Reframing linguistic bootstrapping as joint inference using visually-grounded grammar induction models
This reframes a foundational problem in cognitive science and AI by showing how joint inference can ease language acquisition in constrained settings, though it is incremental in method.
The paper argues that syntactic and semantic bootstrapping in language acquisition are not separate strategies but result from joint learning, demonstrating through neural visually-grounded grammar induction models that simultaneous learning of syntax and semantics improves grammar induction, lexical category learning, and interpretation of novel sentences and verbs.
Semantic and syntactic bootstrapping posit that children use their prior knowledge of one linguistic domain, say syntactic relations, to help later acquire another, such as the meanings of new words. Empirical results supporting both theories may tempt us to believe that these are different learning strategies, where one may precede the other. Here, we argue that they are instead both contingent on a more general learning strategy for language acquisition: joint learning. Using a series of neural visually-grounded grammar induction models, we demonstrate that both syntactic and semantic bootstrapping effects are strongest when syntax and semantics are learnt simultaneously. Joint learning results in better grammar induction, realistic lexical category learning, and better interpretations of novel sentence and verb meanings. Joint learning makes language acquisition easier for learners by mutually constraining the hypotheses spaces for both syntax and semantics. Studying the dynamics of joint inference over many input sources and modalities represents an important new direction for language modeling and learning research in both cognitive sciences and AI, as it may help us explain how language can be acquired in more constrained learning settings.