This article reports on patterns in the production and perception of New Zealand English r-sandhi. We report on two phoneme-monitoring experiments that examine whether listeners from three regions are sensitive to the distribution of r-presence in linking and intrusive environments. The results provide evidence that sound perception is affected by a listener's experience-driven expectations: greater prior experience with a sound in a given context increases the likelihood of perceiving the sound in that context, regardless of whether the sound is present in the stimulus. For listeners with extremely limited prior exposure to a variant, the variant is especially salient and we also observe an experiment-internal effect of experience. We argue that our results support models that incorporate both word-specific and abstract probabilistic representations.

This study investigates how listeners associate acoustically different vowels with a single linguistic vowel quality. Listeners were asked to identify vowel sounds as /æ/ or /ʌ/ and to indicate the size of the speaker that produced them. Results indicate that perceived vowel quality trades off with the perception of speaker size: different vowels can sound the same, and the same vowel can sound different when a different speaker is perceived. These findings suggest that vowel normalization is broadly similar to perceptual constancy in other domains, and that social, indexical, and linguistic information play an important role in determining even the most fundamental units of linguistic representation.

Listeners have a remarkable ability to adapt to novel speech patterns, such as a new accent or an idiosyncratic pronunciation. In almost all of the previous studies examining this phenomenon, the participating listeners had reason to believe that the speech signal was produced by a human being. However, people are increasingly interacting with voice-activated artificially intelligent (voice-AI) devices that produce speech using text-to-speech (TTS) synthesis. Will listeners also adapt to novel speech input when they believe it is produced by a device? Across three experiments, we investigate this question by exposing American English listeners to shifted pronunciations accompanied by either a ‘human’ or a ‘device’ guise and testing how this exposure affects their subsequent categorization of vowels. Our results show that listeners exhibit perceptual learning even when they believe the speaker is a device. Furthermore, listeners generalize these adjustments to new talkers, and do so particularly strongly when they believe that both old and new talkers are devices. These results have implications for models of speech perception, theories of human-computer interaction, and the interface between social cognition and linguistic theory.

Understanding the relation between speech production and perception is foundational to phonetic theory, and is similarly central to theories of the phonetics of sound change. For sound changes that are arguably perceptually motivated, it is particularly important to establish that an individual listener's selective attention—for example, to the redundant information afforded by coarticulation—is reflected in that individual's own productions. This study reports the results of a pair of experiments designed to test the hypothesis that individuals who produce more consistent and extensive coarticulation will attend to that information especially closely in perception. The production experiment used nasal airflow to measure the time course of participants' coarticulatory vowel nasalization; the perception experiment used an eye-tracking paradigm to measure the time course of those same participants' attention to coarticulated nasality. Results showed that a speaker's coarticulatory patterns predicted, to some degree, that individual's perception, thereby supporting the hypothesis: participants who produced earlier onset of coarticulatory nasalization were, as listeners, more efficient users of nasality as that information unfolded over time. Thus, an individual's perception of coarticulated speech is made public through their productions.

This study investigates second language (L2) phonetic categorization and phonological encoding of L2 words (hereafter, phonolexical encoding1) with phonemic and allophonic cross-linguistic mismatches. We focus on the acquisition of Spanish /ɾ/-/l/ and /ɾ/-/t/ contrasts among Spanish learners with American English (AE) and Mandarin Chinese (hereafter, Chinese) as first languages (L1s). [ɾ] and [t] are positional allophones in AE but separate phonemes in Spanish. The phoneme /ɾ/ is absent in Chinese. AE learners showed nativelike phonetic categorization and little between-contrast difference in phonolexical encoding, suggesting that L1 positional allophony does not necessarily impede L2 contrast acquisition. Chinese learners showed persistent perceptual difficulties with both contrasts due to perceptual similarity. Phonetic categorization significantly predicted phonolexical encoding for /ɾ/-/t/ contrasts for Chinese learners bidirectionally, while AE learners showed this relationship only when /t/ was incorrectly replaced by /ɾ/ in Spanish words. This asymmetry can be driven by the fact that [t] is the dominant allophone of /t/ in AE, while [ɾ] is a positional allophone. It suggests L1 allophonic knowledge heightens perceptual monitoring when evaluating substitutions that conflict with L1 phonological expectations. This study calls for more nuanced treatment of L1 influence in L2 phonological acquisition models, especially at the allophonic level.

The development of a sound change can be influenced by linguistic and social factors, both within the language community and from cases of language contact. The present study is an examination of the internally generated ongoing tonogenesis process in Afrikaans, specifically analyzing production and perception of word-initial plosives among different age and gender groups. Results show that female speakers are devoicing significantly more often than male speakers, and the perception of female voices is influenced more by f0 levels than the perception of male voices. This study finds that gender is a larger predictor overall of tonogenetic patterns than age.*

This study revisits the relationship between second-language (L2) learners’ ability to distinguish sounds in non-native phonological contrasts and to recognize spoken words when recognition depends on these sounds, while addressing the role of methodological similarity. Bilingual Catalan/Spanish learners of English were tested on the identification of two vowel contrasts (VI) of diverging difficulty, /i/-/ɪ/ (difficult) and /ɛ/-/æ/ (easy), in monosyllabic minimal pairs, and on their recognition of the same pairs in a word-picture matching task (WPM). Learners performed substantially better with /i/-/ɪ/ in VI than in WPM, and individual scores were only weakly correlated. By replicating previous findings through a more symmetrical design, we show that an account of prior work rooted in methodological dissimilarity is improbable and provide additional support for the claim that accuracy in sound identification does not guarantee improvements in word recognition. This has implications for our understanding of L2-speech acquisition and L2 pronunciation training.

Multilinguals face greater challenges than monolinguals in speech perception tasks, such as processing noisy sentences. Factors related to multilinguals’ language experience, such as age of acquisition, proficiency, exposure and usage, influence their perceptual performance. However, how language experience variability modulates multilinguals’ listening effort remains unclear. We analyzed data from 92 multilinguals who completed a listening task with words and sentences, presented in quiet and noise across participants’ spoken languages (Arabic, Hebrew and English). Listening effort was assessed using pupillometry. The results indicated higher accuracy and reduced effort in quiet than in noise, with greater language experience predicting better accuracy and reduced effort. These effects varied by stimulus and listening condition. For single words, greater language experience most strongly reduced effort in noise; for sentences, it had a more pronounced effect in quiet, especially for high-predictability sentences. These findings emphasize the importance of considering language experience variability when evaluating multilingual effort.

One of the main challenges individuals face when learning an additional language (L2) is learning its sound system, which includes learning to perceive L2 sounds accurately. High variability phonetic training (HVPT) is one method that has proven highly effective at helping individuals develop robust L2 perceptual categories, and recent meta-analytic work suggests that multi-talker training conditions provide a small but statistically reliable benefit compared to single-talker training. However, no study has compared lower and higher variability multi-talker conditions to determine how the number of talkers affects training outcomes, even though such information can shed additional light on how talker variability affects phonetic training. In this study, we randomly assigned 458 L2 Spanish learners to a two-talker or six-talker HVPT group or to a control group that did not receive HVPT. Training focused on L2 Spanish stops. We tested performance on trained talkers and words as well as several forms of generalization. The experimental groups improved more and demonstrated greater generalization than the control group, but neither experimental group outpaced the other. The number of sessions experimental participants completed moderated learning gains.

This study investigated the neural mechanisms underlying bilingual speech perception of competing phonological representations. A total of 57 participants were recruited, consisting of 30 English monolinguals and 27 Spanish-English bilinguals. Participants passively listened to stop consonants while watching movies in English and Spanish. Event-Related Potentials and sLORETA were used to measure and localize brain activity. Comparisons within bilinguals across language contexts examined whether language control mechanisms were activated, while comparisons between groups assessed differences in brain activation. The results showed that bilinguals exhibited stronger activation in the left frontal areas during the English context, indicating greater engagement of executive control mechanisms. Distinct activation patterns were found between bilinguals and monolinguals, suggesting that the Executive Control Network provides the flexibility to manage overlapping phonological representations. These findings offer insights into the cognitive and neural basis of bilingual language control and expand current models of second language acquisition.

Growing evidence suggests that ratings of second language (L2) speech may be influenced by perceptions of speakers’ affective states, yet the size and direction of these effects remain underexplored. To investigate these effects, 83 raters evaluated 30 speech samples using 7-point scales of four language features and ten affective states. The speech samples were 2-min videorecordings from a high-stakes speaking test. An exploratory factor analysis reduced the affect scores to three factors: assuredness, involvement, and positivity. Regression models indicated that affect variables predicted spoken language feature ratings, explaining 18–27% of the variance in scores. Assuredness and involvement corresponded with all language features, while positivity only predicted comprehensibility scores. These findings suggest that listeners’ perceptions of speakers’ affective states intertwine with their spoken language ratings to form a visual component of second-language communication. The study has implications for models of L2 speech, language pedagogy, and assessment practice.

Twenty-five years ago, the publication of an article by Pallier, Colomé, and Sebastián-Gallés (2001) launched a new and rapidly evolving research program on how second language (L2) learners represent the phonological forms of words in their mental lexicons. Many insights are starting to form an overall picture of the unique difficulties for establishing functional and precise phonolexical representations in L2; however, for the field to move forward it is pertinent to outline its major emerging research questions and existing challenges. Among significant obstacles for further research, the paper explores the current lack of theoretical agreement on the concept of phonolexical representations and the underlying mechanism involved in establishing them, as well as the variable use of the related terminology (e.g., fuzziness and target-likeness). Methodological challenges involved in investigating phonological processing and phonolexical representations as well as their theoretical implications are also discussed. To conclude, we explore the significance of L2-specific phonological representations for the bottom-up lexical access during casual, conversational speech and how our emerging knowledge of L2 lexical representations can be applied in an instructional setting as two potentially fruitful research avenues at the forefront of the current research agenda.