Classical NEG Raising: An Essay on the Syntax of Negation

19. Nonfinite Clauses

Nonfinite Clauses 19.1 A Literature Bias toward Finite Hosts In standard Classical NR cases like (1) and (2), the host is finite: (1) He doesn’t seem to be happy. (2) Ryan does not think he can leave until Friday. And, impressionistically, the database found in the decades-old literature on Classical NR as a whole consists overwhelmingly of finite host cases. For instance, we make out that in Horn’s (1978) extensive and widely cited study of Classical NR, none of the more than fifty simple negative sentences taken to illustrate English Classical NR (i.e., those not involving main clause negative quantifier DPs) has a nonfinite host. Similarly, all the Classical NR cases in Gajewski 2007 and Homer 2010 have finite hosts. That fact coexists with the proposals in nonsyntactic approaches that the Classical NR phenomenon is purely a function of the interaction of the meanings of the CNRPs with independent semantic or pragmatic principles. Since it is obscure how the semantic properties of CNRPs could be systematically different in their finite and nonfinite instances, on semantic/pragmatic treatments one expects that the same relations should systematically hold when the NEG sits in a nonfinite host.1 19.2 Basics of Finite/Nonfinite Contrasts But it was long ago briefly observed that the expectation at the end of the previous paragraph does not hold: (3) (Seuren 1985:168) “A further problem is that NR does not seem to operate freely whenever an NR-predicate occurs and not is the highest embedded 19 212 Chapter 19 predicate. NR seems somehow connected with the top of the tree, where it ‘hooks’ on to speech act properties. . . . In some kinds of embedding NR seems altogether out of place. A sentence like (130c) means what it says, not ‘I expect that the minister will think that the plan won’t succeed’: (130) c. I expect the minister not to think that the plan will succeed.” And Seuren merely touched on the tip of an iceberg.2 Classical NR seems to us not to be available in a wide variety of cases where a potential host is nonfinite. In all the following sentences not precedes to, but the contrasts persist as well if not follows to. Each set involves a strict NPI that is ungrammatical when its associated NEG would have raised to a position in a nonfinite host: (4) a. She does not believe that Rodney will leave until 10:00. b. It is possible for her not to believe that Rodney will leave (*until 10:00). c. It is possible for her to believe that Rodney will not leave until 10:00. (5) a. Curtis does not think that they have visited her in ages. b. *For Curtis not to think that they have visited her in ages is strange. c. For Curtis to think that they have not visited her in ages is strange. (6) a. I didn’t think that the play was half bad. b. *They expect me not to think that the play was half bad. c. They expect me to think that the play was not half bad. (7) a. Terry didn’t figure that Rodney would do jackshitA. b. *It was wise for Terry not to figure that Rodney would do jackshitA. c. It was wise for Terry to figure that Rodney would not do jackshitA. (8) a. They don’t expect that she will feel so hot after the operation. b. *It is shocking for them not to expect that she will feel so hot after the operation. c. It is shocking for them to expect that she will not feel so hot after the operation. (9) a. They don’t think that she can help cheating. b. *They are certain not to think that she can help cheating. c. They are certain to think that she cannot help cheating. While (4a) has a Classical NR reading, in (4b) only the short version is grammatical , and it does not intuitively have the Classical NR interpretation. Just Nonfinite Clauses 213 so, while all of the (a) examples in (4)–(9) seem fine, we find the (b) examples ungrammatical. Further examples seem to reveal the same pattern seen in (4)–(9): lack of a Classical NR reading for a NEG raised into a nonfinite clause and the impossibility of a strict NPI in the relevant complement clause: (10) a. That...