Learning to predict and predicting to learn: Before and beyond the syntactic bootstrapper

Abstract

Young children can exploit the syntactic context of a novel word to narrow down itsprobable meaning. This is syntactic bootstrapping. A learner that uses syntacticbootstrapping to foster lexical acquisition must first have identified the semanticinformation that a syntactic context provides. Based on the semantic seed hypothesis,children discover the semantic predictiveness of syntactic contexts by tracking thedistribution of familiar words. We propose that these learning mechanisms relate to a largercognitive model: the predictive processing framework. According to this model, weperceive and make sense of the world by constantly predicting what will happen next in aprobabilistic fashion. We outline evidence that prediction operates within languageacquisition, and show how this framework helps us understand the way lexical knowledgerefines syntactic predictions and how syntactic knowledge refines predictions about novelwords’ meanings. The predictive processing framework entails that learners can adapt torecent information and update their linguistic model. Here we review some of the recentexperimental work showing that the type of prediction preschool children make from asyntactic context can change when they are presented with convincing contrary evidencefrom recent input. We end by discussing some challenges of applying the predictiveprocessing framework to syntactic bootstrapping and propose new avenues to investigatein future work.

Introducing Meta‐analysis in the Evaluation of Computational Models of Infant Language Development

“…(2021)). Central to modeling purposes, APC learning criterion aligns with the idea of predictive processing in human perception (e.g., Friston, 2010; Rao & Ballard, 1999; Babineau, Havron, Dautriche, de Carvalho, & Christophe, 2022) and a similar learning mechanism can thereby be assumed to be available to any mammalian learner. More specifically, APC learning is based on minimizing prediction error of future speech observations, given access to past and current speech stimulus.…”

“…The latent representations APC learns have been shown to be applicable to a range of speech tasks, including phone classification (Chung et al, 2019) and subword modeling (Feng, Żelasko, Moro-Velázquez, & Scharenborg, 2021; see also Yang et al (2021)). Central to modeling purposes, APC learning criterion aligns with the idea of predictive processing in human perception (e.g., Friston, 2010;Rao & Ballard, 1999;Babineau, Havron, Dautriche, de Carvalho, & Christophe, 2022) and a similar learning mechanism can thereby be assumed to be available to any mammalian learner. More specifically, APC learning is based on minimizing prediction error of future speech observations, given access to past and current speech stimulus.…”

Communicative Feedback in language acquisition

“…However, more recent theories on language acquisition do highlight synergies when learning form and meaning (e.g., Abend et al, 2017;Babineau et al, 2022;Christophe et al, 2008;Dupoux, 2018;Feldman et al, 2013;Fourtassi et al, 2020;Landau & Gleitman, 1985;Räsänen & Rasilo, 2015), as well as when leveraging information about how language is used in context to learn various linguistic structures (Bohn & Frank, 2019;E. V. Clark, 2016E.…”

“…Such error-driven learning mechanisms have been proposed to play a major role in human learning more generally (A. Clark, 2015;Friston, 2009) and are increasingly being applied to language acquisition (Babineau et al, 2022;Cox et al, 2020).…”

Communicative Feedback in Language Acquisition

“…However, more recent theories on language acquisition do highlight synergies when learning form and meaning (e.g., Abend et al, 2017;Babineau et al, 2022;Christophe et al, 2008;Dupoux, 2018;Feldman et al, 2013;Fourtassi et al, 2020;Landau & Gleitman, 1985;Räsänen & Rasilo, 2015), as well as when leveraging information about how language is used in context to learn various linguistic structures (Bohn & Frank, 2019;E. V. Clark, 2016E.…”