J Psycholinguist Res DOI 10.1007/s10936-014-9308-4

Meaning Inhibition and Sentence Processing in Chinese Michael C. W. Yip

© Springer Science+Business Media New York 2014

Abstract The present study examined the inhibitory processes of spoken word recognition of Chinese homophones during sentence processing, using a standard cross-modal naming experiment with an innovative design and materials construction. Results confirmed that (1) preceding sentence context has exerted an early effect on disambiguating among different alternative meanings of the homophones; (2) the contextually inappropriate meanings of the ambiguous word were inhibited rapidly during sentence processing; and (3) the present results also demonstrated that the inhibitory mechanism could be sustained to a longer duration following the occurrence of the ambiguous word (homophone). Finally, all these results clearly revealed the dynamics of interaction of context effects and spoken word recognition processes. Keywords Context effects · Spoken homophones · Inhibitory processes · Lexical access · Chinese sentence processing

Introduction Investigation of the interaction between sentence context and spoken word recognition has been a long-studied psycholinguistic topic for the research of lexical ambiguity resolution (Grosjean 1980; Li and Yip 1996, 1998; Marslen-Wilson 1987, 1990; Simpson 1984; Simpson and Kang 1994; Simpson and Krueger 1991; Swinney 1979; Tabossi 1988a,b; Zhang et al. 2006). In normal case, when we listen to a spoken sentence, the accurate meaning of the sentence comes out to our mind naturally although the sentence may have several ambigu-

This research was partially supported by funding from the Research Grants Council of the Hong Kong University Grants Committee [GRF No.: 845613]. I would like to thank Katherine Leung and Carol Chan for their assistance in materials preparation to the present study and the constructive comments from the anonymous reviewers. M. C. W. Yip (B) Department of Psychological Studies, The Hong Kong Institute of Education, Tai Po, N.T., Hong Kong e-mail: [email protected]

123

J Psycholinguist Res

ous words, such as homophones. How can we get the correct meaning and resolve these ambiguities so quickly? This question is exactly one of the fundamental problems of spoken language processing (Gorfein 1989)—lexical ambiguity resolution: When listeners hear an ambiguous word that has multiple meanings, do they use preceding sentence context to help them eliminate (or inhibit) contextually irrelevant meanings, or do they activate all possible meanings in their mental lexicon irrespective of the preceding sentence context? The final outcome of processing the ambiguity is undoubtedly a single contextually appropriate meaning with the aid from the preceding sentence context. The most interesting point researchers would like to figure out in this line of research is the time-course of the context effects operated along with the different stages of lexical access. It is because the answer can reflect the underlying mechanism to our language processor in particular, and the whole cognitive system in general. However, after more than three decades of psycholinguistic research in different languages, there is still no conclusive answer today. So far, there are two competing hypotheses reported in the research literature (Grosjean 1980; Marslen-Wilson 1987, 1990; Simpson 1984; Simpson and Kang 1994; Simpson and Krueger 1991; Swinney 1979; Tabossi 1988a,b). The first one is the exhaustive access hypothesis, which argues that all meanings of an ambiguous word will be accessed momentarily following the occurrence of the ambiguous word, and the sentential context can only help to select the appropriate meaning at a late post-access stage. This hypothesis assumes that language processing is a modular, bottom-up process in which non-lexical, contextual information does not penetrate lexical access (Fodor 1983). The second one is the contextdependency hypothesis, which argues that only the contextually appropriate meaning of the ambiguous word will be accessed early on if the preceding sentence context provides a strong bias to the appropriate meaning due to its increased activation. This hypothesis assumes that language processing is operated by an interactive approach in which information can flow both bottom-up and top-down simultaneously and that lexical access and sentential context can mutually influence each other at a very early stage (McClelland 1987). Many experimental psycholinguistics studies testing on the above hypotheses have been widely researched in different languages (Chinese, Dutch, English, and Italian) during the last three decades with different patterns of results (see Zhang et al. 2006 for a review). In the classical study of Onifer and Swinney (1981), the researchers observed that lexical decision times to both the primary and secondary meanings of the ambiguous word were equally facilitated following the immediate occurrence of the ambiguity during sentence processing. This effect even retained under the biasing context conditions. In his earlier work, Swinney (1979) reported that the sentence context could only exert its influence to select the contextually appropriate meaning of the ambiguous word after four-syllable duration following the ambiguity. Thus, these findings in general support the exhaustive access hypothesis. However, on the contrary, in the study of Simpson (1981), the researcher observed an opposite pattern of results to the issue. Simpson used two lexical decision tasks (by using cross-modal experiment) to investigate the effects of dominance and different types on lexical access. The author found that only the primary (dominant) meaning of the ambiguous word was facilitated relative to the unrelated control word following the immediate occurrence of the sentence-ended ambiguity in the biased sentence condition. In addition to Simpson’s work, Glucksberg et al. (1986) provided further evidence favoring the context-dependency hypothesis. From the results of cross-modal experiments to investigate the time-course of sentence context, they confirmed that context can constrain the initial activation among different meanings of the ambiguous word. Tabossi (1988a,b) obtained similar evidence to consolidate the early effects of different context on selecting the appropriate meaning of the ambiguous word, provided that the ambiguous word occurs in a highly semantically-constrained sen-

123

J Psycholinguist Res

tence context. Taken altogether, empirical findings observed from these researchers strongly support the context-dependency hypothesis. Other than the experimental studies from Indo-European languages, the relevant studies in Chinese language (including both Cantonese Chinese and Mandarin) also yielded divergent results. From their pioneering work, Li and Yip (1996, 1998) have first explored the processing of Chinese homophones using cross-modal and gating paradigms to examine the effects of sentence context on native Cantonese speakers’ access and selection of homophone meanings. The cross-modal experiment demonstrated that context effects took place immediately following the occurrence of the spoken homophone, and the gating experiment also revealed that native Cantonese listeners could recognize the appropriate meaning with less than half of the acoustic information of the homophone. These experiments indicate that native Cantonese listeners are sensitive to the contextually biased meaning at an early stage, probably within the acoustic boundary of the word (and after the tonal differentiation). Their results point to a much earlier context effect than has been previously assumed (e.g., about 1.5 s following the ambiguous word reported in Onifer and Swinney 1981). In addition, in their follow-up studies (Li et al. 2002; Yip 2000, 2002, 2007, 2008a,b, 2009), these researchers further obtained similar patterns of findings on an early influence (both facilitation and inhibition) of the sentence context to the disambiguating process of the ambiguous items by using different approaches (e.g., cross-modal priming; gating; visual-visual priming task). Clearly, these results in general support the context-dependency hypothesis. However, another scholar (Ahrens 2002, 2006) found the opposite evidence to support the exhaustive access of meaning in Chinese spoken word recognition. In her cross-modal experiment (using Mandarin materials), Ahrens observed that there would be a moment of multiple access of different meanings to the ambiguous word even with a constrained sentence context. Apart from the testing language (Cantonese vs. Mandarin), both the research design and tasks demands of Ahrens’s work were different from Li and Yip’s original studies. But, it can still reflect the inconclusiveness of this long-studied psycholinguistic problem. As aforementioned, most of the previous studies in this line of research focused on the facilitative processes of the ambiguous words during sentence processing. Only a few studies were systematically examined the inhibitory processes among different alternative meanings of the ambiguous word during sentence processing so far (Gernsbacher and Faust 1991; Morris and Binder 2001; Simpson and Kang 1994; Wu and Shu 2001). Therefore, the present study attempts to examine this long-studied psycholinguistic question from a different dimension: to trace the inhibitory processes among the different individual meanings of an ambiguous word during sentence processing. In the literature on meanings suppression of ambiguous words, one important study concluded that the contextually inappropriate meanings of the ambiguous word would be suppressed actively and rapidly, within 200 ms, following the occurrence of the ambiguous word with a view to maintain normal sentence comprehension (Gernsbacher and Faust 1991). Simpson and Kang (1994) conducted another experimental study and they demonstrated similar patterns of results. They found that the inhibitory processes of those unselected or those contextually inappropriate meanings would still be sustained for a period of time (might be longer than one second) after the human language processor selected the contextually appropriate meaning. The present study attempts to investigate this rarely addressed question by using Cantonese Chinese as a rigorous testing case. Cantonese Chinese is a language that differs significantly from most other languages (e.g., in its use of lexical tones, its morphemic monosyllabicity) and the language clearly offers many unique and interesting psycholinguistic properties in its phonological, lexical, and syntactic structures (Zhang et al. 2006) to crucially investigate

123

J Psycholinguist Res

this issue. As a matter of fact, lexical ambiguity is pervasive in Chinese (including Cantonese and Mandarin) due to its extensive homophony occurred at the lexical and morphemic level. From the Modern Chinese Dictionary (Institute of Linguistics 1985), about 80 % of the monosyllables (with tonal differentiation) in Chinese are ambiguous between two different meanings, and about 55 % have five or more alternative meanings. For example, a Cantonese monosyllable si1 has up to 14 meanings (e.g., teacher, lion, silk, corpse, poem, private, think, etc.; and 8 of them are common meanings to most native Cantonese listeners) (Zhang and Zhang 1987). Upon hearing the monosyllable si1 in a highly semantically constraint sentence, do native Cantonese listeners activate only the most contextually appropriate meaning out of the total 14 meanings and inhibit the other remaining 13 meanings of the single syllable simultaneously (cf. Neely 1991) or exhaustively activate all the 14 (or at least 8) alternative meanings at that very moment? How the inhibitory processes operate? Will there be other lexical variables affected the inhibitory processes simultaneously? How long the inhibition can be sustained? The present study is designed to answer these aforementioned research questions, by using a cross-modal naming experiment, to seek new evidence on the timecourse picture to the dynamic relationship between the sentence context and other lexical factors operated in the processing of Chinese spoken language.

Experiment Method Participants. Seventy-two native Cantonese listeners (33 male and 39 female, mean age = 20.6) who reported no speech or hearing deficits participated in this experiment. All participants were undergraduate students at the Open University of Hong Kong and the Hong Kong Institute of Education. They were paid HK$30 as a token of gratitude to take part in the experiment. Materials and design. Thirty spoken Cantonese homophones, all nouns, were selected, each with at least two different meanings in the same tone (see Appendix). Each homophone was embedded in four different types of sentences, in a two (Dominance: dominant vs. subordinate) by two (Reiteration: reiterated vs. not reiterated) mixed design. The four different types of sentences are:

(DSR) With the first part of preceding sentence context biased to the dominant meaning of the homophone, and then the homophone will be reiterated again at the end of the sentence in which second part of the sentence context will be accordingly biased to the subordinate meaning of the homophone. (SDR) With the first part of preceding sentence context biased to the subordinate meaning of the homophone and then the homophone will be reiterated again at the end of the sentence in which second part of the sentence context will be accordingly biased to the dominant meaning of the homophone. (DNR) With the first part of sentence context keeping neutral and then the preceding sentence context of the second part biased to the dominant meaning of the homophone at the end of the sentence and without reiterating the homophone at the middle position of the sentence. (SNR) With the first part of sentence context keeping neutral and then the preceding sentence context of the second part biased to the subordinate meaning of the homophone at

123

J Psycholinguist Res

the end of the sentence and without reiterating the homophone at the middle position of the sentence. Therefore, all the sentences have two parts with a 500 ms pause in between, immediately after the first occurrence of the homophone (for the cases of DSR and SDR). Following the norming procedure of relevant studies (Li and Yip 1996, 1998; Yip 2008a,b), a separate group of 20 native Cantonese Chinese speakers was asked to judge the degree of semantic constraint of the sentence context on the homophone. They were given all the test sentences with the preceding biasing context but without the homophone in written form, and were asked to fill in the appropriate blank word in written form. They were told to think of a Chinese word that would naturally complete the sentence. Their responses were scored on a 1–4 scale, based on the scale proposed by Marslen-Wilson and Welsh (1978): 1 was given for a word identical to the test word, 2 for a synonym, 3 for a related word, and 4 for an unrelated word. Responses were pooled across the 20 judges, and the mean rating was 1.9. This score was above the high constraint condition in Marslen-Wilson and Welsh’s study (1978). An effort was also made to have the sentence context of equal length, and the average length of all the test sentences, including the homophones, was 26 words (ranging from 24 to 27 words). In addition, we have made an effort to eliminate any kind of intra-sentential priming from other individual words within the whole sentence frame when constructing the sentence context as much as possible. Apart from the sentence context, we carefully selected the appropriate visual probes as the followings. First, all the visual probes were based on a semantic relatedness norm experiment from another separate group of one hundred and twenty native Cantonese Chinese speakers. In this meaning-association experiment, participants were given a form that included all the 60 homophones in written characters (30 × 2). The participants were then asked to immediately think of three Chinese words that either have the same or closely related meaning to each homophone. The most frequent words they listed were chosen as the related visual probes for each homophone. The set of unrelated visual probes used in the experiment was also carefully selected and matched with the related visual probes in this experiment. All the visual probes in each experimental condition were again carefully matched with the same category of initial phonemes, individual frequency information and the orthographic complexity in order to avoid any possible confounding. Altogether, there were three main variables manipulated in this experiment, namely: (1) Dominance: The sentence context was either biased to the dominant meaning of the spoken homophone or to the subordinate meaning of the spoken homophone. The dominance information was empirically based on the frequency database of Ho and Jiang’s work (1994). (2) Reiteration: The homophones were either reiterated in the middle of the sentence context or not. (3) Relatedness: The visual probes were either related to the dominant meaning of the spoken homophone or to the subordinate meaning, or unrelated. An example si1 (silk/poem) in the four different types of sentence context is given below (* denotes the 500 ms pause). (DSR) Sai3 lo5 lam5 ye5 lam5 do3 tau4 faat3 do1 hei2 saai3 si1 * tai2 lai4 kui5 gam1 yat6 do1 gei2 laan4 jok3 do2 yat1 sau2 si1. Young brother is thinking very hard that his hair seems like gradually becoming silk*, I do not think that today he can finally compose a poem.

123

J Psycholinguist Res

(SDR) Dong1 ga1 je1 hai5 sue1 fong2 lui5 min6 yam4 gan2 sau2 si1 * ma1 me4 jau6 hui3 joh2 sui6 fong2 do6 tong3 gan2 san1 maai5 di1 si1. When elder sister is, inside her studying room, reading aloud a poem*, mother at the same time, in her bedroom, is ironing her newly-bought silk. (DNR) Ngoh5 tung4 taai3 taai2 yap6 joh2 gaan1 uk1 do6 faat3 yin6 * cheung4 gok3 seung6 min6 jung6 yau5 ho2 doh1 ji1 jue1 hai5 do6 to3 gan2 si1. My wife and I entered the house and found that *, and there are so many spiders at the corner that are spinning the silk. (SNR) Wai6 joh2 siu1 moh4 gam1 nin4 sue2 ga3 goh3D si4 gaan3 * ngoh5 lei1 paai4 booi3 joh2 ho2 doh1 sau2 ho2 chut1 ming5 ge3 si1. In order to kill the long time during the summer *, I recently recite a lot of famous poem. The corresponding visual probes to si1 in these sentences are: saam1 “shirt” (SDR and DNR), man4 “prose” (DSR and SNR), and booi1 “cup” and lei4 ‘pear” (unrelated control). Predictions and Rationale Based on the design of these materials, we set out two predictions: [P1] the response latencies to related visual probes of condition (DNR) will be faster than that of condition (SNR) and the response latencies to related visual probes of condition (SDR) will also be faster than that of condition (DSR) due to a simple dominance effect (Hogaboam and Perfetti 1975). According to the Reordered Access Hypothesis (Hogaboam and Perfetti 1975), the dominant meaning of a spoken homophone will always come out first under a neutral sentence context condition during sentence processing. [P2] the response latencies to related visual probes of condition (DNR) will be faster than that of condition (SDR) and the response latencies to related visual probes of condition (SNR) will be faster than that of condition (DSR); while the response latencies to related visual probes of conditions (DSR and SDR) will be slower than the unrelated control due to the inhibitory effects of the inappropriate meanings of the ambiguous words at the first occurrence. [P2] is particularly important to testify the inhibitory effects. The rationale is that by comparing the response latencies to related visual probes of conditions (SDR and DNR) as well as the reiterated conditions of (DSR and SDR) and unrelated control, it is predicted that the activation level of the inappropriate meaning of the homophone (only the contextually appropriate meaning is accessed when hearing the homophone at the first time in the middle of the sentence) will gradually decrease and sustain to the end of the entire sentence frame in which the homophone is reiterated again at the end of the sentence; and hence the inhibition of the inappropriate meaning will arrive at a level that makes a slower response latencies than the time observed in DNR condition, in which no inhibition to the inappropriate meaning of the homophone has occurred. In sum, we can conceive that the inhibitory role of context acts not only as an inhibition effect but also a result of decay in activation. As suggested by Twilley and Dixon (2000), there were two kinds of inhibition: context-relative and meaning-relative. The former one assumes the inappropriate sentence context will suppress the activation of one specific meaning to a level below that occurred in a neutral condition; while the latter one assumes the suppression of one specific meaning to a level below its normal resting state.

123

J Psycholinguist Res

In other words, preceding sentence context facilitates one meaning while inhibits another meaning of the ambiguous word at the same time; thus making the inhibited one difficult for retrieval in the second part of the sentence context, when the ambiguous word is reiterated again. Therefore, if this prediction can be proved, it certainly confirmed the inhibitory effects which take place during spoken sentence processing and that such inhibitory effects can sustain to a certain period of time. Each participant randomly received an equal number of sentences (fifteen) for each condition in a 2 (Dominance) × 2 (Reiterated) × 2 (Relatedness) mixed factorial design. In order to reduce any strategic effect that may occur (McQueen and Sereno 2005), half of the thirty sentences were paired with related visual probes and the other half was paired with unrelated visual probes. The order of presentation for the sentences was pseudo-randomly arranged such that the visual probes did not consecutively bias spoken homophones. The order of presentation was counterbalanced across all participants. Experimental Apparatus. All the test sentences were read by a female native Cantonese speaker at a normal conversational speed and digitally-recorded on a SONY MD. The sentences were then transferred and digitized into a Macintosh PowerBook. A computer program called PsyScope (Cohen et al. 1993) controlled the presentation of the materials. A microphone, which was used to register listeners’ vocal responses and hence calculated the naming latencies, was connected to the notebook computer. A remote controlled SONY IC-recorder was also used and controlled by the experimenter in another partition of the experimental room to check for the accuracy of the naming response. Procedure. All participants did the experiment individually in a quiet experimental room. Before the experiment, the experimenter explained the task in Cantonese to the participants. First, they were told that they would be hearing a sentence in two parts with a brief pause through a pair of headphones, and then at the end of the whole sentence, a Chinese character would be displayed at the centre on the computer screen [the visual probe will occur at the acoustic offset of the spoken sentence: ISI = 0]. Their task was to listen carefully to the spoken sentence and then to name the Chinese character aloud into the microphone as quickly and as accurately as possible. Before the actual experiment began, they were given a practice session in which they heard a set of separate but similar sentences. The whole experiment took about twenty minutes.

Results Mean response latencies, counting from the onset of the visual probe to the vocal response, as a function of dominance, reiteration and relatedness are presented in Table 1. Errors (i.e. listeners named the visual probes with a word that was totally different from the target word) were very rare (approximately 0.04 across all conditions), and therefore error data were not analyzed further in the present study.

Table 1 Mean response latencies (SD) as a function of dominance, reiteration and relatedness

Dominant Related Reiterated

Subordinate Unrelated

Related

Unrelated

839.4 (72.6) 826.8 (66.3) 826.2 (62.5) 827.7 (82.3)

Not reiterated 739.6 (50.5) 843.1 (64.7) 765.9 (64.1) 851.2 (73.5)

123

J Psycholinguist Res

A 2 (Dominance) × 2 (Reiteration) × 2 (Relatedness) repeated-measure ANOVA was conducted on the response latencies to the visual probes. Results revealed a clear main effect on each of the three variables: dominance, reiteration and relatedness. The interaction between reiteration and relatedness was also observed. First of all, the present results revealed that frequency of individual meanings of the homophone clearly affected the response time, F (1,71) = 4.68, p < .05. Collapsed over levels of the other variables, the mean response time to access the dominant homophone meaning was 789.5 ms and the time to access the subordinate homophone meaning was 796.1 ms, and access time to the unrelated visual probe was on average 837.2 ms. These results pointed out that the dominant meaning of a given homophone would be activated first in a highly constrained sentence context. For a direct comparison, the mean response time to related visual probes of conditions (DNR) and (SNR) were 739.6 and 765.9 ms respectively, and a 26.3 ms difference was observed; while the mean response time to related visual probes of conditions (SDR) and (DSR) were 826.2 and 839.4 ms respectively, and a 13.2 ms difference was observed. Hence, these results proved the first prediction [P1]. Secondly, there was a main effect on the variable of Reiteration, F (1,71) = 26.18, p < .05. Collapsed across the other variables, the mean response time on reiterated condition to related visual probe was 832.8 ms and the time on the non-reiterated condition to related visual probe was 752.8 ms with a difference of 80.0 ms. The reiteration effect clearly demonstrated a strong inhibitory effect on the inappropriate meaning following the first occurrence of the ambiguous word, in which contextually appropriate meaning had already been selected due to the preceding constraint context. For a direct comparison, the mean response time to related visual probes of condition (DNR) was 86.6 ms faster than that of condition (SDR)(739.6 vs. 826.2 ms); and the mean response time to related visual probes of condition (SNR) was 73.5 ms faster than that of condition (DSR)(765.9 vs. 839.4 ms). Similarly, the mean response time to related visual probes of conditions (DSR and SDR) was 832.8 ms which were slightly slower than the mean response time of the unrelated control condition, 827.3 ms. Hence, these results supported the second prediction [P2]. Furthermore, these inhibitory effects could even sustain for a long time till the end of the sentence, i.e. more than 2 s, a bit longer than the finding of Simpson and Kang (1994). The main effect of relatedness, F (1,71) = 28.27, p < .05 and the interaction between reiteration and relatedness, F (2,142) = 8.13, p < .05 obviously further strengthen the existence of an inhibition effect on the contextually inappropriate meaning as well as the facilitative effects to the contextually appropriate meaning during spoken sentence processing. Finally, the null interaction effect of dominance and reiteration suggested that there were other lexical variables operating simultaneously during sentence processing which the present study did not control for. One of the possible confounding is the variable of homophone density (Li and Yip 1998; Li et al. 2002; Yip 2008a,b). This variable might more or less affect the competition among different alternative meanings of an ambiguous word. The higher the homophone density, the greater the semantic competition among the competitors and hence reducing the inhibitory effect (Locker et al. 2003; Yates et al. 2003). Other potential confounding variable is the relative frequencies of the dominant and subordinate meanings of an ambiguous word (Zhang et al. 2001).

Discussion The present study examined the relationship between context effects and spoken word recognition during Chinese sentence processing from a different dimension, i.e. to trace the

123

J Psycholinguist Res

inhibitory processes of the ambiguous word (such as homophone) during sentence processing (as compared with the facilitative processes). We used spoken Cantonese Chinese homophone in the present experiment as a rigorous test case because of the pervasive homophony phenomenon in this particular language. In this study, a well-controlled cross-modal naming experiment was used to assess the processing time native Cantonese listeners needed to name a visual target Chinese character either related or unrelated to the meanings of the homophone during sentence processing. Clearly, the main effect of relatedness as well as the interaction effect between context and relatedness shows that preceding sentence context aids the processing of Chinese homophones at an early stage. These findings were actually consistent with other relevant studies in Chinese sentence processing (Chen and Cui 1997; Li and Yip 1996, 1998; Yip 2000, 2002, 2008a,b). Moreover, the results of the present study again indicated a strong dominance effect during the processing of homophones. The dominance effect was strong enough to aid early selection among different alternative meanings of the homophone, as reflected by the faster naming time in the DNR and SDR conditions than in the SNR and DSR conditions respectively. These results are somehow compatible with the reordered access hypothesis (Hogaboam and Perfetti 1975), which emphasized the importance of context and frequency effects in lexical ambiguity processing. Nonetheless, these results are obviously against the assumption of modular approach of language processing, i.e. frequency effect can only occur at a late selection stage (Onifer and Swinney 1981; Swinney 1979). In addition, the important finding on reiteration effect confirmed that the inhibitory processes occurred rapidly following the occurrence of the ambiguous word (homophone). Furthermore, the inhibitory processes can sustain to a relatively longer duration of more than 2 s (about twelve syllables’ duration) after the first occurrence of the ambiguous word (homophone) in the sentence, which was in line with other relevant studies in English language (Gernsbacher and Faust 1991; Simpson and Kang 1994). Therefore, overall speaking, the present results did not support the notion that native Chinese listeners would exhaustively access all possible meanings of a homophone without using contextual information initially to constrain lexical access. On the contrary, when they faced with the extensive homophony situation in this language, they will automatically adopt an interactive processing system that can rapidly disambiguate (through facilitative and inhibitory processes) alternative homophone meanings during Chinese sentence processing (McClelland 1987; McClelland and Elman 1986). Ongoing experiments are being designed in our laboratory to further examine how the other variables, such as homophone density effect, relative frequencies information, and Inter-stimuli-interval (ISI), influence the inhibitory processes of the ambiguous word during sentence processing, with different research designs and adopting different methodologies (such as eye-tracking and ERP measures).

Appendix See Appendix Table 2.

123

J Psycholinguist Res Table 2 Spoken Chinese homophones used in the experiment

The number of each syllable represents the lexical tone

baan1 bo3

baan2 boh1

cheung1

cheung4

chi4

do2

foh3

gau2

jeung3

jeung6

kei4

kwan4

ma5

min6

mo6

ping4

saam1

sau2

seung1

si1

sin3

sui3

tong4

wa6

wan4

woh1

woo4

yuen4

References Ahrens, K. (2002). Timing issues in lexical ambiguity resolution. In M. Nakayama (Ed.), Sentence processing in East Asian languages (pp. 1–29). Stanford, CA: CSLI Publications. Ahrens, K. (2006). The effect of visual target presentation times on lexical ambiguity resolution. Language and Linguistics, 7, 677–696. Chen, Y.-M., & Cui, Y. (1997). The processing of Chinese ambiguous sentences. Acta Psychologica Sinica, 29, 1–7. Cohen, J. D., MacWhinney, B., Flatt, M., & Provost, J. (1993). PsyScope: A new graphic interactive environment for designing psychology experiments. Behavior Research Methods, Instruments, and Computers, 25, 257–271. Fodor, J. A. (1983). The modularity of mind: An essay on faculty psychology. Cambridge, MA: MIT Press. Gernsbacher, M. A., & Faust, M. (1991). The role of suppression in sentence comprehension. In G. B. Simpson (Ed.), Understanding word and sentence (pp. 97–128). Amsterdam: Elsevier Science Publishers. Glucksberg, S., Kreuz, R. J., & Rho, S. H. (1986). Context can constrain lexical access: Implications for models of language comprehension. Journal of Experimental Psychology: Learning, Memory, and Cognition, 12, 323–335. Gorfein, D. S. (Ed.). (1989). Resolving semantic ambiguity. New York: Springer. Grosjean, F. (1980). Spoken word recognition processes and the gating paradigm. Perception and Psychophysics, 28, 267–283. Ho, H.-H., & Jiang, Y.-H. (1994). Word frequency in Hong Kong in the 90’s. Research Institute of Humanities, Chinese University of Hong Kong. Hogaboam, T., & Perfetti, C. (1975). Lexical ambiguity and sentence comprehension. Journal of Verbal Learning & Verbal Behavior, 14, 265–274. Institute of Linguistics, The Academy of Social Sciences. (1985). Xianda Hanyu Cidian (Modern Chinese Dictionary). Beijing: Commercial Press. Li, P., & Yip, M. C. W. (1996). Lexical ambiguity and context effects in spoken word recognition: Evidence from Chinese. In G. Cottrell (ed.) Proceedings of the 18th annual meeting of the cognitive science society, pp. 228–232. Li, P., & Yip, M. C. W. (1998). Context effects and the processing of spoken homophones. Reading and Writing, 10, 223–243. Li, P., Shu, H., Yip, M. C. W., Zhang, Y., & Tang, Y. (2002). Lexical ambiguity in sentence processing: Evidence from Chinese. In M. Nakayama (Ed.), Sentence processing in East Asian languages (pp. 111– 129). Stanford, CA: CSLI Publications. Locker, J. L., Simpson, G. B., & Yates, M. (2003). Semantic neighborhood effects on the recognition of ambiguous words. Memory & Cognition, 31, 505–515. Marslen-Wilson, W. D. (1987). Functional parallelism in spoken word recognition. Cognition, 25, 71–102.

123

J Psycholinguist Res Marslen-Wilson, W. D. (1990). Activation, competition, and frequency in lexical access. In G. T. Altmann (Ed.), Cognitive models of speech processing (pp. 148–172). Cambridge: The MIT Press. Marslen-Wilson, W. D., & Welsh, A. (1978). Processing interactions and lexical access during word recognition in continuous speech. Cognitive Psychology, 10, 29–63. McClelland, J. L. (1987). The case for interactionism in language processing. In M. Coltheart (Ed.), Attention and performance XII: The psychology of reading (pp. 3–36). Hillsdale, NJ: Erlbaum. McClelland, J. L., & Elman, J. (1986). Interactive processes in speech perception: The TRACE model. In J. L. McClelland, D. E. Rumelhart, and the PDP Research Group (Eds.), Parallel distributed processing: Explorations in the microstructure of cognition (vol. II) (pp. 58–121). Cambridge: The MIT Press. McQueen, J. M., & Sereno, J. (2005). Cleaving automatic from strategic effects in phonological priming. Memory & Cognition, 33, 1185–1209. Morris, R. K., & Binder, K. S. (2001). What happens to the unselected meaning of an ambiguous word in skilled reading? In D. Gorfein (Ed.), On the consequences of meaning selection: Perspectives on resolving lexical ambiguity (pp. 139–153). Washington, DC: American Psychological Association. Neely, J. H. (1991). Semantic priming effects in visual word recognition: A selective review of current findings and theories. In D. Besner & G. Humphreys (Eds.), Basic processes in reading: Visual word recognition (pp. 264–336). Hillsdale, NJ: Erlbaum. Onifer, W., & Swinney, D. A. (1981). Accessing lexical ambiguities during sentence comprehension: Effects of frequency of meaning and contextual bias. Memory & Cognition, 9, 225–236. Simpson, G. B. (1981). Meaning dominance and semantic context in the processing of lexical ambiguity. Journal of Verbal Learning and Verbal Behavior, 20, 120–136. Simpson, G. B. (1984). Lexical ambiguity and its role in models of word recognition. Psychological Bulletin, 96, 316–340. Simpson, G. B., & Kang, H. (1994). Inhibitory processes in the recognition of homograph meanings. In D. Dagenbach & T. H. Carr (Eds.), Inhibitory processes in attention, memory, and language (pp. 359–381). London: Academic Press. Simpson, G. B., & Krueger, M. A. (1991). Selective access of homograph meanings in sentence context. Journal of Memory and Language, 30, 627–643. Swinney, D. A. (1979). Lexical access during sentence comprehension: (Re)Consideration of context effects. Journal of Verbal Learning and Verbal Behavior, 18, 645–659. Tabossi, P. (1988a). Accessing lexical ambiguity in different types of sentential contexts. Journal of Memory and Language, 27, 324–340. Tabossi, P. (1988b). Effects of context on the immediate interpretation of unambiguous words. Journal of Experimental Psychology: Learning, Memory, and Cognition, 14, 153–162. Twilley, L., & Dixon, P. (2000). Meaning resolution processes for words: A parallel independent model. Psychonomic Bulletin & Review, 7, 49–82. Wu, N., & Shu, H. (2001). Meaning activation of Chinese syntactic category ambiguous words in Isolation. Acta Psychologica Sinica (in Chinese), 33, 305–311. Yates, M., Locker, L., & Simpson, G. B. (2003). Semantic and phonological influences on the processing of words and pseudohomophones. Memory & Cognition, 31, 856–866. Yip, M. C. W. (2000). Spoken word recognition of Chinese homophones: The role of context and tone neighbors. Psychologia, 43, 135–143. Yip, M. C. W. (2002). Access to homophonic meanings during spoken language comprehension: Effects of context and neighborhood density. In Proceedings of ICSLP’2002, pp. 1665–1668. Yip, M. C. W. (2007). Spoken word recognition of Chinese homophones: A further investigation. In Proceedings of interspeech’2007, pp. 362–365. Yip, M. C. W. (2008a). Lexical and sentential priming of ambiguous words. Psychologia, 51, 196–205. Yip, M. C. W. (2008b). Inhibitory processes of Chinese spoken word recognition. In Proceedings of interspeech’2008, pp. 2873–2876. Yip, M. C. W. (2009). Context effects and the processing of ambiguous words: Further evidence from semantic incongruence. In Proceedings of interspeech’2009, pp. 1871–1874. Zhang, L.-Y., & Zhang, S.-Y. (1987). Guoyin Yueyin Suoyin Zihui (A Pronunciation Dictionary of Cantonese and Putonghua). Hong Kong: Chung Hwa Bookstore. Zhang, Y, Shu, H., Li, P., & Liu, Y. (2001). Inhibitory processes, homophone meanings recognition, and spoken discourse comprehension in Chinese. In Proceedings of the third international conference on cognitive science, pp. 221–226. Zhang, Y., Wu, N., & Yip, M. C. W. (2006). Lexical ambiguity resolution in Chinese sentence processing. In P. Li, L. H. Tan, E. Bates, & O. J. L. Tzeng. (eds.) Handbook of East Asian psycholinguistics (vol. 1: Chinese), pp. 268–278. Cambridge, UK: Cambridge University Press.

123

Meaning Inhibition and Sentence Processing in Chinese.

The present study examined the inhibitory processes of spoken word recognition of Chinese homophones during sentence processing, using a standard cros...
175KB Sizes 2 Downloads 3 Views