Community and lifespan changes in music: Sociophonetic variation in Laurentian French

. Previous studies on English have highlighted various instances where individual singers or small groups change which dialectal features appear in their music (e.g. Trudgill 1997; Beal 2009; Coupland 2011; Eberhardt & Freeman 2015; Lyo n 2019). Whereas corpus studies on music have the option between real-time or apparent-time analyses, most previous research on music has largely been conducted via case studies on change across a singer or group’s career (see Gibson in press a). Focusing on Laurentian French (also known as Quebec French or Canadian French), multiple singers may moderate dialectal traits in music due to their albums being released internationally, where the dialect faces stigma (Szlezák 2015). We further hypothesize that pop singers are especially sensitive to international norms and stigma because they are more likely to market abroad due to pop music’s greater international appeal (Grenier, 1993). We examine non-lengthened high vowel laxing in closed final syllables (e.g. /vit/ [vɪt] vite ‘fast’; Dumas, 1983), a process characteristic of Laurentian French that is categorical (Côté 2012) and non-stigmatized (Lappin 1982; Paradis & Dolbec 1998, Reinke et al. 2006) within the dialect. We expand the study of dialectal traits in music beyond English by using a new corpus of 20 Québécois singers who sing in French and are of the Laurentian French dialect. Additionally, we analyze the patterning of groups of singers across their careers, rather than the patterning of a single singer, and analyze real-time change of groups as opposed to individuals.


Introduction.
Previous studies have highlighted various instances where individual singers or small groups, particularly those who sing in English, change which dialectal features appear in their music (e.g.Trudgill 1997;Simpson 1999;Andres Morrissey 2008;Beal 2009;Coupland 2011;Gibson 2011;Gibson & Bell 2012;Eberhardt & Freeman 2015;Lyon 2019).Additionally, corpus studies on music have the option of analyzing change in real-time or apparent-time, although most previous research on music has largely been conducted via case studies on change across a singer or group's career (see Gibson, in press a).This study fills several gaps in the existing literature on the featuring of dialectal traits in music: (1) we expand the study of dialectal traits beyond English by using a new corpus of 20 Québécois singers who sing in French and are of the Laurentian French dialect (also known as Quebec French or Canadian French); (2) we analyze the patterning of groups of singers, based on genre, across their careers, rather than the patterning of a single singer; and (3) we analyze real-time change of groups of singers, based on genre, as opposed to individuals.
We pay particular attention to differences between pop singers and those of other genres because pop singers appeal to larger, more general audiences in comparison to singers of other genres (Rojek 2011).Further, English pop music has been shown in a recent study to have a supralocal standard, Pop Song English, which is influenced heavily by American Englishes (Gibson, in press b).By focusing on Laurentian French, we aim to test for a similar effect whereby singers in our corpus feature dialectal traits of Hexagonal French (i.e.varieties of French spoken in France).Additionally, because Laurentian French speakers are subject to stigma from speakers of Hexagonal French (Weinstein 1989;Lüdi 1991;Szlezák 2015), we expect singers will accommodate to Hexagonal French norms to varying degrees.
The study addresses three research questions.The first question targets what differences exist between pop and other genres in Quebec: (1) Do pop singers exhibit dialectal features as often as singers of other genres?We postulate that pop will exhibit fewer traits of Laurentian French than singers of other genres since pop singers are more likely to market abroad (Grenier 1993).Therefore, pop singers are more likely to be sensitive to stigma from international audiences and more likely to conform to Hexagonal French norms.
Our second research question focuses on change across the careers of singers: (2) Are dialectal features produced at the same rate across singers' careers?We predict that francophone artists from Quebec will exhibit similar tendencies as those of British singers in the work of Trudgill (1997); they will increase their usage of dialectal variants over the course of their careers.However, we anticipate that any increase will be limited because Laurentian French is still perceived negatively abroad.
Our final research question probes real-time change in the francophone music industry in Quebec: (3) Does year of album release have an impact on the featuring of dialectal traits?We hypothesize that singers who began their careers more recently will produce dialectal variants at higher rates than artists who started releasing albums earlier due to waning effects of the destigmatization of Laurentian French.
In this article, we begin in Section 2 by presenting previous literature on genres and dialectal accommodation in music, the status of Laurentian French in Quebec and France, and the phonological process targeted in this study (high vowel laxing).Section 3 presents the methodology used in this study by detailing the creation of the corpus, the extraction and coding of tokens, and the statistical analysis.We then present the results of the analysis on laxing (Section 4).Section 5 discusses how pop may be particularly sensitive to norms more similar to Hexagonal French.Additionally, we infer that changes across time for pop singers reflect ongoing effects of dialectal destigmatization.The conclusion (Section 6) addresses limitations of the current study and highlights the additions our study brings to questions regarding dialectal accommodation and linguistic norms in music.

Literature review.
In this section, we first review the previous studies examining genre and dialectal accommodation in music (Section 2.1).Section 2.2 then discusses the status of Laurentian French in Quebec along with the attitudes from speakers of Hexagonal French that are directed towards speakers of Laurentian French.Finally, we introduce high vowel laxing in closed final syllables (Section 2.3), which serves as a case study for the present paper.

GENRE AND DIALECTAL ACCOMMODATION. Early work on dialectal accommodation has
shown a tendency for British singers and groups, such as the Beattles and the Rolling Stones, to exhibit dialectal features of American Englishes in their songs (Trudgill, 1997;Simpson, 1999).For example, they would often produce a /t/ that occurs intervocalically as a tap or a flap, characteristic of American Englishes, rather than as a glottal stop, characteristic of British Englishes (Simpson, 1999).Although, both the Beattles and the Rolling Stones exhibited fewer traits of American Englishes in albums which were released later in their careers in comparison to their first few albums (Trudgill, 1997).Additionally, Gibson's (in press b) recent work demonstrates a similar phenomenon is present in singers from New Zealand via a study of the target vowel in BATH and nonprevocalic /r/; New Zealand pop singers exhibited a tendency to produce Americanized versions of these variants rather than those of their native dialect.To this end, we anticipate Québécois singers will be more likely to exhibit dialectal features of Hexagonal French rather than features of Laurentian French, given that Hexagonal French singers represent the majority in the francophone music industry (Grenier 1993).
Although the dominance of a single dialect in music can influence accommodation, other factors, however, may also influence the featuring of dialectal traits in music.For example, Beal's (2009) study on the British indie group, The Arctic Monkeys, indicates that genre may serve as a predictor for the featuring of dialectal traits in music.In the case of the Arctic Monkeys, rather than accommodating to Americanized versions of local variants, they opted to feature traits of Sheffield English in their music.Beal proposed this is due to the fact that The Arctic Monkeys are an indie group, rather than a pop group.Therefore, by singing in the local dialect they demonstrated an association with the local identity and ideals as opposed to those of the mainstream culture.We therefore suspect the same may hold true for indie singers from Quebec; they might feature dialectal traits more frequently than pop singers.
The featuring of dialectal traits, however, can be associated with other genresnot only indie.Other case studies have demonstrated this phenomenon as well.For example, Eberhardt & Freeman (2015) showed that Iggy Azalea, a white Australian woman, raps and performs hip hop while exhibiting features of African American English.Additionally, Lyon (2019) showed that Taylor Swift, from Pennsylvania, featured traits of Southern American English during her era as a country singer, and then stopped featuring these traits when she began producing pop music instead of her traditional country songs.These case studies highlight that singers who exhibit dialectal features in their songs tend to be differentiated as genres other than pop.Gibson's (in press b) recent work also demonstrates this phenomenon.Hip hop singers from both the United States and New Zealand exhibited more dialectal features than pop singers did.
Based on the findings of Beal (2009), Eberhardt & Freeman (2015), Lyon (2019), and Gibson (in press b), we expect Québécois singers who are not of the pop genre may feature dialectal traits more often than pop singers, since pop tends to feature traits that are not as marked dialectally as those featured by singers of other genres.Further, we posit that pop singers will be particularly subject to stigma and dialectal norms based on Hexagonal French because pop singers are more likely to release albums and play shows abroad than singers of other genres (Grenier, 1993).We discuss the status of Laurentian French in Quebec as well as commonly held beliefs regarding Laurentian French by speakers of Hexagonal French in the following section.
2.2.STATUS OF LAURENTIAN FRENCH.Given that singers market to large, diverse audiences to achieve success and lengthy careers (Zwaan & ter Bogt 2009), we consider attitudes towards Laurentian French both in Quebec as well as in France, which are the two main audiences targeted by Québécois singers (Grenier 1993).Since the 1960s, many researchers have documented stigma towards the varieties of Laurentian French spoken in Quebec and in regions throughout the west of Canada (Lambert et al. 1960;Bourhis et al. 1975;Genessee & Holobow 1989;Bouchard 2002;Oakes & Warren 2007;Kircher 2012).Formal varieties of Laurentian French have notably undergone a gradual destigmatization (cf.Lambert et al. 1960;Genessee & Holobow 1989;Kircher 2012).However, informal speech associated with the working class in Quebec remain stigmatized (Corbeil, 1976;Lappin, 1982;Côté 2012).We therefore expect that pop singers in particular may feature non-stigmatized dialectal traits, but that they will not feature stigmatized traits.We anticipate, however, that singers of other genres might feature both types of traits.
However, Québécois singers additionally release albums or perform in France, where the music market is much larger than in Quebec (Grenier 1993).We therefore consider attitudes that speakers of Hexagonal French have towards speakers of Laurentian French.In France, the dialect as a whole is stigmatized, regardless of formal or informal variety, and is often considered a threat to the bon usage of French (Weinstein 1989;Lüdi 1991;Slezák 2015).Because pop singers are more likely to market abroad than singers of other genres, we anticipate pop singers will feature dialectal traits, whether stigmatized or non-stigmatized, at lower rates than singers of other genres.While this study does not directly test the impact that stigma may have on the featuring of dialectal traits in music, we take into account the negative perceptions directed towards the dialect to better understand and predict how genres may vary.In the following section, we detail the target phonological process of the present study: Laurentian French high vowel laxing.

HIGH VOWEL LAXING.
Non-lengthened high vowel laxing in closed final syllables (e.g./vit/ [vɪt] vite 'fast') is a characteristic feature of Laurentian French (e.g.Dumas 1983;Walker 1984, Poliquin 2006;Côté 2012) and distinguishes the dialect from Hexagonal French (Côté & Remysen 2017), given that laxing is categorical in spoken Laurentian French (Poliquin, 2006;Côté 2012;Côté & Remysen 2017).1This process is of particular interest for a study on phonological accommodation not only because the process is categorical in speech, but also because this process is not generally stigmatized by speakers of the dialect (Lappin 1982;Cox 1998;Reinke et al. 2006).To the contrary, should a speaker produce tense variants where lax variants would be anticipated, they may be perceived as arrogant or snobbish (Paradis & Dolbec 1998).However, given that Laurentian French is still viewed negatively by speakers of Hexagonal French (Weinstein 1989;Lüdi 1991;Szlezák 2015), we hypothesize that laxing is also stigmatized by speakers of that dialect.Because this process is non-stigmatized generally non-stigmatized in Quebec, we expect that pop singers will have lower rates of laxing than singers of other genres because pop singers are more likely to interact with international audiences (Grenier 1993).
2.4.SUMMARY.Previous studies on genre and dialectal accommodation suggest pop singers are influenced by norms that are different from those that influence singers of other genres.Given that formal varieties of Laurentian French are relatively destigmatized in Quebec and that the dialect remains stigmatized in France, Laurentian French presents a case with which to expand upon existing work on the dialectal accommodation of Englishes in music.We target high vowel laxing as the target phenomenon in the current analysis because it is a salient dialectal feature whose stigmatization is expected to mirror that of the dialect.In the following section, we describe the methodology used to conduct the present study on the production of dialectal traits based on genre, period in singers' careers, and real-time change in the Quebec music industry.

Methodology.
This study aims to add to existing work on Englishes in the anglophone music industry by examining Laurentian French in the francophone music industry.First, we describe the corpus of Québécois music that we collected for the purposes of this study (Section 3.1).We then present our methodology for extracting high vowels and for coding their tenseness (Section 3.2).Finally, we describe the mixed-effects logistic regression model that we run to address our three research questions (Section 3.3).
3.1.CORPUS.We collected a corpus of music by Québécois artists to test our research questions regarding the patterning of singers as a community, rather than as individuals.To be included in the corpus, a singer had to satisfy four criteria.(1) They must have been born and raised in Quebec.(2) They must be a native speaker of French and speak a variety of Laurentian French, with their dialect confirmed using television interviews or other recorded interactions online.(3) They must release albums either solely or primarily in French.(4) They must have released at least two albums since 2002 and not released any albums before that year.These criteria were developed to ensure the comparability of the artists with respect to the gradual destigmatization of Laurentian French.The corpus includes a total of 20 Québécois artists, eleven women and nine men, and a total of 869 songs in 79 albums (see Table 1 To identify vowels which undergo laxing in speech, the phonemic representations of all words in the corpus were automatically extracted from Lexique 3.81 (New et al., 2004) and then verified manually to ensure the accuracy of transcriptions.Additionally, all borrowings and proper nouns were excluded.Tokens pronounced with a schwa word-finally, as in /abit/ being realized as [abɪtə], were excluded to avoid issues of syllabification, as would occur in spoken language, and a nuanced application of the phonological rule.High vowels followed by a glide were also excluded to ensure reliability of perceptual coding, the methodology described later in this section.
In total, we extracted 4,931 high vowels.Because laxing cannot reliably be categorized on the basis of only one or two acoustic cues (e.g.Arnaud et al. 2011;Sigouin 2013) and because background music could affect the accuracy of formant measurement, tenseness was coded perceptually.The reliability of the main coder was confirmed using Cohen's kappa.Based on a subset consisting of 99 tokens, a high degree of agreement was found between each of the two judges and the main coder (k=0.62 for Judge 1; k=0.75 for Judge 2).
3.3.STATISTICAL ANALYSIS.We ran a mixed-effects logistic regression to predict the rate of laxing using R (R Core Team, 2022).The statistical model includes three independent factors as well as three interactions.We include genre as a factor to test whether pop patterns differently than other genres (our first research question).We also include years since singers' first albums as a factor to test whether the overall community of singers decreases their featuring of dialectal traits over their careers (our second research question).2As our final independent factor, we include year of album release as a main factor to conduct a real-time analysis of laxing in music (our third research question).
As our interactions, we include the interaction between genre and years since singers' first albums to test whether genre may predict the rate and/or trajectory that laxing rates change across careers (our first and second research questions).We also include the interaction between genre and year of album release to test whether genre may predict changes in laxing rate in realtime (our first and third research questions).Finally, we include the interaction between genre, years since singers' first albums, and year of album release to test the effects of genre and change over singers' careers in real time (our first, second, and third research questions).We additionally include singer and word as random intercepts in our model to account for individual variability and lexical effects.

Results.
Our results indicate that genre alone is not a sufficient predictor of laxing rate; however, genre is necessary to predict change over the course of careers and time.In Section 4.1, we present our results based on the second research question regarding the patterning of individuals across their careers and highlight how this is influenced by our first research question regarding distinctions based on genre.In Section 4.2, we present the results based on our third research question regarding how singers pattern in real-time and how this is influenced by our first research question dealing with genre distinctions.Finally, Section 4.3 presents our results from the interaction between genre, years since singers' first albums, and year of album release.4.1.CHANGE OVER SINGERS' CAREERS BY GENRE.The 4,931 tokens in our corpus include 2,753 lax variants (55.83%) and 2,178 tense variants (44.17%).Figure 1 illustrates the results that correspond to our first and second research questions; we see the laxing rates of singers over the course of their careers according to genre.These results reveal a degree of individual variability.Both Cindy Daniel (pop) and Coeur de Pirate (pop) categorically produce tense variants.Additionally, Marc Dupré (pop) increases his laxing rate over the course of his career while all other pop singers decrease or stay the same Singers of other genres tend to increase their laxing rates across time, except for Klô Pelgag (alternative) who decreases.
The model output from our statistical analysis (see Table 2) confirms that pop singers decrease their laxing rates across the course of their careers via the interaction between genre and years since singers' first albums; in the model output, the result for the patterning of pop singers over their careers is reflected in the factor Years since first album (β=-0.5911,p=0.0010).Singers of other genres do not significantly change their rates of laxing over the course of their careers.2 depicts the results that correspond to our first and third research questions; we see laxing rates according to genre over time.As seen in the figure, pop singers produce tense variants more frequently than lax variants throughout the entire time frame in the corpus.Singers of other genres, however, produce lax variants at a rate more similar to the expected rate in speech, although only one album in the corpus features categorical laxing.
The model output in Table 2 (see previous section) does not confirm any significant difference in overall laxing rates between pop and other genres.However, the model output confirms that pop singers marginally increase their rate of laxing in real-time via the interaction between genre and year of album release; the result for the patterning of pop singers in real-time is reflected in the factor Year (β=0.2904,p=0.0900).Singers of other genres do not significantly increase or decrease their rate of laxing over time.3 illustrates the results that respond to all three of our research questions; we see laxing rates according to singers' first albums in real-time by genre.We see that pop singers who released their first album earlier have lower laxing rates than those who released their first album more recently.Additionally, singers of other genres who released their first album earlier do not appear to lax more or less often than those who released their first album more recently.
The model output (see Table 2 in Section 4.1) confirms that pop singers who began releasing albums earlier have lower rates of laxing than those who started doing so more recently via the interaction between genre, years since singers' first album, and year of album release; we see this result in the factor Year, Years since first album, and Pop (β=0.0111,p=0.0022).Singers of other genres who began releasing albums earlier do not have significantly higher or lower rates of laxing than those who released their first album more recently.
Figure 3. Laxing rate of singers' first albums by genre 5. Discussion.We aimed to expand upon previous studies on dialectal traits in music, with a language other than English, by using a new corpus of francophone music from Quebec.Additionally, we analyzed the patterning of singers as a community rather than as individuals and conducted a real-time study to observe change in the industry.Via study of our first research question (Do pop singers exhibit dialectal features as often as singers of other genres?), we find that pop singers do not pattern differently from singers of other genres without the inclusion of other factors.Our results respond to our first and second (Are dialectal features produced at the same rate across singers' careers?) research questions by showing that pop singers' laxing rates typically decrease over the course of the singers' careers.Further, pop singers are distinguishable from other genres based on the tests associated with both our first and third (Does year of album release have an impact on the featuring of dialectal traits) research questions; our results demonstrate that pop singers generally lax less often than singers of other genres.Finally, our results demonstrate that pop singers who released their first album earlier have lower laxing rates than those who released their first album more recently; this is shown via the interaction between the factors which correspond to all three of our research questions.
We first discuss the implications of these results for how the relationship between pop and other genres suggests that pop has a distinct norm that is influenced by Hexagonal French as well as attitudes towards Laurentian French (Section 5.1).We then propose in Section 5.2 that the patterning of pop singers who released their first album earlier as compared to those who released their first album more recently may reflect ongoing effects of dialectal destigmatization in Quebec.

DISTINCT POP MUSIC NORMS.
Based on our analysis of genres in real-time, we see that pop singers have a lower laxing rate than singers of other genres; whereas pop singers produce tense variants more often than lax variants over time, laxing rates are closer to the anticipated rate in speech for singers of other genres.We postulate that differences between pop singers and those of other genres are the result of a phenomenon similar to that in Trudgill (1997) and Simpson (1999), where British singers featured Americanized phonological variants in their music; Gibson (in press, b) shows a similar phenomenon where singers from New Zealand featured Americanized variants for the BATH and nonprevocalic /r/ variables.
Given our results as well as those of Trudgill (1997), Simpson (1999), and Gibson (in press, b), we propose that pop singers, in particular, take on dialectal traits of the dominant dialect in the music industry.For example, in the anglophone music industry the primary dialects are American Englishes, while in the francophone music industry the primary dialect is Hexagonal French.Pop singers are likely more sensitive to dialectal norms than singers of other genres because pop singers appeal to larger, more general audiences than those of other genres (Rojek 2011).We see evidence of this for Québécois singers in particular since pop singers more frequently release their albums and play shows in France than singers of other genres (Grenier 1993).
Our results based on change over singers' careers provide further evidence of pop-specific norms in music; pop singers decrease their rate of laxing over the course of their careers.To achieve long-term career and economic success, singers must appeal to large audiences (Zwaan & ter Bogt, 2009); we therefore postulate that Québécois singers must adjust to norms and expectations set by the majority dialect in the francophone music industry: Hexagonal French.Québécois singers, in particular, are also subject to negative attitudes towards their dialect in France (Weinstein 1989;Lüdi, 1991;Slezák 2015) in ways that British singers do not encounter stigma in the United States (cf.Boberg 1999).Differences in dialectal stigma may then be able to account for why British singers featured dialectal traits more frequently over the course of their careers (Trudgill 1997), while the pop singers in our corpus decrease their usage of dialectal features as they continue releasing albums.
Overall, our results demonstrate that linguistic targets associated with pop in French are distinct from targets associated with other genres; pop singers produce tense variants more often than singers of other genres over time, and pop singers lax less frequently over the course of their careers.Our results suggest the existence of a Pop Song French dialect in music, similar to that of Pop Song English as proposed in Gibson (in press b).However, future research that compares singers of multiple genres from Quebec to speakers of multiple genres from France is necessary to strengthen this particular point.
Although pop singers seem to tend towards Hexagonal French norms, we see they also have moved towards Laurentian French norms; the interaction between genre, years since singers' first albums, and year of album release shows that singers who released their first album more recently feature lax variants more often than those who released their first album earlier.We interpret this finding in the following section as a result of waning destigmatization.

WANING EFFECTS OF DESTIGMATIZATION.
Based on our results depicting real-time change in first albums by genre, we see that pop singers who released their first album earlier have lower laxing rates than those who released their first album more recently.We postulate singers in our corpus who began releasing albums earlier may have been continually impacted by waning effects of stigma towards Laurentian French in Quebec; the dialect underwent a processes of gradual destigmatization in the 1960s in Quebec (cf.Lambert et al. 1960;Genessee & Holobow 1989;Kircher 2012).
We see that the results of Lambert and his colleagues (1960) demonstrated that both anglophones and francophones preferred English to French and suggested that Hexagonal French is preferable to Laurentian French.This reveals stigma was associated with Laurentian French in Quebec, although singers in our corpus were not alive at this time and would not have been subject to these levels of stigma.In the 1980s, Genesee & Holobow (1989) replicated Lambert et al.'s (1960) study to quantify changes in attitudes towards Laurentian French that may have been caused by the passage of 'The Official Language Act' (1974) and 'The Charter of the French Language ' (1977).Attitudes towards French had improved in that neither English nor French was more preferrable.However, Laurentian French speakers were viewed as less intelligent than those of Hexagonal French.
Singers who began releasing albums early on in the time period of our corpus could have been influenced by these negative attitudes towards Laurentian French while growing up, and thereby have negative perceptions towards their own dialect as adults recording albums.Additionally, these singers may have been more strongly influenced by linguistic norms in music during their childhood that were more characteristic of Hexagonal French, given that Laurentian French was stigmatized in both Quebec and France (Weinstein 1989;Lüdi 1991;Szlezák 2015).
Singers who began releasing albums more recently, however, are more likely to have grown up at a time when Laurentian French was relatively non-stigmatized in Quebec.Kircher's (2012) study marked a notable change in positive attitudes towards Laurentian French, thus marking further destigmatization of Laurentian French.Because each subsequent generation faces a decreasing degree of local stigma, pop singers who released their first albums more recently could have been exposed to linguistic norms in music that featured more dialectal traits than those to which previous generations had been exposed.To this end, the destigmatization of Laurentian French may be able to account for why singers who released their first albums more recently lax more frequently than those who did so earlier.However, the present study is limited in this respect given that it cannot describe the norms which existed in the francophone music industry in Quebec during the 20 th century; our corpus only encompasses the years 2002 through 2021.A future study should test for laxing rates over time and over the course of careers in the 20 th century by genre to better understand if those norms could be influencing singers in the 21 st century.

Conclusion.
Whereas previous studies have highlighted various instances where individual singers or small groups, particularly those who sing in English, change which dialectal features appear in their music (e.g.Trudgill 1997;Simpson 1999;Andres Morrissey 2008;Beal 2009;Coupland 2011;Gibson 2011;Gibson & Bell 2012;Eberhardt & Freeman 2015;Lyon 2019), this study analyzes a corpus of 20 Québécois singers who are of the Laurentian French dialect.We analyze change in real-time based on genre and additionally probe how singers tend to pattern over the course of their careers.Results dealing with our first and second research questions demonstrate that pop singers exhibit dialectal traits less often over the course of their careers.Through study of our first research and third research questions, we find that pop singers feature dialectal traits in their music less often than singers of other genres.Based on the results that respond to these two questions, we posit that singers in Quebec follow linguistic norms in singing that are heavily influenced by Hexagonal French pop norms.However, a later study that compares Québécois singers to French singers is necessary to confirm whether or not pop singers in this study are adhering to a norm that exists in France.
Through study of the interaction between factors tested by our three research questions, we find that singers who began their careers more recently produce dialectal variants at higher rates than artists who started releasing albums earlier.We postulate this effect may be due to waning effects of stigma given that Laurentian French underwent a process of gradual destigmatization (cf.Lambert et al. 1960;Genessee & Holobow 1989;Kircher 2012).A future study, however, must be conducted to determine what norms existed in both the Quebec and French music industries in the 20 th century in order to establish whether those norms were more similar to Laurentian French or Hexagonal French.
Although this study expands upon existing literature on English dialectal accommodation in music by focusing on a variety of French, this study is limited in that it only tests one dialect and only concentrates on one dialectal trait.A later study should include other phonological traits characteristic of Laurentian French (e.g.diphthongization and affrication) as well as include singers from France and other varieties of French in the corpus.Additionally, a future study should strive to include singers from genres other than pop who started releasing albums in the period from 2002-2011; this constitutes a gap in the design of the corpus in the current study.
Despite the limitations of this study, it fills several gaps in the existing literature on the featuring of dialectal traits in music.We expand the study of dialectal traits beyond English by using a new corpus of 20 Québécois singers who speak in the Laurentian French dialect.We analyze the patterning of groups of singers, based on genre, across their careers, rather than the patterning of a single singer, and we analyze real-time change of groups of singers, based on genre, as opposed to individuals.

Figure 2 .
Figure 2. Laxing rate across time by genre 4.3.REAL-TIME CHANGE IN FIRST ALBUMS BY GENRE.Figure3illustrates the results that respond to all three of our research questions; we see laxing rates according to singers' first albums in real-time by genre.We see that pop singers who released their first album earlier have lower laxing rates than those who released their first album more recently.Additionally, singers of other genres who released their first album earlier do not appear to lax more or less often than those who released their first album more recently.The model output (see Table2in Section 4.1) confirms that pop singers who began releasing albums earlier have lower rates of laxing than those who started doing so more recently via the interaction between genre, years since singers' first album, and year of album release; we see this result in the factor Year, Years since first album, and Pop (β=0.0111,p=0.0022).Singers of other genres who began releasing albums earlier do not have significantly higher or lower rates of laxing than those who released their first album more recently. Figure

Table 1 .
). Eleven are pop singers and nine are singers of other genres.Singers in the corpus 3.2.EXTRACTION AND CODING.This study analyzes high vowel laxing in closed final syllables.