Special sessions

Special Session 1 (Satellite Event): Sociophonetic variability in the English varieties of Australia (SocioPhonAus4)

A satellite SocioPhonAus4 event will be held as part of SST2024 with a dedicated oral session and a keynote speaker.

Organisers:

Gerry Docherty (Griffith University)

Janet Fletcher (University of Melbourne)

Debbie Loakes (University of Melbourne)

Information about SocioPhonAus4, including invitation for paper submissions, can be viewed here.

Special Session 2: The phonetic expression of phonological length

Organisers:

Francesco Burroni (Ludwig Maximilian University of Munich)

John Hajek (University of Melbourne)

Angelo Dian (University of Melbourne)

Many languages utilise distinctions between short and long consonants and/or vowels to convey lexical differences. For instance, Italian contrasts /papa/ “Pope” with /papːa/ “baby food”, and Thai contrasts /pàk/ “to stick in” with /pàːk/ “mouth”. These differences, known as ‘length contrasts’, are found across diverse language families including Indo-European, Japonic, Austronesian, Afroasiatic, Turkic, Otomanguean, and non-Pama-Nyungan languages.

However, many aspects of how these length contrasts are realised phonetically remain debated. Questions arise about whether speakers focus on relative rather than absolute durational properties within words, if non-durational acoustic and articulatory dimensions are used in language- or dialect-specific ways to enhance length contrasts, and how vowel and consonant length interact, as shown by languages of Australia and Europe among others. Experimental research in this area, especially for Australian languages, is limited.

Scholars are invited to contribute to exploring phonological length and its phonetic realisation. Submissions are encouraged to delve into linguistic, acoustic, perceptual, and physiological perspectives on length contrasts, as well as investigate their implications for language processing and evolution. Through interdisciplinary collaboration, this thematic session aims to advance understanding of phonological length phenomena and contribute to theoretical and empirical insights in phonetics and related disciplines.

Special Session 3: Assisting the courts to determine who says what in forensic speech recordings

Organisers:

ASSTA Forensic Speech Science Committee (FSSC)

Led by Yuko Kinoshita (Australian National University) and Shunichi Ishihara (Australian National University)

Assisted by Michael Carne (Australian National University) and Eleanor Kettle (University of Melbourne)

ASSTA’s Forensic Speech Science Committee (FSSC) presents this special session with the aim of broadening the base of forensic speech science research in our region. We would like more researchers from the ASSTA community to join in discussions on how speech science knowledge can be applied to forensic contexts.

Many criminal trials feature forensic evidence in the form of recordings of spoken language, often of poor quality. Traditional legal procedures for determining who is speaking and what they are saying, developed without input from the language sciences, have been shown to be problematic in various ways, creating actual and potential injustice. Speech science researchers, local and international, have made many useful contributions to improving this situation – but there remains much to be done. Importantly, since forensic speech science asks researchers to address questions that have not traditionally been a major focus in our field, this work has highlighted significant gaps in our collective knowledge about spoken language, some of which are only starting to be filled.   

Presentations are welcome from any branch of speech science with potential relevance to forensic questions, including but not limited to: sociophonetics, speaker characteristics, forensic methodology, signal processing and enhancement, audiology, second language pronunciation, automatic speech and speaker recognition, spoken language processing.

Special Session 4: ASR for better reading education: an interdisciplinary challenge

Organisers: 

Catia Cucchiarini (Radboud University Nijmegen)

Helmer Strik (Radboud University Nijmegen)

Cristian Tejedor García (Radboud University Nijmegen)

Wieke Harmsen (Radboud University Nijmegen)

Learning to read constitutes one of the most fundamental skills children acquire at school, but recent international surveys indicate decreasing levels of reading proficiency in many countries. This calls for innovative solutions that can contribute to making learning to read more efficient and effective within the limitations of current school practice. Automatic Speech Recognition (ASR) technology holds great potential for evaluating reading aloud and possibly providing immediate feedback during reading practice. In addition, systems that incorporate ASR technology enable research to the underlying processes of reading acquisition.

Although research in this direction was initiated many years ago and many commercial devices nowadays incorporate ASR technology, employing ASR in child reading education is still problematic for several reasons. First, despite the impressive recent advancements in ASR, current systems are not yet ready for child speech reading diagnostics. Second, dedicated procedures are required to translate ASR information into interactive user-friendly student-teacher dashboards. Third, important information on prosody and fluency is often neglected. Fourth, privacy and security issues need to be addressed as speech intrinsically contains information on identity and other personal features.

In this special session, we intend to address this interdisciplinary line of research bringing together experts from different disciplines.