Rhythmic Proximity Between Natives And Learners Of French - Evaluation of a metric based on the CEFC corpus

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 2497–2502 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 2497 Rhythmic Proximity Between Natives And Learners Of French Evaluation of a metric based on the CEFC corpus Sylvain Coulangea, Solange Rossatoa,b a LIDILEM, b LIG Université Grenoble Alpes {sylvain.coulange, solange.rossato}@univ-grenoble-alpes.fr Abstract This work aims to better understand the role of rhythm in foreign accent, and its modelling. We made a model of rhythm in French taking into account its variability, thanks to the Corpus pour l’Étude du Français Contemporain (CEFC), which contains up to 300 hours of speech of a wide variety of speaker profiles and situations. 16 parameters were computed, each of them being based on segment duration, such as voicing and intersyllabic timing. All the parameters are fully automatically detected from signal, without ASR or transcription. A gaussian mixture model was trained on 1,340 native speakers of French; any 30-second minimum speech may be computed to get the probability of its belonging to this model. We tested it with 146 test native speakers (NS), 37 non-native speakers (NNS) from the same corpus, and 29 non-native Japanese learners of French (JpNNS) from an independent corpus. The probability of NNS having inferior log-likelihood to NS was only a tendency (p=.067), maybe due to the heterogeneity of French proficiency of the speakers; but a much bigger probability was obtained for JpNNS (p<.0001), where all speakers were A2 level. Eta-squared test showed that most efficient parameters were intersyllabic mean duration and variation coefficient, along with speech rate for NNS; and speech rate and phonation ratio for JpNNS. Keywords: Rhythm modelling, Speech fluency, Foreign accent, Spoken language processing, L2 pronunciation, Variability of speech 1. Introduction Perception of foreign accent is mainly due to a dif- ference of pronunciation between what a speaker said and a norm shared by natives of a target language (Alazard, 2013). This difference has been mostly de- scribed on the segmental level, through theories of lan- guage acquisition, but the role of prosody was proved only recently in accent perception studies (De Meo et al., 2012; Pellegrino, 2012). Today, it is clear that both segmental and suprasegmental levels are important in perception of this difference by native speakers. Among prosodic parameters, rhythm is one that varies noticeably from one language to another. (Gibbon and Gut, 2001) define rhythm as the recurrence of patterns of weak and strong elements in time. Those elements can be short or long syllables, low or high intonation on vowel or consonant segments. (Di Cristo and Hirst, ) define it as the temporal organisation of prominences. (Arvaniti, 2009) talks about perception of repetition and similarity patterns. (Alazard, 2013) refers to phys- ical and psychobiological phenomena like dance and music rhythm, or cardiac rhythm. All these defini- tions put forward the idea of timing patterns; length, height or intensity patterns in time. Many studies try to classify languages based exclu- sively on rhythmic parameters. Most of them rely …

Content truncated to 500 words