Spoken Expression of Individual Identity and the Listener*

personal impression of a talker. The speech of someone familiar can have a subtle impact, and even an unfamiliar voice suggests traits, true or false, of the one who spoke. It should not be surprising that qualitative variation in speech has commonly been identified as the source of these personal impressions, and a substantial technical literature reports investigations of the acoustic correlates of regional and social group, of a talker’s age and sex, and of a talker’s affective state and arousal (classics include: Abercrombie, 1967; Bricker & Pruzansky, 1976; Hecker, 1971; a recent review is offered by Kreiman, VanlanckerSidtis & Gerratt, 2005). But, the custom of ascribing impress ions of persona l qua l i ty so le l y to the extralinguistic properties of speech has led to neglect of a significant dimension of individual variation that can be critical to the expression and perception of personal identity. An individual’s characteristic way of producing the consonants and vowels that compose words can be distinctive even when a spoken expression is otherwise unremarkable, prosaic, and normative. Although an individual’s articulation of speech is a linguistically regulated aspect of communication, personal consistency in this aspect of symbolic expression provides a rich assortment of indexical attributes. Some recent and new studies draw attention to the potential role of idiolect, the individual linguistic characteristics nested within dialect, in individual identification. To expose the argument and evidence, it is useful to consider the precise circumstances in which we speak and listen. A gloss of the production of language spotlights this condition (Levelt, 1989). In a conversation, an urge to take a turn initiates a semantic aim, an intention to create an utterance of a certain dimension. This goal initiates a compositional function to choose words and to assemble them into an order with representational properties that more or less match the semantic aim. There are lots of ways to choose words: by shade of meaning, perhaps, or also commonness or rarity, or by alliteration or rhyme. The dimensions of variation among the words of a language include semantic properties and differential incidence, and variation is well studied in each of these. Alliteration and rhyme are characteristics of the meaningless phonological form by which words are distinguished. Phonemic form amounts to an addressing scheme for items in lexical memory, marking distinctions among words, although this dimension of variation also supplies a set of abstract criteria for articulating the word and making a private linguistic composition public.1 Through this last juncture in production, talkers and listeners within a community share the language forms that are used to compose utterances. This cheerful confluence of symbolic processes is compromised, however, by vocal dispersion when language is spoken. Technical Report Barnard College Speech Perception Laboratory February, 2008