@article{oai:nagoya.repo.nii.ac.jp:00027975,
 author = {Osawa, Satoko},
 journal = {IVY},
 month = {Oct},
 note = {It has been widely assumed that the process of passivization is one of the instance of NP-movement, namely Case-driven movement. However, if we adopt Chomsky's(1995) feature checking theory, passivization cannot be Case-­driven movement since Case feature does not motivate overt movement in this framework. Under the feature checking theory, only strong uninterpretable features, such as the D-feature of T (EPP), cause overt movement and weak uninterpretable features, such as Case-feature, cause covert movement. This entails that passivization in the matrix sentence is caused by the EPP but not Case (cf. Mahajan 1995). A closer observation, however, reveals that in passives in small clauses, where the EPP is irrelevant, the underlying object still undergoes movement, as indicated in (1). (1) a. *John saw [arrested Bill] b. John saw [Billi arrested ti] The examples in (1) suggest that passivization is not caused by the EPP either. Furthermore, I will discuss that the derived subject Bill in (1b) exhibits some subject properties, which is unpredictable if it moves to the matrix Case position due to the Case-theoretic reason. (2) indicates that the derived subjects observe the Subject Condition and (3) shows that they can be modified by subject oriented adverbs. (2) *Whoi did John see [[friends of ti] arrested] (3) Mary saw [the student reluctantly instructed by John] I will propose that the movement in passives is caused by the θ-requirement of the passive morpheme, which is a predicate head forming a complex predicate with vP. Extending traditional analyses of passives, I will put forth a complex predicate analysis of English passive as in (4). (4) A Complex Predicate Analysis The passive morpheme is a predicate head, which (i) introduces an external argument Undergoer, and (ii) takes vP (Event) as its complement. A relevant part of the structure derived from the analysis is given in (5). (5） [PassP Undergoer1 [Pass' Vi-ed [vP by Causer [v' ti [VP ti Undergoer2]]]]] The analysis entails that the passive subject should be a θ-related and that the by-phrase is an argument. There is some cross-linguistic evidence to support our analysis: the existence of passive verbs in Asian languages, which forms a complex predicate with a sentential complement. The existence of such verbs also provides evidence against the analysis advanced by Jeaggli (1986) and Baker, Johnson, and Roberts (1989), under which the passive morpheme is viewed as an argument. The proposed analysis consequently implies that there is movement into a θ-position within a lexical phrase.(cf. Bošković 1994)},
 pages = {89--114},
 title = {A Complex Predicate Analysis of Passives},
 volume = {32},
 year = {1999}
}