27 February 1998

How to get Passive

C.-T. James Huang

University of California, Irvine and Stanford University

This talk will start with an analysis of a class of passive sentences in Mandarin Chinese that integrates two traditional opposing views (i.e., the "complementation approach" and the "movement approach"). It treats this class of passives as involving an experiential verb taking a clausal complement (hence complementation) which turns itself into a null-operator structure by A' movement of its object (hence movement). Evidence for complementation comes from the theta-independent status of the subject and various constituency tests. Evidence for movement comes from the obligatory existence of a null object (in the relevant cases). That this is A'-movement is evidenced by the existence of long-distance passivization, its sensitivity to island conditions, the distribution of object clitic suo, and the distribution of resumptive pronouns.

A different class of Mandarin passives (the "agentless passive") will be shown to lack all of the properties of A'-movement or Agent-deletion, but retain apparent properties of complementation and A-movement. An analysis that likens these passives to English get passives is entertained, but called into question.

I will then compare Mandarin passives to passives in other Chinese languages, then in other East Asian languages (Japanese and Korean), and then in some Western languages (English and French). The questions addressed include the universality of the notion 'passive', sources of cross-linguistic variations, and the proper analysis of "indirect passives" (both the 'inclusive' indirect passives and the 'exclusive'--i.e. adversative--passives). For the last question, it will be suggested that they involve the passivization of the 'outer' or 'outermost' object of a complex predicate.

Time permitting, I will point out that all facts considered amount--for the most part--to an existence argument of the analyses defended, as it is difficult to exclude the possibility that certain passives are derived in a way typically assumed for English be passives. Conversely, it is equally difficult to completely exclude an analysis according to which certain English be passives involve complementation and control.