UC Berkeley

This dissertation explores affective expression in Yaminawa, a Panoan language of Peruvian Amazonia. In this study, ‘affect’ is used to refer broadly to the English language concepts of ‘emotion’ and ‘feeling’. Affective expression is approached as an interactional phenomenon and it is analyzed in terms of affective stancetaking, i.e., the way speakers position themselves to objects in the discourse as well as their interlocutors via linguistic performance. This study considers affective resources at the levels of the lexicon, morphology, prosody, acoustics (voice quality, speech rate and volume, etc.), and interactional features (turn duration, complexity of backchannels, etc.). This study contextualizes affective expression in Yaminawa with a detailed description of Yaminawa ethnopsychology and the lexical resources that describe affective states, as well as behaviors and bodily sensations that are associated with particular affects by the Yaminawa. Using methods from Cognitive Anthropology, I investigate the ways that native Yaminawa speakers categorize emotion terms, and show that prosociality vs. antisociality is a major cultural axis along which emotion terms are conceptually organized. This dissertation also provides both a general ethnographic sketch of daily life among the Yaminawa community of Sepahua and a grammar sketch of the Yaminawa language.

Yaminawa is notable for its rich inventory of bound morphemes that are used in affective expression. Some of the affective categories expressed by these bound morphemes, such as sadness, appear to be typologically unusual. In everyday conversation, certain

morphological, acoustic, and interactional features cluster together in recurrent affective ways of speaking that are identifiable by speakers even when the propositional content of the utterances cannot be clearly heard. This dissertation describes two salient affective ways of speaking in detail: shĩ́nã̀ì ‘sad’ speech and sídàì ‘angry’ speech. Shĩ́nã̀ì ‘sad’ speech is characterized by creaky voice, low speech volume, and high frequency and complexity of backchannelling by co-participants, among other features. Sídàì ‘angry’ speech is characterized by breathy voice, slow and rhythmic speech rate, and scarcity and simplicity of backchannels. I also briefly describe the key features of three additional, minor affective stances: dúì ‘affection’, rátèì ‘surprise’, and bésèì ‘fear’. Some affective resources are used in more than one type of affective speech, for example, high pitch is used in affectionate speech, surprised speech, and commands issued in angry speech. Other affective resources appear to be unique to a single affective type, such as delayed stop release in fearful speech.

While previous descriptions of affective expression in individual languages have tended to focus on single levels of analysis, such as metaphor or morphology, this dissertation aims to provide a model for the holistic description of affective expression in an individual language.