Cheung (1998) identified that Right Dislocation in Cantonese and that in European languages like Italian and English are vastly different. Despite the differences he pointed out, he did not proceed to conclude that what he regarded as the most common type of RD in Cantonese, namely Gap Right Dislocation, is an independent phenomenon. It should be separated with what we commonly refer to as RD, and be analyzed on its own right.

On the other hand, while Cheung regarded Pronominal Right Dislocation to be an equivalent of RD in European languages, I will show that although the two kinds of RD belong to the same phenomenon, Cantonese RD actually has its own specific behaviors, and is not an exact copy of RD in European languages. I will especially compare RD in Italian with that in Cantonese, because of their being pro-drop languages and because Italian RD is more extensively studied.

Huang (1984) proposes that null object is a variable controlled by a null topic, unlike null subject which is a real null pronoun. This paper tries to verify this proposal by investigating the distribution of null object in Cantonese. It is found that while Huang’s proposal is able to explain the majority part of the distribution of null object, it is unable to give an explanation to several cases related to resultativity and indirect object which disallow null object. A possible explanation is given to account for these cases, which is related to feature checking and a modified version of Huang’s Generalized Control Rule.