Wide meanings across narrow contexts: Sakha particles and the grammar of alternatives

Abstract

A central empirical observation that this thesis considers is OBSERVABLE HOMOPHONY IN QUANTIFICATIONAL MORPHOLOGY — namely, in many languages there are morphemes which, depending on context, can perform two or more of the following functions: (i) noun quantifiers expressing meanings like every, any, some, ever, (ii) coordination expressing meanings like and, or, (iii) focus markers expressing meanings like also, even, only, (iv) markers of various types of questions (e.g. yes/no questions, wh-questions). A fundamental question this thesis seeks to answer is whether elements that perform such varied functions can be characterized as having a single, unified definition. When we consider the full array of meanings that such particles express within a given language and compare those meanings to homophonous particles in other languages, it is revealed that the clusters of functions are not picked at random, as they are not arranged in a “pickand-choose” fashion. Rather, there emerge two broad clusters of particles, which are referred to in the semantics literature as MO- and KA-particles, named after Japanese -mo and -ka. KA-particles are those which express meanings like existential some, disjunctive or, and form questions. MOparticles are those which express meanings like negative polarity any, universal every, and conjunction, and additive also, even focus. Based on the cross-linguistic clusters of similar-meaning particles, as well as the semantic plausibility that certain meanings pattern together naturally, I argue that we ought analyze the different functions of particles as semantically unified under a single definition, albeit heavily underspecified. In particular, I focus in large part on MO-particles, drawing on evidence from well-studied languages like Japanese and Hungarian. It is argued that MO-particles in particular are best understood as elements which activate the alternatives of their hosts, resulting in the varied wide array of meanings they can display, which is at once varied and highly contextually restricted. To this literature I add new data from two understudied Turkic languages: Sakha (Yakut) and Tuvan (Tyvan), based on original elicitations with native speakers. A main focus of this dissertation is an in-depth description and semantic analysis of four unique particles observed in Sakha: daɣanï∼da, eme, ere, and baɣarar. When combined with a host interrogative pronoun, these particles form polarity-sensitive indefinites. It is argued that the main semantic contribution the particles make is to activate the alternatives of their host, adopting an exhaustification-based theory of alternative semantics. A new contribution this thesis makes is the argument that particles of this kind can activate subsets of their hosts’ alternatives. By encoding these particles as ALTERNATIVE ACTIVATORS, we can account for their complex distributions. Of these four Sakha particles, the most difficult to account for is daɣanï∼da. This particle displays properties of MO-/TOO-particles. With interrogative pronoun hosts, it forms strict negative polarity items like kim daɣanï/da ‘anybody’ ( ‘who’); however, one interpretation that it lacks which makes it typologically uncommon among particles of this type is a plain additive unary toofocus reading. One main point of complexity is that it displays a binary too-focus reading, i.e. X da Y da/X daɣanï Y daɣanï means ‘both X and Y.’ This presents an intriguing puzzle, which is ultimately answered with reference to a cross-linguistic hypothesis. It is argued that MO-particles induce additivity not via a presupposition, but rather as a post-supposition which is created during the course of exhaustification. It is proposed that in Sakha, daɣanï∼da lacks the crucial toofocus reading because the language has a dedicated additive presupposition particle emie ‘also; again.’ Presupposition-checking happens before exhaustification, thus bleeding the potential for daɣanï∼da to induce additivity; it is only in contexts where presuppositions cannot project that we can observe additivity within it (e.g. in the binary coordination constructions, where presuppositions cannot project from the right to the left). This provides strong evidence for a limited adoption of post-suppositions as a genuine grammatical phenomenon (albeit one derived by another process), in that it demonstrates that the insertion of lexical items can be sensitive to the ordering of presupposition-checking and exhaustification.