-
10. Principle C Phenomena
- The MIT Press
- Chapter
- Additional Information
10 Principle C Phenomena 10.1 Index-Based Principle C Fails Given previous evidence that imposters can antecede either 3rd person or non–3rd person pronominals, at first sight the contrast between (1) and (2) is mysterious. (1) a. I1 think that yours truly1 was treated rather well. b. I1 think that this reporter1 deserves credit. c. I1 think that your faithful correspondent1 should write more often. d. Do you1 think that Madam1 will be able to contact the doctor? (2) a. *He1 thinks that yours truly1 was treated rather well. b. *He1 thinks that this reporter1 deserves credit. c. *He1 thinks that your faithful correspondent1 should write more often. d. *Does she1 think that Madam1 will be able to contact the doctor? In (1), a 1st person pronoun can c-command an imposter whose core DP has AUTHOR as an antecedent. But in (2), where pronoun and imposter match so as to indicate 3rd person agreement with the shell DP, ungrammaticality results. We claim that this contrast reduces to the antecedence-based version of Principle C stated in (45) of chapter 4. (3) Principle C If a DP Q is an antecedent of a pronominal P, then P does not c-command Q.1 Under the view of pronominal agreement in chapter 7, a pronominal receives the values of its relevant ϕ-features from an immediate antecedent (or a secondary source). Underlying that view is the idea that the value of a ϕ-feature— say, person—can have two different types of origin. For lexical DPs (such as the tiger), the values are inherent, not dependent on matching the feature values of other DPs. For nonexpletive pronominal DPs, however, the values are in 132 Chapter 10 general determined by agreement with their immediate antecedents or secondary sources. Returning to the contrast between (1) and (2), he agrees with the 3rd person singular shell of the imposter yours truly. From this, we conclude that yours truly is an immediate antecedent of he; see (4) in chapter 7. Since he c-commands its immediate antecedent, the version of Principle C in (3) blocks that analysis. Why then does this logic not suffice to block (1a) as well? Example (1a) fails to violate the version of Principle C in (3) because the c-commanding pronoun is I, a 1st person form, and the shell DP need not be taken as its immediate antecedent. In such cases, the 1st person subject pronoun I is the antecedent of the imposter DP in the embedded clause, rather than conversely. The 1st person pronoun itself can be taken to be anteceded by the AUTHOR DP posited in chapter 4. Since AUTHOR occurs at a high point in each sentence structure, it will c-command the subject of was treated, not conversely , and no violation of the version of Principle C in (3) is induced. It is worth commenting here on suggestions by Baker (2008, 126). He maintains that a DP can only be 1st or 2nd person if it is bound by one of two special operators (which he calls S ‘speaker’ and A ‘addressee’). He claims, furthermore , that lexical DPs cannot be bound by operators and that therefore “lexical NPs cannot be first or second person.” Evidently, we essentially agree with the claim that 1st and 2nd person pronominals only exist via their relation to special elements, those we called AUTHOR and ADDRESSEE in chapter 4. But our overall multi-DP treatment of imposters is designed to combine that recognition with the unquestionable 3rd person properties of imposters and camouflage DPs. This it does by having such forms anteceded by AUTHOR and ADDRESSEE while nonetheless requiring them to be 3rd person. Unlike Baker’s proposal, which treats imposters as syntactically straightforward, exclusively 3rd person DPs, our analysis yields a description of the properties these DPs can have of jointly determining 3rd person verbal agreement while anteceding and being anteceded by non–3rd person pronominals. 10.2 Antecedence versus Coindexing Discussion of anaphora in recent decades has almost uniformly appealed to indexing and coindexing. This book (like Higginbotham 1983) is an exception , as we have based our account on a primitive relation of antecedence. But the reader will have noticed that our appeal to antecedence, though traditional , has not been buttressed with any argument for the superiority of an antecedence-based account over a purely index-based one. We now consider such an argument. [3.236.18.23] Project MUSE (2024-03-19 11:44 GMT) Principle C...