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

Notes Chapter 1 1. Among the possible syntactic views of Classical NR that are not transformational are those that involve no generative rule at all. This would be the case in any framework assuming a grammar to be a model-theoretic system; see Johnson and Postal 1980, Pullum and Scholz 2001, 2005, Pullum 2007, and Postal 2010. In the framework of Postal 2010, a syntactic version of Classical NR would involve a Type I foreign successor of an arc in the complement whose head node is NEG. This formulation guarantees that the complement arc A headed by the raised NEG, NEG1, is erased and that only the foreign successor arc B of A headed by NEG1 remains unerased. This guarantees that NEG1 appears superficially only in the main clause. 2. Seuren 1985:166–172 represents a rare later defense of syntactic views against nonsyntactic alternatives. McCawley 1998:595–604 is the most recent instance we can cite of support for a syntactic view. 3. See Horn 1971:120, 1975:282, 1978:136–150, Jackendoff 1971:293, Seuren 1974a:184–185, 1974b:121, Smith 1975:68–74, and Prince 1976:405 for various instantiations of this line of argument. Horn (1978:143–150) elaborates many difficulties facing this argument type. 4. The term strict NPI has been used in various senses in the literature; see Horn 1978:136–140, Linebarger 1987:47, Progovac 1994:81, 145 (citing Linebarger 1980), van der Wouden 1997:76, Giannakidou 2011:1680. Our usage (an NPI requiring a local licenser) essentially follows that of Linebarger, Progovac, and Horn. 5. Contrasts like the one in (7) must be segregated from cases of so-called secondary triggering, in which a strict NPI cooccurs in a local domain with a nonstrict NPI; see, for example, Horn 1989 [2001:181]. Such cases are grammatical for some speakers even in NEG-free complement clauses with non-CNRPs; compare (7a) and (i). (i) Calvin did not claim that he had taken any drugs until his 18th birthday. We cannot consider why examples like (i) are well-formed for some speakers. But we suspect that the key aspect is that the nonstrict NPI phrase, here any drugs, and the strict NPI, here until his 18th birthday, form a polyadic quantifier with main clause scope. The issue of polyadic quantification and NPIs is described in detail in chapter 222 Notes 6. The more general confounding factor of high scope is addressed in chapter 9 and later chapters. In terms of those discussions, any drugs and until his 18th birthday take matrix clause scope and are interpreted as a polyadic quantifier with a shared D containing a NEG. Hence, the strict NPI until phrase is locally related to a NEG, but that relation holds in the main clause. We suspect that it is polyadic quantifier formation that allows until to take wide scope in (i). 6. We provide here a list of other strict NPIs that, as far as Classical NR is concerned, we believe function in essentially the same way as those we do cite. Readers may wish to substitute some of these in examples should they find that certain NPIs that we consider strict lack that property in their dialects. We cite them with a preceding not to bring out the often idiomatic meaning: not believe one’s eyes, not for anything in the world, not give it another thought, not be half bad, not long for this world, not miss much, not move a muscle, not be much of (something), not one iota, not be one’s place, not sleep a wink, not feel so hot, not take no for an answer, not touch (something) with a ten-foot pole, not worth a dime. We should stress that the categorization of NPIs as strict in no way precludes considerable diversity among them. To see the possible variety and poorly understood complexity of the matter, it suffices to consider the following paradigm kindly provided to us by Larry Horn (personal communication). The arguably strict NPIs are italicized: (i) a. *Like he’d do that for anything in the world. b. *Like that paper was half-bad. c. *Like she misses much. d. Like he’d do one iota of work/move a muscle for us. e. Like I could sleep a wink with this racket. f. Like I’d touch that with a ten-foot pole. 7. Pullum and Huddleston (2002:825) deny that stop at nothing has a stop at anything variant...

Share