Bon ben enfin fin in non-native speech: the case of Chinese L1 speakers in Paris

Résumé . Basé sur les données de 40 entretiens sociolinguistiques conduits en français avec des locuteurs chinois apprenant le français à Paris, l’article exploite comment des facteurs extralinguistiques influencent l’usage des marqueurs discursifs (MDs) bon ben enfin fin dans le discours non-natif du français. Les résultats montrent que seul le contact extracurriculaire avec les locuteurs natifs est statistiquement significatif pour les trois MDs parmi les quatre formes examinées. Cela prouve que le contact avec les locuteurs natifs facilite l’apprentissage des MDs dans une situation naturelle : le plus souvent on interagit avec les locuteurs natifs, le plus on utilise les MDs. Abstract. Bon ben enfin fin in non-native speech: the case of Chinese L1 speakers in Paris. Based on the data collected from 40 sociolinguistic interviews conducted in French with Chinese L1 speakers in Paris, the current article exploited how some extralinguistic factors influence the use of discourse markers (DMs) bon ben enfin fin in the non-native speech of French. As a result, among five extralinguistic factors examined in the current research, only the extracurricular contact with native speakers of French was proven to be statistically significant to three out of the four DMs examined here. The only exception is enfin , which might be due to the small number of tokens available for analysis. It is proven that extracurricular contact with native speakers facilitates the acquisition of DMs in a natural setting: the more frequent this contact is, the more one would use the DMs.


INTRODUCTION
While discourse markers (DMs) in native speech have been well documented and studied, their use in non-native speech still needs to be completed. Most studies in SLA concerning the acquisition of the DMs are done on L2 students, who acquire the target language in a classroom setting. The choice of this group of speakers may mainly be due to the availability of the participants to the researchers. Nevertheless, we still know little about L2 users, who do not learn French in a traditional classroom setting, but in the natural setting through their contact with native speakers. As part of the target language community, the extralinguistic factors may also influence the use of DMs in non-native speech. Compared to L2 students, whose learning environment is much more artificial, L2 users may provide data more suitable for analyzing the social impact on language use. Meanwhile, we should admit that since L2 users do not acquire the target language in a natural process as native speakers do, even though they may approach the native pattern, there will always be some discrepancies between the non-native and native patterns.
Therefore, by exploring the data collected from 40 semi-directed sociolinguistic interviews conducted in French with L2 speakers in Paris, this article aims to fill in the gap by proposing to use L2 users' data instead of L2 learners' data for the analysis of some functionally interchangeable DMs as well as how some extralinguistic factors influence the use of these functionally interchangeable DMs in non-native speech.
The structure of this article is as follows: First, the previous research on bon ben enfin fin in native speech will be reviewed. Second, the detailed information on the methodology, including the corpus, the participants, the tokens, and the extralinguistic factors to be examined in this article, will be presented. The statistical results concerning the general tendency and social impact will be analyzed and discussed following the methodology. In the end, a conclusion of the current work and potential implications in the future will be drawn.

PREVIOUS WORKS
Bon ben enfin could be viewed as among the most discussed DMs in native speech. Ample research has been done regarding their discursive functions, not only because they appeared relatively early in the evolution of the French language but also due to their frequent use in oral French. However, this seems not to be the case for fin. Even though we know it is the reductive form of the DM enfin and that the native speakers use it as DM, the major works on DMs seem to omit it for unknown reasons. Maybe it is because its discursive function fully overlaps that of enfin, and since enfin is already a well-studied DM, the researchers do not feel the necessity to invest extra effort into it. Nevertheless, since it derives from enfin, the co-existence of the two forms suggests some discursive difference between them, maybe in their frequency or in the contexts under which they are used. For the current work, we distinguish between these two forms. Barnes (1995) viewed bon as a "coherence-building device", claiming that the English particle well is its closest equivalent. Bon could be interchangeable with ben and enfin when used as hesitation particles or pause fillers, in which case it precedes a self-repair or a reformulation. In other words, it marks the point of discontinuity. She suggested that bon is rarely found alone at the head of a speaker's turn. When appearing at the beginning of the turn, it is often in collocation with alors. Beeching (2009) indicated that bon in contemporary spoken French functions either as a "closing-marker" or as a "hesitation marker". In her study, she found that in a very large number of examples, "bon does indeed play a structuring or textual role, marking a stopping or staging point in the unfolding tale". It expresses some kind of reservation on the part of the speaker. She considered this to be the hedging function of bon, even though this hedging use has become less marginal in recent years. By taking Wheeler's notions of "lower-status" varieties (1994), she suggested that bon could assume the "role of politeness marker, suggesting solidarity and creating a sense of intimacy". She also believed that bon is more used by younger speakers and reflects age-grading rather than semantic change even though the real-time data suggested the contrary. Crible (2018) provided four categories of discursive functions for all DMs: sequential, rhetorical, ideational, and inter-personal functions. For each category, there are some sub-categories of functions. Table 1 presents the discursive functions of bon according to her classification. As shown in Table 1, we notice that in native speech, bon could fulfill many different discursive functions except for the ideational functions, such as consequential or concessional functions.

Ben
As mentioned earlier, ben is the reductive form of the adverb bien. Waltereit (2007) suggested that the reduction of bien to ben may not be a diachronic phenomenon due to pragmatisation. Ben existed as a variant of bien as early as the 16 th century. Its first use as a DM could date back to the 18 th century. Ben seems to be a dialectal variant of bien whose function as DM survived while its adverbial function became obsolete. Luzzati (1982) viewed ben as a consecutive marker introducing the second element of a binary structure based on causal relations. Barnes (1995) believed it to have the same function in English suggested by Schiffrin (1985) to mark the beginning of the reported speech. Barnes (1995) argued that it corresponds very closely to the principal functions of the English well. She considered ben to be a response marker that enhances the coherence of the speech and could be interchangeable with bon in many cases. According to her, ben has two main uses: dialogical and monologic use. Table 2 presents the discursive functions of ben according to Crible (2018)'s classification. We could see from Table 2 that ben covers almost the same function as bon when used as a DM. Unlike bon, ben also fulfills the ideational function (consequence and concession), and in its interpersonal use, it could express the disagreeing.

Enfin
The difference between the original use of enfin and its discursive use is relatively evident. Cadiot (1985) suggested that enfin terminates the preceding segment by signaling that its continuation is either unnecessary or inappropriate. Waltereit (2007) believed that the function of enfin involves "a notion of communicative inadequacy" and occurs at "points of discontinuity or junctures in the discourse". Beeching (2001) perceived enfin as a "repair mechanism". It introduces a correction, either by "making the meaning of the preceding utterance more precise or mitigating the force of what is said". This correction could either be a self-correction or a correction to the points of view of others. Hansen (2005) indicated that the use of enfin as a DM could date back to the late 16 th century, after which enfin underwent a gradual process of semantic and pragmatic extension till the 20 th century. She suggested three "prototypical" meanings for enfin in contemporary French: the temporal meaning, a non-truth-conditional synthesizing meaning, and a repairing meaning. Table 3 presents the discursive functions of enfin according to Crible (2018)'s classification: As shown in Table 3, the discursive function of enfin overlaps that of bon and ben in many categories. This functional similarity makes the three DMs quasi-interchangeable in most cases.

Fin
As mentioned earlier, the use of fin as a DM is a relatively recent phenomenon, and few studies on DMs analyzed its discursive functions. It appeared neither on the list of Chanet (2004) of DMs in native speech nor the more recent work of Crible (2018) on DMs. Since fin is the short form of enfin, it could be seen as pragmatisation considering the phonological erosion. As DM, its functions should not vary too much from the enfin presented above in Table 3. However, it might differ from enfin in other ways. We will get back to this point later in this article.

Corpus
The datasets that we used in the current study were extracted from the oral interaction corpus of 40 semi-guided conversations conducted in French with L2 speakers of French in Paris between 2014 and 2017, counting a total of 30 hours of recording.
The length of each interview varies from 30 minutes to 1 hour 30 minutes depending on the talkativeness of the interviewee. All 40 interviews were conducted by the same interviewer so that the interviewer's influence over the interviewee's performance was minimized. A set of questions was prepared before the interview. However, these questions were used merely as support so that the interviewees could freely talk about their experiences. All interviews were entirely recorded and transcribed by using Praat (Boersma and Weenink, 2021). Since the speakers in our corpus are all non-native speakers of French, there might be some grammatical errors in their oral production. For the transcription, these errors were not corrected. All interviews were transcribed as produced. All interviewees are anonymized in our study by being referred only by name initials.

Participants
The 40 participants are all Chinese L1 and French L2 speakers. They were all born in mainland China and only came to France after puberty. All of them resided in Paris, France at the time of the interview. As mentioned earlier, these speakers are not L2 learners in a classroom setting but the actual L2 users living in Paris. These speakers were recruited via the interviewer and interviewees' social networks. Their participation in the interviews is entirely voluntary.
These speakers, 23 women and 17 men are all between 20 and 55 years old. The time that they spent in the community differs significantly. Some speakers have lived in Paris for more than ten years. Others might recently arrive in Paris. Accordingly, their linguistic competence also varies. The following table presents the detailed information on each speaker in our corpus: We should, however, point out that all 40 speakers received 500 hours' intensive French language instruction before coming to Paris as part of the requirement for a visa. Hence, upon their arrival in Paris, they are at about the same level regarding their linguistic competence. However, after their arrival, some chose to take additional language courses either in their universities or in the city hall.

Tokens
According to Chanet (2004), bon is the 4 th most used DM in French native speech with a frequency of 1 occurrence every 1 minute 18 seconds, supposing one speaks continuously 24 hours a day and seven days a week. According to the traditional grammatical category, bon could function either as an adjective (ex. le bon français) or an adverb (ex. bon enfin bref). When used as a DM, bon functions exclusively as an adverb. However, the adverb bon could be an interjection as in ah bon or a DM as in bon enfin bref. For our analysis, only the adverb DM bon was extracted and analyzed. The adjective and interjection bon are therefore excluded from our analysis.
Ben, a short form of bien, is the 9 th most frequent DM in native speech with a frequency of 1 occurrence every 1 minute 59 seconds. (Chanet, 2004) Ben functions adverbially and is used as a DM in all contexts. Therefore, all tokens of ben are taken into consideration.
The adverb enfin could have two different functions, either as temporal markers to indicate the order when used together with d'abord, et puis, etc., or as a DM when used separately or combined with other DMs. It is the 11 th most frequently used discourse marker in native speech (Chanet, 2004), with a frequency of 1 occurrence every 2 minutes 5 seconds. Only the DM enfin was analyzed, and the temporal marker enfin was excluded from the calculation.
Fin was not on Chanet's list. It could either be used as a noun or as a DM. Only tokens of DM fin were included in our calculation, and the noun fin was excluded. Meanwhile, if this is a new emerging form, we should expect to see age-grading effect on the use of this DM in native speech associated more with youth speech. Considering the relatively young age of the speakers in our corpus, fin should be the favored form if it competes against other forms.
A final verification of all coded tokens was done by inspecting the spectrogram in Praat to ensure the coding for all four forms is precise.

Extralinguistic factors
As shown in many other studies, extralinguistic factors are statistically significant to the oral production of DMs. (See, for example, Sankoff et al. 1997) In this study, five extralinguistic factors that might affect the production of bon ben enfin fin were examined to see how the extralinguistic factors influence the use of these four DMs in non-native speech.
Firstly, as shown in a previous study on non-native speech, male and female speakers have different preferences for certain functionally interchangeable DMs. For example, nonnative male speakers were reported to use more voilà while females prefer oui. (Author, 2017) Even though the gender effect was not found in a native speech concerning these four DMs, we still wanted to see if gender impacts the choice of any of these four DMs in nonnative speech.
Secondly, the age factor will be tested. The participants are divided into three age groups: 20-30 years old, 30-40 years old, and 40-50 years old. In previous research, bon was reported to be age-grading (Beeching, 2009). We expect to see if there is any similar effect on non-native speech.
Thirdly, we will also examine the professional status. Some participants are still university students, while others have already joined the workforce for some time. This difference in their social status would provide them with different contact with native speakers. We divided our participants into two groups: professionals and students. Students may have access to a more formal register as they spend more time in the academic environment, while the professionals may have access to a less formal register through their work and social activities. Since DMs appear more in an informal register, we hypothesize that the professionals would use them more than the students' group.
Fourthly, like all acquisition, the mastery of DMs requires a certain time spent in the target community. The length of stay seems to be a relevant factor for non-native speakers. We divided our participants into three groups according to their sojourn time in Paris: 1-3 years, 4-6 years, and more than 7 years (7 years included). We hypothesize that the longer one stays in the target country, the more possible that one would use more DMs. At last, as DMs are not taught explicitly in language classroom settings, their acquisition could only be fulfilled by extracurricular contact with native speakers of French. Therefore, extracurricular contact could be an influential factor. The participants are divided into three groups depending on their self-reported contact frequency with native speakers during the interview: rare, occasional, and frequent. Our hypothesis for this factor is that the more frequent this contact is, the more one would use DMs.

Congrès Mondial de Linguistique Française -CMLF 2022
Before presenting the statistical results, the general tendency of bon ben enfin fin in a nonnative speech in our corpus compared to the native use from the already existed native oral corpus will be discussed.

General tendency
As shown in Figure I, over 30 hours of recording, 567 occurrences of fin, 349 occurrences of ben, 239 occurrences of bon, and 16 occurrences of enfin were produced by the 40 nonnative speakers in our corpus. To provide a better idea of comparing the frequency of these four forms in non-native speech and native speech, we use the frequency and ranking presented in Chanet's study (2004) as a reference for native speech. Since she calculated the frequency by time, even though this might not be the most ideal way to look at frequency, we also calculated the frequency in non-native speech to be comparable to Chanet's results. The following table shows the ranking and frequency of these four forms in native and non-native speech. However, since the DM fin was not included in Chanet's list, we do not have enough information for comparison for this DM.  Table 5, the frequency of bon ben enfin in the native speech was much higher than that in a non-native speech in our corpus. The non-native speakers use these four DMs at a different frequency than native speakers. Even though these four DMs might be functionally interchangeable for most cases, the fin is the preferred form by our nonnative speakers, while enfin is largely disfavored. Table 2. Ranking and frequency of bon ben enfin fin in native speech (Chanet, 2004)

Bon ben enfin fin in non-native speech
To better understand how our non-native speakers use these four DMs, the following are some examples extracted from our corpus. Since we could not illustrate all aforementioned discursive functions for all four DMs, only one example for each discourse marker will be presented and discussed in this section.
Ex. 1 : non ça va parce que fin c'est pas si loin puis ça porte orthographe c'est encore le français XXX modernes les gens utilisent quand même les mêmes noms vocabulaire enfin à peu près le même (FSJ-2015) As shown in Example 1, the interviewee was talking about the orthograph in old French and modern French. Enfin was used to mitigate between the segment "les mêmes noms vocabulaire" and the segment "à peu près le même". This is a typical example of one of the functions mentioned by Beeching (2001) to "mitigate the force of what was said". The force was degraded by using "à peu près" in the second segment, and the DM enfin precedes a self-correction/self-repair. We could see that the interviewee did want to keep a certain reservation. This might also be viewed as a face-saving situation since decreasing the force prevents the interviewer from opposing what was said.
Ex. 2 : c'est fin pour toutes les tous les étudiants en langue donc on a pas de maths on a pas de fin on a pas on a pas de autres choses du coup on est obligé de choisir une langue étrangère (JWH-2015) Here, the interviewee talks about the course options provided by her university. There are two fin in Example 2. The first fin is used to reformulate what was initiated in the previous utterance. It could indeed, under this circumstance, be seen as a repair mechanism as proposed by Beeching (2001). The speaker might have some difficulties finding the proper way of putting together the description, as after fin the speaker also reformulated once by changing from "toutes les" to "tous les". For the second fin, we could see that the speaker staggers over the same utterance "on a pas". Fin fulfills the function of reformulation. The speaker tried to look for the correct term to complete his exemplification but only finished it with a more general term, "autres choses".
Ex. 3 : ça c'est sûr et en en même temps je crois je suis un peu ben je suis très sensible de de la langue (YW-2015) In Example 3, we see ben also fulfills the function of self-correction. It mitigates the force of what was said, as in the two previous examples. We could notice that before ben, the speaker used "un peu" and after ben, she corrected it to "très". She is not satisfied with the force of the utterance preceding ben.
Ex. 4 : donc c'est à dire des fois des je sais pas des mots par exemple en arabe ou je sais pas moi en espagnol par exemple bon y a des mots qui se mélangent on arrive à se comprendre mais on sait qu'ils parlent moins moins (MWL-2014) In Example 4, bon fulfills the topic-resuming function. The speaker mentioned at the very beginning that there are some words, and then he tried to provide some concrete examples in Arabic and Spanish. Nevertheless, he could not think of an example and therefore returned to the general term that he used at the beginning "des mots". By using bon to resume the previous topic, the speaker also saved his face by not giving any potentially bad example.
As shown by these four examples, we could see that all four DMs have the same discursive functions. If we substitute one with another, the meaning of the utterance will not be changed. The four DMs are indeed functionally interchangeable. It also shows that our non-native speakers could vary among these four DMs. However, since all four forms have both discursive and non-discursive use, it is still important to know what proportion these four forms would be used as DMs in native and non-native speech.
Therefore, from a statistical point of view, we compare the distribution of the discursive and non-discursive use in native and non-native speech. For ben and fin, we explore the corpus CFPP2000 (http://cfpp2000.univ-paris3.fr/), which is an online oral French corpus. For bon and enfin, since Chanet has already discussed them in her article (2004), we just used her results as a reference. As shown in Table 6, the non-native speakers approach the native pattern only for their use of ben: both groups use it exclusively as DMs in all four corpora concerned here. The use of the other three DMs in non-native speech differs from the native pattern in different ways. For bon, the native speakers use it mainly as a DM, while the non-native speakers use it more for its original non-discursive function. The use of enfin in non-native speech might be the closest to the native pattern amongst these three DMs. Both native and non-native speakers use it much more as DMs rather than in its original non-discursive use. The most surprising one would be fin. While the native speakers mainly use it in its original nondiscursive function and rarely as a DM, the non-native speakers show the reversed pattern. It seems that the non-native speakers overgeneralize this DM in their speech. By far, we could see that even though the non-native speakers do acquire the discursive use of these four DMs, they are far from approaching the native pattern in their use: they either overgeneralize the forms or underuse them. Unlike the native speakers, the non-native speakers seem to rely heavily on certain forms, even though these forms do have some functionally interchangeable forms available.
What is more, by taking a closer look at our data, only two speakers used enfin. Enfin was not found in most of the speakers' speech in our corpus. The two speakers who used enfin also used a wider variety of other DMs in their oral production compared to other speakers. It is still unknown why enfin is almost avoided by non-native speakers in our corpus while it is favored mainly in native speech. It could be interesting to be explored with more data in future works.

Gender
As shown in Table 7, bon is slightly preferred by female speakers while ben enfin fin were much more preferred by male speakers. The difference between the two gender groups is more significant with ben enfin fin than with bon. Even though the gender difference was not reported in native speech, in our corpus, female speakers and male speakers have different preferences on their choice of these four DMs. However, the P-value indicates that this difference did not reach the statistically significant level.

Age
As presented earlier, bon was reported to reflect age-grading that an increase in the use of bon was observed among the younger speakers (Beeching, 2009). Therefore, we also wanted to see if these four DMs manifest the same pattern in non-native speech.  Table 8 indicates that there is no age-grading effect in the use of bon in that the youngest group use, in fact, less bon than the oldest group, which is contrary to what was observed in native speech. However, the 40-50 years old group seems to largely prefer ben with a significant increase in their use of ben compared to the other two groups. For enfin, since the total number of occurrences in our corpus is very limited, the difference between groups is not obvious. Meanwhile, fin appears to be favored by the younger speakers, though the difference between the youngest and the oldest group was not as drastic as ben.
As the use of fin as a DM is relatively recent, the observed pattern might be viewed as potential proof of age-grading. However, more data will be needed to draw any concise conclusion.

Professional status
For the professional status, we only divided our speakers into two groups: those who are still students and those who have already started their careers after graduation.
As shown in Table 9, the professionals prefer bon ben and fin while the students' group uses more enfin. The statistical result reveals that the professional status is statistically significant to bon. If bon is viewed as a "politeness marker," as suggested by Beeching (2009), then the observed pattern might be explained. As one starts to work, the surrounding environment becomes more complex than the academic environment in school. The need to deal with different situations in real life more appropriately seems to grow, and the resort to the use of some politeness markers such as bon might help in a face-saving situation. As argued by Beeching (2009: 9), by assuming the role of a politeness marker, bon, "associated with informal contexts", suggests "solidarity" and creates "a sense of intimacy". The difference between the two groups in their different access to vernacular French can also be observed by the average frequency of these four forms: professionals use bon, ben and fin more than students, while students largely favor the only form used in formal register, enfin. Since in academic environment, the register is more formal than other environments, this explains well why we see students group adhere to the only formal form, while the professionals group favor more the informal form.

Length of stay
As the acquisition of DMs requires some time spent in the target community, it seems that a longer stay might facilitate more use of the DMs. For this factor, we divided our speakers into three groups according to their length of stay in Paris: 1-3 years, 4-6 years, 7 years, and above. To our surprise, the length of stay was not proved to be statistically significant to any of the four DMs here. However, we notice that the use of ben and fin seems to increase largely with more time spent in the target community. Though functionally interchangeable in most cases, bon and enfin are different from ben and fin in their formalness. The formers could also appear in the written text, while the latter two could only be used in oral production. The result from Table 10 reveals that acquiring some highly informal DMs such as ben and fin requires some time spent in the natural setting.

Extracurricular contact
The last factor to be examined is the extracurricular contact with native French speakers in Paris. The speakers are divided into three groups based on the self-reported frequency of extracurricular contact: frequent contact, occasional contact, and rare contact.  Table 11 shows that frequent contact with native speakers facilitates the use of bon ben enfin fin in non-native speech. The pattern is almost the same for all four DMs. However, due to the small number of occurrences of enfin, the difference between groups for enfin did not reach the statistically significant level. An increase in use correlates with the increase of contact frequency with native speakers is proved to hold for all three other DMs, especially for ben and fin, whose increase in use is quite drastic. The result reveals that frequent contact with native speakers in the target community is indeed needed to acquire DMs, which also corroborates with the results obtained in previous research on other DMs with non-native speakers from different L1 backgrounds. (Sankoff et al., 1997)

SHS Web of Conferences 13
https://doi.org/10. 1051/shsconf/202213812002 8, 12002 (2022) This article proposed to analyze bon ben enfin fin that are functionally interchangeable when used as DMs. The current work differed from the previous work in SLA on DMs in that we used the data from L2 users in a natural setting other than L2 learners in a classroom setting. Meanwhile, it also exploited how some extralinguistic factors influence the use of bon ben enfin fin in the speech of 40 Chinese L1 speakers of French in Paris. The calculation was based on the data from 30 hours' recording collected between 2014 and 2017.
The ranking of the frequency of these four DMs in non-native speech is different from that in native speech reported in earlier work. Fin is the most popular form by non-native speakers, while bon is the preferred form by native speakers. Very few occurrences of enfin were found in our corpus, while it appears to be one of the most frequently used in native speech. The reason why enfin is the avoided form in non-native speech remains unclear for now. More data is needed for further study.
Despite some tendency observed concerning the impact of some social factors, only the extracurricular contact with native speakers of French was indeed proved to be statistically significant to three of the four DMs examined here in this article, except for enfin, which might be due to the small number of tokens that we had for analysis. It is proved that extracurricular contact with the natives facilitates the acquisition of DMs in a natural setting: the more frequent this contact is, the more one uses the DMs.
Decades ago, DMs were a sign of discontinuity in discourse. With more knowledge of DMs, we start to realize that they help guarantee the fluidity of the discourse with various discursive functions. It is also vital to teach DMs in a language class to help the L2 learners better understand the actual use of the language in a natural setting. It is often reported that L2 learners do not vary their registers as native speakers do. The introduction of DMs might serve as an excellent example of the difference between different registers and, therefore, help the L2 learners master the authentic language use better. However, the potential stigmatization of certain forms should be taken into consideration. It would be better to introduce the concept of DMs to advanced learners rather than beginners. However, teaching DMs remains a problem to be explored further in future studies.