The traditional observation that Chinese wh-arguments do not exhibit wh-island effects may be only apparent. With new evidence from “how-many” phrases, it is demonstrated that Chinese has wh-island effects even with wh-arguments. What nullifies such effects is in fact the disguise of D-linkedness. Although the lack of wh-island effects seems to pattern Chinese wh-construals with Japanese ones, further tests show that these two languages are still different with respect to strong island effects, (anti-)crossing effects, and multiple wh-interpretations. The finding leads to the need to reinvestigate the mechanisms underlying the scope-taking wh-elements of wh-in-situ languages on the one hand, and those triggering wh-island effects on the other.
