As is so often the case with ‘facts’ that are taken for granted about language, the ideas of person and number are not, however, common sense but come out of a school teaching tradition going back to the Latin grammars of 100 BC. Because of their origin in classical languages, three persons and two numbers fail to do full justice to English, nor do they necessarily fit languages that are unrelated to Latin, say Chinese. For instance, while all languages have pronouns, they vary over how many persons and numbers they have. Navaho, an indigenous American language, signals a fourth person: when a third person pronoun has already been mentioned in a conversation, a fourth person is used to refer to anybody else.

Furthermore, while number in English pronouns is a two-way choice between singular and plural, many languages interpose a third distinction for ‘two people’—the ‘dual’ number. The contrast is between one person, two people, and more than two people. Old English, the period of English spoken before 1100 AD, had such a three-way system; the first person was the singular ic (I), the dual wit, and the plural we, with similar choices for the other persons in the pronoun system.

Other languages still have a three-way choice of number. For example Tok Pisin, spoken in Papua New Guinea and partly derived from English, has first person singular mi, dual mitupela, and plural yupela. Tongan has three first person forms—singular kita, dual kitaua, and plural kitautolu. Languages exist with even more number contrasts. Fijian has pronouns for singular, dual, trial (three), and plural (as well as four persons). The two-way choice between singular ‘one’ and plural ‘more than one’ is by no means the only possibility for languages.

Nor does the English six-way system exhaust all the possibilities in the first person plural. The English plural we may either include or exclude the person addressed. Suppose Judy says to Richard We’re going out tonight. This sentence might mean that Judy is announcing to Richard that she and Anne are going out or it may be that Judy is reminding Richard that she and Richard are going out together as a couple; one meaning excludes the listener from we, the other includes them. This distinction is signalled in Melanesian Pidgin English by two different pronouns: yumi (you+me) and mipela (me+someone else). Palaung, a language spoken in Burma, has both a dual/plural distinction and a distinction between inclusive and exclusive pronouns for ‘we’: yar (two people, not including the speaker), ar (two people, including the speaker), ye (more than two, not including the speaker) and e (more than two people, including the speaker).

2. Pronouns and status

So, once you know what persons and numbers are used in a language, you should be able to choose the right pronoun to fit a particular speech situation. However, pronouns illustrate something that is true of many aspects of language: once a distinction has been established for one reason, it can be exploited for another. The apparently straightforward concept of singular and plural number can be overlaid with other meanings. Standard Modern English makes no difference between singular and plural you, apart from a few marginal survivals of thou and ye, even if Geordie and some dialects of Irish English distinguish singular you and plural youse.

Other languages, however, regularly make a distinction between singular and plural second person that does not depend on the number of people involved. In French, individuals who are close to the speaker socially are addressed with the singular second person tu, Tu joues bien (you play well); individuals who are more distant from the speaker are addressed with the plural form vous, Vous jouez bien (you play well). The difference between tu and vous, ostensibly one between singular and plural number, is nothing of the kind: a single individual can be referred to by a plural pronoun. The chief modern use of singular tu is to act as a sign of closeness between friends, fellow-students, colleagues, and particularly members of the family. Switching from addressing someone as vous to calling them tu shows a change in the relationship of the speaker and listener or in the circumstances of speaking. Mountain climbers are said to change from vous to tu above a certain altitude when the common danger brings them closer together. In addition tu has demonstrated political solidarity between speakers and listener since at least the days when the French Revolution insisted that all citizens addressed each other as tu. At a French-speaking academic conference in the 1980s, the only form used in the talks and public discussions was tu, still a sign of social solidarity.

A second use of tu is for talking down to another person, say a child or an animal; tu is a sign that the speaker has a higher social status than the person addressed. Teachers for example assert their authority by calling their pupils tu. The sign of this power relation is that the use of tu is not reciprocated; that is to say, one speaker says tu, the other vous. If the person addressed with tu is the speaker’s equal, he or she may respond with the singular tu; if inferior, with the plural vous. A survey in French-speaking Canada showed that children addressed grandparents, priests and employers with vous, but these adults answered them with tu. So it is not just what you call other people that defines the relationship but what they call you. The social status function of tu has diminished in recent years, being partly undermined by the use of tu to show solidarity.

In French, therefore, reciprocated tu may show either social closeness or solidarity with other people. This use of the second person number distinction for status and solidar­ity is common in different languages and is known as the T V system, after the Latin tu (T) and vos (V). Other languages in Europe as well as French use singular T to show social close­ness/ solidarity and plural V to show social distance/status. The Spanish T form is tu, the V form Usted; the German T form is du, the V form Sie; the Russian T form is ty, the V form vy; and so on. In some languages such as Spanish the V form is the same as a singular third person pronoun, Usted; in others such as German, it is the same as a third person plural Sie. In all of these languages, T and V pronouns are used to fit the social relationships between speaker and listener, with slightly different overtones in each language. In the German version of the good guy/bad guy police interrogation technique, the good guy allegedly uses the V form Sie as a sign of politeness, the bad guy uses the T form du to assert power over the suspect. Often the written form of the V pronoun has a capital letter, as in German T du versus V Sie and Italian T tu versus V Lei. Speakers of these languages find it puzzling that English capitalises the first person singular I rather than the second person you.

There has been much speculation how the plural came to suggest distance and high status, the singular familiarity and solidarity. One is that it is an extension of the royal we; kings claiming to be plural established the plural as a sign of power and status, and this usage percolated down to other levels of society. Another is that, during the fourth century AD when there were two Roman emperors at the same time, one in Rome and one in Constantinople, it was diplomatic to address the emperor in the plural, and so power came to be associated with the plural form.

These reasons cannot, however, explain why this T V system occurs in languages unconnected to Latin. For the use of the plural for status also occurs in Turkish, Basque, and Bengali. Nor is it necessarily status and solidarity alone that are involved in the choice between T and V. In Russian, choosing the appropriate T ty or V vy form depends inter alia on the topic of conversation, the age, generation and sex of the speaker, and the precise family relationship.

Some languages use them to reflect a complex range of social roles. In Japanese, the choice between different pronoun forms is dictated by complex social relationships between speaker and listener. The first person singular ‘I’ can be either watakushi, watashi, boku, ore, or atashi, according to the relative status and sex of the speaker and listener. Watakushi is the most formal form, used say to a social superior. The pronoun watashi is more neutral. Next come boku, which is the norm for men under 21, and ore, used only to close friends. In 1957 the Japanese Ministry of Education approved boku for men students but recommended they switched to watashi when they graduated. The choice of Japanese pronoun marks off the speaker and listener on a social scale. One first person form chin, may only be used by the Emperor. A further complication is the difference between the sexes. Men use the first person pronouns watakushi, boku, or ore; either sex use watashi; but only women use atashi, which, from a man, would be a sign of gayness. There are therefore difficulties for male English students who are taught Japanese by a woman and never hear the male status pronouns used naturally.

Japanese is not the only language to signal sex and status through choice of the first person pronoun. In Thai the correct first person singular for addressing a stranger is phoém from a  man, dìchaén from a woman; in informal conversation with family and friends, it is chaén, but kan if all the speakers are the same sex; a child speaking to an adult uses nuéu. Other first person forms are needed for addressing the king, kram˜çm’ coming from a man, kram˜çm’ chá from a woman. Thai speakers in fact have around 25 different first person pronouns to choose from in everyday circumstances.

Languages then use pronouns to reflect diverse social relationships between the participants. Until Japanese speakers know the social relationship between them and the other participants, they do not know which pronouns to use to refer to themselves or to the others. The concept of social relationship is also important in European languages like Polish. Anna Wierzbicka, a Polish linguist living in Australia, tells of two expatriate Poles who had only talked English to each other for many years suddenly having to talk in Polish and not knowing which pronouns to use because they had never before had to think of each other in terms of the Polish pronoun system. To an English speaker, the crucial aspects of the speech situation are person and number. Other languages treat various social roles as being just as important.

3. ‘Missing’ pronouns

The use of English as a starting point conceals another important fact about pronouns in other languages: they do not necessarily have to be present at all. Many languages in effect leave the subject pronoun out of the sentence and let the listener work out the situational roles from the rest of the sentence or from the context. For example an Italian can say Sono di Torino (am from Turin) and the listener knows from the verb sono that the speaker is talking in the first person singular even though no pronoun is present. Or the speaker could say Vende (sells) with no third person pronoun but the ending of the verb -e reveals that the third person singular is intended. Languages like Italian in which the sentence does not require a subject are technically called ‘pro-drop’ languages.

To an extent, the different endings of the Italian verb for person and number provide the clues to the missing subject. Since English has only a few verb endings, there is no way of telling from the verb speak whether the absent subject is I, you, we or they. English is therefore an example of a ‘non-pro-drop’ language in which subjects are compulsory in the sentence. This puts English in a minority of around 7% of the world’s languages, along with German and French.

The pronoun system described above left out a crucial dimension for languages—whether the subject pronoun is actually present in the sentence or not. The pronoun does not necessarily have to be left out in pro-drop languages; there is usually a choice whether to put it in or leave it out. However this choice opens up another potential way of conveying meaning through the pronoun system. If it is usual to have no subject in a pro-drop language, putting it in adds an extra meaning to the sentence, which is absent from non-pro-drop languages.

So, in Spanish, another pro-drop language, the first person yo for instance is sometimes used to underline the person’s responsibility as in Yo espero (I am waiting) compared with Espero. In Greek, omission of the second person in greetings shows the speaker is closer to the listener than leaving it in; Ya! (hello) shows a friendlier relation­ship than including either the T su as in Ya su! or the V sa as in Ya sa! In pro-drop languages the T V system in a sense includes a third possibility — absence of pronoun: it is a three-way T V Ø system.

4. Pronouns and Gender

A widespread contrast in human languages is called ‘gender’. Some German nouns for example are masculine—der Tisch (table) goes with the masculine pronoun er. Other nouns are feminine—die Tür (door) goes with the feminine pronoun sie (she). Others are neuter—das Haus (house) goes with the neuter pronoun es (it). The chief manifestation of gender in many languages is how pronouns are allocated to nouns, grouped into masculine, feminine or neuter.

Some languages do not have pronouns with gender, for example Persian, Turkish, and Hungarian. In English, gender only affects some pronouns, affecting the third person singular he, she, and it, but not I, you, we or they. The choice between he, she, and it depends on sex; a woman is she, a man is he, something without sex is it. English is therefore said to have ‘biological’ or ‘natural’ gender as the noun groups of masculine, feminine, and neuter correspond to the real world categories male, female, and non-sexed. The exceptions beloved by English grammar books are personifications of things as people. Often these are feminine, as in the case of large moving machines such as ships, planes, cars. Less commonly they are masculine, as in rivers, That old man river, he keeps on rolling along. Even these few metaphorical uses of gender seem to be declining: up to the 1930s the Times newspaper invariably referred to countries as she, as in Canada and her great neighbour, but by the 1980s the use of she had waned to less than 10%.

Other languages also have natural gender, such as Tamil. Many languages taught in English schools, however, use a system called ‘arbitrary’ gender, meaning that the gender of nouns is linked arbitrarily to pronouns without reference to their actual sex. In German for instance das Mädchen (girl) is neuter but der Tisch (table) is masculine. In French la table (table) is feminine, le toit (roof) is masculine, and there is no neuter.

It is hard to find a rationale for the gender of nouns in arbitrary gender languages. The way that nouns are grouped into genders is often based on the sounds with which the word ends rather than on their meaning. In French 94% of nouns ending in ‑age //such as voyage (journey) are masculine, and 90% of those ending in /z/ such as heureuse (happy) are feminine. In Italian nouns ending in “-o” are usually masculine, as in vino (the wine), and nouns ending in “-a” usually feminine, casa (house), Such groupings have nothing to do with biolog­ical sex, or indeed with productiveness versus receptiveness.

The celebrated issue in English does not concern gender so much as sex. A natural gender language needs a way of deciding which pronouns should go with nouns that refer to members of both sexes. How should an everyday notice in a doctor’s surgery for example be worded? The patient must keep to the time of his? her? his or her? their? appointment. A practical issue which affects every page of this book is which pronoun to use to refer to a single speaker of indeterminate sex, as in A speaker of English knows that he/she/he or she/they ought to feel guilty about splitting infinitives. He suggests that the writer is ignoring women. She may imply that the writer is falling over backwards to be politically correct or is treating women as a subordinate group; for instance books for teachers by male writers that call all teachers or children she can have a patronising feel. One linguistics book not only calls all speakers he, all hearers she, but also, when there are two or more speakers in the conversation, has the first one male, the second female, surely as discriminatory as the uses it is trying to avoid. The combination he or she is cumbersome and soon becomes irritating. It seems ruled out as denying people’s humanity, even if Wordsworth could write A simple Child ... What should it know of death?

The soft option is to put the noun in the plural, Patients must keep to the time of their appointments, since this enables the non-gender form they to be used. This usage is clearly impossible when a single individual is being talked about. My own preference wherever possible is to use my natural spoken form they with the singular noun and let the fact that they is plural go by the board, as in the third paragraph of this chapter, A speaker of any language has a large store of vocabulary at their beck and call. This was anathema to the Fowlers in The King’s English: it ‘is the popular solution; it sets the literary man’s teeth on edge’.

The use of they for the singular mixed gender was nevertheless the most common form in English till the eight­eenth century, and was in fact used by Shakespeare in Much Ado ‘God send everyone their heart’s desire!’ It has been supported by a minority ever since, going from the grammarian Henry Sweet in 1891 to the American writers Casey Miller and Kate Swift in 1977. The concept of number is not taken literally in many languages but is overridden by considerations of social role, as in the T V contrast seen above. If there is a conflict between number and gender in English pronouns, the easiest solution is to over-ride number. Nevertheless they may continue to be obtrusive in some styles of written English, and the Fowlers’ advice ‘to give the same meaning in some entirely different way’ may still be sound.

An alternative is to propose a new third person singular pronoun that can be coined for humans but has no gender. The proposals started with ne in 1850 and went through thon (1884), hizer (1891), heesh (1940), tey (1972), hann (1984), and many others. The most natural-seeming coinage is perhaps per first used by the psychologist Donald McKay in 1972 and used most memorably in Marge Piercy’s science fiction novel Woman on the Edge of Time, as in We lower per to the ground, It’s per way, and It’s time for per to die.

How people use language must relate to the way that their minds work in some respects. So, if mixed-sex groups such as doctors are always treated as masculine, this usage may indeed marginalise women by permanently associating these roles with men in people’s minds. However, blaming language for sexism is to some extent blaming the messenger for the message; it is primarily the thinking that is wrong.

Trying to change language consciously by, say, adopting a new gender system has often had mixed results on language. The Archbishop of Canterbury in a 1996 radio interview could still blithely say The school only has a child for a fifth of his time. The split infinitive to commonly believe was regarded as unEnglish by Henry Alford in 1864: ‘surely this is a practice entirely unknown to English speakers and writers’. Most commentators ever since have recommended that it should be avoided wherever possible. But split infinitives still seem as healthy as ever.

Yet the attempts to reform English pronouns seem to have had surprisingly potent effects. One study of written American English found that the frequency of he and man fell by about 75% between 1971 and 1979. Many would nevertheless dispute cause and effect; altering people’s language may not so much change their ways of thinking as make them sound as if they have changed. The proof will be whether the apparent decline of he indeed leads to a drop in sexism apart from language usage. Other parts of the pronoun system that have been described here may have similar effects on people’s attitudes in other languages. It would be fascinating to test how speakers’ ideas of number are affected in T V languages, how the concept of self is affected in languages with multiple first person pronouns like Japanese and Thai, and how the role of men in France is affected by the fact that there is a single pronoun for both he and it!

Gender is only one of the possible ways of dividing up nouns into classes, familiar to speakers of English because of its use with English pronouns and its presence in neighbouring languages such as French, German, and Spanish. In Bahasa Melayu, the chief language of Malaysia, there are no gender or number differences. Orang means not only ‘man/men’ but also ‘woman/ women’ (the English word orang-utan therefore in Bahasa Melayu meant ‘person of the forest’). However, the classical language has a complex system of around forty ‘classifiers’ that go with particular nouns. A long, stick-like object needs the classifier batang as in sebatang rokok (a cigarette); if it is ring-shaped, it needs bentuk as in sebentuk cincin (a ring); if it is broad, it needs bidang, as in sebidang tanah (a stretch of land); if it is a type of house, it needs tangga, as in setangga rumah Melayu (a Malaysian house, lit. ‘a house Malaysian’). Systems of classifiers are wide-spread in languages. An English parallel is the phrases used to count ‘uncountable’ nouns—two loaves of bread, not two breads, and three sheets of paper not three papers. 

5. Case forms of pronouns

The English first person singular pronoun does not just have the form I, but can also be me, my, mine, myself, etc; the third person singular is he, him, possessive his, himself, etc. These different forms signal the role that the pronoun plays in the sentence. I has the role of the subject of the sentence, as in I like Jane; me has the role of the object in Jane likes me. This difference is known as ‘case’—signalling roles in the sentence through changes in the form of words, again a traditional analysis going back to classical languages.

In English the case system applies only to pronouns. It distinguishes subject pronouns I, she, they, and so on, from object pronouns me, her, them, and the like, and from possessive pronouns my, her, their, and so on. Other languages have case systems for nouns as well as pronouns and have a greater range of cases. Latin for example has six cases, affecting nouns and pronouns (though, as a pro-drop language, it may also of course leave pronouns out). Let us take the noun amor (love). Using the traditional terms, Latin has nominative (subject) case amor, accusative (object) case amorem, and genitive (possessive) case amoris. In addition to these three cases that overlap with English, Latin expresses the roles of the person who receives something, the dative case (indirect object) amori, and the role of the person something is done by, the ablative case, amore. Some Latin nouns have a further case for addressing people, the vocative: Kennedy’s Latin Primer informs one quaintly that anne means ‘o year’.

Old English had an ampler case system than Modern English, involving both pronouns and nouns, like Latin. The third person singular masculine for example, had four cases he (subject), hine (object), his (possessive), and him (dative), with parallel forms for the neuter (hit, hit, his, him), and for the feminine (heo, hi, hire, hire). Modern Finnish has an even more complex system of fifteen cases.

Simple as the English case system may be, it nevertheless provides problems for its users. In the early stages of language acquisition, children are not apparently aware of the difference between first and second person subject pronouns: John like honey means ‘I like honey’; Mummy like teddy means ‘You like teddy’. Nor do they use the adult case system for pronouns, prefer­ring the object forms for the subject role, Me got bean or Him gone. Children have to learn that English pronouns have three persons and that they have different forms for the object and subject cases.

In English the verb to be is followed by the object case, like other verbs. The pronoun him in It is him has the same form as in I saw him. Latin, however, puts the noun that follows the verb to be in the subject case. Traditional grammar books have taken this to be a model that English should follow. William Cobbett in 1819 for instance said It was me ought to be It is I. It is, however, dangerous to apply the system of one language selectively to another language. Latin is a pro‑drop language which leaves the subject out altogether. Those who support It is I on the model of Latin should therefore also recommend that English speakers say Is raining rather than It is raining or answer Do rather than I do in the marriage ceremony. There is no logical reason why some features of Latin should take precedence over the English forms, others not; languages are independent of each other, however much they have in common beneath the surface. Already in 1580, Edmund Spenser was asking ‘Why a gods name may not we, as else the Greeks, have the kingdome of our owne Language?’

The pressure from the Latin-based school tradition has succeeded, if not in changing people’s behaviour – who says It is I naturally? – at least in making them acutely self-conscious. Lois Lane remarks to Superman Junior in an episode of a TV series, There is you; there is I; there is no we. In the film Camelot Lancelot du Lac proclaims C’est moi, ’Tis I, showing that whatever inhibitions script­writers have about ’Tis me in English, they have no such qualms about the French equivalent C’est moi (it’s me). The Inside Language panel were asked to give the class and age of people who would say It is I (Q6). 79% thought they would be over 50 years old, 81% that they would be middle-class. Lois and Lancelot are apparently trying hard to sound like middle-aged members of the middle-class.

The problem of choosing the appropriate case form has also affected pronouns linked by and, for instance My dog and I went for a walk versus My dog and me went for a walk. My own natural spoken form is my dog and me, but I could never deliberately use it in writing.

The choice of case form is also an issue after prepositions such as to, for, or between. Usually in English prepositions are followed by the object case; He looked at me, rather than He looked at I. The difficulty arises when pronouns are linked together with and, particularly when between is involved—between you and I or between you and me? Between you and I with the subject case I is now widely heard and this usage appears to be spreading to other contexts. A political commentator on the television news for instance said Sadly the money is coming from you and I via the government. At an international conference of English teachers the keynote speaker said One of the things that helped my wife and I was ... According to a survey, the aspect of English that listeners to the BBC were most sensitive about was the use of subject rather than object forms of pronouns.

A plausible reason for moving towards I rather than me is called ‘hypercorrection’: if you are uncertain what to say, it is better to overdo it than underdo it. A person who speaks another dialect of English but is trying to adopt the prestige British accent may not know when to stop. They might, depending on their background, pronounce the ‘h’ in hour if they do not usually have ‘h’ sounds, pronounce Thames with the ‘th’ sound in think if they do not have ‘th’ sounds, or pronounce gas to rhyme with pass. These are signs of people adapting speech for social reasons over-reaching the mark they are aiming at. Between you and I looks like the same phenomenon of the speaker playing safe for social reasons. Hyper­correction will be describ­ed further in Chapter Seven.

The forms of pronouns can also be felt to assert political power over people, in a similar way to gender. Rastafarians prefer the object form I rather than mi because they feel mi puts them in an subservient object position rather than a subject position. The refrain of a song by Lloyd Charmer for example is Rasta never fail I yet.

6. What do pronouns refer to?

The next issue about pronouns is who the third person pronoun refers to. The reference of the first and second persons is usually dictated by the roles in the speech situation itself. It is usually obvious who is speaking (I) and who is being addressed (you). The third person is trickier since by definition it refers to anybody or anything not involved as speaker or listener. This section gives a quick overview of the complex area of language study called ‘binding theory’, to give some flavour of the approach to the language system used by many linguists today.

In a sentence such as They liked the performance, the third person pronoun they refers to some people obvious from the speech situation—the critics, the audience, the stage-hands, etc—either because they have just been mentioned or because they are already well-known to the speaker and listener. In a sentence like John is afraid that he will have to leave, however, the pronoun he could refer to two different people, either someone called John by going with the noun John earlier in the sentence, or someone else not mentioned in the sentence at all—Fred or Frank Sinatra etc. In an actual speech situation, this problem usually takes care of itself and it is obvious who is meant. However, the sentence itself does not make clear whether the third person pronoun he refers to someone already mentioned or to somebody completely new.

Linguists have been interested in seeing how the organisation of the sentence can resolve who the pronoun refers to. Take the pronouns him and himself. In John voted for him the listener knows that, whoever the pronoun him may refer to, it cannot be John: something stops him linking to John. In the sentence John voted for himself on the contrary the listener knows that the reflexive pronoun himself must refer to John and can refer to nobody else: something forces himself to link to John. Speakers of English know that him and himself link to nouns in the sentence in totally opposite ways. They have a system for deciding which pronoun goes with which noun, again so obvious and automatic that people are surprised it needs mentioning.

The system can be seen better in more complicated sentences. In John said Peter voted for him the listener knows that Peter voted for John (or for someone else not mentioned) but did not vote for himself.

  

The pronoun him is linked to the noun John, if not to some unknown.

In the apparently similar sentence John said Peter voted for himself the listener is just as convinced that Peter voted for himself, not for John.

Now himself is linked to Peter within the sentence, not to John or anyone outside.

The vital difference is the position where the pronoun occurs in the sentence. The pronoun himself must link to a noun inside a limited section of the sentence, in fact within the clause Peter voted for himself.

 Him on the other hand has to go with a noun outside this limited area, that is to say, within the whole sentence John said Peter voted for him.

The difference is that himself links to a noun that must be inside the limited area of the clause, him must link to a noun outside this limited area altogether.

Speakers of English know how him or himself link to nouns. Straight pronouns like him link to nouns within the sentence; reflexive pronouns like himself link to nouns within the smaller confines of the clause. Using this clue enables listeners to work out who him refers to in John said that Peter voted for himself and who him refers to in Peter discovered John’s report on him. These sentences in fact form part of an experiment whose purpose was to see whether English people, who had never probably thought about it in their lives before, would agree about the links between pronouns and nouns in a range of sentences. As can be seen, they were in about 90% agreement about whether him or himself went with John or Peter on most sentences, slipping down to the 80s for two sentence types.

In other words, the rule about him and himself is not just an arbitrary invention by linguists but is something all speakers of English already knows, even if they are completely unaware of it.  The language system in their minds includes some highly abstract elements of language, way below their conscious attention but manifest in the answers they come to. The rules given so far are just a pale approximation of this complex knowledge, as we see in later chapters. For example the division into clause and sentence is too rough and ready to work perfectly, one reason why the last two sentence types were more difficult.

The different links between him and himself and nouns become more interesting when other languages are taken into account and have led to a specialised area of research called ‘binding theory’—how noun phrases are ‘bound’ to pronouns. The language system can seldom be understood properly by looking at  just one language. The modern study of language is often concerned with abstract properties that go across languages—how pronouns link to nouns—rather than peculiarities of a single language—how him links to John in English. It is not just the facts of English that are important or those of any language in particular but the general properties of the language system in the human mind, like the general properties of vision that, say, allow us to turn a succession of 24 frames a second into a moving picture. Hence many of the exciting issues about language today are at a high level of abstraction, based on years of developing theories of language and looking at wider and wider ranges of languages. Though this abstraction can cause some difficulty to start with, it leads to far more powerful ideas about language, rather like Newton proposing a theory of gravity rather than a theory of how apples drop to the ground.

III. Grammar and the language system

This piece has explored a range of issues involving pronouns, most of which could be expanded indefinitely: pronouns are involved in the use of language in society, in the acquisition of language by children, and in the loss of language through injury etc etc. The main issues are not the actual sounds of language nor even the meanings of the words, but the system that organises sounds and meaning through number, person, ‘binding’, and so on. This has been called the computational system of language by Noam Chomsky—the bridge between sounds and meanings.

The more usual word for it, which has been avoided up to here, is ‘grammar’. To most people, grammar is a memory of their school years. Their experience of grammar is of the traditional English grammar handed down by the eighteenth century,rather than anything more recent. The idea of grammar makes many people uncomfortable as it is a threat to social status; English speakers fear they are unwittingly giving things away about themselves by splitting infinitives, using between you and I, and so on.

However, this system of grammatical organisation is what gives language its power. Language is not possible without a grammar. While the grammars of human languages vary in many ways, they all have a grammar with many common characteristics. This is not a grammar of the type that is traditionally learnt at school with its lists of parts of speech, rules and exceptions, but the dynamic system that allows one person to organise a sentence The cat is chasing the birds, another Neko ga tori wo oitsuiteiru. The interesting questions about language are how this central computational system works and how such a complex system is acquired by every human child within a few years of birth.

Sources

Much of the general information on pronouns here is based on:
Mühlhäusler, P. & Harré, R. (1990), Pronouns and People: The Linguistic Construction of Social and Personal Identity, Blackwell, Oxford;
Mühlhäusler, P. (1994), ‘Babel revisited’, UNESCO Courier, February, 16-32:
Comrie, B. (ed.) (1990), The Major Languages of East and South-East Asia, Routledge

Speech role and number

Details of person and number are chiefly taken from:
Corbett, G. (1991), Gender, CUP; Burling, R. (1970), Man’s Many Voices, Holt Rinehart Winston

Pronouns and status

The classic article on pronouns and status on which most authors draw is:
Brown, R. & Gilman, A. (1960), ‘The pronouns of power and solidarity’, originally published in Sebeok, T.A. (1960), Style in Language, MIT Press, p.253-76, but reprinted many times, for example in Giglioli, P.P. (ed.), Language and Social Context, Penguin, 1972, 252-281.

Canadian French is reported in Lambert W.E. & G.R. Tucker (1976), Tu Vous Usted: A Social-Psychological Study of Address Patterns, Newbury House, Rowley Mass.
A useful starting book by Wierzbicka is: Wierzbicka, A. (1992), Semantics, Culture, and Cognition, CUP

Missing pronouns (pro-drop)

The pro-drop parameter is described in most recent books on grammar, for example:
Cook, V.J. and M. Newson (2007), Chomsky’s Universal Grammar, Blackwell, Oxford 3^rd edition.

Gender

A full treatment of gender can be found in:
Corbett, G. (1991), Gender, CUP, from which many of the language examples are taken.
The Times example comes from:
Bauer, L. (1994), Watching English Change, Longman, London:
The link between pronunciation and gender in French is based on:
Tucker, G.R., Lambert, W.E., and Rigault, A.A. (1977), The French Speaker’s Skill with Grammatical Gender, The Hague, Mouton.
The problem of gender and English pronouns is dealt with in many sources, such as:
Cameron, D. (1985), Feminism and Linguistic Theory, Macmillan, Basingstoke;
Coates, J. (1986), Women, Men, and Language, Longman.
Male speakers and female hearers feature in
W. Levelt (1989), Speaking, MIT.
The many attempts to coin non-sexist pronouns are documented in:
Baron, D. (1986), Grammar and Gender, Yale University Press;
per comes in:
Piercy, M. (1979)  Woman on the Edge of Time, Woman's Press Ltd.
see also Miller, C. and Swift, K. (1977), Words and Women: New Language in New Times, Anchor Press.
The sources on style are:
Fowler, H.W., and Fowler, F.G. (1918), The King’s English, Clarendon Press:
Strunk, W., & White, E.B. (1979), The Elements of Style, Allyn and Bacon, Boston.
The BBC survey is cited in Crystal, D. (1988), The English Language, Penguin

Pronoun systems

The sources cited for the English pronouns are:

Old English and Middle English: any history of English, say Strang:
New Forest: Wilson, J. (1913), The Dialect of the New Forest, OUP;
British Jamaican: Sutcliffe, D. (1982), British Black English, Oxford, Blackwell

Pronoun reference (binding)

Any standard introduction to grammar covers binding, say Cook and Newson (2007).

Vivian Cook Words website