Listeners generally map continuous acoustic information onto categories in a gradient manner with varying individual differences. Typically, such individual differences in speech categorization have been characterized by the mean slope of the response function, as quantified through the visual analog scaling (VAS) task. However, recent evidence suggests that categorization consistency (trial-by-trial response variability) may be a more robust predictor of real-world outcomes than the mean slope. Despite this, the extent to which these VAS indices represent reliable and stable traits relevant to speech perception remains uncertain. This study investigates the stability and trait-like nature of VAS indices by analyzing VAS responses across a diverse range of speech continua. We also examine the potential associations between differences in speech categorization and broader cognitive differences. American English-speaking adults (n = 68) completed the VAS task, alongside three questionnaires assessing autistic traits, anxiety, and impulsivity. We found that categorization consistency showed stronger correlations across continuum types than the categorization slope. Furthermore, no significant correlations were observed between the VAS indices and broader cognitive factors. These results suggest categorization consistency is a more stable property of individuals over the mean slope, and importantly, it is not an artifact due to higher cognitive factors. Thus, the extent to which listeners are consistent in their speech categorization may provide a more accurate characterization of individual differences in speech perception.
This exploratory study investigates the acoustic correlates of the Korean three-way laryngeal stop distinction in Gyeongsang long-term (LT) transplants who were born in the Gyeongsang region but moved to Seoul to pursue higher education. Acoustic data were collected from eight LT transplants, five short-term (ST) transplants, and 11 Seoul speakers to examine whether exposure to Seoul Korean (SK) affects Gyeongsang speakers’ cue-weighting in distinguishing stops in production. LT transplants produced stimuli in both Gyeongsang and Seoul dialects. A cue-weighting model based on the acoustic data reveals that voice onset time (VOT) is less important to distinguish lenis from aspirated stops for Seoul speakers and for LT transplants’ SK, as compared to ST transplants and LT transplants’ Gyeongsang Korean (GK). In addition, fundamental frequency (F0) is more important for the lenis–aspirated distinction for Seoul speakers and LT transplants’ SK, as compared to ST and LT transplants’ GK, showing that LT transplants rely less on VOT and more on F0 to distinguish lenis from aspirated stops compared to ST transplants. LT transplants’ SK reveals that they rely more on VOT and less on F0 compared to SK speakers. The cue-weighting model of the LT transplants provide empirical evidence that a series of sound changes in GK is due to inter-dialect contact.
A critical aspect of spoken language development is learning to categorize the sounds of the child's language(s). This process was thought to develop early during infancy to set the stage for the later development of higher‐level aspects of language (e.g., vocabulary, syntax). However, many recent studies have shown that speech categorization continues to develop through adolescence. This work, which uses more sophisticated forced‐choice tasks and stimuli that span speech continua (e.g., /b/ to /p/ in small steps) suggests that children show increasingly sharp category boundaries over development. However, this contrasts with the consensus view in adult speech perception that speech categories are graded and this may help listeners be more flexible. Here, our longitudinal study revisited these results using a new Visual Analogue Scaling (VAS) task, which employs a continuous rating scale to potentially unpack the underlying cause of a shallower or more gradient boundary. We tested 225 school‐aged children (in grades 1–3) over 4 years (through grades 4–6). A Bayesian Hierarchical psychometric model was fit to extract the Slope of the categorization function and Response Consistency (the trial‐by‐trial variance). Results show that as children age, they show shallower (more gradient) slopes. It also highlighted a new variable, suggesting that older children exhibit greater trial‐by‐trial categorization consistency. These findings suggest that sound categorization continues to develop through early school age. These findings also suggest that children become increasingly sensitive to fine‐grained gradient detail in the signal. The dramatic changes in consistency represent a challenge to current theoretical models that focus primarily on underlying category representation. Summary: Speech categorization continues to develop through the school years, refining children's ability to categorize speech sounds more precisely and flexibly. A longitudinal study tested 225 school‐aged children using a Visual Analogue Scaling (VAS) task and Bayesian Hierarchical psychometric modeling. With age, children showed shallower (more gradient) categorization slopes and greater trial‐by‐trial response consistency. Findings challenge theoretical models, revealing continued development of speech categorization and increasing sensitivity to fine‐grained gradient detail in the signal.
Purpose: Speech perception continues to develop throughout school age and plays a fundamental role in language and reading development. Recent findings in English-speaking children suggest that speech categorization consistency—the stability of a listener's percept across multiple encounters with a speech sound—predicts both language and reading abilities, with a particularly strong link between reading skills and vowel perception. One hypothesis is that this is due to complex grapheme–phoneme correspondences (GPCs) in English vowels. The present study tested (a) whether the relationship between categorization consistency and language/reading abilities extends to typologically different languages and (b) whether the vowel-specific link observed in English is shaped by GPC complexity, using data from Korean, a language with relatively transparent GPCs. Method: Forty-four first-grade Korean-speaking children completed a visual analog scale task, in which they heard tokens from a speech continuum and rated the correspondence between the stimulus and each word on a continuous scale. Standardized assessments of language and word reading were also conducted. Results: Children with poorer language/reading abilities exhibited lower categorization consistency, which is consistent with the findings from English-speaking children. In contrast to prior findings, however, this relationship was not specific to vowel perception but was held across all contrast types tested. Categorization gradiency (slope of the categorization function) was not significantly associated with any outcome. Conclusions: These findings extend prior work by demonstrating that categorization consistency predicts language and reading abilities, even in a language with transparent GPCs. Importantly, this association was observed across all phonemic contrasts, not just vowels—suggesting that the previously observed vowel-specific link in English may stem from the greater GPC complexity of English vowels. Together, while categorization consistency appears to be a critical predictor of linguistic outcome, the specific pattern may vary across languages depending on the structure of their GPCs. Supplemental Material: https://doi.org/10.23641/asha.31907296