Many current theories of speech production incorporate interaction. In such theories, central cognitive processes (i.e., lexical access: selecting the word you will use to express an idea) modulate or influence more peripheral processes (i.e., phonetic processes that plan and execute the precise articulatory movements that produce speech). These accounts predict cognitive disruptions should also disrupt phonetic processing. For example, if cognitive processing is delayed, there should be corresponding delays in phonetic processing- causing speakers to not only initiate responses more slowly but also take more time to articulate words. However, such interactive effects have not been consistently observed in previous work. To reconcile these inconsistent results with interactive accounts, this project examines the hypothesis that the strength of interaction reflects the instability of processing. Less stable processing results in a higher degree of interaction;more stable processing will lead to weaker effects of interaction. To examine this hypothesis, this project tests for interactive effects in two populations where lexical access processes are less robust than those of the young monolingual speakers examined in previous work: bilinguals and older adults. Bilingual lexical access will be disrupted by requiring participants to mix and switc between their two languages. Interactive theories predict that when mixing and switching result in significant disruptions to lexical access (as indexed by effects in reaction time), similar cost are expected to occur in spoken durations. Similarly, significant disruptions to lexical access are predicted to result in a significant increase in the intrusion of native-language phonetic properties during second language production. Lexical access in older adults will be disrupted using experimental paradigms that induce word substitution errors;these significant disruptions to lexical access are predicted to result in significant phonetic effects. The investigation of gradient phonetic properties under this wide range of processing conditions will be enabled by novel speech analysis tools. These will allow robust, automatic observation of theoretically-motivated phonetic properties that previous work was forced to examine manually.
This project will examine how difficulties in the cognitive processes involved in language production can lead to difficulties in articulating speech. This wil allow us to better understand how language and articulatory disorders interact: when language disorders will produce difficulties in speech articulation as well as cases where articulatory disorders will be exacerbated by difficulties in cognitive processing.
|Goldrick, Matthew; McClain, Rhonda; Cibelli, Emily et al. (2018) The influence of lexical selection disruptions on articulation. J Exp Psychol Learn Mem Cogn :|
|Gustafson, Erin; Goldrick, Matthew (2018) The role of linguistic experience in the processing of probabilistic information in production. Lang Cogn Neurosci 33:211-226|
|Fink, Angela; Oppenheim, Gary M; Goldrick, Matthew (2018) Interactions between Lexical Access and Articulation. Lang Cogn Neurosci 33:12-24|
|Adi, Yossi; Keshet, Joseph; Cibelli, Emily et al. (2017) SEQUENCE SEGMENTATION USING JOINT RNN AND STRUCTURED PREDICTION MODELS. Proc IEEE Int Conf Acoust Speech Signal Process 2017:2422-2426|
|Li, Chuchu; Goldrick, Matthew; Gollan, Tamar H (2017) Bilinguals' twisted tongues: Frequency lag or interference? Mem Cognit 45:600-610|
|Adi, Yossi; Keshet, Joseph; Cibelli, Emily et al. (2016) Automatic measurement of vowel duration via structured prediction. J Acoust Soc Am 140:4517|
|Fricke, Melinda; Baese-Berk, Melissa M; Goldrick, Matthew (2016) Dimensions of similarity in the mental lexicon. Lang Cogn Neurosci 31:639-645|
|Goldrick, Matthew; Keshet, Joseph; Gustafson, Erin et al. (2016) Automatic analysis of slips of the tongue: Insights into the cognitive architecture of speech production. Cognition 149:31-9|
|Adi, Yossi; Keshet, Joseph; Goldrick, Matthew (2015) VOWEL DURATION MEASUREMENT USING DEEP NEURAL NETWORKS. IEEE Int Workshop Mach Learn Signal Process 2015:|
|Fink, Angela; Goldrick, Matthew (2015) The Influence of Word Retrieval and Planning on Phonetic Variation: Implications for Exemplar Models. Linguist Vanguard 1:215-225|