Chru to:m 'meet', Haroi th?m, W.
Cham torn; torn kau?. Restricted to Highlands Chamic plus Haroi andW. Cham.
Restricted
jin paria?, Haroi
xtlen 'money', Chru pria? jen, N. Roglai

< Vietnamese
?en.
thug -f,W. Cham
xthug 'barrel', Rade th?g, Chru thug, N. Roglai

to Highlands Chamic plus W. Cham.
tien
'money'.
The
other
root
thug.
sin,W. Cham
is 'silver'.
xtiog -fvl 'mynah bird', Chru tio:g ira:u, N. Roglai tiog -f, Malay tiung.
This word has been independently borrowed intoMalay and Highlands Chamic.
Roglai
xtra:p 'heavy', Rade katro? -vf, Jarai (PL) trau, k?tr??, Chru tra?, N.
tra:?,Haroi trau?,W. Cham trau?, PR Cham tr?? -1,Wr. Cham trak.
xtu?y 'guest; visitor; stranger', Rade tue -v, Jarai (PL) toai -v, Jarai
(Lee) tuai, Chru tuai, N. Roglai thu?i -i, Haroi t?ai, W. Cham tuai, PR Cham toy,
Wr. Cham tuai; MK: PNB *tamoy.
xtruam -f 'trunk (of animal)',
Jarai (PL) trom, Chru tro:m, N. Roglai
PR
W.
Cham
Cham trom, Wr. Cham traurri;MK:
trom,
trom,
trom -f, Haroi
not
'trunk'
*tam
PMnong
clear
from
gloss
sort
what
. [The
trunk
of
vowel
also
suggests aMK
origin]
xtuki -v 'horn; antler', Rade ki -v 'antler (deer)', Jarai (PL) t?ki, Jarai
(Lee) taki, Chru taki, N. Roglai tuki, Haroi cake -v,W. Cham take -v, PR Cham
take, Wr. Cham take; MK: PNB *ake, PMnong *gke, PSB (Efimov) *gke:, PKa?
tuic
*yake:,
*yaki:.
see
xtu?uay
xtabiat
'go
xcu?uay
out;
appear',
Acehnese
kabi??
Rade
tuiblet,
'go
out',
Jarai
(PL) tabi??r Jarai (Lee) tabi??, Chru ta?ia?, N. Roglai tubia?, Tsat phia?42, Haroi
caphia?, W. Cham tape a?, PR Cham tapy??, Wr. Cham tabiak.
xwa:r
-fl
'stable;
pen',
Acehnese
wuia,
Rade
war
-f,
Jarai
(PL)
war,
Jarai (Lee) war, Chru wa:r 'pen', N. Roglai wa, Haroi wal, W. Cham war -1, PR
Cham wal -f,Wr. Cham wal -f ;MK: PSB (Efimov) *wa:r, *wa:g 'shed, cattle
pen'.
xwa:s
away'.
-f
Restricted
'wipe
away',
to Highlands
Rade
waih
'clean
up,
Chru
wa:ih
-f
Chamic.
'clear
364
Appendix
II: The Chamic
Lexicon
xway -f -vr 'spider', Rade w?k wai, Jarai (PL) w?g wai, Haroi wiai
-vr; MK: PNB *way 'spider web'; *wey 'spider web'. Restricted to Highlands
Chamic plus Haroi.
xweh -vf 'turn aside; visit along the way',
Rade weh 'turn (right or
we
Jarai
N.
Haroi
-f,
(Lee)
w?h; wlh -v, W. Cham weh
left)',
Roglai weh,
'dodge', PR Cham w?h, Wr. Cham waih; MK: PNB *weh 'turn aside', PKatuic
*wih,
*wi:h
'turn'.
xyu:? 'descend', Jarai (PL) y?? 'ouest', y?? gah yang hrai le? ( Ouest),
Jarai (Lee) y?? 'west', Haroi yo? -vr, PR Cham (Lee) yo?, PNB *j?r, PMnong
*j?r. If theMK forms are related, the final -r is unexpected.
Appendix
II: The Chamic
3.
English-Chamic
A
a few, 361
abdomen, 306
able, 305
about to, 320
above, 282, 326
absent, 341
acclaim, 354
according to, 345
ache, 342
after, 290, 292
the modem
e.g.
bamboo
age,
alcohol, 305, 349

alive, 290
all, 280, 354
allow, 326
341
324
amass,
and, 289, 331

ant, 303
363
antler,
348
area,
289
321
around,
arrive, 333
313
arrow,
ascend, 316
ashamed, 297
ashes, 289
ask, 304
301
assemble,
asthma, 338, 362

at, 288, 294
aunt, 282, 307, 350
await, 314
326
awaken,
axe,
357
349
small,
banana,
284
banana
blossom,
357
3 51
bank, 337
bank (river), 319
banyan, 283
bare, 339
bark, to, 355
barkcloth, 321
barking deer, 350
barrel, 363
basket, flat, 353
basket, kind of, 351
basket, large, 336
winnowing,
325
bathe, 297
be on back,
arm, 304, 330

armpit, 342
armspan,
350
sp.,
bamboo,
basket,
363
appear,
330
bamboo strip, 345
agitate, 341
air, 298
always,
Index
back(anat.), 361
bad, 292
bail (water to catch fish),
bald, 339
balete, 283
bamboo, 358
bamboo (medium), 309
again, 318
age,
365
339
afternoon,
Lexicon
316
335
beads,
beak, 333, 354
beam, 341
bean, 344
bear (Malaysian), 352
beat (gong), 310
beautiful, 353
because, 339
become, 357
bed, 353
bedbug, 335
bee, 356
before, 288
behind, 290, 323
belly, 306
below, 318, 334
beneath, 334
betel, 300
betel lime, 354
366
300
betel-nut,
328
big, 282, 302, 318,

bile, 300, 349
bird, 314
bird of prey, 322
bird, mynah, 363
bite, 322
bitter, 300, 349
black, 290, 320
blanket, 323
bleed, 319
blind, 286, 351
blood, 287
blossom,
blow,
away,
blow
e.g.
282
carry
carry
348
suspended
from
brother-in-law,
335
buffalo,
322
312
327
348
net,
358
castrate,
335
320,
337
shoulder,
on side,
353
under arm,
355
casting
360
suspension,
carry
on
casket,
book, 303, 311,

bough, 287
bow, a, 300, 331, 362
bowl, 278, 338, 342
bracelet, 324
brain, 317
branch, 287, 336
brave, 357
break, 327
break, to, 323
breast, 304
breath, 298
water,
objects
stick), 350
carry (wear) on head, 317
carry on back, 318, 351
356
bring,
broken, 312,
broom, 303
(two
carry
300
boat, 317
body, 281, 288,
boil, 344
thunder, 339
bolt, to, 338
bone, 306
bridge,
cadet, 280
cage, 329
calf of leg, 285
call, 310
can, 305
carry, 310, 327, 335
carry (several), 352
blow nose, 299

blow whistle, 282
blue, 290
blunt, 350
blush, 297
boa, 323
board,
358
315
328
the wind,
bug, 330
bunch, 310
bundle, 317, 353,
Lexicon
bum, to, 314

burnt smell, 323
bury, 316
butt, to, 336
buttocks, 323
buy, 285
336
blow
II: The Chamic
burntrns.,
351
banana,
Appendix
cat, 298
cattle, 340
CAUSATIVE,
301
362
cement,
center, 301
centipede, 295
chaff, 319
Cham, 336
Champa, 337
change, 345
charcoal, 290
chase, 328, 345
cheap, 313
cheat,
to,
359
cheek, 326
chest, 287
chew, 297
chicken, 297
child, 281
Appendix
II: The Chamic
298
child-in-law,
324
a,
count,
307
to,
country, 348
cover, to, 318, 340
cow, 340
crab, 335
crack open, 316, 352
330
to,
360
creek, 354
crocodile, 286
crop (of bird), 351
cross, 317
314
crossover,
clay, 359
clean, 355
clear brush, 338
crossbow,
elf. for round objects,

elf. long, thin objects,
climb, 316
close eyes, 327
close, to, 322, 339
cloth, 277, 323
clothes, 353
clothing, 310
cloud, 320, 338
coffin, 312
coiled, 313
cold, 358
331,
crow,
285
313
a,
309
crush, 339
crust at bottom of pot, 323
cry, 310, 319
cry, to, 315
cucumber, 305
cup, 278
curled, 358
cut, 329
cutoff, 324
cut up, 345
D
311
dam,
a,
collapse,
318
dam,
to,
colorful,
336
damp, 284
comb,
362
a,
come,
dance,
297
command,
327
to,
comparable,
344
344
compare,
complete,
326
conserve,
343
contents, 281
cook, 305, 344
cooked, 345
copy, to, 345
cork, 344
com, 322, 363
corpse, 281
362
314
crossroads,
344
cleaver,
cotton, 277, 322, 347

cough, 284, 346
create,
358
to,
310
crawl,
city, 348
civette, 341
clam, 340
clap, 343
claw, 293
claw,
367
correct,
chin, 321
chipped, 312
chisel, to, 300
choke, 346
choose, 330
chop, 284, 314
churning of rapids, 333
citronella, 343
citrus,
Lexicon
312
345
dare, 357
dark, 292
dawn, 350
day, 291
deaf, 345
debt, 362
deer (Sambhur), 302
deer, barking, 350
defecate, 310
deliver, to, 320
delta, 352
descend, 306, 318, 364
desire, 296, 334, 358
368
destroy, 318, 346

dibble stick, 353
die, 298, 326
different, 286, 327
dig, 292, 315
direction, 337
dirty, 353, 355
discard, 326
dish, 278, 342, 356
disk shape, 333
dismantle, 332
distant, 291
dive, 326
divide, 341, 343
dizzy, 333
do, 285, 286, 342
dog, 281
don, 277, 312
earring, 312
earth, 296, 305,
340
emaciated,
empty, 331, 351

encircle, 315
end, 326
enter, 304
280
312
327
feathers, 286
escape, 323
eternally, 330
evening, 296
draw,
314
exclaim,
295
drink, to, 298

drip, 331
drool, 342
drop
drop,
anchor,
a,
3 54
excrement,
310
expensive,
334
explode, 360
extinguish, 299
eye, 297
337
to,
317
error,
draw line, 353

dress,
359
east, 326
easy, 313
eat, 313
eat rice, 320
egg, 285
eggplant, 333
eight, 289
elbow, 357
elbow, to, 326
elephant, 340
eleven, 302
done,
door,
dove,
downy
dream,
344,
Lexicon
340
earthworm,
321
don't,
II: The Chamic
Appendix
294
331
drop, to, 327

drum, 330
drunk, 296
dry, 302, 324, 336,
dry (weather), 323
dry in sun, 357
dry in the sun, 336
dry over fire, 343
duck, 309
dull, 350
dumb, 321
dust, 354
E
ear, 307
ear of grain, 310
345
face, 313
faded, 295
fagot, 345
fall down, 294
fall into, 325
fan,
a,
fan,
to,
337
328
fang, 355
far, 282, 291
fart, 294
fast, 288
fat, 296
father, 280
fear, 356
feed, 353
fell a tree, 316
female, 340
II: The Chamic
Appendix
fence, 299
fence, to, 311,
fetch, 341
few, 288
few,
312
361
a,
field, cultivated, 291

fight (war), 297
fill, 335
finger, 314
fingernail, 293
finished, 280
fire, 281, 332
fire, dead, 343
firebrand, 332
firewood, 310
firm, 358
first (go), 288
firstborn, 321
fish, 282
fish scales, 339
fish trap, 331
fish, to, 346
fishtrap, 362
five, 295
flat, 304, 358
flat (of large rocks), 362
flat object, 312
flatus
ventrus,
294
flesh, 281, 329,

float, 355
flour, 305
flow, 317, 355
flower, 286
flute (front flute),
fly, a, 330
fly, to, 329
fog, 320, 354
fold, 326, 358
fold, to, 326
follow, 345
food, 311
foolish, 289
foot, 292
for (goal), 294
forbid, 343
forearm,
369
forever, 330
forget, 334, 336
fork of tree, 336
form, 348
formerly, 288
four, 299
fowl, 297
fox, 341
free, 331
from, 326
fruit, 285
fry, 316
full, 335, 345
furniture, 353
G
garuda, 355
gasoline, 362
gate(way), 348
gather, 324
gecko, 314, 342
generous,
332
(c. teenage),
287
give, 285
glazed clay, 318
glutinous rice, 355
go, 294, 299
go home, 318
go out, 363
go past, 317
goat, 351
god, 334
gold, 347
gong, 353
gong, small, 315
good, 331, 336
gourd, 312, 328
grab, 315
grain, 322, 363
granary,
337
314
get, 341
get up, 332
ghost, 281
gibbon, 325
ginger, 295
girl
335
330
forehead, 280
forest, 291, 301,
Lexicon
360
grandchild, 315
grandchild, great, 314
grasp, 354
370
348
to,
guest, 363
guide, to, 337
gums, 354
gun, 360
H
hail, 360
hair, body,
half, 340,
I
I (familiar), 293
I (polite), 291, 337
image, 348
imitate, 345
286
341
345
hammer,
hammer,
to,
Lexicon
hole, 295, 312

honest, 307
honey, 356
honorific prefix?, 303
horn, 363
horse, 346
hot, 319, 347
house, 343, 344
how many, 338
how much, 338
hundred, 302
hungry, 295
hurt, 344
husband, 311, 343
hut, 344
grass, 330
grass (congo), 301
grass (thatch), 301
grasshopper, 358
grease, 296
greed, 356
green, 290, 298
group, 318
grow, 304
guard,
II: The Chamic
Appendix
329
277, 312
hammock, 334,
hand, 304
340
IMPERATIVE,
in, 287, 288
hand
362
INADVERTENT, 307
incest, 334
insect, 330
inside, 287
INSTRUMENTAL infix, 280
insufficient, 294
interior, 344
of bananas,
handful, 322
handle (knife), 318
hands on hips, 337
handspan, 352
hangup, 332
hard, 339
harvest (rice), 320
harvest,
to,
intestine,
331
hat (bamboo), 355

hatch, to, 293
have, 280
hawk, 322
he, 299
head, 309
head hair, 287
hear, 356
heavy, 363
heel, 357
here, 306
hide, 354
high, 318
hit with implement,
hold, 310, 335
intoxicated,
360
large,
296
invite, 291
iron, 287
island, 301,
itch, 341
itchy, 289
ivory, 311
352
J
jar, large, 357
jaw, 321, 326
Jew's
310
harp,
351
Jorai, 320
jump, 338
jungle, 291, 301,
just now, 283
337
Appendix
II: The Chamic
K
keep, 343
kettle, 318
kiss, 354
knee, 306
knife, 278
knockdown,
know, 305
Koho, 358
318
L
ladder, 302
ladle, 310
ladle out, 353
lake, 287
lance, 292
land, 344
language, 348
last, 292, 332
lastborn, 332
later, 292
laugh, 339
launder, 311
lazy, 296
lead,
335
to,
leaf, 331
leak, 319
lean,
leech,
340
land,
leech, water,
328
295
left (side), 350

legumes, 311
leisure, 331
lemongrass,
343
leopard, 361
leper, 360
leprosy, 360
less, 294
letter, 303
lick, 341
lid, 332
lie down, 317
lie full length, 345
lie prone, 355, 356
lie suppine, 316
lie, to tell a, 352
lift, 334
lift, to, 334
Lexicon
371
light (fire), 351

light (not heavy), 362
lightning, 339
like (prep.), 338
lime (for betel), 354
lineup, 337
lips, 333, 354
liquor, 349
little, 288, 313, 361
live, 288, 290
liver, 290
living beings, 288
lizard, 314, 342
lobster, 290
lock, to, 338
log, 300
loins, 338
long, 282
longtime, 331
long-legged, 357
look at, 337
look for, 319
loose, 341
lord, 301, 342
lose, 323
lose, to, 323, 325
louse, head, 294
love,
343
lower
part,
318
lungs, 325
lust, 334
M
make,
make
342
a wall,
342
Malaysian bear, 352

male, 294, 311, 332
man, 348
many, 338, 341
mark, 353
317
marrow,
marry, 358
master, 301,
mate,
measure,
to,
342, 343
345
to,
303
meat, 281, 329, 335

medicine, 357
meet together, 342
372
nourish, 329, 353

now, 335
numb, 354, 359
345
metamorphose,
middle, 340, 341

miserly, 339
mist, 320, 354
money, 359, 360,
monkey, 293
month, 286
moon, 286
more, 296
morning, 350
363
280
residence,
much,
338,
mucus,
299
open
362
owe,
ox,
P
341
package, 335
pain, 342
painful, 302, 361
palm, 299
palm (areca), 300
281
name,
317
narrow,
navel,
301
near,
320
neck,
332
335
necklace,
needle,
negative, 277,
nephew, 358
nest, 362
casting,
coconut,
palm,
292
net,
340
329
to,
naked,
287
wide,
318
359
outside,
my nah bird, 363
nail,
335
286,
312
otter,
309
muzzle,
eyes
other,
321
mute,
335
open, 328, 329

open (mouth to say sthg.), 309
341
336
mushroom,
husband,
on time, 344
one, 302
onion, 340
only, 331
mouth, 283
move, 341
move
sister's
older
315
range,
305
front, 281
offer, to, 360
oil, 296
old, 309, 339
old (people), 306
old (things), 330
older brother's wife,
mosquito, 298
mother, 359
mother (animal), 282
mountain
o
ocean,
296
mortar,
Lexicon
not, 313
not yet, 338
305
311
melon,
mend,
II: The Chamic
Appendix
312,
348
new, 283
nice, 331
niece, 358
night, 296, 339
nine, 302
no, 313
nose, 282, 313
313
294
340
pancreas,
pants, 353
papaya, 359
paper, 311
parch, 344
345
in-law,
parent
parrot, 337
PARTICLE,
pass
over,
pasty,
patch,
306,
321,
334
317
359
to,
311
path, 291
Appendix
II: The Chamic
Lexicon
pound, 328
pour, 345
pour out, 319
prepare, 361
press, 293
pretty, 331
prick, 337
prompt, 344
protect, 342
prove, 325
pea, 344
peacock, 350
peck, 322
peck (of bird), 315
peel (with a knife), 314
peel, to, 325
pen, 363
penis, 323
pepper, 347
perforated, 338
perished, 326
permit, to, 285
person, 283, 294, 301,
pestle, 290
pick, 327
pierce, 312, 337, 338
pig, wild, 283
pigeon, 339
pillar, 332
pillow, 335, 344
pinch, 293
pine, 319
pineapple, 348
pipe (for smoking), 354
pit, 295
place, 326, 327
placenta, 325
plains, 352
plane,
to,
300
plank, 300
plant, 300
plant floss, 286
plant with stick, 353
plant,
a,
340
plant, to, 343

plate, 356
play, 296
pluck, 327
pocket, 321
poison, 347
poke, 324
pole, 318
polish rice, 362
poor, 313, 316
porcupine, 357
post, 318, 332
pot, 318
pouch, 321
373
348
pull, 317, 319, 339,

pull out, 331, 362
pull up, 312
pus, 294
put, 326, 327
put away, 361
python, 323
Q
question word,
quit, 326
303,
361
338
R
rabbit, 332
rain, 290
rainbow, 357
raise, 329, 334
304
ransom,
rat,
305
rattan (generic?), 290

raw,
298
receive,
to,
333
red, 296
reflexive (-self), 288
reject, 326
330
remember,
rescue,
304
resin, 289, 352

rest, 342
return,
318
Rhade, 329
rhinitis, 362
rhinoceros, 361
rhinoceros, hornbill, 337
ribs, 302
rice (glutinous), 355
rice (husked), 285
rice dust, 319
374
rice wine,
rice,
349
scoop
295
cooked,
rice, paddy, 299

rich, 359, 360
ridge of (house, mountain),
right, 310, 336
right (side), 356
right hand, 356
rim, 337
ring, 322
ripe, 345
ripple, 361
river, 324
road, 291
roast, 311, 344
Roglai, 301
roll,
289
to,
285
round,
333
rub, 343
run, 317
run after,
355
317,
361
314
salamander,
salt, 303, 349

salted, 297
salty, 297
sand, 315
sap, 289, 352
satiated, 345
337
savage,
saw,
304
to,
311
say, 325
scabbard, 344
scales (fish), 339
scar,
345
319
search,
see, 352
seed, 346, 347
seize, 315, 331
sell, 301
send, 327
separate,
343
separate
a fire,
servant,
291
343
316
294
sesame,
306
seven,
sew, 291
shade, 286
shadow, 286
shake, 359, 361
shake
(blanket),
shake
out,
363
330
shaman,
336
341
sharp, 319
352
sad, 321,
save,
to,
sculpture,
sea, 305
share,
S
sack,
353
up,
shaky, 341
shallow, 354
345
runoff,
335
Lexicon
scrape, 324, 338

scratch, 312, 338, 358
scratch (an itch), 321
scratch (of chicken), 343
serve,
roof, 335
roof thatch, 301
room, 280
root, 283
rope, 304, 320
rotten,
II: The Chamic
Appendix
295
scissors, 357
scold, 313, 328
sharpen,
shave,
281
324
shavings, 335
she, 299
sheath-like, 344
sheep, 351
shellfish, 340
shirt, 310
shoot (bamboo), 301
shoot (bow), 300
shore, 319
short time, 288
shorten, 324
shoulder, 284, 330
shrimp, 290
shy, 297
sibling, elder, 335, 344
sibling, younger, 280
II: The Chamic
Appendix
Lexicon
sick, 302
sickle, 346
side, 337
sister's
358
son,
spherical shape, 333

spicy, 319
spider, 364
spill, 319
spin, 342
spirit, 286, 334
spit, to, 321, 322
spittle, 342
splash, 330
spleen, 340
split, 284, 312, 319
spoilt, 320
spoon, 310
spread out (amat), 3 25
sprout, to, 304
squash, 328
squeeze, 293, 354
squirrel, 328
squirt, 360
stab, 324
363
335
sister-in-law,
sit, 288
six, 299
skin, 293, 325
skin, dead, 331
skirt, 312
sky, 295, 348
slap, to, 327, 343
slave, 291
342
sleep, 317, 327

slow, 331
slurp, 303
small, 355
smell, 354
smell burnt, 323
smoke (of fire), 281
smoke tobacco, 310
snail, 334
stable,
star,
333
sneeze,
336
sniffle,
299
step on,
309
snout,
333
soak,
soil, 305
sole, 299
solid, 358
283
someone,
298
son-in-law,
336
sorcerer,
soul, 286, 298

sound, 348
sound
of
336
stay, 288
steal, 323
stem, 300
stench, 285
322
snap at,
363
a,
stalk, 310
stand, 316
stand up, 332
standing upright, 345
283
snare,
292
spear,
360,
snake,
341
soured,
silver, 359,
sin, 317
sing, 309
sink, 318
sip, 303
slaver,
375
turbulent
rapids, 312
soup, 351
soup, thin, 310
sour, 297
water
in stream
320
stick, 317
stick, dibble, 353
stick, plant with, 353
sticky, 359
stiff, 339
still, 288
sting, to, 316, 344
stinger, 315
stomach, 306, 360
stone, 284
stop, 316
stopper, 344
store,
storm,
343
329
stove, 337
376
straight, 307
straighten, 337
strange, 353
363
stranger,
straw (rice), 328

streaked, 336
354
stream,
strike, 314, 328

strike (snake), 315
string, 304
striped, 286, 336
strong, 339, 360
strong feelings, 328
submerge, 326
subside, 362
suck, 341
suck in, 303
suckle, 341
307
sugarcane,
sun,
291
surf, 361
surplus, 296
swallow,
to,
swarm,
362
359
303
sweep,
297
sweet,
swell, 284, 312

swim, 326
swollen, 284
sword, 278
T
table, 351
tadpole, 324
tail, 282
take, 335, 341
346
take apart,
takeoff, 332, 334

talk, 328
tall, 318
tangerine, 358
taro, 290, 340
341
taste,
tasteless,
tell a lie, 352

ten, 300
tender, 298
tendon, 283
tent, 344
termite, 359
test, 325
testicles (of animal),
that, 281, 337
there, 337
taxes, 357
teach, 311
tell, 327
293
there is, 280

they, 299
thick, 293
thigh, 300
thin, 340
thin (material), 295
think, 344
thirst, 296
this, 282
thorn, 289
thou, 338, 356
thousand, 302
339
thread, 359
three, 293
thunder, 356
tickle, 355
tie, 353
tie together, 301
tie, to,
282
tiger, 361
tired, 317, 325, 340
tired of, 362
to, 294, 342
toad, 350
299
tomorrow,
tongue, 288
tooth, 289
torn, 332
tortoise, 294
towards,
342
trade, 344
transport,
304
Lexicon
280
there are,
tear,
319
sweat,
II: The Chamic
Appendix
327
trap, 333
trap (fish), 331
trap (fish-), 362
tread, 320
Appendix
II: The Chamic
tree, 293
tree species,
turn around,
turn aside,
w
waist, 338
wait, 314
walk, 294, 299
want, 358
war, 297
wash, 311, 329
wasp, 356
watch, 337
water (fresh), 282
333
364
turtle, 294
tusk, 311, 355
twenty, 289
twill, 297
twist, 284, 339,
two, 288
wave,
346
u
uncle, 282, 307,
undress, 332
unlucky, 316
unripe, 298
untie, 333
until, 333
353
unusual,
urinate,
341
use,
335,
to,
346
vegetables, 311
vehicle, 347
vein, 283
347
venom,
verb prefix, 298

very, 316
vessel, 283
Vietnamese,
village, 299
vine, 320
vinegar, 297
virgin, 284
350
348
361
wax (candle), 301

we, 351
we (ex.), 292
we (incl.), 288, 291, 351
weak, 325
wealth, 337
wealthy, 337
wear
wear,
336
ornaments,
337
to,
weasel,
341
weave,
297,
wedge,
upgrade, 326
uproot, 312
364
363
312
turnover,
377
visit, 343, 363

visit along theway,
visitor, 363
voice, 348
vomit, 300, 311
vulture, 355
321
tremble, 363
trim, 345
true, 336
trunk, 300
trunk (of animal),
try, 325
tube, 354
tuber, 290
tumeric, 294
turn, 342
Lexicon
to,
346
316
weed, 338
weeds, 330
weep, 319
well, 339, 360
west, 364
wet, 284
whistle, 356
white, 301, 324
whittle, 340
who, 303, 338
why? what?, 319
wicked, 292
widowed, 283
wife, 326
wild, 337
wind, 281
wine (rice), 305, 349
wing, 303
378
304
winnow,
wipe
wrapped
363
away,
wire, 293
with, 289, 331
withered, 295
wobbly, 332, 341
340
woman,
wood, 293, 317

work, 285, 286
worm, 291
worn, 332
worried, 321
worship, 349
wound, 295
wrap up, 317
wrap,
to,
331
Appendix
II: The Chamic
around,
wring, 339
write, 303,
Lexicon
313
314
Y
yam, 290
yard, 352
yawn, 320
year, 305
yell, 310
yellow, 294
yesterday, 298, 306
yoke, 346
yolk, 335
you, 338, 356
young, 298
References
Kuala
Hamid Mahmood.
1994. Sintaksis dialek Kelantan.
Dewan Bahasa dan Pustaka. Kementarian Pendidikan Malaysia.
Abdul
Lumpur:
Abdul Rahman
al-Amadi.
literature. Actes du
1994 [1987]. Champa in Malay
Seminaire sur le Camp? organise a l'Universit? de Copenhague, Le 23 Mai
1987.
translation in Proceedings
English
of the Seminar on Champa
Dinh
Rancho
trans.,
Te,
Cordova, CA. Pp. 100-109.
(1994), Huynh
on
KA.
More
1988.
Adelaar,
Proto-Malayic. Rekonstruksi dan Cabang-Cabang
Bahasa Melayu Induk. Edited by Mohd. Thani Ahmad and Zaini Mohamed
Zain. Kuala Lumpur: Dewan Bahasa dan Pustaka, Kementerian Pendidikan
-.
Malaysia. Pp. 59-77.

1992. Proto-Malayic,
lexicon and morphology.
the reconstruction
Pacific Linguistics,
of its phonology
Series C-119.
and parts of its

Canberra: The
Australian National
-.
-.
University.
relevance of Salako for proto-Malayic
and for Old Malay
en
tot
de
volkenkunde
148:382-408.
taal-, land-,
epigraphy. Bijdragen
1995. Borneo as a cross-roads for comparative Austronesian Linguistics.
In Peter Bellwood,
James A. Fox, and Darrell Tyron (eds.), The Austrone
sians: Historical
and comparative perspectives.
Pp. 75-95. Canberra: The
Australian National University.
Anttila, Raimo. 1972. An introduction to historical and comparative linguistics.
New York: Macmillan.
Asmah Haji Omar. 1983. The Malay Peoples of Malaysia and Their Languages.
Kuala Lumpur: Dewan Bahasa dan Pustaka. Kementerian Pelajaran Malay?
1993. The
sia.
and Vurnell Cobbey. 1977.

Northern
Roglai:
Roglai vocabulary
[Roglai-Vietnamese-English
vocabulary], Manila, SIL.
Aymonier, Etienne Fran?ois. 1889. Grammaire de la langue chame. In Excusions
Aviong, A-Ty, A-Ly, Maxwell
Suraq vunga sanap Radiai: Ngu-vung
Awai-hathe,
et reconnaissances.
-.
Saigon.
1891. Les Tchames

24:187-237 & 261-315.
and Antoine
Paris,
Cobbey,
XIV.31:1-92.
et leurs religions. Revue de l'histoire des religions
Cabaton.
1906. Dictionnaire
cham-fran?ais.
PEFEO
7.
Leroux.
Studies I. The Lin?

Banker, Elizabeth M. 1964. Bahnar affixation. Mon-Khmer
guistic Circle of Saigon & The Summer Institute of Linguistics. Pp. 99-118.
379
380 References
-.
Baxter, William H. 1989. An Austroasiatic word for cotton in Chinese, ms.

Berlin: Mouton
1992. A handbook of Old Chinese Phonology.
de
Gruyter.
Peter. 1978. Man's conquest of the Pacific. Auckland,

1985. Prehistory
of the Indo-Malaysian
Archipelago.
Bellwood,
-.
demic
-.
Collins.
Aca?
Orlando:
Press.
1991. The Austronesian
dispersal
and the origin of languages. Scientific
American
-.
265.1:88-93.
1992. Southeast Asia
before history. The Cambridge History of South?

Volume 1. From Early Times to c. 1800. Nicholas Tarling (ed.).
Cambridge University Press. Pp. 55-136.
James J. Fox, and Darrell Tryon (eds.). 1995. The Austronesians: Histor?
ical and Comparative Perspectives.
Canberra: The Australian National Uni?
east Asia.
-,
versity.
-.
Benedict, Paul K. 1941. A Cham colony on the island of Hainan. Harvard Jour?
nal of Asiatic Studies 4:129-34.
1984. Austro-Tai parallel: A tonal Cham colony on Hainan. Computa?
tional Analyses of Asian & African Languages
22:83-86.
in the
1976. Austroasiatic
and prehistory
subgroupings
In
N.
Laurence
C.
and
Jenner,
Malay peninsula.
Philip
Thompson,
Stanley
Starosta (eds.), Austroasiatic
Studies Part 1:37-1128. Honolulu: University
Benjamin,
Geoffrey.
of Hawaii
Blagden,
C.
Press.
O.
1929.
Achinese
In Feestsbundel,
and Mon-Khmer.
uitgegeven
van K?nsten en Wetenschap

door het Koninklijk Bataviaasch Genootschap
van zijn 150 Jarig Bestaan
1778-1928.
Vol. 1.Wel
pen bij gelegenheid
tevreden: G. Kolff, 35-38.
units in Cham. Anthropological
Blood, David L. 1967. Phonological
Linguistics
9.8:15-32,
-.
-and
1977. A three-dimensional
analysis of Cham sentences. Papers in South
East Asian Linguistics No. 4: Chamic Studies. Edited by David Thomas,
Ernest W. Lee, and Nguyen Dang Liem. Pacific Linguistics Series A, No.
48:53-76.
Doris W. 1977. East Cham working dictionary. SIL. mf.
in Cham.
of Proto-Malayo-Polynesian
4.9:11-20.
Anthropological
Linguistics
1977. Clause and sentence final particles in Cham. Papers in South East
Asian Linguistics No. 4: Chamic Studies. Edited by David Thomas, Ernest
W. Lee, and Nguyen Dang Liem. Pacific Linguistics Series A, No. 48:39-51.
Blood,
-.
-.
Doris Walker.
1980.
Aspects
1962. Reflexes
of Cham
Culture.
Notes
from
Indochina
on
ethnic
minor?
-and
ity cultures. Dallas, Texas: SIL Museum of Anthropology.

Pp. 11-33.
David Blood. 1977. Eastern Cham working dictionary. SIL. micro?
-,
fiche.
David
Blood,
SIL. microfiche.
and Thien
Sanh Canh.
1976. Eastern Cham
vocabulary.
REFERENCES 381
Blood, Henry F. 1967, 1968, 1974. A reconstruction of Proto-Mnong. MA thesis,
of North Dakota,
U. of Indiana; publ. 1968, SIL/University
115p.; repub
lished 1974, 122p.
Blust, Robert A. 1969. Some new proto-Austronesian
trisyllables. Oceanic Lin?
guistics 8:85-104.
-.
-.
1972.
Proto-Oceanic
with
addenda
Austrone?
in non-Oceanic
cognates
a preliminary list. Working Papers in Linguistics 4.1:1-43.

sian languages ?
Honolulu: Department of Linguistics, University of Hawaii.
a note. Working Papers
in Linguistics
1974. Eastern Austronesian:
Dept. of Linguistics, University of Hawaii.
culture history: some linguistic inferences and their
1976. Austronesian
8:19-43.
relations to the archaeological record. World Archaeology
6.4:101-107.
-.
-.
1977.
The
pronouns
proto-Austronesian
a preliminary
of Hawaii
report. University
and Austronesian
working
subgrouping:
papers
in linguistics
9.2:1-15.
-.
-.
19:1-2:1-181.
1980a. Austronesian
etymologies. Oceanic Linguistics
on the origins of glottalic consonants. Lingua 52:125-156.
1980b. More
-.
-.
1981.
The
-.
Cabang-Cabang
Zaini
-.
-.
of
an
Proto-Malayo-Javanic:
appreciation.
22-23.1 &
III. Oceanic Linguistics

25.1-2:1-123.
1986. Austronesian
etymologies
The linguistic study of Indonesia. Archipel 34:27-47.
1988. Malay historical linguistics: a progress report. Rekonstruksi dan
-.1987.
-.
-.
-.
reconstruction
137.4:456-469.
Bijdragen tot de taal-, land-, en volkenkunde
1983-4. Austronesian
Oceanic
Linguistics
etymologies?II.
2:29-150.
Mohamed
Bahasa Melayu
Zain.
Kuala
Induk. Edited by Mohd.
Lumpur:
Dewan
Bahasa
dan
Thani Ahad
and
Pustaka.
1989. Austronesian
etymologies-IV. Oceanic Linguistics 28.2: 111-180.
1991. Patterns of sound change in the Austronesian
languages. In Pat?
terns of change, change of patterns: linguistic change and reconstruction
Edited by Philip Baldi. Berlin: Mouton de Gruyter. Pp. 129
methodology.
165.
settlement of mainland Southeast Asia. Papers
1992a. The Austronesian
Second
the
Annual
from
Meeting of the Southeast Asian Linguistics Society,
Edited by Karen L. Adams and Thomas John Hudak, pp. 25-83.
in Southeast Asia and South America. Anthropos
1992b. Tumbaga
87:443-457.
-.
-.
1995a. The prehistory of the Austronesian-speaking

peoples: A view
from language. Journal of World Prehistory
9.4:453-509.
1995b. The position of the Formosan languages: Method and theory in
Austronesian
comparative linguistics. Austronesian Studies Relating to Tai?
edited
wan,
by Paul Jen-kuei Li, Cheng-hwa Tsang, Ying-kuei Huang, Dah
an Ho, and Chiu-yu Tseng. Symposium Series of the Institute of History and
Philology, Acad?mica
Sinica, No.
3. Pp. 585-650.
382 References
-.
Comparative
Dictio?
Bahasa Melayu.
Jurnal
1990, 1991, 1992, 1993, 1994, 1995. Austronesian

nary. [Work in progress].
1992. Peminjaman
Brennan, Hugh.
Dewan Bahasa
36.6:552-565.
Briggs, Lawrence
leksikal
dalam
P. 1951. The Ancient Khmer Empire.
Philadelphia:
American
Society.
Philosophical
B?i Kh?nhThe.
1995. T? Di?n Cham-Vi?t.
Nh? Xu?t B?n Khoa Hoc X? hoi.
Burnham, E. C. 1976. The place of Haroi in the Chamic languages. MA thesis,
University of Texas at Arlington.
1901. Nouvelles
recherches sur les Chams.
Publications de
Cabaton, Antoine.
L'?cole Fran?ais D'Extr?me-Orient. Volume IL
Capell, A. 1949. Two tonal languages in New Guinea. Bulletin of the School of
13:184-99.
Oriental and African Studies (London)
Chambert-Loir, H. 1988. Notes sur les relations historiques et litt?raires entre
Campa et monde malais. Actes du S?minaire sur le Campa organise a l'Uni?
versit? de Copenhague, Le 23 Mai 1987. Pp. 95-106. English translation in
Proceedings
of the Seminar on Champa, Huynh Dinh Te, trans., Rancho
Cordova, CA, 1994. Pp. 87-99.
Chao, Yuan Ren. 1930. A system of tone letters. Le Ma?tre Phon?tique.
Russell L. and Jamie James. 1992. Land of the Chams. Archaeology
(May/June) 52-55.
1977. Word lists from ten villages and Du Long word list.
Cobbey, Maxwell.
MSEA: VD 57/58/59-71.
[Note: the Du Long list was completely unread?
able and thus not analyzed].
1979. A statistical comparison of verbs and nouns in Roglai. South-East
Asian Linguistic Studies. Edited by Nguyen Dang Liem. Volume 4:207
212.
1980. A first case of historiography among the Roglai. In Notes from
Indochina on ethnic minority cultures,
edited by Marilyn Gregerson and
Ciochon,
-.
-.
-,
-.
Dorothy Thomas. SIL Summer of Anthropology, Dallas, Texas. Pp. 61-83.

and Vurnell Cobbey. 1980. English ? Northern Roglai Dictionary. Dal?
las: SILMP. 5mf.
1939. La plus anienne inscription en langue cham. New Indian
Coed?s, Georges.
Antiquary. Extra Series I. Bombay. No. 48. Pp. 39-52.
Ed. Walter F. Vella.
1968. The Indianized states of Southeast Asia.
Trans. Susan Brown Cowing. Kuala Lumpur: U. of Malaya Press.
I.Vaughn. 1975. The Austro-Asiatic
substratum inAcehnese. Ph.D. dis?
sertation, University of California, Berkeley.
1969. The position of Atjehnese
among Southeast Asian languages.
Mon-Khmer Studies III, Saigon. Pp. 48-60.
Collins,
-.
James T. 1991. Chamic, Malay and Acehnese:

theMalay world and the
et
In
Le
Le
Monde
Malais.
Actes de la Confer
Camp?
Malayic
languages.
?rence Internationale sur le Campa et leMonde Malais. Kuala Lumpur: Nur
Niaga Sdn. Bhd.
Collins,
REFERENCES 383
Coope, A. E.
1986. Macmillan's
dictionary.
English?Malay
Malay?English
Singapore: Macmillan.
1963. Katu-English-Vietnamese
Costello, Nancy and Judy Wallace.
Dictionary.
87pp.
-and
1976. Katu linguistic articles. 10pp., 24pp., and 8pp.
Judy Wallace.
-and
Kitrunh. 1976. Katu ethnographic texts. 40pp.
-et
-.
-.
al. 1976. Katu notes and word lists. 72pp.

1967. Some areal features of M?ntu Land Dayak. Oceanic
Court, Christopher.
4.1:46-50.
Linguistics
1970. Nasal harmony and some Indonesian sound laws. Pacific Linguis?
tics, Vol. 13: Pacific linguistic studies in honour of Arthur Capell. Linguis?
tic Circle of Canberra, 203-217.
Cowan, H. K. J. 1933. Het Atj?hsch metrum 'sandja" in verband met een Tjam
sch gedieht. Bijdragen tot de Taal-, Land- en Volkenkunde van Nederland
sch-Indi? 90:149-55.
1948. Aanteekeningen
betreffende de verhouding van het Atjesch tot de
Mon-Khmer-talen.
-.
-.
-.
-.
proto-Austronesian.
-.
Bijdragen
tot de Taal-,
Land-
en Volkenkunde
van Neder
landsch-Indi? 104: 429-514.

1974. Evidence of long vowels in early Achehnese. Oceanic Linguistics
XIII. 187-211.
1981. An outline of Achehnese
phonology and morphology. Bulletin of
the School of Oriental and African Studies, LXIV 3:552-49.
1982. The Achehnese metre sanja' and the Thai kloin ps:t\
Journal of
the Royal Asiatic Society of Great Britain and Ireland 2:
156-60.
1983. The Achehnese
diphthong uia and its possible implications for
Acia
Orientalia
44:153-85.
1988. Achehnese
dialects in connection with Chamic migrations. Paper
at
the
fifth
International Conference on Austronesian Linguistics.
presented
ms.
-.
1991. Acehnese dialects in connection with Chamic migrations. VICAL

2: Western Austronesian and contact languages: papers thefifth congress on
Austronesian
Part I: Edited by Ray Harlow. Auckland, New
linguistics.
Zealand. Linguistics Society of New Zealand, vi, 478 p.
-.
element in Chamo-Achehic: Mon-Khmer

n.d. The Mon-Khmer-related
adstratum or direct Austro-Asiatic
heritage, ms.
Crawfurd, John. 1822 [1967]. Journal of an embassy to the courts ofSiam and
Cochin China [reprinted]; with an introduction by David K. Wyatt. Kuala
Lumpur, London, New York: Oxford University Press.
1852. Grammar and dictionary of theMalay language; with a prelimi?
-.
Volume I: Dissertation
nary dissertation.
Elder, and Co., 65, Cornhill.
1976. Proto-Austronesian.
Dahl, Otto Christian.
and Grammar. London:
Smith,
Institute of of
Scandinavian
15. Sweden: Studentlitteratur.
Studies Monograph
Daud, Bukhari and Mark Durie. n.d. Thesaurus of Acehnese. computer disk.
Asian
384 References
-.
voice qual?
Denning, Keith. 1989. The diachronic development of phonological
to
with
reference
Dinka
and
the
other
Nilotic
ity,
languages. Ph. D.
special
Stanford
dissertation,
University.
initial palatals and "substratumized" Austro
Diffloth, G?rard. 1977. Mon-Khmer
Thai. Mon-Khmer Studies VI. Edited by Philip N. Jenner. Pp. 39-58.
1990. Vietnamese
tonogenesis and new data on the registers of Thau
Sino-Tibetan
vang.
Djajadingrat,
Martinus
-.
Conference.
1934. Atj?sch-Nederlansch
Aceh.
Dordrecht:
1990a.
Chamic.
Foris.
Proto-Chamic
Bulletin
and Acehnese
of the School
53.1:100-114.
1990b. A proto-Aceh-Chamic
-.
1996.
Framing
the Acehnese
mid
of Oriental
vowels:
towards
and African
ety
-.
-.
proto-Aceh
Studies
(London)
database. HyperCard program.

text:
Language
choice
and
tures inAceh. Oceanic Linguistics

35.1:113-137.
Dyen, Isidore. 1953. The Proto-Malayo-Polynesian
laryngeals.
-.
'sGravenhage:
Woordenboek.
Nijhoff.
Donegan, Patricia J. 1985. On the natural phonology of vowels. Garland Publish?
ing, Inc. New York & London.
1993. Rhythm and vocalic drift inMunda and Mon-Khmer. Linguistics
the
Tibeto-Burman A rea 16.1:1 -43.
of
W.
G.
J. 1979. Hikajat Potjut Muhamat. An Achehnese epic. Edited and
Drewes,
translated by G. W. J. Drewes. The Hague: Martinus Nijhoff.
Durie, Mark. 1985. A grammar of Achehnese on the basis of a dialect of North
-.
-.
Hoesein.
discourse
Linguistic
struc?
Soci?
of America.
1965. A lexicostatistical
classification
of the Austronesian
languages.
Indiana
in Anthropology
and Lin?
Publications
Bloomington:
University
to
Memoir
the
Joural
American
International
19; supplement
of
guistics,
Linguistics.
1971a. The Chamic languages. In Thomas A. Sebeok (ed.), Current
Trends in Linguistics. Mouton: The Hague: Volume 8.1:200-210.
1971b. The Austronesian
languages of Formosa. In Thomas A. Sebeok
(ed.), Current Trends in Linguistics. Mouton: The Hague: Volume 8.1:171
199.
-.
Austronesian
and inheritance in Austronesianistics.
1995. Borrowing
Studies Relating to Taiwan, edited by Paul Jen-kuei Li, Cheng-hwa Tsang,
Ying-kuei Huang, Dah-an Ho, and Chiu-yu Tseng. Symposium Series of the
Institute of History and Philology, Acad?mica Sinica, No. 3. Pp. 455-519.
1993. Western Cham as a regis?
Edmondson, Jerold A. and Kenneth J.Gregerson.
ter language. Tonality inAustronesian Languages. Edited by Jerry Edmond?
son and Ken Gregerson. Oceanic Linguistics
Special Publication No. 24.
Honolulu, Hawaii: University of Hawaii Press. Pp. 61-74.
REFERENCES 385
Efimov, Aleksandr Yu. 1987. Historical
Ph. D. thesis. Moscow: Academy
phonology of South Bahnaric languages.

of Sciences of USSR, Oriental Institute.
240 pp. [Wordlist translated byWilliam Gage].
Egerod, S0ren. 1978. An English-Rade
vocabulary. The Museum of Far Eastern
Antiquities, Stockholm. Bulletin 50:49-104.
1969. Languages
of aboriginal Taiwan. Taiwan aboriginal
Ferrell, Raleigh.
Institute of Eth?
groups: problems in cultural and linguistic classification.
Academia
Sinica.
No.
17.
63-75.
nology,
Monograph
Pp.
1977. Register inWestern Cham phonology.
Friberg, Timothy and Kvoeu-Hor.
in
East
Asian
South
Papers
Linguistics No. 4: Chamic Studies. Edited by
David Thomas, Ernest W. Lee, and Nguyen Dang Liem. Pacific Linguistics
Series A, No. 48:17-38.
1977. Chru phonemes.
Fuller, Eugene.
No. 4: Chamic Studies.
Edited
Papers in South East Asian Linguistics

by David Thomas, Ernest W. Lee, and
Liem.
Pacific
Nguyen Dang
Linguistics Series A, No. 48:77-86.
1977.
Hella.
Haroi
clauses.
Goschnick,
Papers in South East Asian Linguistics
No. 4: Chamic Studies.
Edited by David Thomas, Ernest W. Lee, and
Nguyen Dang Liem. Pacific Linguistics Series A, No. 48:105-124.
-and
-.
Alice Tegenfeldt-Mundhenk.
1976. Haroi dictionary. SIL. Hunting
ton Beach, Calif. 325pp. ms.
of the Malayo-Polynesian
1500 BC - 500
Grace, George W. 1964. Movements
the
AD;
5:361-368, 403-404.
linguistic evidence. Current Anthropology
1985. Oceanic
subgrouping: retrospect and prospect. In Austronesian
edited by Andrew Pawley
Linguistics at the 15th Pacific Science Congress,
and Lois Carrington. Pacific Linguistics C-88:l-18.
Greenberg,
Joseph
H.
1970.
Some
generalizations
concerning
nants, especially implosives. UAL 36:25-45.

Grimes, Barbara F. (ed.). 1988. Ethnologue:
Languages
Texas: Summer Institute of Linguistics,
Inc.
Gregerson,
-.
Kenneth
J.
1976.
Tongue
root
and
register
glottalic
of the world.
in Mon-Khmer.
conso?
Dallas,
Austroasi?
atic Studies I, edited by Philip N. Jenner, Laurence C. Thompson, and Stan?

ley Starosta. Oceanic Linguistics
special publication no. 13. Honolulu,
Hawaii. Pp. 323-70.
Gregerson, Marilyn and Dorothy Thomas. 1980. Notes from Indochina On Ethnic
Minority Cultures. Dallas, Texas: SIL Museum of Anthropology.
1955. A history of South-East Asia. First edition.
Hall, Daniel George Edward.
New York: St. Martin's Press.
1981. A history of South-East Asia. Fourth edition. New York: St. Mar?
tin's Press.
in early Southeast
Hall, Kenneth R. 1985. Maritme trade and state development
Asia. Honolu: University of Hawaii Press.
1992. Eastern Cham
Han, Phu Van, Jerold Edmondson, and Kenneth Gregerson.
as a tone language. Mon-Khmer Studies 20:31-43.
386 References
-.
-.
1950. Les consonnes pr?glottalis?es en Indochine. Bulle?

Haudricourt, Andr?-G.
tin de la Soci?t? Linguistique de Partis 46:172-182.
1954. De l'origine des tons vietnamien. Journal Asiatique 242:69-82.
le
1956. De la restitution des initiales dans les langues monosyllabiques:
commun.
la
Soi?t?
de
in
Bulletin
de
du
thai
Originally published
probl?me
Linguistiqe
-.
de Paris
52:307-322. Reprinted
1972: 235-51.
Paris,
diachronique.
1965. Problems of Austronesian
comparative
in Probl?mes
philology.
de phonologie
Lingua
14:315
329.
-.
in the northeast. Stud?

1966. The limits and connections of Austroasiatic
in
Austroasiatic
The
ies
Linguistics.
Hague: Mouton & Co.
Comparative
44-56.
Pp.
-.
-.
-.
-.
129:7-9.
? tons. Ma?tre Phon?tique
1968. Deux langues m?lan?siennes
1971. New Caledonia and the Loyalty Islands. In Thomas A. Sebeok,
in Oceania.
editor, Current Trends in Linguistics, volume 8: Linguistics
Pp.
359-396.
1984. Tones of some languages in Hainan. Minzu Yuwen 4:17-25. Also
published in Bulletin de la Soci?t? de Linguistique de Paris as "La tonolo
gie des langues de Hai-nan." 79.1:385-394.
Headley, Robert K. 1976. Some sources of Chamic vocabulary. Austroasiatic
Studies, Parti. Oceanic Linguistics Special Publication No. 13:453-476.
Lan?
of Kompong Thorn Cham. Austroasiatic
1991. The phonology
guages: Essays in honour ofH. L. Shorto. Edited by Jeremy H. C. S. David?
son. University
of London, School of Oriental and African Studies: 105
121.
Henderson, Eug?nie J. A. 1952. The main features of Cambodian pronunciation.
14.1:149
Bulletin of the School of Oriental and African Studies (London)
74.
-.
-.
character?
1965. The topography of certain phonetic and morphological
South
East
15:400-434.
istics of
Asian languages. Lingua
1967. Grammar and tone in South-East Asian languages. Wissenschaftli?
und
Gesellschaftsder Karl-Marx-Universit?t-Leipzig.
che Zeitschrift
Reihe. Part 1/2:171-178.
Sprachwissen-schaftliche
1982. Sons of the mountains:
ethnohistory of the Viet?
Hickey, Gerald Cannon.
namese Central Highlands
to 1954. New Haven: Yale University Press.
Hoang, Thi Chau. 1987. He thong thanh dieu tieng Cham va each ki hieu [The
InNgon-Ngu
[Language]
system of tones in Cham and their symbolization].
the Social Sciences Committee of Viet?
1-2:31-35. Institute of Linguistics,
nam.
Hudson, Alfred B. 1967. The Barito dialects

comparative
reconstruction
and
of Borneo;
lexicostatistics.
a classification
Ithaca,
N.Y,
based on
Department
Asian
of
Studies, Cornell University.

lan?
Huffman, Franklin E. 1976. The register problem in fifteen Mon-Khmer
Studies, edited by Philip N. Jenner, Laurence C.
guages. In Austroasiatic
REFERENCES 387
-.
Thompson, and Stanley Starosta. Special publication of Oceanic Linguistics

no. 13. 2 volumes. Honolulu, Hawaii: University of Hawaii Press, 1:575-90.
1986. Bibliography and index of mainland Southeast Asian Languages
1986. Yale University Press.
and Linguistics.
The
Brian.
1983.
synchrony and diachrony of the Balkan infinitive; A
Joseph,
in
areal,
Cambridge University
general, and historical linguistics.
study
Press.
Jrang, Ja Kuang, JaWi, Ja Dai, Ja Ngai, and Eugene Fuller. 1977. Chru Vocabu?
lary. Summer Institute of Linguistics. Manila.
Keyes, Charles F. 1995 [1977]. The golden peninsula: Culture and adaptation in
mainland Southeast Asia. Honolulu: University of Hawai'i Press.
Kvoeu-Hor and Timothy Friberg. 1978. Western Cham Vocabulary. SIL.
to linguistic analysis. Chicago and Lon?
Ladefoged, Peter. 1971. Preliminaries
-and
don: University of Chicago Press.

IanMaddieson.
1996. The Sounds of the World's languages. Cam?
Blackwell
Publishers.
Massachusetts:
bridge
1968.
Pierre-Bernard.
Lafont,
Lexique Jarai: parler de la province de plei ku.
Volume LXIII. Paris.
de L'?cole Fran?aise D'Extr?me-Orient.
1987a. Aper?u sur les relations entre le Campa et l'Asie du Sud-Est. Pp.
71-82. Actes du S?minaire sur le Campa organise a l'Universit? de Copen?
of the Seminar
hague, Le 23 Mai 1987. English translation in Proceedings
on Champa
Rancho
Dinh
Cordova, CA. Pp. 65
Te, trans.,
(1994), Huynh
Publications
-.
75.
-.
1987b. Research on Champa and its evolution. Proceedings of the Semi?

nar on Champa, Huynh Dinh Te, trans., Rancho Cordova, CA. Pp. 1-20.
-.
1994a.
Copenhagen
Proceedings
on May
of
the seminar
23, 1987.
on
Champa:
held
at
the University
Huynh Dinh Te, trans., Rancho
of
Cordova,
CA.
-.
1994b. Research on Champa and its evolution. Proceedings of the Semi?

nar on Champa, Huynh Dinh Te, trans., Rancho Cordova, CA. Pp. 1-20.
Laver, John. 1994. Principles of Phonetics.
Cambridge: Cambridge University
Press.
-.
1964. Ethnic Groups of

LeBar, Frank M., Gerald C. Hickey, John K. Musgrave.
Mainland Southeast Asia. New Haven: Human Relations Area Files Press.
1966. Proto-Chamic phonologic word and vocabulary. Ph.D.
Lee, Ernest Wilson.
dissertation. Indiana University. 67-3690.
strata of the
areal features in Austronesian
1974. Southeast Asian
13.1-2:643-668.
languages. Oceanic Linguistics
East Asian Linguistics No. 4:
in
South
1977a. Introduction.
Papers
Edited by David Thomas, Ernest W. Lee, and Nguyen
Chamic Studies.
Dang Liem. Pacific Linguistics Series A, No. 48:vi-vii.
1977b. Devoicing,
aspiration, and vowel split in Haroi: evidence for
register (contrastive tongue-root position). Papers in South East Asian Lin
Chamic
-.
-.
388 References
guistics No. 4: Chamic Studies. Edited by David Thomas, Ernest W. Lee,
and Nguyen Dang Liem. Pacific Linguistics Series A, No. 48:87-104.
1996. Bipartite negatives in Chamic. Mon-Khmer Studies 26:291-317.
1998. The contribution of Cat Gia Roglai to Chamic. Papers in South?
east Asian Linguistics No. 15: Further Chamic Studies. Edited by David
Thomas.
Canberra: The Australian
Pacific Linguistics.
Series A-89.
-.
-.
National University. Pp. 31-54.

1960. Moken texts and word-list, a provisional interpre?
Lewis, Martha Blanche.
tation. Federation Museums Journal 4 N.S. (102 pages).
Li, Ce Jin and Tian Guang. 1986. Hainan Dao Yanglan Huizu De Li Shi Ji Xian
Zhuang Diao Zha [The history and current investigation of the Hui national?
ity in Yanglan Hui nationality on Hainan Island]. Minzu Yan Jiu Dong Tai
[Current Nationality Research] 4:26-31.
1977. A handbook of comparative
Li, Fangkuei.
Press of Hawaii.
Li,
Paul
Jen-kuei.
1995.
Formosan
vs.
non-Formosan
Tai. Honolulu:
features
The University
in some
Austrone?
sian languages in Taiwan. Austronesian Studies Relating to Taiwan, edited

by Paul Jen-kuei Li, Cheng-hwa Tsang, Ying-kuei Huang, Dah-an Ho, and
Chiu-yu Tseng. Symposium Series of the Institute of History and Philology,
Acad?mica
Sinica, No. 3. Pp. 651-681.
Ian and Keng-Fong Pang. 1993. Tone
Maddieson,
sian
Edited
Languages.
Linguistics
Hawaii
Maitre, H.
Majumdar,
Press.
Special
Pp.
by
Jerry
Publication
Edmondson
No.
inAustrone?
inUtsat. Tonality
and Ken
24. Honolulu,
Gregerson.
Hawaii:
Oceanic
University
of
75-89.
1912. Les jungles M?i. Paris: Larose.

R. C. 1927. Ancient Indian Colonies
in the Far East,
I: Champa.
Lahore.
Phoeun. 1987. La communaut? cam au Cambodge du XVo au XIXo si?cle.

Actes du S?minaire sur le Campa organise a l'Universit? de Copenhague,
Le 23 Mai 1987. p. 78: English translation in Proceedings
of the Seminar
on Champa, Huynh Dinh Te, trans., Rancho Cordova, CA, 1994. Pp. 76-86.
D. G. (ed.). 1949. Selected writings of Edward Sapir. Berkeley:
Mandelbaum,
Mak,
-.
University of California Press.

1979. Etudes cam. II. L'introduction
de l'Islam au
Pierre-Yves.
Manguin,
Fran?aise
66:255-87.
d'Extr?me-Orient
Bulletin
de
l'Ecole
Champa,
1985. The Introduction of Islam to Champa. Journal of theMalaysian
58.1:1-28. Translation of Manguin
of the Royal Asiatic Society
(1979) by Robert Nicholl.
Manley, Timothy M. 1972. Outline ofSre structure. Oceanic Linguistics Special
Publication No. 12. Honolulu: University of Hawaii Press.
Marrison, Geoffrey E. 1975. The early Cham language and its relationship to
branch of the Royal Asiatic Society
Malay. Journal of the Malaysian
Branch
48.2:52-59.
REFERENCES 389
-.
-.
-.
1985. The Chams and their Literature. Journal of theMalaysian Branch

the
of
Royal Asiatic Society 58.2:45-70.
1928. Le Royaume de Champa. Paris and Brussels.
Maspero, Georges.
1949. The kingdom of Champa: a translation of chapter I of Le Royaume
de Champa. New Haven, Conn.: Yale University. Southeast Asian Studies.
in Southeast Asia. In Consonant Types &
James A. 1973. Tonogenesis
Matisoff,
Tone. Edited by Larry M. Hyman. Southern California Occasional Papers in
Linguistics No. 1. Los Angeles: USC. Pp. 71-96.
the mysterious
connection between nasality
1985. Rhinoglottophilia:
and glottality. InNasalfest,
edited by C. A. Ferguson, L. M. Hyman, and J.
J. Ohala. Stanford University Press. Pp. 265-287.
McNair, J. F. 1972 [1878]. Perak and theMalays. Kuala Lumpur: Oxford Uni?
versity Press.
Medcalf, Anne-Marie.
thesis. Murdoch
-.
-.
University.
Fr. G?rard.
Moussay,
-.
1989. J?rai sentence and phrase structure. BA. Honours

1971.
Talen
van
Indon?sie.
Bernd.
Nothofer,
Phan
Rang:
tot de Taal-,
Bijdragen
en Volkenkunde
Land-
van Ned
1975.
The
Reconstruction
of Proto-Malayo-Javanic.
Verhan?
delingen van Het Koninklijk Instituut Voor Taal-, Land-, en Volkenkunde.

73. The Hague: Martinus Nijhoff.
1985. The subgrouping of the languages of the Javo-Sumatra hesion: a
reconsideration.
landsch-Indi?
Oey, Eric M.
draft],
-.
Cam-Vietnamien-Fran?ais.
40:27-44.
erlandsch-lndi?
-.
Dictionnaire
Centre Culturel Cam.

Dabai.
1988a. The genealogical affiliation of the language of the Hui people
Ni,
in Sanya Hainan. Minzu Yuwen 2:18-25.
1988b. The Kam-Tai of China and the Hui of Hainan. Bulletin of the
Central Institute ofMinorities
3:54-65.
1990a. The origins of the tones of the Kam-Tai languages, ms.
1990b. The Sanya (= Utsat) language of Hainan island: a living speci?
men of a linguistic typological shift, ms.
Niemann, G. K. 1891. Bijdrage tot de Kennis der Verhouding van het Tjam tot de
Bijdragen
tot de
Taal-,
Land-
en Volkenkunde
van Neder
141.2-3:288-302.
n.d. The
linguistic position
of the Chamic
languages,
[preliminary
ms.
1983. The Huihui speech (Tsat) of the Hui

Ouyang, Juey a and Zheng Yiqing.
in
Hainan.
Minzu Yuwen 1:30-40.
Yaxian,
nationality
1980. A brief description of Li (Hainan). Chinese Minority People's
Language, Basic Description Series. Beijing.
1992. The dynamics of gender, ethnicity and state among the
Pang, Keng-Fong.
Muslims
island, People's
(Hui/Utsat) of Hainan
Austronesian-speaking
of
China.
UCLA.
Ph.D.
dissertation.
Republic
390 References
-.
1998. On the ethnonym

'Utsat'. Papers in Southeast Asian Linguistics
No. 15: Further Chamic Studies. Edited by David Thomas. Pacific Linguis?
tics. Series A-89. Canberra: The Australian National University. Pp. 55-60.
Parkin, Robert. 1991. A Guide to Austroasiatic
Speakers and Their Languages.
Oceanic
Hawai'i
Linguistics
Publication
Special
No.
23. Honolulu:
of
University
Press.
Pawley, Andrew and Malcolm Ross. 1995. Prehistory of Oceanic languages: A

current view. In The Austronesians: Historical
and Comparative Perspec?
tives. Edited by Peter Bellwood, James J. Fox and Darrell Tryon. Canberra:
The Australian National University. Pp. 39-74.
Canberra: Pacific Linguistics,
Peiros, Ilia. 1996. Katuic Comparative Dictionary.
Series C-132.
Pham Xu?n Tin. 1955. Lexique Polyglotte. Dran, Vietnam.
Pittman, Richard S. 1959. Jarai as amember of theMalayo-Polynesian
family of
Asian
languages.
Culture
1.4:59-67.
Prachacakij-karacak,
Phraya. 1995. Some Languages of Slam. Translated and
annotated by David Thomas and Sophana Srichampa. Institute of Language
and Culture for Rural Development. Mahidol University.
1966. The German language. Revised. Lon?
Priebsch, R., and W. E. Collinson.
don: Faber and Faber.
11
Ravaisse, P. 1922. Deux inscriptions coufiques du Campa. Journal Asiatique
serie, no. 20, (1922) pp. 247-59.
1993. Islamization and Christianization
in Southeast Asia: The
Reid, Anthony.
critical phase, 1550-1650. InAnthony Reid (ed.), Southeast Asia in the early
modern
-.
era.
Ithaca:
Cornell
University
Press.
Pp.
151-179.
1995. Continuity and change in the Austronesian

transition to Islam and
James J. Fox, and Darrell Tryon (eds.), The
Christianity. In Peter Bellwood,
Austronesians:
Historical
and
Comparative
Perspectives.
Canberra:
The
Australian National University. Pp. 314-322.

evidence for Austric. Oceanic Linguis?
Reid, Laurence A. 1994. Morphological
tics 33-2:323-344.
1988. An Iban-English dictionary. Petaling Jaya: Oxford Uni?
Richards, Anthony.
versity Press, Penerbit Fajar Bakti Sdn. Bhd.
A hunter-gatherer
Rischel, J0rgen. 1995. Minor Mlabri:
language of northern
Indochina. University of Copenhagen: Museum Tusculanum Press.
-.
-.
Ross, Malcolm D. 1992. The sound of Proto-Austronesian: An outsider's view of

the Formosan evidence. Oceanic Linguistics
31.1:23-64.
1995. Reconstructing
Proto-Austronesian
the evi?
verbal morphology:
dence from Taiwan. Austronesian Studies Relating to Taiwan, edited by Paul
Jen-kuei Li, Cheng-hwa Tsang, Ying-kuei Huang, Dah-an Ho, and Chiu-yu
Series of the Institute of History and Philology, Aca?
Tseng. Symposium
d?mica Sinica, No. 3. Pp. 727-791.
1995b.
Linguistics
Proto-Oceanic
terms
for meteorological
phenomena.
Oceanic
34.2:261-304.
REFERENCES 391
on Judeo-German
phonology. The Jewish quarterly

in
Mandelbaum
1949:252-272.
Reprinted
Schaf er, Edward H. 1967. The vermilion bird; Tang images of the south. Berke?
ley: University of California Press.
ein Bindeglied zwischen.
1906. Die Mon-Khmer-V?lker,
Schmidt, Pater Wilhelm.
Sapir, Edward. 1915. Notes
review, n.s. 6:231-266.
V?lkern
schw)
-.
Zentralasiens
n.s.,
und Austronesiens.
of the Austroasiatic
(Language
Data,
Asian-Pacific
series,
no.
109p.
1975. The velar animal prefix relic inVietnam
the Tibeto-Burman
-.
(Braun
language family.
18:206-17.
Language
Bulletin of the
Shorto, Harry L. 1975. Ach?nese and Mainland Austronesian.
School of Oriental and African Studies 38:81-102. University of London.
1977. Proto-Austronesian
*taqan: An anomaly removed. Bulletin of the
School of Oriental and African Studies 40:128-129. University of London.
reconstruction of Proto-North-Bah?
Smith, Kenneth D. 1972. A phonological
naric.
-.
Anthropologie
and Marilou
Stockton, Jr., Elaine M. Murphy,
Schrock, Joann L., William
Fromme. 1966. Minority Groups in the Republic of Vietnam. Ethnographic
Study Series. Department of theArmy.
London: University of London,
Scott, N. C. 1956. A dictionary of Sea Dayak.
School of Oriental and African Studies.
1964. Nasal consonants in Land Dayak (Bukar-Sadong). In David Aber
crombie et al. (eds.), In honour of Daniel Jones. London: Longmans, Green
and Co. Pp. 432-436.
Sebeok, Thomas A. 1942. An examination
-.
Archiv
5:59-109.
Area
2).
Santa
Ana,
CA,
languages. Linguistics
SIL.
of
2.1:1-18.
contact in Ancient South China: the case of

1982. Linguistic
Solnit, David.
8:219
Hainan Chinese, Be, and Vietnamese. Berkeley Linguistics Society
230.
1993. Glottalized consonants as a genetic feature in Southeast Asia. Acta
Ling?istica Hafniensa 25:95-123.
1972. On the natural history of diphthongs. Papers from the
Stampe, David.
Eighth Regional meeting of the Chicago Linguistic Society, pp. 443-454.
Starosta, Stanley. 1995. A grammatical
languages.
subgrouping of Formosan
Austronesian Studies Relating to Taiwan, edited by Paul Jen-kuei Li, Cheng
hwa Tsang, Ying-kuei Huang, Dah-an Ho, and Chiu-yu Tseng. Symposium
Series of the Institute of History and Philology, Acad?mica
Sinica, No. 3.
-,
Pp. 683-726.
Andrew K. Pawley, and Lawrence A. Reid. 1982. The evolution of focus
in Austronesian.
In Amran Halim, Lois Carrington, and S. A. Wurm, eds.,
Papers from the Third Interational Conference on Austronesian Linguis?
tics, vol. 2. Tracking the travellers, 145-170. Pacific Linguistics C-75. Can?
berra: Australian
National University.
392 References
Strecker, David
Ph.D.
gan
1984. Proto-Tai
Shalom.
person pronouns. University
of Michi?
dissertation.
Tadmor, Uri. 1995. Language contact and systematic restructuring: The Malay
dialect of Nonthaburi, Central Thailand. Ph.D. dissertation, University
of
Hawaii.
(ed.). 1992. The Cambridge History of Southeast Asia. Volume

Tarling, Nicholas
1. From Early Times to c. 1800. Cambridge University Press.
Taylor, Keith W. 1922. The early kingdoms: Champa. Nicholas Tarling (ed.). The
Cambridge History of Southeast Asia. Volume 1. From Early Times to c.
1800. Cambridge University Press. Pp. 153-157.
Alice
and Hella Goschnick.
1977. Haroi
Tegenfeldt-Mundhenk,
in
East
Asian
South
4:
No.
Chamic
Studies.
Papers
Linguistics
David Thomas,
Ernest W. Lee, and Nguyen
phonemes.
Edited by
Dang Liem. Pacific Linguistics
SeriesA,No.48:.l-15.
-.
-.
1980. A Rhade-English
Tharp, J. A. and Y-Bham Buoy-Ya.
dictionary with
-No. 58.
Pacific
Series
C
English-Rhade finderlist.
Linguistics.
Thomas, David D. 1964. Comment appended to Grace (1964).
1971. Chrau grammar. Oceanic Linguistics Special Publication No. 7.
Honolulu: University of Hawaii Press.
1979. The place of Alak, Tempuan, and West Bahnaric. Mon-Khmer
Studies VIIL171-186.
Papers in Southeast Asian Linguistics No. 15: Further Chamic
Studies.
Pacific Linguistics.
Series A-89.
Ernest Wilson Lee, and Dang Liem Nguyen.
1977. Papers in South East
Asian Linguistics. No. 4. Chamic Studies. Canberra: Pacific Linguistics.
Series A, No. 48.
reflexes in Rade, Jarai, and
Thomas, Dorothy M. 1963. Proto-Malayo-Polynesian
Chru. Studies in Linguistics
17:59-75.
1967. A phonological
reconstruction of Proto-East-Katuic. MA thesis,
of
North
Dakota.
Dallas: Summer Institute of Linguistics, micro?
University
fiche.
-(ed.).
-,
-.
Sarah
Thomason,
Grey
and
and genetic
olization,
Terrence
linguistics.
Kaufman.
1988.
The University
Language
cre
contact,
Press: Ber?
of California
keley.
Laurence C. 1984-1985. A Vietnamese reference grammar.

Stephan O'Harrow. Mon-Khmer Studies XIII-XIV.
Thompson,
Graham.
1980. Consonants, phonation types, and pitch height. Com?

13:207-19.
putational Analyses of African and Asian Languages
1982. A comparative note on the Indian linguistic area. South Asian
Thurgood,
-.
Review
-.
Edited by
6.3:23-9.
1988a.
Linguistic
Notes
on
the
reconstruction
of Kam-Sui.
Comparative
studies beyond Tai. Edited by Jerold A. Edmondson
Kadai:
and David B.
REFERENCES 393
Summer Institute of Linguistics
in Linguistics No.
Arlington Publications
Solnit.
-.
k-
1988b.
in Kam-Sui
prefixes
and Kadai:
some
notes.
at
of Texas
and The University

86:179-218.
and
Languages
in East Asia: Festschrift for Tatsuo Nishida on the Occasion of his

History
60th Birthday. Edited by Paul K. Eguchi, Yukio Fujimoto, Nobuyoshi Fuku
Atsuchi
Koichi Miyamoto,
Iwamoto, Tatsuo
hara, Masura Hashimoto,
David Sell,
Akihiro
Masaoki
Osamu
Sato,
Kondo,
Sakiyama,
Miyamoto,
Norio Shibata, Ken-ichiro Shirai, Mashiro Shogaito, Shiro Yabu and Kazu
-.
-.
-.
-.
-.
hiko Yoshida. Kyoto:

1991. Proto-Hlai
Shokado. Pp. 229-235.

(Li): A look at the initials, tones, and finals. Kadai:
inKadai and SE Asian Linguistics
3:1-49.
Discussions
1992a. The aberrancy of the Jiamao dialect of Hlai: speculation on its
origins and history. Southeast Asian Linguistics Society I, Edited by Martha
Ratliff and Eric Schiller. Tempe: Arizona State University Southeast Asian
Studies Publications. Pp. 417-433.
1992b. From atonal to tonal in Utsat (a Chamic language of Hainan).
Proceedings
of the Eighteenth Annual Meeting of the Berkeley Linguistics
Society. February 14-17, 1992. Special Session on the Typology of Tone
Jonathan Evans, David
Languages. Edited by Laura A. Buszard-Welcher,
Peterson, Lionel Wee, andWilliam Weigel. Pp. 145-156.
1993. Phan Rang Cham and Utsat: tonogenetic themes and variants.
Edited by Jerry Edmondson and Ken
Tonality in Austronesian Languages.
Oceanic Linguistics
Gregerson.
Special Publication No. 24. Honolulu,
Hawaii: University of Hawaii Press. Pp. 91-106.
1994. Tai-Kadai and Austronesian: The nature of the historical
ship.
-.
-.
relation?
33.2:345-368.
Linguistics
of. The Asian American

1995. Southeast Asian languages?prehistory
Encyclopedia^, vol. 5. New York: Marshall Cavendish. Pp. 1418-1421.
contact and the directionality
of internal 'drift': the
1996. Language
tones
in
Chamic.
of
and
Language 71.1:1-31.
development
registers
-.
-and
Oceanic
1998.
The
Austronesian
and
Mon-Khmer
components
in
the
proto
Chamic vowel system. Papers in Southeast Asian Linguistics No. 15: Fur?
ther Chamic Studies. Edited by David Thomas. Pacific Linguistics. Series
A-89. Canberra: The Australian National University. Pp. 61-90.
Hector Javkin. 1975. An acoustic explanation of a sound change: *
ap
to -o,
*-at
to -e, and
*-ak
to -ae. Journal
1980. The Tan-chou

Ting, Pang-hsin.
Asie Orientale
8:5-27.
of Phonetics
dialect of Hainan.
3.3:161-5.
Cahier de linguistique
Tran, Ky Phuong. 1987. Museum of Cham sculpture inDa Nang. Hanoi: Foreign
Languages Publishing House.
and the major Austronesian
subgroups.
Tryon, Darrell. 1995. Proto-Austronesian
In Peter Bell wood, James J. Fox, and Darrell Tryon (eds.), The Austrone?
sians: Historical
and Comparative Perspectives.
Pp. 17-38.
394 References
-.
Tsuchida, Shigeru. 1982. A comparative vocabulary of Austronesian

languages
of Sinicized ethnic groups in Taiwan, Part I: West Taiwan. Tokyo, Memoirs
of the Faculty of Letters, University of Tokyo, No. 7.
1995. Alienable
and unalienable distinction in Puyuma? Austronesian
Studies Relating to Taiwan, edited by Paul Jen-kuei Li, Cheng-hwa Tsang,
Ying-kuei Huang, Dah-an Ho, and Chiu-yu Tseng. Symposium Series of the
Institute of History and Philology, Acad?mica Sinica, No. 3. Pp. 793-804.
Watson, Sandra K. 1966. Verbal affixation in Pacoh. Mon-Khmer Studies 11:15
30.
Weinreich, Uriel, William Labov, and Marvin

tions for a theory of language change.
Malkiel
(eds.), Directions for Historical
Texas
Wolff,
-.
-.
Press.
I.Herzog.
InWinfred
Linguistics.
1968. Empirical founda?

P. Lehman and Yakov
Austin: University
of
95-195.
Pp.
John U. 1973. Verbal inflection in Proto-Austronesian.

InAndrew Gonza
lea, ed. Parangal kay Cecilio Lopez, 71-91. Quezon City: Linguistic Soci?
ety of the Philippines.
1991. The Proto-Austronesian
phoneme *t and the grouping of the Aus?
tronesian languages. In Currents in Pacific linguistics: Papers on Austrone?
sian languages and ethnolinguistics
in honour of George W Grace,
edited
by Robert Blust, pp. 535-549. Pacific Linguistics C-l 17.
1995. The position of the Austronesian
languages of Taiwan within the
Austronesian
Paul
group.
Jen-kuei
Li,
Chiu-yu Tseng.
Acad?mica
Austronesian
Cheng-hwa
Symposium
Sinica,
No.
3. Pp.
Tsang,
Studies Relating
Ying-kuei
to Taiwan,
Huang,
edited by
Dah-an
Series of the Institute of History
Ho,
and
and Philology,
521-583.
Wyatt, David K. 1967. Introduction. Journal of an embassy to the courts ofSiam

and Cochin China [reprinted], by John Crawfurd. Kuala Lumpur, London,
New York: Oxford University Press. 4pp.
Summer Institute of Linguistics.
Y-Chang, Ni? Si?ng. 1979. Rade vocabulary.
Huntington
Beach, California.
1986. A further discussion
of the position of Huihui speech and

relationship. Minzu Yuwen 6:37-43.
Zorc, R. David. 1978. Proto-Philippine word accent: Innovation or Proto-Hesper
onesian retention? In Second International Conference on Austronesian Lin?
Zheng, Yiqing.
its genetic
-.
edited by S. A. Wurm and Lois Carrington,

guistics: Proceedings,
119. Pacific Linguistics C-61.
accent revisited. Philippine Journal
1983. Proto-Austronesian
14.1:1-24.
guistics
pp. 67
of Lin?
Author
Index
A
Abdul Hamid Mahmood, 25
Abdul Rahman al-Ahmadi, 22, 24
Adelaar, 8,18,34,39,
239, 308
Antilla, 8
Chappell, 225
Ciochon
and James,
Cobbey, 44
Coed?s, 3, 20, 83, 229
81,82,129,147,
Collins,
I. V.,
48
Collins, J.T. 34, 36, 49, 50, 52, 139,
140
Awoi-hathe, 110,271,272
Aymonier, 30, 77, 78, 109, 113, 116,
144, 145, 147, 228, 229, 237, 239,
241, 242, 244, 245, 249, 268, 269,
275, 342
Collinson, 116
Coope, 147, 150,245,278
Court, 156, 165, 172
102, 115, 138, 139, 140,145,

150, 237, 238, 244, 257, 258
Crawfurd, 30, 31,40, 46, 237
Cowan,
Banker, 240, 242, 245

Bastian,
8,48,49,50,52,54,55,56,94,
146,
40
Baxter, 278, 322, 347

Bellwood, 15, 18,32
Benedict, 151, 178, 214, 215, 218, 227
Benjamin, 308
Dahl, 244
Dempwolff, 114, 115,243
Diffloth, 10, 24, 58, 64, 140, 239
Donegan, 61, 115, 116, 117, 126
Blagden, 41,48
Blood, David, 9, 187, 193, 194, 279
Blood, Doris, 7, 39, 60, 62, 63, 88, 105,
187, 190,279
Blood, H., 10, 11, 12, 105
Blust, 8, 12, 17, 25, 32, 33, 34, 35, 36,
37, 39,40, 41,48, 57, 58, 81, 84, 85,
86,87,90,91,
100, 141, 146, 147,
172, 239, 240, 243, 245, 246, 303,
308, 360
Bui Kh?nh The, 268, 275
Drewes,
55
Durie, xi, 7, 8, 23, 48, 49, 51, 56, 58,

63, 68, 70, 77, 85, 88, 94, 98, 99,
100, 102, 103, 117, 127, 130, 145,
147, 214, 217, 225, 237, 244, 262,
263
Dyen, 7, 25, 31, 37,41,87,
105, 115
E
Edmondson, 106, 178, 181, 182, 189,
190, 193
Efimov, 10, 11, 12, 105, 150, 158
Egerod, 89, 269, 270
Bumham, 7, 10, 11, 45, 105, 178, 204,

205, 206, 265, 266
C
Cabaton, 77, 78, 83, 109, 110, 113,
116, 144, 145, 147, 239, 241, 242,
244, 245, 249, 268, 269, 275, 342
Chambert-Loir, 228, 229
F
Friberg, 45, 106, 178, 181, 183, 184,
185, 186, 187, 199,274,275
Fuller, 84
395
396
G
Gage, 10
Goschnick, 69, 155, 198, 210, 265,
Author
Index
Lewis, 58
Li, Fang Kuei, 90
Li, Paul, 141
266
Grace, 254, 256
Greenberg, 60, 63, 86, 87, 88, 89,
188
Gregerson, 28, 106, 178, 181, 182,
189, 190, 193
Grimes, 44, 271
H
Hall,D.G.E.,
17,
18
Hall,K.R.,
14, 18, 19
Han, 178, 189, 190, 193
Haudricourt, 90, 178, 214, 215
Headley, 10, 45, 61, 105, 119, 145,
150, 186,187,
199,322,347,358
Henderson, 93, 179, 181, 196, 235
Herzog, 255
Hoang, 193
Holle, 40
Hudson, 361
Huffman, 198
J
Javkin, 120
Jr?ng, 263, 264
M
Maddieson, 160, 161, 163, 166, 167,
215,218,223,272,273,274
Maitre, 263
Mak Phoen, 22
Manguin, 228, 229
Manley, 68
Marrison, 3, 34, 247, 249, 303, 342
Maspero, 227, 228
Matisoff, 61, 154
Medcalf, 245
Mistry, 70
Morice, 40
Moussay, 113, 188, 193, 268, 269,
275, 279
Mundhenk, 265, 266
N
Ni, 160, 161, 163, 164, 167, 178,
214,215,218,272,273,274
Niemann, 8, 48, 49, 237
Nothofer, 36, 37
K
Kaufman, 254, 256
Kern, 48
Keyes, 15, 27
Kvoeu-Hor, 45, 106, 178, 181, 183,
184,185,186,187,199,274,275
O
Oey, 4, 218, 258,275
Okrand, 161
Ouyang, 160, 161, 163, 164, 165,
167,214,215,218,230,272,273,
274
L
Labov, 255
Ladefoged, 92, 166, 179
Lafont, 22, 40, 90, 196, 238, 266,
267
Lee, 7,10,11,
13,44,45,57,58,61,
62, 69, 84, 86, 87, 88, 89, 92, 93,
96,100,105,106,110,125,140,
151,155,157,158,159,178,204,
205,206,243,244,265,267,270,
279,313
P
Pang, 160, 215, 218, 223, 227, 230,
272, 273, 274
Pawley, 40, 244
Peiros, 11, 12, 106, 151
Pham Xu?n Tin, 7, 266, 267
Pittman, 31, 105
Prachacakij-karacak,
10,
12,
105
Priebsch, 116
Author
Index
397
R
Ravaisse, 228
Reid, Anthony, 228, 229
Reid, L., 20, 240, 241, 244
Relandus, 31
Rischel, 242
Ross, 34, 35,40,
151,243,244
S
Sapir, 116
Sch?fer, 225, 228, 229
Schmidt, 31,237
Schrock, 210
Scott, 172
Sebeok, 31,237
Shorto, 8, 48, 49, 52, 56, 138, 139,
140, 150, 237, 238, 257, 258
Smith, 10, 12, 105, 131, 144,211,
212,213,243,247,293
Solnit, 4, 47, 87,90,91
Stampe, 116
Starosta, 32, 244
Thomas, Dorothy, 7, 60, 87, 88, 89,

105,106,114,115,279
Thomason, 254, 256
Thompson, 196
Thurgood, 32, 33, 90, 106, 120, 178,
183,185,197,199,205,208,214,
215,222,230
Ting, 230
Tryon, 32
Tsuchida, 31
W
Watson, 240
Weinreich, 255
Wolff, 31, 151,243,244
Y
Y-Bham,
89
Y-Chang, 89, 269, 270
Tadmor, 25
Zheng, 2,21,22,160,161,163,164,
165,167,214,215,218,225,226,
227, 229, 230, 272, 273, 274
Zorc, 82, 151
Tegenfeldt, 155, 198,210

Tharp, 89, 269, 270
Thomas, David, 10, 12, 28, 61, 105,
179, 240, 242
Topic Index
A
Abu Hasan, 228
accommodation
area,
Austronesian
evidence,
archaelogical
earlier mainland
Acehnese
Katuic
loss of the presyllable initials, 72
loans
49
loans
contact,
Malay
migration,
MK place
Thai
42-43,47
23-26
an loanwords
in Acehnese,
dating
Kelantanese
and Kelantan,
25
54
Bac-son,
among
the
Be,
among
the
Borneo
50
consonants,
Shared
24,
Chamic
poetry,
innovations
Shared
innovations
MK
Shared innovations in the

epic
poetry,
J. Drewes,
sanja'
patterns,
poetic
borrowings
Cham inChrau, 9
Haroi inHr?, 10
Haroi inKatu, 10
55
55
tradition,
Jarai
Mischsprache,
in Bahnar,
MK intoPC, 10,307-309
Headley (1976), 10
of Arabic origin, 349
of Indic origin, 346-349
55
Annamite Cordillera, 14
Atayalic, 32
Austroasiatische
258
of vowel length, 257-259
54-56
Hikajat Potjut Muhamat, 55

iambic foot, 55
rhyme
18
substratum,
of grammar,
54,
Cowan (1982), 55
G. W.
borrowability, 238
53
lexicon,
15
22
Binh-dinh,
51
vowels,
32
Formosan,
MP, 32
of Nonthaburi,
and mainland
Acehnese
24
traditions,
Pattani,
Acehnese
32
migration into the Pacific, 32

primary subgroups, 32
24
24
place names,
dialects
the Malayic
of, 31
32
Formosa,
the presence,
and early
33
recognition
homeland,
24,
58
Acehnese
scholars
32
family tree, 31
family tree (Figure 5), 35
Hadrianus Relandus (1708), 31
48-50
34-40,
and Kelantan,
Asli
early
to mainland
Chamic,
in proto-Kam-Sui,
in proto-Tai,
32
Thurgood (1994), 32
MK substratum in, 237

Niemann (1891), 47
Acehnese
32-33
33
area,
loans,
58
names,
relationship
evidence,
Guizhou
58
loans,
32
Crawfurd's survey (1852), 31

dating of sites, 32
to a linguistic
258-259
31
399
348
Portuguese,
in eastern Mnong,
Rade
See
borrowings.
continued
voice
with
of
enlongation
Chamic
voiced
See
83
obstruents,
the vocal
tract,
179
Chamic
of
larynx,
179
Phan Rang Cham low

lower
resultant
15-17
prehistory,
also Sa Huynh
reconstruction
lexical,
190-194
tones,
vowel
Fl,
Thomas, Dorothy (1963), 7
179
Chamic
studies,
Chamic
within
Chamic,
early
dialect
Chamic,
of vowels
centralization
larynx lowering, 179
and
the formants
centralized
tract
vowels,
lengthening,
Haroi
179
179
formants,
larynx
raised,
wave
shortened
of
shortening
180
Chinese
in Kelantanese,
first historical reference (137

AD), 20
180
the name
inscription
Tsat,
Jarai,
zenith
Cham Kur (Khmer), 2

Cham Raglai (Roglai), 2
Cham Ro (Chru), 2
earlier
usage,
History of the Song Dynasty (960

1279), 2
Zhan (inMandarin), 2
at Tra-ki?u,
20
northern capital sacked (982), 20

The History of Chin (c.
280AD), 20
Bahnar Cham (Haroi), 2

Cham
19-22
important sites (Fig. 2), 21
180
Cham
and
21-27
history,
peripheral,
24
20
references,
diaspora,
180
more
14
China in north Vietnam (111

AD), 19
179
lengths,
the vocal
26, 27
in assimilation
26-27
patterns,
and Hr?, 27
Champa,
raising the formants (F2), 179
Cepa
affiliations
the Vietnamese,
changes
voice
higher
vowels
31-39
mainland
(F2), 179
tract,
of, 40
and theMK, 26
179
lengths,
lowering
vocal
continuum,
genetic
Chamic,
wave
lengthening
more
history
Austronesian,
Lafont, 238
voice
breathy
6-13
work,
previous
179
raising,
of vowels
centralization
creaky
26
Blood, Doris (1962), 7

Bumham (1976), 7
Dyen (1971), 7
Lee (1966), 7
Haroi, 84, 197

inChru, 84
inHr?, 212
lowering
of
influence
Vietnamese,
breathy
associated
in Vietnam,
arrival
comparative
methodology
See
Zhancheng,
Chamic
Index
Topic
400
in sixth
century,
20
Chong (MK), 69
Chru, 2, 7, 84
etymology
Chru
presyllables
of, 2
from
morphology?, 69-70
Haroi
classification
counterparts,
of Chamic
69
401
Index
Topic
Austroasiatische
31
Mischsprache,
creaky
of Champa,
Malay
179
higher Fl,
of
the
vocal
tract,
shortening
of vowels
See centralization
resultant
30
Schmidt (1906), 31
Sebeok (1942), 31
vs. genetic
criteria,
typological
93
clusters with
-h-, 85-86,
8
methodology,
comparative
"inverted
reconstructions",
D
15
DaNang,
dialect
chain
57
innovations,
34
continuum,
22-23
diaspora,
1662 missionary
22
account,
Aceh, 22
main
62
inventories,
syllable
See
also
See
also
Cambodia, 22
Guangzhou (Canton), 22
consonants
presyllable
word-final
consonants
contact
with
210-213
Hr?,
195-197
Cham,
Rang
also network
22
Java,
22
Annals,
Khmer Royal Chronicles, 22
on Hainan,
230-232
Malaka,
22
Tsat with theLi, 230
See
Western
SejarahMelayu, 22
186-187
Cham,
convergence
Chamic
22
Hainan (See Tsat), 22

highlands (See N. Roglai), 22
Khmer
224
Tsat,
records,
Guangzhou
contact
Haroi
23
Acehnese,
presyllable inventories, 62
See
72
and Rade,
dialect
consonants
Phan
179
30
9, 13
borrowings,
dialect
34
chains,
lexical
evaluating
shared innovations,
voice
Ladefoged, 179
raising of the larynx, 179
Aymonier (1899), 30
Crawfurd (1822), 30
also Western
Southeast
Asia,
4-5,
True
6
consonant
finals,
merger
diphthongization
117
chains,
of, 6
and
5-6
tendencies,
consonants,
4, 6
glottalized
loss of obstruent
6
voicing,
6
monophthongization,
4
Oey,
phonological
restructuring
Solnit, 4
development
vowel
contrasts,
Germanic
Written
115
116-117
parallels,
Cham
115
records,
proliferation
voicing
70
of voiceless
obstruents,
disyllabic tomonosyllabic, 60-66

clusters
inventories,
vowels,
diphthongization chains, 116

Donegan (1985), 115
dissimilative
of, 4, 6
of, 6
vowel
stressed
dipththongs
reflexes of the PMP
diphthongs, 124-126
systems,
of, 6
tones,
of the Emperor
115-117
diphthongization,
Acehnese
6
monosyllabic,
loss of, 6
finals,
general
Records
Xian
Zong of theMing Dynasty

(1368-1644), 22
inventories,
dysyllabioiambio
22
22
Thailand,
with
Cham,
with
-h-, 64
disyllables with liquids > loss of

the unstressed
initial
disyllables with liquids >
Hainanese
Tan-chou
in informal speech (Acehnese), 63

colloquial
62
paths of change,
loss of the vowel before
etymology of, 2, 210

vowel
pre-Haroi
complex,
197
voice,
breathy
-h-, 63-64
cultural
22
Min),
(Southern
Haroi
internal
medial
230
dialect,
Ting (1980), 230

Hainanese
speech
(Cham), 62
Dong-son
70
with
monosyllables
64-65
clusters,
The
21
Guangzhou,
Gujarati,
65-66
syllable,
informal,
Index
Topic
402
15
changes
PC shwa backed, 200

shared with Western
Cham, 199
E
Eastern
Cham.
See Phan
Cham
vowel
sentence,
equative
Rang
velars,
before
raising
200
and vowel
registers
splitting, 201-206
F
final
*-l
(and
to -n, 176
*-r)
14
Funan,
20
Chinese
19
envoys,
with
Borneo,
18
connection
with
Chamic,
17-19
tense
17
contact,
voice
20
Ch'i
history,
envoy
and
influence,
192-194
consonants
Gong Cepa, 24
abrupt
4
terminology,
of the Five Dynasties
History
obstruents
versus
Coastal
46-47
Hindu
geographical setting, 14-15

glottal stop, final
tone splitting inPR
?13
(907
960), 228
History of the Song Dynasty (960
1279), 2, 22
Hoa-binh,
change
themajor Haroi chan
44-45
versus
Chamic,
gradual
197-198
restructured
197, 214
register,
and Cham,
Haroi
to, 18
syllable
high
204
system,
vowels
glottalized
See main
after
assimilation
Highlands
Cham,
197
Donegan (1985), 204
prince fled to Lin-yi, 19

Wu
205-206
and presyllable
quality
and *h, 208-209
vowel
18
Oc-eo,
obstruents,
vowels,
18
location,
205
aspirates,
voiced
voice,
18
K'angT'ai,
202-203
voice,
voiced
vowel
18
18
Jayavarman,
204
197-199
register,
198-199
described,
Hall (1955, 1981), 17

Hsiang-lin,
in Southern
204
restructured
connection
first Austronesian
consonants,
sonorants,
tense
Blust (1992a), 17
Ch'ii-lien,
glottalized
15
Hr?, 2, 210
*s
403
Index
Topic
Li (Hainan), 22
and Proto-North
211-213
Bahnaric,
19, 20
Lin-yi,
loans.
Smith (1972), 211
See
borrowings
15
Hue,
M
I
80-99
consonants,
syllable
85-86
clusters with
-h-,
main
iambic
defined, 61
consonant
1,20
Indrapura,
inscriptions
fourth
Chamic,
century,
oldest,
Malayo-Polynesian,
and external
in Haroi,
178-179
93-94
external
232
general,
231-232
voiced
See
82-84
80-82
obstruents,
*c, 81-82
proto-phoneme
vowels
syllable
main
methodology
comparative
aspirated
84-86
obstruents,
voiceless
reconstructions.
86-93
84
consonants,
paths
Isbukun (Bunan dialect), 141
94-96
clusters,
obstruents,
and voiceless
voiced
Tsat,
inverted
clusters,
sonorants,
213
versus
internal
paths,
96-99
primary
glottalized
PC *w-, 84
Srivijaya, 3
internal
clusters,
secondary
3-4
93-99
clusters,
post-PC
126-137
borrowed,
inherited, 113-126
J
151
summarized,
Jiaozhi, 22
Malagasy
to Acehnese,
parallels
57
languages
Malayic
Adelaar (1988, 1992), 34
K
Kampong
Cepa,
23
Kelantan,
Kelantanese
Malayo-Chamic
shared
24
Malayo-Chamic
25
Malay,
Khmer
Mentawai,
Henderson (1952) on registers, 181

King Pau Kubah, 22
Kufic inscriptions dated 1025
1035, 228
Mentu,
causes,
internal
253-256
Grace, 254, 256

implementation,
relative
stasis,
254
Thomason
and Kaufman,
Labov,
Herzog,
names.
172
165,
sporadic,
methodology
110
for identifying, 278-280
1,20
MK
substratum
inheritance
253
Weinreich,
language
39-43
in Acehnese,
237-238
Cowan (1933 etc.), 237

Durie (1990a), 237
Niemann (1890), 237
Shorto (1975), 237
MK substratum inChamic, 237-238
Aymonier, 237
Crawfurd (1822), 237
L
Land Dayak, 165, 172
language
change
vs.
external
subgroup,
172
metathesis,
Mi-son,
39
innovations,
254,
and
255
See Appendix
256
accounts,
256-??
Malay of Champa, 237

Schmidt (1906), 237
Sebeok (1942), 237
Moken,
58-59
Lewis (1960), 58
Topic
404
typologically similarities to
Chamic, 59
sporadic, 153, 154
monophthongization,
nasals
sandhi forms, 166
inHaroi, 206
morphology
nasals,
"only", 249
*m- verbal prefix,
bipartite
causative
243-244
instrumental
Haroi with MK, 41

Phan Rang Cham with
245-246
infix,
network,
negative
243
imperative,
41
social,
1,20
NhaTrang,
nominalizing prefix, 244-245
Nonthaburi Malay, 24, 25
nouns
northern
and affixes,
244-249
Roglai and Tsat

first migration (986), 15
247-249
pronouns,
quantifier, 249
relative
clause
the
second migration (1486), 15

See
239-241
prefix,
and affixes,
239-244
verbs
N
finals
in Tsat,
164-167
reconstituting
nasalization,
reconstituting
place
articulation,
of
vowel
length,
Sung
late borrowings
into N.
Roglai, 174
and vowel deletion inCham, 155
176-177
inCham, 155-160
inChru, 153-154
inHaroi, 155
inN. Roglai, 170-176
in PC, 152-153
inTsat, 160-170
Matisoff, 154
174
Oc-eo,
15
Oc-eo.
See
Paiwanic,
161
allophonic before final stops, 155
perseverative,
Roglai
Northern
and Tsat,
43-44
22
dynasty,
also
Funan
P
164
163,
with the 42 falling tone, 169
in Acehnese,
Northern
161
nasalization
and
subgrouping
167
168
reconstruction,
reconstituting
also
Northern Roglai
etymology of, 224
161-170
and glottalization,
and subgrouping,
internal
15
Hainanisland,
249
marker,
siy, 249
'inadvertent'
sa-,
Cham
= Northern
prepositions, 246-247
nasal
41
Vietnamese,
249
-kan,
41
interactional,
network,
242
prefix,
249
165,166,167,172
preploded,
Blust, 172
Court, 165
241-242
negatives,
connectives,
176
(Acehnese),
pseverative
Index
32
Pattani Malay, 25
Pengkalan
Cepa,
Phan Rang,
1,20
24
Phan Rang Cham

an incipient
tone
187-197
system,
194-195
phonation
spreading,
187-197
tones,
breathy
tones,
> low
voice
tone,
190
Vietnamese
comparing
196
David Blood (1967), 187

Doris Blood (1962), 187
Greenberg (1970), 188
Han,
Edmondson,
and
Gregerson (1992), 188, 193

instrumental
study,
189
Index
Topic
4?5
Lafont
the nonexistent
and
of
on
literature
remnants
reconstruction
187-188
tones,
connection
of voice
contact-induced
Haroi,
register
type. See
voice
ancient Khmer, 179

Modern Central Khmer, 179
Modern Northern Khmer, 179
presyllable
Acehnese
vowel
Acehnese
voiced
initial
presyllable
*c-, 70
*k- and
70
creaky voice
Western
Cham,
registers
stops
and
voiceless
106-113
of,
preservation
secondary
shifts,
sporadic metathesis,
22
Pu-Luo-E,
4
puny a as a genitive,
also
1,20
Huffman (1976), 198
107-110
of, 2
etymology
110-113
109
Sa Huynh
burial jars, 15
15
dating,
comparison
R
with
Rade
HangGon,
chain,
and Jarai,
raising,
72
45-46
179
lingling-o,
location
of
overview,
pennanular
reciprocal
inHaroi, 69
northern
15
Borneo,
dialect
197-199
comparison with Niah and

Tabon, 15
Rade
complexes
Roglai
carbon
Quang-Tri,
178-179
register
tones
register,
198-199
Haroi,
69-72
obstruents,
vowels,
See
restructured
72-75
breathiness,
also
the array inChamic, 178

the origins, 232-235
the starting point, 178
74-75
sonorants,
See
180-181
tones
and
overview,
79-??
correspondences,
retention
and reduction,
75-76
voiced
of vowels
See
70
'horse',
guda
extra-Chamic
presyllable
proliferation
See breathy
voice
See centralization
166
stops,
67-80
consonants,
voicing,
of
reflexes
of, 179
origins
voice quality
phonetic
See also
final
dissimilative
179
Vietnamese,
quality
post-nasalized
179
manifestations,
pitch, 179
Po Nagar,
complexes
in PR Cham, 194-195
inTsat, 221-223
inWestern Cham, 183-186
phonation
contact,
similarities,
described, 179
235-236
spreading,
206-208
to language
genetically-inherited similarities, 4
quality, 189
tones split by final glottal
stop, 192
phonation
historical,
reconstruction,
(1971), 188
Moussay
See Chamic
Chamic.
reconstruction,
PR tones, 196
15
16
sites,
15
15-17
stone
earrings,
16
Phu Hoa, 15
sites,
dating
of,
16
406
See
also
consonants
presyllable
Southern
Min.
Southern
Vietnam
sporadic
stop, 223-224
from final nasals
See Hainanese
highlands
sub
highlands
sub
or
219-220
vowels,
from final stops, 218-219
Vietnam
from PC
91
area,
of, 214-224
fortition of a final glottal
47
area,
southern
of, 212-21A
description
evolution
sesquisyllabic
defined, 61
SIL, 261
Index
Topic
*-s and
*-h,
in monosyllables,
76-79
dissimilation,
literature
Srivijaya, 3
216
215-216
on, 214-215
secondary
stress
223-224
constriction,
in modern
Tsat and Northern Roglai, 43^4
61
Malay,
in proto-Austroasiatic,
61
61
Tunjung, 172
Southeast
throughout
61
Asia,
62
ultimate,
subgrouping
Northern
32
Tsouic,
penultimate,
and Tsat,
Roglai
120
acoustically-motivated,
upglides,
Utsat = U + Tsat
217,
224, 224-227
Summer
Institute
of Linguistics,
22
261
Indera Berman,
Syah
Vietnamese
Syah Pau Ling, 22

stress,
syllable
stress,
and
reduction,
67
T
Tao
reflexes,
Be, 22
Li (Hainan), 22
20
Cham,
temples,
See also Mi-son,
tense
Phan
Trang,
voice
voice.
See
voice
Rang
vowel
with
voice
creaky
see phonation
spreading
proliferation
199
Haroi,
See Appendix
227-229
Islam,
on Hainan,
230-232
contact with theLi, 230

development
of
179
spreading
voice
creaky
tables.
connection
also
73
71,72,
voice
breathy
centralization
vowel gliding, 179

vowel length, 179
Tsat
of
tones,
214-232
etymology of, 224

migration history, 225-227
phonation
tones
also
See
Transliteration
contact
See
quality
Tra-ki?u,
also
See
Nha
Haroi, 197
inHr?, 212
tense
See
vowels,
Vijaya,
20
1,20,22
Vijaya,
voice quality,
179
and splits in consonant
20
Huang,
Tai-Kadai
migration
to the South",
"push
spreading,
221-223
also
restructured
register
vowels
inventories,
62
inventories,
PMP,
literature,
105-106
overview,
104-105
104
presyllable
vowel
See
also main
reduction,
syllable
62
vowels
vowel length, 138-151
Topic
407
Index
w
Western
word-final
Cham
literature
on register,
179-187
register,
Edmonson
181-182
words
182
Gregerson,
and Kvoeu-Hor
final
sonorants,
100-102
*-s,
of uncertain
102-103
origin,
334
Y
ya relative
(1977), 181
instrumental
yang
description,
phonation
spreading,
sonorants,
vowel
*-h and
and
Friberg
voiced
99-103
consonants,
final
Yeh-po'-t'i,
Yuan-shan
marker,
3
3
18
culture
complex,
32
183-186
182-183
obstruents,
proliferation,
182
marker,
relative
Z
182-183
Zhancheng,
2, 21,22
180-181

20006770

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

20006770

Uploaded by

Copyright:

Available Formats

University of Hawai'i Press

From Ancient Cham toModern Dialects

From Ancient Cham toModern Dialects:

1999 University of Hawai'i Press

Includes bibliographical references (p.

The geographical setting

of the Chamic Languages

The place of Chamic within Austronesian

the Basic Word: From Disyllabic

The iambic syllable: Early Mon-Khmer

Presyliable consonants: The details 67

8 The Origins of Registers

internal paths of change

(and transliteration tables)

II: The Chamic Lexicon

Author Index 395

title of this book From Ancient

has been written

readers, this monograph

from the outset

that these analyses

In part, this reflects the fact that Chamic

they differ, these conclusions

than has been found thus far.

In addition to the main

texts, this work

so that the reconstructions

can usually be done on the basis of the forms

this appendix, both reconstructed forms and borrowings are

forms are included

in this appendix, along with the supporting data, with the

as better and more

in this work for Proto

of the forms found reconstructed

insights, and copies of crucial papers. In particular, the

looking back to Proto-Chamic,

but also to look at them from the top down?from

tives produced amuch more sophisticated and insightful reconstruction

frequency, change positions taken in earlier drafts.

(1995) Some Languages

and a Rade wordlist

from the last century. At a much later point, Peter Ladefoged

reflected in the ideas presented here: Eric Oey, Patricia Donegan,

eta Thurgood. There

to believe, however, that those thanked will

various people have knowingly or unknowingly helped me

general way, I wish to acknowledge my intellectual debts to

and Paul Benedict,

who never lost track of the big picture.

than usual, I fully expect some of the analyses presented here

cited, as well as on the basis of instrumental studies on these same languages. I

refers to the Bahnar forms cited inAy monier

the register complex that

Cham, and Phan Rang Cham

a stage not fully recoverable by reconstruction

a term used to designate

1976; the numbers following

In this work, the Phan Rang Cham forms come

spoken forms. Cited modern spoken forms

Chamic; some of these forms do not reconstruct

Phan Rang Cham

a complex of features thatmay include length,

a term used to designate

Roglai always refers toNorthern Roglai