Professor Daniel Hirst
ISCA Distinguished Lecturer, Directeur de Recherche Emeritus, CNRS & Aix-Marseille University, France
Daniel Hirst is a British linguist and phonetician, who lives and works in the South of France. He has been working in the field of speech prosody and phonology for nearly fifty years. After a PhD and a Habilitation thesis, both on intonation and prosody, he became a researcher at the LPL speech and language research laboratory for the CNRS (French National Scientific Research Centre) and Aix-Marseille University in Aix en Provence until 2016. From 2012 to 2015 he was also Lecture Professor at Tongji University, Shanghai, China. He is currently Emeritus Research Director for the CNRS in Aix-en-Provence.
He has published numerous articles in several major journals and has contributed chapters to several international volumes and was responsible in 1998 for the edition of a major study of the intonation of languages of the world.
He has also developed software for the automatic analysis of speech melody and rhythm, in particular Momel, Intsint and ProZed, which have all been implemented as plugins for the Praat speech analysis software and are all freely available.
In 2002 he organised the first International Conference on Speech Prosody, in Aix en Provence, France. Since then Speech Prosody has become a regular international conference, held every two years (Japan 2004, Germany 2006, Brazil 2008, USA 2010, China 2012, Ireland 2014, USA 2016, Poland 2018).
He is also the chief editor of a collection of books published by Springer, entitled Prosody, Phonology and Phonetics, which publishes monographs and collections of papers on the subject of Speech Prosody.
In 2013 he was elected fellow of ISCA and in 2014 member of the Permanent Council for Organisation of ICPhS. This year he was elected ISCA distinguished lecturer for 2018-2019.
The automatic annotation, display and comparison of speech prosody.
The automatic display, annotation and evaluation of speech prosody.
The automatic annotation of speech prosody is an indispensable step towards better understanding the prosodic structure of languages.
The last few decades has witnessed a search for “prosodic metrics”, measurements derived automatically from the acoustic signal, which can be used for the typological classification of languages in the domain of speech prosody, in particular for rhythm and melody.
This presentation will discuss the general question of the automatic analysis of speech melody, with results from studies using melody metrics for the discrimination of languages, applied to a mutilingual corpus, OMProDat, which contains comparable recordings of read speech in English, French, Mandarin Chinese, and Cantonese as well as L2 recordings of English by both French and Chinese speakers.
Some specific proposals will be made for the automatic visualisation of speech prosody which could be used to provide a visual and auditory feedback as an aid to improve the prosody of L2 speakers.
Professor Yukio Tono, PhD
Tokyo University of Foreign Studies
Dr. Yukio Tono is Professor in Corpus Linguistics and Director of the World Language Center at Tokyo University of Foreign Studies (TUFS). He received his PhD from Lancaster University, UK. His research interests include corpus-based Second Language Acquisition, corpus applications in foreign language teaching, and corpus-informed teaching materials and resources development. He is the project leader of the CEFR-J, an adapted version of the Common European Framework in Japan. He is also President of the Japan Association for English Corpus Studies. He serves as the editorial board members of international journals such as International Journal of Lexicography (Oxford University Press), Corpora (Edinburgh University Press), and Studies in Corpus Linguistics (John Benjamins). He is the author of Corpus-Based Language Studies (with Tony McEnery and Richard Xiao, Routledge, 2006), The Frequency Dictionary of Japanese (with Makoto Yamazaki and Kikuo Maekawa, Routledge, 2013) and Developmental and Crosslinguistic Perspectives in Learner Corpus Research (edited with Yuji Kawaguchi and Makoto Minegishi, John Benjamins, 2012). He is currently working on the CEFR-J x 27 project, where he applies the method of developing pedagogical resources based on the CEFR-J for 27 foreign languages taught at TUFS.
L2 learner profiling research and its application for multilingual pedagogical resource development
Recently there is a growing interest in analyzing second language (L2) learners’ learning processes based upon the Common European Framework of Reference for Languages (CEFR). Methodologies used for profiling research have been increasingly data-intensive. For instance, machine learning techniques have been used to identify linguistic features which serve as criteria for discriminating one CEFR level from the other. In Japan, I have been involved in a large-scale profiling research project called the CEFR-J, in which we developed our own framework based on the CEFR, which is dedicated to English language teaching in our local context and prepared all the related teaching resources for the CEFR-J. I will discuss the processes of such implementation, especially focusing on the development of the CEFR-J Wordlist, the CEFR-J Grammar Profile, and the CEFR-J Text Profile. After the preparation of the teaching resources for English, we went on to start developing the resources for other foreign languages taught at TUFS, including many Asian languages such as Chinese, Korean, Mongolian, Indonesian, Malaysia, Filipino, Thai, Laotian, Vietnamese, Cambodian, Burmese, Urdu, Hindi, and Bengali. This involves the translation of English resources onto various languages using machine translation as well as human intervention. I will discuss the value of creating multilingual pedagogical resources based on the CEFR and some issues of using NLP technologies for creating such resources semi-automatically.