This map shows the extent to which plural markers on full nouns are used in a language (full nouns are nouns which are not pronouns; see Chapter 35 for the plural of personal pronouns). Plural markers can have different formal manifestations (see Chapter 33), which play no role in this chapter. Here we look only at the occurrence of plural markers, which varies along at least two dimensions: animacy and obligatoriness. In the animacy dimension, the most important contrast is between animate (especially human) nouns and inanimate nouns, which are treated differently in many languages (see §2.1 for animal nouns, which are neither human nor inanimate). In the obligatoriness dimension, I distinguish here between non-occurrence, optional occurrence and obligatory occurrence. When these two dimensions are combined, we get the six values shown in the feature-value box.
|No nominal plural||28|
|Plural only in human nouns, optional||20|
|Plural only in human nouns, obligatory||40|
|Plural in all nouns, always optional||55|
|Plural in all nouns, optional in inanimates||15|
|Plural in all nouns, always obligatory||133|
An apparently exceptionless generalization is that human nouns are more likely to have plural marking than non-human (espcially inanimate) nouns. Thus, three further logically possible types, in which inanimate plurals would occur more widely than human plurals, are not attested:
7. Plural only in inanimate nouns, obligatory
8. Plural only in inanimate nouns, optional
9. Plural in all nouns, optional in human nouns
It should be noted that we are talking here about the occurrence of plural marking, not the occurrence of plural meaning. Non-occurrence of plural marking does not mean that only a singular meaning can be expressed, but that a non-number-marked form of the noun is used both for a single referent and for a group of referents. For example, in Shigatse Tibetan, a language with no plural marking on full nouns, the noun ri can be translated into English as ‘mountain’ or ‘mountains’ (Haller 2000: 59, 114). While English speakers must express the number of entities in such a case, Shigatse Tibetan speakers are simply vague with respect to the number of entities (if they wish to be explicit, they can use numerals and quantity words like ‘many’). Similarly, in Baka (Ubangi, Niger-Congo; Cameroon), where all nominal plurals are optional, the noun form gba can be translated as ‘village’ or ‘villages’. Speakers can be explicit and use the plural form gba-o ‘villages’, but they do not have to (Kilian-Hatz 1995: 21, 95).
The first type shown on the map comprises those languages that have no nominal plural at all. This is not uncommon in languages with little or no inflection, but there are also quite a few languages with fairly rich inflectional systems that simply happen to lack nominal number marking. For instance, Gurr-goni (Burarran; Northern Territory, Australia) has nominal case-marking inflection and rich verbal morphology, but no noun plurals (Green 1995: 69).
Languages of the second and third type have a nominal plural only in human nouns and lack a plural of inanimate nouns. In the second type, this human-only plural is optional. An example of such a language is Hatam (West Papuan; Indonesia), where human nouns like munggwom ‘child(ren)’ may but need not take the plural clitic =nya when reference to several people is intended (Reesink 1999: 50-51, 65). In the third type, the human-only plural is obligatory. For instance, in Jamul Tiipay (Yuman; southern California) kinship terms and other human nouns have plural forms (e.g. xechany ‘girl’, xaachaany ‘girls’) which must be used, while inanimate nouns lack plurals (Miller 2001: 114-115).
Languages of the remaining three types have plural marking in all nouns, i.e. human nouns and inanimate nouns. In the fourth type, this marking is always optional (i.e. both in human and in inanimate nouns). An example of this is Tetun (Central Malayo-Polynesian; Timor), where the plural word sia (also used as a 3rd person plural pronoun ‘they’) can combine both with human and with non-human nouns (e.g. asu ‘dog(s)’, asu sia ‘dogs’; kaheke ‘fan(s)’, kaheke sia ‘fans’; van Klinken 1999: 124-125, 135). In the fifth type, plural marking is obligatory in human nouns and optional in inanimate nouns. This is exemplified by Arawak (Arawakan; Suriname), where human nouns require the plural suffix -non (hiaro ‘woman’, hiaro-non ‘women’) while inanimate nouns allow the plural suffix -be (siba ‘stone(s)’, siba-be ‘stones’; Pet 1987: 34-37). Finally, languages of the sixth type have obligatory plural marking in all nouns, human and inanimate. This is exemplified by English and other European languages; another example is West Greenlandic (e.g. nuna ‘land’, nuna-t ‘lands’; Fortescue 1984: 247).
The definitions of the feature values that I have just given involve two pairs of concepts that require further comment.
For the purposes of the map, I looked only at nouns denoting humans and nouns denoting discrete inanimates. Non-human animates, i.e. nouns denoting (non-human) animals, pattern differently in different languages. In some languages, animals are treated like humans; an example is Korku (Munda; central India): siṭa/siṭa-ku ‘dog/dogs’, like koro/kor-ku ‘man/men’, as opposed to pluralless inanimates (Nagaraja 1999: 31, 35). In other languages, animals are treated like inanimates; an example is Taba (Eastern Malayo-Polynesian; Maluku, Indonesia): manik ‘chicken(s)’, like llu ‘leaf/leaves’, as opposed to human nouns with obligatory plural marking, e.g. wang/wang=si ‘child/children’ (Bowden 2001: 190-191).
So this is an interesting parameter of variation, and the main reason animal nouns are disregarded here is that many grammars provide no information about them. In some languages, the class of animals itself is subdivided further, with "higher" animals patterning like humans and "lower" animals patterning like inanimates. For instance, in Tiwi (Tiwian; Northern Territory, Australia), three animal names (‘dog’, ‘dingo’ and ‘goanna’) have plurals like human nouns (Osborne 1974: 52).
In some languages, not all human nouns are treated alike. For instance, in Amele (Trans-New Guinea; Papua New Guinea), only kinship terms have plurals (e.g. cot-i ‘my brother’, cot-i-el ‘my brothers’; Roberts 1987: 171). In other languages, the subset of human nouns with plurals does not fall under any semantic generalization and is purely lexically determined. For the purposes of the map, such cases were not distinguished from languages where all human nouns have plurals.
Finally, in many languages not all inanimates are treated alike: nouns denoting nondiscrete masses such as ‘sand’, ‘salt’, ‘water’, ‘milk’ often lack plurals even in languages that otherwise obligatorily mark the plural of inanimates (e.g. in English). It is possible that this case is somewhat different from the non-occurrence of plurals in discrete nouns, because English speakers find it difficult to imagine what the semantic difference between milk and milks would be, especially if we leave aside the sort reading (‘different sorts of milk’). However, there are quite a few languages in which speakers find it easy to form and interpret plurals of nondiscrete nouns, for instance in Evenki (Tungusic; eastern Siberia), where such plurals are said to denote a large quantity or several pieces (e.g. ulle ‘meat’, ulle-l ‘a lot of meat, pieces of meat’; Nedjalkov 1997: 190).
Thus, it is likely that a more fine-grained study of plurals would find considerable evidence for a richer implicational scale as in (1).
kinship terms > other humans > “higher” animals > “lower” animals > discrete inanimates > nondiscrete inanimates
This scale should be read as follows: if a noun subclass has plurals, then so do all subclasses to its left on the scale. And if a noun class has obligatory plural marking, then so do all noun subclasses to its left on the scale.
In principle, obligatory occurrence of plural marking means that whenever plural reference is intended, the plural must be used. However, languages with otherwise obligatory plural marking very often allow it to be absent when the noun cooccurs with numerals and other quantity words. For instance, in Lezgian (Daghestanian; eastern Caucasus), the plural is normally obligatory, but it is impossible when the noun is accompanied by a numeral (e.g. nük'-er ‘birds’, but c'ud nük' ‘ten birds’, not *c'ud nük'-er ; Haspelmath 1993: 232). In Itzaj (Mayan; Guatemala), the plural is normally obligatory, but it is optional with numerals (Hofling 2000: 227). In Brahui (Dravidian; western Pakistan), the plural is normally obligatory, but is not used with numerals in indefinite NPs (e.g. irā bandagh(*-āk) ‘two persons’, but 'amē bīstangā bandagh-āk ‘those twenty persons’; Andronov 1980: 34-36). When a grammar says that the plural is not used when a quantity expression is present, this is not regarded as evidence for general optionality of the plural; in fact, it can be taken as indirect evidence for obligatoriness, because if the plural were genuinely optional, no special statement about quantity expressions would be necessary. So, like animacy, the dimension of obligatoriness consists of more distinctions than are made here, but since many grammars do not say what happens when a noun is combined with a numeral, it was necessary to simplify the picture and loosen the criteria for obligatoriness. Another special circumstance under which plural marking is often lacking is non-referential or generic use. For instance, Evenki allows singular nouns in contexts like (2).
‘I hunt for squirrels.’
Again, when a grammar mentions optionality of plural marking in such contexts, this is not taken as evidence for genuine optionality here.
Plural marking is classified as optional when it is explicitly described as being optional, and if this optionality clearly goes beyond the non-occurrence of plurals in the two special contexts just mentioned. In many cases, it is probably possible to identify further conditions on the use of plurals. For instance, after noting that the plural word àwọn is optional in Yoruba, Rowlands (1969: 41) goes on to explain that the plural marker is used if the referent is thought of as a group of individuals, whereas no marker is used if it is thought of collectively (e.g. àwọn ìwé mi [book my] ‘my (various) books’, vs. ìwé mi [book my] ‘my (collection of) books’ or ‘my book’). For Udihe (Tungusic; eastern Siberia), Nikolaeva and Tolskaya (2001: 116) observe: “The Plural of a subject, especially when denoting a person, is more or less regularly marked, whereas the Plural suffix of a direct object appears less frequently.” Occasionally, the conditions are described in a very precise way. In Kâte (Finisterre-Huon; Papua New Guinea), a human noun can (and indeed must) have plural marking only if it is possessed (e.g. motec ‘boy(s)’, motec-fâc-ticne [boy-pl-3sg.] ‘his boys’; Pilhofer 1933: 43, 56, 113). In this case, there is strictly speaking no optionality; rather, the plural suffix -fâc is obligatory when the noun is possessed and obligatorily absent when it is not possessed. Still, whenever the variable appearance of plural marking depends on a set of factors unrelated to quantity expressions and non-referentiality, I classify the language as showing optional plural marking, because it would not be possible to systematically distinguish subtypes of optionality.
Unfortunately, a large number of grammatical descriptions do not say whether plural marking is obligatory or optional. I suspect that in most such cases, the absence of a statement about obligatoriness is due to the fact that plural marking is obligatory in the language just as in the European languages that the readers are likely to be familiar with. Authors of such descriptions may not even be aware that obligatoriness of plural marking is an issue in many languages. Grammar writers are much less likely to omit a statement about obligatoriness/optionality if plural marking is optional, because in this case their language is interestingly different from the European languages. For this reason, I decided to take obligatoriness of plural marking as the default situation and to classify a language as obligatorily plural-marking even in the absence of an explicit statement, unless the description is very sketchy or there are other reasons for suspecting that plural marking may not be obligatory (if this was the case, I excluded the language from the sample). It is possible that this decision introduces a bias into the classification, and that some languages classified as obligatorily plural-marking have in fact only optional plural-marking.
Obligatory plural marking of all nouns is found throughout western and northern Eurasia and in most parts of Africa. The rest of the world presents a heterogeneous picture. Optional plural marking is particularly common in Southeast and East Asia and Australia, and complete lack of plural marking is particularly found in New Guinea and Australia. In addition to the areal correlations, there also seems to be at least one correlation with the general morphological type: Isolating languages (i.e. languages with little inflectional affixation; see Chapter 22) appear to favor the lack or non-obligatoriness of plural marking. This can be seen particularly in Africa, where optionality or absence of plural marking is found particularly in the isolating languages of West Africa.
speaker > addressee > 3rd person > kinship terms > other humans > “higher” animals > “lower” animals > discrete inanimates > nondiscrete inanimates
The explanation for the role of animacy in plural marking seems to be the fact that the distinction between one and more than one is more salient for animates than for inanimates, so that speakers are more likely to make use of available plural markers when they refer to a plurality of animates. Through grammaticalization, this preference in language use can lead to obligatoriness in language structure.