Center of Excellence in Southeast Asian Linguistics
Nattanun Chanchaochai

Selected Publications

Anaphoric Expressions in Thai Narratives: A Corpus Study on Accessibility and Distributional Tendency

This study examines the relationship between Thai anaphoric expressions and their antecedents’ Accessibility, analyzing 3,453 expressions from Thai-Zlatev Corpus (Zlatev and Yangklang 2001), Aakanee Website (Aakanee, “Thai Recordings”) and Thai Folktale Database (Princess Maha Chakri Sirindhorn Anthropology Centre, “Folktales”). Grounded in Ariel’s Accessibility Theory (1988, 1990), which asserts that referring expressions are universally arranged on an Accessibility scale but marking systems vary by language, we propose an Accessibility Marking Scale for Thai anaphoric expressions, informed by the factors of Distance, Competition, Saliency and Unity. Some anaphoric expressions showed no significant differences in mean Accessibility, leading them to share ranks. Our newly identified anaphoric expressions align on the scale with those of similar discourse function and pattern. The study also reveals that factors like Saliency and Unity account for the distinctions between the Thai Accessibility Marking Scale and the English Accessibility Marking Scale as proposed by Ariel.

Difficulties with pronouns in autism: Experimental results from Thai children with autism

This paper explores the acquisition of personal reference terms in Thai, a language with a highly complex personal reference system. Two separate studies were conducted for this paper, each featuring two groups of participants: children with typical development (TD) and children with autism spectrum disorders (ASDs). In each study, the participants were asked to complete two tasks on personal reference terms: a production task and a comprehension task. Overall, children with ASD performed on par in production, in terms of overall communicative success. However, an important finding was that they demonstrated a tendency toward pronoun avoidance, being less likely than children with TD to use deictic first-person pronominal forms. Instead, they preferred to use fixed referential terms for self-reference, contrasting with the children with TD’s preference for personal pronouns. The performance of children with ASD was significantly poorer in comprehension than that of children with TD. Children with ASD were generally able to detect lexically encoded person features but struggled with the more pragmatic and socially deictic aspects of personal reference terms. The latter also posed some challenges for children with TD, albeit to a lesser extent. In this regard, our results align with previous claims in the literature that lexical presuppositions are acquired earlier than implicated presuppositions. Our findings also add various new insights in terms of both population-specific effects in a language previously unstudied in this regard and the specific ways in which aspects of implicated presuppositions, i.e., the type of content in play, give rise to particular challenges in acquisition in general and for children with ASD in particular.

Investigating the Corner-Corn effect in Auditory Processing

We present two auditory-auditory priming experiments investigating whether decomposition effects for pseudo- related prime-target pairs like corner → CORN are restricted to early visual word recognition [10] or can also be found in auditory processing. Experiment 1 shows no difference in facilitation effects for pseudo-suffixed pairs and purely phonologically-related pairs

Auditory priming of pseudo-suffixed words

We present results from two auditory experiments that examine priming effects for pseudo-derived pairs of words (e.g., corner-corn), as compared to morphologically, phonologically, and semantically related pairs. Previous work shows facilitation for pseudo-derived pairs in a visual masked priming paradigm; our Experiment 1 presents the novel finding that these effects

GlobalTIMIT: Acoustic-Phonetic Datasets for the World’s Language

Although the TIMIT acoustic-phonetic dataset was created three decades ago, it remains in wide use, with more than 20000 Google Scholar references, and more than 1000 since 2017. Despite antiquity and relatively small size, inspection of these references shows that it is still used in many research areas: speech recognition,