![]() ![]() Participants then completed a categorization task for sounds along /d/–/t/ and /b/–/p/ continua, presented in either the same voice as the lexical decision task or in a novel voice. A similar result was reported by Kraljic and Samuel (2006), who had listeners complete a lexical decision task during which they were exposed to an artificial accent containing phonemes that were ambiguous between /d/ and /t/. (2008) exposed listeners to an artificial accent with lowered front vowels and found that after exposure, listeners showed higher lexical endorsement rates for novel words with lowered front vowels, but not for raised front vowels, showing that listeners were not simply relaxing their vowel category boundaries to accommodate variation. A number of studies using artificial accents have also found evidence that adjustment is accent-specific. Using a six-alternative forced choice task, they found that listeners who were more accurate at classifying an accent also tended to show higher accuracy on a transcription-in-noise task, suggesting that improved perception was due to familiarity with accent-specific phonetics. Atagi and Bent (2015) also found evidence for specificity in the relationship between listeners' accent categorization accuracy and intelligibility scores. For example, Bradlow and Bent, (2008) showed that exposure to multiple talkers of Chinese-accented English led to improved listener perception of a novel Chinese-accented speaker, but not a Slovakian-accented one, showing that adaptation was language-specific. This kind of mechanism has often been suggested to underlie listener adaptation to unfamiliar foreign accents, regional dialects, or speaker idiosyncrasies ( Eisner et al., 2013 Norris et al., 2003 Reinisch and Holt, 2014). As a result, they may actually narrow the space of acceptable realizations of a given category based on these shifted expectations. Under a targeted adaptation mechanism, we assume that listeners exploit the systematic patterns of variation in accented speech and adjust their phonetic category expectations in pattern-specific ways. ![]() Overall, these results provide evidence for a general adaptation mechanism, rather than a targeted mechanism involving accent-specific phonetic adjustments. This basic pattern held even when listeners were primed to expect congruent face-accent pairings (experiment 2). Crucially, however, listeners who correctly perceived the speaker as Chinese-accented showed no additional benefit over those who heard some other foreign accent. Results showed that although there was no significant effect of visual condition, listeners who believed the speaker to be non-natively accented enjoyed significantly improved performance compared to those who reported hearing a native accent. All listeners heard the same voice but were randomly assigned to one of four visual conditions: a blank silhouette, a European face, an East Asian face, or a South Asian face. In experiment 1, native speakers of American English completed a transcription-in-noise task with Chinese-accented English sentences. This study investigates the mechanism that underlies this facilitatory effect of top-down expectation, evaluating between general adaptation (an across-the-board relaxation of phonetic categorization criteria) and targeted adaptation (tuning in to accent-specific phonetics). Foreign-accented speech commonly incurs a processing cost, but this cost can be offset when listeners are given informative cues to the speaker's purported ethnicity and/or language background. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |