INTRODUCTION
Research on the mastery of sociolinguistic variation by second language (L2) learners is an area of study that has enjoyed considerable growth over the last 30 years. Early research on variation (e.g., Dickerson, Reference Dickerson1974; Ellis, Reference Ellis1987; Gatbonton, Reference Gatbonton1978; Huebner, Reference Huebner1983, Reference Huebner1985; Tarone, Reference Tarone1988) investigated L2 learners’ alternation between native and non-native forms to express a given notion (e.g., an apple versus a apple). This type of variation, referred to as Type 1 variation (Rehner, Reference Mougeon, Nadasdi, Rehner, Dewaele and Mougeon2002, Reference Rehner2004), is a manifestation of the L2 learner's incomplete mastery of an aspect of the target language that is invariant in the speech of native speakers and thus represents a transitional stage in the learner's interlanguage development on the way to categorical use of the native form. Important contributions of this early research were to show that L2 learners’ variation in choice of native versus non-native forms evolves over time and that their frequency of use of native versus non-native forms is impacted by a variety of linguistic and extra-linguistic factors. For an overview of such research, the reader is directed to, among others, Adamson (Reference Adamson1988), Beebe (Reference Beebe1988), Ellis (Reference Ellis1999), and Tarone (Reference Tarone1988, Reference Tarone1990).
More recent research on sociolinguistic variation (e.g., Blondeau and Nagy, Reference Blondeau, Nagy, Jensen and Van Herk1998; Dewaele, Reference Dewaele2002; Major, Reference Major2004; Nadasdi, Mougeon, and Rehner, Reference Nadasdi, Mougeon and Rehner2005; Regan, Reference Regan and Dewaele2005), referred to as Type 2 variation (Rehner, Reference Rehner2002, Reference Rehner2004), is focused on aspects of the target language that are variable in the speech of native speakers (e.g., good day versus hello versus hi versus hey versus yo). Studies of Type 2 variation have, for the most part, been conducted within the Labovian framework of variationist sociolinguistics.Footnote 1 The main goal of these studies is to focus on specific cases of sociolinguistic variation in order to determine to what degree L2 learners use a similar range of variants as do L1 speakers, whether they use these variants at similar levels of discursive frequency as do L1 speakers, and if L2 speakers adhere to similar linguistic and extra-linguistic constraints observed by L1 speakers. For an overview of such research, the reader is directed to, among others, Bayley and Regan (Reference Bayley and Regan2004), Mougeon, Nadasdi and Rehner (Reference Mougeon, Nadasdi and Rehner2010), and Regan, Howard, and Lemée (Reference Regan, Howard and Lemee2009).
Within the study of Type 2 variation, a substantial body of research has investigated the mastery of such variation by high school French immersion learners in the Canadian context (e.g., Lyster, Reference Lyster1994; Mougeon, Rehner, and Nadasdi, Reference Mougeon, Rehner, Nadasdi, Bayley and Regan2004; Rehner, Mougeon, and Nadasdi, Reference Rehner, Mougeon and Nadasdi2003; Swain and Lapkin, Reference Swain, Lapkin, Scarcella, Anderson and Krashen1990). From these studies emerges the clear picture that high school French as a second language (FSL) learners in Canadian immersion programs still have a considerable way to go before being able to master the intricate nature of Type 2 sociolinguistic variation. However, what has yet to attract research attention is the role university studies in FSL play in continuing to develop the sociolinguistic competence, in particular the mastery of sociolinguistic variation, of French immersion students once they graduate high school. Also representing a gap in the current literature is the question of how any progress made by former immersion students now enrolled in FSL studies at the university level differs from the learning of sociolinguistic variation by former non-immersion FSL learners enrolled in the same university courses (i.e., those students who graduated from a ‘core French’ program in high school). This type of research is precisely what is being undertaken in the current paper, a paper that is part of a larger research project on the learning of sociolinguistic variation by core French and French immersion high school graduates enrolled in FSL studies at the undergraduate level (cf. F. Mougeon and Rehner, Reference Mougeon, Rehner and Collier2008).
In Canada, French immersion programs offer a form of bilingual education with instruction through the medium of French and the medium of English. These programs are intended for those who do not normally speak French at home, although Francophone parents may enrol their children in such programs for a variety of reasons (e.g., to encourage their development of English language skills while bettering their French skills). French immersion programs in Canada were originally launched as a response to the demands of Anglophone parents who wanted their English-speaking children to be provided with more effective French language instruction than that offered in core French programs. In Ontario, core French programs are offered at all levels in every school in the province and focus on the structure and use of French. In elementary schooling (from Grade 1 to Grade 4), students study French for an average of 20 minutes per day. These classes are often taught in English. From Grade 5 to Grade 8, French instruction is increased to approximately 30 minutes daily, again with many teachers using primarily the medium of English. In Ontario, every high school student must earn at least one credit in French in order to obtain an Ontario Secondary School Diploma. As such, many students leave core French at the end of Grade 9 after fulfilling this requirement.
In contrast to core French, French immersion teaches mainstream subjects (e.g., math, history, science) through the medium of French. Early French immersion programs provide students with French medium instruction from kindergarten to Grade 6 and are delivered via one of two standard approaches: total or partial immersion. Total early immersion provides L2-only instruction from kindergarten to Grade 1, 2, 3, or 4 depending on the program. When English is first introduced, it is used only to teach English language arts for approximately 1 hour per day. By Grade 6, however, up to 60% of the curriculum is taught in English. Partial early immersion, on the other hand, provides instruction in both French and English right from kindergarten to Grade 6. The most common instructional ratio is 50% English and 50% French, with this ratio remaining constant throughout the grades.
In delayed immersion programs, the use of French to deliver mainstream subjects is postponed until Grade 4 or 5, although core French is usually offered prior to the start of this type of immersion. Again, partial and total immersion options are available. Late immersion programs, offering either one or two years of immersion, postpone the use of French as a medium of instruction until the end of elementary school, usually Grade 7 and/or 8. All subjects except for English language arts are taught in French and the immersion period is preceded by core French or by special preparatory courses. The curriculum covered during the total immersion year(s) is the same as for non-immersion classes.
The university learners in the current research graduated from either a core French or immersion high school program and are currently enrolled in first- or fourth-year undergraduate programs in the Social Sciences or Humanities at a bilingual university in Ontario, Canada. The variable currently under study is the use/non-use of the preverbal negator ne (e.g., tu ne peux pas manger ça ‘you cannot eat that’ versus tu Ø peux pas arrêter ‘you cannot stop’).Footnote 2 The study examines the impact of linguistic factors (i.e., the post-verbal negator—pas ‘not’, jamais ‘never’, rien ‘nothing’, plus ‘no more’, personne ‘no one’) and extra-linguistic factors (e.g., core versus immersion program in high school, language(s) spoken at home, frequency of use of French, stays in a Francophone environment) on the students’ frequency of use of the variants. The results of this analysis are compared with those of previous research on ne use/non-use by Grade 9 and 12 students enrolled in French immersion programs in the Greater Toronto Area (Rehner and Mougeon, Reference Rehner and Mougeon1999) and with the results of previous studies of two other sociolinguistic variables in the speech of the same university learners under study here, namely nous versus on (both meaning ‘we’) as first person plural subject pronouns (F. Mougeon and Rehner, Reference Mougeon, Nadasdi, Rehner, Martineau, Mougeon, Nadasdi and Tremblay2009) and donc, alors, ça fait que and so (all meaning ‘therefore’) as markers of consequence (Rehner and Beaulieu, Reference Rehner and Beaulieu2008).
LITERATURE REVIEW
The study of ne use/non-use conducted by Rehner and Mougeon (Reference Rehner and Mougeon1999) focused on Grade 9 and 12 French immersion students and found that these learners were far from approximating native norms in terms of their frequency of non-use of the negative particle ne. They omitted ne only 28% of the time compared to the highly frequent or nearly categorical rates of native speakers, namely 99.5% of the time for speakers of Quebec French (Sankoff and Vincent, Reference Sankoff, Vincent and Sankoff1980), 98.5% of the time for native speakers of Ontario French (Sandy, Reference Sandy1997), and 63% of the time for native speakers of French from France (Ashby, Reference Ashby1981). The immersion students’ rate of ne non-use was also markedly lower than that of other groups of second language learners, for example Irish learners of French – 50% (Regan, Reference Regan, Bayley and Preston1996, Reference Regan2004); American learners of French – over 63% for learners with extended extra-curricular exposure to French (Sax, Reference Sax2003); and Anglophone Montrealers – 89% (Thibault and Sankoff, Reference Thibault and Sankoff1997). What Rehner and Mougeon highlight in this latter finding is the greater levels of exposure to French outside of the school context that the Irish, American, and Montreal L2 learners had compared to the immersion students. The Irish learners participated in a year-long study-abroad experience in France, the American learners had between eight months and four years of extra-curricular exposure to French, while the Anglophone Montrealers live in a bilingual community and have ample opportunities to interact with native speakers of Quebec French in their daily lives.
This key role of exposure to native spoken French for the Irish, American, and Montreal L2 learners was echoed in Rehner and Mougeon's findings for different subsets of the immersion students. Higher frequencies of ne non-use in the immersion students’ speech were correlated with several measures of increased exposure to native spoken French both outside and within the school context, namely with students having the longest stays with a Francophone family (GoldVarb factor effects: no time = 0.40 versus over two weeks = 0.75), the most frequent use of the French language media (no use = 0.45 versus occasional use = 0.57), and the most schooling in French (0–25% schooling in French = 0.49 versus 40% and over schooling in French = 0.59). The authors suggested that these patterns reflected the greater likelihood of being exposed to ne non-use in situations outside the classroom, either in extra-curricular settings or in school hallways.
Higher levels of ne non-use were also found for those immersion students of working class backgrounds (middle class = 0.44 versus working class = 0.62) and for students in Grade 9 (Grade 12 = 0.44 versus Grade 9 = 0.56), with the former result being attributed to the students’ ability to infer the social value of the variant based on in-class treatment and the latter finding being attributed to the inherent complexity of a two-pronged negation proving especially difficult for younger, less experienced learners. A negative correlation with ne non-use was found for students who spoke a Romance language at home as compared to those students speaking English or ‘other’ languages at home (Romance = 0.30 versus English = 0.47 versus ‘other’ = 0.67), a finding the authors explained by highlighting the use of a non-deletable pre-verbal negator in Italian and Spanish, the two Romance languages spoken at home by the students. No correlation was found between the immersion students’ use of ne and formal interview topics and their deletion of ne and informal interview topics, a finding the authors attributed to the subtlety of that particular type of measure of (in)formality.Footnote 3
Finally, in terms of the effect of linguistic factors, Rehner and Mougeon found that the immersion students omitted ne more often with post-verbal negators other than pas (pas = 0.48 versus ‘other’ post-verbal negators = 0.77), a finding they linked to the relative infrequency of the ‘other’ post-verbals as compared to pas and the resulting desire on the part of the students to simplify the complex negation more often in unfamiliar constructions than in familiar ones.
Concerning the previous studies of other sociolinguistic variables in the speech of the same university FSL learners examined in the current research, Rehner and Beaulieu (Reference Rehner and Beaulieu2008) investigated their use of variants to express a consequence, while F. Mougeon and Rehner (Reference Mougeon, Nadasdi, Rehner, Martineau, Mougeon, Nadasdi and Tremblay2009) documented their mastery of first person plural pronouns. In relation to the use of expressions of consequence, Rehner and Beaulieu investigated the university FSL learners’ use of donc, alors, (ça) fait que, and so and found that the students used the hyper-formal variant donc 25% of the time, the formal variant alors 73% of the time, the mildly-marked informal variant so 2% of the time and did not use the vernacular variant (ça) fait que—elles ne parlent pas le français donc c’était moi qui parlais ‘they don't speak French therefore I was speaking’; je n'aimais pas le programme alors j'ai changé mon majeur ‘I didn't like the program therefore I changed my major’; and je travaille aussi so j'ai un peu d'argent ‘I work too therefore I have a little money’.
Compared to high school FSL immersion learners in a study by Rehner and Mougeon (Reference Rehner and Mougeon2003), the university level students make greater use of donc than do the high school students (25% versus 15%), less frequent use of alors (73% versus 78%), and less use of so (2% versus 7%). Neither group of FSL learners used (ça) fait que. Compared to the findings of research on the use of these variants by adult native speakers of Quebec French (Dessureault-Dober, Reference Dessureault-Dober1974) and by adolescent speakers of Ontario French from the community of Hawkesbury (Mougeon, Nadasdi, Rehner, Reference Mougeon, Nadasdi, Rehner, Martineau, Mougeon, Nadasdi and Tremblay2009) the university learners, and for that matter the high school immersion students, are out of step with L1 Canadian French: (ça) fait que was used 55% of the time by the native speakers of Quebec French and 85% by the speakers of Ontario French, alors 43% for Quebec French and 0% for Ontario French, and donc only 2% for Quebec French and 8% for Ontario French. According to Dessureault-Dober, so is not a feature of Quebec French, but according to Mougeon, Nadasdi and Rehner it is used 7% of the time by the adolescent speakers of Ontario French from Hawkesbury.
In terms of the correlations between extra-linguistic factors and the university learners’ use of alors as compared to donc, Rehner and Beaulieu (Reference Rehner and Beaulieu2008) found that greater exposure to French in the school context favoured the use of alors (secondary schooling in English = 0.45 versus French/mixed = 0.96; graduation from core French program = 0.32 versus French immersion program = 0.57), while greater exposure to the language in extra-curricular settings displayed a negative correlation with alors in favour of donc (no time in a Francophone environment = 0.65 versus semester or more = 0.09). The authors explain these patterns by pointing out that while the students attended school in Canada where research has found the most frequent of the variants to be alors, their stays in French-speaking environments have been almost exclusively in Europe, where they suggest that future research would find donc to be the more frequently used variant, a pattern of variant choice they suggest is being picked up on by the students. In keeping with the findings of previous research on the use of markers of consequence by French immersion high school students (Rehner and Mougeon, Reference Rehner and Mougeon2003), Rehner and Beaulieu found that it is the female students who show a clear preference for alors (female = 0.61 versus male = 0.02). However, in contrast to the study by Rehner and Mougeon, Rehner and Beaulieu found that those students from a Romance language background did not favour alors on the basis of their familiarity with allora as expected, but rather showed a very strong preference for donc (Romance language background = 0.10).
Finally, in relation to the use of so versus donc/alors, Rehner and Beaulieu (Reference Rehner and Beaulieu2008) found that use of this variant was correlated with two distinct student groups. First, so was favoured by those students who had the least exposure to French (core French = 0.79 versus immersion = 0.36; first year = 0.81 versus fourth year = 0.21; time in a Francophone environment was a knockout with use of so only in the two bottom categories with less time spent and no variation in the top category of a semester or more with 100% alors/donc). The authors suggested that it was those students with the least amount of exposure to French who were most likely to use the English variant so in place of its French counterparts due to lack of familiarity with the French variants. However, the second group of students who favoured so were those students who had the highest levels of French or mixed French/English schooling at the secondary level (French/mixed = 0.96 versus English = 0.45). The authors explained that this French/mixed schooling had taken place primarily in Ontario French-medium high schools where the students would have had exposure to Franco-Ontarians who use so as an integrated part of their spoken French.
Turning to the study of first person plural pronouns, F. Mougeon and Rehner (Reference Mougeon, Nadasdi, Rehner, Martineau, Mougeon, Nadasdi and Tremblay2009) examined the use of nous, on, nous-autres on, and nous on to mean ‘we’ in the speech of the same FSL university students. The authors found that the hyper-formal variant nous was used 25% of the time (chaque été ma sœur et moi nous départons pour la Russe ‘every summer my sister and I we leave for Russia’) and the mildly-marked informal variant on was used 75% of the time (ma famille on est ah anglicanes ‘my family we are Anglican’), while the vernacular variants nous-autres on and nous on were used only once each. Compared to the high school FSL immersion learners in a study by Rehner, Mougeon, and Nadasdi (Reference Rehner, Mougeon and Nadasdi2003), we see that the university level students make markedly less frequent use of nous than do the high school students (25% versus 44%) and sharply greater use of on (75% versus 54%). Both groups of FSL learners made nil to highly marginal use of the vernacular variants. Compared to the findings of research on the use of these variants by native speakers of Quebec French (Laberge, Reference Laberge1977), we see that the university learners, and for that matter the high school immersion students, are again out of step with L1 Quebec French: on was used 98.4% of the time by the native speakers and nous was used only 1.6% of the time.Footnote 4 The vernacular variants, according to Blondeau (Reference Blondeau2001), were used with a combined frequency of just under 3%.
Concerning correlations with linguistic and extra-linguistic factors, F. Mougeon and Rehner (Reference Mougeon, Nadasdi, Rehner, Martineau, Mougeon, Nadasdi and Tremblay2009) found that the university FSL students’ use of on compared to nous was associated with several measures of greater exposure to and use of French (French immersion = 0.59 versus core French = 0.30; use of French at levels equal to English = 0.93 versus rare use of French = 0.56 versus no use of French 0.19; French/mixed elementary schooling = 0.80 versus English elementary schooling = 0.45; and a semester or more in a Francophone environment = 0.84 versus 2–3 weeks = 0.53 versus no time = 0.41). In keeping with past research on FSL learners in high school immersion programs (Rehner, Mougeon, and Nadasdi, Reference Rehner, Mougeon and Nadasdi2003), the authors found that the university level FSL students adhered to the linguistic factor examined in previous L1 research, namely that on was most strongly associated with referents that were non-specific and unrestricted (e.g., all of humanity, people in general), while nous was most strongly associated with referents that were specific and restricted (i.e., a limited group of people who the speaker can count and name – e.g., the members of the speaker's family) – (non-specific/unrestricted = 0.72 versus specific/restricted = 0.45). In contrast with Rehner et al.'s previous research, F. Mougeon and Rehner found that the extra-linguistic factor of speaking a Romance language at home did not favour the use of nous over on. Instead, students from a Romance language background favoured on more than those students from English speaking homes (Romance = 0.60 versus English = 0.46).
CHARACTERISTICS OF THE STUDENT SAMPLE
The FSL learners under study here are, as mentioned, the same learners as those examined in the research on expressions of consequence and first person plural pronouns described in the literature review above. These learners are enrolled in their first or fourth year of study in undergraduate programs in the Social Sciences or Humanities at a bilingual university in Ontario, Canada. All students at this college are required to take a minimum of one full course at the second year level taught through the medium of their L2 (French in the case of these learners). Since students at this college cannot study exclusively in their L1, all programs offer a variety of courses in both English and French from which students can choose required courses for their major in their L2. As such, each of the fourth-year students in the current study, regardless of their major subject, will have taken courses through the medium of French, while the same cannot yet be said for the first-year learners.
Further details of the characteristics of the speaker sample are provided in Table 1. As can be seen, 52% of these 61 students were enrolled in a high school French immersion program, while 48% had taken core French courses. Among the 52% of students from immersion programs, 6 students had attended either an elementary or secondary French language school. None of the 48% of students who had attended core French courses had ever been enrolled in a French language school or in an immersion program. Participants were considered as former immersion students if they had attended immersion programs or French language schools for longer than regular English schools. As Table 1 also shows, these 61 learners represent both males and females, though there are more females than males (90% versus 10%, respectively), particularly at the fourth-year level, reflecting an uneven sex distribution among the FSL learners in these programs at the college. Their first languages include Indo-European and non-Indo-European languages, but the students are predominantly native speakers of English (73%). Their elementary and high school learning was undertaken primarily in English language schools (91% and 98%, respectively), though, as mentioned above, some students attended French medium schools for either all or part of their prior schooling. Approximately two thirds of the students have never stayed in a Francophone environment, but among the remaining third, some students have had such stays for as long as a semester (three to four months), usually via a ‘study abroad’ program in Europe.
Table 1. Characteristics of the Student Sample
†One 1st year former immersion student did not indicate an elementary school language.
Finally, while the levels of overall fluency and grammatical proficiency of these students is not the focus of the current research, it should be noted that these learners, whether former core French or former immersion students, produced longer stretches of sustained speech in their interviews than did the high school French immersion students reported on in the research by Mougeon and his colleagues, despite the fact that the interviews were of the same length and conducted by the same interviewer according to a nearly identical set of questions. Specifically, the average number of words produced during the interviews by the high school French immersion students was approximately 3500 words, while the approximate averages for the university learners were 2300 words for the first-year former core French students, 3200 words for the fourth-year former core French students, 3400 words for the first-year former immersion students, and 4700 words for the fourth-year former immersion students. These numbers show an interesting difference between high school and university learners, between university learners from former core French versus immersion programs, and between first-versus fourth-year university learners. Having said this, it is also important to point out that the university learners continue to make some of the same grammatical errors during the course of their interview that were present in the interviews with the high school immersion learners (e.g., occasional mismatches in agreement between subject and verb, errors in assigning grammatical gender, incorrect verb tense).
METHODOLOGY
Following the methodology of previous sociolinguistic studies, notably those to which the current results will be compared, the 61 students under study here participated in a Labovian-style interview and completed a language background questionnaire. The questionnaire was administered prior to the interview and focused on the students’ patterns of use of their L1 and of their L2-French both within and outside of the university setting. The questionnaires also provided general sociological information about the students. The data generated by these questionnaires are used here as independent variables in a factor analysis that correlates these variables with the students’ use/non-use of ne. The individual, semi-directed, semi-formal interviews were 40 to 50 minutes long and were conducted by the same individual who interviewed the high school French immersion students in Mougeon, Nadasdi, and Rehner's research. The interviewer is a native speaker of French from France who has lived in Canada for over 30 years. The interview design and transcription protocol are based on those used by Mougeon and Beniak (Reference Mougeon and Beniak1991) in their sociolinguistic research on Franco-Ontarian high school students and, subsequently, by Mougeon, Nadasdi, and Rehner (Reference Rehner2002) in their sociolinguistic research on the high school French immersion students. The interview questions tapped the students’ experience at university, family activities, ideas on current issues, and their thoughts about their own bilingual competence. Specifically, students were asked questions about their favourite past-times, the clubs they belong to, recent films they had seen and books they had read, what types of music they listen to, what sports they play, and what they do during their summer vacations. Students were also asked to talk about family holiday traditions, the role religion plays in their lives, how they view child/parent relationships, funny or amusing situations they have experienced, frightening events they have lived through, tricks they have played on teachers or friends, what they would do if they won the lottery, and where they have or would like to travel. Other questions asked students to elaborate on times they have stayed in a Francophone environment, on the differences they perceive between French in Quebec and in France, on how they learned French at school, on how they would change FSL teaching if they could, and on whether or not they consider themselves bilingual. Finally, the interviews also asked students to discuss what their plans are after graduation, what they know of global politics, and what their views are concerning the future of the world.
In order to identify tokens of ne use and non-use in the students’ interview speech, the computerised concordance program MonoConc Pro (Barlow, Reference Barlow1998) was employed. Searches were conducted for pas, jamais, rien, plus, and personne used both with and without ne. Several instances of negation with two post-verbal negators other than those listed above were also found in additional searches, namely aucun(e) and guère. A search for aucun(e) revealed 13 tokens, of which 8 were used with ne and 5 without ne. A search for guère revealed only one token and it was used with ne. Given that these two post-verbal negators were not included in previous sociolinguistic studies of ne use/non-use and since there were so few tokens identified in the present corpus, these two post-verbal negators were excluded from the study.
Tokens were also excluded if they represented unfinished thoughts (e.g., j'ai vu une mais je ne. . . ‘I saw one but I didn't. . .’), false starts or repetitions (e.g., alors l'ami ne pouv ne pouvait pas sortir ‘so the friend couldn't couldn't leave’); if they were followed by inaudible words (e.g., non je n’ai XXX ‘no I don't have (inaudible)’); or if the surrounding sounds made it impossible to determine if ne was used or not (e.g., on (n’?) a pas beaucoup de famille qui . . . ‘we don't have a lot of family who. . .’). Once the tokens were identified, they were coded according to the post-verbal negator used (i.e., pas, jamais, rien, plus, personne). Coded tokens were then analyzed using GoldVarb 2001 (Robinson, Lawrence, and Tagliamonte, Reference Robinson, Lawrence and Tagliamonte2001). This statistical software runs a multivariate analysis in order to obtain frequency counts and factor weightings indicating which linguistic and extra-linguistic factors are significantly correlated with the students’ variant choice. Specifically, this program performs a step-wise logistic regression analysis and yields an ordered selection of the factors that are most closely associated with variant choice. Factor effects vary between 0 and 1, with values greater than 0.5 indicating that a sociolinguistic variant is favoured and values less than 0.5 indicating that the variant is disfavoured.Footnote 5
HYPOTHESES
Given the findings of previous research related to studies of ne use/non-use by other Ontario FSL learners and related to the use of other sociolinguistic variables by the university level FSL learners under study here, the following hypotheses can be drawn. These hypotheses address the frequency of use/non-use of ne by the university FSL learners, the impact on this use/non-use exerted by various linguistic and extra-linguistic factors, and how the patterns obtained in relation to the above will compare to the results of previous research on high school FSL learners.
• First, in terms of frequency, one can expect the learners to make frequent use of the mildly-marked informal variant of ne non-use, since these same learners have been shown to make extensive use of the mildly-marked informal variant associated with first person plural subject pronouns, namely on. Further, one can expect that this rate will be higher than that of the high school FSL learners, but at levels not yet approaching native norms. This hypothesis reflects the fact that these university FSL learners outperformed the high school FSL learners in their use of on, but that these levels did not approach native norms.
• Second, regarding the impact of the linguistic factor, namely the nature of the post-verbal negator, one can expect that the learners will adhere to the same constraint observed by the high school learners, namely they will omit ne more often before less frequent post-verbal negators than before pas. This hypothesis is in keeping with the fact that the university learners observed the same linguistic constraint on nous/on variation as was documented for the high school learners. However, given that this pattern of influence of post-verbal negator is based on degree of difficulty of maintaining a complex two-pronged negation in unfamiliar constructions, it is likely that the degree to which there is a difference between deletion rates before pas and before the other post-verbal negators will be lessened for the university level learners compared to their high school counterparts. This reflects the fact that the university level learners have had more exposure to and practice with the language than have the younger learners and therefore may be more likely to have mastered this dimension of French negation.
• Third, concerning the influence of extra-linguistic factors measuring exposure to and opportunities to use French, it can be hypothesised that those university level learners with higher scores on these measures will make greater use of the mildly-marked informal variant of ne non-use than will the remaining university learners, as was the case for the university learners in relation to their use of the mildly-marked informal variant on. Further, given that the university learners’ amount of exposure to and opportunities to use French in such settings has been greater than was reported for the high school learners, it can be expected that the degree to which these factors will impact on the non-use of ne will be more marked for the university learners than it was for the high school learners.
• Finally, in relation to the factor of languages spoken at home, it can be expected that the tendency for students from a Romance language background to strongly favour ne use that was documented for the high school FSL learners will not be found for the university level students. This hypothesis reflects the fact that the expected associations between Romance language background and alors use and nous use found for the high school learners were not found for the university level learners. In fact, based on the high level of correlation between the university level students from a Romance language background and donc and on, one could hypothesize that these students will strongly favour ne deletion.
RESULTS
With these hypotheses in mind, let us turn now to the results of the analysis of ne use/non-use by the 61 university level FSL students under study. Table 2 provides the results of the GoldVarb analysis.
Table 2. Effects of Factors on Ne Use and Non-Use by University FSL Learners
First, as expected, the students make frequent use of the mildly-marked informal variant of ne non-use (42%).While Table 2 shows that the students still favour ne use over its non-use their frequency of non-use is, as expected, considerably higher than that documented for the high school FSL learners. As Table 3 shows, the university level learners’ rate of 42% is well above the 28% documented for the high school students. Also as expected, the students’ rate of 42% ne non-use is still well below the rates documented for native speakers of French, which, as the reader will recall, were 99.5% for speakers of Quebec French, 98.5% for speakers of Ontario French, and 63% for speakers of French from France.
Table 3. Effects of Factors on Ne Use and Non-Use by University FSL Learners versus High School FSL Learners
Second, concerning the impact of the linguistic factor, Table 2 shows that, as expected, the university learners do indeed omit ne more often before post-verbal negators other than pas (pas = 0.49 versus ‘others’ = 0.58). However, as expected, Table 3 shows that the influence of ‘other’ post-verbal negators on ne non-use is less significant among the university level learners than among the high school learners (university = 0.58 versus high school = 0.77). Interestingly, the importance of pas as the post-verbal negator is almost identical across the two groups (university = 0.49 versus high school = 0.48). These findings suggest that the degree of difficulty in maintaining a complex two-pronged negation in less familiar contexts does not present as much of a challenge for the more advanced university learners as it does for the high school students. As a result, it appears that learning continues to take place between the high school and university levels with regard to the formation of negation in French, the use of less frequent post-verbal negators, and the sociolinguistic ability to control the use of variants in challenging linguistic contexts.
Third, in relation to the impact of extra-linguistic factors measuring curricular and extra-curricular exposure to and opportunities to use French, Table 2 shows that, with the exception of one measure (year of study), it is, as expected, in those categories indicating the most contact with French that the students demonstrate the highest levels of ne non-use. For instance, regarding curricular contact with French, Table 2 shows that immersion graduates omit ne significantly more often than do graduates of a core French program (0.54 versus 0.34, respectively) and students with French or mixed French/English elementary schooling omit ne more often than do those students with English elementary schooling (0.88 versus 0.43, respectively). Concerning measures of extra-curricular exposure to French, Table 2 shows that students who report using French in equal proportion to English omit ne more often than do students who report rarely or never using French (0.81 versus 0.51 versus 0.44, respectively),Footnote 6 as do students with at least a semester long stay in a Francophone environment, as compared to students with only a two to three week long stay or with no such stay at all (0.80 versus 0.60 versus 0.43).Footnote 7
The exception to this trend represented by year of study which displayed higher levels of ne non-use by first-year rather than fourth-year students is interesting and unexpected. A closer examination of this factor reveals an important distinction between former core French and former immersion students. Among former core French students, fourth-year students are indeed deleting ne more often than first-year students, as one might expect (factor weight 0.79 versus 0.37, respectively), whereas the former immersion students match the overall trend reported in Table 2.Footnote 8 This finding suggests that continued university study is a more significant factor in promoting the development of informal sociolinguistic markers for former core French students than for former immersion students.
One possible explanation for this finding is that the fourth-year former core French students have now had the opportunity to learn through the medium of French (recall the requirement that students take at least one full course taught through the L2 at the college), like both the first- and fourth-year former immersion students have had who learned through the medium of French in their high school immersion programs. It is possible that this experience of learning through the medium of the L2 promotes greater use of the informal features of the target language, such as the non-use of ne. This is an experience that the first-year former core French students have not yet necessarily had.
Another possible explanation for this findings may lie in the anecdotal reports by former core French students who say that core French programs prepare them well for formal uses of the language, but not informal ones; whereas anecdotal reports by former immersion students reveal the opposite. These anecdotal reports seem to be borne out by the low levels of ne non-use for first-year former core French students (17%) and relatively ‘high’ levels of ne non-use for first-year former immersion students (56%). Interestingly, fourth-year former core French students display an almost identical rate of ne non-use to that of the first-year former immersion students (57% versus 56%, respectively), while the rate of ne non-use among fourth-year former immersion students falls to 44%. Perhaps what is at issue here is the differential learning goals of former core versus former immersion learners upon entry into university study. It is possible that a desire among former core French learners to improve their informal register leads them to focus on ne non-use as an informal marker and ‘improve’ this aspect of their spoken French by the end of their university studies. Conversely, it is also possible that a desire among former immersion students to improve their formal register leads them to focus on ne use as a formal marker and ‘improve’ this aspect of their spoken French by the end of their university studies. Clearly, the two possible explanations proposed here await further study in order to be confirmed or refined.
Notwithstanding the interesting exception discussed above, it remains clear that, as expected, increased contact with French has a favourable effect on ne non-use in the spoken French of these university learners. Further, as expected, the impact of curricular and extra-curricular contact with French appears to have a greater impact on the non-use of ne by the university learners than by the high school learners. As Table 3 shows, students at both levels of study with the least amount of elementary education in French display remarkably similar factor effects for ne non-use (university = 0.43 versus high school = 0.49), while in the category marking more curricular contact with French in the elementary grades there is clearly a difference between the two levels (university = 0.88 versus high school = 0.59). This pattern is even more striking when one looks at the extra-curricular contact with French that learners obtain through stays in a Francophone environment. In the categories marking no time or less than one day, the factor effects are the same across the two levels of study (university = 0.43 versus high school = 0.43), while in the top categories of a stay of a semester or longer or a stay of three weeks or longer the difference between the two levels is immense (university = 0.80 versus 0.31).Footnote 9 These patterns not only reinforce the idea that contact with French in a naturalistic environment allows learners to refine their choice of sociolinguistic variants, but also highlight the fact that prolonged contact of a semester or more is needed in order to make real progress. While the differences in the impact of continued contact with French between the two levels of study are clearly notable, the fact that the university level learners are still not approximating native norms suggests that even the current university experience combined with a semester in a Francophone environment is not enough to provide learners with sufficient exposure to and opportunities to use the language in order to fully master the subtleties of sociolinguistic variation. More intense contacts appear to be needed, especially when one recalls the even higher levels of ne non-use documented for the Irish FSL learners in Regan's (Reference Regan, Bayley and Preston1996) study who participated in a year abroad program, for the American FLS learners in Sax's (Reference Sax2003) study with between eight months and four years of extra-curricular exposure, and for the Montreal Anglophones in Thibault and Sankoff's (Reference Thibault and Sankoff1997) study who interact daily with native speakers of French in their community.
Finally, concerning the factor of languages spoken at home, Table 2 shows that, as expected, the university learners from a Romance language background do not favour ne use, as did the high school learners. Instead, as expected, these students show a strong and marked preference for ne non-use (0.80). Interestingly, this striking difference between the sociolinguistic choices of students across the levels of study applies only to students of Romance language backgrounds, not to students from English language backgrounds.Footnote 10 Recall that the high school students from a Romance language background showed a marked preference for ne use (0.30), likely based on the presence of a non-deletable preverbal negator in Spanish and Italian, while the high school students who spoke only English at home showed only a very weak preference for ne use (0.47). What the above findings suggest is that there may be a threshold of competence beyond which the impact of the L1 is lessened when the original impact of the L1 was strongly marked, as was the case with the students from a Romance language background.
CONCLUSION
The present study has investigated the sociolinguistic competence of 61 former core French and French immersion students enrolled in their first or fourth year of undergraduate studies at a bilingual university in Ontario, Canada by examining their ability to vary their spoken French in sociolinguistically appropriate ways. Specifically, it has explored their variable choice of ne use versus ne non-use in the context of a semi-formal, semi-directed interview and the set of linguistic and extra-linguistic factors correlated with this variable use. The results have been compared with findings on ne use/non-use by high school FSL learners in Ontario. In sum, the results of the study are as follows:
• First, the university students have been shown to make frequent use of the mildly-marked informal variant of ne non-use (42%), a rate of frequency that is considerably higher than that documented for high school FSL learners, but still considerably lower than that documented for native speakers.
• Second, the students were found to omit ne more often before post-verbal negators other than pas, but this effect was less pronounced than was the case for high school learners.
• Third, with only one interesting exception the students demonstrated the highest levels of ne non-use in relation to those factor categories indicating the most curricular and extra-curricular contact with French. Further, the impact of increased contact with French was more pronounced for the university learners than for the high school learners.
• Finally, the students from a Romance language background showed a strong and marked preference for ne non-use, whereas high school students from similar backgrounds clearly favoured ne use.
These findings suggest that L2 learning continues to take place between the high school and university levels, specifically with regard to the formation of complex L2 structures, the use of less frequent features, and the sociolinguistic ability to control the use of variants in challenging linguistic contexts. The findings also suggest that the nature of students’ goals for this continued learning may be impacting the type of progress they make. Further, the present study suggests that there may be a threshold of L2 competence beyond which the impact of the L1 is lessened when the original impact of the L1 was strongly marked. Finally, the results also serve to reinforce the idea that contact with the L2 in both curricular and naturalistic environments allows learners to refine their choice of sociolinguistic variants.
Interestingly, the findings in relation to ne use/non-use are similar to those of two other studies of sociolinguistic variation in the speech of these same 61 university FSL learners. As Table 4 shows, the university students also make frequent use of the mildly marked informal variant on (75%, compared to the hyper-formal variant nous) and of the formal variant alors (75%, compared to the hyper-formal variant donc). As Table 4 further shows, factor categories indicating the most curricular and extra-curricular contactFootnote 11 with French also favour the students’ use of on and alors over nous and donc: immersion versus core program; use of French equal to use of English versus rare or nil use of French; French or mixed elementary schooling versus English elementary schooling; and a semester or longer stay in a Francophone environment versus a stay of 3 weeks or less.Footnote 12 Finally, Table 4 shows that the university students from a Romance language background display a marked preference for ne non-use, on, and donc, while high school learners, as the reader will recall, strongly favour ne use, nous, and alors.
Table 4. Effects of Factors on Ne Non-Use, Use of On, and Use of Alors by University FSL Learners
The results of this three-study comparison in the speech of the university learners clearly show that increased contact with the L2 both within and outside the school context serves to improve the students’ sociolinguistic competence insofar as it further develops their mastery of sociolinguistic variation, a finding in keeping with the role of this factor in the mastery of sociolinguistic variation by high school L2 learners. However, the results of the three-study comparison also highlight the fact that even when continued L2 study at the university level is combined with a semester abroad, it is not enough to provide learners with sufficient exposure to and opportunities to use the language in order to fully master the subtleties of sociolinguistic variation. More intense contacts appear to be needed to reach native-like levels of variant use, especially when one recalls the findings of Regan's (Reference Regan, Bayley and Preston1996) study examining the impact of a year-abroad program on the learning of sociolinguistic variation in a L2, of Sax's (Reference Sax2003) study highlighting the role of extended extra-curricular target language exposure on the learning of sociolinguistic variation, and of Thibault and Sankoff's (Reference Thibault and Sankoff1997) study investigating the role played in the development of L2 sociolinguistic competence by daily interactions with native speakers in the local community.
A study by F. Mougeon and Rehner (Reference Mougeon, Rehner and Collier2008) provides additional evidence that it takes more than university L2 study and a semester abroad in order to come close to approximating the native norms of sociolinguistic variant use. This study, based on the same university FSL learners under study here, correlated the fourth-year students’ lifestyle choices with their use of on and their non-use of ne. The lifestyle choices examined concerned the students’ engagement in or plans to engage in activities or initiatives that brought them into direct and prolonged contact with French, such as choosing a French-speaking partner in life, looking for a bilingual job, qualifying for a position where French is the language of work, living in a French-speaking country or city, enrolling in courses taught through the medium of French, and traveling in Francophone countries. F. Mougeon and Rehner showed that those students enrolled in university FSL classes who had been abroad for a semester or longer and who had also made or also planned to make two or more such choices demonstrated native-like levels of on use (99%) and near-native-like levels of ne non-use (87%).
Taken together, the results of the present study and the findings of this past research show clearly that university students have to rely heavily on extra-curricular contacts with their L2 to reach native-like levels of sociolinguistic variant use. The direct implication of this finding is that the university classroom learning they are engaging in is not providing them with the information and opportunities they need to develop this aspect of their L2 competence. Previous research has shown that high school classroom input is not providing L2 learners with this type of information or these types of opportunities either (cf. Rehner and Mougeon, Reference Rehner and Mougeon2003). This research demonstrates that high school French immersion teachers’ in-class speech and the textbooks and workbooks used in these settings are not presenting L2 learners with a sufficiently rich linguistic model, set of instructions, or range of activities. These types of in-class input are not offering learners sufficient exposure to a wide range of variants, to information on relative discursive frequencies, or to insights into the linguistic and extra-linguistic correlates of variants in order for them to master native-like levels of sociolinguistic variation. Having immersion teachers, in particular, offer this type of rich and diversified sociolinguistic input would be especially important given the heavy reliance of immersion learners on their teachers for modelling their sociolinguistic variation. For instance, Mougeon, Nadasdi and Rehner (Reference Mougeon, Nadasdi and Rehner2010) report that the Ontario French immersion teachers examined in their research delete ne 29% of the time while teaching (contrasted, as mentioned earlier, with levels nearing categorical for native speakers of Canadian French), a figure nearly identical to the 28% non-use of ne reported by Mougeon and Rehner (Reference Mougeon and Rehner1999) for the French immersion high school students. However, similar research into the nature and importance of the educational input offered to university L2 learners has not yet been undertaken and would appear to be the next step in helping to develop the sociolinguistic competence of university L2 learners who are not able to undertake a semester abroad and who are not able to gear their lifestyles toward intense and prolonged contact with their L2.
