A Crosslinguistic Account on Speech Information Rate

François Pellegrino - Dynamics of Language Lab & Laboratory of Excellence ASLAN (CNRS & University of Lyon, France)

A Crosslinguistic Account on Speech Information Rate

François Pellegrino
Dynamics of Language Lab & Laboratory of Excellence ASLAN
(CNRS & University of Lyon, France)

Each language system exhibits some specificities on what is important, mandatory, or conversely superfluous to explicitly encode/decode in speech communication. One language would consider that kind of social relationship between the speaker and hearer is essential to encode, while another would rather consider that the physical configuration between them is highly important. Such examples can be listed with no limits, illustrating that what is a relevant information may dramatically differ among languages. However, once it turns to the capacity to pack such information into a speech stream, one can ask whether such large differences are still salient or whether a convergence occurs, because of functional motivations. In this talk I address issues related to the notion of speech information rate and the hypothetical existence of a regulation of information rate from one language to another. Starting from the existence of cross-linguistic differences in speech rate (in terms of syllables uttered per second), I examine how information rate can be defined using the syllable as an information brick and show an assessment based on 15 languages. I then discuss these results in a broader picture based on perceptual (speech rate perception) and cognitive (connection between speech rate and brain oscillatory rhythms) considerations.

If you would like to meet with the speaker, please contact Bernd Möbius.

François Pellegrino

I have been a CNRS Senior Researcher in Linguistics (DR2 CNRS) since  2009 and the Director of the laboratory DDL ‘Dynamique Du Langage’ in Lyon from 2004 to 2012. I am also the coordinator of the ‘Laboratory of Excellence’ ASLAN (2011-2019, Advanced studies on language complexity). I received a Master Degree in electronic engineering and signal processing from the ‘Institut National Polytechnique de Toulouse’ in 1994 and a Ph.D. in Computer science from the University of Toulouse (France) in 1998. I defended my ‘habilitation à diriger des recherches’ on language identification and phonological complexity in 2009. I joined DDL in Lyon in 1999 and led its “Language Identification” group for seven years. My research focuses primarily on speech communication and my interests range from the automatic modelling of rhythm to the study of the structure and dynamics of phonological systems in the light of the science of complexity. I am also involved in the study of the interaction between neurocognitive, perceptual and phonetic processes during the comprehension of degraded speech. Over the last dozen years, I have authored or co-authored about 15 journal papers or book chapters and about 65 papers in the proceedings of national and international conferences in speech and signal processing, linguistics and cognition. In 2013, I was the local chairman (General Co-Chair) of the Interspeech 2013 conference (in Lyon, 25-29 August).