INTRODUCTION
There is a long tradition of viewing French as a monocentric language, that is, a language that is homogeneous enough to have only one norm – namely the variety spoken in the Île-de-France (see e.g. Lüdi, Reference Lüdi and Clyne1992: 149). As Bourhis (Reference Bourhis, Coupland and Jaworski1997: 306–308) notes, the monocentric ideology around French is a consequence of the social history of the language in France, and particularly the notion of le bon usage. Defined by Vaugelas in the 17th century as the language of the most sensible part of the court, that is, the elite in Paris, le bon usage was conceived in stark contrast to the many other ‘bad usages’ – that is, any social and regional deviations from the norm. The elite in Paris considered changing aspects of grammar and vocabulary to lead inevitably to the corruption of the French language; it was felt that only the variety used by the elite could be considered as ‘the vehicle of French civilization’ throughout the francophone world, and all other varieties were dismissed as ‘degenerate patois’ (Bourhis, Reference Bourhis, Coupland and Jaworski1997: 308). Yet while the ideology of le bon usage persists to this day, it is in fact somewhat of a myth since it is extremely doubtful that this ‘correct usage’ actually corresponds to an existing sociolect. It is impossible to find a body of speakers whose variety fully and unwaveringly represents the standard of the language (Lodge, Reference Lodge1993: 189). Instead, ‘there is plenty of evidence that we are facing an illusion here, an imaginary object of utopian discourse’ (Lüdi, Reference Lüdi and Clyne1992: 155). Nevertheless, as Lüdi (Reference Lüdi and Clyne1992: 155) notes, the ideology of le bon usage appears to endure: ‘this phantasm belongs to the linguistic representations of most speakers of French, even of those who do not know it’. By many, the variety used by the cultivated Parisian bourgeoisie is thus still considered to be the only standard norm of the French language (see e.g. Picoche and Marchello-Nizia, Reference Picoche and Marchello-Nizia1989: 26; Lodge, Reference Lodge1993: 189). It is the persistence of this ideology that has led to the denigration of many other varieties of French.
One variety that has been particularly affected by this is Quebec French, which, for many years, was strongly disparaged, with the ‘myth of French Canadian Patois’ and the debates about joual (an account of which will be given in the next section) being the most significant instantiations of this disparagement. However, in the context of growing nationalist sentiment in Quebec, the 1970s saw the beginning of the debate about an autonomous and legitimate standard for Quebec French, and in 1977 the Association québécoise des professeurs de français (1977: 11) programmatically declared that the French taught and used in Quebec schools should be ‘le français standard d'ici’, which they defined as ‘the socially prestigious variety of French that the majority of Quebecers tend to use in formal communication situations’.Footnote 2 Since then, many have come to accept this standard norm of Quebec French – yet, surprisingly, there has been very little conclusive research into the manner in which this has affected the attitudes that Quebecers hold towards their own variety of French compared to that from France. Have they come to think of French as having several centres, each of which provides a national variety with its own norms – that is, have they come to regard French as a pluricentric rather than a monocentric language?
In order to shed light on this, the study presented here provides a contemporary perspective on the attitudes that Quebecers hold towards Quebec French (henceforth ‘QF’) compared to European French (henceforth ‘EF’), particularly with regard to the two main evaluative dimensions of language attitudes, that is, status and solidarity. It should be noted, however, that due to the non-representative nature of the subject sample, the study makes no claims regarding the generalisation of its findings to the Quebec population at large. The aim is simply to gain an insight into the attitudes of the subject sample investigated, which can then serve as a basis for further, more comprehensive research.
BACKGROUND AND PREVIOUS STUDIES
During the French regime, that is, from 1608 onwards, the variety of French spoken in Quebec was regarded very positively (see e.g. Pöll, Reference Pöll2005: 161). This is generally assumed to be a consequence of the relative uniformity of the French spoken by the first settlers – a uniformity that was reputed to be lacking in France itself, where large parts of the population did not speak françoys, as it was then called, but instead used patois, that is, other regional dialects (see e.g. Poirier, Reference Poirier, Mougeon and Beniak1994; Certeau et al., Reference Certeau, Julia and Revel1975). The Conquest in 1759, however, marked a turning point both in the history of the French language itself as well as in the perception of its use in Quebec: the province became a British possession and was cut off from France, and as a consequence of this rupture, the varieties of French spoken in the two locations began to differ significantly from each other. Not only was QF removed from the linguistic effects of the French Revolution (such as the spread of françoys throughout France and the subsequent changes within this variety) but the return to France of many members of the French elite also resulted in ‘a greater emphasis on the language of the masses’ in Quebec, which not only triggered the revival of many regionalisms but also opened the door to an influx of anglicisms (Oakes, Reference Oakes and Ayoun2008: 369). By the middle of the 19th century, Anglo-Canadians, Americans and the French from France had come to regard QF as an incomprehensible patois – and by the late 19th century, this so-called ‘myth of French Canadian Patois’ had resulted in a deep sense of linguistic insecurity amongst Quebec francophones (see e.g. Bouchard, Reference Bouchard2002: 95–98). Over the following decades, QF was strongly denigrated by foreigners and Quebecers alike. The 1960s saw the beginning of the debates about joual, a distortion of the word cheval that was taken from the expression parler cheval – that is, ‘to speak badly’ (Oakes and Warren, Reference Oakes and Warren2007: 111). While it was in fact little more than a (strongly anglicised) variety of urban, working-class French spoken mainly in Montreal, at the time, many equated joual with QF in its entirety and considered it to symbolise the degeneration of QF (see e.g. Bouchard, Reference Bouchard2002: 220; 222). These debates about joual evidently did nothing to alleviate the francophones’ linguistic insecurity. Two further factors that are believed to have contributed to this insecurity are the policy of successive French governments of establishing standard EF as the prestige norm across all parts of la francophonie, as well as the early policy of the Office de la langue française (see e.g. Bourhis and Lepicq, Reference Bourhis, Lepicq, Posner and Green1993: 366–368). The latter was established in Quebec in 1961 with the mandate of overseeing the ‘révalorisation’ of QF, of ensuring its correctness and enrichment (Levine, Reference Levine1990: 53), and the first bulletins of the Office clearly reflected the desire to align QF with the normative demands of standard EF (see e.g. Office de la langue française, 1965).
These historical and social circumstances serve as an explanation for the much more favourable attitudes towards EF than QF, particularly on the status dimension, that were evidenced by the findings of voice-evaluation experiments conducted amongst francophone and anglophone Quebecers between the late 1950s and the mid-1970s (Lambert et al., Reference Lambert, Hodgson, Gardner and Fillenbaum1960; Preston, Reference Preston1963; d'Anglejan and Tucker, Reference d'Anglejan, Tucker, Shuy and Fasold1973; Bourhis et al., Reference Bourhis, Giles and Lambert1975). In the late 1960s and the 1970s, as nationalist sentiment grew in Quebec, a much greater sense of the province's distinctiveness as a society began to emerge. Linguistically, this manifested itself in the francophones’ abandonment of the self-descriptor ‘Canadiens’ in favour of the new term ‘Québécois’ – which firmly rooted their sense of identity within their province (Conrick and Regan, Reference Conrick and Regan2007: 30). Moreover, in the mid-1970s, a more realistic perception of the French language began to take hold as Quebecers became more aware that French is not, in fact, homogeneous but that it is spoken differently by members of different social classes, in different regions and in different circumstances – and that this variation exists not only in Quebec but also in France (see e.g. Corbeil, Reference Corbeil, Plourde and Georgeault2008: 381). In 1977, the aforementioned declaration of the Association québécoise des professeurs de français sparked the debate about an autonomous and legitimate standard for QF, and since then, this standard has come to be accepted by a significant segment of the Quebec population. As Oakes (Reference Oakes and Ayoun2008: 373) notes, with regard to phonology, standard QF avoids the most stigmatised features of French as spoken in Quebec, such as the diphthongisation of long vowels, but it includes many traits that are now considered socially neutral – such as, for instance, the affrication of /t/ and /d/ before high front vowels (as in [ptsi] for petit, and [dzir] for dire) (see e.g. Ostiguy and Tousignant, Reference Ostiguy and Tousignant1993, for more detail on the phonology of standard QF). Moreover, in addition to certain features concerning morphology (e.g. the feminisation of profession titles, such as professeure for a female teacher) and syntax (e.g. the use of periphrastic verb phrases, such as être à faire for EF être en train de faire), standard QF is defined in particular by distinct lexical items: archaisms (e.g. s'abrier, which is now EF s'abriter), neologisms (e.g. poudrerie for fine snow in a blizzard), and borrowings (e.g. atoca, ‘cranberry’, from an Amerindian language) (see e.g. Conrick and Regan, Reference Conrick and Regan2007, for more detail).
With the emergence of this standard of QF, it appears that the linguistic referent has ‘at least partially been repatriated and linguistic correction is no longer exclusively measured by the yardstick of the bon usage from France’ (Pöll, Reference Pöll2005: 171). However, while there are many who accept this standard, Lockerbie (Reference Lockerbie, Lockerbie, Molinaro, Larose and Oakes2005: 16) notes that there are still linguistically more conservative Quebecers who do not agree that EF should be abandoned as the model of reference for French speakers in the province. There is thus an ongoing debate between the so-called endogénistes, who advocate the Quebec-internal standard (notably Pierre Martel and Hélène Cajolet-Laganière; see e.g. Martel and Cajolet-Laganière, 1996), and the exogénistes, who believe in the superiority of the European standard and in whose opinion the promotion of a Quebec-internal standard will lead to the linguistic ghettoisation of Quebec francophones (notably Annette Paquot and Lionel Meney; see e.g. Paquot, Reference Paquot2008; Meney, Reference Meney2010).
The continuing debate regarding the legitimacy of a standard norm for QF serves as an explanation for the fact that in the 1980s, the findings of a voice evaluation experiment by Genesee and Holobow (Reference Genesee and Holobow1989) revealed the persistence of more favourable attitudes towards EF on the status dimension, amongst both francophones and anglophones. On the solidarity dimension, however, both groups had shifted from a preference for EF to equally favourable attitudes towards EF and QF. This can be interpreted as a consequence of the newly emerging perception of Quebec's distinctiveness as a society, as well as Quebecers’ incipient sense of belonging to their province. Regrettably, no attitudes research was conducted amongst Quebec allophones up until this point. (‘Allophones’ is the term used in the Quebec context to describe those individuals who have a mother tongue other than French or English, excluding the First Nations. The allophones living in Quebec are immigrants – and individuals of immigrant descent – from countries where a vast variety of different mother tongues is spoken.Footnote 3 )
Not many studies were carried out in the more recent past, with the exception of two surveys by Bouchard and Maurais (Reference Bouchard and Maurais1999) and Laur (Reference Laur2001) and a set of interviews by St-Laurent (Reference St-Laurent2008). While the allophones in these studies openly admitted to preferring the normative EF model, the francophones and anglophones claimed they no longer considered this to be a norm to which they aspired, thus suggesting improved attitudes towards QF on the status dimension. It should be noted, however, that the outcomes of studies that employ only direct methods of attitude elicitation such as questionnaires or interviews have to be viewed with caution since they are often influenced by social desirability biases, that is, the tendency of respondents to reply in a manner that will be viewed favourably by others (see e.g. Hamers and Blanc, Reference Hamers and Blanc2000: 222–23). Consequently, further research by means of indirect methods such as voice evaluation experiments, which are less susceptible to such biases, was deemed necessary for clarification. Furthermore, regrettably, none of the more recent studies revealed much about attitudes towards QF and EF with respect to the solidarity dimension, thus also necessitating further research, and no previous study at all appears to have investigated the attitudes of French-English bilinguals.
The study presented here therefore aims to update and expand the existing body of research by shedding more light on the current attitudes that young francophones, anglophones, allophones as well as French-English bilinguals hold towards QF and EF in terms of status and solidarity, and the manner in which these attitudes might have been shaped by factors such as the continuing debate about a standard for QF and the burgeoning of Quebecers’ sense of belonging to their province.
Evidently, and as noted above, more than one variety of French exists in Quebec as well as in France, with variation depending on factors such as region, social class and level of education. The terms QF and EF are thus simplifications of a rather complex situation. ‘EF’ is used here to stand for the variety of French that is traditionally spoken by the bourgeoisie in the Île-de-France region. In previous attitudes studies, this has been labelled ‘Parisian French’ (Preston, Reference Preston1963), ‘standard European French’ (d'Anglejan and Tucker, Reference d'Anglejan, Tucker, Shuy and Fasold1973) or simply ‘European French’ (Bourhis et al. Reference Bourhis, Giles and Lambert1975; Genesee and Holobow, Reference Genesee and Holobow1989). ‘QF’, on the other hand, here refers to the educated, middle-class variety of French typically spoken in Montreal (see Genesee and Holobow, Reference Genesee and Holobow1989). This has variously been referred to as ‘educated’ or ‘cultured’ QF (Lambert et al., 1966), ‘a middle class Montreal accent’ (Genesee and Bourhis, 1982), ‘français soigné’ (Laberge and Chiasson-Lavoie, 1971) and ‘le français académique’ (Méar-Crine and Leclerc, 1976). Essentially, this is the standard QF described above.Footnote 4
METHOD
As indicated above, direct and indirect methods of attitude elicitation frequently yield rather different results since they pertain to different ‘levels of analysis’ (Ryan et al., Reference Ryan, Giles, Hewstone, Ammon, Dittmar and Mattheier1987: 1076). The purpose of direct methods is typically recognisable, and as most individuals try (consciously or unconsciously) to respond in an acceptable manner, the findings obtained by means of such direct methods tend to reveal what is considered to be socially desirable. The major strength of indirect methods, on the other hand, lies in the spontaneous elicitation of data that are significantly less sensitive to reflection and social desirability biases. It is thus assumed that in studies of this kind, more private reactions are revealed. The use of both direct and indirect methods is thus advisable since it is impossible to make any definitive statements about language attitudes if only one type of measurement is employed (see e.g. Ryan et al., Reference Ryan, Giles, Hewstone, Ammon, Dittmar and Mattheier1987: 1076). The present study therefore made use of the combination of a questionnaire and a matched-guise experiment.
Participants
The majority of previous language attitudes studies were conducted amongst youngsters from Montreal, the urban centre of Quebec. While the rest of the province is rather homogeneously francophone, the Montreal metropolitan region is home to comparatively large anglophone and allophone communities (12.1% and 21.6% respectively; Statistics Canada, 2007), which lends it a very distinct character. For the sake of comparability with earlier research, young Montrealers were thus also decided upon as the subject sample for this study.
All participants were students at Collèges d'enseignement général et professionnel, typically abbreviated to ‘CEGEPs’ – a type of post-secondary education institution exclusive to the province of Quebec. Usable data was collected from 160 students, whose mean age was 18.3 years. Based on Statistics Canada's (2009) definition of mother tongue as the first language a person learned at home in childhood and still understands, and taking into account that some respondents may declare that they learned two or more languages simultaneously, the students’ self-evaluations were used to classify them into four groups: 55 whose mother tongue was French (and in some cases another language) (henceforth: francophones), 44 whose mother tongue was English (and in some cases another language) (henceforth: anglophones), 13 who had both French and English as their mother tongues (henceforth: French-English bilinguals), and 48 whose mother tongue was a language/languages other than French or English (henceforth: allophones).Footnote 5 It should be noted that the term ‘French-English bilinguals’ was chosen for brevity's sake to refer to those respondents who claimed to have learned both languages in childhood, at home, and were still proficient in them. The majority of respondents from all mother tongue groups (overall: 93.1%) were in fact multilingual, and many of them stated that they were fluent in both French and English. Consequently, the variable under investigation here really is mother tongue rather than knowledge of languages.
As many Montrealers have an immigrant background, the participants were also divided into actual immigrants (26.8%), Montrealers of immigrant descent (36.3%), and non-immigrants (36.9%). However, the variable immigrant background did not have a significant effect on the participants’ attitudes. Consequently, it is only mentioned briefly in the discussion of the results. The overall subject sample contained more than twice as many females (70.6%) than males (29.4%); however, there was no evidence of significant non-orthogonality, and consequently, the influence of the variable sex is excluded from the discussion. Since, for reasons of space, the questionnaire did not inquire about the respondents’ socio-economic or ethnic backgrounds, the influence of these variables could not be ascertained. The focus will thus be on the attitudes held by the four different mother tongue groups.
Over 97.5% of the subject sample had knowledge of French, and as evidenced by Table 1, the trend was for the majority of respondents from each mother tongue group to describe their own variety of French as QF (or close to it), followed by those claiming to speak a variety in between QF and EF, and a minority who described their own variety as EF (or close to it).Footnote 6
Cross-tabulation of mother tongue (L1) and variety of French spoken.
Observed numbers (N), expected numbers (Ne), percentages (%) and results of Chi-square test.
N.B. The participants were originally asked to rate their own variety of French on a scale from 1 to 5, with 1 meaning ‘QF’ and 5 meaning ‘EF’ – it was only for simplicity's sake that the values were recoded for presentation in this table.
Procedure
Testing was conducted in different classes at two French-medium and two English-medium CEGEPs in the autumn of 2007. The research instruments were administered to all students in a given class at the same time, and the procedure took one lesson in total. At the beginning of the lesson, the matched-guise experiment was conducted in accordance with the standard procedure contrived by Lambert et al. (Reference Lambert, Hodgson, Gardner and Fillenbaum1960): the students were played recordings of three speakers who each read the same passage of prose three times, once in QF, once in EF, and once in English.Footnote 7 The subjects remained unaware of the fact that they were hearing the same speakers three times, in matched guises, but were instead under the impression that they were listening to a series of nine different speakers. They were asked to rate personality characteristics of these speakers on semantic scales for certain traits (such as intelligence, dependability, and kindness). To avoid the influence of social desirability biases, the real purpose of the experiment was withheld. With the subjects remaining unaware that they were hearing the same speakers in matched guises, any differences in reaction to the different text fragments read by the same speaker can be presumed to be based on the subjects’ attitudes towards the different varieties spoken.
Following the matched-guise experiment, the participants were debriefed and finally, the questionnaire was administered. The rationale behind the decision to use the same subject sample for the matched-guise experiment and the questionnaire was that this allows for a greater degree of comparability between the results obtained by means of the two methods, which leads to a more complete understanding of the participants’ attitudes. As one JFLS reviewer noted, it could be argued that the use of matched subject samples for the different methods might have had the advantage that those participants who only filled in the questionnaire would not have been as acutely aware of the issue under investigation, that is, language attitudes. However, the actual wording of the questionnaire would have revealed this in any event. Furthermore, the advantage gained by the greater comparability of the results was considered to outweigh the possible advantage that might have been gained by using matched subject samples.
MATERIALS
Questionnaire
The background section of the questionnaire was designed to ascertain the participants’ age, sex, mother tongue and immigrant background, their own variety of French, as well as the amount of time that the immigrants to Quebec had spent in the province. The main section contained three items each that were designed to elicit attitudes towards QF versus EF on the status and solidarity dimensions. These items were closed questions whose response options were five-point, interval, Likert-like scales (1 = completely agree, 2 = partially agree, 3 = undecided, 4 = don't really agree, 5 = don't agree at all).Footnote 8
A variety that is perceived to have much status is defined as one that is, inter alia, associated with both economic opportunity and upward social mobility (see e.g. Echeverria, Reference Echeverria2005: 249). The item ‘Speaking QF (rather than EF) increases my opportunities to find employment’ was thus designed to elicit attitudes with respect to the first component of this definition. The aim of the item ‘QF speakers are likely to get farther in life than EF speakers’ was to ascertain attitudes in terms of the second component. Based on the assumption that a variety's suitability to modern society implies its usefulness, the item ‘QF is better suited to modern society than EF’ sought to investigate the utilitarian value that is attributed to QF and EF – another important characteristic of varieties that are associated with status and social recognition (see e.g. Gardner and Lambert, Reference Gardner and Lambert1959: 267).
Conversely, a variety that is evaluated highly on the solidarity dimension is one that ‘elicits feelings of attraction, appreciation and belonging’ – which is typically the case for the variety of one's family life and intimate friendships since this ‘acquires vital social meaning and comes to represent the social group with which one identifies’ (Ryan et al., Reference Ryan, Giles, Sebastian, Ryan and Giles1982: 9). As it can be assumed that it is mainly in their family life and/or intimate friendships that individuals share their joys as well as their concerns, the item ‘QF lends itself to expressing feelings and emotions more readily than EF does’ sought to find out to what extent this is the case for QF and EF amongst the participants. Based on the assumption that cultural heritage constitutes a significant part of a group's distinctiveness, the item ‘Speaking QF (rather than EF) constitutes a significant part of Quebec's cultural heritage’ aimed to find out to what extent the varieties elicit feelings of attachment and belonging at the level of group identity. In order to obtain a more complete picture, the item ‘Speaking QF (rather than EF) is important for my sense of personal identity’ was used to ascertain the extent to which the varieties are important to the respondents at the level of individual identity.
Additionally, one more open-ended item was included in the questionnaire: ‘What five words do you think are the most suitable to describe QF?’. The words obtained by means of this item were categorised into five groups: ‘status-related – favourable’, ‘solidarity-related – favourable’, ‘general – favourable’, ‘general – negative’ and ‘other’. (The ‘other’ words were descriptors that were neither clearly favourable/negative, nor did they unequivocally pertain to the status or the solidarity dimension; see Appendix A for the classification of the descriptors.) The objective of this item was to investigate the respondents’ own views of what QF is and what they associated it with. Therefore, in order not to bias their choice of descriptors, no definition of QF was provided in the questionnaire.
Stimulus recordings
The stimulus recordings employed for the matched-guise experiment were the same as those used by Genesee and Holobow. They describe their selection process as follows (Genesee and Holobow, Reference Genesee and Holobow1989: 23–24): ‘Recordings in each of three guises – Canadian English, Quebec French and European French – were made of five “trilingual” males [. . .]. All speakers spoke middle class language varieties [which essentially conform to the descriptions of standard QF and EF provided above]. [. . .] The present recordings were then presented in random order to 20 male and female undergraduate psychology students from McGill University who were asked to judge the first language and nationality of each speaker. These judgements were used to select three of the five speakers for inclusion in the study. The three speakers who were selected were judged to be native speakers in each of their guises by at least 85% of the students.’ The recordings of these speakers were then arranged so that no two guises of the same speaker and no two recordings of the same variety occurred consecutively (see Appendix B for the order).
The text employed in the matched-guise experiment originated from a bilingual airline magazine (private communication with Fred Genesee; for the text, see Appendix C). Evidently, neither texts nor the topics they deal with are ever really neutral. Some might therefore consider the topic of this particular text to be emotionally charged. However, the manner in which it is delivered by the speakers is unemotional, and moreover, the passage does not contain any socially, ideologically or politically charged information, and it is in no way language-related. It can therefore be hoped that it had only a negligible influence upon the subjects.
Studies such as Laberge and Chiasson-Lavoie (1971) and Lappin (1982) have shown that individuals’ judgements of speakers are affected by lexical and syntactic as well as phonological factors. In order to avoid a bias in favour of either EF or QF, the text did not contain any lexical items or syntactic constructions that were clearly European or québécois – and consequently, the evaluations made of the speakers revealed attitudes based on pronunciation only. Since the respondents were asked to evaluate each recording in its entirety (rather than each linguistic variable in each of the recordings), it is regrettably not possible to ascertain whether there are differences in the extent to which different phonological variables affect attitudes.
The age of the recordings is evidently a drawback. However, while they originate from the 1980s, they do not contain any phonological variables that were not considered to be part of standard QF – either at the time the recordings were made or at the time this study was conducted. Moreover, even when the respondents were asked directly whether they thought that there was anything unusual about the recordings (see below), they made no comments indicative of the recordings not sounding current. This suggests that the age of the recordings was not overly noticeable or problematic.
Evaluation sheet
The subjects were asked to give their impressions of each language sample on 16-point rating scales with 0 meaning ‘not at all. . .’ and 16 meaning ‘very . . .’.Footnote 9 Five of the scales pertained to status-related traits: intelligence, dependability, education, ambition, and leadership. The other five scales regarded solidarity-related traits: kindness, humour, warmth, likeability, and sociability. All of these evaluation traits had previously been employed in numerous other investigations of language attitudes in Quebec (e.g. Genesee and Holobow, Reference Genesee and Holobow1989; Lambert et al., Reference Lambert, Hodgson, Gardner and Fillenbaum1960).
On the final page of the evaluation sheet, the participants were asked for their opinion regarding the purpose of the study, and whether they thought that there was anything unusual about the voices they had heard. The purpose of these questions was to enable the researcher to ascertain whether any of the subjects had guessed the actual aim of the experiment – and 4 out of the originally 164 subjects did indeed guess the real objective. As with matched-guise studies, the participants’ ignorance with regard to the methodology is crucial for the elicitation of valid results, these participants were removed from the study, resulting in an overall number of 160.
STATISTICAL ANALYSES
The data obtained from the closed survey questions was analysed by means of analyses of variance (ANOVAs) that used mother tongue and immigrant background as the independent, between-subject variables. Duncan PostHoc tests were used to ascertain statistically significant differences. Since the categorisation of the words obtained by means of the open-ended survey question was a subjective one, based purely on the researcher's own assessment of matters, it was not ascertained whether the percentages of words in the different categories were in fact statistically significant.
After combining the ratings of the speakers, the matched-guise data was analysed by means of repeated measures ANOVAs that also used mother tongue and immigrant background as the independent, between-subject variables.
Furthermore, for the data obtained by means of both methods, correlations with two-tailed tests of significance were performed in order to investigate firstly, the relation between the participants’ own variety of French and their evaluations of QF versus EF, and secondly, the relation between the amount of time that the immigrants had spent in Quebec and their evaluations of QF versus EF.
RESULTS
Attitudes on the status dimension – results of the questionnaire
Table 2 presents the results of the ANOVAs that were performed on the questionnaire data to establish attitudes towards QF compared to EF on the status dimension. It appears that the participants did not hold more favourable attitudes towards QF than EF in this respect: none of the mother tongue groups regarded QF to be better suited to modern society than EF. Neither did they think that speaking QF rather than EF was likely to increase their opportunities to find employment, with a PostHoc test revealing that the allophones evaluated QF even less favourably in this respect than the other mother tongue groups. Moreover, the participants from all mother tongue groups alike judged QF speakers to be unlikely to get farther in life than EF speakers.
Results of the questionnaire: evaluations of QF versus EF on the status dimension.
Item, mother tongue groups (L1), absolute numbers (N), means, F value (F), degrees of freedom (df) and level of significance (sig.). Means: 1 = completely agree, 2 = partially agree, 3 = undecided, 4 = don't really agree, 5 = don't agree at all.
Due to the comparative nature of the items, it is not possible to ascertain from this data whether the participants simply evaluated QF and EF equally favourably on the status dimension or whether their ratings of EF were in fact more favourable. (This is a weakness of the research instrument. Ideally, it should have elicited separate evaluations of QF and EF.) However, the findings obtained by means of the open-ended item ‘What five words do you think are the most suitable to describe QF’ suggest that the latter was the case: only 3.4%, 3.1%, 6.8% and 0.6% of the words chosen by the anglophones, the francophones, the French-English bilinguals, and the allophones respectively could unequivocally be classified as indicating favourable attitudes towards QF on the status dimension. (Examples of these few words include ‘useful’, ‘practical’ and ‘necessary’.) Furthermore, it is noteworthy that the majority of the words that the anglophones, the French-English bilinguals, and the allophones chose to describe QF – and even a relatively large number of the words selected by the francophones – were clearly negative (64.0%, 52.3%, 66.2%, and 23.2% respectively).Footnote 10 Many of these words, such as ‘patois’, ‘joual’, ‘Fran/glais’, ‘accented’, and ‘anglicised’, imply that the participants considered QF to be non-standard and impure. Other frequently used words, including ‘farmer’, ‘hick’, ‘hillbilly’, ‘dumb’, and ‘ignorant’, either imply or clearly state a lack of education. The choice of these descriptors can be seen to suggest that many of the respondents still associated (and possibly even equated) QF with joual, the aforementioned anglicised variety of urban working class French that was considered to symbolise the degeneration of QF in the debates of the 1960s. Overall, the findings obtained by means of the questionnaire thus suggest that the participants from all mother tongue groups had rather negative attitudes towards QF in terms of status, and the allophones’ attitudes appeared to be even more negative than those of the others.
Attitudes on the status dimension – results of the matched-guise experiment
As evidenced by Table 3, the results of the repeated measures ANOVAs that were performed on the matched-guise data confirm the impression that the participants regarded QF less favourably than EF in terms of status. The results show that on four out of the five status traits, the EF guises were rated more favourably than the QF guises, and for three of these traits, the difference was statistically significant (i.e. ‘dependability’, ‘intelligence’ and ‘education’; the difference for ‘ambition’ was not significant). While the QF guises were rated slightly more favourably in terms of one trait, namely ‘leadership’, this was not statistically significant. Overall, the results of the matched-guise experiment thus clearly indicate more favourable attitudes towards EF than QF on the status dimension amongst the participants from all mother tongue groups.
Results of the matched-guise experiment: evaluations of speakers in different guises in terms of status traits.
Absolute numbers (N), means for guises (QF, EF), F value (F), degrees of freedom (df) and level of significance (sig.). Means: 0 = not at all, 16 = very.
N.B. Since mother tongue did not have a significant effect on the ratings, only the total means are presented here.
Attitudes on the solidarity dimension – results of the questionnaire
Table 4 shows the outcome of the ANOVAs that were performed on the questionnaire data to establish attitudes towards QF compared to EF on the solidarity dimension. These findings indicate that while none of the mother tongue groups regarded QF to lend itself better to expressing feelings and emotions than EF, they all judged QF (rather than EF) to be a significant part of Quebec's cultural heritage. The francophones also agreed that speaking QF rather than EF was important for their sense of personal identity; the remaining mother tongue groups, however, did not consider this to be the case.
Results of the questionnaire: evaluations of QF versus EF on the solidarity dimension.
Item, mother tongue groups (L1), absolute numbers (N), means, F value (F), degrees of freedom (df) and level of significance (sig.). Means: 1 = completely agree, 2 = partially agree, 3 = undecided, 4 = don't really agree, 5 = don't agree at all.
It is not possible to ascertain from the data whether the participants simply evaluated QF and EF equally favourably with regard to the expression of feelings as well as the varieties’ importance for their sense of personal identity, or whether their ratings of EF were in fact more favourable. Nevertheless, overall, these results suggest that the participants from all mother tongue groups held more favourable attitudes towards QF in terms of solidarity than they did in terms of status, and that the francophones’ attitudes were even more favourable than those of the other mother tongue groups. The results obtained by means of the more open-ended question corroborate the latter notion: a sizeable 17.1% of the words selected by the francophones could unequivocally be classified as indicating favourable attitudes towards QF on the solidarity dimension (compared to only 5.7%, 4.5% and 1.3% of the descriptors chosen by the anglophones, the French-English bilinguals, and the allophones, respectively). Many of these words concerned feelings and emotions (e.g. ‘emotional’, ‘moving’) or a sense of community (e.g. ‘belonging’, ‘propre à nous’). The francophones also selected words connected to their family as well as their own identity (e.g. ‘personnel’, ‘identité’, ‘moi’).
Attitudes on the solidarity dimension – results of the matched-guise experiment
The findings of the repeated measures ANOVAs that were performed on the matched-guise data (see Table 5) indicate that all four mother tongue groups held more favourable attitudes towards QF than EF in terms of solidarity. The findings show that the QF guises were rated more favourably than the EF guises in terms of all solidarity traits, and that this difference was statistically different for four out of the five traits (i.e. ‘humour’, ‘warmth’, ‘likeability’, and ‘sociability’; only the difference for ‘kindness’ was not significant).
Results of the matched-guise experiment: evaluations of speakers in different guises in terms of solidarity traits.
Absolute numbers (N), means for guises (QF, EF), F value (F), degrees of freedom (df) and level of significance (sig.). Means: 0 = not at all, 16 = very.
N.B. Since mother tongue did not have a significant effect on the ratings, only the total means are presented here.
Correlations
Furthermore, a statistically significant correlation was found to exist between the variety of French spoken by the respondents themselves and their evaluations of QF in the questionnaire: the more québécois the participants judged their own variety of French to be, the more favourably they evaluated QF on both dimensions (see Table 6). Regarding the status dimension, the more the respondents felt they spoke QF, the better suited they deemed this variety to be to modern society and the more they thought that speaking QF rather than EF would increase their opportunities to find employment. (Moreover, the closer to QF the participants judged their own variety of French to be, the more likely they thought QF speakers to get farther than EF speakers; however, this was not statistically significant.) With regard to the solidarity dimension, the respondents who claimed to speak a more québécois variety of French deemed QF to lend itself more readily to expressing feelings and emotions. They also judged speaking QF rather than EF to be a more significant part of Quebec's cultural heritage and considered it more important for their sense of personal identity.
Results of the questionnaire: correlations (with two-tailed tests of significance) between variety of French spoken and evaluations of QF versus EF.
Dimension, item, absolute numbers (N), Pearson's correlation (r) and level of significance (sig.).
A correlation was also found to exist between the participants’ own variety of French and their evaluations of QF in the matched-guise experiment, suggesting that the more québécois the respondents of all backgrounds judged their own variety of French to be, the more favourable the evaluations they made of the QF guises – however, this trend was not found to be statistically significant (see Table 7).
Results of the matched-guise experiment: correlations (with two-tailed tests of significance) between variety of French spoken and evaluations of the QF guises.
Dimension, trait, absolute numbers (N), Pearson's correlation (r) and level of significance (sig.).
As evidenced by Table 8, there did not appear to be a correlation between the number of months that those respondents who were immigrants to Quebec had spent in the province on the on hand, and their evaluations of the matched-guise speakers in their QF guises on the other hand.
Results of the matched-guise experiment: correlations (with two-tailed tests of significance) between number of months spent in Quebec and evaluations of the QF guises amongst immigrants.
Dimension, trait, absolute numbers (N), Pearson's correlation (r) and level of significance (sig.).
Interestingly, however, there was a statistically significant correlation between the number of months the immigrants had spent in Quebec and their ratings of QF in the questionnaire: as evidenced by Table 9, the more time the immigrants had spent in the province, the more favourable were their evaluations of QF compared to EF. This correlation was statistically significant for one of the status-related items (‘QF is better suited to modern society than EF’) and one of the solidarity-related items (‘QF lends itself to expressing feelings and emotions more readily than EF does’).
Results of the questionnaire: correlations (with two-tailed tests of significance) between number of months spent in Quebec and evaluations of QF versus EF amongst immigrants.
Dimension, item, absolute numbers (N), Pearson's correlation (r) and level of significance (sig.).
DISCUSSION
Regarding the status dimension, the overall results of both the questionnaire and the matched-guise experiment are indicative of more favourable attitudes towards EF than QF amongst all four mother tongue groups. It is commonly assumed that the overt status that is attributed to a particular variety is indicative of the official acceptance that this variety enjoys. The overall results of this study can thus – tentatively – be interpreted as indicating that, at least amongst the young Montrealers who took part in this study, QF has not yet come to be considered a national variety in its own right, and that EF is still considered the standard norm to be aspired to. When interpreted in the broader historical and social context, it is not entirely unexpected that these young Montrealers should continue to conceive of French as a monocentric language with EF as the only standard. As mentioned above, for well over a century, QF was denigrated by foreigners and Quebecers alike. It is true that since the beginning of the debate about an autonomous and legitimate standard for QF in the 1970s, many have come to accept and advocate this standard. Nevertheless there are still linguistically more conservative Quebecers who do not agree that EF should be abandoned as the model of reference for French-speakers in the province. Furthermore, standard QF has not yet been codified by means of an officially sanctioned dictionary – and codification is a process that usually has the effect of legitimising a variety. Even in official circles, it has long been recognised that an officially sanctioned dictionary of QF could improve the prestige of the variety by providing ‘a solid, formally documented reference which would demonstrate the existence and legitimacy of a standard Quebec French’ (Conseil de la langue française, 1991). However, it is only now that such a dictionary is being produced (see FRANQUS), and in the absence of a codified form of QF, reference continues to be made to works produced in France – thus perpetuating the conception of EF as the only standard form of the French language. These circumstances can be seen to account for the young Montrealers’ negative attitudes towards QF on the status dimension as evidenced by the results of both the questionnaire and the matched-guise experiment.
As mentioned above, the outcome of the questionnaire suggests that the allophones in this study held even less favourable attitudes towards QF in terms of status than the other mother tongue groups. This, too, can – albeit only tentatively – be explained in terms of the social context. As Lockerbie notes (2005: 28, 53–54), outside the province of Quebec, the notion that there is a standard of QF is still very recent and not yet in general circulation. EF remains the target norm taught to those learning French as a second language in most parts of the world (see e.g. Bourhis, Reference Bourhis, Coupland and Jaworski1997: 307) – and the conception of French amongst the majority of newcomers to Quebec is therefore likely to be even more monocentric than that of those who were born in the province. The majority of the allophones who took part in this study were immigrants (66.7%, compared to only 5.6% of francophones, 11.9% of anglophones, and 15.4% of French-English bilinguals). While the variable immigrant background did not have a statistically significant effect on the results, it was nevertheless noted that the immigrants consistently evaluated QF less favourably in terms of status than the other participants. It is thus possible that the allophones’ especially negative attitudes towards QF in terms of status are a reflection of the aforementioned even stronger monocentric conception of the French language amongst newcomers to Quebec. This notion is supported by the fact that the more time those participants who were immigrants had spent in Quebec, the more favourable their attitudes were towards QF compared to EF – a finding that is statistically significant. It is likely that with time, these newcomers had become increasingly aware of the fact that there is a complex hierarchy of social variation within QF – and that QF in its entirety is not synonymous with joual, as some still assume (see e.g. Turcotte, Reference Turcotte2009). Moreover, it is probable that the more time the newcomers had spent in Quebec, the more they had become aware of the fact that there is in fact a relatively broad consensus concerning a standard for QF – at least in the province itself.
Regarding the solidarity dimension, the survey results were indicative of more favourable attitudes towards QF amongst the participants from all four mother tongue groups, and particularly amongst the francophones. The outcome of the matched-guise experiment was suggestive of more favourable attitudes towards QF than towards EF on the solidarity dimension amongst all participants, regardless of their mother tongue. As mentioned above, individuals typically evaluate the variety of their family life and/or intimate friendships most favourably on the solidarity dimension, as this comes to represent the social group with which they identify. Since the majority of participants from all four mother tongue groups claimed to speak either QF or a variety close to it, it can be assumed that it was also QF rather than EF that they used in their intimate friendships. This circumstance serves as a likely explanation for the more favourable attitudes towards QF than EF on the solidarity dimension. It also lends itself as a possible account for the observation that the more québécois the respondents of all backgrounds judged their own variety of French to be, the more favourable their attitudes were towards QF. As noted above, amongst those participants who were immigrants, more favourable attitudes towards QF on the solidarity dimension were displayed by those who had been in the province for longer. This could be due to the fact that the more time they had spent in Quebec, the more they had come to use QF as the variety of their close friendships. Finally, the aforementioned more favourable attitudes towards QF amongst the francophones might be a reflection of the fact that in their case, QF was presumably not only the variety used in intimate friendships but also in their family life – therefore assuming an even more important role.
Evidently, it is possible that those tendencies that emerged solely from the questionnaire results – that is, the more favourable attitudes towards QF amongst the francophones, the participants whose own variety of French was more québécois, as well as those immigrants who had lived in Quebec for longer – might be reflections of social desirability biases. The overall trend, however – that is, the generally more favourable attitudes towards QF than EF on the solidarity dimension – surfaced in the findings obtained by means of both the questionnaire and the matched-guise experiment. Consequently, it is highly unlikely that this trend is solely a result of social desirability biases.
As noted above, this overall trend of more favourable attitudes towards QF in terms of solidarity was not evidenced by the findings of earlier studies. This can be explained in terms of the broader social and political context. As mentioned previously, the nationalist movement in the 1970s led to the emergence of a sense of Quebec's distinctiveness as a society – distinctive not only from the rest of Canada but also from the country of the francophones’ ancestors, that is, France. Since the variety typically evaluated most favourably on the solidarity dimension is the one that represents the social group with which one identifies most and that elicits the strongest feeling of belonging, this emergent sense of distinctiveness constitutes a probable account for the shift amongst the francophones and anglophones from more favourable attitudes towards EF than QF (as evidenced between the late 1950s and the mid-1970s: Lambert et al., Reference Lambert, Hodgson, Gardner and Fillenbaum1960; Preston, Reference Preston1963; d'Anglejan and Tucker, Reference d'Anglejan, Tucker, Shuy and Fasold1973; Bourhis et al., Reference Bourhis, Giles and Lambert1975) to equally favourable attitudes towards both varieties in terms of solidarity (as evidenced in the 1980s: Genesee and Holobow, Reference Genesee and Holobow1989). The burgeoning of many Quebecers’ feeling of belonging to this distinctive society that has occurred since then lends itself as a likely explanation for the shift to more favourable attitudes towards QF than EF, as evidenced in the findings of the study presented here.
CONCLUSION
Evidently, the subject sample employed in this study is too small to allow for a generalisation of the findings. More work with larger and more diverse subject samples is desirable, and such work should also examine in more detail the influence that variables such as age, social background and ethnic background have on individuals’ language attitudes. Furthermore, since studies such as Laur (Reference Laur2008) indicate that the sex of speakers has a significant impact on listener evaluations, future matched-guise experiments should examine this issue in more detail. Such investigations would enable a more complete understanding of language attitudes, in Montreal as well as in general.
Yet while they might not allow for generalisations, the findings of this study nevertheless shed light on the language attitudes held by the young Montrealers who participated in this particular study. In summary, the results presented here suggest that the francophone and anglophone Montrealers’ attitudes towards QF compared to EF on the status dimension were the same as those of the francophones and anglophones who took part in the last significant attitudes study in Montreal, namely Genesee and Holobow (Reference Genesee and Holobow1989), while their attitudes on the solidarity dimension were different. (While little is known about the attitudes held by allophones and French-English bilinguals in the past, the findings of the present study indicate that at least now, these groups share the attitudes held by the francophones and anglophones.) The shift from neutrality to a preference for QF on the solidarity dimension can be explained in terms of the burgeoning of Quebecers’ feeling of belonging to their society. The fact that more status continues to be attributed to EF than to QF, which is indicative of the persistence of the monocentric ideology surrounding the French language, is accounted for principally by the lack of codification of standard QF by means of an officially sanctioned dictionary. However, attitudes towards QF in terms of status may well improve in the future since such a dictionary is currently being created by the FRANQUS team at the University of Sherbrooke. Inter alia, this dictionary will take québécismes to be the norm and mark helvétismes as such. Regrettably, the production of this dictionary has proved to be extremely lengthy and its publication has been postponed several times already. Nevertheless, once it has been published, this new dictionary will constitute a milestone in Quebec language planning and it is expected to assist strongly in the promotion of a new, pluricentric conception of the French language (Oakes and Warren, Reference Oakes and Warren2007: 126). As Pöll (Reference Pöll2005: 161) points out, this is even more meaningful when one considers that Quebec is the only francophone territory to have attempted the codification of its own linguistic norms, thereby defying the traditional hegemony of French from France as the supra-regional norm for all French-speaking countries and territories.
APPENDIX A: CLASSIFICATION OF THE WORDS USED TO DESCRIBE QUEBEC FRENCH
APPENDIX B: ORDER OF STIMULUS RECORDINGS IN THE MATCHED-GUISE EXPERIMENT
speaker 2 – EF, speaker 3 – English, speaker 1 – QF,
speaker 3 – EF, speaker 2 – QF, speaker 1 – English,
speaker 3 – QF, speaker 1 – EF, speaker 2 – English
APPENDIX C: TEXT EMPLOYED IN THE MATCHED-GUISE EXPERIMENT
English version
On a cold January day when the temperature was minus twenty degrees Celsius, a thirteen-year-old girl was buried under debris when the heating system of her parents’ home exploded. When firemen arrived on the scene, she was soaked with water and waited two hours before she could be rescued and taken to hospital. Her body temperature was well below thirty-three degrees Celsius and she was shivering violently. She had lost her sense of balance, spoke incoherently, and was suffering from extreme fatigue.
French version
Un froid matin de janvier, par une température de moins vingt degrés Celsius, une jeune fille de treize ans se trouva enfoui sous les débris produits par l'explosion du système de chauffage de ses parents. Lorsque les pompiers arrivèrent sur les lieux, elle était trempée jusqu'aux os et il a fallu deux heures à ses sauveteurs pour la sortir de là et la conduire à l'hôpital. La température de son corps était au-dessous des trente trois degrés Celsius; de plus, elle était secouée de violents frissons, avait perdu le sens de l’équilibre, parlait de façon incohérente et souffrait de fatigue extrême.
