1. INTRODUCTION
There are some 550,000 Francophones in Ontario,Footnote 1 the largest French-speaking population in Canada outside Quebec, but they represent only about 4% of the total population of the province (Statistics Canada, 2017). In most communities where French is spoken, Francophones represent a distinct minority. However, there are a few enclaves where French speakers are in the majority and where English, although present, is not the dominant language of social or commercial interactions. For example, in towns such as Hearst (Northern Ontario), Casselman (Southeastern Ontario) and Hawkesbury (on the Quebec-Ontario border), Francophones make up to 87%, 79% and 78% of their respective populations (Statistics Canada, 2017). French language and culture are therefore particularly vigorous in these communities. Nonetheless, as contact with English is also intense, the use of lexical and grammatical borrowings is a common phenomenon in spoken French in Ontario, as in the other varieties of French spoken outside Quebec.
The use of so, in alternation with (ça) fait que vs donc vs alors, is a typical example of grammatical borrowing and has been the subject of numerous studies.Footnote 2 The purpose of our research is to examine the internal and external factors that determine the use of these forms in the light of recent data extracted from the Casselman corpus, a majority Franco-Ontarian community. Furthermore, the original dimension of this research is that we also seek to statistically correlate the production of these connectors by Franco-Ontarian speakers with the ethnocultural and linguistic identities they claim.Footnote 3
Our approach is motivated by the fact that the issue of the identity of Ontario’s minority Francophones has been raised in a number of studies. According to Heller (Reference Heller1994: 156), language has always been the central element in the construction of a Franco-Ontarian identity and in the political motivation of Franco-Ontarians generally, especially since religion and the concept of ‘race’ began to lose their importance throughout French Canada during the 1960s. It should be remembered that, historically, all French-speaking communities outside the Maritimes considered themselves as being French-Canadian, one of the three co-founder groups of the country (Aboriginal, French, English).Footnote 4 As early as the 1960s, at the time of the Quiet Revolution and the beginning of Quebec’s independence movement, a distinct ‘Québécois’ identity, in opposition to the more general and historical ‘French-Canadian’ identity, began to surface. This inevitably forced French-speaking minority communities west of Quebec to redefine themselves in terms of their own historical experiences, traditions and even linguistic characteristics, and to look towards their own regional or provincial political and community structures, leading eventually to new identities such as Franco-Ontarian (or even Ontarois – on the ‘Québécois’ model), Franco-Manitoban, Fransaskois, etc. (Aunger, Reference Aunger and Thériault1999).
In this article, we briefly discuss the relationship between language and identity and review some of the relevant research on Franco-Ontarian identity. We then describe the community of Casselman, the corpus under study, and the methodological aspects of our research. Subsequently, we deal with the different identities claimed by the participants. We present the results of the analysis of the variation between (ça) fait que vs donc vs alors vs so, taking into account the internal, external and identity factors conditioning their use. Finally, we conclude by mentioning points of convergence between our results and those of previously published research.
2. ON THE RELATIONSHIP OF LANGUAGE AND IDENTITY
Omoniyi and White (Reference Omoniyi and White2006: 1) consider that: ‘Across the social and behavioral sciences in general there has been an increased interest in identity as a subject of inquiry […]. In particular, as a dimension of linguistic inquiry, identity has moved to the fore as a priority subject for investigation’. Zenker (Reference Zenker2018:1) adds that the relationship between language and identity has emerged as a transdisciplinary field of research that is concerned with how languages ‘shape and are shaped by diverse identities’. Block (Reference Block, Paltridge and Phakiti2013) also considers that speakers’ identities are indexed in how they draw on repertoires of linguistic resources, which include language choice, accent, lexical choice and morpho-syntax.
A number of studies dealing specifically with language and identity have been published in the past forty years (Omoniyi and White, Reference Omoniyi and White2006; Coupland, Reference Coupland2007; Riley, Reference Riley2007; Edwards, Reference Edwards2009; Eckert, Reference Eckert2012; Heller, Reference Heller2013; Preece, Reference Preece2016, among many others). For our immediate purpose, Omoniyi and White (Reference Omoniyi and White2006: 2) conveniently present the most common positions on language and identity:
-
Identity is not fixed;
-
Identity is constructed within established contexts and may vary from one context to another;
-
These contexts are moderated and defined by intervening social variables and expressed through language(s);
-
Identity is a salient factor in every communicative context, whether given prominence or not;
-
Identity informs social relationships and therefore also informs the communicative exchanges that characterize them;
-
More than one identity may be articulated in a given context – in which case there will be a dynamic of identities management.
According to Drummond and Schleef, identity has been a concern in variationist sociolinguistics (VS) since the early 1960s (Labov Reference Labov1963); however, they explain that the role played by identity in conditioning language variation and change, as well as how the notion of identity itself is defined, has changed throughout the decades and ‘continues to be at the very centre of contention in VS.’ (Drummond and Schleef, Reference Drummond, Schleef and Preece2016: 51)
Eckert (Reference Eckert2012) theorizes that there have been three waves of research on the relationship between language and identity in VS. The first wave (Labov, Reference Labov1966; Wolfram, Reference Wolfram1969; Trudgill, Reference Trudgill1974; Thibault and Sankoff, Reference Thibault and Sankoff1993; etc.) treated social variables as indexed directly to broad social categories such as age, gender, socioeconomic status and ethnicity. Identity per se is usually not treated nor even mentioned. If these broad macro-sociological labels are conceived as identity labels, they are mere reflections of language use and as such, identities are considered as stable, unified and ‘essential’, since they are based on membership of individuals in specific social categories (Drummond and Schleef, Reference Drummond, Schleef and Preece2016: 51).
Second-wave studies (Labov, Reference Labov1963; Milroy and Milroy, Reference Milroy, Milroy and Trudgill1978; Eckert, Reference Eckert2000; etc.) attribute social agency to the use of vernacular or standard features. The vernacular is seen as an expression of identity, particularly local, ethnic or class identity. In second-wave VS studies, researchers make use of ethnography and other qualitative methods to examine how certain linguistic forms are locally meaningful to social groups. Furthermore, these social categories are not pre-formulated frameworks of analysis (as in first-wave studies) but are based on participant observation: ‘Rather than imposing identity categories on speakers, ethnographic observation enables us to work with identity categories that emerge from the data and that we know are salient to the speakers themselves at the local level’ (Drummond and Schleef, Reference Drummond, Schleef and Preece2016: 52).
If linguistic features index social categories in waves one and two, in third-wave VS, linguistic features index social meanings. According to Drummond and Schleef (Reference Drummond, Schleef and Preece2016: 55) social meanings are exemplified by stances (mental or emotional positions adopted with respect to something), personal characteristics, personae (the role(s) that one assumes or displays in public or society; one’s public image or personality, as distinguished from the inner self) and social types. Third-wave research views language use as not reflecting identities per se but constituting them through linguistic practice, thus putting an emphasis on the social meaning of variable features. In third-wave VS, in order to study and understand identity, one cannot simply focus on one particular linguistic feature; rather, one must focus on something larger, namely style (Podesva, Reference Podesva2011). For others, the focus will be on stance (Bucholtz and Hall, Reference Bucholtz, Hall, Llamas and Watt2010).
Even though our study is firmly embedded in the first-wave VS tradition, it differs from earlier research in that it specifically bears on ethnocultural and linguistic identities as categories potentially influencing language practice.
3. ON FRANCO-ONTARIAN IDENTITY MARKERS
Since the 1970s, much has been published on Franco-Ontarian identity.Footnote 5 However, little or nothing has been published on identity in Ontarian French-majority communities. Furthermore, the only published research papers which deal specifically with Franco-Ontarian identity markers as such, at least that we are aware of, are those of Boissonneault (Reference Boissonneault1996) and Dallaire (Reference Dallaire2004).
Boissonneault (Reference Boissonneault1996) surveyed by written questionnaire 137 collegiate and 37 university-level students aged between 18 and 24 years, studying in French-language institutions in Northern Ontario. She was interested in obtaining data pertaining to their linguistic behavior relative to their linguistic and sociocultural attitudes. Among other questions, respondents were asked the following: 1) Do you consider yourself to be Francophone, Anglophone or Bilingual?Footnote 6 2) Do you consider yourself to be Franco-Ontarian, French-Canadian, Canadian, Anglo-Canadian or something else? For both questions, participants were asked to explain their choices.
Boissonneault (Reference Boissonneault1996) considers the first question to reveal indicators of a ‘linguistic’ nature and the second question to reveal indicators of what she calls a ‘structuro-cultural’ nature. Boissonneault’s structuro-cultural identity is based on work by Juteau-Lee and Lapointe (Reference Juteau-Lee, Lapointe and Elliot1983). It refers to the fact that, on the one hand, identity may be based on ‘cultural’ concepts such as language, religion, life-style and culture, and on the other hand, a Franco-Ontarian identity may be largely based on ‘structural’ criteria where underlying group boundaries are emphasized. These structural criteria outline different institutional spheres and refer to specific local, territorial or provincial realities.
Results of Boissonneault’s research were as follows: as many as 74% of the students consider themselves to be Bilingual while only 23% to be Francophone. Boissonneault (Reference Boissonneault1996: 184) explains this result by the fact that, outside Quebec, in minority communities, bilingualism is the most important identity marker. But we hasten to note that for 83% of the participants, bilingualism is defined as the mastery and frequent use of both languages, thus constituting a ‘linguistic’ identity rather than a strictly structuro-cultural one; only 17% of the respondents underscore other values of bilingualism, such as economic and professional advantages, equality between the two languages and cultures, etc. (op.cit.: 185). An aggregate of more than 90% of the students identify as being Franco-Ontarian, French-Canadian or simply Canadian (op.cit.: 184). Those claiming a Franco-Ontarian identity justify their choice either according to linguistic criteria (e.g. having French as mother tongue) or according to structuro-cultural criteria (e.g. being born in Ontario) (op.cit.: 186). Those identifying as French-Canadian underscore their belonging to the Canadian nation (op.cit.: 186); those identifying as Canadian also refer to their belonging to the Canadian nation, but, in this case, the criterion of language is devoid of all cultural values (op.cit.: 187). By cross-indexing linguistic and structuro-cultural indicators, Boissonneault (op.cit.: 188) reports that 67.8% of participants consider themselves mainly to be bilingual.
Dallaire’s (Reference Dallaire2004) study bears on a large group of teenagers, interviewed during the 2001 edition of the Jeux franco-ontariens (Franco-Ontarian Games).Footnote 7 Using data obtained through participant observation, oral interviews, drawings and answers to written questionnaires from a selected group of volunteers, Dallaire compiled results allowing her to establish an inventory of the main identities claimed by the participants. She was interested in knowing how participants conceive their Francophone identity and how they describe the Francophone community in Ontario.
Dallaire (Reference Dallaire2004) considers that there are two types of narratives which contribute to the contemporary construction of Francophone identities and communities in Canada: a linguistic narrative and a cultural one. She obtained 587 completed questionnaires, representing 77.3% of the 759 participants. They were asked a number of questions related to their identity.Footnote 8 The results are as follows:
-
37.6% of participants identify themselves as being Franco-Ontarian;
-
19% identify themselves as being bilingual;
-
18.3% identify themselves as being Canadian;
-
10.5% identify themselves as being French-Canadian;
-
6.6% identify themselves as being Francophone;
-
3% identify themselves as being bilingual Canadian;
-
5.1% opt for other identities.
Dallaire (Reference Dallaire2004: 142) points out that Franco-Ontarian, French-Canadian and Francophone identities, all three of which refer to ‘being a speaker of French’, constitute 54.7% of the answers. However, the percentage of participants who identify themselves as having an ‘other’ ethnic background (24.5%) is vastly greater than the percentage of participants who opt for an ‘other’ identity (5.1%). This implies that a great number of French-speaking immigrant youths in Ontario have abandoned their original ethnic identity and have adopted new ones, more closely related to an Ontarian or Canadian reality.
Dallaire’s study also allows us to better understand some of the participants’ identities. The reasons motivating their choices are as follows:
-
For those claiming a Franco-Ontarian identity: belonging to the province of Ontario (29.2%); using French spontaneously and/or routinely (21.6%); pride in one’s local community (14.4%);
-
For those claiming a French-Canadian identity: belonging to the Canadian nation (31.3%); using French spontaneously and/or routinely (18.1%); pride in the French-Canadian community (14.4%);
-
For those claiming a Francophone identity: using French spontaneously and/or routinely (25.7%); other reasons (unspecified) (17.2%); pride in one’s local community (11.4%);
-
For those claiming a bilingual identity: active French-English bilingualism (70.1%); other reasons (unspecified) (9.4%); belonging to the Canadian nation (4.3%);
-
For those claiming a Canadian identity: belonging to the Canadian nation (38.7%); active French-English bilingualism (16.9%); other reasons (unspecified) (11.3%).
Dallaire (Reference Dallaire2004: 138) concludes that even in a large group of teenagers mostly favouring a Francophone identity, there does not seem to be a consensus as to the terminology chosen to express this identity.
4. RESEARCH QUESTIONS
Located in the United Counties of Prescott and Russell, Ontario, 56 kilometers east of Ottawa and 145 kilometers west of Montreal, the village of Casselman is a majority French-speaking community. Indeed, its population is made up of 79% of French mother-tongue speakers. As we pointed out earlier, the use of (ça) fait que vs donc vs alors vs so has already been examined in Ontario and Quebec French. Given the specific linguistic situation of this small community, it seems legitimate to wonder whether the linguistic practices of speakers observed in our corpus reflect those revealed in previous studies.
The relevance of our study is twofold. We report data from a recent corpus of unrestricted speakers belonging to a Franco-Ontarian community that is not only majoritarian, but which is also located geographically close to Quebec and is therefore likely to be influenced by the linguistic practices of both Franco-Ontarians as well as Quebecois.
Moreover, we will attempt to answer the following questions: 1) Does the use of (ça) fait que vs donc vs alors vs so by speakers of the Casselman community differ from that of Quebec and Ontario speakers examined in previous studies? 2) Can we correlate the use of these markers to the ethnocultural and linguistic identities claimed by the interviewees of our corpus?
5. METHODOLOGY
First, we present our database. We then deal with the dependent and independent variables taken into account in our analyses.
5.1 The Casselman corpus
The foundation of the town of Casselman dates back to 1830, when Martin Casselman, a descendant of a Loyalist family, moved to the United Counties of Prescott and Russel, in eastern Ontario, to develop the lumber trade. Although the first settlers recruited by the contractor were of English or Scottish origins, Francophones arriving from neighbouring Quebec rapidly formed the bulk of the population (Brault, Reference Brault1965).
Since then, Casselman has steadily grown and the town now numbers 3,548 inhabitants, Francophones being in the majority. Since 2006, the percentage of Francophones has tended to decrease slightly. In ten years, mother-tongue French speakers have dropped from 83% of the population to 79% (Statistics Canada, 2007; Statistics Canada, 2017). During the same time period, the number of Anglophones has increased from 15% to 17%. Nevertheless, the community of Casselman has the second highest proportion of mother-tongue French-speakers in the province and French remains the dominant language in homes, schools, commerce and public services.
Our corpus consists of 62 informal semi-directed one-hour long at-home interviews with French-speaking interviewees, conducted between 2009 and 2010 by a young woman native of Casselman and known by most of the participants. Each participant was recruited by the ‘snowball’ technique (Biernacki and Waldorf, Reference Biernacki and Waldorf1981). This technique ensures a certain proximity between the interviewer and the interviewee. In recruiting participants, the interviewer had to make certain that gender, age and socio-economic status were more or less evenly distributed. Although speakers were aware of being recorded, their degree of self-monitoring was minimized as much as possible. All interviews were recorded with a Sony ICD-UX70 MP3 device, then transcribed in Word format according to a common protocol allowing easy data analysis through CasualConc, a free concorder software for MacOs.
Social data were collected orally at the beginning of each interview. All the adults were interviewed using the same questionnaire/guide. The questionnaire/guide was somewhat adapted for participants aged 20 years or less, but the questions nevertheless dealt with the same themes as those of the adult version. These dealt with personal data and family background, ethnocultural and linguistic identities claimed, educational background, including language of instruction, work experience, relationship with the province of Quebec and the rest of Canada, rate of consumption of French- or English-language media, level of language competence, degree of bilingualism, etc. The social class attributed to each speaker was determined from Blishen et al.’s (Reference Blishen, Carroll and Moore1987) classification, which is based on the speaker’s employment, or that of his/her parents, for the younger generation. Table 1 presents the distribution of speakers according to the social factors considered in our study.
Table 1. General distribution of the Casselman corpus speakers
Finally, while 75% of adolescents in the corpus are born in Casselman, native-born Casselman adults represent only 53%. Nonetheless, all of the adult speakers born elsewhere have spent most of their lives in Casselman and consider themselves as full members of the community.
5.2 The dependent and independent variables
Our dependent variables are the connectorsFootnote 10 and discourse markers of consequence (ça) fait que vs donc vs alors vs so. We follow most authors reviewed in Blondeau et al. (Reference Blondeau, Mougeon and Tremblay2019) in distinguishing between their grammatical connecting function, that of joining together clauses, phrases or words, and their discursive function, that of managing the flow and structure of discourse, by showing turns, joining ideas together, showing attitude and generally controlling communication.Footnote 11 As in Blondeau et al. (Reference Blondeau, Mougeon and Tremblay2019), we only take into account the occurrences of each of the variants expressing consequence. For example we eliminated occurrences such as ‘C’est donc ben beau…’ or ‘Ah ça alors!’ and so occurring in code-switches as in ‘I’m so sorry for you’. Finally, the four markers can appear with ‘là’ as in ‘(ça) fait que là…’, ‘donc là…’, ‘alors là…’ and ‘so là…’, however, the number of occurrences of these forms is quite limited (1 for ‘so là’, 1 for ‘alors là’, 30 for ‘donc là’ and 185 for ‘(ça) fait que là’ out of a total of 3 509 occurrences). We therefore do not discriminate forms with or without the adverbial particle.
As do Blondeau et al. (Reference Blondeau, Mougeon and Tremblay2019: 47), we invoke the notion of functional equivalence and while a strict semantic equivalence between the four forms cannot be totally presupposed, we postulate that they possess sufficient common features to fulfill the same linguistic functions.Footnote 12
The speakers are grouped according to the following external factors:
-
Social class: working class vs middle-middle class vs upper-middle class.
-
Gender: male vs female.
-
Age: participants aged under 20 vs aged 20 to 25 vs aged 26 to 54 vs aged 55 and over.
-
Language dominance: French-dominant bilingual vs Balanced bilingual vs English-dominant bilingual.
-
Self-declared ethnocultural identity: Franco-Ontarian vs French-Canadian vs Canadian vs Ontarian.
-
Self-declared linguistic identity: Francophone vs Bilingual vs Anglophone.
The last two independent variables were determined by two questions explicitly addressing ethnocultural and linguistic identities, as in the following examples:
-
(1) Interviewer: Et puis, comment est-ce que tu te définis? Est-ce que tu te définis comme une Nord-Américaine, une Canadienne, une Ontarienne, une Canadienne-française, une Franco-Ontarienne ou même une Québécoise… ?Footnote 13
Speaker 11: Ah mon dieu! Je suis pas Québécoise. Je suis Canadienne.
-
(2) Interviewer: Toi, est-ce que tu te considères comme…quelqu’un de bilingue?
Speaker 28: Oui. Je suis… disons je suis ben bilingue; comment dire… disons, je suis… un des plus bilingues dans la famille.
The above categories are relevant to our analysis because they are those claimed by Francophone minority communities (Boissonneault, Reference Boissonneault1996; Dallaire, Reference Dallaire2004).
We are well aware that the ethnocultural identities taken into account here appear relatively static. Indeed, ethnocultural identities may vary according to a given interlocutor. For example, someone could claim to be Franco-Ontarian in the presence of other Franco-Ontarians but claim to be French-Canadian in front of Anglo-Canadians, or even to be simply Canadian in the presence of Quebecois. However, we believe that these identities may well reveal how speakers represent themselves, at least, in the context of the interviews on which our corpus is based. All of the identities claimed were spontaneously given in the presence of the same interviewer, thus eliminating the problem noted above. These avowed identities may therefore be relevant to verify whether this type of factor potentially conditions the use of a particular variant. Note that only two interviewees hesitated in answering the ethnocultural identity question. In these two cases, we took into account the claimed identity declared most often during the interview. Also, notice that our ‘ethnocultural identity’ label is more or less equivalent to Boissonneault’s (Reference Boissonneault1996) ‘structuro-cultural’ identity and to Dallaire’s (Reference Dallaire2004) ‘cultural’ narrative.
As shown in Section 3, claiming a ‘bilingual’ identity can theoretically be interpreted as an ethnocultural identity. However, in all cases, participants who claim to be bilingual refer only to their proficiency in both French and English and never to an ethnocultural identity. For this reason, we consider bilingual competence to constitute a linguistic identity (what languages one knows and uses) and not as an ethnocultural identity (who one is or to what ethnocultural community ones belongs), as for Boissonneault (Reference Boissonneault1996).
As in Blondeau et al. (Reference Blondeau, Mougeon and Tremblay2019), we also measure the impact of the grammatical function of the variants: connector vs. discourse marker, as in the following examples:
Connector function:
-
(3) Speaker 25: […] je suis pus dans la région à cause, justement, à cause de mes études, ça fait que c’est dur de savoir.
-
(4) Speaker 6: L’hôpital est resté ouvert, donc je pense que ça, ça démontre que les Franco-Ontariens peuvent avoir une voix en Ontario.
-
(5) Speaker 9: Il était une autorité dans le domaine pis je voulais avoir M. Paccioco, alors j’étais prêt à prendre le cours en anglais, ça me dérangeait pas du tout, du tout.
-
(6) Speaker 12: Carry, elle a rencontré, comme, un de ses anciens chums, là-bas, so elle a cheaté sur son mari.
Discourse marker function:Footnote 14
-
(7) Speaker 19: Des fois les parents sont pas là; mon frère, lui, il est à l’université; ça fait que… Avant, des fois, il venait me chercher à l’école.
-
(8) Speaker 13: J’ai de la parenté que…sont anglophones donc…Footnote 15
-
(9) Speaker 30: Ben, là, je commence à douter mes connaissances, parce que tu me demandes des choses ça me vient pas alors…
-
(10) Speaker 5: Je veux toujours aller en Afrique, comme juste aider pis, tout ça, so…
Our analysis consists of three parts. Through a multivariate analysis using the Goldvarb Yosemite freeware,Footnote 16 we first measure the impact of internal factors (associated with the grammar), then we measure the external or social constraints (associated with social meaning),Footnote 17 and finally we measure the impact of identity factors.Footnote 18 In our opinion, identity factors are neither internal (Labov, Reference Labov1994) nor external (Labov, Reference Labov2001). Rather, they are cognitive factors (Labov, Reference Labov2010). We believe what languages speakers consider they know and use (their linguistic identity) and to what national, regional or local ethnocultural group they feel they belong (their ethnocultural identity) is revealed by certain linguistic forms they tend to favour, which does not mean that the above identities are necessarily fixed. We thus formulate the hypothesis that the specific use of (ça) fait que, donc, alors or so may be apprehended as an act of identity (Labov, Reference Labov2010: 193).
6. RESULTS AND ANALYSES
We begin by observing the identities claimed by the participants of the Casselman corpus. Subsequently, we present the results of the multivariate analysis we used to determine the internal, external and identity factors that condition the use of (ça) fait que vs donc vs alors vs so.
6.1 Identities claimed by the Casselman participants
The results regarding the linguistic identities are summarized in Table 2.
Table 2. Linguistic identities of the Casselman participantsFootnote 19
Thirty-three out of 42 participants (79%) declare themselves to be bilingual, and nine out of 42 (21%) declare themselves to be Francophone. As expected, none claim to be Anglophone. This result contrasts sharply with the percentages of interviewees we obtained for language dominance. Indeed, the distribution of speakers according to their language dominance shows that of 59 speakers,Footnote 20 37 (62.7%) are French-dominant bilinguals, 20 speakers (33.9%) are balanced bilinguals and two interviewees (3.4%) are English-dominant bilinguals. On the one hand, this may show that for some speakers, the frequency and preference of use of French does not necessarily determine their linguistic identity and that this identity reflects more what they think they are than what language they actually speak in their everyday life. On the other hand, it is possible that terms such as ‘Francophone’, ‘Anglophone’ and ‘bilingual’ may be interpreted differently by individual speakers. For example, many speakers believe that to be bilingual one must master both languages equally well, or to be ‘Francophone’, one must speak only French, etc.
Finally, the distribution of linguistic identities by age group is relatively balanced. Four out of nine Francophones are under the age of 25, and five are aged 25 and over. Regarding the bilingual speakers, 16 out of 33 speakers are under the age of 25, compared to 17 aged 25 and over.
We now consider the ethnocultural identities, as well as the justifications that are provided. Results are summarized in Table 3.
Table 3. Ethnocultural identities and their justifications of speakers of the Casselman corpus
Five ethnocultural identities are declared: 59% of interviewees consider themselves to be Franco-Ontarian, 21% define themselves as being Canadian, 14% of the speakers identify as being French-Canadian. One female participant sees herself as being an Ontarian-Canadian, and one young male claims to be simply an Ontarian.Footnote 22
In Boissonneault (Reference Boissonneault1996) and Dallaire (Reference Dallaire2004), the main ethnocultural identities identified are also Franco-Ontarian, Canadian and French-Canadian, in that order. In Boissonneault (Reference Boissonneault1996), these categories constitute 90% of the responses vs. 94% for the Casselman corpus vs. 66.4 % for Dallaire (Reference Dallaire2004). In our opinion, the difference with Dallaire (Reference Dallaire2004) can be explained by the fact that in her study, one of the choices of ethnocultural identity was being Bilingual (19% of the subjects). In our study, no participant claimed to be Bilingual as an ethnocultural identity, even if they could have chosen it (see note 13).
The reasons motivating the identities professed by the speakers in Casselman also resemble quite closely those given in Dallaire (Reference Dallaire2004). For example, in both cases, being born in Ontario, using French on a daily basis and pride in one’s community were the most frequent reasons invoked by those professing to be Franco-Ontarian.
Not only does a Franco-Ontarian identity rank highest among the Casselman subjects, it is also the one that receives the highest number of specific justifications or markers. Indeed, we distinguish seven identity characteristics, in order of importance (frequency): 1) Ability to speak French, 2) Being a native of Ontario, 3) Having a distinct identity/culture, 4) Being bilingual, 5) Not being Quebecois, 6) Pride of being Franco-Ontarian and 7) Being a native of Canada. With regards to the Canadian identity, it includes the following aspects: 1) Residing in Canada, 2) Not being Quebecois and 3) Being bilingual. A French-Canadian identity boils down to the following aspects: 1) Ability to speak French, 2) Residing in Canada and 3) Having French-Canadian origins. The female speaker selecting the Ontarian-Canadian identity focuses on three elements: 1) Ability to speak French, 2) Residing in Ontario and 3) Not being a Quebecois. Finally, the participant who claims to be an Ontarian explains his identity by the fact of living in Ontario and not being a Quebecois.
The ability to speak French is clearly the central element of both Franco-Ontarian and French-Canadian identities. It should also be noted that these identities include marks of cultural identification (such as the community origins or the identification of a culture distinct from that of other communities), linguistic identification (such as having French as a mother tongue or the criterion of bilingualism), but also geographical identification (living in Ontario and/or in Canada).
Franco-Ontarian, Canadian, Ontarian-Canadian and Ontarian identity characteristics include a negative one: not being Quebecois. Much as many Anglo-Canadians often identify themselves as not being American (Resnick, Reference Resnick2012), many Francophones outside Quebec and the Maritimes tend to identify themselves as not being Quebecois. The reasons for this are numerous and diverse, an important one being that most majority-English provincial governments have for a long time ignored the existence of French-speaking communities within their borders and it has been part of the Canadian myth that French speakers in Canada largely come from Quebec, and to a lesser extent, from the Maritimes. Therefore, a number of French-speakers outside Quebec still feel the need to specify that while they speak French, they are not Quebecois.
The development of a new Franco-Ontarian identity has taken quite some time and many older speakers still tend to consider themselves as French-Canadian rather than Franco-Ontarian. Indeed, 50% of Casselman speakers 55 years old or older identify as French-Canadian while 60% of speakers 25 years old or younger identify as Franco-Ontarian.
6.2 Factors conditioning the use of (ça) fait que vs donc vs alors vs so
We first attempt to measure the impact of the function of the variants: connector (3,426 occurrences) vs. discourse marker (83 occurrences). The total number of occurrences observed is 3,509. Results are presented in Table 4.
Table 4. Internal factors conditioning (ça) fait que/donc/alors/so
In the Casselman corpus, internal variation is significant for three out of the four variables. On the one hand, (ça) fait que (eff. = 0.51) and donc (eff. = 0.51) are slightly more frequent to express the grammatical function of consequence; on the other hand, so (eff. = 0.84) is used more often as a discourse marker than as a connector of consequence. The results for alors are not significant since it is almost systematically used as a connector.
These results diverge somewhat from two previous studies. Dessureault-Dober (Reference Dessureault-Dober1974) shows that the connectors (ça) fait que, donc and alors mainly express a grammatical function in Montreal French.Footnote 23 In Moncton French, Roy (Reference Roy1979) demonstrates that the choice of (ça) fait que and so is not readily determined by linguistic function. However, Blondeau et al. (Reference Blondeau, Mougeon and Tremblay2019:50) show that in the small French-speaking minority community of Welland (Ontario), the forms (ça) fait que and alors are more often used as connectors, while so is more frequent as a discourse marker, much as is the case in Casselman. They also note that in Montreal French, the language of the majority, (ça) fait que is more often used as a discourse marker than as a connector. In other words, in terms of internal variation, our results converge with those recently collected in Welland, but diverge from those of Montreal. Blondeau et al. (Reference Blondeau, Mougeon and Tremblay2019: 62) argue that in the Welland corpus: ‘the discursive use of so may have been a gateway for this connector. In turn, there would have been a strengthening of the association of the two main competitors of so with the grammatical function’ (our translation).
Our second analysis concerns the external or social factors conditioning the variation between the four connectors.
These results provide us with the following information: (ça) fait que is the form most often used, as it accounts for nearly half of the total occurrences (47.2%) and has a dispersion rate of 97%. This converges with Martineau and Séguin (Reference Martineau and Séguin2016), who have argued that this variant is typical of most Francophone communities in Canada. However, the results of Blondeau et al. (Reference Blondeau, Mougeon and Tremblay2019: 48) show that if in Montreal, (ça) fait que is indeed the most frequent variant, in Welland, this variant is only the second most frequent, after alors (40%). It must be remembered that in Casselman nearly 80% of the community are French-speakers and that it is geographically close to Quebec, favouring linguistic contacts with Quebecois and this may directly influence Casselman speakers’ speech. It is therefore not surprising that (ça) fait que is the most common variant in Casselman, as it is in Montreal.
Regarding donc and alors, the analysis reveals a number of interesting facts. First, both variants have general results very close to those for Montreal: donc represents 21.3% of the occurrences of the Casselman corpus and 19% of Montreal corpus, and alors counts for 4.9% of the occurrences in Casselman and 5% in Montreal (Blondeau et al. Reference Blondeau, Mougeon and Tremblay2019: 48). In the Casselman corpus, both forms are rejected by the working class (eff. = 0.13 for donc and eff. = 0.06 for alors) and both are favoured by women (eff. = 0.57 for donc and eff. = 0.70 for alors), which confirms their normative value. In Welland and Montreal, both forms are also rejected by the working-class. However, while alors is favoured by women in Welland (eff. = 0.61), the difference between men and women is non-significant in Montreal (NS). Moreover, while in Casselman, men tend to reject donc (eff. = 0.44), this connector is favoured by men in Welland (eff. = 0.73) and Montreal (eff. = 0.65) (Blondeau et al., Reference Blondeau, Mougeon and Tremblay2019: 51–53). Blondeau et al. (Reference Blondeau, Mougeon and Tremblay2019: 61) argue that in Montreal French (ça) fait que may have lost its covert prestige, since women now use it more often than men. In counterpart, men would now seem to prefer donc. In the case of Casselman, the social reconfiguration of (ça) fait que has not yet taken place.Footnote 26 We know that according to Labov’s Principle 2: ‘For stable sociolinguistic variables, women show a lower rate of stigmatized variants and a higher rate of prestige variants than men’ (Labov Reference Labov2001: 266). This may explain why, in Casselman, donc and alors are still favoured by women.
In our Casselman corpus, alors is much more characteristic of speakers aged 26 and above (eff. = 0.98 for speakers aged 26 to 55 and eff. = 0.92 for speakers aged 55 and above) than donc. In fact, alors is almost absent in the discourse of speakers 25 and younger (only eight occurrences produced by the two youngest generations).
It is quite possible that we are witnessing a change in progress, where donc (21.3%) is gradually replacing alors (4.9%). Notice that this also converges with the results of Blondeau et al. (Reference Blondeau, Mougeon and Tremblay2019: 48), which show that, in Montreal French, alors (representing only 5% of the total occurrences) is also becoming marginal.
Furthermore, we note that all social factors are significant. First, language dominance is the most important factor (rank = 1). Indeed, (ça) fait que decreases as the use of so increases, as in Blondeau et al. (Reference Blondeau, Mougeon and Tremblay2019: 55) for Welland, but not for Montreal, where so is not attested. In fact, (ça) fait que is almost absent among English-dominant bilinguals (eff. = 0.01) and is less frequently used by balanced bilinguals (eff. = 0.47) than by French-dominant speakers (eff. = 0.61). The effect of age is also very salient (rank = 2), and we find the opposite phenomenon for the variant so. Speakers aged 26 and over regularly use (ça) fait que (eff. = 0.76 for the 26 to 54 age group, eff. = 0.68 for those 55 and over), as opposed to the youngest (eff = 0.39 for 20–25 years and eff. = 0.40 for those under 20). This result supports the hypothesis of a change in progress, where the French vernacular variant seems to be gradually replaced by borrowed so. Lastly, as in most previous studies,Footnote 27 the social value of (ça) fait que remains strong, since this form is representative of the working class (eff. = 0.70) and of men (eff. = 0.60). Its status as the typically vernacular variant seems therefore constant across Laurentian French varieties.
Regarding so, several previous studies have shown that this variant is characteristic of French minority communities in general,Footnote 28 of working/middle classes,Footnote 29 of balanced bilinguals or English-dominant speakersFootnote 30 and of the youngest speakers of Welland.Footnote 31 Our results show similar trends. First, the global percentage of use of so (26.6%) is almost identical to that of Welland (26%) (Blondeau et al., Reference Blondeau, Mougeon and Tremblay2019: 48). So is rejected by the upper-middle class (eff. = 0.28); English-dominant and balanced bilinguals are those who use it most often (eff. = 0.99 and eff. = 0.68), as expected. Finally, Table 5 indicates that speakers aged 26 and over disfavour the English variant (eff. = 0.05), unlike the younger age groups (eff. = 0.91 for those aged under 20 and eff. = 0.73 for those aged 20–25.). The age constraint is the most important one in conditioning the use of so (rank = 1 (0.86)), and it is followed by language dominance (rank = 2 (0.66)) and social class (rank = 3 (0.37)). This result reinforces the hypothesis of a change in progress in favour of the English variant, hypothesis that Blondeau et al. (Reference Blondeau, Mougeon and Tremblay2019: 59) have recently formulated: ‘Without losing its non-standard character, so has gained in value through the latent prestige it holds within the Franco-Ontarian community, which tends to promote bilingualism’ (our translation).
Table 5. Social factors conditioning (ça) fait que/donc/alors/so
Table 6. Identity factors conditioning (ça) fait que/donc/alors/so Footnote 32
Our final analysis addresses the identity factors conditioning the use of (ça) fait que vs donc vs alors vs so.
The results presented in Table 6 reveal a complex situation. Regarding (ça) fait que, we observe the following: speakers identifying as French-Canadian (eff. = 0.92) use it more than those identifying as Franco-Ontarian (eff. = 0.40) or Canadian (eff. = 0.35). On the other hand, it is slightly favoured by Bilinguals (eff =0.53), but not by Francophone speakers (eff =0.40). We believe the reason behind this distribution lies in the fact that (ça) fait que is the prototypical French-Canadian vernacular variant and therefore characteristic of the speakers claiming a traditional French-Canadian ethnocultural identity. Furthermore, the weight of ethnocultural identity favouring (ça) fait que (rank = 1 (0.52)) is much higher than that of linguistic identity (rank = 2 (0.13)). In other words, ethnocultural identity seems to influence the choice of the variant more than linguistic identity.
The case of the normative variants donc and alors is equally complex. Both forms are favoured by participants identifying as Franco-Ontarians or as Canadians, two ethnocultural characteristics. However, donc and alors are clearly distinguished when linguistic identities are taken into account: donc is characteristic of Bilinguals (eff. = 0.63), alors of Francophones (eff = 0.93). In the case of donc, linguistic identity is the most important factor (rank = 1 (0.49)), while ethnocultural identity is more powerful for alors (rank = 1 (0.69)). How can this be explained? Thanks to a Crosstab analysisFootnote 33 of both social and identity factors, the results show that the majority of occurrences of donc are produced by balanced bilinguals belonging to the two higher social classes (53.7% of utterances). In the case of alors, the Crosstab analysis also shows that 57% of the utterances are produced by French-dominant speakers, mostly aged 26 to 54 years. In our opinion, these results seem to show that social class and age outweigh identity factors.
As for so, it is characteristic of speakers who mainly identify as Franco-Ontarian (eff. = 0.70) and is also significant for those identifying as Canadian (eff. = 0.57). On the other hand, it is rejected by speakers identifying as French-Canadian (eff. = KO). Table 3 indicates that speakers identifying either as Franco-Ontarian or as Canadian were the only ones justifying their identity by referring to the fact of being bilingual, as opposed to the other identities, where bilingualism is absent in the list of justifications. These results thus seem logical to us.
In terms of linguistic identity, it is surprising to note that Francophone speakers make greater use of so (eff. = 0.71) than Bilinguals (eff = 0.44), where the opposite would be expected, since balanced bilinguals are equally at ease in both languages. However, it must be remembered that the majority of Casselman speakers consider themselves to be Franco-Ontarians, an ethnocultural identity whose major characteristic is the primary use of the French language. Moreover, the weight of linguistic identity (rank = 2 (0.27)) is much lower than that of ethnocultural identity (rank = 1(0.70)). The English variant so therefore appears to have become a shibboleth of Franco-Ontarian identity.
7. DISCUSSION
Our objectives were to identify the ethnocultural and linguistic identities of the Casselman speakers, and to determine the internal and external factors that condition their use of (ça) fait que vs donc vs alors vs so. We also sought to determine if the identities claimed by our participants could potentially explain the use of forms among (ça) fait que vs donc vs alors vs so.
First, we showed that a Franco-Ontarian identity is predominant for Casselman speakers, and that it is followed by Canadian and French-Canadian identities, as in Boissonneault (Reference Boissonneault1996) and Dallaire (Reference Dallaire2004). Despite the use of a different methodology, we nevertheless found markers of structural and cultural identification similar to those of Boissonneault (Reference Boissonneault1996) and Dallaire (Reference Dallaire2004), which confirms the unifying character of ethnocultural identities.
Our study has also highlighted several elements converging with earlier papers, and particularly with the recent research of Blondeau et al. (Reference Blondeau, Mougeon and Tremblay2019). Indeed, we note that the linguistic practices of Casselman speakers share similarities with both those of Montreal and of Welland, Ontario. Our general results concerning donc (21.3% of occurrences) and alors (4.9%) are particularly close to those of Montreal (donc =19% and alors = 5%). Regarding so, its percentage of occurrences among the four markers in Casselman (26.6%) is nearly identical to the one for Welland (26%). Moreover, the percentage of occurrences of (ça) fait que for Casselman (47.2%) is almost exactly half-way between that of Montreal (76%) and that of Welland (27%).
The particular linguistic practices of Casselman speakers, partially reflecting both majority Franco-Montreal and minority Franco-Welland speakers, is also highlighted when we take into account internal and external variation. Our research emphasizes the singularity of communities such as Casselman, whose linguistic practices and norms straddle those of minority- and majority-French language communities. If, in the case of communities such as Welland and Montreal, the minority/majority dichotomy is demographically and linguistically quite evident, the case of Casselman demonstrates that such a dichotomy needs to be nuanced, according to the social and identity dynamics that characterize specific communities. The French spoken in Casselman reflects the fact that it is a majority-French language community but situated in a majority-English language province, while Montreal is a majority-French language city in a majority-French language province and Welland is a minority-French language community in a majority-English language province.
Moreover, as we have seen, the data extracted from the Casselman corpus (although incomplete for some speakers) demonstrate the importance of considering ethnocultural identities as a complementary factor conditioning linguistic variation. While it is difficult to measure the degree of awareness that speakers have of their ethnocultural identities when speaking, given the high degree of significance for (ça) fait (que) and a French-Canadian identity (eff. = 0.92), and for so and a Franco-Ontarian identity (eff. = 0.70), it seems these forms play a role as linguistic markers confirming the membership of a specific community.
However, although our results show that there is indeed a statistical link between linguistic variation and ethnocultural identities, for the moment, our methodology does not allow us to account for conflicting choices as in (11):
-
(11) Interviewer: Pis, toi, tu t’identifies comme quoi…comme personne ? Tu t’identifies-tu comme étant un…mettons, un Franco-ontarien, un Canadien, un Canadien-français… ? Comment tu t’identifies ? Speaker 26: Moi, un Canadien-français. Je suis fier d’être Franco-ontarien mais je suis plus… je me battrais plus pour fran/un Canadien-français.
Speaker 26 displays both a French-Canadian and a Franco-Ontarian identity, with a strong tendency to favour the former. Such an example is an illustration of Omoniyi and White (Reference Omoniyi and White2006: 2), who point out that identity is constructed within established contexts, that it may vary from one context to another and that more than one identity may be articulated in a given context. It also shows that there is still much work to be done to tease out the various possibilities. Indeed, we think it will be necessary to develop a scale to more precisely measure the management dynamics of identities claimed by speakers, in order to better exploit the weight of (socio)linguistic and ethnocultural identities as cognitive factors that can potentially explain linguistic variation. Furthermore, because identities are not fixed and may change over a lifespan, it would be quite interesting to take into account ethnocultural identities in real-time sociolinguistic studies.
It is important to note that much of the research on Franco-Ontarian identities deals with an increasingly important ‘Bilingual’ ethnocultural identity among young Franco-Ontarians (Heller, 1994; Reference Heller2013), where they live in Franco-Canadian, Anglo-Canadian and, in fact, Anglo-American cultures. This is mostly true for the majority of Franco-Ontarians who live in decidedly small French-speaking minority communities. Because Casselman is a heavy majority French-speaking community, this may explain why such a bilingual ethnocultural identity has not been revealed in our corpus, since all of the participants who declare themselves as ‘bilingual’ only refer to their linguistic abilities and not to their ethnocultural adhesion.
As a final point, we fully realize that first attempts often raise methodological questions, particularly when using corpora that are not specifically designed to broach sensitive and problematic subjects such as identities, as in the case of our Casselman corpus. Nonetheless, we hope we have opened a new gate for forthcoming variationist analyses.
