Norway: Language Situation 709
stops, /b/, /d/, and /g/, such as Guinaang Bontok [f], [ts],
and [ h] (Himes, 19841985), which because of the in-
fluence of English in the schools are losing their environ-
mental conditioning and are now becoming separate
phonemes in the languages (Reid, 2005).
Published text resources are available on a number
of Cordilleran languages. Moreover, one of the finest
dictionaries of a Philippine language is that of the
central Cordilleran language Ifugao (Newell, 1993).
See also: Austronesian Languages: Overview; Ergativity;
Kapampangan; Malayo-Polynesian Languages; Philip-
pines: Language Situation; Tagalog.
Bibliography
Barlaan R R (1999). Aspects of focus in Isnag. Manila:
Linguistic Society of the Philippines.
Bellwood P, Anderson A, Stevenson J & Dizon E (2003). Ar-
chaeological and palaeoenvironmental research in Batanes
and Ilocos Norte provinces, northern Philippines. Bulletin
of the Indo-Pacific Prehistory Association 23, 141161.
Brainard S (1994). The phonology of Karao, the Philippines.
Canberra: Pacific Linguistics.
Brandes C O & Scheerer O (19271928). On sandhi in the
Ibana g language. In Scheerer O (ed.) The Archives 6.
Manila: University of the Philippines. 542.
Norway: Language Situation
A R Linn, University of Sheffield, Sheffield, UK
2006 Elsevier Ltd. All rights reserved.
Norwegian is a North Germanic language, spoken by
around 4.5 million inhabitants of Norway and a small
number of speakers elsewhere, principally emigrant
workers in other Scandinavian countries and the
United States, although Ethnologue also notes groups
of speakers in Canada, Ecuador, and the United Arab
Emirates. The spoken language is characterized
by rich dialectal variation, resulting historically
from small groups of speakers spread across a large
area with poor internal communications. Systematic
dialectology has been a fertile branch of linguistic
enquiry in Norway since the second half of the 19th
century. Dialectal variation is a feature of all social
groups and is relatively little stigmatized. The
written language situation is more complex, and writ-
ten Norwegian provides the classical example of lan-
guage planning.
Grimes B F & Grimes J E (eds.) (2000). Ethnologue: languages
of the world (14th edn.). Dallas: SIL International.
Himes R S (19841985). Allophonic variation and the
Bontok-Kankanaey voiced stops. In Philippine Journal
of Linguistics. 1516, 4956.
Himes R S (1998). The southern Cordilleran group
of Philippine languages. Oceanic Linguistics 37(1),
120177.
Newell L E (1993). Batad Ifugao dictionary with ethnographic
notes. Manila: Linguistic Society of the Philippines.
Reid L A (1974). The central Cordilleran subgroup of
Philippine languages. Oceanic Linguistics 13, 511560.
Reid L A (1979). Towards a reconstruction of the pronominal
systems of Proto-Cordilleran, Philippines. In Nguyen D L
(ed.) Southeast Asian linguistic studies 3. Canberra: Re-
search School of Pacific Studies, Department of Linguistics,
Australian National University. 259275.
Reid L A (1987). The early switch hypothesis: linguistic
evidence for contact between Negritos and Austronesians.
Man and Culture in Oceania 3(Special Issue), 4160.
Reid L A (2005). A cross-generational view of contact-
related phenomena in a Philippine language: phonology.
In Quakenbush S & Dayag D (eds.) Sociolinguistics and
language education in the Philippines and beyond:
festchscrift in honor of Ma. Lourdes S. Bautista. Manila:
Linguistic Society of the Philippines / Summer Institute of
Linguistics.
Reid L A & Liao H-C (2004). A brief syntactic typology of
Philippine languages. Language and Linguistics 5(2),
433490.
Following political independence from Denmark
in 1814, several proposals for the establishment of a
native written language were put forward. The two
most vigorously pursued were the work of Ivar Aasen
(18131896) on the one hand and Knud Knudsen
(18121895) on the other. Aasen, with the backing of
the State, created a new written language based on ex-
tensive and pioneering dialect fieldwork, and Knudsen
pursued a policy of Norwegianization of the inherited
Danish. Both policies had their opponents, but in 1885
the two varieties were placed on an equal footing by
Parliament, leading to the written standards initially
known as Landsmaal and Riksmaal, respectively, but
renamed Nynorsk and Bokma l in 1929. Norwegian
language politics in the first half of the 20th century
were characterized by a series of official reforms in-
tended to bring the two varieties closer together. The
reforms of 1917 and 1938 in particular were under-
pinned by the ultimate goal of creating one common
Norwegian, Samnorsk. This policy was never popular,
and protests were especially vehement in the 1950s,

stops, /b/, /d/, and /g/, such as Guinaang Bontok [f], [ts],
and [ h] (Himes, 19841985), which because of the in-
fluence of English in the schools are losing their environ-
mental conditioning and are now becoming separate
phonemes in the languages (Reid, 2005).
Published text resources are available on a number
of Cordilleran languages. Moreover, one of the finest
dictionaries of a Philippine language is that of the
central Cordilleran language Ifugao (Newell, 1993).
See also: Austronesian Languages: Overview; Ergativity;
Kapampangan; Malayo-Polynesian Languages; Philip-
pines: Language Situation; Tagalog.
Bibliography
Barlaan R R (1999). Aspects of focus in Isnag. Manila:
Linguistic Society of the Philippines.
Bellwood P, Anderson A, Stevenson J & Dizon E (2003). Ar-
chaeological and palaeoenvironmental research in Batanes
and Ilocos Norte provinces, northern Philippines. Bulletin
of the Indo-Pacific Prehistory Association 23, 141161.
Brainard S (1994). The phonology of Karao, the Philippines.
Canberra: Pacific Linguistics.
Brandes C O & Scheerer O (19271928). On sandhi in the
Ibanag language. In Scheerer O (ed.) The Archives 6.
Manila: University of the Philippines. 542.
Norway: Language Situation
A R Linn, University of Sheffield, Sheffield, UK
2006 Elsevier Ltd. All rights reserved.
Norwegian is a North Germanic language, spoken by
around 4.5 million inhabitants of Norway and a small
number of speakers elsewhere, principally emigrant
workers in other Scandinavian countries and the
United States, although Ethnologue also notes groups
of speakers in Canada, Ecuador, and the United Arab
Emirates. The spoken language is characterized
by rich dialectal variation, resulting historically
from small groups of speakers spread across a large
area with poor internal communications. Systematic
dialectology has been a fertile branch of linguistic
enquiry in Norway since the second half of the 19th
century. Dialectal variation is a feature of all social
groups and is relatively little stigmatized. The
written language situation is more complex, and writ-
ten Norwegian provides the classical example of lan-
guage planning.
Norway: Language Situation 709
Grimes B F & Grimes J E (eds.) (2000). Ethnologue: languages
of the world (14th edn.). Dallas: SIL International.
Himes R S (19841985). Allophonic variation and the
Bontok-Kankanaey voiced stops. In Philippine Journal
of Linguistics. 1516, 4956.
Himes R S (1998). The southern Cordilleran group
of Philippine languages. Oceanic Linguistics 37(1),
120177.
Newell L E (1993). Batad Ifugao dictionary with ethnographic
notes. Manila: Linguistic Society of the Philippines.
Reid L A (1974). The central Cordilleran subgroup of
Philippine languages. Oceanic Linguistics 13, 511560.
Reid L A (1979). Towards a reconstruction of the pronominal
systems of Proto-Cordilleran, Philippines. In Nguyen D L
(ed.) Southeast Asian linguistic studies 3. Canberra: Re-
search School of Pacific Studies, Department of Linguistics,
Australian National University. 259275.
Reid L A (1987). The early switch hypothesis: linguistic
evidence for contact between Negritos and Austronesians.
Man and Culture in Oceania 3(Special Issue), 4160.
Reid L A (2005). A cross-generational view of contact-
related phenomena in a Philippine language: phonology.
In Quakenbush S & Dayag D (eds.) Sociolinguistics and
language education in the Philippines and beyond:
festchscrift in honor of Ma. Lourdes S. Bautista. Manila:
Linguistic Society of the Philippines / Summer Institute of
Linguistics.
Reid L A & Liao H-C (2004). A brief syntactic typology of
Philippine languages. Language and Linguistics 5(2),
433490.
Following political independence from Denmark
in 1814, several proposals for the establishment of a
native written language were put forward. The two
most vigorously pursued were the work of Ivar Aasen
(18131896) on the one hand and Knud Knudsen
(18121895) on the other. Aasen, with the backing of
the State, created a new written language based on ex-
tensive and pioneering dialect fieldwork, and Knudsen
pursued a policy of Norwegianization of the inherited
Danish. Both policies had their opponents, but in 1885
the two varieties were placed on an equal footing by
Parliament, leading to the written standards initially
known as Landsmaal and Riksmaal, respectively, but
renamed Nynorsk and Bokmal in 1929. Norwegian
language politics in the first half of the 20th century
were characterized by a series of official reforms in-
tended to bring the two varieties closer together. The
reforms of 1917 and 1938 in particular were under-
pinned by the ultimate goal of creating one common
Norwegian, Samnorsk. This policy was never popular,
and protests were especially vehement in the 1950s,
710 Norway: Language Situation
when the policy was being most actively pursued.
Since the 1960s a series of developments has led to
a gradual stalling of this policy, and in 2002 it was
abandoned. Policy is now to leave the two varieties to
develop in their own way, and while the Language
Council has been disbanded, a new language agency is
to be set up in its place, although its role and powers
are, at the time of writing, unclear.
The majority of Norwegians use a moderately con-
servative form of Bokma l in writing, a variety that
opponents of the reform process identify as Riksma l
to distinguish it from the more radically engineered
Bokma l. Riksma l has its own dictionary and an influ-
ential pressure group, The Norwegian Academy for
Language and Literature. Advocates of a more conser-
vative variety of Nynorsk, closer to the form original-
ly proposed by Aasen, identify and champion another
variety still, known as Hgnorsk. This has no official
status, and its pressure groups are, at the time of
writing, in some disarray.
Norwegian
T A Afarli, Norwegian University of Science and
Technology, Trondheim, Norway
2006 Elsevier Ltd. All rights reserved.
Sociohistorical Setting
Norwegian together with Danish and Swedish consti-
tute the Mainland Scandinavian languages, which,
together with the Insular Scandinavian languages,
Faroese and Icelandic, constitute the Scandinavian
languages. The Scandinavian languages belong to
the Germanic family of the Indo-European languages.
Norwegian is exceptional in having two officially
recognized written standards. These are called Bokmal
book language and Nynorsk New Norwegian. Bok-
mal is used by more than 80% of the population, where-
as Nynorsk is used by less than 20%, mainly in the area
stretching from the interior of southern Norway to the
western coast. As for the spoken language, there is a rich
variety of dialects, although all of them are mutually
intelligible. The use of (nonstandard) dialects is widely
accepted, even in more formal contexts. There is no
officially recognized standard for the spoken language.
It is important to note that Bokmal and Nynorsk
are written standards that individuals choose largely
irrespective of which dialectal variety they speak.
See also: Aasen, Ivar Andreas (18131896); Danish; Ger-
manic Languages; Norwegian.
Language Maps (Appendix 1): Map 136.
Bibliography
Bra tveit K, Jones W G & Gade K (1995). Colloquial
Norwegian: a complete language course. London &
New York: Routledge.
Haugen E (1966). Language conflict and language planning:
the case of modern Norwegian. Cambridge, MA: Harvard
University Press.
Haugen E (1976). The Scandinavian languages. London:
Faber & Faber.
Jahr E-H (1989). Utsyn over norsk sprakhistorie etter 1814.
Oslo: Novus forlag.
Linn A (1997). Constructing the grammars of a language:
Ivar Aasen and 19th-century Norwegian linguistics.
Mu nster: Nodus Publikationen.
Strandskogen A -B & Strandskogen R (1986). Norwegian:
an essential grammar. London & New York: Routledge.
The two written standards have their background
in the history of Norway. From 1380 to 1814,
Norway was in a political union with Denmark, and
Danish was used as a written language. Eventually,
during the flourishing of nationalism after 1814, a
Dano-Norwegian written standard was developed in
the 19th century, bringing Norwegian elements into
the Danish language. This standard gradually evolved
into present-day Bokma l. Also, in the 19th century, a
more radical approach was followed by Ivar Aasen
(18131896). He developed a written standard based
on the spoken dialects. This standard gradually
evolved into present-day Nynorsk.
Today, Bokma l and Nynorsk are quite similar, but
there are still certain spelling differences both regard-
ing content words and grammatical morphology. In
the description of Norwegian below, I will concen-
trate on the common structural features of the spoken
language as a whole. All examples are written in the
Nynorsk standard unless otherwise stated.
Morphology and Phonology
Finite verbs in Norwegian only show tense distinctions
(past, present). There are two nonfinite verb forms, the
infinitive and the past participle, and an adjectival
710 Norway: Language Situation
when the policy was being most actively pursued.
Since the 1960s a series of developments has led to
a gradual stalling of this policy, and in 2002 it was
abandoned. Policy is now to leave the two varieties to
develop in their own way, and while the Language
Council has been disbanded, a new language agency is
to be set up in its place, although its role and powers
are, at the time of writing, unclear.
The majority of Norwegians use a moderately con-
servative form of Bokmal in writing, a variety that
opponents of the reform process identify as Riksmal
to distinguish it from the more radically engineered
Bokmal. Riksmal has its own dictionary and an influ-
ential pressure group, The Norwegian Academy for
Language and Literature. Advocates of a more conser-
vative variety of Nynorsk, closer to the form original-
ly proposed by Aasen, identify and champion another
variety still, known as Hgnorsk. This has no official
status, and its pressure groups are, at the time of
writing, in some disarray.
Norwegian
T A Afarli, Norwegian University of Science and
Technology, Trondheim, Norway
2006 Elsevier Ltd. All rights reserved.
Sociohistorical Setting
Norwegian together with Danish and Swedish consti-
tute the Mainland Scandinavian languages, which,
together with the Insular Scandinavian languages,
Faroese and Icelandic, constitute the Scandinavian
languages. The Scandinavian languages belong to
the Germanic family of the Indo-European languages.
Norwegian is exceptional in having two officially
recognized written standards. These are called Bokmal
book language and Nynorsk New Norwegian. Bok-
mal is used by more than 80% of the population, where-
as Nynorsk is used by less than 20%, mainly in the area
stretching from the interior of southern Norway to the
western coast. As for the spoken language, there is a rich
variety of dialects, although all of them are mutually
intelligible. The use of (nonstandard) dialects is widely
accepted, even in more formal contexts. There is no
officially recognized standard for the spoken language.
It is important to note that Bokmal and Nynorsk
are written standards that individuals choose largely
irrespective of which dialectal variety they speak.
See also: Aasen, Ivar Andreas (18131896); Danish; Ger-
manic Languages; Norwegian.
Language Maps (Appendix 1): Map 136.
Bibliography
Bratveit K, Jones W G & Gade K (1995). Colloquial
Norwegian: a complete language course. London &
New York: Routledge.
Haugen E (1966). Language conflict and language planning:
the case of modern Norwegian. Cambridge, MA: Harvard
University Press.
Haugen E (1976). The Scandinavian languages. London:
Faber & Faber.
Jahr E-H (1989). Utsyn over norsk sprakhistorie etter 1814.
Oslo: Novus forlag.
Linn A (1997). Constructing the grammars of a language:
Ivar Aasen and 19th-century Norwegian linguistics.
Munster: Nodus Publikationen.
Strandskogen A-B & Strandskogen R (1986). Norwegian:
an essential grammar. London & New York: Routledge.
The two written standards have their background
in the history of Norway. From 1380 to 1814,
Norway was in a political union with Denmark, and
Danish was used as a written language. Eventually,
during the flourishing of nationalism after 1814, a
Dano-Norwegian written standard was developed in
the 19th century, bringing Norwegian elements into
the Danish language. This standard gradually evolved
into present-day Bokmal. Also, in the 19th century, a
more radical approach was followed by Ivar Aasen
(18131896). He developed a written standard based
on the spoken dialects. This standard gradually
evolved into present-day Nynorsk.
Today, Bokmal and Nynorsk are quite similar, but
there are still certain spelling differences both regard-
ing content words and grammatical morphology. In
the description of Norwegian below, I will concen-
trate on the common structural features of the spoken
language as a whole. All examples are written in the
Nynorsk standard unless otherwise stated.
Morphology and Phonology
Finite verbs in Norwegian only show tense distinctions
(past, present). There are two nonfinite verb forms, the
infinitive and the past participle, and an adjectival

present participle form. Example (1) shows the inflec-
tional paradigm for the weak verb kjpe buy.
(1) kjpe (infinitive) - kjper (present) - kjpte (past)
- kjpt (past participle)
Nouns have no productive case distinctions. How-
ever, they inflect for number (singular, plural) and
definiteness (definite, indefinite), and there are
three genders (masculine, feminine, neuter). Example
(2) shows the paradigm for the feminine noun geit
goat.
(2) sg pl
indef geit goat geiter goats
def geita the goat geitene the goats
As shown, definiteness is marked inflectionally as a
suffix (-ene is a portmanteau morph expressing both
definiteness and plurality). There is also a free-form
indefinite article, varying according to gender; ei is
the feminine form, as in ei geit a goat.
Adjectives show agreement in gender (m/f and n)
and number (sg, pl) in predicative and attributive
positions.
As for phonological properties, Norwegian dia-
lects have relatively rich vowel and consonant sys-
tems. Also, Norwegian dialects make extensive use
of diphthongs. In most dialects, tones may have
distinctive function.
Syntax
Norwegian is a SVO language with fixed word order,
cf. (3).
(3) Mannen kjper geita.
The man buys the goat.
Norwegian is also a verb second (V2) language, with
the finite verb in second position in main declarative
clauses. Thus, topicalization of the direct object in (1)
yields geita kjper mannen the goat the man buys
with the finite verb in the V2 position.
Auxiliary verbs are positioned between the subject
and the main verb, as in mannen har kjpt geita the
man has bought the goat. Finite auxiliary verbs fol-
low the V2 pattern, cf. geita har mannen kjpt the
goat the man has bought.
Adverbials are positioned in the middle field (sen-
tence adverbials) or toward the end of the clause
(predicate adverbials).
(4) Mannen har aldri kjpt geiter pa ein sndag.
The man has never bought goats on a Sunday.
Like other V2 languages, Norwegian shows an asym-
metry between main and embedded clauses as to the
relative distribution of sentence adverbial and finite
Norwegian 711
verb, compare (4) to the embedded clause in eg veit at
mannen aldri har kjpt geiter pa ein sndag I know
that the man has never bought goats on a Sunday.
Norwegian has a strict subject requirement in finite
clauses. If there is no semantic subject, an expletive
subject must be inserted.
(5a) Det regnar.
It rains.
(5b) Det sta r ei flaske pa bordet.
it stands a bottle on table-the
There is a bottle standing on the table.
In most dialects, det it is used as expletive subject in
both meteorological (5a) and presentational (5b) sen-
tences. Some dialects allow der there as an expletive
subject in addition to det it.
Norwegian has two main types of passive, namely
periphrastic passive and reflexive (s-) passive. Exam-
ple (6b) is written in the Bokma l standard.
(6a) Geita blir kjpt (av mannen) i dag.
goat-the becomes bought (by man-the) to day
The goat is bought by the man today.
(6b) Geita kjpes (av mannen) i dag.
goat-the buy-s (by man-the) to day
The goat is bought by the man today.
Both types allow impersonal passive, as e.g., in det
blir kjpt ei geit i dag there is bought a goat today.
There are two main types of interrogative clauses:
yes/no questions (formed by placing the finite verb in
initial position) and questions with a question word
(with the question word placed in initial position).
As for relative clauses, the most common type is intro-
duced by the complementizer som. The complementi-
zer is optional if the relativized position is a
nonsubject, as in geita (som) mannen kjpte the goat
(that) the man bought. Infinitival clauses are intro-
duced by the infinitive marker a to, as in mannen
prver a kjpe geita the man tries to buy the goat.
See also: Aasen, Ivar Andreas (18131896); Danish; Inflec-
tion and Derivation; Language/Dialect Contact; Language
and Dialect: Linguistic Varieties; Morphological Typology;
Norse and Icelandic; Norway: Language Situation; Scan-
dinavian Lexicography; Subjects and the Extended Pro-
jection Principle; Swedish.
Bibliography
Askedal J O (1994). Norwegian. In Ko nig E & van der
Auwera J (eds.) The Germanic langugaes. London:
Routledge. 219270.
Faarlund J T, Lie S & Vannebo K I (1997). Norsk referanse-
grammatikk. Oslo: Universitetsforlaget.
712 Noun Phrases
Noun Phrases
J R Payne
2006 Elsevier Ltd. All rights reserved.
This article is reproduced from the previous edition, volume 5,
pp. 28482855, 1994, Elsevier Ltd.
The term noun is used to describe the class of lexical
items whose prototypical members refer to entities
(chair, film, leg), substances (butter, concrete, blood),
and named individuals or locations (John, Paris,
France). Traditionally, the term noun phrase has then
been used to describe any phrase consisting of a noun
as its central constituent (or head) and capable of
functioning as an argument (e.g., subject or object) in
a sentence. Examples, in which the head nouns are
in bold type, are: my big red chair, the butter on the
table; John, who is a lion-tamer. The structure of noun
phrases, which in many respects appears to be as
complex as the structure of sentences, has become a
controversial issue. While some linguists maintain that
noun phrases are headed by nouns, others insist that
the head of a noun phrase should be the determiner
(e.g., the definite article the in the butter on the table).
According to these theorists, noun phrases should
really be called determiner phrases.
Nouns
Nouns can be divided into a number of important
subclasses based on their grammatical and semantic
properties. These subclasses relate to definiteness,
person, number, and gender, and they play a major
role in determining what categories can cooccur with
a given noun in a noun phrase.
Definiteness
One subclass of nouns, the subclass of proper nouns
whose prototypical members refer to named indivi-
duals or locations, contains nouns whose reference is
exclusively definite (i.e., when the speaker uses such a
noun, he or she indicates that the hearer will be able
to identify the proposed referent by locating some
mutually manifest set which consists uniquely of the
named individual or location). In some languages,
proper nouns are used in conjunction with definite
articles (e.g., Greek o Yanis (literally the John)),
which gives a noun phrase whose definiteness is indi-
cated both by the noun and by the article. In other
languages, this redundancy may be avoided and a
noun phrase may consist of the proper noun alone.
The majority of proper nouns in English are of this
type (e.g., John, Paris, France); however, some proper
nouns (e.g., the Hague, the Netherlands) require the
use of the article.
Opposed to proper nouns are common nouns, for
which a distinction between definite and indefinite
reference is possible. In indefinite reference, the speaker
does not indicate to the hearer that the hearer will
necessarily be able to locate a mutually manifest set
in which the referent can be found. Indefinite refer-
ence is therefore either used when some new entity
or substance is being introduced into the discourse
(Yesterday I saw a film in which Meryl Streep played
a New York psychiatrist), or, if a set mutually manifest
to both speaker and hearer can be found, to indicate
that the referent is not unique within that set. In the
sentence John lost a leg in the war, a set of two
legs belonging to John may be mutually manifest to
speaker and hearer, but if only one leg has been lost,
indefinite reference is forced (see Hawkins, 1991).
In languages which distinguish between definite
and indefinite articles, common nouns are able to oc-
cur with both. English has definite article the and
indefinite articles a and some, giving a contrast be-
tween the film and a film, and the butter and some
butter. With proper nouns, on the other hand, the use
of contrasting definite and indefinite articles involves
special interpretations: I met a John at the party, but
I dont think its the person you mean; The John that
I met at the party isnt present today. In this case, it can
be said that John has been reclassified as a common
noun meaning something like person named John.
Person
Distinctions of person are typically carried by a spe-
cial class of nouns called personal pronouns, for ex-
ample, I (first person singular), you (second person
singular), he/she/it (third person singular), and they
(third person plural) in English. These are similar to
proper nouns in that they have definite reference, and
do not cooccur with the definite article: *the you.
Common nouns are neutral with respect to person.
When nouns are used with standard determiners such
as the articles or demonstratives, the noun phrase is
usually interpreted as having third-person reference,
but, as pointed out by Postal (1966), common nouns
can also be combined in English with first-person
plural we and second-person plural you to give noun
phrases such as we students and you students. We and
you can therefore be analyzed as first- and second-
person person determiners in English, rather than as
personal pronouns.
The neutrality of common nouns with respect
to person is especially clear in Nama (Namibia). In
Nama, all noun phrases contain a final marker

showing the person (as well as the number and gen-
der) of the phrase. These markers are as follows (ex-
cluding special plural forms for mixed masculine and
feminine phrases):
Masculine Feminine
Sing Dual Pl Sing Dual Pl
First person -ta -khom -ge -ta -m -se
Second person -ts -kho -go -s -ro -so
Third person -b -kha -gu -s -ra -di
All these markers can be attached to any common
noun, giving forms such as ao-b (man-third person
singular) or ao-khom (man-first person dual). The
third-person pronoun ||i behaves like any other
noun, except of course that it can only be followed
by third-person markers: ||i-b he, ||i-s she, etc. In ad-
dition, Nama also has determiners representing first
and second person:
Sing Dual/Plural
First person ti si
Second person sa sa
These occur in initial position in first- and second-
person noun phrases to give forms such as si ao-khom
we (two) men.
The first- and second-person personal pronouns
of Nama are formed simply by combining the rele-
vant determiner and noun-phrase suffix, for example,
ti-ta I, sa-kho you (dual), si-khom we (dual, ex-
cluding the hearer), sa-khom we (dual, including
the hearer). In this case, as for English we and you
(plural), it can be argued that no noun root as such
is present in the noun phrase.
Number
Common nouns can be divided into those which can
be counted (count nouns, like chair), and those which
cannot (noncount nouns, like butter). Count nouns
are able to cooccur with low numerals (one if they are
singular, and two if they are plural), for example, one
chair, two chairs. The use of low numerals with non-
count nouns leads to unacceptability, for example,
*one butter, *two butters, unless the noun in question
is reclassified as a common noun with a different and
countable interpretation, for example type of butter
or pat of butter.
Count nouns may refer to sets consisting of one en-
tity, or more than one entity, and this is the basis for
number distinctions in language such as singular and
plural. It is necessary, however, to distinguish be-
tween grammatical and semantic number. In English,
collective nouns such as family are grammatically
singular since they cooccur with grammatically sin-
gular forms of noun-phrase modifiers such as
the singular indefinite article a and the singular
Noun Phrases 713
demonstratives this and that: a family, this family,
that family. Equally, family has a grammatically plu-
ral form families which cooccurs with the grammati-
cally plural indefinite article some and the
grammatically plural demonstratives these and
those: some families, these families, those families.
However, the grammatically singular form family
can be thought of as referring semantically to a set
consisting of its individual members, and this is
reflected in the possibility of plural verb agreement:
All the family are coming to the wedding.
Conversely, in some languages it is possible to find
grammatically plural nouns which are semantically
singular. For example, the Russian noun casy clock
is grammatically plural since it is modified by plural
forms of the demonstratives tot that and etot this:
te casy (that-plural clock) that clock, eti casy (this-
plural clock) this clock. Its semantic singularity is
shown by its ability to cooccur with the numeral odin
one, which also has a grammatically plural form
especially for use with such nouns: odni casy (one-
plural clock) one clock! English nouns like scissors,
which appear similar (e.g., those scissors, these scis-
sors), are grammatically plural, but noncount. One
cannot say *one scissors, but must add the count
noun pair for the numeral to be possible: one pair of
scissors.
While count nouns typically have both gram-
matically singular and grammatically plural forms in
English, there are a few like remains which are only
plural (*this remains, these remains). Note that nouns
like sheep are count nouns with both a singular and
plural form (this sheep, these sheep): the singular
and plural forms in this case just happen to be
the same.
Gender
Nouns have different genders if they fall into different
agreement classes, that is, if they cooccur with differ-
ent agreeing forms of other categories such as noun-
phrase modifiers, predicates, or pronouns.
In English, gender distinctions are absent from
noun-phrase modifiers like adjectives, articles, and
demonstratives, and from predicates. However, rela-
tive pronouns distinguish between personal nouns
which are referred to by who and nonpersonal
nouns which are referred to by which, and singular
personal pronouns distinguish between masculine
nouns (referred to by he), feminine nouns (referred
to by she), and neuter nouns (referred to by it). The
distinction between personal and nonpersonal is
not quite the same as that between human and non-
human, since familiar animals such as pets can be
referred to by who, and unfamiliar babies can
714 Noun Phrases
be referred to by which. Nor is it the same as the
distinction between animate and inanimate, since
both animate nouns like sheep and inanimate nouns
like box are referred to by which.
This then gives essentially the following gender
classes in English:
(a) personal masculine nouns (who/he): uncle
(b) personal feminine nouns (who/she): aunt
(c) nonpersonal neuter nouns (which/it): box
(d) nonpersonal feminine nouns (which/she): ship
Some nouns belong to more than one class. For ex-
ample, doctor can be both personal masculine and
personal feminine; ship can be nonpersonal neuter
as well as nonpersonal feminine. In the plural, the
distinction between masculine, feminine, and neuter
is neutralized (the personal pronoun is always they,
and the only gender distinction which is preserved is
between personal and nonpersonal).
In some languages, meaning alone suffices to deter-
mine distinctions of gender (English is basically of
this type). In other languages, however, information
about the form of the noun (e.g., declension class,
phonological form) is also required. In Russian, for
example, nouns fall into a few major declensional
classes, each of which is associated with a particular
gender: dom (first declension) house is masculine,
ruka (second declension) arm, hand is feminine,
skorost (third declension) speed is feminine, and
okno (fourth declension) window is neuter. This may
lead, as with number, to discrepancies between gram-
matical and semantic characteristics of particular
nouns. For example, the Russian noun djadja uncle
belongs to the second declension class, but in this
case the semantic characteristic takes precedence and
agreeing forms are invariably masculine (e.g., moj
djadja (my-masculine uncle) my uncle). The noun
vrac doctor belongs to the first declension, but when
referring to a female doctor may be used with either
masculine or feminine agreeing forms. In this case,
feminine agreeing forms are more likely with pro-
nouns and predicates than with noun-phrase modifiers
like adjectives (Corbett, 1991). The similarity with the
number agreement of collective nouns in English is
striking, since here, too, semantic agreement is found
with pronouns and predicates.
Noun-Phrase Modifiers
Noun phrases are traditionally thought of as consist-
ing minimally of a head noun, together with any num-
ber of noun-phrase modifiers. Typical noun-phrase
modifiers are determiners, quantifiers and quantifier
phrases, adjectives and adjective phrases, nouns and
noun phrases, adpositions and adpositional phrases,
and clauses.
Determiners
Determiners form a closed class of functional words
which have the general property of not themselves
permitting modification. The class of determiners in-
cludes articles, personal determiners, demonstratives,
interrogative determiners, exclamatory determiners,
and quality determiners.
Articles Articles have the function of simply indicat-
ing definite or indefinite reference, without contain-
ing any identification within themselves of the place
where the referent(s) can be located. The English
articles are definite the and indefinite a or some. In
indefinite reference, a is used with singular count
nouns, and some with noncount and/or plural nouns.
Examples are: the film, the butter, the films, a film,
some butter, some films. Many languages lack articles
altogether (e.g., Russian, Lithuanian).
Personal Determiners Personal determiners are de-
terminers which indicate the person of a noun phrase.
In English, the personal determiners are we and you
(plural), which can cooccur with grammatically plu-
ral common nouns. Examples are: we students, you
students. Nama (see Person) has both singular and
plural first- and second-person determiners. Since per-
sonal determiners invariably seem to indicate defi-
niteness, it might be argued that they are simply
first- and second-person forms of the definite article.
However, the definite article in English does not nec-
essarily imply third-person reference: for example,
in an appositional construction like we, the students,
the first-person personal determiner and the definite
article can cooccur. It seems better to treat the usual
interpretation of third-person reference induced by
the definite article as an implicature which can be
canceled if the context requires it.
Demonstratives Demonstratives serve generally to
indicate the location of the intended referent(s) of
the noun phrase with respect to the context of utter-
ance. English demonstratives are this/these and that/
those, which cooccur with common nouns and have a
distinction of grammatical number. Examples are:
this box, these boxes, that box, those boxes. The dis-
tinction between this/these and that/those is one of
closeness to and distance from the speaker. In some
languages, further distinctions based on closeness to
and distance from the hearer can also be found, as
well as more recondite distinctions based on visibility

and topological factors such as uphill and down-
hill, upriver and downriver. For example, Dyirbal
(Australia) has basic forms giyi (meaning that the
referent is visible and here), bayi (meaning that
the referent is visible and there), and nayi (meaning
that the referent is not visible). To any basic form can
be suffixed one of a set of twelve morphemes indi-
cating distance uphill or downhill, distance upriver
or downriver, and/or one of a set of three morphemes
indicating up (vertically), down (vertically), or out
in front (with respect to which way either speaker
or hearer is facing; Dixon, 1972). It should be
noted that demonstratives may indicate definiteness,
as in English, or may be neutral with respect to
definiteness, as in Dyirbal.
Interrogative Determiners Interrogative determiners
are forms like which and what in English. Examples
are: which box?, what box?
Exclamatory Determiners Exclamatory determiners
are forms like what in English. Examples are: what a
fool!, what fools! Note that the exclamatory deter-
miner what, unlike the interrogative determiner what,
must cooccur with the indefinite determiner a when
determining a singular count noun.
Quality Determiners Quality determiners are forms
like such in English. Examples are: such a fool,
such fools. Like the exclamatory determiner what,
this form requires the indefinite determiner a when
determining a singular count noun.
Quantifiers and Quantifier Phrases
Quantifiers have the general function of indicating
the quantity of elements referred to by the noun
phrase. Unlike determiners, they permit various
kinds of modification and therefore have their own
phrasal structure.
In many cases, quantifiers permit the same kinds of
modification as adjectives, and might therefore be
treated as a special subclass of adjectives. The quanti-
fier many in English, for example, permits premodifi-
cation by the adverb very (e.g., very many boxes), and
has comparative and superlative forms (more boxes,
most boxes).
Typical quantifiers in English are: all, both, half,
every, each, any, either, some, much, enough, several,
many (more, most), a few, few (fewer, fewest), a
little, little (less, least), neither, no, together with the
cardinal numerals one, two, three, etc. Examples are:
all the boxes, enough boxes, few boxes, no boxes,
two boxes. With modification of the quantifier, it is
Noun Phrases 715
possible to form quantifier phrases: virtually all the
boxes, not nearly enough boxes, indescribably few
boxes, absolutely no boxes, almost two boxes.
Adjectives and Adjective Phrases
Adjectives and adjective phrases are noun-phrase
modifiers par excellence. As well as general adjectives
like red, large, round, and brave, the class of adjec-
tives in English includes ordinal numerals such as first
and second, related adjectives such as next and last,
adjectives such as same and other, and arguably the
whole class of quantifiers (see Determiners).
In some languages, the class of adjectives may be
extended to include possessive adjectives, for exam-
ple, Italian mia in la mia casa (the my-feminine singu-
lar house-feminine singular) my house, and in others,
according to Dixon (1977), it may be severely reduced
or nonexistent (the concepts being expressed by forms
which can be classified as nouns or verbs).
Adjectives in English are generally positioned be-
tween determiners and the head noun, for example,
that first brave attempt, a large red apple. While there
is a certain natural ordering relationship between the
adjectives themselves, orders which deviate from this
ordering are typically possible: that brave first at-
tempt, a red large apple. Even quantifiers, which in
English typically occupy positions close to (or in com-
plementary distribution with) the determiners, may
occur in differing orders with ordinary adjectives:
those two first brave attempts, those first two brave
attempts, those brave two first attempts.
Adjective phrases in English may be formed by
premodification of the adjective by adverbs, or post-
modification by prepositional phrases and clauses:
very proud, proud of her achievements, proud that
she had won so easily. The type of modification has
an influence on the order of noun head and adjective-
phrase modifier within the noun phrase. Only pre-
modified adjective phrases pattern with single adjec-
tives in occurring before the head noun: a very proud
woman. Postmodified adjective phrases must occur
after the head: a woman proud of her achievements, a
woman proud that she had won so easily.
Nouns and Noun Phrases
Nouns themselves may act as noun-phrase premodi-
fiers. For example, it can be argued that, in English
noun phrases like a plastic factory, the item plastic is
a noun rather than an adjective. The evidence for this
is that plastic can itself be modified by an adjective
such as corrugated to give a corrugated plastic fac-
tory ( a factory which makes corrugated plastic).
Adjectives themselves do not permit modification by
716 Noun Phrases

adjectives, so there is no alternative but to consider (c) the Eskimos who(m) you met
plastic as a noun. the Eskimos that you met
It should be noted that noun-phrase modifiers the Eskimos you met
of this type cannot contain determiners, although The relative clause who live in igloos in (a) is non-
they may contain quantifiers: a party committee, restrictive, in that it provides additional information
*a that party committee, a two-party committee, an about the referents of a noun phrase, the Eskimos,
all-party committee. whose identity has already been established. By con-
trast, the relative clauses in (b) and (c) are restrictive,
Adpositions and Adpositional Phrases in that the information contained in the relative
clause serves to restrict the intended referents to
Adpositions are prepositions or postpositions, typi-
those who satisfy the condition expressed in the rela-
cally taking noun-phrase complements. Prepositional
tive clause. For example, the relative clauses in (b)
phrases are a common form of noun-phrase postmo-
restrict the reference to a subset of the set of Eskimos,
difier in English: the man on the roof, a box without a
those who happen to live in igloos.
lid, the room underneath the eaves. Some preposi-
Nonrestrictive relative clauses in English always
tions can occur alone as postmodifiers: the room
begin with a wh-phrase containing a wh-word
underneath.
like who or which. Restrictive clauses may, however,
On the other hand, postpositional phrases in
begin with a wh-phrase, or the subordinating conjunc-
English with the genitive postposition -s are pre-
tion that, or zero (symbolized by ). Zero, however, is
modifiers with a variety of functions including the
not permitted when the position relativized (the
possessor function (the kings crown), the subject
position in the relative clause which represents
function (the kings decree that windows should be
the role played by the intended referents) is the sub-
taxed), the object function (the kings execution), and
ject position as in (b), rather than the object position
the descriptive function (a womens hall of residence).
as in (c).
Phrases of the first three types in English induce the
The relative clauses in (a), (b), and (c) are based on
so-called definiteness effect, that is, the kings crown
the finite verb forms live and met. However, in many
means the unique crown belonging to the king, and
languages, relative clauses are essentially based on
not a crown belonging to the king. In order to ex-
nonfinite verb forms, for example, Turkish John un
press the indefinite meaning, the double genitive is
Mary-ye ver-dig-i patates (John-genitive Mary-dative
used in which the preposition of governs a postposi-
give-participle-his potato) the potato John gave to
tional phrase with -s: a crown of the kings.
Mary (Keenan, 1975). Here, the position relativized
While phrases of the first three types are in comple-
is the object position. Such relative clauses may be
mentary distribution in English with the determiners
called participial relative clauses. The closest En-
(i.e., Edwards richly jewelled crown and not *that
glish equivalents are clauses like living in igloos in
Edwards richly jewelled crown), this is not universal-
the noun phrase all Eskimos living in igloos. Howev-
ly the case. In Finnish, for example, the premodifying
er, in English participial relative clauses, the position
possessor and demonstrative are compatible: nuo
relativized is restricted to the subject position, and a
Aimon koirat (those Aimo-genitive dogs) those dogs
full finite relative clause must be used when the posi-
of Aimos. There therefore seems to be no general
tion relativized is the object or any other sentential
basis for considering premodifying possessor, subject,
function.
and object phrases as themselves determiners.
Complement clauses, unlike relative clauses, do not
Descriptive genitive phrases of the fourth type
contain a position relativized. Rather, the clause
have an entirely different distribution. They are com-
represents the propositional content of a thought or
patible with all the determiners and typically occur
utterance expressed by abstract nouns like belief,
close to the head noun, after any adjective phrases: a
statement, rumour, etc. Examples are: the belief that
very pleasant womens hall of residence.
linguistics is easy, the statement that the pound
Clauses would not be devalued, the rumour that the prime
minister would resign.
Clause modifiers can be divided into two major types:
relative clauses, and complement clauses. Noun-Phrase Structure
Examples of relative clauses in English are:
Configurational versus Appositional Structures
(a) the Eskimos, who live in igloos
(b) the Eskimos who live in igloos There is considerable evidence that the structure of
the Eskimos that live in igloos noun phrases in English is configurational, that is,

Figure 1 Q Quantifier, D Determiner, A Adjective, PP
Postpositional or Prepositional Phrase, N Noun, S Relative
Clause.
that the structure of a noun phrase like all those five
pleasant womens halls of residence which are de-
scribed in the University prospectus can be described
by a hierarchical constituent structure as in Figure 1.
For example, to give just one argument for one aspect
of the structure in Figure 1, the circled node can act
as the shared node in shared constituent coordina-
tion: all those five and all these three pleasant
womens halls of residence which are described in the
University prospectus.
By contrast, the structure of noun phrases in many
languages can clearly be appositional, with little or no
hierarchical structure. In Nama, for example, a noun
phrase like [6kini]-di [ne !nona]-di [!Gombates di]-di
([book]-third person feminine plural [these three]-
third person feminine plural [!Gombate of]-third
person feminine plural) these three books of !Gom-
bates can be considered as consisting of three minia-
ture noun phrases in apposition, each marked as third
person feminine plural by the postposition of -di and
each capable of occurring in isolation. The flavor of
the construction is captured by an English version
like: books, these three, the ones of !Gombate. The
whole construction, however, forms a single phrase to
which a case-assigning postposition may be attached
(Figure 2).
In Nama, an alternative method of constructing
noun phrases is hierarchical, and in this case the
modifiers are always unmarked for person, number,
and gender, and always precede rather than follow
the head noun. A similar distinction between frac-
tured and nonfractured phrases has been reported
in the Australian language Gooniyandi (McGregor,
1989). In Gooniyandi, however, the elements of a
fractured phrase are each independently marked for
case, and while they may form a single appositional
Noun Phrases 717
Figure 2
unit, they may also be split and placed in different
clause positions according to principles of informa-
tional structure.
The X-bar Framework
One very influential framework for labeling the con-
stituent structures of hierarchical noun phrases is the
X-bar framework introduced by Chomsky (1970) in
order to capture the kind of similarity between the
structure of sentences and noun phrases which is
seen most clearly in nominalization, the formation of
nouns from verbs. According to Chomsky, the noun
phrase Euclids proof of the theorem is similar in
structure to the sentence Euclid proved the theorem:
like the verb prove, the noun proof takes a subject
(Euclid) and a complement (the theorem), the main
differences between the two structures (the introduc-
tion of the adpositions -s and of ) being related to the
choice of a nominal rather than a verbal head.
In one early version of the X-bar theory (see
Jackendoff, 1977 for discussion), the verb is taken
to be the head of a sentence, and the parallel be-
tween sentence and noun phrase emerges very clearly
(Figure 3). The maximal (phrasal) level of any cate-
gory is taken to be X00 (where X represents any word-
level category N(oun), V(erb), P(reposition), or
A(djective)). N00 is therefore equivalent to noun
phrase, P00 to prepositional phrase, A00 to adjectival
phrase, and V00 is equivalent to sentence. Comple-
ments are placed as sisters of an intermediate level
node X0 , and subjects (treated invariantly as N00 de-
spite the presence of the adposition-s in the noun
phrase subject) are placed in the so-called specifier
position as daughters of N00 or V00. The parallelism is
however spoilt in this early version only by the pres-
ence of a tense node T in the sentence, and the fact
that in noun phrases the specifier position may be
occupied by a determiner node D, capturing the com-
plementary distribution between -s genitive phrases
and determiners which holds in English (Figure 4).
In the model of the noun phrase expressed by
Figures 3 and 4, the N0 node is recursive and permits
718 Noun Phrases
Figure 3
the addition of any number of additional modifiers
such as adjective phrases, prepositional phrases, and
relative clauses. A noun phrase such as those difficult
proofs of the theorem from Euclid which are set out in
our textbook would have the structure in Figure 5.
Structures like Figure 5 seem fairly successful in
capturing the structure of complex noun phrases.
However, there are some potential problems. First,
there are elements which occur higher in the structure
and do not seem to be well catered for: for example,
the quantifiers all, both, and half, which in English
can precede the definite determiners (all the boxes),
and genitive phrases in languages (e.g., Finnish and
Hungarian) which allow the genitive phrase to pre-
cede the determiner rather than occur in complemen-
tary distribution with it. Second, the distinction
between complements and other kinds of modifier
seems related to the nouns ability to assign a seman-
tic (or thematic) role to the complement, but it is
not clear that this entails the existence of a different
syntactic structure for complements involving the
existence of a distinct N0 node (for example, the ten-
dency of complements to occur closer to the noun
than other modifiers could be explained by a principle
such as the domain adjacency condition of Travis
(1984), which requires elements assigned a thematic
role by a head to be adjacent to that head).
These issues remain to be developed thoroughly.
One possible direction in which solutions might lie
is in abandoning bar structures altogether and treat-
ing all modifiers as sisters of a single recursive node
with no distinctions of bar level (a move proposed in
principle by, for example, Speas, 1990). In the case of
noun phrases, the single node would be the node N,
Figure 4
Figure 5
and all modifiers, without exception, would be sisters
of N.
Further independent principles would then be re-
quired to account for the possible orders in which
modifiers can occur. Work on the typology of noun
phrases has begun to show, however, that the relative
order of some basic types of premodifiers is extremely
stable across the worlds languages, the order being:
demonstrative numeral adjective head noun (see
Rijkhoff, 1992). This suggests that, at least for these
categories of modifier when they precede the head,
very general principles governing the relative applica-
tion of modifiers are involved.
The DP Hypothesis
Discussions of noun-phrase structure which maintain
the noun as the head of the noun phrase have been
overshadowed by the DP (determiner phrase) hy-
pothesis, which maintains that determiners are the
ultimate heads (Abney, 1987). A survey of the devel-
opment of the DP hypothesis may be found in
Ouhalla (1991: ch. 4).
A typical DP structure for the noun phrase the
proof of the theorem is Figure 6 Instead of being
 Noun Phrases 719

Figure 6
Figure 7 t is the trace of the D00 Euclid after movement from
specifier position in N00 , and e is the empty D node.

a modifier, the determiner the is the head of the

construction, taking an N00 proof of the theorem as
its complement. In a noun phrase like Euclids proof of
the theorem, the specifier positions of N00 and D00 can
both be exploited by assuming that Euclid originates
in the specifier position of N00 , where it can be assigned
the appropriate thematic role (agent) by the noun.
Euclid can then move to the specifier position of D00 ,
where it is assigned the genitive case (surfacing as
the -s postposition) by the empty element D Figure 7.
This structure, in which the subject and object of
the noun are originally both within N00 and movement
of the subject from within N00 up to the specifier
position of a nominal functional category D, then
parallels an analysis of sentences in which the subject
and object of the verb are originally both within V00 Figure 8 t is the trace of the D00 Euclid after movement from
specifier position in V00 , and e is the empty I node.
and movement of the subject places the subject in
the specifier position of a sentential functional cat-
egory I (which contains information about the tense the sentence (for example, the inclusion of aspect
and agreement properties of the sentence). The and negation), further functional categories will
parallelism between noun phrase and sentence emerge within the noun phrase.
originally sought by the X-bar theory is then complete While the DP hypothesis looks attractive, it too has
(see Figure 8). a number of problems (see Payne, 1993 for discus-
Work in the DP hypothesis has sought to refine the sion). For example, an analysis of the noun phrase in
analysis of the functional categories in the noun which there is a skeleton of N nodes from the phrasal
phrase by including D (determiner) and AGR (agree- to the lexical level naturally accounts for the percola-
ment) as heads, paralleling the separation of I into tion along this skeleton of morphological agreement
T (tense) and AGR (agreement) in the analysis of sen- features (e.g., case, number, gender, definiteness) in
tences. This may account for noun phrases in which languages like Russian, in which nearly all noun
nouns have markers of agreement with their genitive phrase constituents are morphologically marked. In
phrases (e.g., Turkish s u benim s apka-m (this my Figure 9, for example, the features of instrumental
hat-first person singular) this my hat. case and plural number percolate to every constituent
In addition, the node K (case) has been proposed of the noun phrase.
as the highest functional head in noun phrases which If, however, categories like case and definiteness
are marked for case, reflecting the idea that case are represented by single individual nodes K and
determines the external distribution of noun phrases. D, and these are the highest nodes in constituency
Cased noun phrases are then KP rather than DP structure trees under the DP hypothesis, the natural
or NP. It can be predicted that, parallel to the prediction is that only the highest nodes in the tree
further refinement of functional categories within will be morphologically marked for those categories.
720 Noun Phrases
Figure 9
See also: Articles, Definite and Indefinite; Case; Definite
and Indefinite; Gender; Nouns; Number; Pronouns; Word
Classes/Parts of Speech: Overview; X-Bar Theory.
Bibliography
Abney S P (1987). The English noun phrase in its senten-
tial aspect. Ph. D. dis., Massachusetts Institute of
Technology.
Chomsky N (1970). Remarks on nominalization. In
Jacobs R A & Rosenbaum P S (eds.) Readings in English
transformational grammar. Waltham, MA: Ginn.
Nouns
M Koptjevskaja-Tamm, Stockholm University,
Stockholm, Sweden
2006 Elsevier Ltd. All rights reserved.
Prototypical nouns refer to things, persons, places,
and other more or less concrete objects, but the noun
class itself normally includes many more items. Nouns
always constitute an open class, even in languages
that have relatively few nouns compared with the
European languages, such as Iroquoian languages
(Mithun, 2000). Languages have normally multiple
ways of deriving nouns, both from other word classes
by means of nominalization (see Nominalization) and
from other nouns. Nouns or their different forms can
develop into words of other classes, e.g., adverbs and
adpositions. Nouns seem to exist as a distinct word
Corbett G (1991). Gender. Cambridge: Cambridge Univer-
sity Press.
Dixon R M W (1972). The Dyirbal language of North
Queensland. Cambridge: Cambridge University Press.
Dixon R M W (1977). Where have all the adjectives gone?
Studies in Language 1, 1980.
Hawkins J A (1991). On (in)definite articles: implicatures
and (un)grammaticality prediction. JL 27(2), 405442.
Jackendoff R (1977). Constraints on phrase structure
rules. In Culicover P W, Wasow T & Akmajian A (eds.)
Formal syntax. New York: Academic Press.
Keenan E (1975). Relative clauses. In Shopen T (ed.)
Language typology and syntactic description, vol. 2.
Cambridge: Cambridge University Press.
McGregor W (1989). Phrase fracturing in Gooniyandi.
In Maracz L & Muysken P (eds.) Configurationality:
the typology of asymmetries. Dordrecht: Foris Publi-
cations.
Ouhalla J (1991). Functional categories and parametric
variation. London: Routledge.
Payne J R (1993). The headedness of noun phrases: slaying
the nominal hydra. In Corbett G, Frase N M &
McGlashan S (eds.) Heads in grammatical theory.
Cambridge: Cambridge University Press.
Postal P (1966). On so-called pronouns in English. In
Reibel D & Schane S (eds.) Modern studies in English.
NJ: Prentice Hall, Englewood Cliffs.
Rijkhoff J (1992). The noun phrase: a typological study
of its form and structure. (Dissertation, University of
Amsterdam.)
Speas M (1990). Phrase structure in natural language. Dor-
drecht: Kluwer Academic.
Travis L (1984). Parameters and effects of word order
variation. (Doctoral dissertation, Massachusetts Institute
of Technology.)
class in most languages, even though languages vary
as to how pronounced their nounverb distinction is
(Evans, 2000; also see Word Classes/Parts of Speech:
Overview).
Nominal Subcategories
Nouns can be classified in various ways. It has become
usual in linguistic literature to order nouns according
to one or another version of what has been termed
the animacy hierarchy or the topicality hierarchy
(Comrie, 1981: 178193), the hierarchy of individua-
tion (Sasse, 1993), or the empathy hierarchy (Kuno and
Kaburaki, 1977; Lehmann and Moravcsik, 2000).
The basic idea is that these hierarchies reflect speak-
ers different degrees of empathy with different types
of entities and their different needs and/or abilities
720 Noun Phrases
Figure 9
See also: Articles, Definite and Indefinite; Case; Definite
and Indefinite; Gender; Nouns; Number; Pronouns; Word
Classes/Parts of Speech: Overview; X-Bar Theory.
Bibliography
Abney S P (1987). The English noun phrase in its senten-
tial aspect. Ph. D. dis., Massachusetts Institute of
Technology.
Chomsky N (1970). Remarks on nominalization. In
Jacobs R A & Rosenbaum P S (eds.) Readings in English
transformational grammar. Waltham, MA: Ginn.
Nouns
M Koptjevskaja-Tamm, Stockholm University,
Stockholm, Sweden
2006 Elsevier Ltd. All rights reserved.
Prototypical nouns refer to things, persons, places,
and other more or less concrete objects, but the noun
class itself normally includes many more items. Nouns
always constitute an open class, even in languages
that have relatively few nouns compared with the
European languages, such as Iroquoian languages
(Mithun, 2000). Languages have normally multiple
ways of deriving nouns, both from other word classes
by means of nominalization (see Nominalization) and
from other nouns. Nouns or their different forms can
develop into words of other classes, e.g., adverbs and
adpositions. Nouns seem to exist as a distinct word
Corbett G (1991). Gender. Cambridge: Cambridge Univer-
sity Press.
Dixon R M W (1972). The Dyirbal language of North
Queensland. Cambridge: Cambridge University Press.
Dixon R M W (1977). Where have all the adjectives gone?
Studies in Language 1, 1980.
Hawkins J A (1991). On (in)definite articles: implicatures
and (un)grammaticality prediction. JL 27(2), 405442.
Jackendoff R (1977). Constraints on phrase structure
rules. In Culicover P W, Wasow T & Akmajian A (eds.)
Formal syntax. New York: Academic Press.
Keenan E (1975). Relative clauses. In Shopen T (ed.)
Language typology and syntactic description, vol. 2.
Cambridge: Cambridge University Press.
McGregor W (1989). Phrase fracturing in Gooniyandi.
In Maracz L & Muysken P (eds.) Configurationality:
the typology of asymmetries. Dordrecht: Foris Publi-
cations.
Ouhalla J (1991). Functional categories and parametric
variation. London: Routledge.
Payne J R (1993). The headedness of noun phrases: slaying
the nominal hydra. In Corbett G, Frase N M &
McGlashan S (eds.) Heads in grammatical theory.
Cambridge: Cambridge University Press.
Postal P (1966). On so-called pronouns in English. In
Reibel D & Schane S (eds.) Modern studies in English.
NJ: Prentice Hall, Englewood Cliffs.
Rijkhoff J (1992). The noun phrase: a typological study
of its form and structure. (Dissertation, University of
Amsterdam.)
Speas M (1990). Phrase structure in natural language. Dor-
drecht: Kluwer Academic.
Travis L (1984). Parameters and effects of word order
variation. (Doctoral dissertation, Massachusetts Institute
of Technology.)
class in most languages, even though languages vary
as to how pronounced their nounverb distinction is
(Evans, 2000; also see Word Classes/Parts of Speech:
Overview).
Nominal Subcategories
Nouns can be classified in various ways. It has become
usual in linguistic literature to order nouns according
to one or another version of what has been termed
the animacy hierarchy or the topicality hierarchy
(Comrie, 1981: 178193), the hierarchy of individua-
tion (Sasse, 1993), or the empathy hierarchy (Kuno and
Kaburaki, 1977; Lehmann and Moravcsik, 2000).
The basic idea is that these hierarchies reflect speak-
ers different degrees of empathy with different types
of entities and their different needs and/or abilities

to perceive those as individuals. The common part in
all of them is the basic distinction humans other
animates inanimate entities. The highest (left) part
of the proposed hierarchies differs; thus, an additional
split may single out personal proper names and occa-
sionally certain kin terms. The highest positions
belong to pronouns for speech-act participants (first-
and second-person pronouns), i.e., not nouns. Actual-
ly, opinions and descriptive praxis vary on whether
(personal) proper names should be considered a
subclass of nouns or a category of their own. Thus,
they often have inflectional and distributional pecu-
liarities stemming from their inherent uniqueness and
definiteness, e.g., resistance to pluralization and non-
compatibility with usual articles. In a number of lan-
guages, personal proper names require special proprial
articles, distinct from those used with normal nouns.
The lower part of the hierarchies differs even more
substantially among the different scholars. Thus, ab-
stract nouns, such as love, beauty, or democracy, are
opposed to all the others, concrete nouns and are
clearly lower in their degree of individuation. The
same goes for words denoting substances (prototypi-
cal mass nouns), such as air or water, compared with
words denoting individual objects (prototypical count
nouns), e.g., girl, cat, and cup. However, these two
distinctions are rather orthogonal to each other (for a
different opinion, see Sasse, 1993).
Various structural phenomena and grammatical
rules across languages are in one way or another
sensitive to the animacy/empathy/individuation hier-
archy, e.g., the distribution of number distinctions
(see Number), the number of case distinctions (see
Case), direct versus inverse marking of subjectobject
relations (e.g., in Navajo), etc., (see Dahl and
Fraurud, 1996 for an overview). However, since the
hierarchy looks different for the different phenomena
and is often operationally defined as to whatever
determines the phenomena under investigation
(Dahl, 2000: 40), one might suspect that it is in fact
an illusion, hiding several different, more basic dis-
tinctions. And in fact, the seemingly self-explanatory
concept of individuation is itself in need of inde-
pendent psycholinguistic research (Comrie, 1981:
178193).
Classifications along the above given lines are sen-
sitive to the type of entity denoted by a noun, i.e., to
its correlate in the physical world, the class of its
possible referents. However, what often counts even
more is the mental reflection of these possible refer-
ents, the way they are construed, i.e., the properties of
the corresponding concepts in the mental lexicon.
Take, for instance, the opposition between count and
mass nouns. In English, the count noun cow easily
combines with numerals, e.g., one cow, three cows,
Nouns 721
etc., with the indefinite article a, and with the denu-
merating measure words many, several, etc. The mass
noun petrol cannot combine with numerals and with
the indefinite article a, takes the measure word much
instead of many, and lacks the number opposition
otherwise typical of English nouns (cf. cow vs. cows
with petrol vs. *petrols). The underlying semantic
difference is often described as the difference between
denoting (or referring to) discrete entities with a well-
defined shape and precise limits versus homogeneous,
undifferentiated stuff without any certain shape or
precise limits. The count/mass distinction in lan-
guages such as English, French, or Russian shows a
strong correlation between the properties of referents
and their proneness to be referred to by a count or a
mass noun, more or less in accordance with the ani-
macy hierarchy. Thus, humans and animals favor a
high degree of individuation and are normally coded
by count nouns, but stuffs are more often than not
expressed by mass nouns. However, discreteness
(boundedness) of count-noun referents and homoge-
neity of mass-noun referents reflect not so much the
objective reality itself, but rather its construals by
language speakers. Thus, under appropriate circum-
stances several cows can alternatively be referred
to as (several) cows; as a herd (of cows), by means
of a collective count noun; or as cattle, by means of
a collective mass noun. There are amazing cross-
linguistic differences in the lexical preferences for
count or mass uses even within the European Indo-
European languages, which deserve profound study;
it is also far from clear to what extent the count/
mass distinction is useful crosslinguistically (Behrens,
1995; Rijkhoff, 2002: 4456; Koptjevskaja-Tamm,
2004b). Personal proper names are another example
of an ontologically based, rather than phenomenologi-
cally based, classification: thus, for example, organiza-
tions are often given proper names that may behave
linguistically more or less similarly to personal proper
names (Fraurud, 1996).
Language-specific classifications of nouns (or
of the corresponding entities in our conceptual
world) underlie their different combinability with
various linguistic elements. The most grammatica-
lized form of nominal classification is gender (see
Gender, Grammatical). Several properties distin-
guish prototypical genders from prototypical classi-
fiers: numeral classifiers, occurring in numeral
constructions, e.g., in many Southeast Asian lan-
guages (see Numerals), or possessive classifiers, oc-
curring in adnominal possession, e.g., in Melanesian
languages (see Possession, Adnominal; Classifiers and
Noun Classes: Semantics). There are, however, many
intermediate cases (cf. Aikhenvald, 2000; Rijkhoff,
2002).
722 Nouns
Relationality
While almost all verbs are relational in the sense of
relating participants to an event (see Verbs), most
prototypical nouns are absolute, or sortal (Lo bner,
1985): they designate concrete individual objects,
e.g., a cup or a book, not obviously related to other
entities. A lot of nouns are relational for example,
body-part nouns and nouns referring to other parts,
kin terms and nouns referring to other interpersonal
relations, or words like consequence or results. They
invoke a special relation to another entity and can
often not be understood without this entity: a person
cannot be a neighbor or a son, but is always some-
ones neighbor or son. Many more nouns are relation-
al in a weak sense, in that their meaning itself
presupposes or invokes specific relations. Thus, for
example, since books are strongly associated with
their authors, with their content, and with a specific
purpose, the corresponding relations are readily in-
voked by the semantic structure of the word book.
The term qualia structure (Pustejovsky, 1998) refers
to a system of relations that characterize the seman-
tics of nominals, in a way parallel to the argument
structure of a verb, and includes the four basic roles:
(1) the constitutive role (the relation between an ob-
ject and its constituents, or proper parts), (2) the
formal role, distinguishing the object within a larger
domain (color, shape, etc.), (3) the telic role (purpose
and function of the object), and (4) the agentive role
(factors involved in the origin or bringing about of
an object).
There has been relatively much crosslinguistic re-
search on the issues of inalienable versus alienable
possession (see Possession, Adnominal), or how typi-
cally body-part terms and/or kin terms are singled out
in the ways their relations to other entities are
encoded. Otherwise, the nominal qualia structure
has received very little attention compared with the
extensive work on the verbal argument structure and
its manifestations across languages. There are, how-
ever, many interesting issues and amazing cross-
linguistic variation here. Thus, for example, Russian
and Swedish mark the telic role of key in different,
not to say opposite, ways: the key that belongs to the
door, i.e., the key the purpose of which is to open the
door corresponds to nyckeln TILL dorren, key.the TO
door.the, in Swedish, but to kljuc OT dveri, key FROM
door:GEN, in Russian.
Nominal Inflectional Morphology and
Functions
Nouns typically head noun phrases that are used
for reference, i.e., as arguments and adjuncts to
predicates (e.g., The little boy is playing with the
new toy) or as attributes to other nouns (the boys
new toy). They can also be used in nonreferring func-
tions, e.g., as predicates (Peter is a boy), and as non-
referential attributes, e.g., a childrens hospital. In
these functions nouns often show fewer of their typi-
cal inflectional properties and sometimes even ac-
quire certain properties of other word classes
(Hopper and Thompson, 1984; Stassen, 1997;
Croft, 2001; Mauri, 2003; Koptjevskaja-Tamm,
2004, and Nonverbal Predication).
Typical nominal inflectional markers include mar-
kers of number (see Number), case (see Case), posses-
sion (see Possession, Adnominal), and definiteness
(see Definite and Indefinite). The noun in (1), from
Hungarian, carries the plural number marker, the
third person plural possessive marker, and the accu-
sative case marker, while the above-mentioned Swed-
ish example demonstrates inflection for definiteness.
(1) bara t-ai-k-at
friend-PL-POSS.3PL-ACC
their friends (ACC)
Both case markers and possessive markers usually
signal syntactic relations between nouns (or noun
phrases headed by them) and other expressions and
thus constitute nominal dependent marking and head
marking, respectively (Nichols, 1986). Possessive
markers are normally chosen in accordance with
the person and number of the possessor, that is either
explicitly mentioned in the same construction or not.
In many languages nouns can also attach person mar-
kers when used as nominal predicates. Example (2),
from Chukchi (Chukot), illustrates a polite introduc-
tion routine.
(2) qun waj enmen CekwaNaqaj-eXem
INTERJ INTERJ so personal.name-1SG.ABS
[CekwaNaqaj:] Well, then Im CekwaNaqaj.
eej CekwaNaqaj-eXet
INTJ personal.name-2SG.ABS
[The woman:] Ooh, youre
CekwaNaqaj. (Dunn, 1999: 318)
As predicates, nouns can sometimes attach tense-
aspect-mood markers. However, even nouns used in
other functions as arguments and adjuncts can
occasionally inflect for these categories, independent-
ly of the clause they belong to, as in (3) from Guaran.
(3) O-va-ta che-roga-kue-pe.
3-move-FUT 1SG-house-PST-in
He will move into my former house.
(Nordlinger and Sadler, 2004: 781)
All of the categories expressed by nominal inflectional
morphology can also be expressed by other means,
such as grammatical words, and the different

categories differ in their crosslinguistic proneness to be
encoded by inflection. Thus, while case is a relatively
common inflectional category, bound definiteness
markers are much more rare.
Languages also differ considerably in whether they
have nominal inflectional morphology at all (and
how much). It has been suggested (Lehmann and
Moravcsik, 2000; Moravcsik, 2003) that nominal
inflection in a language implies verbal inflection.
Thus, while there are languages with both, with
neither (Vietnamese), and with only verbal but no
nominal inflection (Japanese), nominal inflection
without verbal does not yet seem to have been
attested. There are other crosslinguistic generaliza-
tions relating nominal and verbal morphology. For
example, according to Nichols (1986: 75), If a lan-
guage has major, salient head-marking morphology
anywhere, it will have it at the clause level, or in
other words, if a language uses head-marking in pos-
sessive noun phrases (primarily possessive affixes), it
will use head-marking to relate arguments to verbs
(i.e., primarily verbal agreement markers). Siewierska
(1998) confirms further that the majority of lan-
guages in her large global sample show close similar-
ity between nominal possessive affixes and verbal
person marking.
There are also possible generalizations on interac-
tion among different nominal inflectional categories.
According to Aikhenvald and Dixon (1998) and
Corbett (2000), there are many examples of case
systems with a larger number of distinctions in the
singular than in the plural, but no examples are given
of the opposite possibility. In Basque, number distinc-
tions hit only definite nouns. In other languages defi-
nite nominals sometimes show more elaborate case
and/or number paradigms than indefinite ones even
though languages with both case and bound definite-
ness markers are relatively rare: Matthew Dryers
700-language sample contains only 25 instances
(Matthew Dryer, personal communication).
The order of nominal inflectional (and derivational)
markers with respect to the stem and to each other
obeys certain principles, basically reflecting the degree
of semantic relevance of the corresponding categories
to the meaning of the stem (Bybee, 1985). However,
languages sometimes show contradictory orders be-
tween inflectional markers. Thus, while in (1) from
Hungarian, the possessive suffix precedes the case
suffix, Finnish nouns show the opposite order in their
nominal suffixes.
See also: Adpositions; Articles, Definite and Indefinite;
Case; Classifiers and Noun Classes: Semantics; Definite
and Indefinite; Gender, Grammatical; Kinship Expressions
Nouns 723
and Terms; Mass Expressions; Nominalization; Nonverbal
Predication; Noun Phrases; Number; Numerals; Objects,
Properties, and Functions; Possession, Adnominal; Predi-
cation; Pronouns; Proper Names: Philosophical Aspects;
Tense; Verbs; Word Classes/Parts of Speech: Overview.
Bibliography
Aikhenvald A Y (2000). Classifiers: a typology of
noun categorization devices. Oxford: Oxford University
Press.
Aikhenvald A Y & Dixon R M W (1998). Dependencies
between grammatical systems. Language 74, 5680.
Behrens L (1995). Categorizing between lexicon and
grammar. Lexicology 1, 1112.
Bybee J L (1985). Morphology: a study of the relation
between meaning and form. Amsterdam: Benjamins.
Comrie B (1981). Language universals and linguistic
typology. Oxford: Blackwell.
Corbett G (2000). Number. Cambridge: Cambridge Univer-
sity Press.
Croft W (2001). Radical construction grammar: syntactic
theory in typological perspective. Oxford: Oxford
University Press.
Dahl O (2000). Egophoricity in discourse and syntax.
Functions of language 7.1, 3770.
& Fraurud K (1996). Animacy in grammar
Dahl O
and discourse. In Fretheim T & Gundel J K (eds.) Refer-
ence and referent accessibility. Amsterdam: Benjamins.
4764.
Dunn M J (1999). A grammar of Chukchi. Canberra:
Australian National University.
Evans N (2000). Word classes in the worlds languages. In
Booij G, Lehmann C & Mugdan J (eds.) Morphology:
an international handbook on inflection and word-
formation, vol. 1. Berlin: Walter de Gruyter. 708732.
Fraurud K (1996). Cognitive ontology and NP form. In
Fretheim T & Gundel J K (eds.) Reference and referent
accessibility. Amsterdam: Benjamins. 6588.
Hopper P & Thompson S A (1984). The discourse basis for
lexical categories in universal grammar. Language 56,
251299.
Koptjevskaja-Tamm M (2004a). Marias ring of gold:
adnominal possession and non-anchoring relations in
the European languages. In Kim J-Y, Lander Y &
Partee B H (eds.) Possessives and beyond: semantics and
syntax. Amherst, MA: GLSA Publications.
Koptjevskaja-Tamm M (2004b). Mass and collection. In
Booij G, Lehmann C & Mugdan J (eds.) Morphology:
a handbook on inflection and word formation, vol. 2.
Berlin: Walter de Gruyter. 10671074.
Kuno S & Kaburaki E (1977). Empathy and syntax.
Linguistic Inquiry 8, 627672.
Lehmann C & Moravcsik E (2000). Noun. In Booij G,
Lehmann C & Mugdan J (eds.) Morphology: an interna-
tional handbook of inflection and word-formation, vol. 1.
Berlin: Walter de Gruyter. 732757.
Lo bner S (1985). Definites. Journal of Semantics 4,
279326.
724 Nouns
Mauri C (2003). I lessemi nominali in funzione
non referenziale: uno studio interlinguistico. Facolta`
di lettere e filosofia. Pavia: Universita` degli studi di
Pavia.
Mithun M (2000). Noun and verb in Iroquoian languages:
multicategorisation from multiple criteria. In Vogel P &
Comrie B (eds.) Approaches to the typology of word
classes. Berlin: Mouton de Gruyter. 397420.
Moravcsik E (2003). Inflectional morphology in the
Hungarian noun phrase: a typological assesment. In
Plank F (ed.) Noun phrase structure in the languages of
Europe. Berlin: Mouton de Gruyter. 113252.
Nichols J (1986). Head-marking and dependent-marking
grammar. Language 62, 56119.
Number
G G Corbett, University of Surrey, Guildford, UK
2006 Elsevier Ltd. All rights reserved.
Number is one of the most common features of
the worlds languages. While it is common, it is cer-
tainly not uniform. The situation in languages like
English is often treated as normal. Yet the number
feature reveals many surprises, and the system found
in English is a rather extreme one, when seen against
the possibilities established from a survey of the lan-
guages of the world. We shall consider characteristics
of the English-type systems in turn, showing some of
the ways in which they are typical or atypical and the
contrasting types of system found elsewhere. We shall
give particular prominence to the question of number
values. This entry is based on Corbett (2000), where a
much fuller account is given.
Nominal and Verbal Number
English makes a semantic and a formal distinction
between, for instance, cat and cats, child and chil-
dren. In English, this number distinction reflects a
nominal feature. It is realized on nouns via relatively
simple inflectional morphology. (There are many
languages with more complex morphological reali-
zation of number, see, for instance, the account of
Nilo-Saharan languages in Dimmendaal, 2000.) In
English, when the number feature is realized on
verbs, this is by agreement with a subject noun
phrase, as in the children were playing together. The
verb is plural to indicate the number of children, and
not the number of playing events. That is, we have a
nominal feature which can be realized where
expected, on the noun phrase, and also on the verb.
Nordlinger R & Sadler L (2004). Nominal tense in cross-
linguistic perspective. Language 80(4), 776806.
Pustejovsky J (1998). The generative lexicon. Cambridge,
MA: MIT Press.
Rijkhoff J (2002). The noun phrase. Oxford: Oxford
University Press.
Sasse H-J (1993). Syntactic categories and subcategories.
In Jacobs J, von Stechow A, Sternefeld W & Vennemann
T (eds.) Syntax: an international handbook of contempo-
rary research. Berlin: de Gruyter. 646686.
Siewierska A (1998). On nominal and verbal person mark-
ing. Linguistic Typology 2(1), 156.
Stassen L (1997). Intransitive predication. Oxford: Oxford
University Press.
While there are many languages like English in this
respect, we also find instances of verbal number,
that is, where number is realized on the verb to in-
dicate the number of events or the number of partici-
pants. For an example we turn to Rapanui (an
Oceanic language spoken on Easter Island) (Veronica
Du Feu, personal communication):
(1) ruku
dive
(2) ruku ruku
go diving
The form in (2) is appropriate if there is more than
one dive; it does not require more than one person to
be diving. The reduplicated form indicates verbal plu-
rality: there are plural events. Verbal number may also
be concerned with the number of participants rather
than the number of events (see, for instance, Durie,
1986). Verbal number may be restricted to relatively
small numbers of verbs, and it rarely shows more than
a two-way opposition (for details see Corbett, 2000:
243264). Nominal number shows a greater range of
possibilities and we shall concentrate on that type.
Number as an Obligatory Category
Number is an obligatory category in English. If some-
one says theres a goat in the garden when there are
several, this is misleading and inappropriate. In many
languages, however, the use of the plural is not oblig-
atory, but it is available when it is important to mark
number. Perhaps the clearest example is provided
by the Cushitic language Baiso, spoken on Gidicho
Island in Lake Abaya (southern Ethiopia) and on the
western shore of the lake. Baiso nouns can mark
general meaning, that is, they have a form which
724 Nouns
Mauri C (2003). I lessemi nominali in funzione
non referenziale: uno studio interlinguistico. Facolta`
di lettere e filosofia. Pavia: Universita` degli studi di
Pavia.
Mithun M (2000). Noun and verb in Iroquoian languages:
multicategorisation from multiple criteria. In Vogel P &
Comrie B (eds.) Approaches to the typology of word
classes. Berlin: Mouton de Gruyter. 397420.
Moravcsik E (2003). Inflectional morphology in the
Hungarian noun phrase: a typological assesment. In
Plank F (ed.) Noun phrase structure in the languages of
Europe. Berlin: Mouton de Gruyter. 113252.
Nichols J (1986). Head-marking and dependent-marking
grammar. Language 62, 56119.
Number
G G Corbett, University of Surrey, Guildford, UK
2006 Elsevier Ltd. All rights reserved.
Number is one of the most common features of
the worlds languages. While it is common, it is cer-
tainly not uniform. The situation in languages like
English is often treated as normal. Yet the number
feature reveals many surprises, and the system found
in English is a rather extreme one, when seen against
the possibilities established from a survey of the lan-
guages of the world. We shall consider characteristics
of the English-type systems in turn, showing some of
the ways in which they are typical or atypical and the
contrasting types of system found elsewhere. We shall
give particular prominence to the question of number
values. This entry is based on Corbett (2000), where a
much fuller account is given.
Nominal and Verbal Number
English makes a semantic and a formal distinction
between, for instance, cat and cats, child and chil-
dren. In English, this number distinction reflects a
nominal feature. It is realized on nouns via relatively
simple inflectional morphology. (There are many
languages with more complex morphological reali-
zation of number, see, for instance, the account of
Nilo-Saharan languages in Dimmendaal, 2000.) In
English, when the number feature is realized on
verbs, this is by agreement with a subject noun
phrase, as in the children were playing together. The
verb is plural to indicate the number of children, and
not the number of playing events. That is, we have a
nominal feature which can be realized where
expected, on the noun phrase, and also on the verb.
Nordlinger R & Sadler L (2004). Nominal tense in cross-
linguistic perspective. Language 80(4), 776806.
Pustejovsky J (1998). The generative lexicon. Cambridge,
MA: MIT Press.
Rijkhoff J (2002). The noun phrase. Oxford: Oxford
University Press.
Sasse H-J (1993). Syntactic categories and subcategories.
In Jacobs J, von Stechow A, Sternefeld W & Vennemann
T (eds.) Syntax: an international handbook of contempo-
rary research. Berlin: de Gruyter. 646686.
Siewierska A (1998). On nominal and verbal person mark-
ing. Linguistic Typology 2(1), 156.
Stassen L (1997). Intransitive predication. Oxford: Oxford
University Press.
While there are many languages like English in this
respect, we also find instances of verbal number,
that is, where number is realized on the verb to in-
dicate the number of events or the number of partici-
pants. For an example we turn to Rapanui (an
Oceanic language spoken on Easter Island) (Veronica
Du Feu, personal communication):
(1) ruku
dive
(2) ruku ruku
go diving
The form in (2) is appropriate if there is more than
one dive; it does not require more than one person to
be diving. The reduplicated form indicates verbal plu-
rality: there are plural events. Verbal number may also
be concerned with the number of participants rather
than the number of events (see, for instance, Durie,
1986). Verbal number may be restricted to relatively
small numbers of verbs, and it rarely shows more than
a two-way opposition (for details see Corbett, 2000:
243264). Nominal number shows a greater range of
possibilities and we shall concentrate on that type.
Number as an Obligatory Category
Number is an obligatory category in English. If some-
one says theres a goat in the garden when there are
several, this is misleading and inappropriate. In many
languages, however, the use of the plural is not oblig-
atory, but it is available when it is important to mark
number. Perhaps the clearest example is provided
by the Cushitic language Baiso, spoken on Gidicho
Island in Lake Abaya (southern Ethiopia) and on the
western shore of the lake. Baiso nouns can mark
general meaning, that is, they have a form which

is noncommittal as to number, along with singular,
paucal (example (8) below), and plural (Corbett and
Hayward, 1987). The form lu ban lion does not
specify a number of lions (it could be one, or more
than that) (Dick Hayward, personal communication).
(3) lu ban foofe
lion.GENERAL watched.1SING
I watched lion
Having a distinct form for general number is highly
unusual. There are many languages which can ex-
press general meaning, but they do so by means of
a form shared with the singular. This more usual
situation, with general identical to singular, can be
illustrated from Arbore (also a Cushitic language).
We find pairs like the following (Hayward, 1984:
159183).
(4) a. kele h
gelded goat(s)
b. keleh-me
gelded goats
c. goran
heifer cow(s)
d. gor-no
heifer cows
e. ke r
dog(s)
f. ker-o
dogs
Alhough the morphology may appear comparable
to English, the meanings of the forms are different:
kero guarantees more than one dog, and it would be
used when to indicate this is important; ke r does not
imply only one: it might be one, it might be more than
that. (There are other, less frequent number pairings
in Arbore.) Systems like that of Arbore, in which
number is not an obligatory category (and so is argu-
ably not inflectional) are common in the worlds lan-
guages. At the opposite extreme from English we find
Piraha , the only remaining member of the Mura fam-
ily, spoken in 1997 by some 220 people along the
Maici River (Amazonas, Brazil). It has been described
by Everett (1986), who stated that it had no system of
grammatical number, even in the pronouns. This is
extremely rare.
The Nominals Involved in the Number
System
In English and similar languages, the majority of
nominals can mark number, from the personal pro-
nouns and nouns denoting persons (e.g., nurse) down
to those denoting inanimate objects (e.g., blanket),
and even to some denoting abstracts (e.g., week).
Many languages restrict the number opposition to
Number 725
fewer nominals, namely those which are high in
animacy. Thus, in Slavey (an Athapaskan language
spoken in parts of the Northwest Territories, British
Columbia, and Alberta, Canada), number may be
marked on the noun phrase, provided the noun
denotes a human or a dog (Rice, 1989: 247).
We wish to investigate the countability preferences
(Allan, 1980) of nominals (including pronouns as well
as nouns). There is considerable variety in the worlds
languages. Consider the following Warrgamay exam-
ple (Queensland, Australia) (Dixon, 1980: 266268):
(5) yibi-yibi Nulmburu-Ngu
child-REDUP.ABS woman-ERG
wurrbi-bajun-du buudi-lgani-y
big-very-ERG take-CONT-UNMARKED.TENSE
malan-gu
river-ALL
the very big woman/women is/are taking the
children to the creek
We see that a noun can be marked for number
in Warrgamay, as in yibi-yibi child. However, this
is not required; forms like nulmburungu woman/
women are quite normal. Dixon (1980: 267) stated
that a noun in this language is not normally specified
for number and suggested that this is the typical
situation in Australia (1980: 22). Warrgamay is
particularly interesting, since in (5) the verb does
not determine number either. As for pronouns, the
first and second persons (singular, dual, and plural)
and the third dual and plural are strictly specified for
number and are used only for reference to humans,
and occasionally tame dogs (Dixon, 1981: 3940).
The form filling the third singular slot can range
over all persons and all numbers (it can have nonhu-
man as well as human reference) but its unmarked
sense (1981: 40) is third-person singular.
Thus, in Warrgamay, the word for woman is not
normally specified for number, while in English it
must be. Yet the first and second persons are. This is
part of a wider pattern. The observation that the sets
of nominals involved in number distinctions are
related to animacy was taken up and investigated
further by Smith-Stark (1974), who proposed the
version of the animacy hierarchy shown in Figure 1.
Smith-Stark suggested that plurality splits a lan-
guage if it is a significant opposition for certain
categories but irrelevant for others (1974: 657).
The type of evidence he produced concerned mark-
ing of the noun phrase for number (usually by marking
on the noun itself) and agreement in number
(mainly verbal agreement but with some instances of
agreement within the noun phrase).
As postulating a typological hierarchy implies, var-
ious languages make the split at different points. An
726 Number
Figure 1 The Smith-Stark (animacy) hierarchy.
example of a split between pronouns and nouns is
found in Bengali, since according to Masica (1991:
225226) number is obligatory for pronouns, while
other plural suffixes are optional. In Gunwinggu, a
Gunwingguan language of western Arnhem Land,
pronominal affixes on the verb mark number for
humans but not normally for nonhumans (Evans,
2003: 234235, 417418). A lower split is found in
agreement in Mundari (a Munda language of east
India); verbs agree in number with nominals on the
hierarchy as far down as animate nouns, but not with
inanimates (Bhattacharya, 1976: 191192).
Interesting information on which nouns have a
plural, and whether its use is obligatory, can be
found in Haspelmaths (2005) map for the World
atlas of language structures. His data, from a sample
of 290 languages, are in accord with the animacy
hierarchy. In addition, he mapped the distribution
of the different types. Obligatory number marking
for different types of noun (the English pattern) is
found in western and northern Eurasia and much of
Africa. Elsewhere, Haspelmaths sample reveals a
more mixed picture. He found optional marking of
number to be common in Southeast and East Asia and
Australia, while a lack of number marking is wide-
spread in New Guinea and Australia. For comparable
data on pronouns see Daniel (2005).
Even in languages that mark number as extensively
as English does, there are usually some nouns which
are off the bottom of the scale, those which do not
enter into number oppositions. In English, they typi-
cally pattern with the singulars, thus reliability has
the form of a singular and takes singular agreements.
On the other hand, in Manam (Lichtenberk, 1983:
269), mass nouns are treated as plural (unless they
refer to a single quantity):
(6) daN di-e no
water 3PL-exist
there is water
(available)
In various Bantu languages we often find both
possibilities: some mass nouns are singular and
some plural.
We have concentrated on the basic singular/plural
opposition. An important goal of the typological in-
vestigation of number is to integrate the typology of
the nominals involved in number with the typology
of values (see Number Values). This is a significant
challenge. Smith-Stark considered only plurals, sug-
gesting that other values, such as the dual, would
follow the plural. This situation is certainly not the
only possibility. For instance, in Avar, some nouns
have a paucal, and in modern Hebrew and in Maltese,
some nouns have a dual (in addition to singular and
plural). The nouns involved are relatively few in com-
parison with those with a plural, and they are not those
at the top of the animacy hierarchy. These paucals and
duals can be analyzed as minor numbers; they do not
pattern according to the animacy hierarchy, but they
are counterexamples of a narrowly definable type.
Then there are other apparent number values that
appear to run counter to the hierarchy; for example,
associatives, as in Bagvalal (Nakh-Daghestanian) jas
girl, daughter, jas-i (normal plural) girls, daughters,
jas-ari (associative plural) daughter and her family
(Danie`l, 2001: 135136). Associatives involve an ad-
ditional category and so fall outside the scope of
the constraints above (Corbett and Mithun, 1996;
Moravcsik, 2003; Moravcsik and Daniel, 2005).
A discussion of these various complications is beyond
the scope of this entry; see Corbett (2000: 5488)
for more detailed data on the animacy hierarchy
and Corbett (2000: 89132) for an integration of
the typology of the nominals involved in number
oppositions with the typology of values.
The Semantics of Number
So long as we stay with innocent examples like
English cat and cats, the semantics of number look
clear, since cats is appropriate for more than one
instance of cat. In fact, there are many concealed
difficulties: a concise overview of work on number
within formal semantics can be found in Link (1998),
who showed how the study of plurals related to dif-
ferent movements in linguistic semantics and
provided a substantial bibliography. Ordinary plurals
like cats are only one possible type of number; if we
consider I versus we, it becomes clear that here
we have an associative use of the plural; we is most
often used for I and other(s) associated with me
(Moravcsik, 2003; Corbett, 2004). These different
uses are bound up with the nominals position on
the animacy hierarchy. Thus, in English, while the
items at the very top of the hierarchy have associative

uses, nouns lower on the hierarchy, like coffee, have a
number opposition provided they are recategorized.
If coffee denotes a type, or a typical quantity of
coffee, then there is an available plural, as in: she
always keeps three different coffees, and weve
ordered two teas and three coffees.
Number Values
A striking way in which English is limited in its num-
ber system is in terms of the number values. It can
distinguish only between singular and plural. Many
languages have more distinctions.
The Dual
A common system is one in which singular and plural
are supplemented by a dual to refer to two distinct
real-world entities. Examples are widespread. We
will take Central Yupik, as spoken in southwestern
Alaska; the number forms are: singular arnaq
woman, dual arnak two women and plural arnat
three or more women. The addition of the dual has
an effect on the plural, which is now used for three or
more real-world entities. More generally, a change in
the system gives the plural a different meaning. As
with the plural, the use of the dual need not be oblig-
atory; Slovenian is a language with a dual, but one
whose use is subject to interesting conditions (see
Derganc, 2003). And in Kxoe (Central Khoisan) the
plural is used of a man and wife; to use the dual
would be an insult, because it would suggest they
were together by chance (Treis, 2000).
The Trial
The trial refers to three distinct real-world entities,
and it occurs in systems with the number values:
singular, dual, trial and plural. A particularly clear
instance is found in Larike (LarikeWakasihu), a
Central Moluccan language spoken on the western
tip of Ambon Island, Indonesia. (Central Moluccan
forms part of the Central MalayoPolynesian sub-
group of Austronesian.) This example is from Laidig
and Laidig (1990: 96):
(7) Matidui-tue au-huse nusa
3TRIAL-live at-there island
those three live on the island
over there
The reason for citing Larike is that it has a genu-
ine trial, which is available for use for exactly three
(Laidig and Laidig, 1990: 92). The dual and trial
forms of Larike can be traced back to the numerals
two and three and the plural comes historically
from four. This is a frequent development in Austro-
nesian languages. Accordingly, some linguists use the
Number 727
term trial more widely, according to the form of the
inflections (derived historically from the numeral
three), even for languages where the forms are used
synchronically for small groups, including those
greater than three (for which paucal would be
more accurate). The trial of Larike is not a paucal,
but strictly a trial according to Laidig and Laidig.
The Paucal
As just noted, paucals are used to refer to a small
number of distinct real world entities, somewhat
akin to the English quantifier a few in meaning.
There is no specific upper numerical limit on the use
of paucals (and their lower boundary varies accord-
ing to the system in which they are found). Let us
return to the Cushitic language Baiso. Besides the
general number forms (as in (3), where number is
not specified), nouns also have singular, paucal and
plural. Here is the paucal (Dick Hayward, personal
communication).
(8) luban-jaa foofe
lion-PAUCAL watched.1SING
I watched a few lions
The paucal is used in Baiso when referring to a
small number of individuals, from two to about six.
The Baiso system (for nouns) of singular/paucal/
plural, also with a general form, is rare. The paucal
is usually found together with a dual, so that the
system of number values is singular/dual/paucal/
plural. This widespread system is found, for example,
in Manam, an Oceanic language spoken on islands
off the north coast of New Guinea (Lichtenberk,
1983: 108109).
The Largest Systems
It is interesting to ask what is the largest system of
number values. There are claims in the literature
for languages with a quadral, for four distinct real-
world entities, as part of a system of five number
values. Systems with five number values have indeed
been found, and these are the largest systems, but
the evidence specifically for a quadral is less certain.
All the claims come from within the Austronesian
family, and the best documented case is Sursurunga
(Hutchisson, 1986, and personal communications),
spoken in southern New Ireland. The forms in ques-
tion are restricted to the personal pronouns, but are
found with all of them, the first person (inclusive and
exclusive), the second and the third. Given the con-
clusion to be reached, we shall label the forms greater
paucal (Table 1).
Here is an example of what we shall call a greater
paucal form (Hutchisson, 1986; and personal
communications).
728 Number
Table 1 Emphatic pronouns in Sursurunga (Based on Hutchisson, 1986, and personal communications)
Singular Dual
1 exclusive iau giur
1 inclusive gitar
2 iau gaur
3 -i/on/ai diar
Note that a is used to indicate schwa (e).
(9) gim-hat ka wa n
1EXCL- maternal.uncle:nephew/
GREATER_PAUCAL niece
we four who are in an unclenephew/niece
relationship
The greater paucal has two further uses. First, Sur-
surunga plural pronouns are never used with terms
for dyads (kinship pairings like unclenephew/niece).
Here the greater paucal is used, for a minimum of
four, and not just for exactly four (Hutchisson, 1986:
10). Second, a speaker may use the first-person-
inclusive greater paucal in hortatory discourse to sug-
gest joint action involving more than four persons,
including the speaker. These two uses account for
the majority of instances of the use of the forms in
question. In terms of meaning, the term quadral is
therefore not appropriate, and so paucal is more
accurate.
Why then have we suggested greater paucal? To
see why, we need to examine the rest of the number
system of Sursurunga (using data and judgments
from Don Hutchisson, personal communications).
The regular use of the dual is quite strictly for two
people. It also has one special use for the singular
when the referent is in a taboo relationship to the
speaker. The so-called trial can be used for three,
and it can also refer to small groups, typically
around three or four, and to nuclear families of any
size. It is, therefore, not strictly a trial, as discussed
earlier: it could be glossed as a few and it should be
labelled paucal. The development of trial into a
paucal is a frequent one, as we noted above. The
so-called quadral, as we have just seen, is primarily
used in different ways with larger groups, of four or
more (an appropriate gloss would be several). This
too would qualify as a paucal. There are therefore
two paucals in Sursurunga, a paucal (traditionally
trial) and a greater paucal (traditionally quadral).
The plural is for numbers of entities larger than what
is covered by the greater paucal; however, there is no
strict dividing line. Using semantic labels, we can say
that the number feature in Sursurunga has the
values: singular, dual, paucal, greater paucal, and
plural. Thus, Sursurunga has a well-documented
five-value number feature. There are certainly other
Paucal Greater paucal Plural
gimtul gimhat gim
gittul githat git
gamtul gamhat gam
ditul dihat di
languages with five number values, namely Lihir,
Tangga and Marshallese; we do not have such de-
tailed information as for Sursurunga and there is no
certain case of a quadral: it seems that in all cases the
highest value below plural in such systems can be
used as a paucal.
Besides the split in the paucal, we may also find
languages that split the plural, with greater plurals
of different types, as for instance in Arabic (Ojeda,
1992). Greater plurals may imply an excessive number
or else all possible instances of the referent.
Number Mismatches and the Agreement
Hierarchy
Nominal number may be realized directly on the
noun phrase; it may also be marked on the verb
(agreement). In the straightforward cases the values
of number match. There are also many instances
where we find a mismatch, and English is an especial-
ly rich source here. An interesting mismatch involves
nouns which are singular morphologically and (typi-
cally) have a normal plural and yet, when singular,
may take plural agreement, particularly in British
English:
(10) the committee have decided.
Having an agreement choice here is made possible by
the semantics of the noun. However, the likelihood of
examples like (10) being acceptable depends on the
variety of English involved, and speakers from differ-
ent parts of the world have opposing judgments.
Speakers of different varieties were asked to consider
sentences apparently produced by nonnative speakers
of English and to correct them where necessary (data
from Johansson, 1979: 203; Bauer, 1988: 254). The
relevant test sentence was:
(11) The audience were enjoying every minute of the
show.
(Table 2) For each variety, n is the number of respon-
dents, and the other figures are the percentage giving
each response. There is substantial divergence be-
tween the varieties. Most speakers of British English
were happy with the example as given, as were the
 Number 729

Table 2 Judgments on agreement with Committee-Type Table 3 Agreement in written and spoken American English
nouns in three varieties of English (Levin, 2001: 76)

Response (%) Variety Verbs Relative Personal

pronouns pronouns
GB (n 92) US (n 93) NZ (n 102)
n % plural n % plural n % plural
No correction 77.2 5.4 72.5
NYT 3233 3 702 24 1383 32
Was enjoying 15.2 90.3 20.6
LSAC 524 9 43 74 239 94
Other response 7.6 4.3 6.9
NYT New York Times; LSAC Longman Spoken American
GB Great Britain; US United States; NZ New Zealand.
Corpus.

majority of New Zealand English speakers (for fur-
ther comparative data see Levin, 2001; and Corbett,
2006).
Another source of variation is the agreement tar-
get; while Table 2 is concerned with the predicate, it
is known that the relative frequency of plural agree-
ment is different for other agreement targets such as
the personal pronoun. So let us examine the different
agreement targets. In attributive position there is no
choice, only singular is possible:
(12) that committee . . . / *those committee.
In the predicate, as we know, there are two possibi-
lities; both numbers are possible in British English:
(13) The committee has/have agreed to reconvene.
The English relative pronoun does not mark number,
but it controls number agreement in its clause and so
its number can be inferred:
(14) the committee, which has decided . . . / the
committee, who have decided . . .
Both numbers are possible. This is also the case with
the personal pronoun, and here speakers of American
English admit the plural too:
(15) The committee . . . It . . . / They . . .
Thus, with such nouns, for speakers of British
English, syntactic or semantic agreement is possible
for all agreement targets, except attributive modi-
fiers, where only syntactic agreement is acceptable.
Levin (2001: 76) counted examples of these agree-
ments with nouns like committee in American
English, in a corpus from the New York Times
(NYT) and from the Longman Spoken American
Corpus (LSAC) (Table 3).
The picture is remarkably clear. Plural agreement
is more common in spoken rather than written
language. Moreover, the choice depends on the syn-
tactic position. In attributive position we find only
singular agreement. Plural agreement is increasingly
possible in the predicate, the relative pronoun and the
personal pronoun.
These data fit into a much more general pattern
based on the agreement hierarchy, which comprises
four target types (Figure 2). Possible agreement
patterns are constrained as follows:
For any controller that permits alternative agreement
forms, as we move rightwards along the agreement hier-
archy, the likelihood of agreement with greater semantic
justification will increase monotonically (that is, with no
intervening decrease).
The agreements with committee-type nouns con-
form to this typological claim, and there is a good
deal of further evidence, both from number, and from
gender (Corbett, 1991: 225260, forthcoming).
The English examples are well known, but
the phenomenon is found much more widely. For
instance, Samoan has several such nouns, some mor-
phologically simple, like aiga family, and some
derived, like au-puaa group of pigs:
(16) Ua momoe uma le aiga
PERF.V sleep.PL all ART.SING family
the whole family were asleep
(O lou igoa o Peko)
(17) A tetei le
but suddenly.awake.PL ART.SING
au-puaa . . .
COLL-pig
When the pigs suddenly awake . . .
(Moyle)
Note that the article is singular, but the verb is
plural; the use of the verb in the plural is frequent
but not obligatory (Mosel and Hovdhaugen, 1992:
91, 443, details of sources there). The reason for the
potential choice of agreement in these cases is clear.
We have seen that the choice may be influenced by
language variety but it is also constrained by linguis-
tic factors, notably the agreement hierarchy.
Number and Numerals
In the discussion of number distinctions, we left out
any mention of numerals. This is because they can
730 Number
Figure 2 The agreement hierarchy.
have a strangely distorting effect on number use.
There are languages like English, where numerals
above one typically require the plural (two cats,
not *two cat). On the other hand, there are languages
where the indication of number by the numeral is
sufficient, and the noun stands in the singular, as in
Bagvalal and Hungarian, among others. And then
there are languages which have both options, as is
generally true of Slavonic languages. Taking Russian
as our example, odin one takes a singular noun; dva
two, tri three and c etyre four, when themselves in
the nominative, take a noun in the genitive singular (a
situation which has arisen through the loss of an
earlier dual); and pjat five and above, when in the
nominative, take a noun in the genitive plural. Com-
plex numerals govern the noun according to the last
element of the numeral. The numeral phrases that
result from the rules given can in turn produce inter-
esting mismatches when predicate agreement is
involved.
Conclusion
Number is indeed a more complex category than
is generally recognized, and the English system,
although a common one, is at one extreme of the
typological space. We noted that, number may be a
verbal category as well as a nominal one (but the
possibilities are rather more restricted for verbal
number). We also saw that number need not be oblig-
atory, that the range of items that distinguish number
varies dramatically from language to language, that
the semantics of number is affected by the particular
item involved, and that the values of the number
category can be just two or, in various combinations,
up to five. When number is marked by agreement, the
values of controller and target do not always match,
but the patterns of mismatching are constrained.
Complex issues that we have not been able to address
here include various interactions, notably between
the forms of marking of number, the particular nom-
inals that distinguish number in a given language, and
the systems of number values. For these see Corbett
(2000).
Acknowledgments
The support of the Economic and Social Research
Council (UK) under grant RES051270122 is
gratefully acknowledged.
See also: Mass Nouns, Count Nouns, Non-count Nouns:
Philosophical Aspects.
References
Allan K (1980). Nouns and countability. Language 56,
541567.
Bauer L (1988). Number agreement with collective nouns
in New Zealand English. Australian Journal of Linguis-
tics 8, 247259.
Bhattacharya S (1976). Gender in the Munda languages.
In Jenner P N, Thompson L C & Starosta S (eds.)
Oceanic linguistics special publication 13: Austroasiatic
studies I. Honolulu: University Press of Hawaii.
189211.
Corbett G G (1991). Gender. Cambridge: Cambridge
University Press.
Corbett G G (2000). Number. Cambridge: Cambridge
University Press.
Corbett G G (2004). Suppletion in personal pronouns:
theory versus practice, and the place of reproducibility
in typology. Linguistic Typology 8(3).
Corbett G G (2006). Agreement. Cambridge: Cambridge
University Press.
Corbett G G & Hayward R J (1987). Gender and number
in Bayso. Lingua 73, 128.
Corbett G G & Mithun M (1996). Associative forms in a
typology of number systems: evidence from Yupik.
Journal of Linguistics 32, 117.
Danie`l M (2001). Imja suscestvitelnoe. In Kibrik A E (ed.)
Bagvalinskij jazyk: grammatika: Teksty: Slovari. Mos-
cow: Nasledie. 127150. [Co-editors Kazenin K I,
Ljutikova E A & Tatevosov S G.]
Daniel M (2005). Plurality in independent personal pro-
nouns. In Haspelmath M, Dryer M, Gil D & Comrie B
(eds.) World atlas of language structures. Oxford:
Oxford University Press.
Derganc A (2003). The dual in Slovenian. Sprachtypologie
und Universalienforschung 56, 165182.
Dimmendaal G J (2000). Number marking and noun cate-
gorization in Nilo-Saharan languages. Anthropological
Linguistics 42, 214261.
Dixon R M W (1980). The languages of Australia.
Cambridge: Cambridge University Press.
Dixon R M W (1981). Wargamay. In Dixon R M W &
Blake B J (eds.) The handbook of Australian languages:
II: Wargamay, the Mpakwithi dialect of Anguthimri,
Watjarri, Margany and Gunya, Tasmanian. Amsterdam:
John Benjamins. 1144.
Durie Mark (1986). The grammaticization of number as
a verbal category. In Nikiforidou V, VanClay M,
Niepokuj M & Feder D (eds.) Proceedings of the
Twelfth Annual Meeting of the Berkeley Linguistics Soci-
ety: February 1517, 1986. University of California.
Berkeley, California: Berkeley Linguistics Society.
355370.
Evans N (2003). Pacific linguistics 541: Bininj Gun-wok: a
pan-dialectal grammar of Mayali, Kunwinjku and Kune
(2 vols). Canberra: Pacific Linguistics.

Everett D (1986). Piraha. In Derbyshire D C & Pullum G K
(eds.) Handbook of Amazonian languages: 1. Berlin:
Mouton de Gruyter. 200325.
Haspelmath M (2005). Occurrence of nominal plurality.
In Haspelmath M, Dryer M, Gil D & Comrie B (eds.)
World atlas of language structures. Oxford: Oxford
University Press.
Hayward D (1984). Cushitic language studies 2: The
Arbore language: a first investigation: including a vocabu-
lary. Hamburg: Buske.
Hutchisson D (1986). Sursurunga pronouns and the special
uses of quadral number. In Wiesemann W (ed.) Pronomi-
nal systems (continuum 5). Tu bingen: Narr. 217255.
Johansson S (1979). American and British English
grammar: an elicitation experiment. English Studies 60,
195215.
Laidig W D & Laidig C J (1990). Larike pronouns: duals
and trials in a Central Moluccan language. Oceanic
Linguistics 29, 87109.
Levin M (2001). Lund studies in English 103: Agreement
with collective nouns in English. Stockholm: Almqvist
and Wiksell.
Lichtenberk F (1983). Oceanic linguistics special publica-
tion 18: A grammar of Manam. Honolulu: University of
Hawaii Press.
Link G (1998). Ten years of research on plurals-where
do we stand? In Hamm F & Hinrichs E (eds.) Studies
Numbers: Calculation
C Semenza and A Grana`, University of Trieste,
Trieste, Italy
2006 Elsevier Ltd. All rights reserved.
Human beings, as well as animals, seem to be geneti-
cally endowed with some mathematical abilities in-
cluding a fundamental sense of quantity: most
researchers, in fact, now believe that evolution devel-
oped a specialized mechanism for number appreciation
and processing (Gelman and Gallistel, 1978; Meck and
Church, 1983; Fuson, 1988; Wynn, 1990; Dehaene,
1997; Butterworth, 1999; Brannon and Terrace,
2000). This conclusion rests in part on research on
human infants: infants are at least able to discriminate
between small numbers of entities, can enumerate
physical actions, actions sequences, and sounds, and
can match a number of sounds to the same number of
material objects (Wynn, 1998).
Numbers are properties of objects in the external
world that are mentally represented as abstract con-
cepts. The numerosity of a collection of elements is
in fact independent of physical attributes. Most re-
searchers (Meck and Church, 1983; Gelman and
Numbers: Calculation 731
in linguistics and philosophy 69: Plurality and quantifi-
cation. Dordrecht: Kluwer. 1954.
Masica C P (1991). The IndoAryan languages. Cam-
bridge: Cambridge University Press.
Moravcsik E (2003). A semantic analysis of associative
plurals. Studies in Language 27, 469503.
Moravcsik E & Daniel M (2005). Associative plural. In
Haspelmath M, Dryer M, Gil D & Comrie B (eds.) World
atlas of language structures. Oxford: Oxford University
Press.
Mosel U & Hovdhaugen E (1992). Samoan reference gram-
mar. Oslo: Scandinavian University Press.
Ojeda A E (1992). The semantics of number in Arabic. In
Barker C & Dowty D (eds.) The Ohio State University
working papers in linguistics 40: Proceedings of the
Second Conference on Linguistics and Semantic Theory,
held at the Ohio State University May 13, 1992. Colum-
bus: The Ohio State University Department of Linguis-
tics. 303325.
Rice K (1989). Mouton grammar library 5: A grammar of
Slave. Berlin: Mouton de Gruyter.
Smith-Stark T C (1974). The plurality split. In La Galy M W,
Fox A R & Bruck A (eds.) Papers from the Tenth Regional
Meeting, Chicago Linguistic Society, April 1921.
Chicago: Chicago Linguistic Society. 657671.
Treis & Yvonne (2000). NP coordination in Kxoe (Central
Khoisan). Afrikanistische Arbeitspapiere 63, 6392.
Gallistel, 1992) agree that numerosity is mapped
into a memory representation via a accumulator, a
device that sums impulses coming from the elements
of a collection. In long term memory the appreciation
of numerosity is confronted with previous experience.
The internal representation of a number is thought to
be analogical and its precision decreases with numer-
osity. Quantification, i.e., the process of capturing
numerosity and accessing to the correspondent men-
tal representation, is performed via three basic tasks:
subitizing, counting, and estimation.
Subitizing refers to the ability to correctly ap-
preciate small quantities, apparently without serial
processing. The idea that subitizing differs from
counting, not unanimously accepted, rests mainly on
the finding that reaction times to number apprecia-
tion are the same for quantities inferior to four ele-
ments and increase exponentially thereafter, where
counting or estimation become necessary.
Counting consists of pairing each element of a
collection to a number word and considering the
last assigned word as representing the numerosity of
the collection: this latter property is called cardinal-
ity The concept of ordinality acquired at the age of

Everett D (1986). Piraha. In Derbyshire D C & Pullum G K
(eds.) Handbook of Amazonian languages: 1. Berlin:
Mouton de Gruyter. 200325.
Haspelmath M (2005). Occurrence of nominal plurality.
In Haspelmath M, Dryer M, Gil D & Comrie B (eds.)
World atlas of language structures. Oxford: Oxford
University Press.
Hayward D (1984). Cushitic language studies 2: The
Arbore language: a first investigation: including a vocabu-
lary. Hamburg: Buske.
Hutchisson D (1986). Sursurunga pronouns and the special
uses of quadral number. In Wiesemann W (ed.) Pronomi-
nal systems (continuum 5). Tubingen: Narr. 217255.
Johansson S (1979). American and British English
grammar: an elicitation experiment. English Studies 60,
195215.
Laidig W D & Laidig C J (1990). Larike pronouns: duals
and trials in a Central Moluccan language. Oceanic
Linguistics 29, 87109.
Levin M (2001). Lund studies in English 103: Agreement
with collective nouns in English. Stockholm: Almqvist
and Wiksell.
Lichtenberk F (1983). Oceanic linguistics special publica-
tion 18: A grammar of Manam. Honolulu: University of
Hawaii Press.
Link G (1998). Ten years of research on plurals-where
do we stand? In Hamm F & Hinrichs E (eds.) Studies
Numbers: Calculation
C Semenza and A Grana`, University of Trieste,
Trieste, Italy
2006 Elsevier Ltd. All rights reserved.
Human beings, as well as animals, seem to be geneti-
cally endowed with some mathematical abilities in-
cluding a fundamental sense of quantity: most
researchers, in fact, now believe that evolution devel-
oped a specialized mechanism for number appreciation
and processing (Gelman and Gallistel, 1978; Meck and
Church, 1983; Fuson, 1988; Wynn, 1990; Dehaene,
1997; Butterworth, 1999; Brannon and Terrace,
2000). This conclusion rests in part on research on
human infants: infants are at least able to discriminate
between small numbers of entities, can enumerate
physical actions, actions sequences, and sounds, and
can match a number of sounds to the same number of
material objects (Wynn, 1998).
Numbers are properties of objects in the external
world that are mentally represented as abstract con-
cepts. The numerosity of a collection of elements is
in fact independent of physical attributes. Most re-
searchers (Meck and Church, 1983; Gelman and
Numbers: Calculation 731
in linguistics and philosophy 69: Plurality and quantifi-
cation. Dordrecht: Kluwer. 1954.
Masica C P (1991). The IndoAryan languages. Cam-
bridge: Cambridge University Press.
Moravcsik E (2003). A semantic analysis of associative
plurals. Studies in Language 27, 469503.
Moravcsik E & Daniel M (2005). Associative plural. In
Haspelmath M, Dryer M, Gil D & Comrie B (eds.) World
atlas of language structures. Oxford: Oxford University
Press.
Mosel U & Hovdhaugen E (1992). Samoan reference gram-
mar. Oslo: Scandinavian University Press.
Ojeda A E (1992). The semantics of number in Arabic. In
Barker C & Dowty D (eds.) The Ohio State University
working papers in linguistics 40: Proceedings of the
Second Conference on Linguistics and Semantic Theory,
held at the Ohio State University May 13, 1992. Colum-
bus: The Ohio State University Department of Linguis-
tics. 303325.
Rice K (1989). Mouton grammar library 5: A grammar of
Slave. Berlin: Mouton de Gruyter.
Smith-Stark T C (1974). The plurality split. In La Galy M W,
Fox A R & Bruck A (eds.) Papers from the Tenth Regional
Meeting, Chicago Linguistic Society, April 1921.
Chicago: Chicago Linguistic Society. 657671.
Treis & Yvonne (2000). NP coordination in Kxoe (Central
Khoisan). Afrikanistische Arbeitspapiere 63, 6392.
Gallistel, 1992) agree that numerosity is mapped
into a memory representation via a accumulator, a
device that sums impulses coming from the elements
of a collection. In long term memory the appreciation
of numerosity is confronted with previous experience.
The internal representation of a number is thought to
be analogical and its precision decreases with numer-
osity. Quantification, i.e., the process of capturing
numerosity and accessing to the correspondent men-
tal representation, is performed via three basic tasks:
subitizing, counting, and estimation.
Subitizing refers to the ability to correctly ap-
preciate small quantities, apparently without serial
processing. The idea that subitizing differs from
counting, not unanimously accepted, rests mainly on
the finding that reaction times to number apprecia-
tion are the same for quantities inferior to four ele-
ments and increase exponentially thereafter, where
counting or estimation become necessary.
Counting consists of pairing each element of a
collection to a number word and considering the
last assigned word as representing the numerosity of
the collection: this latter property is called cardinal-
ity The concept of ordinality acquired at the age of
732 Numbers: Calculation
about 30 months, refers to the ability to judge which
of two different quantities is larger.
According to Gelman and Gallistel (1978, 1992),
cardinality is part of innate mathematical knowledge.
Such knowledge would also include realizing that:
(a) each element must correspond to one verbal label
only; (b) a conventional order rules verbal labels;
(c) counting applies to any collection of objects; and
(d) changing the labeling order does not modify the
numerosity of the collection. A separate ability
involves separating already counted objects from
those that remain to be counted. Not all authors
believe that counting rests on innate knowledge: the
alternative view (Wynn, 1998) is that children deduce
principles of counting from experience.
Estimation is a quick but less accurate process that
applies also in the manipulation of internal represen-
tations. Several variables, besides objective numeros-
ity, have been shown to influence estimation: the
physical properties of the stimuli, their spatial dispo-
sition, practice in the tasks, visual clarity, etc. Estima-
tion, however, heavily depends on psycho-physic
laws, like Fechners and Webers, that also affect inter-
nal representations (Krueger, 1984) like the number
line a content of imagery that, in the majority of
subjects, is thought to support number understanding.
Number processing and calculation have a com-
plex relation with processing of language. As a conse-
quence, neuropsychological disorders affecting the
two domains, acalculia and aphasia, are often over-
lapping and sometimes intimately intertwined. Both
linguistic and mathematical functions are largely sus-
tained, in right-handed subjects, by the left hemi-
sphere, whose lesions often feature aphasia as well
as acalculia (in a series of 50 left hemisphere patients
studied by Delazer et al., 1999, none of the 37 apha-
sics was error free in calculation). This connection
may not be the result of chance evolutionary events.
Semenza et al. (2004) recently showed that a series of
patients suffering from aphasia as a result of right
rather than left hemispheric lesions were nonetheless
all affected by some form of acalculia. This finding
lead to the speculation that linguistic and mathemati-
cal faculties, no matter how anatomically and func-
tionally independent from each other, developed from
the one hemisphere endowed with the recursive fac-
ulty, a combinatory capacity at the basis of both
faculties (Hauser et al., 2002).
This said, one should be cautious in drawing firm
conclusions about a stronger relationship between
representation and processing of language and
math. First, statistical associations between deficits
do not allow to infer causal relationships between
cognitive functions: associations may stem from com-
mon underlying resources (such as attention) or just
because two functions have neighboring neuroana-
tomical implementations. Second, empirical evidence
argues against an obligatory verbal basis of calcu-
lation processes: indeed reliable double dissocia-
tions between calculation and language have been
reported. Thus global aphasics have been described
still demonstrating remarkable arithmetical skills,
while severe acalculics have been observed with no
detectable language impairment. The observation, in
particular, of a nonaphasic patient who entirely lost
her knowledge of numbers and number concepts
above the numeral four was held as evidence that
numbers constitute an independent semantic do-
main. The same sort of double dissociation may be
found when written rather than oral language skills
are considered. In the same way, neuropsycho-
logical dissociations have been reported of math
faculties with general cognitive skills: acalculic
patients can still show a higher than average I.Q.,
while, in contrast, patients with very poor intelligence
(including idiot savants) may yet be able to solve
complex arithmetical problems.
The deduction here is that while it is possible that
some subcomponents impinge on a common basic
function, the independence between mathematical
skills and other cognitive skills, including language,
seems proved to a large extent. This fact is reflected
in neuropsychological literature where, since the very
early studies on acalculia, the intuitive distinction
has been commonplace between primary acalculia
and acalculia secondary to disorders (pivotally
alexia, agraphia, and spatial deficits) interfering
with peripheral aspects of number and calculation
processing.
Cognitive analyses of neuropsychological deficits,
undertaken in the past three decades, allow nowadays
extremely fine-grained distinctions among different
processing and representational subcomponents of
both the number system and calculation. The anatomi-
cal correlation of functions has also made some
progress from the original, still basically correct,
idea, that, with the exception of some spatial compo-
nents located in the right hemisphere, most math
skills are located in the left parietal lobe. One cogni-
tive model, the triple code model (Dehaene and
Cohen, 1995), is based on recently discovered anatomi-
cal correlations. The model, while assuming that
Arabic and magnitude representation of numbers
are available to both hemispheres in the vicinity of
the parieto-occipito-temporal junction (but extending
to the occipito-ventral on the left), assigns the verbal
representation underlying the retrieval of arithmetical
facts (i.e., single digit problems, like tables, that are,
in learned calculators, retrieved from long term mem-
ory rather than calculated) exclusively to the left

hemisphere; only the left hemisphere would possess
the representation of the sequence of words corre-
sponding to verbal numerals and lexical mechanisms
for identifying and producing spoken numerals; these
lexical mechanisms, not specific for numbers, would
be implemented in the classic language (Brocas
and Wernickes) areas, including the inferior frontal
and the superior and middle temporal gyri, as well as
basal ganglia and thalamic nuclei. Within the left
parietal lobe, three major circuits for number proces-
sing were recently identified (Dehaene et al., 2003): a
core system, called into action whenever numbers
are manipulated and quantity processing is required,
would be located in the horizontal segment of the
intra-parietal sulcus; a left angular gyrus area, in
connection with other left hemisphere areas, would
support the retrieval of arithmetic facts, and a bilat-
eral posterior superior parietal system would sup-
port attention on spatial dimensions (including the
number line).
Number processing is now considered, on the basis
of evidence brought about by cases of double dissoci-
ation, to be independent from calculation. Processes
dealing with numbers are very often transcoding
processes, that transform a number from a given
representational format into another format. Thus,
in the contemporary Western culture (other codes,
e.g., the Roman code, have been used or are used in
other cultures), a task like reading aloud presupposes
transcoding Arabic or alphabetically written numbers
into spoken number words; in writing on dictation
the reverse transcoding process is at work. Other
common transcoding tasks are, for example, number
repetition or number transformation from one code
into the other (e.g., three into 3 or vice versa). How
does the cognitive system perform such tasks?
Patients committing revealing patterns of errors in
these tasks have been described. The most interesting
findings concern two main types of transcoding
errors: lexical and syntactic errors (Deloche and
Seron, 1982). These errors most often cooccur but
are sometimes observed in relative isolation from
each other, reflecting independent lexical and
syntactic processing mechanisms.
The number lexicon constitutes the collection
of the symbols and of simple number words. The
number syntax, shown, on the basis of neuropsycho-
logical dissociations, to be independent from lan-
guage syntax, defines the set of acceptable structures
formed by combining the basic lexical elements. A se-
mantic component in the system allows attribution
of meaning to each sequence.
The verbal numeral system is composed of different
lexical primitives (i.e., words) organized in four dif-
ferent classes: the Units or Ones (from one to nine),
Numbers: Calculation 733
the Teens (from eleven to nineteen), the Decades
(from ten to ninety), and the Multipliers. These last
includes words like hundred, thousand, or million,
which according to their position in a sequence are
in sum (e.g., hundred and four [100 4]) or are in a
product relationship (e.g., four hundred [4  100])
with the lexical primitives. The Arabic numeral sys-
tem has a limited lexicon that includes the digits from
1 to 9, whose position within numerals determines
the power of 10 by which it must be multiplied (e.g.,
487 has the value of 4  102 8  101 7  100) and
the digit 0, that indicates the absence of a given power
in the base of the number (e.g., in 407, 0 stands for
0 decades). Zero has a unique role: it never appears
in the leftmost position, where it would not make a
well-formed integer number.
The lexical processes are thus retrieved on the
basis of knowledge of the correct number class and
of the position within the class. Lexical errors con-
sist of the incorrect production of one or more of
the individual elements in a number (e.g., 5 instead
of 6, 45 instead of 55, 1305 instead of 1405).
The combination of lexical primitives leads to the
formation of complex numerals. Complex numerals
can be conceived as multiplicative-additive structures
that can be described as sequences of embeddings of
multiplicative and additive relationships. Thus every
nonprimitive number is build up by the application of
syntactic rules (Power and Dal Martello, 1990): the
concatenation rule and the overwriting rule. The
concatenation rule is applied when the multiplier
(e.g., hundred) is in a product relation with the primi-
tives (e.g., five hundred [5  100]), while the over-
writing rule is applied when the multiplier is in a
sum relation (e.g., hundred and five [100 5]) with
the primitive numbers. Evidence of these distinct
mechanisms comes from both developmental and
neuropsychological studies. The overwriting rule is
generally more difficult to apply, and in some cases
are the almost exclusive type of error, but cases have
been described with relatively more errors involving
the concatenation rule.
Syntactic errors are thus errors whereby the power
of 10 is wrong with respect to the target and are often
violations of the concatenation (e.g., 610043 instead
of 643) or of the overwriting rules (60043 instead
of 643); deletions of one or more zeros also occur
(46 instead of 406) or deletions of a multiplier
(1305 instead of 5305). A recently described pa-
tient (Grana` et al., 2003) directed the attention of
researchers to an important distinction, that be-
tween lexical zeros and syntactic zeros, the errors the
patient made were exclusively on syntactic zeros.
Lexical zeros are semantically derived zeros, like
those in tens (e.g., the zero in 20, and the first zero
734 Numbers: Calculation
in 20104), that originate from a numerical concept,
while syntactic zeros are syntactically produced as the
result of syntactic rules, a concatenation rule plus an
overwriting rule, leaving one or more intermediary
zeros (e.g., the zeros in 12004). Syntactic zeros are
thus more difficult to manipulate.
Number manipulation deficits can selectively con-
cern both production and comprehension and
the Arabic or the verbal code. Patients have been
reported with selective deficits for particular numbers
such as 7 or 9. One still-debated point is whether
or not number transcoding and calculation necessari-
ly imply semantic mediation via a central abstract
semantic representation and whether asemantic,
code-specific routes exist besides the semantic one
(different models have been proposed on this sub-
ject: McCloskey, 1992; Campbell and Clark, 1992;
Cipolotti and Butterworth, 1995; Seron and Noel,
1995; for reviews see Noel, 2001 and Cipolotti and
van Harskamp, 2001).
The nature of the semantic representation is also
controversial. While, for instance, McCloskey (1992)
assumed a precise representation even for large num-
bers, Dehaene and Cohen (1995) hypothesized an
analogue magnitude representation coding for ap-
proximate quantity. The existence of this last repre-
sentation and of a spatial mental number line would
be documented by the SNARC (Spatial-Numerical
Association of Response Codes) effect (Dehaene
et al., 1993), whereby, within a given interval, smaller
numbers tend to be preferentially responded to with
the left hand, whereas the reverse is true for large
numbers. This effect has been shown to hold at least
with one-digit numbers, and in speakers of languages
that are written from the left to the right, whereas
the effect is inverted in speakers of languages with
right-to-left writing.
Ordinality and cardinality were also shown to dis-
sociate in patients. A selective impairment with num-
ber cardinality has in fact been reported: this patient
could correctly order any sequence of numbers, while
unable to add 1 to a given number!
With respect to the number system, the calcu-
lation system is relatively less studied. However,
with a major contribution from neuropsychology,
the relatively independent domains of arithmetical
signs, arithmetical number facts and rules, calcula-
tion procedures, approximate calculation, and con-
ceptual knowledge have been investigated to a large
extent.
An independent system for arithmetical signs is sug-
gested by the relatively rare finding of patients unable
to understand and name arithmetical signs and
performing otherwise correct calculation according
to their misidentification.
An independently stored system containing over-
learned, quickly retrievable single digit problems
(arithmetical facts) is also thought to be acting in
calculation (Ashcraft, 1992). This system is subject
to the widely investigated problem-size effect, which
refers to the observation that difficulty of simple
arithmetic combinations generally increases with
numerical size. Arithmetical facts, vis-a`-vis neuropsy-
chological evidence, seem to be segregated accord-
ing to the type of operation. Thus patients are there,
for example, with selective disturbances of table facts
only. This segregation appears also at the level of more
complex operations (Cipolotti and van Harskamp,
2001). Disturbances of arithmetical facts can occur
both in the memory storage and in the retrieval.
Arithmetical rules are distinct from facts and in-
clude problems where one of the elements (e.g., the
multiplier or the addend) can be substituted by any
number (for example, n 1, n  1, n  1, n/1, n 0,
n  0, n  0). Rules involving zero are particularly
difficult to learn and can be selectively disturbed.
Arithmetical procedures, concerning the algo-
rithms necessary for each type of operation, are
also stored independently from each other. As a con-
sequence, acquired disturbances of procedures consti-
tute a rather heterogeneous picture and selective
deficits of multiplication or subtraction only have
been described. Work on normal learners (Van
Lehn, 1986) highlighted systematic, consistent errors
known as bugs at certain points of the proce-
dure. Semenza et al. (1997) on the basis of selective
cases, also distinguished memory problems for the
procedural algorithm (typically resulting in systemat-
ic misapplication of some specific rule) and monitor-
ing problems, deriving from the inability to devote
attention and keep track of each specific step in the
procedure. Thus, while memory patients (Girelli and
Delazer, 1996) commit, just like normals, errors of
the bug type, monitoring patients, while always start-
ing the operation in the correct way, would then make
an increasing and inconsistent number of errors of
all sorts, have little awareness about the precision of
the performance and end the operation with difficul-
ty. It has been recently suggested the possibility that
selective right hemisphere lesions might compromise
the ability to use a spatially based schema that, in
complex operations, would help in keeping the single
substeps together by representing the information of
where exactly each substep should be placed. This
theory is argued to be in line with proposals made
by psychologists in the early decades of last century
(see Hartje, 1987 for a review).
While the application of procedural knowledge
does not require conceptual understanding of each
step of the procedure and is applied only in already

familiar tasks, what has been called conceptual
knowledge (Hittmair-Delazer et al., 1994), in con-
trast, implies understanding of arithmetical opera-
tions, and the laws and principles pertaining to
these operations. Also labeled adaptive expertise,
it may or may not be explicitly known as such to
the calculator, it allows one to make inferences
and can be flexibly adapted to new tasks. Examples
of this sort of knowledge are potentially infinite.
Delazers et al. (2003) battery for the evaluation
of mathematical deficits, for example, includes the
following principles, allowing a calculator to answer
the question given X, what is the result of Y?: the
commutativity principle (e.g., X: 22  31 682;
Y: 31  22 ?), repeated addition (12  4 48;
12 12 12 12 ?), 10a  10b (45  8 360;
450  80 ?), a1  b (94  5 470; 93  5 ?),
multiplication/division invariance (71  9 639;
639/9 ?). Such knowledge is independent from
other mathematical notions and abilities. Hittmair-
Delazer and coworkers described patients unable to
retrieve number facts who brightly used their residual
knowledge to overcome their difficulties with tables
(e.g., 7  8 ?; 7  10 70; 7  2 14; 70  14 56),
or, while unable to solve facts as simple as 3 2 5,
had no difficulty with algebraic expressions such
as (ac bc)/c a b! Conversely, there are patients
who flawlessly retrieve arithmetical facts while
showing no understanding of mathematical concepts
whatsoever.
Bibliography
Ashcraft M H (1992). Cognitive arithmetic: a review of
data and theory. Cognition 44, 75106.
Brannon E M & Terrace H S (2000). Representation of
the Numerosities 19 by Rhesus Macaques (Macaca
mulatta). Journal of Experimental Psychology: Animal
Behavior Processes 26(1), 3149.
Butterworth B (1999). The mathematical brain. London:
Macmillan.
Campbell J I D & Clark J M (1992). Cognitive number
processing: an encoding-complex perspective. In
Campbell J I D (ed.) The nature and origins of mathemat-
ical skills. Amsterdam: Elsevier. 457491.
Cipolotti L & Butterworth B (1995). Toward a multi-
route model of number processing: impaired number
transcoding with preserved calculation skills. Journal of
Experimental Psychology: General 124(4), 375390.
Cipolotti L & van Harskamp N (2001). Disturbances of
number processing and calculation. In Berndt R S (ed.)
The handbook of neuropsychology. Amsterdam: Elsevier.
305334.
Dehaene S (1997). The number sense: how the mind creates
mathematics. New York: Oxford Press.
Numbers: Calculation 735
Dehaene S & Cohen L (1995). Towards an anatomical and
functional model of number processing. Mathematical
Cognition 1, 83120.
Dehaene S, Bossini S & Giraux P (1993). The mental repre-
sentation of parity and numerical magnitude. Journal of
Experimental Psychology: General 122(3), 371396.
Dehaene S, Piazza M, Pinel P et al. (2003). Three
parietal circuits for number processing. Cognitive
Neuropsychology 20(36), 487506.
Delazer M, Girelli L, Grana` A et al. (2003). Number
processing and calculation: normative data from healthy
adults. Clinical Neuropsychologist 17(3), 331350.
Delazer M, Girelli L, Semenza C et al. (1999). Numerical
skills and aphasia. Journal of the International Neuro-
psychological Society 5(3), 213221.
Deloche G & Seron X (1982). From one to 1: an analysis of
a transcoding process by means of neuropsychological
data. Cognition 12(2), 119149.
Fuson K C (1988). Childrens counting and concepts of
number. New York: Springer-Verlag Publishing.
Gelman R & Gallistel C R (1978). The childs understand-
ing of number. Oxford, England: Harvard University
Press.
Gelman R & Gallistel C R (1992). Preverbal and verbal
counting and computation. Cognition 44, 4374.
Girelli L & Delazer M (1996). Subtraction bugs in an
acalculic patient. Cortex 32(3), 547555.
Grana` A, Lochy A, Girelli L et al. (2003). Transcoding
zeros within complex numerals. Neuropsychologia
41(12), 16111618.
Hartje W (1987). The effect of spatial disorders on arith-
metical skills. In Deloche G & Seron X (eds.) Mathemat-
ical disabilities. Hillsdale, NJ: Lawrence Erlbaum
Associates. 121125.
Hauser M D, Chomsky N & Fitch W T (2002). The faculty
of language: what is it, who has it and how did it evolve?
Science 298, 15691579.
Hittmair-Delazer M, Sailer U & Benke T (1995). Impaired
arithmetic facts but intact conceptual knowledge: a
single-case study of dyscalculia. Cortex 31(1), 139147.
Hittmair-Delazer M, Semenza C & Denes G (1994). Con-
cepts and facts in calculation. Brain 117(4), 715728.
Krueger L E (1984). Perceived numerosity: a comparison of
magnitude production, magnitude estimation, and dis-
crimination judgments. Perception and Psychophysics
35(6), 536542.
McCloskey M (1992). Cognitive mechanisms in
numerical processing: evidence from acquired dyscalcu-
lia. Cognition 44, 107157.
Meck W H & Church R M (1983). A mode control model
of counting and timing processes. Journal of Experi-
mental Psychology: Animal Behavior Processes 9(3),
320334.
Noe l M P (2001). Numerical Cognition. In Rapp B
(ed.) The handbook of cognitive neuropsychology.
Philadelphia: Psychology Press. 495518.
Power R J D & Dal Martello M F (1990). The dictation of
Italian numerals. Language and Cognitive Processes 5,
237254.
736 Numbers: Calculation
Semenza C, Miceli L & Girelli L (1997). A deficit for
arithmetical procedures: lack of knowledge or a lack of
monitoring? Cortex 33, 483498.
Semenza C, Delazer M, Bartha L et al. (2004). Mathemat-
ics in right hemisphere aphasia: a case series study. Brain
and Language 91, 164165.
Seron X & Noe l M P (1995). Transcoding numbers from
the Arabic code to the verbal one or vice versa: how many
routes? Mathematical Cognition 1, 215245.
Numerals
J Gvozdanovic, Universitat Heidelberg, Heidelberg,
Germany
2006 Elsevier Ltd. All rights reserved.
Numeral Systems: Their Structure and
Development
Quantification is essential for our cognition of the
world. To quantify is to establish the limits of an
entity in a relative way. Each entity is defined by its
limits and identified by the relativity of its limits. It is
discrete. For example, a piece of sugar is a discrete
entity, which can have further spatial properties not
attributable to its source, the amorphous mass of
sugar.
Discreteness emerges from quantification. Discrete
entities can be grouped together so as to form sets.
Sets can be homogeneous or heterogeneous; they can
contain many entities or be empty, yet they always
have clear limits. By having clear limits, a set is an
entity, too, an entity of a higher order.
Set formation presupposes entities, themselves a
cognitive product of either quantification or (if the
entities have already been quantified) identification.
The primary act of set formation involves abstraction
from all but a classifying property. An example is
found in tally sticks with notches on bones, which
originated in Africa around 37 000 years B.C. and
in Europe some millennia later. The notches repre-
sent quantified entities; the tally stick contains a set
(Figure 1).
Counting is based on classification, accumulation,
and grouping. Cognitive units of counting are numbers;
their symbolic expressions in language are numerals.
Figure 1 African tally stick (37 000 B.C.), probably representing
the moon cycle (29 notches).
Van Lehn K (1986). Arithmetic procedures are
induced from examples. In Hiebert J (ed.) Conceptual
and procedural knowledge: the case of mathematics.
Hillsdale, NJ: Lawrence Erlbaum Associates. 133179.
Wynn K (1990). Childrens understanding of counting.
Cognition 36(2), 155193.
Wynn K (1998). Numerical competence in infants. In
Donlan C (ed.) Development of mathematical skills.
Hove, England: Psychology Press/Taylor & Francis. 325.
Grouping of entities is an essential feature of
number systems, with the effect of establishing bases
(in addition to digits such as 1). Across cultures,
primitive counting on fingers and toes led to the
establishment of the group of 5 symbolized by a
hand as a basis for further counting. Significantly,
counting on fingers and toes operates with the base
of 5 (i.e., five fingers or a hand), the base of 10
(i.e., two hands), and the base of 20 (i.e., fingers and
toes or hands and feet). Contemporary examples in
active use are found in Papua New Guinea, where,
e.g., the Aghu language symbolizes 5 by the word
for hand, whereas Kombai, Korowai, and Wambon
symbolize 5 by the word for thumb. Similar systems
are found, e.g., in Mullukmulluk in Australia (cf.
Greenberg, 1978: 257) and in Bantawa in east Nepal
(where the numeral for 5 is u. kchuk one hand, 10
is hu. achuk two hands, and 20 is re kachuk four
hands; cf. Gvozdanovic , 1985: 137). Many systems
known as advanced started from a quinary base, for
example, Sumerian from the 4th millennium B.C. in
Mesopotamia (Table 1).
Comparative studies of Indo-European have shown
that Proto-Indo-European (PIE) words for 5 and 10
are also based on counting by hand. Thus, PIE
*penkw(e) 5 is connected, e.g., with Old
High German (OHG) fu st and Russian pjast fist,
Table 1 Sumerian numerals above 5 composed with the
base 5a
Number Sumerian Number Sumerian numeral
numeral
1 dis, as 6 as < *ia as (5 1)
2 min 7 imin < *ia min (5 2)
3 es 8 ussu < *ia es (5 3)
4 limmu 9 ilimmu < *ia limmu
(5 4)
5 ia 10 u
a
Cf. Justus (1996).
736 Numbers: Calculation
Semenza C, Miceli L & Girelli L (1997). A deficit for
arithmetical procedures: lack of knowledge or a lack of
monitoring? Cortex 33, 483498.
Semenza C, Delazer M, Bartha L et al. (2004). Mathemat-
ics in right hemisphere aphasia: a case series study. Brain
and Language 91, 164165.
Seron X & Noel M P (1995). Transcoding numbers from
the Arabic code to the verbal one or vice versa: how many
routes? Mathematical Cognition 1, 215245.
Numerals
J Gvozdanovic, Universitat Heidelberg, Heidelberg,
Germany
2006 Elsevier Ltd. All rights reserved.
Numeral Systems: Their Structure and
Development
Quantification is essential for our cognition of the
world. To quantify is to establish the limits of an
entity in a relative way. Each entity is defined by its
limits and identified by the relativity of its limits. It is
discrete. For example, a piece of sugar is a discrete
entity, which can have further spatial properties not
attributable to its source, the amorphous mass of
sugar.
Discreteness emerges from quantification. Discrete
entities can be grouped together so as to form sets.
Sets can be homogeneous or heterogeneous; they can
contain many entities or be empty, yet they always
have clear limits. By having clear limits, a set is an
entity, too, an entity of a higher order.
Set formation presupposes entities, themselves a
cognitive product of either quantification or (if the
entities have already been quantified) identification.
The primary act of set formation involves abstraction
from all but a classifying property. An example is
found in tally sticks with notches on bones, which
originated in Africa around 37 000 years B.C. and
in Europe some millennia later. The notches repre-
sent quantified entities; the tally stick contains a set
(Figure 1).
Counting is based on classification, accumulation,
and grouping. Cognitive units of counting are numbers;
their symbolic expressions in language are numerals.
Figure 1 African tally stick (37 000 B.C.), probably representing
the moon cycle (29 notches).
Van Lehn K (1986). Arithmetic procedures are
induced from examples. In Hiebert J (ed.) Conceptual
and procedural knowledge: the case of mathematics.
Hillsdale, NJ: Lawrence Erlbaum Associates. 133179.
Wynn K (1990). Childrens understanding of counting.
Cognition 36(2), 155193.
Wynn K (1998). Numerical competence in infants. In
Donlan C (ed.) Development of mathematical skills.
Hove, England: Psychology Press/Taylor & Francis. 325.
Grouping of entities is an essential feature of
number systems, with the effect of establishing bases
(in addition to digits such as 1). Across cultures,
primitive counting on fingers and toes led to the
establishment of the group of 5 symbolized by a
hand as a basis for further counting. Significantly,
counting on fingers and toes operates with the base
of 5 (i.e., five fingers or a hand), the base of 10
(i.e., two hands), and the base of 20 (i.e., fingers and
toes or hands and feet). Contemporary examples in
active use are found in Papua New Guinea, where,
e.g., the Aghu language symbolizes 5 by the word
for hand, whereas Kombai, Korowai, and Wambon
symbolize 5 by the word for thumb. Similar systems
are found, e.g., in Mullukmulluk in Australia (cf.
Greenberg, 1978: 257) and in Bantawa in east Nepal
(where the numeral for 5 is u. kchuk one hand, 10
is hu. achuk two hands, and 20 is re kachuk four
hands; cf. Gvozdanovic, 1985: 137). Many systems
known as advanced started from a quinary base, for
example, Sumerian from the 4th millennium B.C. in
Mesopotamia (Table 1).
Comparative studies of Indo-European have shown
that Proto-Indo-European (PIE) words for 5 and 10
are also based on counting by hand. Thus, PIE
*penkw(e) 5 is connected, e.g., with Old
High German (OHG) fust and Russian pjast fist,
Table 1 Sumerian numerals above 5 composed with the
base 5a
Number Sumerian Number Sumerian numeral
numeral
1 dis, as 6 as < *ia as (5 1)
2 min 7 imin < *ia min (5 2)
3 es 8 ussu < *ia es (5 3)
4 limmu 9 ilimmu < *ia limmu
(5 4)
5 ia 10 u
a
Cf. Justus (1996).

originally referring to the hand clenched to form a fist
as a symbol of five (cf. Winter, 1992: 17), whereas
PIE *de 10 has been reconstructed as *de-
two hands, and * om < d om 100 as either
the genitive plural of the neuter collective noun
*d - (i.e., * d d om a decade of decades,
cf. Szemere nyi, 1960: 139140), or a derivative with
to-, signifying the limiting point in the series of
decades (cf. Coleman, 1992: 404). Indeed, the base
of 100 was the highest decimal base in Proto-Indo-
European (as the reconstruction of 1000 *ghes-lo-
on the basis of Aryan and Greek fails to account
for the other Indo-European languages; cf. also
Szemere nyi, 1960: 1).
Historically seen, the development of advanced nu-
meral systems was connected with astronomy, geome-
try, and calendar construction, in addition to stock
administration and trade. This applies to the ancient
Egyptian decimal system (since the 5th millennium B.C.)
as well as to the Sumerian and later Akkadian and
Babylonian sexagesimal system (cf., e.g., Neugebauer,
1947: 37 concerning the calculation of lunar phases).
Apart from bases of unrestricted productivity, ru-
dimentary bases have been recorded, for example,
in duodecimal counting of Benue-Congo languages
(involving a base of 12), later replaced by a decimal
system, or in Sora, a Munda language (cf. Greenberg,
1978: 270) (related to it were probably also the
big hundreds transitionally present in Northern
Germanic [equalling 120]). Smaller rudimentary
bases have been attested in rural communities, such
as base 4 in Barbareno (a Chumash language of
California; cf. Comrie, 1999: 89), in Tunisian, or in
Celtic (cf. Justus, 1999: 61 f.). Yuman languages of
Southern California, Arizona, and Baja California,
next to Coahulteco a Hokan language, provide
examples of limited use of the base 3. A rudimentary
base 2 is found in the central Wintun branch of
Californian Penutian (cf. Greenberg, 1978: 279). Ru-
dimentary bases coexist with productive bases, usually
of later origin.
The lowest productive base by which all higher
productive bases can be divided is called fundamental
(cf. Greenberg, 1978: 270). Apart from the rare ex-
ceptional types mentioned above, languages are clas-
sified by their fundamental bases as quinary (with the
fundamental base 5), decimal (with the fundamental
base 10), vigesimal (with the fundamental base 20), *quarante trente [40 30]). The point at which a
or sexagesimal (with the fundamental base 60).
The capacity to undergo recursive mathematical
operations makes bases into the backbone of numer-
al systems. The relevant operations and their asym-
metrical implications (in the sense of Greenbergs
hierarchies (1978: 257272) are the following:
Numerals 737
. addition (e.g., French dix-neuf ten-nine, i.e., 19),
. subtraction (e.g., Latin un-de-viginti one-off-twen-
ty, i.e., 19); it implies the existence of addition
(every minuend is a base of the system or a multiple
of a base; a subtrahend is never larger than the
remainder);
. multiplication (e.g., English fifty [5  10]); it
implies the existence of addition (a rare exception
is found in a subgroup of the Yuman languages,
affiliated to the Hokan stock, which has the first
decade only, analyzed as 1, 2, 3, 4, 5, 3  2,
7, 4  2, 3  3, 10);
. division (e.g., Welsh hanner cunt half hundred,
i.e., 50); it implies the existence of multiplication
(division is expressed as multiplication by a frac-
tion, and the denominator of the fraction is always
2 or a power of 2);
. (rarely) overcounting (e.g., older Danish halv-tred-
sinds-tyve half-third-time-twenty, i.e., 50) and
going-on operation (e.g., in Mayan and in Finno-
Ugric, cf. Ostyak 18 as (8 going-on 20));
. exponentiation (e.g., Sumerian ges 60, sa`r 3600,
sa`r-gal sa`r-big 216,000).
Bases show up as significant boundaries in lan-
guage decay, at least at intermediate stages. This
may be illustrated by the Tibeto-Burman languages
of Nepal (cf. Gvozdanovic, 1985: 136), which have
been overlayered by the (Indo-European) standard
language Nepali, but have preserved their original
numerals either up to the fundamental base 5 (sym-
bolized by one hand or by the Proto-Tibeto-Burman
form ngaji) or up to 3. The observed regularity
demonstrates the cognitive relevance of numeral
bases.
Morphology and Syntax of Numeral
Expressions
Numeral morphosyntax shows a combination of
mathematical and linguistic properties, relating to
semantic, morphosyntactic, and pragmatic levels.
For numeral morphosyntax Hurford (1987) formu-
lated a so-called packing-strategy principle that
states that complex numeral expressions are formed
by combining the highest-valued simpler numerals
available (e.g., the French numeral for 70 is soixante
dix [60 10], not *cinquante-vingt [50 20] or
language changes methods for signaling addition is
as a rule indicative of a base break. It is typical of
Indo-European that such a break occurs between the
-teens (cf. English thirteennineteen, expressed as
digit base) and the upper decades (starting from
20, regionally limited also above 60). Outside of
738 Numerals
Indo-European, pattern discontinuity may be illus-
trated by Hebrew, where the numeral plurals of 3
to 9 denote the corresponding decades 30 to 90,
but the plural of 10 denotes 20. The next-higher
base is 100, for which Hebrew has a separate noun,
me a.
Numeral bases are treated differently from digits.
In systems with more than one base, there is a base
above which certain regularities hold, e.g., complex
expressions consist of a product and a remainder, and
the remainder never has a value larger than the next
lower base of the total expression (cf. French soixante
douze 72, i.e., 60 12 is possible because the next
lower base is 20).
Numeral syntax is usually highly complex, and
only a part of this complexity is directly connected
with numeral morphology. Typologically interesting
is the tendency of the numeral to precede the noun if
the descriptive adjective is also in the preceding posi-
tion, as opposed to variation in the position of the
numeral when the adjective follows the noun. An
example of the latter is found in Malay, in which the
descriptive adjective regularly follows the noun, but
the numeral and the classifier may be placed either
before or after the noun (the latter option seems to be
pragmatically marked).
Numeral morphosyntax exhibits a set of further
regularities:
. the numeral one is syntactically an attribute of its
denominator;
. numerals expressing bases either are syntactic
heads or share headedness with their denominator
(cf. Hebrew, in which these numerals have inherent
gender); in the development of Indo-European, the
lower bases have arguably developed from adjec-
tives to nouns;
. the numerals two, three, and four have an inter-
mediate status; they may differentiate gender; lan-
guages with grammatical dual have a special
treatment of two;
. many languages have discontinuity of expression
around the numerals three and four, relating
especially to distinct marking for gender, case, in-
definiteness, or word order; these phenomena are
presumably related to short-term memory limita-
tions (cf. Cowan, 2001);
. in rule-governed variation between the singular
and the plural with numerals, the singular is fa-
vored with higher numbers, in measure construc-
tions, in indefinite constructions, and with nouns
that are inanimate or impersonal (Greenberg,
1978: 283);
. the order noun-numeral is favored in indefinite and
approximative constructions (Greenberg, 1978:
284);
. predicate agreement with quantified subjects above
the cut-off point around three or four mentioned
above is sensitive to animacy and topicality: the
singular agreement is preferred with inanimate
and non-topical subjects, the plural agreement
with animate and topical subjects.
Although quantification takes effect in the logical
structure of language, its morphology and syntax
depend in part on pragmatics. This shows that quan-
tification permeates all the central levels of language.
See also: Classifiers and Noun Classes: Semantics;
Number.
Bibliography
Coleman R (1992). Italic. In Gvozdanovic J (ed.) Indo-
European numerals. Berlin: Mouton de Gruyter.
Comrie B (1999). Haruai numerals and their implications
for the history and typology of numeral systems. In
Gvozdanovic J (ed.) Numeral types and changes world-
wide. 8194.
Corbett G G (1983). Hierarchies, targets and controllers;
agreement patterns in Slavic. London: Croom Helm.
Corbett G G (1993). The head of Russian nominal expres-
sions. In Corbett G G, Fraser N M & McGlashan S (eds.)
Heads in grammatical theory. Cambridge: Cambridge
University Press. 1135.
Cowan N (2001). The magical number 4 in short-term
memory: A reconsideration of mental storage capacity.
Behavioral and Brain Sciences 24(1), 87114.
Greenberg J H (1978). Generalizations about numeral
systems. In Greenberg J H (ed.) Universals of human
language, vol. 3: Word structure. Stanford: Stanford
University Press. 249295.
Gvozdanovic J (1985). Language system and its change; on
theory and testability. Berlin: Mouton de Gruyter.
Gvozdanovic J (ed.) (1992). Indo-European numerals.
Berlin: Mouton de Gruyter.
Gvozdanovic J (ed.) (1999). Numeral types and changes
world-wide. Berlin: Mouton de Gruyter.
Gvozdanovic J (1999). Types of numeral changes. In
Gvozdanovic J (ed.). 95111.
Halle M (1994). The morphology of numeral phrases. In
Avrutin S, Franks S & Progovac L (eds.) Formal
approaches to [Slavic] linguistics. Ann Arbor: Michigan
Slavic Publications. 176215.
Hurford J R (1975). The linguistic theory of numerals.
Cambridge: Cambridge University Press.
Hurford J R (1987). Language and number; the emergence
of a cognitive system. Oxford: B. Blackwell.

Ifrah G (1981). Histoire Universelle des Chiffres. Paris:
Seghers.
Justus C F (1996). Numeracy and the Germanic upper
decades. Journal of Indo-European Studies 24, 4580.
Justus C F (1999). Pre-decimal structures in counting and
metrology. In Gvozdanovic J (ed.). 5579.
Menninger K (1934). Zahlwort und Ziffer. Breslau: Ferdi-
nand Hirt.
Numerical Notation
J Lawler, University of Michigan, Ann Arbor, MI, USA
2006 Elsevier Ltd. All rights reserved.
Notation is a conventional written system for encod-
ing a formal axiomatic system. Notation governs
. the rules for assignment of written symbols to ele-
ments of the axiomatic system;
. the writing and interpretation rules for well-formed
formulae in the axiomatic system;
. the derived writing and interpretation rules for
representing transformations of formulae, in accor-
dance with the rules of deduction in the axiomatic
system.
All formal systems impose notational conventions on
the forms. Just as in natural language, to some extent
such conventions are matters of style and politics,
even defining group affiliation. Thus, notational con-
ventions display sociolinguistic variation; alternate
conventions are often in competing use, though
there is usually substantial agreement on a classical
core notation taught to neophytes.
Excluding logic (see Logical and Linguistic Nota-
tion), mathematics as a whole is divided into three
parts, of which one is algebra (roughly, the realm of
the discrete, including number theory and combina-
torics), another geometry (roughly, the realm of
the visibly codable, including topology), and the
third what ordinary people call calculus, but which
is spoken of by mathematicians as analysis (roughly,
the realm of the continuous, including statistics, dif-
ferential and integral calculus, theory of functions,
and complex variables). All these areas overlap and
grow into one another fractally; in particular, they
crucially involve the concept of number.
Numbers and Arithmetic Symbols
Though there appear to be some cultures where
counting and numbers are unknown (see Everett,
2005), almost all languages contain some numbers,
Numerical Notation 739
Neugebauer O (1947). Studies in ancient astronomy.
VIII. The water clock in Babylonian astronomy. ISIS 37,
1, 3743.
Szemere nyi O (1960). Studies in the Indo-European system
of numerals. Heidelberg: Winter.
Winter W (1992). Some thoughts about Indo-European
numerals. In Gvozdanovic J (ed.). 1128.
and almost all cultures practice counting and other
mathematical activities. The written representation
of quantities and numbers is as old as, and
may even predate and prefigure (see Schmandt-
Besserat, 1992; Zimansky, 1993), writing itself.
After alphabetic writing was developed, it was
normal to use individual letters to represent particu-
lar numbers in an additive system such as Roman
numerals, where, for instance, MDCCVII represents
1000 500 100 100 5 1 1, or 1707. Simi-
lar systems existed for ancient Greek and Hebrew;
even after the widespread use of Hindu/Arabic
numerals, the Hebrew system of alphabetic number
representation survived as a Kabbalistic numerologi-
cal system known as Gematriya, in which the number
associated with any Hebrew word has gnostic signifi-
cance. See Ifrah (1991: ch. 2) in Barrow (1992),
and Pettersson (1996) for history. The Hindu/Arabic
numerals, with their positional notation, were origi-
nally developed in India during the Gupta empire and
spread to the Muslim world, whence they made their
way to Europe after the Crusades. In positional nota-
tion, the position of each numeral in a series indicates
the power of the base that it enumerates; thus, 1707
in base ten represents (1  103) (7  102)
(0  101) (7  100). Such a notation (including
zero, which is necessary for it to work) represents an
enormous advance in numeration over additive sys-
tems, because it makes possible algorithmic calcula-
tions such as long division.
Hindu/Arabic numerals in base ten are all but uni-
versal in the writing systems of the modern world.
While the decimal (base ten) system is the norm, in
principle any numeric base can be used, and several
have been: the Babylonian numeric system was sexa-
gesimal (base 60, from which we inherit the 360
circle and the 60-minute hour); the Mayan system
was vigesimal (base 20); vestiges of a duodecimal
(base 12) system remain in English vocabulary
(dozen, gross); and several systems based on powers
of 2 binary (base 2, digits 0 and 1), octal (base 23,

Ifrah G (1981). Histoire Universelle des Chiffres. Paris:
Seghers.
Justus C F (1996). Numeracy and the Germanic upper
decades. Journal of Indo-European Studies 24, 4580.
Justus C F (1999). Pre-decimal structures in counting and
metrology. In Gvozdanovic J (ed.). 5579.
Menninger K (1934). Zahlwort und Ziffer. Breslau: Ferdi-
nand Hirt.
Numerical Notation
J Lawler, University of Michigan, Ann Arbor, MI, USA
2006 Elsevier Ltd. All rights reserved.
Notation is a conventional written system for encod-
ing a formal axiomatic system. Notation governs
. the rules for assignment of written symbols to ele-
ments of the axiomatic system;
. the writing and interpretation rules for well-formed
formulae in the axiomatic system;
. the derived writing and interpretation rules for
representing transformations of formulae, in accor-
dance with the rules of deduction in the axiomatic
system.
All formal systems impose notational conventions on
the forms. Just as in natural language, to some extent
such conventions are matters of style and politics,
even defining group affiliation. Thus, notational con-
ventions display sociolinguistic variation; alternate
conventions are often in competing use, though
there is usually substantial agreement on a classical
core notation taught to neophytes.
Excluding logic (see Logical and Linguistic Nota-
tion), mathematics as a whole is divided into three
parts, of which one is algebra (roughly, the realm of
the discrete, including number theory and combina-
torics), another geometry (roughly, the realm of
the visibly codable, including topology), and the
third what ordinary people call calculus, but which
is spoken of by mathematicians as analysis (roughly,
the realm of the continuous, including statistics, dif-
ferential and integral calculus, theory of functions,
and complex variables). All these areas overlap and
grow into one another fractally; in particular, they
crucially involve the concept of number.
Numbers and Arithmetic Symbols
Though there appear to be some cultures where
counting and numbers are unknown (see Everett,
2005), almost all languages contain some numbers,
Numerical Notation 739
Neugebauer O (1947). Studies in ancient astronomy.
VIII. The water clock in Babylonian astronomy. ISIS 37,
1, 3743.
Szemerenyi O (1960). Studies in the Indo-European system
of numerals. Heidelberg: Winter.
Winter W (1992). Some thoughts about Indo-European
numerals. In Gvozdanovic J (ed.). 1128.
and almost all cultures practice counting and other
mathematical activities. The written representation
of quantities and numbers is as old as, and
may even predate and prefigure (see Schmandt-
Besserat, 1992; Zimansky, 1993), writing itself.
After alphabetic writing was developed, it was
normal to use individual letters to represent particu-
lar numbers in an additive system such as Roman
numerals, where, for instance, MDCCVII represents
1000 500 100 100 5 1 1, or 1707. Simi-
lar systems existed for ancient Greek and Hebrew;
even after the widespread use of Hindu/Arabic
numerals, the Hebrew system of alphabetic number
representation survived as a Kabbalistic numerologi-
cal system known as Gematriya, in which the number
associated with any Hebrew word has gnostic signifi-
cance. See Ifrah (1991: ch. 2) in Barrow (1992),
and Pettersson (1996) for history. The Hindu/Arabic
numerals, with their positional notation, were origi-
nally developed in India during the Gupta empire and
spread to the Muslim world, whence they made their
way to Europe after the Crusades. In positional nota-
tion, the position of each numeral in a series indicates
the power of the base that it enumerates; thus, 1707
in base ten represents (1  103) (7  102)
(0  101) (7  100). Such a notation (including
zero, which is necessary for it to work) represents an
enormous advance in numeration over additive sys-
tems, because it makes possible algorithmic calcula-
tions such as long division.
Hindu/Arabic numerals in base ten are all but uni-
versal in the writing systems of the modern world.
While the decimal (base ten) system is the norm, in
principle any numeric base can be used, and several
have been: the Babylonian numeric system was sexa-
gesimal (base 60, from which we inherit the 360
circle and the 60-minute hour); the Mayan system
was vigesimal (base 20); vestiges of a duodecimal
(base 12) system remain in English vocabulary
(dozen, gross); and several systems based on powers
of 2 binary (base 2, digits 0 and 1), octal (base 23,
740 Numerical Notation
digits 07), and hexadecimal (base 24, digits 09 plus
AF) are in common use in computing contexts.
By contrast with the numerals, all of the usual
symbols for arithmetic operations (, , , ) are
of European origin, originating among a congeries
of symbols used between the 15th and 17th centuries.
is simply the stylized letter T of et, an early amper-
sand. The symbol for division (), now seen as an
obvious mnemonic of the fraction notation with a
horizontal bar introduced by the Arabs, was in fact
originally used in Europe for subtraction, and it is
from this sign that the modern  sign comes. Finally,
 is only one of several symbols for multiplication,
introduced and promoted by William Oughtred
(15741660) in the Clavis Mathematicae (1631).
Other multiplication notation in use today includes
the raised dot (a  b), which Leibniz preferred, and
simple juxtaposition (ab), which appears to have
been the original notation in India, dating from the
10th century and appearing sporadically in Arabic
and European sources afterward. By the 18th century,
these symbols had become general in European
mathematical notation.
Other Symbols
Mathematics is a gigantic intellectual area, famously
unapproachable by the uninitiated, replete with mul-
tiple notational conventions and the materials and
will to invent new ones. Considering the size of the
field and the elaborate profusion of its notation, this
article can do no more within its space constraints
than touch on the notations of several areas (besides
logic) that impinge on natural language, and remark
on general notational tendencies in mathematics.
One example is the notation used in modern inte-
gral and differential calculus, which descends from
Leibnizs original invention in the 17th century;
Newton famously invented calculus independently
before Leibniz, but published later. More importantly
for our purpose, he used a different notation which,
while patriotically employed in English mathematics
for the next century, proved awkward and was
eventually abandoned for Leibnizian differentials.
The basic concepts in calculus are the derivative of
a function y f(x) and its integral. The derivative is
another function f 0 whose value at any point x is the
instantaneous rate of change with respect to x of
the value y of f(x). This function in turn has its own
derivative f 00 , which is thus the second derivative of f,
and so on. Newton used overdot y for the derivative.
The integral of a function is another function whose
value represents the area between the curve and an
axis. Newton used overbar y to represent the integral.
While first and second derivatives are still quite com-
monly indicated with prime marks, replication of dots
and primes becomes difficult when speaking of higher
derivatives, or when using a variable. Leibnizs nota-
tion, unlike Newtons, used a ratio between the differ-
entials dy and dx (infinitesimal increments in y and x)
to represent the derivative. Second and higher deriva-
tives are indicated by exponents on the differentials;
thus f 0 (x) dy=dx, and f 00 x d2 y=dx2 , producing a
natural operator notation d/dx for differentiation, in
contrast to Newtons notation, which did not lend itself
to such generalization.
For integrals Leibniz used a special operator, the
long-S integral symbol f(x) dx. The differential dx
after the function in this notation is a mnemonic
for multiplication, just as the ratio between differen-
tials symbolizes division in the derivative notation.
Leibnizs convention recalls the fundamental theorem
of calculus (i.e., derivative and integral are inverses)
by use of the inverse operations multiplication and
division to represent them; by contrast, Newtons
notation was arbitrary and had no mnemonic value.
Following the invention of the calculus, mathe-
matics endured for several centuries a disconnect be-
tween axiomatic theory and actual practice. Calculus
worked, for reasons which mathematicians could not
explain, and about which engineers did not care.
The result was that practice got very far ahead
of axiomatic theory, until at the beginning of the
19th century, mathematicians resolved to put such
axiomatization at the front of their goals.
By axiomatization, mathematicians mean the es-
tablishment of a system of undefined elements,
axioms, and rules of inference for deriving theorems
from the axioms, based on the model of Euclidean
geometry, which for millennia has stood as the exem-
plar of pure mathematical thought. This is what is
meant in mathematics as a formal system; in linguis-
tics this term has a different sense, since linguists do
not establish axioms or prove theorems.
The most significant area of mathematics in this
axiomatization effort has been set theory, which
now, as a result of the 19th-century work, is taken
to underlie all areas of mathematics and logic. Set
theory has only three undefined terms: set, element
of a set, and inclusion in a set. These are universally
symbolized with capital letters to name sets, lower-
case letters to name elements in them, and the relation
symbol E for set inclusion, thus a E A indicates that a is
an element of the set A. Sets may also be specified by
using curly braces, e.g., the set of English vowel let-
ters is {a, e, i, o, u}, and the set N of natural numbers is
{1, 2, 3, . . .}. Since most interesting sets are infinite,
or at least very large, simple enumeration is rarely

sufficient; thus one finds qualified set descriptions,
e.g., the set of perfect squares can be described as
{x2: x E N}, pronounced the set of all x squared,
such that x is in N.
From this set convention come all the various uses
of parentheses and brackets for inclusion and specifi-
cation, such as the ordered pair (0, 1) in algebra,
which is simply a special set consisting of two ele-
ments in order. Exactly the same notation (0, 1) in
topology is used to denote an open set, i.e., the set of
all points on a line lying between 0 and 1, but not
including the end-points; if an end-point is included,
the notation changes to a square bracket, so that the
closed set [0, 1] has a largest and a smallest member,
while the open set (0, 1) has neither. In analysis, (0, 1)
following a function symbol such as f implies that f is
a function of two arguments to be applied to 0 and 1.
Fonts play a more important role in mathematical
notation than one might expect, given the arbitrari-
ness of font choice. The problem in mathematics is
that there is an infinite number of concepts that one
might wish to represent with a simple symbol, but
there is a limited number of simple symbols that can
be printed easily. Hence, it is common to mix and
match Latin, Greek, and Fraktur alphabets (among
others), in boldface, italic, and plain variants, each
indicating some systematic feature of the concept
under discussion. Such use is quite idiosyncratic,
with considerable individual variation and specific
traditions in each subarea. For instance, in relativity
theory, tensor indices use Greek letters when they
vary from 0 to 3 and Latin letters when they vary
from 1 to 3.
General Tendencies
Mathematical notation, like all mathematics, is an
exemplification of the grounding metaphors that
structure the field (see Lakoff and Nun ez, 2000).
Nuuchahnulth
J Stonham, University of Newcastle, Newcastle upon
Tyne, UK
2006 Elsevier Ltd. All rights reserved.
Nuuchahnulth, also known as Nootka, Nutka, Aht,
Takhaht, and taataaqsapa, and by the various dia-
lect designations, is one of the earliest documented
languages of Western Canada, with first contact going
back to the 1770s. Early work focused mostly on
Nuuchahnulth 741
These, in turn, are mostly based on the basic human
cognitive activity of predication, instantiated by con-
cepts such as function, operator, or relation, which
name various symbols and types of sentencelike
symbolic syntax in all areas of mathematics and
logic: a function f(x) is a clause with a predicate and
argument(s); an equation a b is a topiccomment
clause. All mathematical notation is ultimately pro-
nounceable, a subset of written language, though like
Chinese characters it is not language-specific.
See also: Assertion; Mood and Modality in Grammar; Ne-
gation; Operators in Semantics and Typed Logics; Panini;
Predication; Truth Conditional Semantics and Meaning.
Bibliography
Barrow J D (1992). Pi in the sky: counting, thinking, and
being. Boston: Little, Brown.
Cajori F (19281929). A history of mathematical notation
(2 vols). Lasalle, IL: Open Court. (Reprinted in 1993
(1 vol.). New York: Dover.)
Everett D L (2005). Cultural constraints on grammar and
cognition in Piraha : another look at the design features of
human language. Current Anthropology, AugOct.
Ifrah G (2000). The universal history of numbers: from
prehistory to the invention of the computer (3 vols. 2nd
revised English Edition, originally in French). London:
Havill.
Lakoff G & Nun ez R (2000). Where mathematics comes
from: how the embodied mind brings mathematics into
being. New York: Basic Books.
Pettersson J S (1996). Numerical notation. In Daniels P T
& Bright W (eds.) The worlds writing systems. New
York: Oxford University Press. 795806.
Schmandt-Besserat D (1992). How writing came about.
Austin: University of Texas Press.
Zimansky P (1993). Review of Schmandt-Besserat (1992).
Journal of Field Archaeology 20, 513517.
the vocabulary, with word lists being documented
by a number of early explorers. The earliest detailed
grammatical investigations include those of Knipe
(1868), Boas (1890), and Sapir (1924). The name
Nuuchahnulth is an anglicized version of nuuc aan u
along the mountains (nuci mountain(s),-a nu
along: indicates variable length of the vowel).
Geographically, the Nuuchahnulth people occupy
the west coast of Vancouver Island, from Brookes
Peninsula in the north to Bamfield in the south.

sufficient; thus one finds qualified set descriptions,
e.g., the set of perfect squares can be described as
{x2: x E N}, pronounced the set of all x squared,
such that x is in N.
From this set convention come all the various uses
of parentheses and brackets for inclusion and specifi-
cation, such as the ordered pair (0, 1) in algebra,
which is simply a special set consisting of two ele-
ments in order. Exactly the same notation (0, 1) in
topology is used to denote an open set, i.e., the set of
all points on a line lying between 0 and 1, but not
including the end-points; if an end-point is included,
the notation changes to a square bracket, so that the
closed set [0, 1] has a largest and a smallest member,
while the open set (0, 1) has neither. In analysis, (0, 1)
following a function symbol such as f implies that f is
a function of two arguments to be applied to 0 and 1.
Fonts play a more important role in mathematical
notation than one might expect, given the arbitrari-
ness of font choice. The problem in mathematics is
that there is an infinite number of concepts that one
might wish to represent with a simple symbol, but
there is a limited number of simple symbols that can
be printed easily. Hence, it is common to mix and
match Latin, Greek, and Fraktur alphabets (among
others), in boldface, italic, and plain variants, each
indicating some systematic feature of the concept
under discussion. Such use is quite idiosyncratic,
with considerable individual variation and specific
traditions in each subarea. For instance, in relativity
theory, tensor indices use Greek letters when they
vary from 0 to 3 and Latin letters when they vary
from 1 to 3.
General Tendencies
Mathematical notation, like all mathematics, is an
exemplification of the grounding metaphors that
structure the field (see Lakoff and Nunez, 2000).
Nuuchahnulth
J Stonham, University of Newcastle, Newcastle upon
Tyne, UK
2006 Elsevier Ltd. All rights reserved.
Nuuchahnulth, also known as Nootka, Nutka, Aht,
Takhaht, and taataaqsapa, and by the various dia-
lect designations, is one of the earliest documented
languages of Western Canada, with first contact going
back to the 1770s. Early work focused mostly on
Nuuchahnulth 741
These, in turn, are mostly based on the basic human
cognitive activity of predication, instantiated by con-
cepts such as function, operator, or relation, which
name various symbols and types of sentencelike
symbolic syntax in all areas of mathematics and
logic: a function f(x) is a clause with a predicate and
argument(s); an equation a b is a topiccomment
clause. All mathematical notation is ultimately pro-
nounceable, a subset of written language, though like
Chinese characters it is not language-specific.
See also: Assertion; Mood and Modality in Grammar; Ne-
gation; Operators in Semantics and Typed Logics; Panini;
Predication; Truth Conditional Semantics and Meaning.
Bibliography
Barrow J D (1992). Pi in the sky: counting, thinking, and
being. Boston: Little, Brown.
Cajori F (19281929). A history of mathematical notation
(2 vols). Lasalle, IL: Open Court. (Reprinted in 1993
(1 vol.). New York: Dover.)
Everett D L (2005). Cultural constraints on grammar and
cognition in Piraha: another look at the design features of
human language. Current Anthropology, AugOct.
Ifrah G (2000). The universal history of numbers: from
prehistory to the invention of the computer (3 vols. 2nd
revised English Edition, originally in French). London:
Havill.
Lakoff G & Nunez R (2000). Where mathematics comes
from: how the embodied mind brings mathematics into
being. New York: Basic Books.
Pettersson J S (1996). Numerical notation. In Daniels P T
& Bright W (eds.) The worlds writing systems. New
York: Oxford University Press. 795806.
Schmandt-Besserat D (1992). How writing came about.
Austin: University of Texas Press.
Zimansky P (1993). Review of Schmandt-Besserat (1992).
Journal of Field Archaeology 20, 513517.
the vocabulary, with word lists being documented
by a number of early explorers. The earliest detailed
grammatical investigations include those of Knipe
(1868), Boas (1890), and Sapir (1924). The name
Nuuchahnulth is an anglicized version of nuucaanu
along the mountains (nuci mountain(s),-a nu
along: indicates variable length of the vowel).
Geographically, the Nuuchahnulth people occupy
the west coast of Vancouver Island, from Brookes
Peninsula in the north to Bamfield in the south.
742 Nuuchahnulth

Table 1 Nuuchahnulth consonant inventory

Labial Alveolar Lateral Alveopalatal Velar/labiovelar Uvular/labiouvular Pharyngeal Glottal

p t k q
p t k q
ko qo
ko qo
s L s x x. h. h
Lo xo x. o h. o
c | c
c | c
m n y w
m n y
w

Nuuchahnulth has a relatively large consonant in- vowels labeled variable length; such vowels appear
ventory as shown in Table 1. Conversely, the vowel long in the first foot of the word, but are short after
system is quite simple, involving just the three vowels, that. These are characteristically Wakashan in nature
/i, u, a/, and a length distinction. There are, in addi- and constitute a special class of vowels.
tion, two mid vowels that are only encountered in Morphologically, Nuuchahnulth is extremely com-
the long variety, /e / and /o /, at least phonologically, plex, with upward of 500 bound derivational mor-
and even then only under certain special circum- phemes, along with numerous inflectional paradigms.
stances, such as in foreign borrowings. Primary stress The creative genius of the language is demonstrated
is predictable and appears on one of the first two by the following words:
syllables of a word, depending on weight.
(2) cicih. aq|mapt crab-apple tree (< cih. sour - aq|
Important morphophonological processes include
inside -mapt plant)
glottalization and lenition, which affect the preceding ciciyupkuk spaghetti (< ciyup intestines -kuk
consonant and depend on the individual suffix trig- resemble)
gering the effect (abbreviations: LOC, locative; DUP, niikmaLiicack guitar (< n ik scratch -maL about -iicu
duplication; SUF, suffix, MOM momentaneous aspect, tool) at strings -cako
NOW contemporaneous, MC momentaneous causativ e,
Morphological processes include reduplication, in-
SUB subordinate, L lengthening, R reduplication. All
fixation, and suffixation, although lexical compound-
data are from Sapirs unpublished fieldnotes, (Sapir,
ing is absent. There is a class of lexical suffixes that
no date), unless otherwise indicated.
requires reduplication of some part of the
(1a) hiitah. ti|si a| root as a concomitant of the attachment (3a),
hita-h. t -i| -si| -a|
in addition to reduplication indicating plurality (3b),
LOC-exit woods-go for[L]-MOM-NOW
distributivity (3c), repetition (3d), and other aspectu-
They started out of the woods.
al categories. In fact, double reduplication may occur
(1b) wiskum
in contexts in which both derivational and inflec-
wisk -um
angry -on the rocks
tional triggers to reduplication cooccur (DISTRIB,
angry on the rocks distributivity; REP, repetition; PASS, passive)
(1c) aayimki|qas (3a) kuukuh. inqiL
aya -miik -i| -qa s DUP-kuh . -inqiL
many -getter of -MC -1S.SUB SUF-hollow -at the ribs [RL]
may I be a getter of many . . . with a hole in the side
(1d) titinkum (3b) taataayi
DUP-ti -nuko -im DUP-taayii
SUF-wipe -at the hand [R] -thing PL-older brother
handwiper older brothers, seniors
(3c) nunupqimLayii at
As can be seen from (1a) and (1b), a glottalizing
DUP-nup -qimL -ayi -at
suffix will trigger glottalization of the preceding DISTRIB-one -classifier -give -PASS
consonant, whereas a nonglottalizing suffix ((1c) He gave a dollar to each.
and (1d)) will not. Other common morphophonolo- (3d) tuuxtuuxoa
gical processes include labialization, delabialization, DUP-tux -(y)a
vowel coalescence, and -epenthesis. One further REP-jump -DUR
morphophonological process involves a class of jumping
 Nuuchahnulth 743

Predicates are typically marked for aspect and (7b) iih. iiL |uL c apic
often for location (e.g., -iL in the house, -as on the iih. o -iiL |uL c apac
ground, in the village, and -is on the beach), and very -make good canoe
bear one of a set of paradigmatic mood/person/ He made a really nice canoe.
number suffixes and other possible markers. Tense is (Rose, 1981)
(7c) iih. naak capic
optionally marked on predicates, as is plural for
iih. o -na ko c apac
nouns. Nouns may be marked for number, diminutive,
big -have canoe
augmentative, former/future state, and possession. He has a big canoe.
Syntactically, Nuuchahnulth may be described as (Rose, 1981)
head-initial and head-marking. The most common (7d) ayasuu| nam int ath.
word order is verb initial, either verbsubjectobject aya -sawi| nam int -ath.
or verbobjectsubject, although arguments may be many -die Namint -tribe
omitted. In keeping with its head-marking nature, pos- Many Namint people died.
sessors follow possessees (4a) and relative clauses
Coordination may involve noun phrases, verb
follow their heads (4b):
phrases, or intermediate units and may employ one
iLuk i
(4a) h. aw qoayaciiktaqimL of several conjunctions, including is and uh. is , both

h. awiL-uk i qoayaciik -taqimL meaning and. Subordination is marked by special
chief-POSS DEF wolf -tribe inflectional paradigms and may or may not involve
the chief of the wolves subordinating conjunctions such as ani that, uyi
(4b) Luucsme i yaqoac itq tan a when, etc. There is a marker, - at, which has been
Luucsma- i yaqo-ac - i tq tan a labeled variously passive, switch reference, or dis-
woman-DEF REL-belong to -3S.REL child
course marker. (The debate over the role of this mor-
the woman whose child he was
pheme continues, for which see Kim (2004) and
In the case of relative clauses, there are two types references therein.) For further information on Nuu-
headed (4b), involving either yaqo who(m) or qoi chahnulth, the reader is referred to the bibliography
which, or headless (5), in which the specifier i is appended to this article, in particular the discussions
attached to the predicate of a relative clause: in Davidson (2002), Jacobsen (1979a,b), Kim (2003),
Nakayama (2001), Rose (1981), and Stonham (1999,
(5) qah. s i| i 2004).
qah. -s i| i
die -MOM DEF
See also: Canada: Language Situation; Wakashan.
the dead (ones)
A process reminiscent of noun incorporation exists,
but with a rather broad range of implementation: Bibliography
(6a) uutyaapa| [[ suuh. aa ]N]NP Boas F (1890). The Nootka. In Report of the 60th
u-ityaap -a| suuh. aa meeting of the British Association for the Advancement
REF-bring . . . as -NOW spring of Science (BAAS-R 60). 582604.
gift silver salmon Davidson M (2002). Studies in Southern Wakashan
He brought a gift of silver spring salmon. (Nootkan) grammar. Ph.D. diss., State University of
(6b) New York at Buffalo.
Jacobsen W H Jr (1979a). Wakashan comparative studies.
suuh. aa -ityaap -at In Campbell L & Mithun M (eds.) The Languages of
silver spring salmon -bring . . . as gift -PASS Native America. Austin, TX: University of Texas Press.
He brought a gift of silver spring salmon. 766791.
Jacobsen W H Jr (1979b). Noun and verb in Nootkan. In
Elements incorporated may include numerals (7a), Efrat B S (ed.) The Victoria conference on northwestern
adverbs (7b), adjectives (7c), or nouns (6b), always languages. Victoria, BC: British Columbia Provincial
from the object argument. Subjects of unaccusative Museum. 83155.
verbs may also be involved (7d): Kim E-S (2003). Theoretical issues in Nuu-chah-nulth pho-
nology and morphology. Ph.D. diss., University of British
(7a) h. aayumiikuk uh. is muu iih. tuup Columbia.
h. ayu -m i ko -uk uh. is muu iih. tuup Kim E-S (2004). The morphological status of - at in
ten -capture -POSS and four whale Nuu-chah-nulth. In Proceedings of BLS 30. Berkeley:
He captured fourteen whales. University of California Press.
744 Nuuchahnulth
Knipe C (1868). Some Account of the Tahkaht Language,
as Spoken by Several Tribes on the Western Coast of
Vancouver Island. London: Hatchard.
Nakayama T (2001). Nuuchahnulth (Nootka) morphosyntax.
Berkeley: University of California Press.
Rose S M (1981). Kyuquot grammar. Ph.D. diss., University
of Victoria.
Sapir E (no date). Fieldnotes on Nootka. Boas collection of
the American Philosophical Society [W2a.18]. Philadel-
phia, PA.
Sapir E (1924). The rival whalers, a Nitinat story (Nootka
text with translation and grammatical analysis). Interna-
tional Journal of American Linguistics 3, 76102.
Sapir E & Swadesh M (1939). Nootka texts, tales and
ethnological narratives, with grammatical notes and
Nyanja
S Mchombo, University of California, Berkeley,
CA, USA
2006 Elsevier Ltd. All rights reserved.
History and Politics
ChiNyanja (Nyanja) is a language of the Bantu group
of the Niger-Kordofanian language family, and is
spoken in parts of eastern, central, and southern
Africa. It is spoken in Malawi, where, from 1968
until recently, under the name of Chichewa, it served
as the national language. It is also spoken in Mozam-
bique, especially in the provinces of Tete and Niassa,
as well as in Zambia and Zimbabwe. In the latter,
according to some estimates, it ranks as the third most
widely used local language, after Shona and Ndebele.
The countries of Malawi, Zambia, and Mozambique
overwhelmingly constitute the central location of
chiNyanja.
The language derives its name from the lake that is
shared by Malawi, Mozambique, and Tanzania, with
most if it as part of Malawi. The local word for a large
expanse of water is nyanja. The people who lived
along the shores of the lake and the banks of the
Shire River called themselves aNyanja (lake dwellers).
The Shire River flows from the southern extremity
of the lake, formerly called Lake Nyasa, but now
known as Lake Malawi, through southern Malawi
into southern Mozambique to join the Zambezi
River. The aNyanja (the singular form of which is
mNyanja, where the prefix m is a syllabic nasal)
spoke the language called chiNyanja (henceforth,
Chinyanja).
Like most languages, Chinyanja has a number of
regional dialectal variations. One of these, spoken in
lexical materials. Philadelphia & Baltimore, MD: Lin-
guistic Society of America.
Sapir E & Swadesh M (1955). Native accounts of
Nootka ethnography. Bloomington, IN: Indiana Univer-
sity Research Center in Anthropology, Folklore and
Linguistics.
Stonham J (1999). Tsishaath Nootka phonetics and pho-
nology. [Lincom Studies in Native American Linguistics
32.] Munich: Lincom Europa.
Stonham J (2004). Linguistic theory and complex words:
Nuuchahnulth word formation. London: Palgrave
Macmillan.
Swadesh M (1939). Nootka internal syntax. International
Journal of American Linguistics 9, 77102.
the hinterland of Malawi, is called chiChewa (hence-
forth Chichewa). This dialectal variation is the one
that was spoken by the first president of Malawi, the
late Dr Hastings Kamuzu Banda (cf. Watkins, 1937).
The ascendancy of a Chewa to the presidency in
independent Malawi had repercussions on language
issues. President Banda argued that the classification
of Chichewa as a dialect of Chinyanja was erroneous,
deriving from unfortunate aspects of the history of
missionary activity in the country, whose early activ-
ities were concentrated along the lake. Banda invoked
aspects of history, plausible in some ways, to argue
that Chichewa was the language of which Chinyanja
was the dialectal variation (for pertinent observations,
see Marwick, 1963).
The version of the history of the Chewa that Banda
espoused was that the people who speak Chichewa,
known as aChewa, trace their origins to a group
of people known as the Maravi (according to some
Portuguese records) who migrated from the lower
basin of the Congo in central Africa and eventually
settled in the land mass now covered by Malawi,
Zambia, and Zimbabwe. Pushed by wars, disease,
and other maladies from the Congo area, the Maravi
were the first group of Bantu peoples to move and
settle in-present day Malawi in the 16th century.
Other Bantu groups such as the Tumbuka, Tonga,
Yao, Lomwe, and Ngoni moved into Malawi long
after the Maravi group had successfully established
itself (see Kalipeni, 1996).
The Chewa were led by a powerful leader called
Kalonga. He founded in Malawi what later came to be
called the Maravi empire. In Malawi he established
his headquarters or seat in a place called Mankhamba.
Once settled, he decided to extend his influence by
744 Nuuchahnulth
Knipe C (1868). Some Account of the Tahkaht Language,
as Spoken by Several Tribes on the Western Coast of
Vancouver Island. London: Hatchard.
Nakayama T (2001). Nuuchahnulth (Nootka) morphosyntax.
Berkeley: University of California Press.
Rose S M (1981). Kyuquot grammar. Ph.D. diss., University
of Victoria.
Sapir E (no date). Fieldnotes on Nootka. Boas collection of
the American Philosophical Society [W2a.18]. Philadel-
phia, PA.
Sapir E (1924). The rival whalers, a Nitinat story (Nootka
text with translation and grammatical analysis). Interna-
tional Journal of American Linguistics 3, 76102.
Sapir E & Swadesh M (1939). Nootka texts, tales and
ethnological narratives, with grammatical notes and
Nyanja
S Mchombo, University of California, Berkeley,
CA, USA
2006 Elsevier Ltd. All rights reserved.
History and Politics
ChiNyanja (Nyanja) is a language of the Bantu group
of the Niger-Kordofanian language family, and is
spoken in parts of eastern, central, and southern
Africa. It is spoken in Malawi, where, from 1968
until recently, under the name of Chichewa, it served
as the national language. It is also spoken in Mozam-
bique, especially in the provinces of Tete and Niassa,
as well as in Zambia and Zimbabwe. In the latter,
according to some estimates, it ranks as the third most
widely used local language, after Shona and Ndebele.
The countries of Malawi, Zambia, and Mozambique
overwhelmingly constitute the central location of
chiNyanja.
The language derives its name from the lake that is
shared by Malawi, Mozambique, and Tanzania, with
most if it as part of Malawi. The local word for a large
expanse of water is nyanja. The people who lived
along the shores of the lake and the banks of the
Shire River called themselves aNyanja (lake dwellers).
The Shire River flows from the southern extremity
of the lake, formerly called Lake Nyasa, but now
known as Lake Malawi, through southern Malawi
into southern Mozambique to join the Zambezi
River. The aNyanja (the singular form of which is
mNyanja, where the prefix m is a syllabic nasal)
spoke the language called chiNyanja (henceforth,
Chinyanja).
Like most languages, Chinyanja has a number of
regional dialectal variations. One of these, spoken in
lexical materials. Philadelphia & Baltimore, MD: Lin-
guistic Society of America.
Sapir E & Swadesh M (1955). Native accounts of
Nootka ethnography. Bloomington, IN: Indiana Univer-
sity Research Center in Anthropology, Folklore and
Linguistics.
Stonham J (1999). Tsishaath Nootka phonetics and pho-
nology. [Lincom Studies in Native American Linguistics
32.] Munich: Lincom Europa.
Stonham J (2004). Linguistic theory and complex words:
Nuuchahnulth word formation. London: Palgrave
Macmillan.
Swadesh M (1939). Nootka internal syntax. International
Journal of American Linguistics 9, 77102.
the hinterland of Malawi, is called chiChewa (hence-
forth Chichewa). This dialectal variation is the one
that was spoken by the first president of Malawi, the
late Dr Hastings Kamuzu Banda (cf. Watkins, 1937).
The ascendancy of a Chewa to the presidency in
independent Malawi had repercussions on language
issues. President Banda argued that the classification
of Chichewa as a dialect of Chinyanja was erroneous,
deriving from unfortunate aspects of the history of
missionary activity in the country, whose early activ-
ities were concentrated along the lake. Banda invoked
aspects of history, plausible in some ways, to argue
that Chichewa was the language of which Chinyanja
was the dialectal variation (for pertinent observations,
see Marwick, 1963).
The version of the history of the Chewa that Banda
espoused was that the people who speak Chichewa,
known as aChewa, trace their origins to a group
of people known as the Maravi (according to some
Portuguese records) who migrated from the lower
basin of the Congo in central Africa and eventually
settled in the land mass now covered by Malawi,
Zambia, and Zimbabwe. Pushed by wars, disease,
and other maladies from the Congo area, the Maravi
were the first group of Bantu peoples to move and
settle in-present day Malawi in the 16th century.
Other Bantu groups such as the Tumbuka, Tonga,
Yao, Lomwe, and Ngoni moved into Malawi long
after the Maravi group had successfully established
itself (see Kalipeni, 1996).
The Chewa were led by a powerful leader called
Kalonga. He founded in Malawi what later came to be
called the Maravi empire. In Malawi he established
his headquarters or seat in a place called Mankhamba.
Once settled, he decided to extend his influence by

acquiring more land and having it settled by his sub-
jects. To achieve those objectives he dispatched a
number of his matrilineal relatives to establish settle-
ments in various parts of the country. Among the
relatives who traveled on were such chiefs as Mwase,
who moved into the area called Kasungu, Kaphwiti
and Lunda, who settled in the lower Shire valley. As
they spread throughout the central and southern part
of the country, into eastern Zambia, and into parts
of Mozambique, including along the Zambezi River,
their language spread too. The dispersion of Kalongas
relatives and the ensuing Chewa diaspora resulted in
a proliferation of regional varieties of the language.
The distinct names that the regional varieties acquired
created the impression of the existence of a multiplic-
ity of ethnic groups. Some of the groups identified
themselves by making reference to significant features
of their habitat.
Malawi is a country dominated by a huge lake that
ranks as the third largest in Africa, after Victoria
Nyanza (Lake Victoria) and Lake Tanganyika, and as
the 12th largest in the world. As indicated, the word
for a large expanse of water in Chichewa is nyanja,
and the word for tall grass (savanna) is chipeta. The
people who settled along the lakeshores and along
the banks of the Shire River began to call them-
selves aNyanja, the lake people, and their particular
variety of Chichewa was called chiNyanja, or simply
Nyanja, the language of the lake people. Those who
moved into the interior, the land of tall grass, were
called aChipeta, the dwellers of the savanna land.
These names began to obscure the nature of the rela-
tionship among the people. This was further compli-
cated by the introduction of yet other labels. Thus,
the advent of the Portuguese, entering the area from
southern Africa in the 17th century, was accompanied
by the introduction of new labels. They had been in
contact with such ethnic groups as the Xhosa, the
Nyika, the Tchangani, etc. These referred to them-
selves as amaXhosa, amaNyika, amaTchangani, etc.
Banda claimed that when the Portuguese encountered
the Chewa living in southern Malawi and southern
Mozambique, who referred to themselves as aNyanja,
they referred to them as amaNyanja (see Banda,
19741975). Under the influence of Portuguese pho-
nology, the sound ny, a palatal nasal, got velarized to
ng. This gave rise to an ethnic group of amaNganja,
whose language was called chiManganja, definitely
not distinct from the Nyanja. Meanwhile, the Chewa
who had settled around the southern end of Lake
Malawi, spreading to the area surrounding Lake
Chirwa, encountered another ethnic group, the Yao.
The Yao predominated in Mozambique but had flowed
into the southeast part of Malawi. The Yao word for a
large expanse of water is nyasa, and they referred to the
Nyanja 745
Nyanja people as aNyasa. The people had by then come
to be grouped into aChewa, aChipeta, amaNganja,
aNyanja, and aNyasa. The last designation appears to
have contributed to British colonialists eventual desig-
nation of the lake as Lake Nyasa, and of the country as
Nyasaland. This is the name that the country had until
independence in 1964, when the name of Malawi,
apparently derived from Maravi, was restored.
The multiplicity of labels under which the Chewa
came to be identified was something that received
some comment from various scholars. Thus, Young
remarks about the language Nyanja that
it is the language of a people scattered over a large
South-east-central African area, the aMaravi, who to-
day live under at least six different names according to
the area in which Europeans found them in the closing
decades of the last century. And they were more or less
on the same ground at least 300 years earlier since the
Portuguese records give some of them the same names as
they bear to-day. (Young, 1949: 53)
Earlier, Hetherwick had stated that
On the Shire River they are called Manganja, a merely
local pronunciation of the word A-Nyanja. Around
Lake-Shirwa they are best known by their Yao name
A-Nyasa. (Hetherwick, 1901: 15)
Although Chichewa is widely spoken in Southern,
eastern, and central Africa, spreading over the land-
mass that includes Zambia, Malawi, Mozambique,
and Zimbabwe, Greenberg does not mention it in his
classification of African languages. In the works of
Guthrie, Chichewa and Chimanganja are listed as
two dialect variations of Nyanja. He classifies Chi-
chewa as belonging to zone N31b, being identified as
the second dialect of the main language.
Chinyanja served as the main linguistic medium for
the mass media in Malawi and was taught as a subject
in educational institutions at both primary and sec-
ondary levels. In 1968, under political pressure from
President Banda, a resolution was passed at the annu-
al convention of the Malawi Congress Party, then the
ruling and sole political party in Malawi, to have the
name of the language officially changed to Chichewa.
From that point the language Chinyanja became
known as Chichewa in Malawi. Simultaneously,
it was elevated to the status of national language.
English remained in use as the official language.
The language policy adopted in Malawi that made
Chichewa the national language contributed to the
promotion of Chichewa through active educational
programs, media usage, and other research activities.
With the exception of work carried out within the
University of Malawi, tied to contributions to, and
adaptations of, advances in linguistic theory, the pro-
motion and standardization of Chichewa were placed
746 Nyanja
under the oversight of the Chichewa Board. The
terms of reference of the Chichewa Board included
monitoring of proper usage of Chichewa in the
media, revising and updating the orthographic con-
ventions, as well as engaging in lexicographic work.
The sustained effort over many years to boost the
status of Chichewa as the main language resulted
in increased functional literacy in that language. Out
of a population of around 11 million in Malawi,
upwards of 65% have functional literacy or active
command of this language.
The political directive that led to the change of
name of the language in Malawi, from Chinyanja to
Chichewa, did not carry over to the neighboring coun-
tries of Zambia and Mozambique. Political factors
were definitely relevant. During the period when the
language issue was being addressed in Malawi, polit-
ical relations between Malawi and Zambia reached a
nadir. This deserves comment.
Regional Politics and Language Issues
Soon after gaining independence from the United
Kingdom in 1964, Malawi had a cabinet crisis.
A number of radical political activists with more
nationalist fervor broke ranks with Kamuzu Banda,
then Prime Minister. These were among the political
leaders who had invited Kamuzu Banda to return
from his exile in Britain, and subsequently Ghana, to
join in, and assume command of, the fight for inde-
pendence. Following the granting of independence
the young radicals got disillusioned with the direction
taken by Kamuzu Bandas policies. The policies were
seen as aiming to oppress the masses, to practically
deify Banda himself as a cult figure to whom the
people were to pay homage, to undermine the efforts
to promote Pan-Africanism, and to maintain the sta-
tus quo, such that the colonialists would still enjoy
the privileges that they had before independence. In
brief, Bandas policies brought out the dictatorial
tendencies that were eventually to characterize his
rule and laid the groundwork for his tyrannical grip
over the countrys political development. In the revolt
that resulted from the rift, the dissident politicians left
the country and sought refuge in the neighboring
countries of Tanzania and Zambia. This resulted in
strained relations between Malawi on the one hand
and those other countries on the other.
The situation became aggravated when, around
1967, there was an unsuccessful effort by a group of
insurgents under the leadership of Yatuta Chisiza, a
dissident politician, to unseat Kamuzu Banda mili-
tarily. The military escapade was viewed as having
been perpetrated with the connivance or complicity, if
not open support, of those neighboring countries. At
the minimum, the insurgents seemed to have received
logistical support from Zambia (for relevant details
see Mchombo, 1998a, 1998b).
The worsening relations between the two countries
meant that the policy about language, in this regard
relating to change in the name, was not merely
viewed as an issue internal to Malawi but, further,
as yet another instance of Kamuzu Bandas grandiose
scheme to identify the country or region with his
cultural and linguistic heritage, a version of Chewa
hegemony. As noted by Matiki, Banda carried this
idea of Chewa supremacy a little further by claiming
that the dialect spoken by his clan is the best. This
is nothing but linguacentrism (Matiki, 1997: 529).
Even if the logic or historical accuracy or factual basis
for the name change were to prove impeccably sound
and unimpeachable, Zambia was not ready to take
hints or, worse, orders, from Malawi. A subsequent
dispute concerning the proper borders of the coun-
tries, again perpetrated by the Malawi regime around
the same time, merely exacerbated an already grave
situation. Thus, in Zambia, as well as in Mozambique,
the language has always remained as Chinyanja.
In Mozambique Chinyanja is native to 3.3% of a
population numbering approximately 11.5 million.
In Tete Province it is spoken by 41.7% of a popula-
tion of 777 426, and it is the first language of 7.2% of
the population of Niassa Province, whose population
totals 506 974 (see Firmino, 1995). In Zambia, with a
population of 9.1 million, Chinyanja is the first lan-
guage of 16% of the population and is used and/or
understood by at least 42% of the population,
according to a survey conducted in 1978 (cf. Kashoki,
1978). It is one of the main languages of Zambia,
ranking second after Chibemba (Bemba). In fact, out
of the 9.1 million people of that country, it is estimated
that 36% are Bemba, 18% Nyanja, 15% Tonga, 8%
Barotze, and the remainder consisting of other ethnic
groups including the Mombwe, the Tumbuka, and the
Northwestern peoples (see Kalipeni, 1996). The fig-
ures show that at least six million people are fluent in
Chinyanja.
The Ascendancy of Chinyanja
A recurrent joke in linguistics courses about the dis-
tinction between language and dialect is the quip that
a language is a dialect with an army and a navy. The
rise of Chichewa in Malawi was intimately connected
to the tenure of Kamuzu Banda, a Chewa, as presi-
dent of Malawi. With altered political dispensation
through the shift to democratic practice, and Bandas
subsequent demise, Chichewa effectively lost the
army and navy that protected it from the status of
dialect. Without formally or openly introducing a

new language policy, Malawian scholars have felt it
prudent to fall in line with the other countries in the
region by restoring the name of Chinyanja to the
language. This restoration of Chinyanja to its former
status goes beyond mere efforts to promote regional
linguistic harmonization. Within Malawi the national
language policy adversely affected the other lan-
guages. Once again, as noted by Matiki, the change
of name [from Chinyanja to Chichewa] angered some
people because there was no justification in changing
the name other than the fact that President Banda was
an ethnic Chewa (1997: 527). Thus, the other ethnic
groups in the country felt alienated, more so given the
identification of Chichewa with political power and
the relegation of their languages to relative obscurity
(cf. Kishindo, 1994). The political transition to dem-
ocratic practice and the departure of Kamuzu Banda
from the political helm provided opportune occasion
for implementation of more equitable access to polit-
ical participation and the recognition of the cultural
and linguistic heritage of the various segments of the
nation. Kamuzu Bandas departure from the political
scene was accompanied by the ascendancy of Bakili
Muluzi, a Yao, as the second president of a democratic
Malawi. Inevitably, the political changes witnessed
shifting fortunes for Chichewa.
On Reverting to the Name of Chinyanja
Although use of the label Chichewa is likely to re-
main, there is systematic diminution of its former
status. Thus, the Chichewa Board was subsequently
dissolved and, in 1996, the Center for Language Stud-
ies (CLS), with a broader scope of activities, was
established to replace it. The center is affiliated to
the University of Malawi, stressing its new mandate
as a locus of research and scholarship, not as an organ
of political ideology or an instrument of political
hegemony. The activities carried out under the aus-
pices of the CLS have included concerted efforts to
document and provide linguistic descriptions of some
of the endangered languages in Malawi.
Further, the erstwhile Department of Chichewa and
Linguistics at the University of Malawi, pioneered
by the present author, under political directive
from President Banda to the University of Malawi
to contribute to the enhancement of Chichewa, was
renamed the Department of African Languages and
Linguistics, thereby degrading further the profile of
Chichewa.
The politically correct stance of diminishing the
profile of Chichewa in the university through the
removal of direct reference to it in the name of
the department, together with the establishment
of the CLS, should not, however, be (mis)construed
Nyanja 747
as indicating a real diminution in linguistic work on
it. Given the historical circumstances which account
for the preponderance of the available trained
linguists working on Chichewa, and the headstart
Chichewa was given in getting material prepared for
educational and other purposes, it will continue to
function in the capacity of the major language of
the country. Of some significance, at least to public
perception, is the recent introduction of radio news
bulletins in the other languages such as Chilomwe
(Lomwe), Chiyao (Yao), Chitumbuka (Tumbuka),
Chitonga (Tonga), and Chisena (Malawi Sena). The
news bulletins have served to increase peoples
awareness of these other Malawian languages, sub-
tly contributing to the subjection of Chichewa to
competition.
There are grounds for the restoration of Chinyanja
as the main language. These include the literary tradi-
tion that Chinyanja has enjoyed (see Made et al.,
1976). The description of Chinyanja goes back to
at least 1875; the first significant work can be traced
to Alexander Riddels publication in 1880 of A gram-
mar of Chinyanja as spoken at Lake Nyasa, with
Chinyanja-English vocabulary. This work, while not
linguistically very significant, was followed in 1891
with the publication of George Henrys A grammar of
Chinyanja: a language spoken in British Central
Africa on and near the shores of Lake Nyasa. This
was more comprehensive than the work by Riddel. In
1892 David Scotts A cyclopaedic dictionary of the
Manganja language spoken in British Central Africa
appeared, a work that was later to be revised
and enlarged by Alexander Hetherwick in 1929 as
Dictionary of the Nyanja language. This still remains
an authoritative dictionary of the language. Previous-
ly, in 1901, Hetherwick had produced A practical
manual for the Nyanja language.
These descriptions of Chinyanja and the functional
utility that the language enjoyed underscored its legit-
imacy to the status of a major language or lingua
franca. The subsequent adoption of the colonialists
language policy, which recognized the position of
Chinyanja in Central Africa, especially in Malawi,
eliminated further detractions from its status. The
establishment of the colonial administration in
Malawi at the turn of the 20th century provided
extra impetus to the promotion of Chinyanja. This
led to the appearance of more works on Chinyanja
and the emergence of more literary works and
newspapers, such as Msimbi, a weekly Chinyanja
newspaper that flourished for several years from
1951.
As Chinyanja gets rehabilitated, and some recent
publications such as dictionaries now bear that
name instead of Chichewa, the latter still remains
748 Nyanja

undeniably the most familiar label for the language A lingering perennial question relates to the semantic
that is spoken and understood by more than half the or cognitive basis for the classification of nouns.
population of Malawi. The reversion to the label of A definitive response to the question remains forth-
Chinyanja, just like the prior change to Chichewa, coming. Noun modifiers are marked for agreement
may lack a linguistic basis but it definitely fulfills with the class features of the head noun, and these
political objectives. features are also reflected in the subject marker (SM)
and object marker (OM) in the verbal morphology.
This can be illustrated by the following:
Linguistic Aspects of Chinyanja
(1a) chi-soti ch-a nga lch-a -tso pano
Chinyanja manifests typical aspects of the linguistic 7-hat 7SM-my 7SM-ASSOC-now
structure of Bantu languages. Its nominal system com- chi-ja ch-ma-sangalats-a a-lenje
prises a number of gender classes that are involved in 7SM-REL.PRO 7SM-HABIT-please-FV 2-hunters
the agreement patterning of the language, character- that new hat of mine pleases hunters
istic of Bantu languages. Thus, nominal modifiers (1b) m-kondo w-anga w-a-tsopano
3-spear 3SM-my 3SM-ASSOC-now
agree with the head noun in the relevant features of u-ja u-ma-sangalats-a a-lenje
gender and number, as will be illustrated below. 3SM-REL.PRO 3SM-HABIT-please-FV 2-hunters
In its verbal structure, Chinyanja, just like other that new spear of mine pleases hunters
Bantu languages, displays an elaborate agglutinative
In these sentences, the words in construction with the
structure. The verb comprises a verb root or radical,
nouns are marked for agreement with that head noun.
to which suffixes or extensions are added to form the
(The actual agreement markers in these examples are
verb stem (cf. Guthrie, 1962). The extensions affect
chi and u. The i vowel in chi is elided when followed
argument structure or the number of expressible nom-
by a vowel, and the u is replaced by the glide w in
inal arguments that the stem can support. The verb
a similar environment.) Chinyanja is a head-initial
stem also has proclitics which encode such syntacti-
language. Within the noun phrase the head noun
cally oriented information as negation, tense/aspect,
precedes its modifiers. The formal patterns that mark
subject and object markers, modals, conditional
singular and plural number are traditionally identi-
markers, directional markers, etc. (cf. Mchombo,
fied by a numbering system, now virtually standard in
2004).
Bantu linguistics (Bleek, 18621869). Consider the
Chinyanja is also a tonal language, displaying
following data:
features of lexical and grammatical tone. It has two
tone levels, high (H) and low (L). Contour tones are (2a) m-nyamata boy a-nyamata boys
attested but result from a combination of these tone m-lenje hunter a-lenje hunters
levels, usually on long syllables (Mtenje, 1986). In its m-kazi woman a-kazi women
segmental phonology, Chinyanja has a basic organi- (2b) m-kondo spear mi-kondo spears
zation of five vowel phonemes. The verb stem, i.e., mu-nda garden m-nda gardens
the domain comprising the verb root and the suffixes m-kango lion mi-kango lions
(2c) tsamba leaf ma-samba leaves
changing argument structure, is also the domain of
duwa flower ma-luwa flowers
vowel harmony. In its syllable structure, Chinyanja phanga cave ma-panga caves
has the basic CV structure common in Bantu (Mtenje, (2d) chi-sa nest zi-sa nests
1980). chi-tosi chicken zi-tosi chicken droppings
dropping
Classification of Nouns
chi-putu grass zi-putu grass stubble
The nominal morphology of Chinyanja displays the stubble
paradigmatic case of nouns maintaining, at the mini- These classes show part of the range of noun clas-
mum, a bimorphemic structure, which consists in a sification that is characteristic of Bantu languages. In
nominal prefix and a nominal stem. The prefix the examples above, the singular forms of the first
encodes grammatically relevant information of gen- group above, dominated by nouns that denote ani-
der (natural) and number, involved in agreement be- mate things, constitutes Class 1, and its plural coun-
tween the nouns and other grammatical classes in terpart is Class 2. Of course, not all animate things
construction with them. This is illustrated by the are in this class. In fact it does also include some
following: inanimate objects. The groups in (2b), (2c), and (2d)
m-lenje hunter a-lenje hunters constitute Classes 3, 4, 5, 6, 7, and 8 respectively. The
chi-soti hat zi-soti hats odd numbers indicate the singular forms and the even
m-ko ndo spear mi-ko ndo spears numbers their plural counterparts. This runs on to
 Nyanja 749

Classes 5, 6, 7, and 8. There is a further class, 1a. This Table 1 Full range of noun classes in Chinyanja
class consists of nouns whose agreement patterns are Classes Prefixes Subject Object
those of Class 1 but whose nouns lack the m(u) prefix marker marker
found in the Class 1 nouns. The plural of such nouns
SG PL SG PL SG PL SG PL
is indicated by prefixing a to the word. For instance,
the noun kalu lu hare, whose plural is akalu lu, typi- 1 2 m(u)- a- a- a- m(u) wa
fies this class. Each of these classes has a specific class 3 4 m(u)- mi- u- i- u i
marker and a specific agreement marker. Beginning 5 6 *li- ma- li- a- li wa
with Class 2, the agreement markers are respectively 7 8 chi- zi- chi- zi- chi zi
9 10 *N- *N- i- zi- i zi
a, u, i, li, a, chi, and zi. Class 1 has the agreement 12 13 ka- ti- ka- ti- ka ti
markers mu (or syllabic m), u, and a, depending on 14 6 u- ma- u- a- u wa
the category of the modifier. Consider the following: 15 ku- ku- ku-
16 pa- pa- pa-
(3) m-lenje m-mo dzi a-na-bwe la nd m-ko ndo 17 ku- ku- ku-
1-hunter 1SM-one 1SM-PAST- with 4-spears 18 m(u)- m(u)- m(u)-
come-FV
one hunter came with spears

Here, the numeral mo dzi one is marked with the

agreement marker m but the verb has a for the subject
marker. The u is used with demonstratives and when
the segment that follows is a vowel. This seems to
apply to most cases regardless of whether the vowel in
question is a tense/aspect marker, an associative
marker, or part of a stem, such as with possessives.
Consider the following:
(4) m-lenje w-anu u-ja w-a
1-hunter 1SM-your 11SM-that SM-ASSOC
ntha bwala w-a-thyol-a mi-ko ndo
10-humor 1SM-PERF-break-FV 4-spears
that humorous hunter of yours has broken the
spears
In this sentence, the glide w replaces u when a vowel
follows regardless of the function associated with that
vowel.
Although most of the nouns are bimorphemic,
there are a number of cases where a further prefix,
marking either diminution or augmentation, is added
to an already prefixed noun. This is shown in the
following:
(5) ka-m-lenje k-a nu ka-ja k-a
12-1-hunter 12SM-your 12SM-that 12-ASSOC
ntha bwala k-a-thyol-a ti-mi-ko ndo
10-humor 12SM-PERF-break-FV 134-spears
that small humorous hunter of yours has broken
the tiny spears
In this sentence, the preprefixes ka for singular and ti
for plural get attached to nouns to convey the sense
of diminutive size. These preprefixes then control
the agreement patterns (cf. Bresnan and Mchombo,
1995), providing the rationale for regarding them as
governing separate noun classes. One significant point
to be made is that locatives also control agreement
patterns. Consider the following:
(6) ku mudzi kw-a nu ku -ma-sangala ts-a
17-at 3-village 17SM-your 17-HABIT-please-FV
ale ndo
2-visitors
your village [i.e., the location] pleases visitors
This gives such locatives the appearance of being class
markers, giving rise to the view that in Chinyanja
locatives are not really prepositions that mark
grammatical case but, rather, class markers (for some
discussion, see Bresnan, 1991, 1994).
The full range of noun classes in Chinyanja, togeth-
er with their restive subject and object markers, is
given in Table 1. Some of the classes have prefixes
that are starred. These classes consist of nouns that,
normally, lack the indicated prefix in the noun mor-
phology. Samples of Class 5 nouns are provided
above. Most of the nouns in Classes 9 and 10 begin
with a nasal but there are no overt changes in their
morphological composition that correlate with
number. The number distinction is reflected in the
agreement markers rather than in the overt form of
the noun. Examples of Class 9/10 nouns are nyu mba
house(s), nthenga feather(s), mphni tattoo(s),
nkhondo war. Class 15 consists of infinitive verbs.
The infinitive marker ku- regulates the agreement
patterns, just like the diminutives (Classes 12 and
13) and locatives. The infinitives are thus regarded
as constituting a separate class although, just as with
the locatives, there are no nouns that are peculiar to
this class. There are minor exceptions to locatives.
These have to do with such words as pansi down,
kunsi underneath, panja outside of a place, kunja
(the general) outside, pano here (at this spot), kuno
here (hereabouts), muno in here. With these, the
locative prefixes pa, ku, and mu are attached to
the stems -nsi, -nja, and -no, which are bound. The
750 Nyanja
agreement pattern regulated by the infinitive marker
ku- is exemplified by the following:
(7) ku-mba kw-anu ku -ma-sangala ts-a Mackenzie D (1911). Notes on Tumbuka syntax. Nyasaland:
15INF-sing 15SM-your 15SM-HABIT-please-FV
a-lenje
2-hunters
your singing pleases hunters
See also: Bantu Languages; Kordofanian Languages;
Malawi: Language Situation; Mozambique: Language
Situation; Niger-Congo Languages; Tanzania: Lan-
guage Situation.
Bibliography
Banda H K (19741975). Lectures on Chichewa at Chan-
cellor College by His Excellency the Life President
Ngwazi Dr. H. Kamuzu Banda. Blantyre, Malawi.
Bleek W H I (18621869). A comparative grammar of
South African languages. London: Tru bner.
Bresnan J (1991). Locative case vs. locative gender. Paper
delivered at the Berkeley Linguistics Society. Berkeley:
University of California.
Bresnan J (1994). Locative inversion and the architecture
of universal grammar. Language 70(1), 72131.
Bresnan J & Mchombo S A (1987). Topic, pronoun and
agreement in Chichewa. Language 63(4), 741782.
Bresnan J & Mchombo S A (1995). The lexical integrity
principle: evidence from Bantu. Natural Language &
Linguistic Theory 13, 181254.
Chichewa Board (1980). Chichewa orthography rules.
Zomba: Chichewa Board.
Chinyanja orthography rules (1931). Zomba: The Secretar-
iat.
Firmino G D (1995). Revisiting the language question
in post-colonial Africa: the case of Portuguese and indig-
enous languages in Mozambique. Doctoral dissertation,
University of California, Berkeley.
Guthrie M (1962). The status of radical extensions in Bantu
languages. Journal of African Languages 1, 202220.
Hailey F (1938). An African survey: a study of problems
arising in Africa south of the Sahara. London: Oxford
University Press.
Hetherwick A (1901). A practical manual of the Nyanja
language. London: Christian Knowledge Society.
Kalipeni E (1996). Chewa and related groups in Malawi/
Zambia. In Gall T (ed.) Worldmark Encyclopedia
of Cultures and Daily Life 1: Africa. Pepper Pike, OH:
Eastword Publications Development. 98102.
Kashoki M (1978). Between-language communication
in Zambia. In Ohannessian S & Kashoki M (eds.)
Language in Zambia. London: International African
Institute.
Kishindo P J (1994). The impact of a national language
on minority languages: the case of Malawi. Journal of
Contemporary African Studies 12(2), 127150.
Livingstonia Mission.
Made S M, Mangoche Mbewe M V B & Jackson R
(1976). 100 years of Chichewa in writing, 18751975.
Zomba.
Malawi Congress Party Convention (1983). Malawi
Congress Party Convention resolutions, 19651983.
Blantyre, Malawi: Blantyre Print.
Marwick M G (1963). History and tradition in east central
Africa through the eyes of Northern Rhodesian Cewa.
Journal of African History 4, 375390.
Matiki A (1997). The politics of language in Malawi: a
preliminary investigation. In Herbert R K (ed.) African
linguistics at the crossroads: papers from Kwaluseni.
Cologne: Ru diger Ko ppe. 521540.
Mchombo S (1998a). Democratization in Malawi: its
roots and prospects. In Gros J-G (ed.) Democratization
in late twentieth-century Africa: coping with uncertainty.
Westport, CT: Greenwood Press. 2140.
Mchombo S (1998b). National identity, democracy, and
the politics of language in Malawi and Tanzania. Journal
of African Policy Studies 4(1), 3346.
Mchombo S (2004). The syntax of Chichewa. Cambridge:
Cambridge University Press.
Mtenje A (1980). Aspects of Chichewa derivational phonol-
ogy and syllable structure constraints. M.A. dissertation,
Southern Illinois University.
Mtenje A D (1986). Issues in the nonlinear phonology of
Chichewa. Doctoral dissertation, University College
London.
Price T (1940). Nyanja linguistic problems. Africa 13,
125137.
Price T (1946). The elements of Nyanja. Blantyre:
Nyasaland Church of Scotland Mission.
Sanderson M M & Bithrey W B (1939). An introduction
to Chinyanja. Glasgow: African Lakes Corporation.
Edinburgh/London: Oliver & Boyd.
Scott D C (1892). A cyclopaedic dictionary of the
Manganja language spoken in British Central Africa.
Edinburgh: Foreign Mission Committee of the Church
of Scotland.
Thomson T D (1942). Military Chinyanja. Zomba: Gov-
ernment Printer.
Thomson T D (1955). A practical approach to Chinyanja.
Zomba: Government Printer.
Watkins M H (1937). A grammar of Chichewa, A Bantu
language of British Central Africa. Philadelphia: Linguis-
tic Society of America.
Young C (1949). A review of A practical approach to
Chinyanja with English-Nyanja vocabulary by T D
Thomson, Salisbury, 1947. Africa 19, 253254.