The Grammar of Negative Polarity
, M. Linebarger 1980
This study attempts to characterize the distribution of negative polarity items in English, and to delimit the role of NPIs within sentence grammar.
the recalcitrance of NPIs to a purely syntactic analysis has long been observed, notably in Baker (1970). It is argued there that the primary trigger is negation, a sufficient condition on NPI acceptibility being a syntactically defined relationship to negation; and that all other triggers are to be defined in terms of a semantic relation to negation, logical entailment. The structure of this proposal is maintained, but with major alterations.
First, the SYNTACTIC (i.e. structural) condition must be stated on logical form rather than on surface structure: in the LF of a sentence containing an NPI the NPI must be represented as occurring in the "Immediate scope" of the negation operator. It is shown that this notion of immediate scope, which also plays a role in grammatical processes such as government and in "pragmatic quantification", similarly constrains the distribution of "free choice" any and other quantifiers. This constraint cannot be defined on surface syntactic structure: NPIs thus provide empirical evidence about the existence and syntax of LF, a level of linguistic representation at which logical structure is represented and which is the interface between sentence grammar and semantics. Further evidence about the syntax at LF is provided by arguments that the negative polarity quantifier any must be represented as existential rather than a universal quantifier.
Second, the SEMANTIC condition on sentences with NPIs which do not meet this structural condition is to be stated in terms of their literal meanings; the availability of such sentences for utterance, however, is affected by conversational intent and real world beliefs.
This analysis of NPIs, which posits as the primary trigger a structural relation to negation and defines other triggers in terms of their semantic relation to this primary trigger, is constrasted with a quite different account of NPIs proposed by Ladusaw (1979), who attempts a unified an semantic analysis of all triggers on the basis of logical entailment alone.
Table of Contents
Introduction 1
Footnotes 4
Chapter 1 Surface structure accounts 7
1.1 Lasnik (1975) 8
1.2 Jackendoff (1969, 1972) 10
1.3 Critique of surface structure accounts 12
Summary of Chapter 1 20
Footnotes 21
Chapter 2 Baker"s conjecture 23
Chapter 3 Part (A) " the paradigm CASE 27
3.1 Sentence (1a) and the predicate CAUSE 27
3.2 The immediate scope constraint 29
3.2.1 Additional violations of the ISC 31
3.2.2 Immediate scope in Kroch (1974) 39
3.2.3 Immediate scope and free choice any 43
3.2.4 Immediate scope and "nonreferentiality" 45
Summary of Section 3.2 48
3.3 Mechanics of the proposal 48
3.3.1 Logical form 49
3.3.2 SS Ã LF mapping rules 49
3.3.3 The application of the NPI rule 56
3.4 Possible counterexamples 57
3.4.1 Acceptable interveners 57
3.4.2 NPI quantifiers with wide scope 60
Summary of Chapter 3 62
Footnotes 63
Chapter 4 Part (B) " "Allusions" 67
4.1 A data set 67
4.2 Insufficiency of a purely semantic account 77
4.3 Insufficiency of a purely conversational account 80
4.4 A tentative proposal 81
4.5 Availability 84
4.5.1 Implicature denied by speaker 85
4.5.2 Implicature "background" 87
4.5.3 Implicature uninformative 89
4.5.4 Implicature otherwise unavailable 90
Summary fo Section 4.5 93
4.6 Part (b) and sentence grammar 94
Summary of Chapter 4 97
Footnotes 100
Chapter 5 The representation of any 103
5.1 Arguments based upon truth conditions 104
5.1.1 "Not Æ" contexts 104
5.1.2 Verb phrase deletion 107
5.2 Arguments based upon other grammatical rules 108
5.2.1 Tag questions 108
5.2.2 Restrictions on quantifier scope 109
5.3 Distributional arguments 110
5.4 Arguments based upon simplicity 112
5.4.1 Interaction with every 112
5.4.2 Other NPIs 114
5.4.3 Surface structure constraints 115
5.5 Arguments based upon "Æ not" contexts and the immediate scope
constraint 116
Summary of Chapter 5 120
Footnotes 122
Chapter 6 Ladusaw (1979) 123
6.1 Belief contexts and downward entailment 129
6.1.1 Ladusaw"s notion of entailment 129
6.1.2 Interaction with other entailments 132
6.2 Acceptable NPIs not in the scope of a DE operator 133
6.2.1 Incorrect belief 134
6.2.2 Loss and forgetfulness 135
6.2.3 Only 135
6.3 Unacceptable NPIs in the scope of a DE operator 138
6.3.1 Intervening logical elements 138
6.3.2 Promises 139
6.3.3 Comparatives 139
6.3.4 Squishiness in other contexts 139
6.3.5 Queclaratives 140
Summary of Section 6.3 140
Summary of Chapter 6 141
Footnotes 143
Chapter 7 Surface structure residue 143
7.1 The desirability of eliminating explicit surface structure restriction
on NPIs 143
7.2 SS Ã LF mapping rules and NPIs: the Easy Cases 144
7.3 Problems 147
7.3.1 Nonspecific NPs 147
7.3.2 Raised quantifiers 150
7.3.3 Part (b) triggers 151
7.3.4 Wh-words and NPIs 153
Footnotes 154
Chapter 8 What is an NPI? 155
8.1 Scalar endpoints 156
8.2 Degrees of dependence upon NOT 158
8.3 Understaters 161
8.4 Negated verbs 162
Footnotes 164
Conclusions 165