LINGUIST List 22.3794 Thu Sep 29 2011 Diss: Philosophy of Lang/pragmatics: Lassiter: 'Measurement and ...' Editor for this issue: Xiyan Wang <xiyanlinguistlist.org> To post to LINGUIST, use our convenient web form at http://linguistlist.org/LL/posttolinguist.cfm. Directory         1.     Daniel Lassiter , Measurement and Modality: The scalar basis of modal semantics Message 1: Measurement and Modality: The scalar basis of modal semantics Date: 22-Sep-2011 From: Daniel Lassiter <danlassiterstanford.edu> Subject: Measurement and Modality: The scalar basis of modal semantics E-mail this message to a friend Institution: New York University Program: Department of Linguistics Dissertation Status: Completed Degree Date: 2011 Author: Daniel Lassiter Dissertation Title: Measurement and Modality: The scalar basis of modal semantics Dissertation URL: http://www.stanford.edu/~danlass/Lassiter-diss-Measurement-Modality.pdf Linguistic Field(s): Philosophy of Language                             Pragmatics                             Semantics Dissertation Director: Seth Yalcin Chris Barker Dissertation Abstract: This dissertation argues that modal expressions and gradable expressions in English can and should be treated using the same semantic apparatus. I present a number of serious empirical problems for standard theories of epistemic, deontic, and bouletic modality -- including the dominant theory due to Kratzer -- which treat them as quantifiers over possible worlds. In particular, the fact that many modals are gradable presents a serious problem, as do a variety of differences between epistemic and deontic modals in the inferences that they license. Most importantly, 'p is better than q' and 'p is better than r' together imply 'p is better than (q or r)', but the same inference is clearly invalid if 'better' is replaced by 'more likely'. However, Kratzer's theory wrongly predicts that this inference should be valid with epistemic and deontic modals alike. This difference turns out to be closely related to a similar division among non-modal adjectives between additive properties such as height and weight and non-additive properties such as temperature and danger. I develop an algebraic semantics of degree built on the Representational Theory of Measurement (Krantz et al. 1971) which is able to model both kinds of properties in a straightforward manner, along with differences in boundedness and other properties of scales which have been the subject of recent work on gradability. By examining a wide range of data involving degree modification and entailments with gradable modal adjectives and verbs, I show that this model provides a good fit for epistemic, deontic, and bouletic modals, and explains a number of old and new puzzles in the semantics of epistemic and deontic modals while avoiding the incorrect predictions of quantificational theories which I discuss. The most important result involving epistemic modals is that the distribution of degree modifiers with epistemic modals and their behavior on other standard tests for scale structure suggests that this scale is both upper- and lower-bounded, a fact which -- in combination with additivity -- makes this scale provably equivalent to a probability measure. Given this result, we have little choice but to embrace a probabilistic semantics for the modal auxiliaries as well. I show that this semantics avoids the incorrect predictions made by quantificational approaches, including Kratzer's. Turning to deontic and bouletic modals, a number of independent arguments show that these items are non-monotonic and exquisitely information-sensitive, two facts which are deeply troubling for quantificational theories but simple to implement in a scalar approach. In addition to gradability and comparison, the model gives a simple explanation of the possibility of conflicts of obligation and desire, the division between 'weak' and 'strong' necessity modals which has been the subject of much recent work, and the focus- and alternative-sensitivity of weak necessity modals along with 'likely' and 'probable'. The dissertation contains new empirical and formal results which cast serious doubt on the dominant theory of modality, or indeed any attempt to treat modals as quantifiers over possible worlds. Instead, I argue that modals are scalar expressions, and in particular that epistemic modals differ from deontic and bouletic modals in that the former are upward monotonic (additive) while the latter are non-monotonic (non-additive). Read more issues|LINGUIST home page|Top of issue Page Updated: 29-Sep-2011       About LINGUIST    |   Contact Us       While the LINGUIST List makes every effort to ensure the linguistic relevance of sites listed on its pages, it cannot vouch for their contents.