This dissertation examines the syntax and semantics of intensification and attenuation in English through four cases studies. These case studies provide a way of addressing two questions on the nature of intensification and attenuation. First, what components can intensification and attenuation be decomposed into, and are these components shared across various constructions? Second, can instances of intensification and attenuation be unified under one theoretical framework, or are intensification and attenuation broad terms for disparate phenomena?Chapter 2 focuses on the modifiers sorta and kinda. These modifiers are of interest due to their cross-categorial nature, being able to modify noun phrases, verb phrases, and adjective phrases. When composed with a gradable category, such as a gradable adjective (e.g., sorta tall), these degree words weaken entailments to the standard. When used with a non-gradable category (e.g., sorta swim), they weaken the conditions when the non-gradable category can be used, allowing it to be used imprecisely. I adapt the framework in Morzycki 2011, supposing that natural language expressions have flexible denotations corresponding to pragmatic halos, in the sense of Lasersohn (1999). These halos are linked to a degree of precision on the interpretation function. Typeshifting mechanisms allow this degree of precision to be accessed through grammatical meanings, coercing predicates from being non-gradable into gradable, with the degree of precision providing the scale along which to grade the predicate.The analysis of sorta in chapter 2 is extended to very in chapter 3. Canonically, when very is used with a gradable adjective, it asserts that the adjective holds to a high degree. However, there exist other cases where very is used with a nominal, such as in the very center of the Earth and I spoke with this very person, as well as with ordinals (the very first person in line). I argue that these are imprecision-related uses of very, and that, like with sorta/kinda, an implicit typeshift is used toconvert these noun phrase into predicates that are graded by their degree of precision. In keeping with its use in the adjectival domain, very also asserts that these predicates are to hold to a high degree—in this case, a high degree of precision.In chapter 4, I examine the use of some as a numeral modifier, as in twenty-some people were at the party. These cases commit the speaker to ignorance about which particular number satisfies a claim. Moreover, these examples have both a lower bound, coming from the modified numeral and an upper bound due to the syntax of the numeral. I build a syntax for these constructions, and adapt Alonso-Ovalle & Menéndez-Benito’s 2010 analysis of algún in order to show how the ignorance effect is derived from presuppositions on some.Finally, chapter 5 focuses on some in a type of exclamative construction using the determiner some. These are examples such as John is some lawyer!. I show that these some-exclamatives are constrained in that the noun phrase that some combines with must be able to be construed so that subkinds can be associated with it. In analyzing these exclamatives, I adopt a question-theory of exclamatives in the style of Zanuttini & Portner (2003), where exclamatives underlyingly make use of an alternative semantics in the style of Hamblin 1973. The existence of exclamatives being built from an indefinite such as some provides additional support for exclamatives more generally being an alternative-sensitive construction.These case studies shed light on various components that underly intensification and attenuation. First, chapters 2 and 3 show how imprecision and slack regulation can be modeled using a degree semantics, as well as a special typeshifting mechanism that transforms non-gradable predicates into gradable predicates by grading them based on precision. Chapter 4 shows how properties of the epistemic determiner some are used in generating ignorance effects with numerals and building approximate meanings. Finally, chapter 5 shows how speakers exclaim about kinds and subkinds, and how exclamative constructions depend on alternative-generating constituents (whether they are questions or indefinites). The variety of analytical tools used suggests that intensification and attenuation are not primitive theoretical notions and should not be unified.
Read
