1. INTRODUCTION
The purpose of this article is to consider the main contributions of Mougeon et al.’s study and to examine the implications of their findings for future research on Canadian French. We first give a brief overview of results obtained, then present a critical analysis of their findings. In our comments, we consider alternative methodological approaches with a view to reflect on how such alternatives would enhance Mougeon et al.’s results and those of future research.
2. SUMMARY OF RESULTS
Mougeon et al.’s article presents a sociolinguistic analysis of restrictive adverbs in four historically-related Laurentian French speech communities in Canada: a) Bonnyville, Alberta; b) Welland, Ontario; c) Saint-Boniface, Manitoba; and d) Saint-Laurent, Manitoba. It undeniably constitutes a novel and an ambitious undertaking that should pave the way for future comparative research on Canadian French. Before considering the strengths and limitations of their results, let us briefly summarize the main findings of their study. These are:
a) General distribution of variants: juste is the most common at 41%, followed by rien que, 36%, then seulement (que), 21% and finally ne . . . que, 2%.
b) Regional differences: the real outlier is Saint-Laurent since it is the only community where rien que dominates (it is used here 76% of the time). In all other communities, juste is the most frequent variant. Of note, the only community where ne . . . que has a non-negligible rate of occurrence is Saint-Boniface where it is used six percent of the time.
c) Social factors: juste is favoured by younger speakers and females, but social class results are inconsistent or unavailable; rien que is favoured by older speakers, males and working class speakers (except in Saint-Laurent); seulement (que) is favoured by older speakers in Montreal and Bonnyville, while social class results are inconsistent or not available; ne . . . que: is associated with educated speakers in Saint-Boniface.
d) Linguistic factors: internal constraints are consistent across communities; juste is favoured by verbs and circumstancial complements; rien que is favoured by nouns; seulement (que) is favoured by nouns.
These overall results suggest general homogeneity across Laurentian French speaking communities, in spite of the fact that some social results are inconsistent.
3. METHODOLOGICAL CONSIDERATIONS
It needs to be stated from the onset that the authors of this article have made an admirable effort to draw conclusions about the variable based on extant data. In other words, they did not work with an ideal data set, but rather used available corpora gathered over the last forty years. It would be instructive to consider what an ideal Laurentian corpus might look like by considering some of the shortcomings of the one used.
3.1 Age
The first drawback of the composite corpus used concerns the age-distribution of speakers. Although there is a total of 190 speakers, few younger speakers are represented. While this is compensated in part by reference to previous work on the variable based on Mougeon and Beniak's adolescent Ontario corpus (cf. Mougeon & Beniak 1991), it is relatively difficult to make claims about language change in the Western corpora (i.e., Bonnyville, Saint-Boniface and Saint- Laurent). If younger generations were better represented, we could expect higher rates of juste (assuming increased contact with English) and lower rates of rien que (assuming devernacularisation due to greater contact with standard French in the French language school system). This is particularly true for the Saint-Laurent corpus since speakers are over 40 years old. One might also predict increased use of seulement through greater exposure to standard French in the education system.
3.2 Education
A second limitation of the corpus, and one alluded in the authors' conclusions, is that very little is known about the speakers’ contact with standard French via the education system. I won't dwell on this here, since the authors are keenly aware of this. It deserves mentioning though since any future corpus should be constructed with a view to gather information about speakers' French language education and not just about their general education level. This is particularly important when trying to make sense of the distribution of standard variants (like seulement).
3.3 Language use restriction
Probably the most glaring shortcoming of the composite Laurentian corpus used is the lack of information concerning language use restriction. Mougeon and Beniak's pioneering work on language use restriction (cf. Mougeon and Beniak Reference Mougeon and Beniak1991) has shown that minority language speakers (or bilinguals) should not be treated in a monolithic fashion. Some speakers use French on a daily basis in a variety of settings while other use it almost exclusively at school. This is not only important for understanding the distribution of an English-like variant like juste. It is also relevant for both social class distributions (cf. Nadasdi Reference Nadasdi2000) and even for the relevant effects of internal factors (cf. Mougeon and Nadasdi Reference Mougeon and Nadasdi1998; Grimm and Nadasdi Reference Grimm and Nadasdi2010). We can expect that in the Western provinces represented in the Laurentian corpus, speakers have high and variable levels of contact with English. It is unfortunate that this hallmark of minority French variation is not taken into account.
3.4. Social class
As in most Labovian-inspired variationist studies, the socio-economic status of speakers is examined in Mougeon et al.’s study. This is, however, a challenge since the social representation of speakers across corpora is disparate. This is particularly obvious in Saint-Laurent and Saint-Boniface. On the one hand, the working class is over-represented in Saint-Laurent, on the other hand, the upper-middle class is over represented in Saint-Boniface. As such, it is perhaps not surprising that rien que dominates in Saint-Laurent and that ne . . . que has relatively high currency in Saint-Boniface when compared to other communities (though the fact that the Saint-Boniface interviewer was from France may also have contributed to this result). The different social constitutions of these two corpora may in part be a reflection of the local communities. However, it reminds us that an analysis based on socio-economic status produces less than desirable results in “small places”. Indeed, the distribution of standard and non-standard variants can sometimes make little sense according to standard SES measures. This is not to say, however, that interesting patterns are not to be found. For example, King and Nadasdi (Reference King and Nadasdi1999) examined two cases of variation in Prince Edward Island French, namely the use of first person plural Je . . . ons (versus subject on with standard third person singular morphology) and the use of third person plural -ont (versus standard third person plural verbal morphology). Within the PEI French-speaking community, there is actually a great deal of variation, even though speakers roughly belong to the same social class.
A linguistic marketplace analysis reveals very clear patterns. Frequent use of both standard variants is found in the speech of individuals who are involved in activities where normative French has status. Speakers who are not involved in such activities/occupations make frequent use of the non-standard variants. King and Nadasdi found that speakers in the highest linguistic marketplace category never use the non-standard variant Je . . . ons, while those in the lowest category use it 27% of the time. Similar results have been reached by various scholars using either a linguistic marketplace or social network approach to examining the distribution of standard and non-standard variants. For example, Milroy (Reference Milroy1987) examines a variable involving a standard and non-standard pronunciation among working-class speakers (backing of /æ/). Her study shows that while no social class variation is found (reminiscent of lack of a SES effect for rien que in Saint-Laurent), important differences are found according to speakers' social networks. It would appear that such an approach would be fruitful in communities like Saint-Laurent and Saint-Boniface. The authors in fact do make passing mention of the importance of individuals' community activities. For example, two older speakers account for 2/3 of the occurrences of ne . . . que and are said to be involved in cultural and educational associations. The importance of social cohesion is also mentioned various times in reference to Saint-Laurent, but unfortunately this notion does not seem to have been operationalised.
4. FINAL REMARKS
Mougeon et al.’s study constitutes an important contribution to the study of linguistic variation in Canada. It draws on varied Canadian corpora to arrive at a unified analysis of restricted adverbs. The relative consistency of linguistic results supports the use of the label Laurentian to describe the varieties in question. It would be interesting to gather data on both inter-community contact as well as subjective reactions to different variants to determine the extent to which one could push the notion of a Laurentian speech community. Perceptual dialectological approaches would also be helpful in this regard.
The present study of restrictive adverbs is, of course, only a starting point for future research. There are numerous sociolinguistic variables that should be studied in the western corpora and then compared with existing studies of Quebec, Ontario and Acadian French. Would the same patterns for Saint-Laurent and Saint-Boniface be reproduced with other variables? Some predictions to consider are:
a) In Saint-Laurent, we might expect above-average use of non-standard variants like ca fait que (cf. Beniak and Mougeon Reference Beniak, Mougeon, Mougeon and Beniak1989), rester (meaning ‘live’, cf. Nadasdi Reference Nadasdi2005), subject doubling (cf. Nadasdi Reference Nadasdi2000), use of avoir with être verbs (cf. Beniak & Mougeon Reference Beniak, Mougeon, Mougeon and Beniak1989).
b) In Saint-Boniface, we might expect speakers to favour standard and even hyperformal variants, e.g.: ne retention, /l/ retention, habiter, the simple future and donc.
c) A variety of contact-induced forms might be particularly prevalent in Bonnyville and Saint-Boniface since francophones constitute very weak minorities provincially. This is indeed suggested by the juste results of the present study as this variant dominates in both communities. It would be interesting to see if similar results obtain for variants like à la maison de, sur la télévision (cf. Mougeon & Beniak 1991) and postposted pronouns (cf. Nadasdi 2000).
Finally, it would also be of value to ascertain whether or not the linguistic constraints found to influence the variable use of other variants are observed by Bonnyville speakers. This is not always the case (cf. Mougeon and Nadasdi Reference Mougeon and Nadasdi1998) and if it turns out that some key linguistic constraints are not respected for a number of variables, it would weaken the case for cohesion among Laurentian varieties.
For all the merits of Mougeon et al.’s study, it becomes clear that a new corpus of Canadian French would be most welcome (and is long overdue). Such a corpus would ideally have representations in western Canada (at least Alberta, Manitoba and Saskatchewan), Quebec and Acadia. A recent corpus gathered by Mougeon, Nadasdi and Rehner (2005) already exists for Ontario French and could serve as a model for other corpora gathered in bilingual settings. This would allow for meaningful comparisons and ensure that measures of language use restriction are used in the western provinces. Such a corpus could perhaps benefit from an approach that considers both social class information as well as information regarding social networks and the linguistic marketplace. These latter methodologies would be particularly useful in small communities where traditional social class rankings are not found. A better age distribution would also be welcome in order to gain greater understanding of changes in progress. Finally, their suggestion that comparisons with hexagonal French be undertaken is an important one which, if pursued, will help us better understand features typical of Laurentian French and features shared by a larger community.
You have
Access
1. INTRODUCTION
The purpose of this article is to consider the main contributions of Mougeon et al.’s study and to examine the implications of their findings for future research on Canadian French. We first give a brief overview of results obtained, then present a critical analysis of their findings. In our comments, we consider alternative methodological approaches with a view to reflect on how such alternatives would enhance Mougeon et al.’s results and those of future research.
2. SUMMARY OF RESULTS
Mougeon et al.’s article presents a sociolinguistic analysis of restrictive adverbs in four historically-related Laurentian French speech communities in Canada: a) Bonnyville, Alberta; b) Welland, Ontario; c) Saint-Boniface, Manitoba; and d) Saint-Laurent, Manitoba. It undeniably constitutes a novel and an ambitious undertaking that should pave the way for future comparative research on Canadian French. Before considering the strengths and limitations of their results, let us briefly summarize the main findings of their study. These are:
a) General distribution of variants: juste is the most common at 41%, followed by rien que, 36%, then seulement (que), 21% and finally ne . . . que, 2%.
b) Regional differences: the real outlier is Saint-Laurent since it is the only community where rien que dominates (it is used here 76% of the time). In all other communities, juste is the most frequent variant. Of note, the only community where ne . . . que has a non-negligible rate of occurrence is Saint-Boniface where it is used six percent of the time.
c) Social factors: juste is favoured by younger speakers and females, but social class results are inconsistent or unavailable; rien que is favoured by older speakers, males and working class speakers (except in Saint-Laurent); seulement (que) is favoured by older speakers in Montreal and Bonnyville, while social class results are inconsistent or not available; ne . . . que: is associated with educated speakers in Saint-Boniface.
d) Linguistic factors: internal constraints are consistent across communities; juste is favoured by verbs and circumstancial complements; rien que is favoured by nouns; seulement (que) is favoured by nouns.
These overall results suggest general homogeneity across Laurentian French speaking communities, in spite of the fact that some social results are inconsistent.
3. METHODOLOGICAL CONSIDERATIONS
It needs to be stated from the onset that the authors of this article have made an admirable effort to draw conclusions about the variable based on extant data. In other words, they did not work with an ideal data set, but rather used available corpora gathered over the last forty years. It would be instructive to consider what an ideal Laurentian corpus might look like by considering some of the shortcomings of the one used.
3.1 Age
The first drawback of the composite corpus used concerns the age-distribution of speakers. Although there is a total of 190 speakers, few younger speakers are represented. While this is compensated in part by reference to previous work on the variable based on Mougeon and Beniak's adolescent Ontario corpus (cf. Mougeon & Beniak 1991), it is relatively difficult to make claims about language change in the Western corpora (i.e., Bonnyville, Saint-Boniface and Saint- Laurent). If younger generations were better represented, we could expect higher rates of juste (assuming increased contact with English) and lower rates of rien que (assuming devernacularisation due to greater contact with standard French in the French language school system). This is particularly true for the Saint-Laurent corpus since speakers are over 40 years old. One might also predict increased use of seulement through greater exposure to standard French in the education system.
3.2 Education
A second limitation of the corpus, and one alluded in the authors' conclusions, is that very little is known about the speakers’ contact with standard French via the education system. I won't dwell on this here, since the authors are keenly aware of this. It deserves mentioning though since any future corpus should be constructed with a view to gather information about speakers' French language education and not just about their general education level. This is particularly important when trying to make sense of the distribution of standard variants (like seulement).
3.3 Language use restriction
Probably the most glaring shortcoming of the composite Laurentian corpus used is the lack of information concerning language use restriction. Mougeon and Beniak's pioneering work on language use restriction (cf. Mougeon and Beniak Reference Mougeon and Beniak1991) has shown that minority language speakers (or bilinguals) should not be treated in a monolithic fashion. Some speakers use French on a daily basis in a variety of settings while other use it almost exclusively at school. This is not only important for understanding the distribution of an English-like variant like juste. It is also relevant for both social class distributions (cf. Nadasdi Reference Nadasdi2000) and even for the relevant effects of internal factors (cf. Mougeon and Nadasdi Reference Mougeon and Nadasdi1998; Grimm and Nadasdi Reference Grimm and Nadasdi2010). We can expect that in the Western provinces represented in the Laurentian corpus, speakers have high and variable levels of contact with English. It is unfortunate that this hallmark of minority French variation is not taken into account.
3.4. Social class
As in most Labovian-inspired variationist studies, the socio-economic status of speakers is examined in Mougeon et al.’s study. This is, however, a challenge since the social representation of speakers across corpora is disparate. This is particularly obvious in Saint-Laurent and Saint-Boniface. On the one hand, the working class is over-represented in Saint-Laurent, on the other hand, the upper-middle class is over represented in Saint-Boniface. As such, it is perhaps not surprising that rien que dominates in Saint-Laurent and that ne . . . que has relatively high currency in Saint-Boniface when compared to other communities (though the fact that the Saint-Boniface interviewer was from France may also have contributed to this result). The different social constitutions of these two corpora may in part be a reflection of the local communities. However, it reminds us that an analysis based on socio-economic status produces less than desirable results in “small places”. Indeed, the distribution of standard and non-standard variants can sometimes make little sense according to standard SES measures. This is not to say, however, that interesting patterns are not to be found. For example, King and Nadasdi (Reference King and Nadasdi1999) examined two cases of variation in Prince Edward Island French, namely the use of first person plural Je . . . ons (versus subject on with standard third person singular morphology) and the use of third person plural -ont (versus standard third person plural verbal morphology). Within the PEI French-speaking community, there is actually a great deal of variation, even though speakers roughly belong to the same social class.
A linguistic marketplace analysis reveals very clear patterns. Frequent use of both standard variants is found in the speech of individuals who are involved in activities where normative French has status. Speakers who are not involved in such activities/occupations make frequent use of the non-standard variants. King and Nadasdi found that speakers in the highest linguistic marketplace category never use the non-standard variant Je . . . ons, while those in the lowest category use it 27% of the time. Similar results have been reached by various scholars using either a linguistic marketplace or social network approach to examining the distribution of standard and non-standard variants. For example, Milroy (Reference Milroy1987) examines a variable involving a standard and non-standard pronunciation among working-class speakers (backing of /æ/). Her study shows that while no social class variation is found (reminiscent of lack of a SES effect for rien que in Saint-Laurent), important differences are found according to speakers' social networks. It would appear that such an approach would be fruitful in communities like Saint-Laurent and Saint-Boniface. The authors in fact do make passing mention of the importance of individuals' community activities. For example, two older speakers account for 2/3 of the occurrences of ne . . . que and are said to be involved in cultural and educational associations. The importance of social cohesion is also mentioned various times in reference to Saint-Laurent, but unfortunately this notion does not seem to have been operationalised.
4. FINAL REMARKS
Mougeon et al.’s study constitutes an important contribution to the study of linguistic variation in Canada. It draws on varied Canadian corpora to arrive at a unified analysis of restricted adverbs. The relative consistency of linguistic results supports the use of the label Laurentian to describe the varieties in question. It would be interesting to gather data on both inter-community contact as well as subjective reactions to different variants to determine the extent to which one could push the notion of a Laurentian speech community. Perceptual dialectological approaches would also be helpful in this regard.
The present study of restrictive adverbs is, of course, only a starting point for future research. There are numerous sociolinguistic variables that should be studied in the western corpora and then compared with existing studies of Quebec, Ontario and Acadian French. Would the same patterns for Saint-Laurent and Saint-Boniface be reproduced with other variables? Some predictions to consider are:
a) In Saint-Laurent, we might expect above-average use of non-standard variants like ca fait que (cf. Beniak and Mougeon Reference Beniak, Mougeon, Mougeon and Beniak1989), rester (meaning ‘live’, cf. Nadasdi Reference Nadasdi2005), subject doubling (cf. Nadasdi Reference Nadasdi2000), use of avoir with être verbs (cf. Beniak & Mougeon Reference Beniak, Mougeon, Mougeon and Beniak1989).
b) In Saint-Boniface, we might expect speakers to favour standard and even hyperformal variants, e.g.: ne retention, /l/ retention, habiter, the simple future and donc.
c) A variety of contact-induced forms might be particularly prevalent in Bonnyville and Saint-Boniface since francophones constitute very weak minorities provincially. This is indeed suggested by the juste results of the present study as this variant dominates in both communities. It would be interesting to see if similar results obtain for variants like à la maison de, sur la télévision (cf. Mougeon & Beniak 1991) and postposted pronouns (cf. Nadasdi 2000).
Finally, it would also be of value to ascertain whether or not the linguistic constraints found to influence the variable use of other variants are observed by Bonnyville speakers. This is not always the case (cf. Mougeon and Nadasdi Reference Mougeon and Nadasdi1998) and if it turns out that some key linguistic constraints are not respected for a number of variables, it would weaken the case for cohesion among Laurentian varieties.
For all the merits of Mougeon et al.’s study, it becomes clear that a new corpus of Canadian French would be most welcome (and is long overdue). Such a corpus would ideally have representations in western Canada (at least Alberta, Manitoba and Saskatchewan), Quebec and Acadia. A recent corpus gathered by Mougeon, Nadasdi and Rehner (2005) already exists for Ontario French and could serve as a model for other corpora gathered in bilingual settings. This would allow for meaningful comparisons and ensure that measures of language use restriction are used in the western provinces. Such a corpus could perhaps benefit from an approach that considers both social class information as well as information regarding social networks and the linguistic marketplace. These latter methodologies would be particularly useful in small communities where traditional social class rankings are not found. A better age distribution would also be welcome in order to gain greater understanding of changes in progress. Finally, their suggestion that comparisons with hexagonal French be undertaken is an important one which, if pursued, will help us better understand features typical of Laurentian French and features shared by a larger community.