Transdisciplinary & innovation Grants announced for 2017
The following projects have been awarded funding this round:
Project: Sociophonetic variation in first- and second-generation Australians’ English and community language
Ksenia Gnevsheva, Australian National University
The proposed project investigates how first- and second-generation bilingual immigrants vary in their use of English and community language. Documenting how different language varieties interact is important because such interactions shape the Australian linguistic landscape, and it helps us understand how multiple languages work within one individual. I will apply variationist methods and employ innovative data collection and analysis methods to build a unique spoken bilingual corpus, and I will run a perception experiment which will tap into the effect of different linguistic and social variables on bilinguals’ language processing.
Project: Brainwaves and clause chains: Testing EEG in PNG
Hannah Sarvasy, Australian National University
Models of sentence processing have largely relied on experiments with speakers of a small sample of familiar languages—English, Spanish, Dutch, etc. Hundreds of other languages, however, involve multi-clausal sentences—‘clause chains’—that differ syntactically from English sentences. Speakers likely also process clause chains differently. This project will use mobile electroencephalography (EEG) data from speakers of the clause chaining language Nungon in remote Papua New Guinea to establish how the brain processes clause chains. Of special interest are: correlates of working memory, syntactic agreement, and attentional switching. The project will give strong evidence on how hierarchical structures shape language processing.
Project: Building a social robot for the Ngukurr Language Centre
Gautier Durantin, University of Queensland
This project is a new collaboration between CoEDL social robotics lab and the Ngukurr language centre in the Northern Territory to design and evaluate new technologies for language teaching and recording. The project will involve the collaborative design and development of a social robot capable of interacting with children in Ngukurr (Northern Territory) through language games, as well as the development of tools for community language workers to create tablet-based language teaching materials easily. The project takes CoEDL technology research into the field, and seeks to develop and evaluate engaging ways of teaching and preserving indigenous languages.
Project: Using Brain-Computer Interfaces to measure, predict and boost individual language learning
Alba Tuninetti, Western Sydney University
Tailoring individual neural technology has long been the playground of prosthetic engineering and brain-computer interfaces (BCIs). Such interfaces use electrical signals from the brain to control a device (active BCIs) or estimate the mental state of the user (passive BCIs). The rise of passive BCIs brings new opportunities for language learning, as predicting an individual’s capabilities for language learning could be used to adapt current language-learning technologies. Our study will combine individuals’ neural measurements in language perception and production to predict individual language learning. We will analyse EEG and behavioural data concurrently, focusing on within-participant analyses to develop adaptive technologies.
Project: Contingency in infant-directed speech: Neurophysiological and psycho-social responses in pre-linguistic infants
Christa Lam-Cassettari, Western Sydney University
Recent research has shown that the quality and quantity of speech provided to young infants has long lasting effects on child language learning and processing. Whether the quality or quantity of speech play a more powerful role are yet to be ascertained. This study investigates how the quality of speech affects infant responses in the pre-linguistic period, by comparing infant responses to contingent versus non-contingent infant-directed speech. It is expected that contingent speech will elicit more advanced neurophysiological and psycho-social responses than non-contingent speech in infant and adult listeners.
Project: Are super-complex words represented like sentences in speakers' minds?
Brett Baker, University of Melbourne & Rikke Bundgaard-Nielsen, Western Sydney University
Many Australian languages defy a seemingly straightforward distinction between 'words' and 'phrases' (groups of words acting as a unit of meaning) by having very complex words with meanings that English would require a sentence to express. This project aims to examine speakers' knowledge of ‘super-complex’ words in the Australian language Wubuy, using innovative 'low tech' experimental approaches suited for remote field research. The results will experimentally investigate what a 'word' is in Wubuy, and thereby inform not only language description/typology but also likely challenge current models of speech processing and language acquisition, which are built on this fundamental distinction.
Project: Look who’s talking: Perceived ethnic background of speaker impacts language learning
Karen Mulak, Western Sydney University
Preliminary data revealed that Caucasian participants showed poorer word learning when told novel Dutch words were spoken by someone of Middle Eastern than Caucasian appearance, despite everyone hearing the same stimuli. The possibility that knowledge about the speaker’s ethnic appearance affects language learning has clear application to understanding multicultural communication. To determine whether this stems from knowledge of European versus Middle Eastern languages, or general or specific out-group biases, we will compare language learning from perceived narrators of different ethnic backgrounds by AusE-speakers of different backgrounds, and a population from PNG with limited exposure to people not indigenous to PNG.
Project: ‘I can hear that language but I can’t speak it’: assessing receptive competence in Mawng at Warruwi Community (Australia)
Ruth Singer, University of Melbourne
‘Do you speak French?’ we might be asked, to assess our knowledge of French. The word ‘speak’ is used as shorthand for linguistic knowledge. However it is possible understand a language well without being able to speak it. Hidden language knowledge like this plays an important role in many multilingual settings. At Warruwi Community in Arnhem Land, people often understand a number of Indigenous languages very well, in addition to those they can speak. We know very little about what kind of linguistic knowledge people have, who only ‘hear’ a language. This research will help us find out more about how the brain processes language and how languages change over time.
Project: A Pan-Australian Model for the Transcription Acceleration of Australian Indigenous Languages
Hywel Stoakes, University of Melbourne
The aim of this project is to adapt existing software tools that can segment the Indigenous languages of Australia making recorded speech searchable. This tool will greatly assist with and accelerate linguistic analysis of languages which are endangered and disappearing. With the rapid increase in data generation, spurred on by cheaper and more available digital storage, there is a growing need for methods that allow searching and cataloguing speech data. There have been many tools available for languages with many speakers and many centuries of linguistic study e.g. English and German. There are fewer tools for languages which are endangered.