Exploring the Use of an Artificial Accent of English to Assess Phonetic Learning in Monolingual and Bilingual Speakers

Differences between Monolinguals and Bilinguals in Phonetic and Phonological Learning and the Connection with Auditory Sensory Memory

“…The aim of the current study was to address the prediction put forth in the previous section, which postulates a link between phonetic and phonological learning and auditory sensory memory, an experiment was designed to include a novel accent learning task, following [ 28 , 29 ] and a digit span task with a suffix [ 46 ]. The experiment was conducted with monolingual and bilingual speakers in person in a quiet room, inside a sound-attenuated booth, on the CUNY Kingsborough Community College campus and comprised the following parts: a language background questionnaire, a translation task for bilingual participants (from English into their other language)—not discussed here, a novel accent learning task that included three blocks (i.e., baseline, training, and testing), and a digit span task with a suffix.…”

“…Although the two groups performed very similarly during the training when they were asked to imitate what they heard immediately, bilinguals produced their glottal stop significantly more frequently than monolinguals during the post-training session. A follow-up study [ 29 ] employed a novel accent that was artificially created to have four phonetic features that differed from standard American English. The decision to use an artificially constructed accent instead of a natural one was made to allow better control over the measurements of the input and the output in the experiment.…”

Underpinnings of explicit phonetic imitation: perception, production, and variability

“…Phonetic imitation is a complex behavior, requiring accurate perception, identification, and (re-) production of the relevant characteristics of the target of imitation. Previous work has shown that listeners are sensitive to, and can to some extent spontaneously reproduce, phonetic properties of familiar accents when asked to mimic them from memory (e.g., Flege and Hammond, 1982;Mora et al, 2014), and that they can imitate properties of novel (D'Imperio et al, 2014;Spinu et al, 2018) or artificial (Spinu et al, 2020) accents, or phonetic variation in words or syllables (Olmstead et al, 2013;Dufour & Nguyen, 2013). The variability found within these studies reflects the wide range of imitative ability in the real world.…”

“…Compared to the small number of studies exploring perceptual or articulatory predictors of imitation, there has been considerable interest in examining which characteristics of individual speakers, and their attitudes toward the interlocutor or the target speech, might predict imitation. A multitude of factors, including bilingualism (Spinu et al, 2020), musical experience (Coumel et al, 2019), and general cognitive processes (working memory: Reiterer et al, 2011; neurocognitive flexibility: Reiterer et al, 2013) have been shown to correlate with individuals' extent of imitation. The list of social or personality-related factors proposed to condition imitation expands even further when considering the better-studied domain of phonetic convergence (see Wade et al, 2020 for a review).…”

“…To maintain full control over the acoustics of the material presented to participants, while still allowing for the voice-related variability that is present in naturally-occurring accents, we created artificial accents using sentence-level stimuli that differ minimally in specific phonetic features. Artificial accents can be constructed in a variety of ways: by using a speech synthesizer (e.g., Maye et al, 2008), by manipulating natural speech (Yu et al, 2013), or by recording natural productions produced with systematic phonetic/phonological differences (Adank & Janse, 2010;Spinu et al, 2020), and they have been shown to elicit phonetic imitation in past work. For example, using an explicit imitation task, Spinu et al (2020) found that bilinguals showed more imitation than monolinguals of an artificial accent of English which varied from a canonical accent in four features (/ɛ/ -> /jɛ/, intervocalic /l/ -> /ɾ/, /ə/-epenthesis in s+stop clusters, and a novel intonation pattern in tag questions), and Yu et al (2013) showed that participants produced longer VOTs after exposure to a narrative with systematically lengthened VOT of voiceless stops.…”