Yiu-Kei Tsang

ERP correlates of pre-attentive processing of Cantonese lexical tones: The effects of pitch contour and pitch height

The pre-attentive processing of Cantonese tones was studied with an auditory passive oddball paradigm. Event-related potentials to standard and deviant auditory stimuli were recorded as participants watched a silent movie attentively. The standards and deviants differed in either pitch level or pitch contour. Mismatch negativity (MMN) and P3a were elicited by all types of deviant tones, suggesting that lexical tone was processed pre-attentively. In addition, the size and latency of MMN were sensitive to the size of pitch level change, while the latency of P3a captured the presence of pitch contour change. These results indicate that pitch contour and pitch height are two important dimensions in sensory processing of lexical tones.

Do position-general radicals have a role to play in processing Chinese characters?

Over 90% of Chinese characters are compounds, comprising two or more constituents called radicals. Two experiments employed a character matching task to examine the contribution of radical position in Chinese character processing. The task was to decide whether the target character had appeared in two briefly and sequentially presented preceding source characters. Experiment 1 discovered significantly more false matching when the target (e.g., ‘’) shared radicals with the source stimuli (‘’, ‘’) than when the target and the source shared no radical, indicating that radicals contribute to character processing. Experiment 2 replicated this finding and further demonstrated that sharing a single radical between target and source characters, regardless of its radical position, was sufficient to generate false matching. More importantly, participants took significantly longer time to correctly reject those target characters with two shared radicals (one at the same position and another at the changed position) relative to those with only a single shared radical at the same position. Furthermore, false matching rates were significantly affected by lexical variables such as character frequency. These results suggest that position-general radicals play a significant role in character recognition and processing.

Morphemic ambiguity resolution in Chinese: Activation of the subordinate meaning with a prior dominant-biased context

In the present study, we examined how morphemic ambiguity is resolved using the visual-world paradigm. Participants were presented with Chinese bimorphemic words containing an ambiguous morpheme (analogous to the suffix -er in teacher and taller) and performed a visual search task. Their eye-movement patterns during target detection showed that (1) without a prior context, the dominant meaning of an ambiguous morpheme was more available than the subordinate one; (2) with a dominant-biased prior context, the subordinate meaning was still activated; and (3) a subordinate-biased prior context could inhibit the dominant interpretation. Therefore, both the frequency of the intended meaning and the prior contextual biases play a role in morphemic ambiguity resolution. The results are discussed with reference to models of ambiguity resolution and recent proposals of the graded nature of morphological effects.

ERPs reveal sub-lexical processing in Chinese character recognition

The present study used ERPs and a lexical decision task to explore the roles of position-general and position-specific radicals and their relative time courses in processing Chinese characters. Two types of radical frequency were manipulated: the number of characters containing a specific radical irrespective of position (i.e., radical frequency or RF) and the number of characters containing a specific radical at a particular position (i.e., position-specific radical frequency or PRF). The PRF effect was found to be associated with P150, P200, and N400, whereas the RF effect was associated with P200. These results suggest that both position-general and position-specific radicals could influence character processing, but the effect of position-specific radicals appeared earlier and lasted longer than that of position-general radicals. These findings are interpreted in terms of the specific orthographic properties of the sub-lexical components of Chinese characters.

Eye movement control in reading: Logographic Chinese versus alphabetic scripts.

Eye tracking is widely used to study the reading process in different languages. Given the unique properties of written Chinese, it is important to identify the similarities and differences in eye-movement controls between readers of Chinese and alphabetic scripts. In this article, we review the data available for comparison in four domains: (a) the perceptual span, (b) parafoveal processing, (c) the basic encoding unit, and (d) computer modeling. Our review indicates that, while some aspects of eye movement control in reading appear to be universal (e.g., the availability of orthographic information in the parafoveal preview), others are more script-specific (e.g., the size of the perceptual span). It is concluded that the study of cross-language differences is important for the development of reading models that are generalizable across readers of different scripts.

Processing Cantonese lexical tones: Evidence from oddball paradigms

Two event-related potential (ERP) experiments were conducted to investigate whether Cantonese lexical tones are processed with general auditory perception mechanisms and/or a special speech module. Two tonal features (f0 direction and f0 height deviation) were manipulated to reflect acoustic processing, and the contrast between syllables and hums was used to reveal the involvement of a speech module. Experiment 1 adopted a passive oddball paradigm to study a relatively early stage of tonal processing. Mismatch negativity (MMN) and novelty P3 (P3a) were modulated by the interaction between tonal feature and stimulus type. Similar interactions were found for N2 and P3 in Experiment 2, where more in-depth tonal processing was examined with an active oddball paradigm. Moreover, detecting tonal deviants of syllables elicited N1 and P2 that were not found in hum detection. Together, these findings suggest that the processing of lexical tone relies on both acoustic and linguistic processes from the early stage. Another noteworthy finding is the absence of brain lateralization in both experiments, which challenges the use of a lateralization pattern as evidence for processing lexical tones through a special speech module.

Activation of Morphemic Meanings in Processing Opaque Words

Two masked priming experiments were conducted to examine the activation of morphemic forms and meanings during opaque word processing. In Experiment 1, opaque primes significantly facilitated the recognition of transparent targets, which was consistent with previous results. However, transparent primes did not influence the recognition of opaque targets. This asymmetry could not be explained solely by morpho-orthographic processing, but it was consistent with models that have assumed early morpho-semantic activation. Experiment 2 directly tested whether the meanings of the constituent morphemes in opaque words were activated. In the critical condition, the targets were unrelated to the opaque primes at the lexical level, but were semantically related to their morphemes (e.g., “butterfly–bread”). Facilitation was observed in this condition, providing strong evidence of morpho-semantic activation during opaque-word recognition. These findings indicate that although initial morphological decomposition is determined by surface morphological form, it does not necessarily imply that morphemic meanings will be activated at later stages of processing. Rather, the morphemic meanings may be available automatically once segmentation is complete.

Morpho-orthographic and morpho-semantic processing in word recognition and production: evidence from ambiguous morphemes

Two sets of experiments were conducted to investigate the role of morphemes in word recognition and production. These experiments employed three priming procedures (i.e., masked, unmasked and long lag) to study the relatively early to late stages of morphological processing. Targets were Chinese compound words containing an ambiguous morpheme (analogous to “chair” in “chairman” vs. “armchair”). Primes and targets shared the same ambiguous morpheme with the same interpretation (S), a different interpretation (D) or were completely unrelated (U). For word recognition, the facilitation by the S and the D primes was statistically identical in the masked priming procedure. But only the S primes continued to facilitate word recognition in the unmasked and the long-lag priming procedures. In contrast, for word production, only the D primes produced significant facilitation in masked priming. In unmasked priming, both the S and D primes facilitated the naming reaction times, as compared with the unrelated baseline. But the facilitation was stronger in the S than in the D conditions. Finally, in the long-lag priming procedure, both the S and the D primes produced facilitation of equal strength. These results indicate that the processing of ambiguous morpheme involves both morphemic form and meaning, and that the temporal dynamics of the two effects differ in recognition and production.

Right hemisphere advantage in processing Cantonese level and contour tones: evidence from dichotic listening.

The brain lateralization pattern of Cantonese tonal processing was examined with the dichotic listening (DL) paradigm. Three factors were manipulated systematically in the study. First, the processing of level tones was compared with that of contour tones. Second, the influence of a linguistic context in tonal processing was studied by contrasting the patterns of brain lateralization for real syllables, pseudo-syllables, and hums. Finally, the discrimination and the identification tasks were used to test how processing depth might modulate the results obtained. A right hemisphere advantage (RHA) was obtained regardless of tone type, stimulus type, and task. In addition, the performance on level tones was in general better than that on contour tones. These findings suggest that Cantonese speakers are highly sensitive to the acoustic features of lexical tones, which supports the acoustic view about tonal processing.

Morpho-Semantic Processing in Word Recognition: Evidence from Balanced and Biased Ambiguous Morphemes.

The role of morphemic meaning in Chinese word recognition was examined with the masked and unmasked priming paradigms. Target words contained ambiguous morphemes biased toward the dominant or the subordinate meanings. Prime words either contained the same ambiguous morphemes in the subordinate interpretations or were unrelated to the targets. In addition, the relative frequency of the alternative meanings of ambiguous morphemes could be balanced (i.e., the alternative meanings are of similar frequency) or biased (i.e., one of the meanings is used much more frequently). The recognition of subordinate targets was facilitated by the subordinate primes for both balanced and biased items, regardless of the priming procedure. However, the subordinate primes did not facilitate the recognition of dominant targets, except for biased items in masked priming. These results are interpreted as supporting the hypothesis that morphemic meaning is activated to constrain morphological priming even at the early stage of processing. Yet, morpho-semantic activation is modulated by the frequency of the intended morphemic interpretations. Therefore, because of the high frequency of use, the dominant meanings of biased ambiguous morphemes can nevertheless be activated by the subordinate primes.