Cross-linguistically, expressive language tends to diverge in various theoretically-significant ways relative to plain (non-expressive) language. Studying it has therefore been useful in compelling us to adapt our theoretical assumptions, leading to innovations in understanding the morphological dimension of expressives (e.g. Fortin 2011) and multidimensional semantics (e.g. Potts 2005), for instance. In this talk I investigate some distributional and semantic properties of a set of expressive clitics in Kwak'wala, a Wakashan language spoken on the central coast of British Columbia, and ask whether they present any anomalies that our theory must adapt to account for. Specifically, I discuss the clitics =bidu 'diminutive (singular)', ='məniXw 'diminutive (plural)', =dzi 'augmentative', =Gas 'poor thing', and =kəna'l 'how nice'. Along the way, I also chronicle key aspects of the fieldwork methodology I used to study them.
Within words, expressive clitics in Kwak'wala follow a common cross-linguistic pattern of appearing on the border between derivational and inflectional material. Within sentences, their distribution gets more interesting: expressive clitics can appear on the predicate, on arguments, on modifiers, and on auxiliaries, and may re-occur in several different places within the same sentence (though apparently not within the same prosodic word). While this distributional mayhem may suggest that expressive clitics are indeed anomalous within the grammar, I'll question to what extent this is actually the case given somewhat comparable distributional properties of the future tense clitic (=tl).
Next we'll consider some basic semantic properties of expressive clitics, including the question of whether (and if so, how) the syntactic position of expressive clitics constrains semantic interpretation. At this point we'll observe a basic split between the diminutives and augmentative, which appear to express descriptive content in certain contexts, and the other expressives which only convey speaker-oriented meaning. Finally, I'll consider whether Kwak'wala expressives raise any new puzzles for semantic theory.
Along the way we'll look at methods designed to elicit expressive clitics in a fieldwork context, recognizing that there are unique challenges involved in the elicitation of the expressive dimension.