Learning Attitude Verb Meanings in a Morphosyntactically-Poor Language

Word meanings are learned under less-than-ideal conditions: in principle, a word can have many possible meanings, and learners must choose the correct one with limited explicit instruction. This problem seems to be exacerbated in the context of attitude verbs, like think, believe, want, which describe abstract mental states. These verbs lack reliable physical correlates in the real world, so the non-linguistic context provides very limited information to the learner about their meanings. An influential proposal addressing this learning problem posits that learners rely on linguistic context to learn their meanings, via syntactic bootstrapping (Gleitman, 1990; Gleitman et al., 2005; Papafragou et al., 2007; Gillette et al., 1999, etc.). In this paper, we investigate the feasibility of syntactic bootstrapping in a morphosyntactically poor language. We are interested in the fact that cross-linguistically, children seem to master at a relatively early age the semantic differences between “belief verbs” like think, believe, which express judgments of truth, and “desire verbs” like want, love, that express preferences (e.g. Bolinger, 1968; Hooper, 1975; Villalta, 2008). In particular, we explore the Declarative Main Clause Syntax Hypothesis (e.g. Hacquard and Lidz, to appear; Hacquard, 2014; Lidz et al., to appear; Harrigan et al., 2016), a syntactic bootstrapping account for attitude verb meanings, taking Mandarin Chinese as a case study. In principle, Mandarin offers relatively few cues for syntactic bootstrapping purposes, as it has minimal verbal and nominal morphology and allows null arguments. If the Declarative Main Clause Syntax Hypothesis is a viable learning strategy in Mandarin, then it is likely to be just as viable in languages that are similarly morphosyntactically impoverished, and even more viable in languages with richer morphosyntax. This paper is organized as follows: Section 2 presents the Declarative Main Clause Syntax Hypothesis and the syntactic properties of Mandarin belief and desire verbs. Sections 3 and 4 present and discuss the results of two corpus studies. We argue that there are enough morphosyntactic cues in Mandarin