Questions and entailment

蕴涵是语义学中的一个基本概念：例如，每个英语使用者都知道“John read War and Peace”包含“John read a book”，这是一个核心语义事实。此外，蕴涵对理解诸如负极性项和数量含义等语法现象也很重要。在陈述句中，它们的分布可以从句子的蕴涵属性中预测（例如Gazdar 1979，Ladusaw 1979）。即，观察到这些现象的出现与特定蕴涵性质之间的强相关性。但负极性项和数量含义也出现在问句中。蕴涵是以陈述句之间的语义关系为基础的，这是理解疑问句中这些现象的一个众所周知的问题。因此，例如，“约翰是否读过《战争与和平》？”和“约翰读书了吗？“是通过蕴涵而联系在一起的。因此，人们对疑问句中的负极性项和数量含义的具体许可条件并不清楚。但是，这一悬而未决的问题使人们对蕴涵和最初提到的现象之间的因果关系产生了怀疑。这个项目以一种新颖的方式解决了30年来关于问题中蕴涵敏感现象的争论。它通过两个方面来解决这个问题：首先，我在Guerzoni和Sharvit（2007）的工作基础上，详细研究了蕴涵反转和问题的穷举性之间的相关性。其次，基于这些见解，我发展了一个新的问题语义理论。这个新理论的结果是：（i）问题是由蕴涵联系起来的，（ii）对一个问题的强穷举回答以否定的方式颠倒了蕴涵。换句话说，对“约翰读过一本书吗？”“约翰读过《战争与和平》吗？”'.这项调查对其他一些重要的辩论产生了重大影响。特别是，我希望该建议的证据来自五个经验领域：（i）反加法和修辞问题，如“约翰举起一根手指来帮助你吗？（ii）某些疑问句中数量隐含的反转和另一些疑问句中数量隐含的缺失，（iii）说话人对强穷举答案的直觉，（iv）答案中的焦点分布，以及（v）嵌入疑问句及其嵌入谓词的分析。

Entailment is a fundamental notion in semantics: e.g., it is a core semantic fact that every English speaker knows that 'John read War and Peace' entails 'John read a book'. Moreover, entailment is also important for the understanding of grammatical phenomena such as negative polarity items and scalar implicatures. In declarative sentences, their distribution is predictable from the entailment properties of the sentence (e.g. Gazdar 1979, Ladusaw 1979). I.e., a strong correlation between the appearance of these phenomena and specific entailment properties is observed. But negative polarity items and scalar implicatures also occur in questions. A well-known problem for the understanding of these phenomena in questions is that entailment is based on semantic relations between declarative sentences. It is therefore unclear whether, for instance, 'Did John read War & Peace?' and 'Did John read a book?' are related by entailment. As a result the exact licensing conditions of negative polarity items and scalar implicatures in questions are ill-understood. But this open issue has cast into doubt the causal link between entailment and the phenomena mentioned in the first place. This project addresses the 30-year-old debate about entailment-sensitive phenomena in questions in a novel way. It seeks a solution to the problem by addressing it on two fronts: first I study in detail the correlation between entailment reversal and exhaustivity in questions, building on work by Guerzoni & Sharvit (2007). Second I develop a new theory of question semantics based on these insights. This new theory has the consequence that (i) questions are related by entailment, and (ii) the strongly exhaustive answer to a question reverses entailment in the same way as negation. In other words, the strongly exhaustive answer to 'Did John read a book?' entails the one to 'Did John read War & Peace?'. The investigation has significant consequences for a number of other important debates. In particular, I expect evidence for the proposal to come from five empirical domains: (i) anti-additivity and rhetorical questions such as 'Did John lift a finger to help you?', (ii) the reversal of scalar implicatures in certain questions and the absence thereof in others, (iii) speaker's intuitions about the strongly exhaustive answer, (iv) focus-distribution in answers, and finally (v) the analysis of embedded questions and their embedding predicates.