Visioseminar (LPL, PolyU, Macau U)
The third Variamu Seminar was held on Tuesday 24th of March.
We listen a talk by:
Typological variations in the acquisition of relative clauses: evidence from child Cantonese
Relative clause (RC) is a widely recognized topic of investigation in the fields of language typology, acquisition and processing. The Noun Phrase Accessibility Hierarchy (NPAH, a “putative” linguistic universal) has a long history in accounting for the acquisition of RCs in European languages. Recently however, conceptual changes in rethinking the typology of RCs and child language data from certain Asian languages led researchers of language acquisition to question how far NPAH could account for the learning patterns observed. For instance, Chen & Shirai (2014) hypothesized that in attributive clause languages, multiple factors including input frequency and similarity to canonical clauses, rather than purely structure-based accounts, determine the learning trajectory of RCs. Currently, there is little direct testing of these competing theoretical perspectives. Thus far the evidence from child first acquisition of attributive clause languages (e.g. Japanese, Mandarin, Cantonese) suggesting that the developmental patterns do not follow predictions from NPAH is largely based on naturalistic speech production or comprehension data which have their own methodological limitations.
As an attributive clause language, Cantonese provides an interesting test case. In this talk, I will first review the current literature including recent and ongoing works by myself and collaborators, highlighting phenomena of apparent conflicting findings (or “variations”) that also happen in other Asian attributive clause languages, but appear to rarely occur in European languages. I will then outline a proposed project that uses multiple methods to better test the alternative theoretical perspectives. The findings offer a rich testing ground to evaluate a range of theoretical perspectives. An important innovation of the proposed project is the extension of experimental studies to not only Cantonese, and eventually to other attributive clause languages such as Japanese, Mandarin and Thai. Moreover, computational linguists can make use of these new datasets to model the acquisition and processing of RCs in attributive clause languages. The project will broaden and deepen our understanding of RC acquisition/processing in an Asian linguistic context.