In lieu of an abstract, here is a brief excerpt of the content:

NEG Raising 3.1 NPIs This study of Classical NR should be understood against a background assumption about negative polarity items (NPIs), sketched in Postal 2005. This view makes the entirely nonstandard assumption that what are normally called NPIs are expressions underlyingly associated with a NEG, which has raised away from the NPI. Such a view contrasts radically with the range of consensus views about NPIs. Consider a standard NPI example like ever in (1b): (1) a. *Chloe ever tasted beer. b. Chloe did not ever taste beer. On the standard view, because ever is an NPI, it can only occur in contexts where it relates in specific ways to a form standardly called its licenser. This licenser must have certain semantic characteristics (usually taken to be decreasingness ; see Ladusaw 1979) and needs to be in an appropriate syntactic position vis-à-vis the NPI, for example, one c-commanding the NPI. In (1b), the licenser would be not, while (1a) would be ungrammatical because of the absence of a legitimate licenser. With respect to (1b), we claim that the not found immediately following the Aux originates as part of the adverbial form. In effect, we posit a single original structure common to the ever of (1b) and the clearly negative never of (2): (2) Chloe never tasted beer. Artificially ignoring scope structures for the moment, a first schematic account of the structure underlying both (1b) and (2) would be (3a) or, a bit more explicitly, (3b): (3) a. Chloe [NEG ever] tasted beer b. Chloe [[NEG SOME] ever] tasted beer 3 18 Chapter 3 The relation between structures (3a,b) and (2) is relatively straightforward, involving the morphological specification that a [[NEG SOME] ever] constituent (out of which its NEG has not raised) takes the form never. NEG in this case is realized simply as n- and SOME is not pronounced. But more is required to relate (3a,b) to (1b). In particular, the NEG must raise out of the adverbial form to a position immediately adjacent to the Aux. Such an analysis requires an overall syntactic framework in which separation of NEG from its original structural locus is both well-defined and permitted— that is, in effect, a syntactic framework that countenances some notion of NEG raising. The resulting analysis of (1b) is given in (4): (4) Chloe did NEG1 [[ SOME] ever] taste beer In this structure, NEG has raised from a position adjacent to SOME, to a position right-adjacent to the Aux did. The higher occurrence of NEG1 in (4) is realized as not or n’t. The lower occurrence is covert. (We note in passing that the movement of NEG1 in (4) violates Ross’s (1967 [1986:127]) Left Branch Condition. We return to a detailed discussion of the locality constraints on NEG raising in chapters 11 and 12.) The NEG-raising analysis of (1b) accounts for the truth-conditional equivalence between (1b) and (2). Both sentences have the same underlying structure, and the movement of the NEG does not contribute to the semantic interpretation of (1b). That is, even though the NEG’s superficial position immediately follows the Aux in (1b), its interpretation is a function of its preraising position (internal to the DP). There are distinct ways of capturing the fact that the raising of the NEG in (1b) is semantically vacuous. In the Metagraph framework of Johnson and Postal 1980 and Postal 2010, this conclusion follows from the defining assumption that semantic interpretation ignores noninitial arcs. In Minimalist syntax, the natural treatment of the semantic vacuity of NEG raising would be in terms of reconstruction. We do not pursue these issues here. This brief discussion illustrates that within a framework accepting the view taken in Postal 2005 about NPIs, the NEG raising required by a syntactic view of Classical NR (see chapter 1) should be viewed as a special instance of a much broader syntactic phenomenon of NEG raising. We view NEG raising in general as being possible from any position. Many, no doubt a majority, of the resulting structures are ill-formed. We assume these are blocked by various filtering conditions. One class of these will specify possible landing sites where NEGs can remain. In English, raised NEGs that are not ultimately deleted (see chapters 7 and 8 on NEG deletion) mostly end up as right sisters of the finite Aux. [18.225.149.32] Project MUSE (2024-04-19 21:54 GMT) NEG Raising 19 3.2 Some/Any A similar...

Share