ChiSense-12: An English Sense-Annotated Child-Directed Speech Corpus

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 5198–5205 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 5198 ChiSense-12: An English Sense-Annotated Child-Directed Speech Corpus Francesco Cabiddu1, Lewis Bott1, Gary Jones2, Chiara Gambi1 1Cardiff University, 2Nottingham Trent University {cabidduf, bottla, gambic}@cardiff.ac.uk, gary.jones@ntu.ac.uk Abstract Language acquisition research has benefitted from the use of annotated corpora of child-directed speech to examine key questions about how children learn and process language in real-world contexts. However, a lack of sense-annotated corpora has limited investigations of child word sense disambiguation in naturalistic contexts. In this work, we sense-tagged 53 corpora of American and English speech directed to 958 target children up to 59 months of age, comprising a large-scale sample of 15,581 utterances for 12 ambiguous words. Importantly, we carefully selected target senses that we know - from previous investigations - young children understand. As such work was part of a project focused on investigating the role of verbs in child word sense disambiguation, we additionally coded for verb instances which took a target ambiguous word as verb object. We present experimental work where we leveraged our sense-tagged corpus ChiSense-12 to examine the role of verb-event structure in child word sense disambiguation, and we outline our plan to use Transformer-based computational architectures to test hypotheses on the role of different learning mechanisms underlying children word sense disambiguation performance. Keywords: Word Sense Disambiguation, Verb-Event Structure, Child-Directed Speech 1. Introduction Although theories of word learning predict that young children do not map word forms to multiple meanings (Markman, 1989; Trueswell et al., 2013), recent evidence suggests that the speech children hear early in development is rich in word sense ambiguity, and also that children’s early vocabularies are populated by ambiguous words (Meylan et al., 2021). This raises a number of questions: Which environmental factors and which learning mechanisms might help children deal with sense diversity in real-life contexts? Developing annotated corpora is crucial to answering such questions: developmental researchers can test the effect of naturalistic variables by constructing experimental stimuli based on corpus statistics, and they can implement and test the plausibility of hypothesized learning mechanisms by using computational architectures trained on naturalistic conversations (Monaghan and Rowland, 2017). In this work, we focus on the role that verbs might play in word sense disambiguation during sentence processing. Verbs play a key role in different aspects of language acquisition (unambiguous word processing, Mani, Daum, and Huettig, 2016; syntactic ambiguity resolution, Kidd, and Bavin, 2005), but their role has received less attention in child word sense disambiguation. To help fill this gap, we share a large corpus of child-directed speech, ChiSence- 12, where we tagged all verb-sense occurrences of 12 ambiguous words used in previous developmental studies. The corpus can be freely downloaded at https://gitlab.com/francescocabiddu/chisense-12. We present preliminary analyses of experimental work based on ChiSense-12, and outline our plans to train Transformer- based computational architectures on this corpus to test hypotheses about the learning mechanisms involved in learning ambiguous …

Content truncated to 500 words