摘要：This paper addresses the theoretical and descriptive relevance of pragmatic inference with the purpose of exploring the syntax-semantics interface in secondary predication in English. Most standard syntactic analyses proceed on the assumption that there is a uniform, one-to-one correspondence of the postverbal NP (NP2 henceforth) and the following XP with the semantic interpretation of these phrases as direct object and complement/modifier (or adjunct), respectively. However, upon closer examination of sentences like "I found John gone/out of sight" or "I have a tooth missing", it appears that such an assumption can be challenged on both theoretical and descriptive grounds. Since these pragmatically marked predication structures involve the cancellation of the conventional implicature of the NP1 V NP2string, it is suggested that the syntactic analysis of these instances cannot be established solely on the basis of formal properties alone, but rather needs to be seen in terms of the interaction of the inherent meaning and form properties of the syntactic constituents of the construction with the actual interpretation of these properties by the subject/speaker in a given discourse scenario. At a descriptive level, it is argued that these marked instances of secondary predication can be more aptly analyzed, both syntactically and semantically, as involving complex predicates taking the postverbal NP2as their sole object argument
