Professional Documents
Culture Documents
. ,
N?%UIZED
SOUNDS IN YORUBA
I.
After an inquir: 17i G the phonemic status of the nasalized consonants ?, ti, j as
opposed to oral V, -! . . the great number of Yoruba nasalized vowels is inspected.
The expression tkct a certain oral vowel also occurs nasalized is shown to be
dangerous for a &H insight into a system of nasalized vowels. To establish the
number of n=alized vowel phonemes it is not enough to find minimal pairs in
which an oral vowel is contrasted with its nasalized variety; the point is to find
minimal pairs with mutually contrastive nasalized vowels. Thus Yoruba with its
7 oral vowel phonemes is shown to have not more than 4 nasalized ones. e
356
357
One of these points is the phonemic interpretation
of nasalized
soundis. There is a great variety of nasalized sounds in Yoruba
questil,n which of them are phonemic and to which phoneme a cxrtain
sound is to be assigned is not always easy to answer?)
Nasahzed Consonants
In lxr Introduction to the Yomba Langzlage, Ida Ward gives the
following lists of phonemes (pp. 6,12) :
Vozvels;: i, e, &:,a, 3, 0, u, 3, 5, 5, ti; written : i, e, e, a, I), 0, u, in, an, OR,
*
UT
3).
kre noteworthy
have been discussed in detail from a phonetic
CI f vie-w. The unusual
sound k$ has been dweit on at iarge there and,
to Wards statements and those of others, has been described
as botIi1 in- and explosive, its essential characteristic
being a certain
suction. Yoruba pronunciation,
therefore, will be touched upon in the
presenll paper only in as far as it gives rise to questions in the phokemic
field.
From a phonemic point of view the only di
able things in Wards
list of consonants are 6 and 9.5) There is a t
nasalized consonant,
zed initial consonant :
P. Exa:nples
of the three sets are: non-nas
9% [ri],
to see; ZQI [Iwo], to drag; yi [ji], to turn; nasalized initial
consonmt:
rirt
[:?I,
to laugh; ZQVZ[CS], to be dear; yin [ji], tolay
eggs ; ; ti with high tone.
Warrif readily assumes nasalized r to be the result of the neighbouring nasal vowel, adlding by way of reason for this explanation:
No
attemp, t has been made to z&e a nasalized Y . . . . (p. 23, italics BS).
cussions: and helpful suggestions regarding the problems dealt with in this paper.
3) Th P n is not sounded but serves merely as mark of nasalization of the
precedir .g vowel.
4) 33. Siertsema, Some Notes on Yomba Phoraetics arrd SpedZing, to appear in
Bulletin de Hnstitut IFranGaisdAfrique Noire, B, 1958. Further referred to as
YOY.Phm. ,a& Sp.
5) To1avoid confusion, Wards indication of the sound, y, is taken ov(:r. The
phonetill: symbol would be ul, i;hisis used in the transcriptions of whole words.
But for 8 and 9, often w&m (with preceding s), her treatment is
different and she raises an interesting problem: . . L . it is possible,
even probable, that they are weakened forms of the nasal consonants
3~ and by, the former a combination of q with the lip-rounding
belonging to v and the latter a pure palatal consonant. gw and tiy are
sounds common to many West African languages. The pronunciation
ny, i.e. the froht: Gf the tongue in contact with the hard palatt.e, was
found in one dialect (Ilesha) : it would be interesting to discover if it
ut the pronunciation GUY,
i.e. the back of the
occurs in any other.
tongue actuaJly touching the soft palate, was not found in any speaker.
The informants I have worked with require a strong nasalization of the
consonant from the very beginning of its articulation: this l$oints to the
probability of the consonantal origin of these nasalized semLvowels
. . . . (p. i3). The last sentence evidently means to say tha:t 5 anirl 5 are
to be looked upon as originally, and perhaps still so phonemically, two
more nasaks: the labialized velar one and the palatal one. Further on
this is specified: Considerably more research is necessary to settle the
question d the o~igis of the nasalized ZQ,y, r ar d their methods of
representation (p. 25, italics BS).
It is true that the representation of nasalized ze,and :j is not consistent : the two words [w3], they, them, their and [ii]) you, your
(~1.) may be written won, yin, in which the final a indicates nasal zation
of the whole group, dr ~wprt, @in, in which the first n -indicates the
nasalization of the consonant, the last that of the vowel. With [%6] the
spelling titian is often used to indicate the subject pronoun, wgrt for the
object and possessive pronoun, but there is no difference in pronunciAonn
The idea that the spelling nw, ny should indicate that these two
consonants are to be looked upon as nasak is at first sight rather
attractive. In view of the structure of the u:onsonant system there is
even much to be said in favour of it, for it would make the system more
regular with a nasal corresponding to each voiced plosive 6).Historically it is supported by Westermanns statement that k$ and gb have
) The ffj, spelled j, Ward puts in a class by itself called affkative.
Mowever,
Yoruba j is different from English [d3]. The fricative is very slight and often
is hardly heard. It seems appropriate to describe it as a $alatd fdxive - the
palatai
articulation (instead of palatoalveolar as in English [dj]) being clearly
audible.
359
developed out of kst and gti respectively 7). Thus the series of labiallized
velars, too, would originally have been perfectly re@ar : kw - p - p,
and tht.: present system would be:
b
d
.
dl
g:
kP
gb
m
n
J1
V
?9ur
Firs.ly, 6 [gw] and 9 [Jl] would be the only nasals occurring before
nasalized vowels olziy and-never before oral ones. If the nasalization of
the vo,Nel should be colnsidered as caused by the nasal consonant, tlhere
would be the inexplic:able situation that some of the nasals always
cau:se? nasalization of the following vowel whereas others did not: the
a ;q WWO, to spend money, the e in m&z, three, the o in 90 SWZ, I
shctll Edeep, are not nasaliried; but any vowel after ~8 and 9 shows
strong nasalization.
onsonant nasalized
Sect jndly, nasalized r would remain as the so
by a :Iallowing nasalized vowel. This would n
e impossible, the r *
being a very open consonant with little friction, into which a nasal
resomnce would be more likely to extend than in a narrow fricative
such GS, e.g., x. But such open consonants are 6 and 5 too. Whatever
may llave been their pronunciation
in former days, at present there is
no nazial here phone tically spe L:&mg*), as the tongue does not touch the
pa.latt,l with these sounds 9). As far as I am able to observe, apart from
the 1c:wered soft palate and uvula, the oral formation of [j] and [dj] is
exact ;y the same as that of [j] and [w].
In view of these facts and in spite of the loss of symm.etry in the
-7) I). Westermann, Die! Westlickeri Sudansfwachm und ihre Reziehungen zum
Baultzlc, 1927, pp. 8,9.
8) Ef there should be, VJBIwould still not be a true labio-velar nasal; at mast it
would be a Zabialized velar nasal, i.e. a velar nasal pronounced with lip-roundbing,
not with lip-cZos%ve.As it is, it is mostly a nasalized labio-velar semi-vowel, Le.
with :IOclosure either at the lips or at the velum.
9) At least not in Standard Yoruba (Cf. the quotation from Ward on p. 358
abovci:).
360
consonant system, the above considerations rather underline the conclusion arrived at before, that in Yoruba only m, 92 and y are nasal in
their own right : they occur both before nasalized vowels 10)and before
oral ones (see a,bove, p. 359) and naturally have so orn2 ~ariat~ts.
Nasalized w, y, r, on the contrary, occur only before nasalized vowels
and are thus ~nviGww&ally determind variants of oral w, y, r, which
*cur only before oral vowels. 11). Thus one might formulate the
pfi~nornenon ais d simple case of assimilation by stating that w, yJ r are
nasalized before nasalized vowels. No indication in the spelling, like the
present rczlv,ny, would be necessary and the phoneme inventory could
be reduced by three items 11)- or rather, seeing that Ward adopts the
above #explanation for nasatized r: by MO items.
361
S#kdling
and
Sounds
362
souds
-an
--on
-un
15) Comparative
363
In this amorphous mass of nasalized vowels, those whose tongue
position corresponds to or&lones (underlined in the scheme) tend to be
singled out as phnemes for no other reason than this purely phonetic
relationship. Thus Olmsted has readily found minimal pairs of words
differentiated by oral versus nasal vowel: e.g. /&/, liver - /E&/,
grief ; /idol, canax seed - /ids/, craving; /ikti/, death - /ikG/,
squimel leb). ut it would be difficult to discover minimal pairs in
which all Olmsteds nasal&d vowels are mutually contrasted and
distinctive, His no. 5, /:F/in /&dS/,grief, lzannot be contrasted with his
no. 6,161, in the same setting: /edi?/, nor with his no. 7, /a/ : /d3/, /a@,/ or
/&diiiiare variant pronunciations of the same word, spelled in Yontba
gdtm, grief.
In- view of all this it would be more realistic not to relate Yoruba
nasalized vowels to any of the oral ones and to refrain from the expression that a certain vowel occurs nasalized. In Yoruba the nasarized vowels form a class of phonemes by themselves in addition to the
seven oral vowels.
1
Phonemes
is simpler than one
From the phonemic point of view the mat
would expect from the complicated list of correspondences. A considleration of the nasalized vowe:Lphonemes leaves a system of comparatively simple structure.
Starting from the most obvious data, it can be established that the
spelling -2cn represents u WHWor Jess close back romded phoneme,
realized as [g] frequently after velar consonants, and as [ii] or [6], or in
fast speech sometimes [15],after labials and alveolars.
The spelling -Qn is superfluous. It occurs only after labials where it
represents the open phoneme, realized as the back rounded sound
after labial consonants, as [a] or [Zj after other consonants; with some
speakers still more fronted,as almost [%I after Y.The spelling -an co1
be uged in all cases to represent this phoneme.
The sound [Z], represented by -en occupies a peculiar place ationg
Yoruba nasalized vowels. It would be very convenient and it would
make the phoneme syst.em so much simpler if one could discard it, like
[5J# as a mere environmentally determined realization of the third
l
y&J
firmly)
365
366
what Doke calls the etrt~a-grammaticapphonetics of onomatopoebr $8).
If asked, therefore, how many nasalized vowel phonemes there are in
Yoruba, one might venture, although at some risk, to suggest with
Ward that there are indeed only three. Considering the wide r
possible realizations for each of them, these phonemes could
safely indicated as:
f . a more - or less - &se front wwmmded cm,
2. :a more - or less - class back rounded one,
one,
2. a more - or lee - front (or back) O@VS
covering the ranges of tongue positions roughly T outlined in the scheme :
University College
Ibadan, Nigeria
B. SIERTSEMA