Dawei Jin (University at Buffalo/ Shanghai Jiao Tong University)
Rethinking the argument-adjunct asymmetry in Mandarin Chinese island effects
Unlike the (strong) island sensitivity in overtly displaced filler-gap dependencies (e.g. fronted wh-questions), the island mechanism in wh-in-situ is known to apply selectively, because in situ islands depend on the choice of wh-phrase. It has been a controversial topic in the East Asian literature which type of wh-phrase incurs island violations. There is agreement that wh-arguments are free from island effects. Meanwhile, some analyses argued that island-sensitive items comprise wh-adjuncts (Huang 1982; Tsai 1994; Fujii & Takita 2007), defined in terms of structural category (Stepanov & Tsai 2008), restriction of nominal variables (Reinhart 1998), semantic referentiality (Szabolcsi & Zwarts 1993) or D-linking (Pesetsky 1987). Such proposal predicts that why-questions and manner how-questions, as well as amount phrases (how much/how many), pattern similarly in island contexts, their degradations being non-significant from each other. This is due to the reason that these phrases fit in with the criteria of adjuncthood regardless of which definition above is adopted. An alternative proposal claims that the why-adjunct uniquely gives rise to island effects due to its idiosyncratic attachment position (Tang 1988; Nishigauchi 1990; Lin 1992; Jin 2015; Murphy 2017). Hence, it predicts that why-questions induce a significantly more severe degradation compared with manner how-questions and amount questions in island contexts. This talk goes through the theoretical background of the above debate, and presents experimental evidence in support of the latter proposal wherein the in situ island effects are associated with why-questions.