Skip to navigation – Site map

Plurality in English and other languages: does it add up?

Wilfrid Rotge
p. 101-120

Abstract

Cet article a pour objectif d’analyser le fonctionnement du pluriel nominal en anglais et dans quelques autres langues, en montrant systématiquement les limites de l’opposition traditionnelle singulier / pluriel. Certaines langues n’ont pas de pluriel nominal, d’autres possèdent un système ternaire, voire quaternaire. La définition du pluriel peut varier d’une langue à l’autre. L’anglais ne le conçoit pas de la même façon que le français, par exemple. Bien que cet article ne s’inscrive pas explicitement dans le cadre de la grammaire cognitive, certaines implications cognitives du pluriel et de son marquage seront considérées et certains principes remis en cause, notamment celui de la motivation et du caractère non-arbitraire des nombres.

Top of page

Full text

Introduction

1My aim in this paper is to examine how plurality functions in English and other languages. I will start off with an analysis of the singular / plural dichotomy and point out its limitations. Such a binary system provides an incomplete picture, which requires refining, if only because some languages have a ternary or even quaternary system or because there is no strict overlapping between the grammatical singular / plural distinction and the semantic one.

2Next, the various markings of the plural will come under scrutiny, which will allow me to point out their cognitive implications, that is, we will see how the plural can be envisaged from a cognitive point of view in various languages.

3Plurality is not just a question of morpho-syntax or word-endings. In the final section, I will deal with numerals and what they tell us from a cognitive point of view.

4More generally, this paper will also examine the body-based model championed by cognitivists and a typically cognitive question that may be asked about number and numerals: what do they tell us about the way we view our surrounding reality? The problem of motivation and non-arbitrariness will also be raised, as will the question of universals: is the singular / plural dichotomy universal? Are singular and plural necessary categories?

1. The singular / plural dichotomy

1.1. Plurality and non-singularity

5The English word plural comes from the Old French plurel, itself derived from the Late Latin pluralis, which means “concerning many”. In plural we find the Latin word plus, which of course means “more”. In other words, the notion of plurality is based on the idea that when using a plural we mean “more” and probably “more than one”, thus suggesting that plurality is a relative concept, which is not entirely quantitative and from which degree is not absent. More precisely, the plural is understood in relation to the singular; it is “more” relative to the singular, a word which is a cognate of the word single: singular comes from the Latin singularis, meaning “single”.

6We may infer from this that the notion of plurality is less stable and more vague than that of singularity, which is not surprising given that plurality encompasses anything from ‘2’ to infinity. In English, as will be seen later, the plural also encompasses ‘0’. This relative instability may account for some erratic plural markings that will also be dealt with later.

7The relationship between plural and singular is more complicated than the often postulated singular vs. plural dichotomy would have us believe.

1.2. The dual, duelling with the plural

8Many languages do not restrict themselves to the two-term system opposing the singular and the plural and include a third term, called the dual, which is used to refer to two entities. Some verbs can also have a dual agreement form. In languages that have a dual form the plural is generally restricted to entities greater than two, although in Ancient Greek dual forms are optional.

9For example, Hopi distinguishes mima maanat (those girls - dual form) and mima mamant (those girls – plural form, necessarily more than two girls)2 and Sanskrit also has a dual (Croft 2003:141).

10There seems to be a tendency for languages to drop the dual as they evolve, in which case the plural takes over the dual. Conversely, in less developed societies, smaller figures, especially from ‘1’ to ‘10’, are more central to their culture and it therefore makes more sense to fully distinguish singularity, duality and plurality. It is useful to recall that some Amazonian or Aboriginal languages contain few number words. The Andamanese, who live on the Andaman Islands, a union territory of India, are said to count no higher than ‘2’ (Heine 1997:24). Dixon (1980:107-8) points out that “[t]he one obvious gap in Australian vocabularies is the lack of any system of numbers”, and notes that most aboriginal languages only have numbers for ‘1’, ‘2’, ‘several’ and ‘many’.

11It would even seem that one Amazonian language, spoken by the Pirahã tribe, has no number words, not even for ‘1’, which is extremely rare3. It is more common for a language to have a small quantity of number words, which refer to low numbers, in which case ‘1’, ‘2’ and ‘3’ acquire special status, which justifies all the more the existence of the dual.

12This tendency for languages to drop the dual is evidenced by Slavic languages. The dual was very common in these languages at the beginning of the second millennium, but only remnants of this are still to be found in modern Slavic languages, except in Slovene, which has retained full grammatical use of the dual: ena miza (= one table); dve mizi (= two tables); tri mize (= three tables).

13The same can be said of Germanic languages insofar as the dual was present in the early Germanic languages. In Old English, the dual was only present in some pronouns used in reference to two entities. In other words, there was no dual inflection for nouns.

14Some Old English texts distinguish between wit (= we two) and we (= we more than two), as in the lyrical poem “The Wife’s Lament” (Marsden 2004:339). The object pronoun and possessive determiner are also distinguished along these lines: unc [dual] vs us [non-dual] and uncer [dual] vs ure [non-dual].

1.3. The dual’s days are numbered, in English

15Despite the general tendency for the dual to disappear, remnants of it are found in contemporary English, although none of these concern noun endings. One such remnant is the word both, which implies a totality restricted to two items. It can be contrasted with all, which also expresses totality but applied to a set higher than two: All three decided to go vs *All two decided to go (OK Both decided to go). This dual word competes with the two, but that is beside the point. In Romance languages, such as French, a literal translation of all two is syntactically possible: tous (les) deux.

  • 4 -tween in between is related to two, twain.

16The other remnants are either / neither, whether, between and latter. Whether it is used as a conjunction or a determiner, either implies a choice between two items: Either you pass or you fail (only two possibilities exist); You can park on either side of the road (there are only two possible sides). Whether is a cognate of either. The dual form latter can only compare two items, unlike the non-dual word last which can only compare more than two items. Despite its etymology4, the preposition between is not strictly dual as such because it is not restricted to two (Between us four it’s over 1,000 euros), but it can be contrasted with among, which is non-dual (*Among us two vs Between us two). The preposition between can be described as “paucal”, a term which will be dealt with later.

17To the list of anti-dual items (among, all...) may be added the quantifier every, in that it is not compatible with duality (*every one of my parents /OK every one of my relatives /OK each of my parents).

18As was suggested before, the use of the dual is “slipping”. This is borne out by the two constructions the nicer of the two and each other. The dual construction the nicer / bigger / more intelligent of the two tends to be replaced by the more regular the nicest / biggest / most intelligent of the two, to the extent that the superlative used for a set of two is no longer considered as grammatically flawed by some grammarians. Be that as it may, the construction “the + comparative” could be referred to as a dual superlative in that the nicer of the two can be construed as a superlative reduced to a set of two items.

  • 5 Some style guides in the United States still advocate this distinction, although Merriam Webster’s (...)

19Until twenty or thirty years ago, students were taught that each other was the dual reciprocal pronoun (Luke and the boss were talking to each other) and one another its non-dual counterpart (The ten women were talking to one another). Even though some teachers still cling to this distinction they have to face the fact that it is no longer valid. Nowadays, both each other and one another are used equally for a set of two or a larger set5.

20Even the quantifier a couple of, which typically refers to two items, is sometimes applied to a slightly larger set of items, as in A couple of people voted against it, where it is more or less synonymous with a few. Originally it was restricted to a dual use.

21Whether in Old or Modem English, the dual never concerned the number inflection of nouns.

22Unlike English, contemporary French does not seem to have dual forms. It should be remembered that the dual was present in all the early Germanic languages and so remnants are more likely to be present in these languages than in Romance languages.

1.4. Dual, trial, paucal, plural et al.

  • 6 Tolomako is a language of the Oceanic subgroup of Austronesian languages. Lihir and Manam arc langu (...)

23Beyond the three-term system (singular, dual, plural), some languages have developed an even more complex one with the existence of the trial number, that is, a plural restricted to a set of three, thus evincing a four-term system: singular, dual, trial and plural. In this case, the plural is used for reference to four or more items. There does not seem to be a number based on four items, a “quadral” (see Corbett 2000:26-30). Tolomako, Lihir, Manam and Tok Pisin (though only in its pronouns) have trial number6. It must be added that the trial is rather unusual.

24Some languages have a paucal number, which co-exists alongside the singular and the plural. The paucal is used to refer to some or a few items, usually from ‘3’ to around ‘10’. This word is a cognate of the noun paucity and both are derived from the Latin paucus, which means “few”. Manam, a language spoken in Papua New Guinea has a paucal (Lichtenberk 1983:276). The English determiners some or (a) few could be termed “paucal” in theory but this notion usually refers to the plural marking of the noun.

25The counting system of Russian could be said to include a paucal, even though to the best of my knowledge this metalinguistic term is not usually applied to Russian. Indeed, ‘1’ is followed by a noun in the nominative; after ‘2’, ‘3’ and ‘4’ the noun goes into the genitive singular; all other numbers are followed by the genitive plural. In other words, number inflection in Russian varies depending on how many referents there are. ‘2’, ‘3’ and ‘4’ are apart in that, although they do not refer to a singular item, they are followed by a noun in the singular, but contrary to ‘1’ the noun takes the genitive, so that ‘1 child’, ‘2/3/4 children’, ‘5 children’ read as follows in Russian: ‘1 child’, ‘2/3/4 of child’, ‘5 of children’. It can be seen that ‘2’, ‘3’ and ‘4’ constitute an intermediate category between the singular (‘1’) and the “real” plural (‘5’ and beyond) and, in a sense, can be considered as a paucal.

26As I mentioned before, only one known language seems not to have numbers, that is, the Amazonian language spoken by the Pirahã tribe, in which counting is reduced to a three-term system, i.e. ‘few’, ‘some’, ‘more’7. According to Franck, Everett, Fedorenko and Gibson (2008:819-824), it is the only known language not to have a word for ‘1’. The notions of singular, dual, trial, paucal or plural seem totally alien to the speakers of this language.

27Even though one of the aims of this paper is to question the supremacy of the singular/plural distinction, it must be added that no language has a trial number unless it has a dual (Croft 2003:50) and no language has a dual unless it has a plural. In other words, the plural is far more widespread than the dual, let alone the trial.

1.5. Natural number? The cognitive hypothesis

28The singular/plural distinction is undeniably very common but it is not universal. To start with, some languages do not have it, like Japanese or Malagasy. In Malagasy nouns do not inflect for number, although pronouns and demonstratives do (Montagne 1931:22). Furthermore, we have just seen how the distinction is challenged by intermediate categories like the dual, trial or paucal.

29When the evolution of languages is taken into account, it might even be tempting to say that a three-term system (singular / dual / plural) would be more “natural” than a two-term one (singular / plural), because of the seminal role of the body (especially the hands) in the development of counting. According to many cognitive linguists (see Heine 1997:18-34), the two hands of the human body provided the conceptual source for lower base numbers. It is therefore easy to imagine that numbers like 2 or 5 played a central role in the early stages of counting, a central role that was “naturally” translated into the grammatical marking of the dual or even the trial or the paucal.

30My hypothesis is that with the evolution of society, dealing with larger numbers became the norm and the central role of ‘2’ was challenged to the extent that the dual became extraneous and disappeared in many languages to merge with the plural.

31In languages in which the dual has disappeared, it is possible to view this change as a simplification of the system to a binary one, that is, singular vs plural, which, again, is neither universal nor “natural”.

2. Plural inflections

2.1. The fishy plural in grammar books

32Number is not something grammarians of the English language are particularly interested in, probably because the expression of number in that language seems fairly straightforward.

33Some basic grammar books limit their references to the existence of the plural marker -s, a marker added to the majority of nouns, and fail to point out that this marker is only one of the nominal plural markers of the English language, albeit by far the most common. When it comes to the frequently used noun children, it is treated only as an exception and not as an “Ν + plural suffix

34, which makes sense, given the very restricted number of nouns that fall into the “Ν + -en plural marker” category. Even the plural form children is increasingly challenged by the regular plural noun kids. The tendency of the English language to make words enter a regular pattern is at work here too.

35In the best-case scenario, grammarians will list the rather limited number of nouns which, for historical reasons, are invariable, like most fish names (two fish, three trout) or the well-known aircraft, sheep, deer, swine and a few others such as bear or buffalo which can inflect for the plural marker or not. The invariability of many nouns is attributable to the Old English declensions they belonged to. However, it is difficult to explain why so few nouns retained their Old English invariability, while the vast majority of Old English invariable nouns (i.e. invariable in the nominative and accusative plural) became regular, like the nouns word or thing. This regularisation process of English nouns is partly due to the influence of the French language during the Middle English period (Stévanovitch 2008:67). Paradoxically, some animal nouns like trout or salmon (from Old French truite / salmun) borrowed from the French after 1066 are also invariable, probably because they belong to that category of animal nouns that is so typically invariable, i.e. fish.

  • 8 The use of a singular in He’s six foot tall may be attributed to an adjectival interpretation of th (...)

36Two nouns are invariable in spoken English but variable in a more standard variety of the language: pound (I can give you fifty pound / fifty pounds) and foot (He’s six foot / feet tall8).

37The noun head is traditionally invariable when used to count animals: thirty head of oxen. The animal kingdom (with its fish, hunted animals and head of oxen) is once again grammatically singled out.

38Grammarians who bother to mention invariable nouns also tend to draw up more or less exhaustive lists of irregular nouns, some of which can also be regular like cactuses vs cacti or formulas vs formulae. There are very few native irregular nouns left. The best-known patterns in the native irregular form is the stem vowel change oo > ee (foot > feet), with boc (Mod. Engl, book) for example being inflected as bec (Mod. Engl. books), but which only survives in foot, goose and tooth and the - ouse > -ice change, which is only found in mouse and louse in modern English, the former being somewhat challenged by mouses, which exists alongside mice in computer parlance. I will come back to the stem vowel change to point out what it may teach us from a cognitive point of view.

39The -en plural only survives in oxen, children and brethren. As this latter noun is now archaic, it could even be argued that this plural morpheme only survives in two nouns. The noun children actually includes two overt number markers from an etymological point of view: <r> in this word is the trace of an Old English plural marker, not a highly productive one. In Old English, the form was cildru, to which the more productive -n plural marker was added, possibly because <r>marker was no longer felt to be a plural marker. By contrast <r>plural marker is highly productive in German and remains one of the several plural markers in modern German (Kind > Kinder; Buch > Bücher). As a result, a plural like *Kinderen could not be coined in that language.

40Interestingly, there are many more non-native irregular nouns (phenomena, cherubim, indices...) than native ones.

2.2. Collective nouns: e pluribus unum?

41Still on the subject of number, most grammar books will mention the ambiguous status of collective nouns, when it comes to the use of a singular or plural verb form. Here again things are not as clear-cut as the singular / plural distinction would have us believe.

42We all know that collective nouns like government or team take a singular or a plural verb depending on whether they are conceived as global entities or a set of individuals. This obtains mainly in British English. In American English, things are clearer: a singular collective noun takes a singular verb. The following pairs can be contrasted:

It appears the government are selectively choosing to privatise parts of the road network. [Brit. Engl.]
It appears the government is selectively choosing to privatise parts of the road network.
[Am. Engl, or Brit. Engl.]
I was looking forward to one of those quiet, soothing journeys of the kind that British Rail are always promising. [Notes from a Small Island,
B. Bryson]
I was looking forward to one of those quiet, soothing journeys that British Rail is always promising.
[Am. Engl.]

43National teams are typically plural in British English, but not in American English:

England have failed to reach the World Cup final.
Italy qualify for the Euro.
Who are Aston Villa playing today?

44This plural use of a verb with a singular collective noun is impossible in many languages: *La France se sont qualifiés vs La France s’est qualifiée.

45The collective noun family stands out in American English, since, unlike its other collective counterparts, it may be followed by a plural verb, even though the singular seems more usual:

My family have moved to Texas.

  • 9 Things are slightly more complicated with the noun police. Most grammar books consider that it cann (...)

46To add to this confusion, it may be added that four collective nouns are always plural, in all varieties of English: people, police9, cattle, vermin, so that they can be considered as intrinsically plural nouns, even though they are not plural nouns from an etymological point of view and do not betray any plural features in their morphology. These four nouns are different from government or family, which are intrinsically singular, in that the latter have a plural form (governments, families). Conversely, people, police, cattle and vermin do not have a plural form: the words * cattles or *vermins do not exist. The plural form peoples does exist but with a different meaning. The word polices does exist but as a verb.

47Collective nouns challenge the apparent clarity and simplicity of the singular / plural distinction, in that they show that singular nouns do not always behave singularly.

48Furthermore, if some singular collective nouns may take a plural verb, some plural nouns take a singular verb, as in:

Is the United States going broke?
The Netherlands is a parliamentary monarchy, [sometimes used with a plural verb]
Where is that ten dollars I gave you?

Two and three is [or are] five.
More than one person is going to resign.
There’s people who wouldn’t agree there, [informal]

49This state of things can easily be accounted for (put briefly, unity prevails behind the apparent plurality) but that would be beyond the scope of this article. What is more, some nouns that are historically plural can behave like singular nouns, like darts, skittles, customs or means: Darts is a predominantly British game.

50It is well known that potentially plural nouns used adjectivally are not n umber-inflected, even with nouns that are always intrinsically plural, i.e. pluralia tantum, like trousers in trouser-Ø press. Nevertheless, the current tendency is on the contrary to add the plural marker to nouns used adjectivally as in antiques shop (or antique shop) and drugs problem, which is replacing drug problem.

51This linguistic phenomenon betrays a discrepancy between the grammatically-motivated singular and the semantically-motivated plural and further challenges the neatness of the singular / plural dichotomy.

2.3. The plurality of the plural

52As I mentioned earlier, plural marking in English goes beyond the -(e)s ending, with plural markers in -Ø (sheep), -e (formulae), -i (cacti), -im (cherubim), -x (tableaux), -en (oxen), stem vowel change (mice) or more radical change (penny > pence). However, some of these nouns have a regular plural: formulas, cactuses, cherubs, pennies, tableaus.

53Despite these different markers, the tendency of English since 1066 has been towards the simplification of its plural system, with -(e)s (and its three different phonetic realizations) representing the quintessential plural marker, especially when compared with German, where plural markers are scarcely more frequent than in English but have all remained extremely productive.

54As said before, the influence of the French language precipitated this simplification process. In Old French, the -s marker was pronounced. It was originally mainly used for the plural accusative, the nominative plural being in -Ø: li pere, sing. nom. or pl. nom.; les peres, pl. acc. (Ο.F. pere =father).

55Many languages have several plural markers, be they native, as in German sg. Mann > pl. Männer, or borrowed, as in Handy (= mobile in German) > Handys. However, that does not mean that all languages perceive the plural the same way. For example, the conception of the plural is more grammatical in English than in French, as can be seen with decimals. In English, it is the presence of any numeral other than ‘1’ that triggers the plural, as in 1.5 kilometres. In French a singular noun would be used here: 1,5 kilomètre. More surprisingly, even with figures below Ί’the plural is used in English: 0.5 kilometres, which is grammatically justified by ‘5’ even though semantically 0.5 should be treated as a singular, as in French 0,5 kilomètre.

  • 10 In the same way the quantifier no is followed by a plural count noun: Mirren has no children and sa (...)

56The plural is even used with zero: zero degrees Celsius, which can be contrasted with zéro degré (singular) in French10.

57We thus see two slightly different conceptions of the plural: in English the plural is reserved for entities other than ‘1’; in French the plural starts with ‘2’.

58It can be argued from this fact that the-s marker cannot be defined just as a plural marker in English; it is more strictly speaking a non-singular marker used whenever the referent is not limited to ‘1’. Croft’s more usual definition of the English plural “in English ‘plural’ means ‘more than one’” (2003:141) is therefore questionable.

59The distribution of the singular and plural forms can be represented as follows, using the “structure of the conceptual number space” devised by Croft (2003:141) as applied to Sanskrit:

sing. Form

plural form

French

1- l.xxx

2 – 3 – 4 – 5 –....

English

1

0.-notl, l.-notl 2 – 3 – 4 – 5 –....

60This means that the singular in French goes from 1 to 1.9999...

61In English, the singular is restricted to 1 and does not include 0.9 (which corresponds to “0.notl”).

62l.-notl” in this diagram means that 1.07 for instance is followed by a plural noun (1.07 miles) and that 1.1 takes a singular noun (1.1 mile).

2.4. The plural makes its mark

63It is often said that the singular is the least marked of the various number categories: the dual, trial, paucal, plural (Croft 2003:156). In Indo-European languages we can even say that the singular is unmarked. We can thus compare book, Bitch [German], livre [French] with books, Bücher, livres. French provides two exceptions to this generalization, boeufs and oeufs: the singular of these two nouns is more marked than their plural, as will be seen later, which is extremely unusual for any language (Croft 2003:89).

64Some singular nouns bear the plural marker because they were plural nouns originally, like means, customs, darts but that does not mean that the singular is marked in these nouns.

65In Hungarian, a non-Indo-European language, the same phenomenon can be observed: the singular is the unmarked form to which a plural marker in -(e)k may be added: gyerek (= child) > gyerekek (= children).

66By contrast the plural is expressed by an overt inflection in Indo-European languages, even though we know that there are invariable nouns in these languages, like sheep in English or Messer (knife) in German. This distinction between an unmarked singular noun and a marked plural noun is often considered to be a universal tendency, despite the fact that in some languages, like Japanese or Malagasy, nouns are not marked for number. It should be noted that in some languages both singular and plural are marked, but this is still a matter for debate, as in Zulu (Doke 1930: 36, quoted in Croft 2003: 88). The singular sometimes has to be marked when a noun is intrinsically plural in a language, as in a pair of trousers for the intrinsically plural noun trousers, but this does not invalidate the unmarked nature of the singular in English.

67The plural may make its mark but it is “no match” for the singular, given that singular noun phrases far outweigh plural noun phrases. This idea put forward by Greenberg (1966) has led to another one posited by Croft (2003: 111): “human beings talk more about singleton sets of entities than plural sets of entities”. In other words, human beings would tend to view the world as a place in which “singletons” outweigh plural sets.

2.5. Quiet French

68Few people would question the fact that Indo-European languages have a marked plural. However, this is only partly true of French, which, though it has a visible prototypical plural marker in -s, no longer has an audible plural marker, unlike all the other Indo-European languages. Indeed the NPs train rapide (= fast train) and trains rapides (= fast trains) are pronounced exactly the same way.

69The few nouns that do change in their plural form are nouns ending in -al or -ail: cheval > chevaux (horse); travail > travaux (work), but not all of them (festival > festivals; portail > portails). Surprisingly, two nouns drop their final consonant in the plural: boeufs (= oxen) and oeufs (= eggs) are pronounced as boeu or oeu whereas the -f is sounded in the singular.

70As it is very difficult for most people to dissociate their language from writing, most French speakers would instinctively deny that their language has simplified its nominal system to the extent that the plural can be said to be unmarked in contemporary French. That is however exactly how a linguist would describe French if it was not a written language.

71Remnants of the -s plural, which as I mentioned earlier was pronounced in Old French, are found in the linking -s before a vowel: trains électriques is pronounced trains-z-électriques in careful French, with a sounding of the plural marker -s, but trains électriques is also very often pronounced like the singular NP train électrique.

72Oral French can therefore be described as a language with unmarked plural nouns, with a few exceptions (chevaux, trains-z-électriques). In other words, the vast majority of French nouns no longer have overt inflections for the singular and the plural.

73That being said, it could be argued that there is a covert plural marker in oral French, because arbres in deux arbres (= two trees) is definitely felt to be a plural noun and because the plural is definitely a grammatical category in French. Also, determiners are often present to remind us of the plural category (le vs les or un vs des) and the adjective is clearly heard in its plural form in les beaux arbres, pronounced les beaux-z-arbres: the linking /z/ is not optional between a plural adjective and a noun that starts with a vowel sound, which adds to the idea that a French noun in a plural NP is felt to be plural even if the plural marker is not heard. Hence this notion of covert plural marker to be found in oral French.

74Consequently, French provides a slightly different grammatical scenario from the languages that have no singular-plural distinction regarding nouns.

2.6. Not stating the obvious, especially in Hungarian

75Mass nouns in many languages are singular in form even if they refer to several countable referents, which provides evidence that a singular form does not necessarily convey singular meaning. Famous examples of uncountable nouns are nouns like hair or furniture. In German the noun Haar (hair) is also used as a non-count noun.

76At the same time hair and furniture do not behave in the same way, in that the latter cannot be number-inflected (*furnitures) whereas the former may be: Darling, why are there three blond hairs on your black jacket?

77It is easy to imagine why the referent(s) of the noun hair may be conceptualized as a mass noun: the hair that grows on the head or body is visualized as a whole rather than as tiny parts making up a whole as in Romance languages (your hair > tes cheveux [pl.]).

  • 11 Hungarian always treats nouns as mass nouns when they come after numerals or quantifiers, as in sok (...)

78What is more puzzling with a language like Hungarian is that when the plural meaning of a noun is obvious then the noun is not number-inflected. Thus after the quantifier sok (= a lot of) the noun appears in its singular form, even though the referents are numerous: sok gyerek (= lit. many child) vs *sok gyerekek (gyerekek = children)11.

79The same logic is found with numerals. Nouns are not inflected after numerals irrespective of the number: két gyerek (lit. two child); *két gyerekek is ungrammatical.

80When referents are identical or very similar, the singular is used too, as with haj (= hair). From a cognitive point of view, this suggests that similarity is perceived as one, despite the obvious numerous items involved.

81Furthermore, there is no inflection for number with “natural duality”, i.e., items that naturally come in twos in nature, like feet, legs or eyes. Such items are always treated as singular items. In order to refer to only one item of the two, the “half + noun” construction has to be used, so that the equivalent of one eye is literally “one half-eye”. The cognitive logic here is that the language does not dissociate what nature provides in pairs. Language has created some sort of “compact duality” based on the human body, thus providing another body-based model as advocated by cognitivists.

82At the same time, this natural duality cannot be treated as a dual because Hungarian does not have dual inflection. These nouns are obviously semantically non-singular but this is not stated grammatically. Cognitively they are perceived as one, given that “one eye” in Hungarian translates as “half an eye”.

83This section on number in Hungarian shows that the language stands slightly apart when it comes to noun inflection insofar as the proportion of singular nouns in Hungarian is even higher than in other languages, which is partly due to this lack of number agreement between numerals and nouns or between duality and nouns or between quantifiers and nouns. It also stands apart in following the principle that what is obviously plural does not need to be stated grammatically.

2.7. The plural at heart

84In Indo-European languages, the singular is often perceived as the bare form and the plural as a modified form. However, the modification provided by the plural should not be viewed as peripheral despite the fact that the plural marker often seems added to the bare form, as in English: book+s or Spanish: libro+s.

85My contention at this stage is that the plural is not so much added to the noun as an intrinsic part of it. To prove this point it can be said first that in languages like Italian the plural marker is not added to the noun but alternates with the singular: mano [masc.] > mani; donna [fem.] > donne. As said before, in Old French the nominative plural was the same as the nominative singular and in Old English many nouns were invariable.

86As an intrinsic part of the noun, the plural should be construed as modifying the noun intrinsically, which accounts for the internal modification of the noun which was prevalent in Old English as in fot >fet [feet] or boc [book] > bec [books]. This plural marking is still prevalent in German: Bruder [brother] > Brüder [brothers], even though the internal modification is often supplemented by a suffix as in Mann [man] > Männer [umlaut + -er suffix].

87Similarly, the diminutive in German often entails the internal modification of the noun with an umlaut as in Frau [woman] > Fraülein [young woman / miss]. The diminutive in -lein modifies intrinsically our perception of the bare noun, in that a Fraülein is not the same as a Frau, which also explains the internal modification by means of the umlaut.

88In all languages, the number (and probably also the diminutive) is more intrinsic to the noun than case, if we are to believe Greenberg (1966:95), who postulates the following universal: ‘Where morphemes of both number and case are present and both follow or both precede the noun base, the expression of number almost always comes between the noun base and the expression of case’. This universal obtains of course for Hungarian, as in gyerekeket (gyerek = child; ek = pl.; et = acc.) or végtagaimat (végtag = limb; i = pl. with possessives; m=1st pers. sing, poss.; t = acc.).

89In languages in which the determiner is added to the noun as a suffix, the order is “noun + plural + determiner”, as in Icelandic steinarnir (= the stones < stein + pl. ar + det. nir). This indicates once more that the plural is more intrinsic to the noun than all other suffixes.

2.8. Distant and close relations

90Greenberg’s universal tends to prove that number modifies the meaning of the noun intrinsically, from within: the referents are not the same whether you say book or books, if only because you refer to several items with books and only one with book.

91By contrast, an accusative does not change the semantics of the noun; it just provides the function of the noun phrase within the sentence.

92The distance between the accusative and the noun base is iconic of this lack of change in meaning: the accusative marker is most often a suffix, that is, it is added and it always comes after number.

93Conversely, the proximity between the plural marker and the noun base is iconic of the internal change in meaning, which sometimes leads to fusion between plural marking and the noun base: the base itself is altered and changed into a new noun base, as in foot > feet. Even though French no longer has an overt plural, there still exists a pair in which the plural is totally different from the singular as in œil [eye] > yeux [eyes].

94Bybee (1985: 24-25) coined the phrase “the iconic distance hypothesis”, whereby the distance between two morphemes is a reflection of the degree of alteration in meaning. This hypothesis is supported by the differential between case and number.

3. Numerals

3.1. A cardinal point: numerals and arbitrariness

95The plural is intimately connected with numerals. Linguists have often expressed interest in numeral systems and especially in cardinal numerals, 1, 2, 3, etc, paying special attention to the origin of these words.

96Any discussion about cardinal numerals often revolves around the question as to whether numerals are motivated or not. Are words like ‘1’, ‘2’, ‘3’ arbitrary or not in the languages of the world?

97Even though Saussure showed that most words are arbitrary, some linguists, especially cognitivists like Heine (1997:19), qualify the arbitrariness usually attached to the signifier: “I claim that numeral systems across languages are motivated – that is, they are non-arbitrary.”

98The higher a numeral, the more likely it is to be non-arbitrary. According to Heine, it seems that no society in the world has developed a totally non-arbitrary word for ‘13’. If the words three in English or drei in German may seem arbitrary, thirteen or dreizehn are clearly not. The German word is even more transparent than its English equivalent in that dreizehn can be broken up into drei (= three) and zehn (= ten). Old English was also more transparent in the old form for ‘13’, i.e. threotene, which can also be broken up into threo (= 3) and tene (= 10). In French, the /z/ phoneme found in numerals after ten, from onze (‘11’) to seize (‘16’), appears to be a contraction of the Latin decem (= ten). The word treize (= 13) can thus be construed etymologically as “three-ten”.

99The equivalent of thirteen in Hungarian is also totally transparent: tizenhárom clearly combines tizen (= 10) and három (= 3). Innumerable examples could be quoted from other languages.

100In most Indo-European languages the apparent arbitrariness of numerals stops at number ‘10’, as seen with onze, douze, treize... In English, the etymology of eleven would have been lost, had not the language been recorded for several centuries, which bears out what cognitive linguists say about the non-arbitrariness of numerals. Its etymology is not totally clear but it seems to hinge on “one”.

101According to the Chambers Etymological Dictionary (1912), eleven comes from Old English en(d)luf-on, in which <d>is an excrescent, i.e. haphazardly added to the word and -on a dative plural suffix. So we are left with enluf, i.e. en, which is one variant of one in Old English + luf (or lif) < tak (= ten) weakened to dak, lik, lip and lif (cf. Gothic ain-lif). This shows that eleven corresponds to “one-ten”. However Skeat (1993) provides a different etymology for eleven: endlufon is a cognate of Gothic ainlif, in which ain = one and lif is related to Lithuanian lika, which means “remaining”. Thus eleven would mean “one remaining” (probably “one left from ten”, i.e. “one above ten”). It might also be possible to relate lif and leave.

102German elf (‘11’) comes from Old High German einlif which meant “one left”, possibly “one left once you have counted ten with your fingers”. Only specialists could work out this type of etymology, which can only be traced for a very tiny proportion of the world’s languages.

103By comparison, the numeral twelve appears far more transparent: any native speaker could easily be convinced that twelve is connected to two. More precisely, it can be broken up into twa + lif The connection between zwei (‘2’) and zwölf (‘12’) in German is even easier to see and to hear: <zw>is pronounced in the same way in the two words, unlike <tw>in two and twelve.

104All other numerals, except for hundred and thousand, are easy to deal with in terms of transparency: it can be easily understood that the suffix -ty in twenty, thirty... is one of the three ways of expressing ‘10’ in English, alongside ten and teen.

3.2. Size matters

105The transparency of higher numbers is further evidenced by the fact that high numerals tend to be nouns originally (like Latin cent-um), which have retained some of their nominal properties (Heine 1997: 22). Thus in French the numeral cent is number-inflected just like an ordinary noun: trois cents ans (= three hundred years).

106The use of score as a numeral is now archaic, even though it survives in the phrase by the score and the quantifier scores of: People are resigning by the score·, They’ve received scores of letters. Eighty replaced four-score long ago. Score is actually a noun, albeit invariable (She lived to be three score years and ten), which meant “a mark or notch for keeping count”, “the number twenty, once represented by a larger notch” (Chambers Etymological Dictionary). In Old Norse the cognate word skor meant both “notch” and “twenty”.

107What is also important to cognitive linguists, in their quest for non-arbitrariness, is the fact that many numerals are constructed by means of addition, subtraction or multiplication. Numerals like quatre-vingts (literally “four-twenty”, with the plural marker -s added to vingt) or quatre-vingt-dix (“four-twenty-ten”) in French are a case in point. The Belgians and Swiss much prefer the more logical nonante (‘90’), based on the sequence quarante, cinquante, soixante, but octante or huitante for ‘80’ is less common than quatre-vingts. The origin of quatre-vingt-dix is, of course, totally transparent but, when used in ordinary conversation, this numeral is not thought of as meaning “four-twenty-ten” but just ‘90’.

3.3. Pointing fingers (and toes)

  • 12 This is in keeping with Benveniste’s (1971: 45) famous analysis of the linguistic sign as propounde (...)

108Heine’s contention that no society has developed a totally non-arbitrary word for ‘13’ and beyond raises the following question: what about smaller numerals? Are they arbitrary or not? In a Saussurean approach, numerals like one, two, three are considered arbitrary in that there is nothing intrinsically motivated about the use of these words, other than the fact that to an English speaker three means intrinsically ‘3’- it is difficult not to closely associate the signified with the signifier - and that the arbitrariness of the word is not felt in spontaneous conversation12.

109Heine’s argument is that all numeral systems are non-arbitrary. The reason why we think the signifiers one, two, three are arbitrary is because of “a gap in our knowledge that remains to be filled” (1997:19). The implicit idea behind this reasoning is that the origin of small numerals is not accessible, which leaves room for the non-arbitrariness of numerals or at the very least for the debatable nature of the posited arbitrariness of numerals. In other words, if we could go back in time, we could prove that basic numerals are actually motivated, as in Mamvu, a central African Nilo-Saharan language. They are motivated in that language insofar as ‘6’ means literally “the hand seizes one”; ‘10’ means “all hands”, ‘11’ means “the foot seizes one”. Very interestingly, ‘20’ means literally “one whole person” and ‘100’ “five whole persons”. A whole person can be equated with twenty, in that a whole person normally has ten fingers and ten toes, i.e. a total of twenty extremities.

110The problem is that even in a language like Mamvu, the numerals ‘1’ to ‘5’ are etymologically opaque, which leads one to conclude that the lowest numerals are opaque in all languages but the higher up one goes in numerals the more likely they are to be transparent and therefore motivated.

111Does opaque mean arbitrary? We are faced with two options here. A Saussurean approach would of course conclude that opacity is akin to arbitrariness. Heine’s contention is that opacity just means that a gap in our knowledge needs to be filled. To him, arbitrariness is open to debate, but it could also be argued that non-arbitrariness is open to debate too, in that it cannot be proven in the case of the signifiers one, two, three... and their equivalents in other languages.

112We can only conjecture that the Indo-European equivalents of one, two, three... may have been motivated but this is impossible to prove.

113Having said that, there might seem to be a connection between ‘1’ and the word finger if we are to believe Merritt Ruhlen, who, in his highly controversial book The Origin of Language (1994), claims that *tik was a word that was used both for “one” and “finger” in the hypothetical original mother tongue.

114The word “opaque”, as used by Heine, is in itself questionable insofar as it implies that we cannot yet see through something. He explicitly opposes transparent linguistic forms, where genetic motivation has already been reconstructed, to opaque forms (1997:32). This implies that the motivation behind these forms is awaiting reconstruction, which can be considered as a misuse of the word “opaque”.

115Furthermore, Heine’s theoretical position is open to criticism: one may wonder why a lexeme that was supposedly motivated once, but no longer is, should be regarded as non-arbitrary now.

3.4. Decimal systems: plenty of body

  • 13 It is of course far more difficult to explain the 12 base number system used in many languages. Sta (...)

116What seems less debatable in the cognitivist approach to numerals is their relation to the body. The fact that decimal systems are widespread in the world can probably be accounted for by the body-part model championed by cognitivists like Heine (1997: 19-24). The body-part model is also referred to as “mimesis” (or bodily mimesis) in cognitive parlance (Evans 2007). Mimesis involves the ability to use a body part in the expression of a more abstract notion. It is at work in counting in decimal systems, the source of which is to be found in the two hands of every human being. It is easy to imagine the importance of hands as a visual aid when counting and to infer that this visual aid provided the basis of the abstract decimal systems prevalent in the world13.

117This anatomic characteristic of the human body probably also explains why numerals beyond ‘10’ are very often etymologically transparent, and as a general rule much more so than all numerals from ‘1’ to ‘10’.

118Evidence from many languages (Heine 1997: 18-34) suggests that numerals follow a decimal system conceptually derived from the total sum of fingers and sometimes of fingers and toes, as in Mamvu (see above). The system exemplified in Mamvu is “widespread in the languages of the world” (Heine 1997: 20). Stampe (1976: 596) calls it the ‘one-hand, two-hand, one-foot, whole-man variety’.

119This shows that the decimal system is far from arbitrary but rests on our physical experience. To Heine, the human hand “provides the most important model for structuring the numeral system”, which leads him to conclude that “The most common structure of numeral systems appears to be one in which ‘5’ derives from ‘hand’, ‘10’ from two hands’, and ‘20’ from either something like ‘hands and feet’ or ‘whole person’.”

120It is easy to see why cognitivists set great store by numerals: they can be used to confirm that human experience informs language and syntax. This idea lies at the heart of the cognitive agenda. One aim of cognitive grammar is to link language and other aspects of human sensor-motor experience. In other words, linguistic knowledge is not to be dissociated from other types of knowledge.

Conclusion

121The fact that some languages do not have the singular / plural distinction shows that it is not a necessary category of language. The context is usually quite sufficient to express singularity or plurality. In the same way, when we use an unmarked plural noun in English, as in Look at the sheep, the context is sufficient to interpret sheep as referring to one or several items.

122In many languages the two-term system “singular / plural” is challenged by intermediate categories, like the dual, even though these intermediate categories have proven less stable than the singular or plural over time.

123The singular / plural distinction is not watertight: the singular may entail a plural verb and vice versa. Also what appears plural (morphologically) may be treated as singular (syntactically) or what was once plural may now be singular and, again, vice versa. Language is very flexible in dealing with these two notions.

124What seems extremely widespread is the general tendency for each language to have developed a variety of plural forms. In Hopi, the plural is multifarious. German and English have several plural markers, although relatively few nouns are irregular in contemporary English, a language that has undergone a regularization process. By contrast, the general tendency for the singular is to be stable, i.e. it is often the bare form of the noun.

125The plurality of the plural tends to show that the plural cannot be reduced to a singular noun with a plural suffix. Cognitively, it is more than that: the plural modifies intrinsically our perception of the bare noun.

126The plural is not perceived in the same way in languages like English, French or Hungarian. In English, the plural is really a non-singular rather than a number that starts with, ‘2’. What is more, each language has its own way of defining what is ‘singular’ or ‘plural’ even with similar referents, like somebody’s hair or eyes.

127Positing a binary singular / plural system can therefore be seen as an oversimplification. A plural approach is required.

Top of page

Bibliography

Benveniste, Emile (1971 [1966]). Problems in General Linguistics. Miami: University of Miami Press.

Bybee, Joan L. (1985). Morphology: A Study of the Relation between Meaning and Form. Typological studies in language, v. 9. Amsterdam: J. Benjamins.

Cauquil, Gérard & François Schanen (2000). La Grammaire allemande. Collection Bescherelle. Paris: Hatier.

Corbett, Greville (2000). Number. Cambridge: Cambridge University Press.

Croft, William (2003). Typology and Universals. Cambridge Textbooks in Linguistics. Cambridge: Cambridge University Press.

Croft, William & D. Alan Cruse (2004). Cognitive Linguistics. Cambridge Textbooks in Linguistics. Cambridge: Cambridge University Press.

Dixon, Robert (1980). The Languages of Australia. Cambridge: Cambridge University Press.

Evans, Vyvyan (2007). A Glossary of Cognitive Linguistics. Edinburgh: Edinburgh University Press.

Everett, Daniel (2005). “Cultural Constraints on Grammar and Cognition in Pirahã: Another Look at the Design Features of Human Language”, in Current Anthropology 46:621-646.

Findlater, Andrew (1912). Chambers’s Etymological Dictionary. London: W & R Chambers.

Franck, Michael, Daniel Everett, Evelina Fedorenko & Edward Gibson (2008). “Number as a Cognitive Technology: Evidence from Pirahã Language and Cognition”, in Cognition 108: 819-824.

Gordon, Peter (2004). “Numerical Cognition without Words: Evidence from Amazonia”, in Science Vol. 306: 496-499.

Greenberg, Joseph (1966). Language Universals, with Special Reference to Feature Hierarchies. The Hague: Mouton.

Heine, Bernd (1997). Cognitive Foundations of Grammar. New York/Oxford: Oxford University Press.

Kiss, Katalin (2002). The Syntax of Hungarian. Cambridge Syntax Guides. Cambridge: Cambridge University Press.

Kor, Chahine, Irina & Robert ROUDET (2003). Grammaire russe. Les structures de base. Paris: Ellipses.

Laverne Masayesva, Jeanne. Aspects of Hopi Grammar. Thesis MIT, http://dspace.mit.edu/handle/1721.1/16325

Lichtenberk, Frantisek (1983). A Grammar of Manam. Oceanic Linguistics Special Publication 18. Honolulu: University of Hawaii Press.

Malavieille, Michèle & Wilfrid Rotge (2008). Grammaire anglaise. Collection Bescherelle. Paris: Hatier.

Marsden, Richard (2004). The Cambridge Old English Reader. Cambridge: Cambridge University Press.

Mitchell, Bruce (1995). An Invitation to Old English and Anglo-Saxon England. Oxford: Blackwell.

Montagne, Lucien (1931). Essai de grammaire malgache. Paris: Société d’Editions Géographiques, Maritimes et Coloniales.

Radden, Günter & René Dirven (2007). Cognitive English Language. Amsterdam: John Benjamins Publishing Company.

Ruhlen, Merritt (1994). The Origin of Language. Tracing the Evolution of the Mother Tongue. New York: John Wiley & Sons.

Sauvageot, Aurélien (1971). L’Edification de la langue hongroise. Paris: Editions Klincksieck.

Stampe, David (1976). « Cardinal Numeral Systems ». Chicago Linguistic Society 12: 594-609.

Stevanovitch, Colette ([1997], 2008). Manuel d’histoire de la langue anglaise des origines à nos jours. Paris: Ellipses.

Swan, Michael. (2005). Practical English Usage. Oxford: Oxford University Press.

Skeat, Walter (1993). The Concise Dictionary of English Etymology. Ware, Hertfordshire: Wordsworth Editions.

Szende, Thomas & Georges Kassai (2007). Grammaire fondamentale du hongrois. Paris: Langues & Mondes – L’Asiathèque.

Talmy, Leonard (1986). The relation of grammar to cognition. Duisburg: Linguistic Agency University of Duisburg.

Wierzbicka, Anna (1988). The Semantics of Grammar. Amsterdam: John Benjamins Publishing Company.

Top of page

Notes

2 Jeanne LaVerne Masayesva, Aspects of Hopi Grammar, thesis MIT, downloadable at http://dspace.mit.edu/handle/l 721.1/16325

3 http://www.mindhacks.com/blog/2008/06/counting_in_the_lang.html.
See also Gordon 2004 or Everett 2005.

4 -tween in between is related to two, twain.

5 Some style guides in the United States still advocate this distinction, although Merriam Webster’s Dictionary of English Usage (1994) quotes Fowler (1926) as saying “the differentiation is neither of present utility nor based on historical usage; the old distinction of two as opposed to several was not e[ach], but either, & either other, which formerly existed beside e[ach] o[ther] & one another would doubtless have survived if its special meaning had been required”. Footnote suggested by Linda Pillière.

6 Tolomako is a language of the Oceanic subgroup of Austronesian languages. Lihir and Manam arc languages of Papua New Guinea, Tok Pisin is a Melanesian English pidgin. Source: http://www.nationmaster.com/encyclopedia/Grammatical-number#Trial

7 Bruce Bower, “Numbers Beyond Words”, http://sciencenews.org/view/generic/id/33300/title/Numbers_beyond_words

8 The use of a singular in He’s six foot tall may be attributed to an adjectival interpretation of the noun foot, grammatically interpreted as He’s a six foot tall man.

9 Things are slightly more complicated with the noun police. Most grammar books consider that it cannot be followed by a singular verb. For instance, the sentence the police is looking... is crossed out in Swan 2005:517. However there is much debate over the supposed ungrammaticality of the construction the police is... as in We have no clue what the police is up to (private conversation).

10 In the same way the quantifier no is followed by a plural count noun: Mirren has no children and says she has no maternal instinct.

11 Hungarian always treats nouns as mass nouns when they come after numerals or quantifiers, as in sok gyerek (= lit. many child)· In marked contexts this happens in English too, as in There is cat all over the driveway (Talmy 1986). The difference is that this construction is not optional in Hungarian and very unusual in English.

12 This is in keeping with Benveniste’s (1971: 45) famous analysis of the linguistic sign as propounded in “The Nature of the Linguistic Sign”: “The concept (the “signified”) bæuf is perforce identical in my consciousness with the sound sequence (the “signifier”) böf. How could it be otherwise? Together the two are imprinted on my mind, together they evoke each other under any circumstance. There is such a close symbiosis between them that the concept of bœuf is like the soul of the sound image böf”.

13 It is of course far more difficult to explain the 12 base number system used in many languages. Stampe (1976) provides an explanation for the duodecimal system used in Sora, a Munda language spoken in Eastern India: it is the result of a reanalysis of the decimal system, with a shift from ‘10’ to ‘12’.

Top of page

References

Bibliographical reference

Wilfrid Rotge, « Plurality in English and other languages: does it add up? », Anglophonia/Sigma, 13 (26) | 2009, 101-120.

Electronic reference

Wilfrid Rotge, « Plurality in English and other languages: does it add up? », Anglophonia/Sigma [Online], 13 (26) | 2009, Online since 13 December 2016, connection on 23 August 2017. URL : http://anglophonia.revues.org/875 ; DOI : 10.4000/anglophonia.875

Top of page

About the author

Wilfrid Rotge

Centre de recherches anglophones, CREA, EA 370, Université de Paris Ouest Nanterre La Défense

By this author

Top of page

Copyright

Licence Creative Commons
Anglophonia – French Journal of English Linguistics est mis à disposition selon les termes de la licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.

Top of page
  • Logo Presses universitaires du Midi
  • Revues.org