Chapter Order of Numeral and Noun

by Matthew S. Dryer

1. Defining the values

This map shows the order of cardinal numerals with respect to a noun they modify. Cardinal numerals are words denoting the number of things referred to, as in English four chairs, in contrast to ordinal numerals, which specify the rank in some order of objects, as in English the fourth chair. See also Map 131A on numeral bases and Map 53A on ordinal numerals.

Values of Map 89A. Order of Numeral and Noun
Go to map
 ValueRepresentation
Numeral precedes noun (NumN) 479
Numeral follows noun (NNum) 608
Both orders of numeral and noun with neither order dominant 65
Numeral only modifies verb 2
Total: 1154

The first type shown on the map consists of languages in which the numeral precedes the noun. English is an instance of such a language, as is Selknam (Chon; Tierra del Fuego, Argentina), illustrated in (1).

(1) Selknam (Najlis 1973: 28) 

şèàwken-èn 

naʔ 

three-f

woman 

Num 

‘three women’ 

The second type consists of languages in which the numeral follows the noun. Examples of such languages include Pumi (Tibeto-Burman; China), as in (2a), and Ju|’hoan (Northern Khoisan; Namibia), as in (2b).

(2)

a. 

Pumi (Ding 1998: 191) 

 

qüa 

xüé 

 

pig 

eight 

 

Num 

 

‘eight pigs’ 

b. 

Ju|’hoan (Dickens 1992: 61) 

 

jú 

tsán 

gèà 

tjù 

n!áng 

 

person 

two 

be 

house 

in 

 

Num 

     
 

‘There are two people in the house.’ 

The third type shown on the map are languages in which both orders of numeral and noun occur with neither order dominant (see “Determining Dominant Word Order”). An example of such a language is Acoma (Keresan; New Mexico), as in (3).

(3) Acoma (Maring 1967: 119) 

a. 

tʸának’a 

wáaštíṭa 

b.

kákʰana 

tʰúw̓é 

 

four 

fawn 

 

wolf 

two 

 

‘four fawns’ 

 

‘two wolves’ 

In most languages with both orders of numeral and noun, it is not clear whether there is any difference in meaning associated with the two orders. In Nias (Austronesian; Sumatra, Indonesia), the noun phrase is interpreted as indefinite when the numeral precedes the noun, as in (4a), but definite when it follows, as in (4b). 

(4) Nias (Brown 2005

a. 

öfa 

geu 

m-baβi 

s=afusi 

 

four 

clf 

abs-pig

rel=white

 

‘four white pigs’ 

b. 

baβi-ra 

s=afusi 

si=öfa 

geu 

 

pig-3pl.poss

rel=white

rel=four

clf 

 

‘their four white pigs’ 

In some languages there is a split among numerals, some preceding the noun, others following. For example, in Egyptian Arabic, the numerals for ‘one’ and ‘two’ follow the noun, while higher numerals precede the noun, as in (5).

(5) Egyptian Arabic (Gary and Gamal-Eldin 1982: 111) 

a. 

binteen 

ʔitneen 

b. 

talat 

banaat 

 

girl.du

two 

 

three 

girl.pl

 

‘two girls’ 

 

‘three girls’ 

This variety of Arabic is shown on the map as lacking a dominant order of numeral and noun. In Moroccan Arabic, only the numeral for ‘one’ follows the noun, as in (6a), while all other numerals precede the noun, as in (6b). 

(6)Moroccan Arabic (Harrell 1962: 206) 

a. 

mṛa 

weḥda 

b. 

žuž 

ktub 

 

woman 

one.f

 

two 

book 

 

‘one woman’ 

 

‘two books’ 

When only one numeral differs from the others in position, as in Moroccan Arabic, the position of the other numerals is treated as dominant. Hence, Moroccan Arabic is shown on the map as having numeral before noun as the dominant order. Iraqi Arabic is intermediate between Egyptian and Moroccan Arabic: the numeral for ‘one’ follows, the numeral for ‘two’ either precedes or follows, and numerals higher than two normally precede (Erwin 1963: 348). It is shown on the map as lacking a dominant order of numeral and noun. 

In most languages in which some numerals precede the noun while others follow, both of the two sets of numerals represent continuous sequences of numbers, so that there is some number n  such that numerals from 1 to n  occur on one side of the noun while numerals greater than n  occur on the other side. For Egyptian Arabic, illustrated above in (5), the value of n  is 2. For Tauya (Madang, Trans-New Guinea; Papua New Guinea), the value of n  is 4 (MacDonald 1990: 117-118). For Nivkh (isolate; Sakhalin Island, Russia), the value of n  is 5 (Panfilov 1962: 129-130). For Runga (Maban; Chad), the value of n  is 6 (Nougayrol 1990: 40-42).

There are other languages which do not exhibit a pattern with a single cutoff point. The majority of exceptions appear to be languages in which the numeral for ‘two’ occurs on one side of the noun, while the numeral for ‘one’ and higher numerals occur on the other side. Examples of languages of this sort include Shipibo-Konibo (Panoan; Peru; Pilar Valenzuela, p.c.) and Coptic (Egyptian; Lambdin 1983: 56). A more unusual case is provided by Obolo (Benue-Congo; Nigeria), in which the numerals for 10, 20 and 400 precede the noun, while other numerals consisting of one word follow. The special treatment of 20 and 400 reflects the vigesimal system (see Map 131A). With numerals consisting of more than one word, the first word (which will normally be one of the words for ‘ten’, ‘twenty’, or ‘four hundred’) will precede the noun, while the remainder will follow the noun, as in (7c).

(7) Obolo (Faraclas 1984: 27, 37) 

a. 

óbóp 

úwù 

b. 

úwù 

ílé 

ítá 

 

four.hundred 

house 

 

house 

big 

three 

 

‘four hundred houses’ 

 

‘three big houses’ 

c. 

étîp 

úwù 

mè 

gò 

 

twenty 

house 

and 

five 

 

‘twenty-five houses’ 

Welsh is similar to this in that numerals consisting of more than one word place the first word in the numeral phrase before the noun, and the remainder following the noun (Williams 1980b: 41). Because a numeral consisting of a single word precedes the noun, and these include the most frequent numerals, Welsh is shown on the map as NumN; for analogous reasons, Obolo is coded as NNum. 

The final type shown on the map is languages in which the numeral does not combine syntactically with the noun at all, but occurs only as a modifier of the verb. There are only two languages in my data in which this construction is the only normal way to express numerals, in close proximity to each other in western Brazil, namely Karó (Arára) (Tupian) and Wari’ (Chapacura-Wanham; Everett and Kern 1997: 154-155, 221, 347), the former illustrated in (8).

(8) Karó (Arára) (Gabas 1999: 172) 

maʔwɨt 

ip 

ʔɨy 

matet 

cagárokōm=nem 

man 

fish 

catch.ind

yesterday 

two=advz

‘The man caught two fish yesterday.’ 

This construction is somewhat similar to the English construction illustrated in (9), in which the word all  appears in a position associated with adverbial elements.

(9) 

The guests have all left. 

There are many other languages that allow this option alongside a construction in which the numeral occurs in the noun phrase. For example, in Kutenai (isolate; western North America), there are two sorts of numeral words, one illustrated in (10a), in which the numeral occurs in the noun phrase with the noun (ki=ʔas ni¢tahaⱡ  ‘two men’), the other illustrated in (10b), in which the numeral is a preverb modifying the verb, somewhat analogous to the Karó (Arára) construction in (8).

(10) Kutenai (own data) 

a. 

xa-s 

ki=ʔas 

ni¢tahaⱡ 

ʔat 

mitx-ni 

niʔ-s 

numa-s 

 

then-obv

subord=two

young.man 

hab 

shoot.at-indic

the-obv

thunder-obv

 

‘Then two young men would shoot at the thunder.’ 

b. 

n=ʔas-iⱡ 

ʔupiⱡ-iⱡ-ni 

ʔa•qⱡ¢maknik’ 

 

ind=two-preverb

kill-pass-ind

Indian 

 

‘Two Indians were killed.’ 

Languages like Kutenai which have both constructions are coded according to the construction in which the numeral combines with the noun; hence Kutenai is shown as placing the numeral before the noun. 

2. Geographical distribution

The two orders of numeral and noun show very clear geographical patterns. There is a large area in which almost all languages place the numeral before the noun, which covers most of Europe and Asia except for an area in Southeast Asia described below, extending southeastward into Indonesia, the Philippines and western Micronesia and extending southwestward into the Middle East and across North Africa. It is also the dominant order in the Americas, especially in the westernmost regions of the United States and Mexico. 

The order with the numeral after the noun is the overwhelmingly preferred order in sub-Saharan Africa, and the majority of exceptions are geographically peripheral within this region, namely in the far south of Africa, in the extreme northwest (in the vicinity of Senegal), and in the northeastern region of Ethiopia and Eritrea. This order also predominates in an area within Southeast Asia, but the boundaries of this area are different from the boundaries of the Southeast Asian area found with other word order characteristics. In particular, it excludes the languages of Vietnam and languages in southern China to the north and northeast of Vietnam and it excludes the languages of Indonesia and the Philippines. Conversely it extends further to the northwest than some other word order characteristics do, including Tibeto-Burman languages of northeast India, Burma, and adjacent areas of China. Even among the more western Tibeto-Burman languages, there is a mixture of the two orders of numeral and noun. Noun-numeral order is also the dominant type in New Guinea and among the languages of the easternmost islands of Indonesia.

Both types are interspersed amongst each other among the languages of the Pacific, including the eastern islands of New Guinea and in Australia. The two orders are about equally common among the languages of Canada and the United States east of the Rocky Mountains, with a notable pocket of noun-numeral languages in the southeastern United States. 

Languages lacking a dominant order of numeral and noun are somewhat more common in Australia and in western Indonesia. 

3. Theoretical issues

In many languages, numerals must occur with classifiers, as in (11). 

(11) Nusu (Tibeto-Burman; southern China; Sun and Lu 1986: 44) 

va⁵³za⁵⁵ 

nga⁵⁵ 

gɔ³⁵ 

little.pig 

five 

clf 

‘five little pigs’ 

In such languages, the numeral is not really modifying the noun, at least not directly, but rather is modifying the classifier. For such languages, the map shows the order of the noun relative to this phrase consisting of numeral and classifier. Whether this phrase actually modifies the noun is itself something that is often unclear. It has been proposed for some languages that the classifier is the real head of the noun phrase and the numeral and noun are both dependents of the classifier. See Map 55A on numeral classifiers.

The issue of what element is the head also sometimes arises sometimes in languages without numeral classifiers. It has been proposed that some numerals in Russian serve as heads (Babby 1987). Similarly, in Rif Berber (Morocco), most modifiers follow the noun, while numerals precede; but the construction they occur in is the same as the genitive construction, as illustrated in (12), suggesting that the numeral is the head, not a modifier. 

(12) Rif Berber (Kossmann 2000: 160, 108) 

a. 

ṯlaṯa 

təwrar 

b. 

axxam 

wəryaz 

 

three 

gen 

hill 

 

house 

gen 

man 

 

‘three hills’ 

 

‘the man’s house’ 

The issue of which element is the head is ignored in coding languages for this map. 

There are other languages in which the numeral can be argued not to be a modifier of the noun. One type is represented by Kutenai (isolate; British Columbia etc.), in which the numeral is the verb in an internally-headed relative clause. This is illustrated above in (10a), where the noun ni¢tahaⱡ  ‘young man’ is grammatically the subject of the verb ʔas  ‘two’. For the purposes of this map, languages are coded according to the order of numeral and noun, regardless of whether the relationship is one of modification or not; since the normal order is that shown in (10a), Kutenai is coded as placing the numeral before the noun. The fact that the numeral is grammatically a verb in Kutenai (as it is also in Nias in (4) above) is ignored for the purposes of this map.