Individual variation in phonetic and phonological grammars

Individual variation is ubiquitous in both speech perception and production. Such variability is often assumed, especially within the variationist sociolinguistic tradition, to be the results and reflection of sociolinguistic factors such as age, gender, socio-economic status, dialect background, and language ideology. Nonetheless, individuals of the same age, gender, dialect, and socio-economic background may differ in patterns of speech perception and production in significant and systematic ways. While traditional phonetic research often dismisses such variability as random or a matter of idiosyncratic behavior of individual talker, we seek to leverage this type of variability to better understand the nature of language variation and change. Specifically, we aim to explicate a view of sound change where phonologization is understood to be an individual-level phenomenon; it takes place whenever an individual acquires a so-called “phonetic precursor” as an intended, and controlled, pattern in the language, even if the pattern exhibits gradient properties. Different phonologization outcomes often go unnoticed until there emerges a variation that is large enough to generate different descriptions in the coarse coinage of our shared language. Our work has demonstrated systematic variation in speech perception and production among native speakers of the same language, gender, and socioecononmic background.

To understand the source(s) of individual variation, we focus on elucidating what we referred to as “intake biases”, that is, intrinsic systematic attributes each speaker-listener contributes to the production and decoding of speech and language. The traditional appeal to channel biases (i.e., the type of articulatory, acoustic, auditory and perceptual constraints inherent to the vocal tract, along with the auditory and perceptual apparatus as alluded to above) is, perhaps paradoxically, not sufficient to answer this question since channel biases that are presumed to be universal are thus always present and not unique to any particular individual. Phonetic precursors that are shared by all members of a community are not expected to produce sound change because of their commonness. An idiolectal approach to linguistic change necessitates the identification of innovative individuals and the motivations underlying their exceptional behaviors. Our research has revealed that individual differences in autistic-like traits (e.g., attention switching skills, social skills, etc) significantly affect variability in perceptual compensation for coarticulation, phonotactic effects on speech perception, and coarticulation in production. Individuals with certain personality traits, such as high degree of openness and more focused attentional style, are more likely to phonetically accommodate to their interlocutor in their own production. Our recent work has focused on understanding the neuro-cognitive mechanisms underlying observed individual variabilities in speech perception and production.

Selected publications:

Sound systems in contact

Beyond the influence of language-internal factors, sound changes are, more often than not, the results of languages in contact. For example, some of our research in this area have focused on the contact of Cantonese with other languages. Together with collaborators in Hong Kong, we have investigated the phonetic systems of Cantonese as spoken by speakers of South Asian descent and by recent immigrants who settled in Hong Kong.

Selected publications:

The Phonetic Characteristics of Autim

The current understanding of the nature of speech production and perception in individuals with autism spectrum disorder (ASD) is quite impoverished, even though atypical language use is a core symptom in the clinical presentation and evaluation of ASD. The situation is even more acute for the speech production and perception of ASD individuals whose native language is not English. Our research aims to improve understanding and recognition of the clinical presentation of ASD, which has the potential to inform the design of diagnostic and assessment tools to better capture the presence and severity of ASD symptoms.

Selected publications: