Navigation auf uzh.ch
Navigation auf uzh.ch
The Crosslinguistic Perspectives on Processing and Learning Workshop (X-PPL) brings together the growing community of researchers working to expand the diversity of languages in the scope of experimental or corpus research on adults or language acquisition. This research is driven by the recognition that structural / typological and socio-cultural diversity represents different opportunities to see processing and learning mechanisms at work. The bulk of processing and acquisition research represents only a small fraction of linguistic diversity, and this risks skewing both our theories and research questions.
The Crosslinguistic Perspectives on Processing and Learning Workshop (X-PPL) aims to fill this gap and provide a platform for cross-linguistic research on language processing and learning. X-PPL 2024 will take place on September 2-3, 2024.
A key aim of psycholinguistics is to uncover the nature of linguistic representations and processes that subserve comprehension and production across typologically distinct languages. Do typologically distinct languages (e.g., SOV vs SVO languages) employ different processing strategies? Can certain processing strategies or constraints be deemed universal that hold across languages? In this talk I weigh in on these questions by presenting new data on various aspects of processing in an SOV language Hindi.
From a comprehension perspective, I will present data on syntactic prediction and syntactic
priming. From a production perspective, I will present data on word order choices. The results highlight processing tendencies that can be understood as adaptations driven by the typology of the language — (a) there is robust prediction of the clause final verb during sentence comprehension, (b) syntactic priming during comprehension is constrained by parser behaviour, (c) there is a tendency for formulating a long-before-short word order during sentence production.
Interestingly, the results also provide evidence that these processes are influenced by working-memory constraints that can be deemed universal — (a) robust prediction of the clause final verb during sentence comprehension decreases when the preverbal linguistic complexity increases, (b) syntactic priming fails when the prime is relatively more complex while the target can be parsed using a simpler syntactic structure, (c) in addition to a long-before-short word order, a significant amount of short-before-long word order is also visible during sentence production.
Together, these results shed new light on processing of SOV languages. The results highlight how processing strategies are shaped by both typologically-driven features as well as by universal cognitive constraints.
Traditional laboratory experiments have significantly advanced our understanding of speech and language use, but a reliance on controlled environments limits our ability to ensure observations generalize to real-world communication. In this talk, I summarize some recent work on young children's speech development to introduce a novel approach that leverages long-form recordings from real-world interactions and machine learning to uncover the dynamics of speech behavior as it naturally occurs. For instance, in a 13-author collaborative study, we analyzed over 40,000 hours of audio from 1,001 children across 12 countries, revealing maturation and speech exposure as more important predictors of infants' speech development than gender and socioeconomic status. Furthermore, we use state-of-the-art self-supervised learning models to argue that tailored biases are needed to face the rich variability of naturalistic audio. This talk aims to foster an interdisciplinary exchange on leveraging wearable technology to study language and communication, highlighting both the breakthroughs and the hurdles of venturing beyond the lab.
