Abstract Allen et al.
Cross-linguistic developmental differences in the expression of Manner and Path: Evidence from speech and gesture
Languages differ in how they typically express manner and path elements that occur simultaneously in a motion event (Talmy 1985). Satellite-framed languages like English typically express them together in one clause (flew out), while verb-framed languages like Japanese and Turkish typically use two separate clauses (ucarak cikti ‘exited flying’). Patterns for iconic co-speech gesture mirror those in speech. Adult speakers who express manner and path together in one clause in speech tend to produce one gestural unit which represents manner and path simultaneously (Kita & Özyürek 2003). Adult speakers who express manner and path in separate clauses tend more often to produce a separate gestural unit for each of manner and path.
In this talk, I explore whether children follow these language-specific patterns in the syntactic and gestural encoding of manner and path in early stages of language development, or if they use a universal default pattern. Data were collected from speakers of three languages: English, Japanese, and Turkish. In each language, 3 groups of 20 children each participated (mean ages 3;8, ) as well as 20 adults. Each subject narrated 10 video clips of motion events incorporating both manner and path.
Speech results show that children are sensitive to the predominant syntactic packaging patterns used by adults in their language as early as age 3. However, universal patterns are also evidenced: across all three languages, 3-year-olds show a tendency to encode both elements in one verbal clause even if this pattern is not typical of the adult language. This tendency disappears by age 9. Gesture results are similar. Children at age 3 follow broadly language-specific patterns, but there is less difference between Turkish- and English-speaking children at age 3 than there is between Turkish- and English-speaking adults.
Various explanations for these results can be entertained. Speech results alone suggest that children might have a general learning bias to encode simultaneous aspects of an event in one clause in early stages of language development. Gesture data allow us to ask whether children might also conceptualize motion events differently from adults.