1.1 Public interest, professional indifference?

It is paradoxical that, whereas many laypersons have a great deal to say about French T/V, most of the major overviews of spoken French have barely mentioned this topic (e.g. Blanche-Benveniste and Jeanjean, Reference Blanche-Benveniste and Jeanjean1987; Blanche-Benveniste, Reference Blanche-Benveniste1997; Gadet, Reference Gadet1997; Ball, Reference Ball2000). Even more surprising perhaps is the fact that French textbooks of sociolinguistics have devoted little, if any, attention to this important aspect of communicative competence (e.g. Marcellesi and Gardin, Reference Marcellesi and Gardin1974; Bachman et al., Reference Bachman, Lindenfeld and Simonin1981; Baylon, Reference Baylon1991; Gadet, Reference Gadet2007). In contrast most English-language overviews of sociolinguistics devote several pages to honorifics (‘polite forms’), including the T/V pronouns and other terms of address, in French and other languages (e.g. Fasold, Reference Fasold1990; Trudgill, Reference Trudgill2000; Romaine, Reference Romaine1994; Holmes, Reference Holmes2001; Mesthrie et al., Reference Mesthrie, Swann, Deumert and Leap2000; Wardhaugh, Reference Wardhaugh2005). It is noteworthy, however, that this material is usually located in different chapters from those dealing with ‘Labovian’ sociolinguistic variables.

In academic journals based in France which are oriented towards contemporary French or to sociolinguistics, the coverage of T/V has similarly been remarkably thin. In Le Français moderne, published since 1933, the only article on this topic appears to have been a brief study of usage in the French spoken in Mali, contributed by a British scholar (Constable, Reference Constable1984). The journals Langages and Langue française (founded in 1966 and 1969, respectively) do not appear to have published any article specifically on this topic. Part of Anscombre (Reference Anscombre1985) in Langages is devoted to a discussion of the meaning of the delocutive verbs tutoyer and vouvoyer, but not strictly to T/V as such. Alvarez-Pereyre (Reference Alvarez-Pereyre1977) in Langue française and Grimaud (Reference Grimaud1989) in Le Français moderne are on terms of address such as Madame, Monsieur, not T/V. However La Linguistique did publish an article on this topic in 1978, by a researcher based in Switzerland (Schoch, Reference Schoch1978), and more recently has published a contribution from Canada, devoted to problems of translating certain pronouns, including those of the second person (Schogt, Reference Schogt2004). The French Review, intended mainly for teachers of French in the United States, published a cluster of articles on T/V in the early 1970s, stimulated perhaps in part by the post-1968 increase in tutoiement (Bryan, Reference Bryan1972; Maley, Reference Maley1972; Bustin-Lekeu, Reference Bustin-Lekeu1973; Ford, Reference Ford1974), and two more subsequently (Vassallo-Villaneau, Reference Vassallo-Villaneau1991; Williams and van Compernolle, Reference Williams and van Compernolle2007). Le Français dans le Monde, which, like the French Review, has a readership of language teachers rather than linguistics researchers, has devoted at least two short articles to T/V (Calvet, Reference Calvet1976; Ibrahim, Reference Ibrahim1984). Langage et Société, founded in 1977, and probably the best-known sociolinguistics periodical in France, published nothing specifically on French T/V prior to a special issue in 2004 which was devoted to second-person pronouns (Pires, Reference Pires2004), and included some French-specific contributions.

These observations are not of course intended to belittle the significance of any of the publications referred to, or indeed the research on T/V by those French or other native francophone scholars who have published elsewhere. But the very modest amount of attention given to this topic in French linguistics journals is surprising, and one can readily sympathise with the view expressed recently by Gardner-Chloros: ‘Il est temps que le tutoiement/vouvoiement [. . .] reprenne sa place parmi les variables sociolinguistiques les plus centrales et révélatrices.’ (2003–2004: 98).Footnote ⁴

Why has there not been more research on T/V within linguistics or sociolinguistics in France? One possible explanation is that it is felt to be a topic that is not properly linguistic, one that lends itself rather to study within neighbouring disciplines, such as social psychology or anthropology. Another reason may be that the discipline of linguistics within France has traditionally been less receptive to sociolinguistics than has been the case in other western countries, perhaps due to the greater prestige attached in French academic circles (and in the French educational system) to theorising and abstract thought, rather than to research based on data.Footnote ⁵ In addition it seems to be generally recognised that the complexities of T/V usage arise from socio-psychological factors, rather than matters of linguistic distribution or constraints. This is of course precisely why T/V is a topic of such great interest to the general public. Even a relatively salient grammatical topic such as past participle agreement receives far less public attention than T/V, to judge by the number of mentions on the Internet: 26,200 ‘hits’ for l'accord du participe passé, compared with about 80,000 for tutoiement (on 8.2.08).

1.2 The disciplinary diversity of T/V research

Despite the paucity of articles on T/V in French academic journals, a large number of studies have in fact been conducted, especially by scholars based in other countries. Much of this research since the 1960s has taken as a major point of reference the celebrated article by Brown and Gilman (Reference Brown, Gilman and Sebeok1960), ‘The pronouns of power and solidarity’, but research on T/V had been carried out long before that, notably from the perspectives of philology and literary criticism: e.g. Schliebitz (Reference Schliebitz1886), Fay (Reference Fay1920) (both cited by Brown and Gilman), Foulet (Reference Foulet1918–19) and Bakos (Reference Bakos1955). A good deal of this early work concentrated on the complex patterns of switching between tu and vous in Old French texts. Post-1960 studies have adopted a broader range of disciplinary approaches, methods and analytical concepts, including the following:

• – philological and literary study of texts (e.g. Kennedy, Reference Kennedy, Barnett, Crow, Robson, Rothwell and Ullmann1972; Field, Reference Field1986, Reference Field1987; Lebsanft, Reference Lebsanft1987; Morrison, Reference Morrison1988; Hunt, Reference Hunt, Taavitsainen and Jucker2003);

• – cultural history (Wolff, Reference Wolff1994; Coffen, Reference Coffen2002);

• – socio-psychological questionnaires and experiments (Lambert, Reference Lambert1967, Reference Lambert and Puhvel1969; Lambert and Tucker, Reference Lambert and Tucker1976);

• – ethnographic interviews and long-term participant observation (Guigo, Reference Guigo1991; Morford, Reference Morford1997; Eckert, Reference Eckert, Sankoff and Cedergren1981); Sherzer (Reference Sherzer, Jazayery and Winter1988) analyses various types of recorded narratives about T/V;

• – socio-pragmatic analysis of conversation (Kerbrat-Orecchioni, Reference Kerbrat-Orecchioni1992; Martiny, Reference Martiny1996);

• – sociolinguistic questionnaires, interviews and, most recently, discussion- or focus-groups (Bustin-Lekeu, Reference Bustin-Lekeu1973; Schoch, Reference Schoch1978; Gardner-Chloros, Reference Gardner-Chloros1991, Reference Gardner-Chloros, Jones and Ayres-Bennett2007; Hughson, Reference Hughson2003, Reference Hughson, Peeters and Ramière2009; Warren, Reference Warren2006, Reference Warren, Peeters and Ramière2009).

Peeters (Reference Peeters2004) has provided a wide-ranging critical review of much of this work. Some researchers have recently begun studying T/V use in films (Abecassis, Reference Abecassis2002, Reference Abecassis2005; Hirvonen and Sutinen, Reference Hirvonen and Sutinen2005; Havu and Sutinen, Reference Havu, Sutinen, Broth, Forsgren, Norén and Sullet-Nylander2007) and in electronic written discourse, such as Internet ‘chat’, discussion fora and on-line dating advertisements (Williams and van Compernolle, Reference Williams and van Compernolle2007, Reference Williams and van Compernolle2009; van Compernolle, Reference van Compernolle2008). Researchers in Second Language Acquisition have also turned their attention to T/V usage in learners’ language, borrowing from the range of data types and approaches referred to above (e.g. Lyster, Reference Lyster1996; Dewaele, Reference Dewaele2002, Reference Dewaele2004; Liddicoat, Reference Liddicoat2006; Planchenault, Reference Planchenault, Peeters and Ramière2009).

The vast majority of empirical studies of T/V in spoken French have focused on reported usage (sometimes referred to as ‘perceptions’) rather than actual observation of behaviour. The main reason for this is that it has generally been thought to be impossible to gather much useful data on T/V by classic techniques such as the Labovian sociolinguistic interview, and that observing even one individual speaking with a wide range of addressees would be excessively time-consuming. Researchers using questionnaires, written or oral, have made the generally reasonable assumption that informants are able to report accurately on their own T/V usage with actual interlocutors. What is more open to doubt is whether informants can always report accurately on other people's usage, or on hypothetical situations. In contrast, studies that have drawn their data from written sources, such as novels, plays, film-scripts and Internet ‘chat’, have been able to focus much more on specific examples of T/V. The analysis of such data has produced many valuable insights, but it is unlikely that these sources reflect everyday usage in spontaneous speech – the principal object of study in sociolinguistics.

One exception to the general research pattern regarding the spoken language is Eckert (Reference Eckert, Sankoff and Cedergren1981), who was able to conduct long-term observation of usage of the second-person pronouns in a bilingual Gascon-French village (Saint Pierre de Soulan in the Ariège), during fieldwork for her main project on phonological variation and change. Similarly Guigo (Reference Guigo1991) observed the use of address terms (including pronouns) towards the end of his period of ethnographic fieldwork in the offices of a company based in Paris. Rapid Anonymous Observation and Language Diaries have both proved their worth as data-collection techniques in researching the related areas of language choice and code-switching (cf. Milroy, Reference Milroy1987; Gardner-Chloros, Reference Gardner-Chloros1985), but it would appear these have been employed only to a very limited extent for research into French T/V usage. Lambert and Tucker (Reference Lambert and Tucker1976: 52–3) report on a small-scale study in which they sent two teenage boys into department stores in Montreal: one boy used tu when addressing a saleswoman (inappropriately, given that this was still in the 1960s), while the second boy surreptitiously recorded the interaction, and, after the first boy had moved away, asked the saleswoman whether she was shocked by the inappropriate tutoiement. (In most cases, she was.) More recently, as part of Williams and van Compernolle's (Reference Williams and van Compernolle2007) study of T/V in Internet chat, the ‘fieldworker’ became a participant observer and conducted a simple experiment: having established that reciprocal tutoiement is the norm in this type of electronic discourse, he set out to discover how other members of the ‘chat room’ would react to sustained vouvoiement from him. (Result: they expressed surprise.) As for Language Diaries, I am not aware of any study that has used these in researching French T/V, though Gadet (Reference Gadet2007: 170) proposes a practical exercise along these lines for students using her sociolinguistics textbook. Although one might consider journalists’ articles on T/V as anecdotal or superficial, they do sometimes provide useful observations of actual behaviour and can thus, to a modest extent, help fill the gap left by academic research on the subject. For example, Négroni (Reference Négroni2007) describes some cases of upper-class French people who are maintaining vouvoiement within the family, though she also comments that the majority of young parents in this class now use reciprocal tu with their children.Footnote ⁶

In addition to Brown and Gilman's (Reference Brown, Gilman and Sebeok1960) ‘semantics’ of power and solidarity, several other concepts have been invoked to interpret and account for patterns of T/V usage, notably politeness (both as a pre-theoretical concept and as a technical notion within the Politeness Theory of Brown and Levinson (1978/Reference Brown, Levinson and Goody1987). Brown and Levinson characterise reciprocal V usage as expressing negative politeness (approximately equivalent to ‘social distance’, or the absence of solidarity), while reciprocal T conveys positive politeness (roughly, ‘solidarity’). The fact that both patterns are widespread within France, in different domains of linguistic behaviour, suggests that the country is located somewhere between the two poles of a ‘positive-politeness culture’ and a ‘negative-politeness culture’.Footnote ⁷ T/V has also been interpreted (e.g. by Gardner-Chloros, Reference Gardner-Chloros, Jones and Ayres-Bennett2007) in terms of the ‘acts of identity’ concept, originally developed by Le Page and Tabouret-Keller (Reference Le Page and Tabouret-Keller1985) to help account for sociolinguistic variation more generally, especially in multilingual and Creole-speaking communities. Since different social groups vary in their tendency to favour T or V, the individual's behaviour in this respect may reflect her/his desire to take on a particular type of identity.

The data produced by studies of T/V have been both quantitative and qualitative, and have been analysed with the help of a wide range of techniques and tools, from the schematic diagrams of Brown and Gilman (Reference Brown, Gilman and Sebeok1960), to Guigo's (Reference Guigo1991) chart representing the progression from formal to informal patterns and Béal's complex systemic network (Reference Béal1989, Reference Béal, Peeters and Ramière2009). Béal's network diagram is rather more intricate than the computer-style flow charts first produced to represent choices of address by Ervin-Tripp (Reference Ervin-Tripp, Pride and Holmes1972).Footnote ⁸ Naturally the various disciplinary approaches and data-collection methods have different strengths and weaknesses and these will not be reiterated here. But one general observation is that psychologically oriented work, including research conducted within the framework of Second Language Acquisition, tends to emphasise the statistical treatment of quantitative data, whereas ethnographic and pragmatic work has stressed instead qualitative data and its interpretation.

3.1 V as the unmarked variant

Although T/V is not then a stochastic sociolinguistic variable, it is nevertheless convenient to refer to T and V as ‘variants’, and in this section I will discuss the concept of markedness in relation to them. Many writers have described vous as the unmarked variant (e.g. Brown and Levinson, Reference Brown, Levinson and Goody1987: 292; Maingueneau, Reference Maingueneau1994; Hughson, Reference Hughson2003; Halmøy, Reference Halmøy, Peeters and Ramière2009). The fact that most journalistic pieces on this subject include tu or tutoiement in their title (rather than vous or vouvoiement) implies perhaps a view that tu is the marked variant, or, in media terms, that it is more newsworthy than vous.Footnote ¹² Peeters is one of the few writers who have, on the contrary, described T as unmarked:

In this section I will examine the evidence for considering either T or V as the unmarked variant.Footnote ¹³ In an important recent article, Haspelmath (Reference Haspelmath2006) has subjected the term ‘markedness’ to rigorous scrutiny, identifying twelve senses of the word that are commonly used in the linguistics literature. The discussion here will review those senses of markedness referred to by Haspelmath that are relevant to T/V.

As far as I am aware, there has been no in-depth justification of why vous should be considered the unmarked variant.Footnote ¹⁴ However one writer who has offered at least a brief rationale for this position is Maingueneau (Reference Maingueneau1994: 27), who describes vous as the background from which tu detaches itself. He points out that, while one may quite readily suggest a switch to T (e.g. by saying, Tutoyons-nous!), it is almost inconceivable to say Vouvoyons-nous!. Other, perhaps more frequent, expressions that are used to suggest switching to T include Si on se tutoyait? and especially On se dit ‘tu’?, while the corresponding phrases suggesting a move to V do indeed seem quite improbable: Si on se vouvoyait? and On se dit ‘vous’?. Although the latter are not quite unacceptable pragmatically, it is difficult to imagine them being used except as a joke, or in some rather strange situation.Footnote ¹⁵ With or without overt linguistic marking of the change, permanent switches from T to V seem to be very rare in the conversational history between any two adults. A rare real-life case of a seemingly permanent switch from T to V is reported to have occurred between an elderly man and his adult daughter in 2004, after she succeeded him as Mayor on the conseil municipal of Puteaux (Hauts de Seine, France), and he stayed on as Deputy Mayor. In addition to using reciprocal vouvoiement, it is reported that they have vigorously opposed each other in council debates since the change of roles.Footnote ¹⁶ Even temporary switches from T to V are very uncommon, although they do sometimes occur when there is a major change in the nature of the audience (i.e. the totality of participants present) and/or the situation (e.g. going ‘on-air’ on radio or television). Carton (Reference Carton2003) discusses the case of certain television interviewers and politicians who switch from reciprocal tutoiement when speaking off-air to reciprocal V once they go on-air. He claims that the switch to V amounts to an attempt to deceive the public, in that it presents the relationship between journalist and interviewer as a purely professional one, which will enable the journalist to ask challenging and difficult questions. Another rare example of temporary T-to-V switching is discussed by Bilger and Cappeau (Reference Bilger and Cappeau2004: 23) in their analysis of style-shifting in a non-media interview. The interviewee uses T to the old friend who is interviewing him, but then shifts to V when he starts talking about an association of which he is a member:

1. (1) je vais donc te présenter l'Association des Petits Frères des Pauvres [. . .] les problèmes dont je vous parlais c'est euh l'analphabétisme l'illettrisme

This case involves rather more than a mere change of topic: as Bilger and Cappeau point out, the interviewee seems at this point to assume the role of spokesperson of the association. In terms of Bell's theory of Style Shifting (Reference Bell1984, Reference Bell, Eckert and Rickford2001), one could describe this T-to-V switch as an instance of ‘referee design’ (rather than the more common ‘audience design’), whereby a speaker modifies their speech so as to take on temporarily another identity or persona. As Bell (Reference Bell, Eckert and Rickford2001) points out, referee design is similar to ‘acts of identity’ in the approach of Le Page and Tabouret-Keller (Reference Le Page and Tabouret-Keller1985).

A further point made by Maingueneau (Reference Maingueneau1994) is that vous is the spontaneously used form – except among children of course. He does not elaborate on what he means by ‘spontaneously used’, but it seems reasonable to suppose that he is thinking of cases where one adult is addressing another adult for the first time. This corresponds to the advice commonly given in textbooks for L2 learners of French: ‘En règle générale, il est conseillé d'attendre que votre interlocteur/trice vous tutoie.’ (Gervais and Sanders, Reference Gervais and Sanders1986: 22). One piece of evidence that perhaps suggests that laypersons consider tu to be marked is the fact that tutoiement has a far higher number of mentions on the Internet than does vouvoiement: c. 80,000 and c. 47,400, respectively (on 8.2.08). Similarly avoir le tutoiement facile is a fairly common expression (1,370 mentions on the Internet), while avoir le vouvoiement facile is not (just 6 ‘hits’). This implies that not everyone likes to use T with a wide range of addressees, including those outside one's own network of friends. In a comparative perspective, Kerbrat-Orecchioni observes that ‘the number of situations in which one chooses to use vous rather than tu is greater in France than in most neighbouring countries’ (2005: 41). One can readily accept that V is ‘unmarked’ in the sense that most French people opt for T only if they are addressing an adult belonging to their network of family, friends and colleagues – and even then, not all of these. But rather than the much-overused terms ‘marked’ vs ‘unmarked’, it is preferable to follow Haspelmath's advice to adopt a more specific term, such as ‘rare/frequent in the world’, or perhaps ‘situationally rare/frequent’.

Another sense of ‘markedness’ (Haspelmath, Reference Haspelmath2006) according to which one could designate vous as unmarked is ‘semantic markedness’. The example of this given by Haspelmath is the lexical pair dog/bitch, the former being the unmarked item since it can refer to the animal in general, regardless of sex. Similarly one could argue that vous is unmarked, since it can be used as the address pronoun, regardless of number.Footnote ¹⁷

3.2 T as the unmarked variant

Contrary to the majority opinion, I would like to argue that vouvoiement is the marked member of the T/V pair in French, and I will endeavour to provide supporting evidence for this viewpoint by examining a number of other senses of ‘markedness’. Not all the senses identified by Haspelmath are relevant to T/V, notably those relating specifically to phonetic or phonological phenomena. ‘Textual markedness’ (or ‘markedness as rarity in texts’) is also omitted here, because the relative frequencies of T and V are heavily dependent on the nature of the texts in question, and it seems impossible to make any meaningful generalisation about this that would be valid for all genres or text-types.

For the sake of conciseness of presentation, the relevant senses of ‘markedness’ are listed below, together with a brief indication of why French T can be said to be unmarked in that particular sense. (Words enclosed in quotation marks are of Haspelmath's formulation.)

• – ‘Formal markedness: markedness as overt coding.’ In the spoken form of the present indicative of regular first-conjugation verbs (the ‘–er’ category that includes the vast majority of verbs), the T verb form has no inflection, whereas the V form has /e/.

• – ‘Markedness as morphological difficulty or unnaturalness.’ In both spoken and written French, V forms (pronouns, determiners, verbs) are nearly always at least as long as the corresponding T forms, and are often longer. The only exceptions to this are the written pairs la tienne/la vôtre, les tiennes/les vôtres.Footnote ¹⁸ Moreover the V verb form in the present indicative is highly irregular in three of the most frequent verbs (êtes, faites, dites) and this is reflected in the fact that vous faisez and vous disez are not uncommon errors, even occurring sometimes in the speech of adult native speakers of French. The following example is from the Orléans corpus:Footnote ¹⁹

1. (2) alors quand le: – le maçon m'a dit hier mais vous saviez pas que si votre po votre tube avait éclaté vous faisez – tout tout écl à ce qu'il paraît –

(In this extract, the colon symbol represents a lengthened vowel, while the dash signifies a silent pause.)

• – ‘Cognitive markedness: markedness as conceptual difficulty’. Being the plural category, as well as the polite singular, V ‘is marked because it requires more mental effort and processing time than the singular’. There are occasions when a hearer needs to pause to decide whether V is intended as plural or singular.

• – ‘Situational markedness: markedness as rarity in the world’. Taking account of all interactions, and not just those involving adults outside one's social networks, it is almost certain that V is rarer than T in contemporary French. (This point will be further discussed below.)

• – ‘Typological markedness: markedness as typological implication or crosslinguistic rarity’. In the World Atlas of Language Structures, the majority of languages sampled (136/207) have no polite second-person pronoun such as vous (Helmbrecht, Reference Helmbrecht, Haspelmath, Dryer, Gil and Comrie2005).Footnote ²⁰

Of the various senses discussed by Haspelmath, it is ‘situational markedness’ that is the most sensitive to sociolinguistic factors. From a socio-historical viewpoint, it seems clear that, among the lower social strata – the overwhelming majority of the population down the centuries – reciprocal T has always been the norm in conversations between peers, and that such exchanges have almost certainly always represented the vast bulk of spoken language: children and young people talk mainly with their peers, and this surely has usually been true for most adults too. (For a sociological study of the number and type of people's interlocutors in contemporary France, see Héran, Reference Héran and Marpsat1990.) It has long been recognised that reciprocal T is normal among children, adolescents and often – though not universally – students, but Gadet has recently observed also that ‘the use of tu is now pretty much standard in the workplace in many professions’ (Reference Gadet, Ammon, Dittmar, Mattheier and Trudgill2006: 1789). However articles in business magazines suggest that behaviour is by no means uniform in the workplace: some managers encourage reciprocal T – or even try to impose it as a ‘management tool’ –, while others maintain the virtues of reciprocal V (cf. Beyer and Jouan, Reference Beyer and Jouan2006; Lévy, Reference Lévy2006; Telfizian, Reference Telfizian2006).

From the acquisitional viewpoint too, it seems clear that polite singular V is the marked variant: it is certainly acquired well after T, and although young children no doubt vary considerably regarding the age at which they master V, it seems uncontroversial to suggest that most start using vouvoiement systematically only after they go to school. Over a decade ago, Auger (Reference Auger and Moreau1997: 17) commented that there were no empirical studies of children's use of T/V in French, and this still appears to be the case. Studies of the acquisition of personal pronouns typically concern children aged less than 5, and tend to make no mention of polite V or indeed the plural persons (e.g. Girouard et al., Reference Girouard, Ricard and Gouin Décarie1997).Footnote ²¹ Evidence from other languages seems somewhat contradictory regarding the age at which children master V: on the one hand, Auger mentions a study of a five-year-old Spanish-speaking child who showed an ability to use the singular V pronoun (usted) while playing the role of a teacher; but on the other hand, Hollos (Reference Hollos, Ervin-Tripp and Mitchell-Kernan1977) found that, in a group of Hungarian children, even nine-year-olds were unable to sustain reciprocal use of V. Notwithstanding the lack of empirical research into children's acquisition of French V, there is no doubt that most very young children hear V forms regularly, notably from stories in books or on television and DVDs.Footnote ²²