R Fung

An ERP Study of Good Production vis-à-vis Poor Perception of Tones in Cantonese: Implications for Top-Down Speech Processing

This study investigated a theoretically challenging dissociation between good production and poor perception of tones among neurologically unimpaired native speakers of Cantonese. The dissociation is referred to as the near-merger phenomenon in sociolinguistic studies of sound change. In a passive oddball paradigm, lexical and nonlexical syllables of the T1/T6 and T4/T6 contrasts were presented to elicit the mismatch negativity (MMN) and P3a from two groups of participants, those who could produce and distinguish all tones in the language (Control) and those who could produce all tones but specifically failed to distinguish between T4 and T6 in perception (Dissociation). The presence of MMN to T1/T6 and null response to T4/T6 of lexical syllables in the dissociation group confirmed the near-merger phenomenon. The observation that the control participants exhibited a statistically reliable MMN to lexical syllables of T1/T6, weaker responses to nonlexical syllables of T1/T6 and lexical syllables of T4/T6, and finally null response to nonlexical syllables of T4/T6, suggests the involvement of top-down processing in speech perception. Furthermore, the stronger P3a response of the control group, compared with the dissociation group in the same experimental conditions, may be taken to indicate higher cognitive capability in attention switching, auditory attention or memory in the control participants. This cognitive difference, together with our speculation that constant top-down predictions without complete bottom-up analysis of acoustic signals in speech recognition may reduce one’s sensitivity to small acoustic contrasts, account for the occurrence of dissociation in some individuals but not others.

A database for investigating the logographeme as a basic unit of writing Chinese

Chinese script is non-alphabetic and a Chinese graph is basically syllabic which may consist of phonetic and semantic radicals with no representation of phonemes. The logographeme, a unit smaller than a radical, has been suggested to be the basic unit of Chinese writing based on data collected on people with aphasia. To better understand the role of logographemes in Chinese writing development, a data corpus of logographemes based on characters appearing in primary school textbooks is established. Logographemes are analysed in terms of features that are believed to influence writing development. A total of 249 logographemes were identified: 151 logographemes with no meaning and sound (NMS), 84 logographemes with both sound and meaning which could also stand alone to serve as a character (SA) and 14 logographemes with meaning only (MO). At each grade, the frequencies of NMS logographemes were relatively lower than those of SA and MO logographemes, and the frequencies of SA and MO logographemes were similar; 94% of logographemes were present in the characters taught to grade one students. Students learnt all the pronounceable logographemes by grade three, while they finished all the logographemes without sound until grade six. Characters with left-right, top-bottom and enclosing configurations constituted about 94% of all single-unit characters acquired in primary school years. Statistics derived from the data corpus regarding these features across grades enable us to make specific predictions about stages of literacy development and suggestions for investigation into processes involved in character production.

Perception and production of Cantonese consonant endings

Abstract This paper reports findings from an experimental study of the perception and production of Cantonese consonant endings [m, n, η, p, t, k] by 40 male and female subjects who were university students at the time of the study. The subjects were presented with a series of tasks that had been specially designed for investigating their ability to perceive, produce and identify Cantonese, consonant endings. For the perception part of the study, subjects listened to recordings of pairs of Cantonese syllables and indicated whether the two syllables were the same or different. In the production part of the study, subjects listened to a series of individual syllables and repeated what they had heard. In the identification part, subjects matched the syllable they heard in a recording with its romanized form by choosing the appropriate letterstring from among three presented to the subject on the computer screen. Among the major findings of this study are the following: (1) A positive relationship was found between the perception and production of endings [t]∼[k] and [n]∼[η];that is, if they could perceive the difference in the endings they were also likely to be able to produce it. (2) Subjects made significantly more errors in perceiving, producing, and identifying stop endings [t, k] than nasal endings [n, η]. (3) Subjects made the most errors in perceiving and repeating syllables with nuclear vowel [œ]. (4) Subjects erroneously interchanged stop endings [t, k] ([t] with [k] and [k] with [t]) at about the same rate. (5) In contrast, subjects tended to change nasal ending [η]to [n]. (6) Male and female subjects showed no difference in perception, production, and identification of consonant endings.

Relationship between individual differences in speech processing and cognitive functions

A growing body of research has suggested that cognitive abilities may play a role in individual differences in speech processing. The present study took advantage of a widespread linguistic phenomenon of sound change to systematically assess the relationships between speech processing and various components of attention and working memory in the auditory and visual modalities among typically developed Cantonese-speaking individuals. The individual variations in speech processing are captured in an ongoing sound change–tone merging in Hong Kong Cantonese, in which typically developed native speakers are reported to lose the distinctions between some tonal contrasts in perception and/or production. Three groups of participants were recruited, with a first group of good perception and production, a second group of good perception but poor production, and a third group of good production but poor perception. Our findings revealed that modality-independent abilities of attentional switching/control and working memory might contribute to individual differences in patterns of speech perception and production as well as discrimination latencies among typically developed speakers. The findings not only have the potential to generalize to speech processing in other languages, but also broaden our understanding of the omnipresent phenomenon of language change in all languages.

Automatic word segmentation for spoken Cantonese

Though Cantonese is the most influential variety of Chinese other than Mandarin, there are only a limited number of Cantonese corpora available for linguistic studies. Among the essential steps of building a corpus, word segmentation is a necessary but highly challenging task due to the lack of clear word boundary in Cantonese. This paper reports the construction and evaluation of an open-source automatic Cantonese word segmenter developed for Cantonese. The tool is a component of the multilingual SPPAS program designed to be used directly by linguists. It is a free software distributed under a GPL license. The effectiveness of the tool was evaluated by comparing the result of segmenting some samples of a spoken Cantonese corpus manually and automatically using the tool developed. High precision and recall were found in our study. Upon completion, the tool would definitely promote the development of more Cantonese corpora for language related studies.

Feedback Consistency Effect on Writing-to-Dictation Task in Chinese

Abstract The influence of feed-forward consistency (FFC) (i.e., the consistency of mapping from the orthographic form to the pronunciation of the whole character) on character recognition and reading aloud tasks have been well documented in a sizable literature in which subjects performed better on feed-forward consistent characters than inconsistent characters. Analogous to the FFC effect in reading aloud, the feedback consistency (FBC) effect (“consistent” when a pronunciation always maps onto one orthographic form, and “inconsistent” if the pronunciation maps onto multiple orthographic forms) was observed in spelling tasks at the phoneme-grapheme levels in English (e.g., Weekes, Castles, & Davies, 2006) and French (e.g., Alegria & Mousty, 1996), where subjects had better spelling performances on feedback consistent words than inconsistent words. Although Ziegler and Muneaux (2007) showed that the size of FBC effect was predicted by the reading level (determined by a feed-forward task) of children in a spoken word recognition task, the words that are phonologically similar, in fact, are also orthographically similar in alphabetic scripts. Therefore, the poor performance on feedback inconsistent words may be attributed to the difficulty in choosing the correct answers among similar orthographic forms. Unlike alphabetic scripts, homophones in Chinese characters do not always share similar orthographic forms and orthographically similar characters may have different pronunciations. These properties allow researchers to investigate the FBC effect without confounding with orthographic similarity. The aims of the present study are three-fold: first, to find out whether similar FFC effect exists in both reading aloud and writing-to-dictation in Chinese; second, to evaluate the influence of the number of homophones of a character on writing-to-dictation across grades; and, third, to observe the possible interactions between FFC and the number of homophones. Phonetic semantic compound characters were selected as stimuli. A reading- aloud task was administered to 1,590 students and a writing-to-dictation task to 2,194 students from grades 1 to 6. The stimulus characters were categorized in terms of homophone numbers and consistency values. Data collected were analyzed using ANOVA. As expected, significant positive FFC effect (consistent characters better than inconsistent characters) was observed in reading aloud across grades. In the writing-to-dictation task, significant homophone effect was found from grade 2 to grade 6. Post hoc analyses revealed that a positive FFC effect occurred when the characters had many homophones, whereas negative FFC effect occurred when the characters had fewer homophones. The findings clarify the influences of feed-forward and feedback consistencies on writing without the possible confounding orthographic similarity factor. The significant interaction effect between FFC and homophone number suggests that the orthographic similarity of the family members in a homophone family might affect the FBC effect in different tasks.