Orthographic congruency effects in the suprasegmental domain: Evidence from Thai

Abstract

The influence of orthographic knowledge on lexical tone processing was examined by manipulating the congruency between the tone and the tone marker of Thai monosyllabic words presented in three metalinguistic tasks. In tone monitoring (Experiment 1) and same–different tone judgement (Experiment 2)—that is, tasks that require an explicit analysis of tone information—an orthographic congruency effect was observed: Better performance was found when both tone and tone marker led to the same response than when they led to opposite, competing responses. In rhyme judgement (Experiment 3), a metaphonological task that allows tone to be processed in a more natural way since it does not require explicit analysis of tone, the orthographic effect emerged only when the interstimulus interval was lengthened from 30 to 1,200 ms. In addition to demonstrating the generalization of the orthographic congruency effect to the suprasegmental domain in Thai, the present data also suggest relatively late and task-dependent activation of orthographic representations of tone.

Keywords:

• Metaphonology
• Orthographic congruency effect
• Suprasegmental element
• Thai lexical tone

Acknowledgments

This research was supported by two grants from the Belgian French community (Direction Générale de la Recherche Scientifique–Communauté Française de Belgique, and ARC 06/11–342), as well as by the Human Frontier Science Program RGP 53/2002 “An interdisciplinary approach to the problem of language and music specificity”. The first and the second authors are Postdoctoral Researcher and Senior Research Associate, respectively, of the Fonds de la Recherche Scientifique–FNRS. Special thanks to Patavee Charnvivit and the other members of the Centre for Research in Speech and Language Processing, Chulalongkorn University, for their extensive assistance during the first author's stay in Thailand. We also warmly thank Paulo Ventura, Guy Van Orden, and an anonymous reviewer for helpful comments on an earlier version of this article.

Notes

¹ The absence of a reliable interaction between orthographic congruency and material should, however, be considered with caution. Indeed, this statistical result does not truly reflect the mean RTs observed in Table 7 (i.e., a strong advantage of the congruent compared to incongruent trials for Material A, but a small advantage of the incongruent compared to congruent trials for Material B). This discordance may be explained by a lack of statistical power: At least 34% of the RT data in the P − O + condition and nearly 20% of the RT data in the P − O − condition had to be discarded from the analyses due to an exceptionally high error rate observed in these conditions.

² Similar analyses could not be performed on the different tone trials (i.e., P − O + and P − O − ) due to the high rate of missing RT data in the P − O + condition of the same–different tone judgement task. In addition, such a comparison would have been biased by the fact that it involves negative responses in the tone judgement task but positive responses in the rhyme judgement task.

³ Computed by using the necessary adjustment procedure for repeated measures (Dunlap, Cortina, Vaslow, & Burke, 1996; Johnson & Eagly, 2000).

⁴ This comparison could not be performed for Material B, in which there was no P + O − condition.