Effects of the variation of phoneme duration on word processing

Abstract

Contextually predictable, high frequency, competitor-dense words are often produced with less contrastive categories in informal conversation (Plug, 2011; Gahl et al., 2012; Tucker & Ernestus, 2016). One of the more frequent ways this manifests is through changes in phoneme duration, with shorter duration related to less careful speech (Gahl et al., 2012). However, initial observations point to large temporal variation occurring even in isolated words produced in controlled settings. The present study investigates how temporal variation affects processing speed for single words. A number of measures of temporal variation (e.g., word mean standardized phoneme duration) are compared, while controlling for a variety of psycholinguistic variables. Data from the Massive Auditory Lexical Decision project (Tucker & Brenner, 2016) was used. 232 native speakers of English responded to a subset of 26800 words from a full range of word types produced in isolation by a single speaker. Temporal variation measures are assessed based on their contribution to models predicting participant response latencies. Results offer insights into different operationalizations of temporal variation at the word level and how this durational variation influences speed of processing.