| 研究実績の概要 |
The present project aimed to create an infrastructure to measure and analyze infants social-communicative environment across cultures. Theories and data on language acquisition suggest a range of cues are used, ranging from information on structure found in the linguistic signal itself, to information gleaned from the environmental context or through social interaction. We developed a framework to extract such data from the input in a unified way and to link it to known learning algorithms in order to make explicit the connection between the kinds of information available to the social learner and the computational mechanisms required to extract language-relevant information and learn from it. As a second step, we proposed to collect infant data and apply our framework. Although we were able to do some pilot data collection, due to the pandemic, we were not able to collect a full dataset in Japan. Instead, we moved on to the cross-linguistic part of the project and applied our framework to Tseltal, a language spoken by a hunter-gatherer society.. These data reveal a high quantity of input that could be used to extract language structure, but less input that would permit the learner to link this structure to events or entities in the surrounding world. Together, this approach will allow us to make precise recommendations for future large-scale empirical research.
|
| 備考 |
Tsuji, S., Cristia, A., & Dupoux, E. (under review). SCALa: A blueprint for computational models of language acquisition in social context.
Havron, N., & Lovcevic, I., & Tsuji, S. (in preparation).
|
