Identification of regional French accents in (northern) France, Belgium, and Switzerland

Abstract

This paper investigates whether European Francophone subjects are able to distinguish between regional French accents from (northern) France, Belgium, and Switzerland, and at what level of granularity. In total, samples from 120 speakers (from five different areas in each country under study) were presented to hundreds of native French listeners from these three countries. In a first set of experiments, listeners were asked to identify the speakers’ country of origin: they achieved 60% correct identification on average, with significant effects of listeners’ region of origin, speakers’ age, socioeconomic status, and region of origin. In a second set of experiments, listeners from Belgium, France, and Switzerland were asked to identify the speakers’ region of origin within each country (5-alternative forced choice). Results, albeit above chance, proved to be poorer than they were in the first set of experiments (31% correct identification on average). Complementary analyses were conducted to evaluate the role of listeners’ region of origin, speakers’ age, speakers’ region of origin, and their interaction. They showed asymmetrical response patterns across the three countries under investigation: France (or, within France, Paris, which represents the norm) seems to act as a magnet and a catalyst of unification. Younger generations, especially, are more often associated with its way of speaking when their accent is not clearly identifiable. Switzerland, though, resists this homogenizing process better than Belgium does.

The Role of Acoustic Distance and Sociolinguistic Knowledge in Dialect Identification

“…The experiment also provided more general insights into the process of dialect identification. At first sight, listeners appeared to better identify their own as opposed to the other dialect, as reported in several dialect recognition studies on other languages ( Williams et al, 1999 ; Gooskens, 2005 ; Boomershine, 2006 ; Baker et al, 2009 ; Yan, 2015 ; Avanzi and Boula de Mareüil, 2017 ). Our results and, in particular, the use of signal detection theory, offer an alternative interpretation.…”

“…However, there is also some evidence for variation in the own-dialect response bias. Avanzi and Boula de Mareüil (2017) , who compared the effect of speaker and listener dialect on accent identification in French, found that the tendency to better identify one’s own as opposed to other dialects was more marked in Swiss than in Belgian and French participants. Although this interpretation remains speculative, it is possible that the federal, non-centralistic political organisation of Switzerland, together with the lack of an oral standard variety with high prestige, further enhances this own-dialect bias.…”

“…Not surprisingly, previous experience with the tested dialects, for instance, through moving or traveling, facilitates their recognition ( Clopper and Pisoni, 2004a ; Díaz Campos and Navarro Galisteo, 2009 ). The finding that listeners more easily identified their own over other dialects ( Williams et al, 1999 ; Gooskens, 2005 ; Boomershine, 2006 ; Baker et al, 2009 ; Yan, 2015 ; Avanzi and Boula de Mareüil, 2017 ) has also been explained with exposure and experience ( Clopper and Pisoni, 2004a ; Baker et al, 2009 ; Yan, 2015 ). An alternative explanation, however, is that an own-dialect bias accounts for the higher accuracy (see Ruch , in press ), similar to an own-age bias in the recognition of speaker age ( Moyse et al, 2014 ).…”

“…Apart from a potential own-dialect bias, listeners have been found to ascribe ambiguous stimuli to the most prestigious variety that was offered as an answer. For instance, in experiments on American English varieties ( Purnell et al, 1999 ; Perrachione et al, 2010 ), Mandarin dialects ( Yan, 2015 ), and French regional accents ( Avanzi and Boula de Mareüil, 2017 ), listeners were biased towards ascribing most stimuli to Caucasian American English, Enshi Mandarin, and Parisian French, respectively. Furthermore, regional varieties that are present in the media seem to be identified more easily, especially when they appear in a clearly identifiable regional context, for instance, in the speech of well-known public persons ( Lameli et al, 2008 ; Purschke, 2012 ).…”

Peut-on distinguer perceptivement huit accents régionaux en français parlé en Europe ? Une réponse à base de crowdsourcing

“…Une méthodologie à base de crowdsourcing (production participative à grande échelle) a depuis permis de passer à quelques centaines voire quelques milliers d'auditeurs, auxquels étaient présentés des accents français d'Amérique du Nord, d'Afrique et d'Europe (Boula de Mareüil et al, 2017), mais chaque variété était représentée par un seul locuteur. D'autres expériences ont été conduites, centrées sur la France, la Suisse et la Belgique, touchant également une large audience, à travers des sites grand public ou les réseaux sociaux (Avanzi & Boula de Mareüil, 2017). Au total, des échantillons (d'une durée comparable aux tests précédents, soit une douzaine de secondes), provenant de 120 locuteurs (de 5 régions différentes dans chacun de ces pays) ont été présentés à des centaines d'auditeurs de ces mêmes pays.…”