ISSN 0439-755X
CN 11-1911/B


    25 April 2013, Volume 45 Issue 4 Previous Issue    Next Issue

    For Selected: Toggle Thumbnails
    The Representation and Processing of Subjective Words for Chinese Readers
    YAN Guoli;ZHANG Lanlan;SUN Shasha;BAI Xuejun
    2013, 45 (4):  379-390.  doi: 10.3724/SP.J.1041.2013.00379
    Abstract ( 1029 )   PDF (419KB) ( 2014 )  
    Given that there are no visual spaces between words in written Chinese and characters are the basic perceptual unit, it is intriguing how Chinese readers segment and represent words. Many studies have indicated that words play an important role in Chinese reading (Yan et al., 2006; Li et al., 2005; Rayner et al., 2007; Bai et al., 2008; Yan et al., 2009), while some other research has demonstrated that characters are more important than words in reading (Chen et al., 2003). However, some researchers have found that native Chinese readers have no clear concept of words, and they often disagree on how to divide the continuous string of characters within a sentence into words (Hoosain, 1992; Tsai et al., 1998; Miller et al., 2007). So there is a discrepancy here: If words play an important role in Chinese reading, what are the characteristics of these words? We propose that Chinese readers segment text into words according to complex cognitive representations: “subjective words” (e.g., representing the phrase “economic development” as a single word). Three studies were conducted to explore 1) whether subjective words are psychologically real in readers’ minds and 2) how these subjective words are processed. The first study examined how well the readers could recognize the two-character combination as a word or a phrase. The results showed that the readers tend to judge phrases as subjective words. which indicated that they represented words according to their own complex cognition about words, which demonstrated that subjective words were real in readers’ minds. The second and third experiment investigated the characteristics of subjective word processing. The second experiment explored whether subjective words had a “word superiority effect” — whether the readers found the position of a character in subjective words more efficiently than in non-words. The third experiment examined whether the reaction time of classifying subjective words as words was shorter than nonwords by using a lexical decision task. The results showed that there was no significant difference between words and subjective words in reaction time and accuracy. The accuracy rate of searching for the position of a character in subjective words was higher than in nonwords in the second experiment. In the third experiment, we didn’t find any significant difference between subjective words and nonwords in accuracy rate. To sum up, the present study indicated that native Chinese readers don’t judge a two-character combination as a word or a phrase accurately, and that they often confused them. It also demonstrated that subjective words are psychologically real in readers’ minds, which shows that the readers represented subjective words, and that some subjective words violated the Chinese grammar rules. The second experiment found that the word superiority effect in subjective words is the same as in the words compared with nonwords under compulsive choice task. The third experiment indicated that subjective words are processed more efficiently than nonwords. Subjective words are more likely to be processed as a whole.
    Related Articles | Metrics
    Effect of Semantic Decomposability on Understanding of Chinese Three-Character Idioms
    MA Lijun;ZHANG Jijia;DU Kai
    2013, 45 (4):  391-405.  doi: 10.3724/SP.J.1041.2013.00391
    Abstract ( 734 )   PDF (422KB) ( 1230 )  
    Idiomatic expressions are widely used as one of figurative languages when people sensed that idiom contained more than its surface meaning. For a long time, researchers have done a lot of work studying the mechanism of processing idioms, and two models were proposed: noncompositional model and compositional model. Those two models, however, can only explain one aspect of processing idioms. Therefore, hybrid model (Cutting & Bock, 1997) came into being in explaning how people process idioms: top-down process and bottom-up process co-exist. Compositionality was introduced in psycholinguistics by Gibbs and his colleagues (Gibbs, Nayak, & Cutting, 1989), and later developed into the so-called idiom semantic decomposition hypothesis, which made great contribution to generating hybrid model. According to this view, idioms are at least partly decomposable, and speakers have shared intuitions about how the meanings of the parts contribute to the idiomatic meaning. The degree of analyzability determines the access of idioms’ figarative meaning. During recent years, semantic analyzability has become very popular among psycholinguists, and the claims of the idiom semantic decomposition hypothesis are now widely accepted (Libben & Titone, 2008). Some questions remain unsettled, however, including: (1) whether semantic decomposability affects understanding idioms independently; (2) if so, how semantic decomposability affects it; (3) whether semantic decomposition hypothesis is suitable in processing Chinese modifier-noun structure; (4) which model can be applied in accessing Chinese three-character idioms. The present study tried to answer those questions by exploring semantic priming task to investigate how Chinese three-character idioms with verb-object structure and with modifier-noun structure were processed. We intended to reveal the role semantic decomposability played in processing idioms in early (SOA=300ms in Experiment 1) and later (SOA=800ms in Experiment 2) period. The results showed that semantic decomposability affected semantic activation independently. In early period, as to high decomposable idioms with verb-object structure and with modifier-noun structure, we found significant activation in their literal meanings; on the other hand, in later period, as to low decomposable idioms with modifier-noun structure, we found significant accuracy in activating literal and figurative meanings. The results also showed that the degree of semantic decomposability resulted in different effects of semantic priming. In early period, as to low decomposable idioms with verb-object structure, we found significant activation in both literal and figurative meanings, and we found activation in literal meanings when idioms with verb-object structure and with modifier-noun structure were high decomposable. In later period, significant activation in both literal and figurative meanings was found in low decomposable idioms with verb-object structure. But significant activation in figurative meanings was only found in high decomposable idioms with modifier-noun structure. These results supported the notion that semantic decomposability plays a role in processing idioms in early period; while in later period, as to high decomposable idioms with verb-object structure, the simultaneous activation of literal and figuratvie meanings weakens the advantage of semantic priming. So we made conclusions that (1) semantic decomposability independently affects understanding idioms; (2) figurative meanings are activated more easily when idioms are low semantic decomposable; whereas literal meanings tend to be easily activated when idioms are high semantic decomposable; (3) the privilege of activation in literal meaning exists in processing idioms. All those results partly support the Graded Salience Hypothesis and hybrid hypothesis.
    Related Articles | Metrics
    The Mental Representation of Color Information in Sentence Comprehension
    ZHANG Pan;LU Zhongyi
    2013, 45 (4):  406-415.  doi: 10.3724/SP.J.1041.2013.00406
    Abstract ( 1059 )   PDF (284KB) ( 2470 )  
    The mental representation of implicit color information during sentence comprehension remained to be answered in embodied cognition. In the current study, we adopted a sentence–picture matching paradigm with different settings of temporal intervals and sentences containing typical and atypical color information to address the question. A hundred and three college students with normal vision participated, with thirty-eight in Experiment 1, twenty-one in Experiment 2 and forty-four in Experiment 3. In Experiment 1, we determined the process of mental representation of color after sentence integration. In each trial, a sentence that implied a particular color for a given object was first presented, followed by an interval (0 ms or 1500 ms), then an object either matched or mismatched to the implied color appeared. Participants were required to judge whether the pictured object was mentioned in the preceding sentence or not and reaction time was registered. In Experiment 2, online measure was adopted to explore whether the mental representation of static color occurs immediately and locally during sentence processing. Subjects were required to self-regulate their readings by pressing the Space key if they finish one word, which will also terminate the presentation of the current word and initiate next word. The object words were replaced by a picture. After finishing one sentence, subjects should make a sensibility judgment on it. In Experiment 3, offline measure (identical to Experiment 1) was adopted to explore the mental representation of dynamic color information. The sentences included two types: one changes typical color to atypical and the other reversed. Participants were required to judge whether the pictured object was mentioned in the preceding sentence. Experiment 1 is a 2 colors × 2 intervals × 2 picture condition mixed design; Experiment 2 is a 2 colors × 2 picture condition within-subjects design; Experiment 3 is a 2 colors × 2 intervals × 2 picture condition mixed design. All the data were analyzed with repeated-measure MANOVA. In Experiment 1, response time was shorter in the mismatched condition than that in the matched condition, at the earlier period of sentence comprehension (interval of 0 ms); the result was reversed at the later period (interval of 1500 ms). In Experiment 2, the reading time was shorter for the picture in the mismatched condition. For the first word following the picture object, there was a mis-match facilitation in the atypical color condition. In Experiment 3, when the sentence implied typical color, response time was shorter in the matched condition as opposed to that in the mismatched condition at the later period of sentence comprehension. Our findings provide abundant evidence for how the implicit perceptual information on object color is represented during sentence comprehension. Together with other empirical findings, the current study indicates: (1) Under time-limited condition, the appearance of matched or mismatched facilitation depends on whether the two tasks compete for common cognitive resource; (2) The representation of implicit static typical color information is an immediate and local phenomenon. On the contrary, the representation of implicit static atypical color information is an immediate but non-local phenomenon; (3) The dynamic color information could not be immediately represented at the early period of sentence comprehension, only the dynamic typical color information could be represented at the later period of sentence comprehension.
    Related Articles | Metrics
    Neurophysiological Mechanism of Implicit Processing of Vocal Emotion Transition
    CHEN Xuhai;YANG Xiaohong;YANG Yufang
    2013, 45 (4):  416-426.  doi: 10.3724/SP.J.1041.2013.00416
    Abstract ( 770 )   PDF (4560KB) ( 2168 )  
    Successful decoding of vocal emotion is critical for social interaction. Specifically, it is adaptively important to detect vocal emotion transition in time, since changes in vocal emotion are important signals for detecting the speakers’ emotion change and are common in spoken interactions. Despite the fact that human beings are born with competence to process vocal emotion transition efficiently, the underlying neurophysiological mechanism remains largely unclear. To answer this question, the present study acquired electroencephalogram (EEG) during implicit processing of two types of vocal emotion transitions (neutral to angry and angry to neutral) and their control unchanged vocal emotion from 15 healthy volunteers. Fifty sentences of neutral content produced by a trained native male actor of Mandarin Chinese in neutral and angry prosodies served as original materials. Then we constructed sentences with vocal emotion transition through cross-splicing. The participants were required to perform sound intensity change judgment when the EEG was recorded. In addition to the ordinary ERP analysis in time domain, we also conducted the analysis in frequency domain including event related spectral power (ERSP) and inter-trail coherence (ITC) to specify the neurophysiological source of vocal emotion transition processing. The results indicated that vocal emotion transition elicited N2/P3 complex, as well as theta (4~6 HZ) band power (ERSP) and inter-trail coherence (ITC) increase, irrespective of transition types. In addition, vocal emotion transition induced beta band power decrease and ITC increase. Moreover, these effects were modulated by transition types, specifically, the more intense transition of emotion resulted in shorter latency of N2/P3 complex, and much stronger increase of theta band ERSP and ITC enhancement, and different features in beta band power change. These findings suggested that the processing of vocal emotion transition is a process consisting of change detection and reintegration, similar to the processing of other auditory materials, such as spoken language and music. Moreover, theta and beta band power change and ITC increase appear to be the main neurophysiological source of the processing of vocal emotion transition. In addition, the present study implies that the analysis in frequency domain can be a useful tool for verifying the neurophysiological mechanism of cognitive processing.
    Related Articles | Metrics
    Emotional Prosody Modulates the Recognition of Affective Word: An ERP Study
    ZHENG Zhiwei;HUANG Xianjun;ZHANG Qin
    2013, 45 (4):  427-437.  doi: 10.3724/SP.J.1041.2013.00427
    Abstract ( 757 )   PDF (539KB) ( 1770 )  
    An important difference between spoken language and written language is that spoken language carries prosodic information. Previous studies investigated how speech prosody modulated syntactic processing and discourse comprehension, but few studies examined the roles of emotional prosody in speech processing. This study investigated the time course of the processing of emotional spoken speech of Mandarin, and explored that how emotional prosody modulated the processing of spoken word using the technique of event-related potential and prosody/word interference paradigm. Thirty-two college students participated in this study, sixteen of them for the first experiment, and the rest for the second one. We used positive and negative words spoken with affectively congruent or incongruent prosody as study materials. Participants completed a delayed matching task. They were asked to judge that whether the spoken word matched the subsequent presented visual word. We recorded the EEG evoked by emotional spoken words. The emotional prosody was blocked presented (Experiment 1) or mixed presented (Experiment 2). The ERP results were analyzed by four-way repeated-measure ANOVA. The ERP results of Experiment 1 indicated that at the time window of 150~250ms after the onset of spoken word, the congruent affective word evoked a more positive P2 than incongruent affective word under both conditions of happy and sad prosody, and the scalp distribution of the P2 effect was broad. At the time window of 250~550ms, the incongruent affective word evoked more negative N300 (early N400) and N400 than congruent affective word. The N300 effect spread widely and the scalp distribution of N400 effect was modulated by the emotional prosody. The N400 effect of happy prosody distributed over the whole scalp while that of sad prosody mainly distributed at the parietal-occipital area. Experiment 2 showed similar results. The incongruent affective word evoked P2 effect, but only at the condition of sad prosody. The incongruent affective word evoked N300 and N400 effect independent of valence of emotional prosody, and both effects were bilaterally distributed over the whole scalp. These findings suggest that the emotional prosody modulates the processing of affective word at real-time. The emotional prosody can facilitate both of the processing of phonology and the retrieval of semantics. This study emphasizes the effects of facilitation of emotional prosody on the processing of spoken word, which begin at the stage of phonological processing.
    Related Articles | Metrics
    Neural Basis of Behavioral Sex Difference during Local Image Generation
    ZHAO Qingbai;ZHANG Xiaofei;SUI Danni;ZHOU Zhijin;CHEN Qicai;ZHOU Zongkui
    2013, 45 (4):  438-445.  doi: 10.3724/SP.J.1041.2013.00438
    Abstract ( 732 )   PDF (520KB) ( 1212 )  
    Previous behavioral studies reported that females performed faster than males in generating the local details of image. However, the neural basis of behavioral sex difference during local image generation is not clear. Generally, reaction time is considered as an index of speed of brain information processing. And the small-world topology of functional brain connectivity could provide high global and local efficiency of parallel information processing. Therefore, it was speculated that stronger small-world effect underlay females’ faster reaction time during local image generation. Total 28 people (14 males and 14 females) participated in the experiment and they were informed to observe a set of experimental pictures and then generate corresponding mental image to the global or local cue. The reaction time and event-related potential were recorded. The behavioral and electrophysiological data obtained in the experiment were analyzed by ANOVA and small-world measures, respectively. Results showed that females had the shorter reaction time, the smaller P300-650 amplitude and the shorter average path length of brain network in local image generation. Additionally, there was a positive correlation between the reaction time and the average path length. The present findings suggested that the lower cost and higher efficiency of brain information processing supported the better behavioral performance.
    Related Articles | Metrics
    Inhibition of Return in Trait Anxieties Under Different Cues
    ZHANG Yu;ZHENG Xifu;HUANG Shanshan;LI Yue;DU Xiaofen;ZHOU Wei
    2013, 45 (4):  446-452.  doi: 10.3724/SP.J.1041.2013.00446
    Abstract ( 1014 )   PDF (262KB) ( 2290 )  
    Inhibition of return (IOR), as an important part of attention bias, was generally considered as an adaptive mechanism during human evolution. Previous studies on the peculiarities of anxiety individual’s attention mainly focused on the attention orientation and inhibition ability; the effects of the natures of the stimuli on post-orientation inhibition, and the relation between the anxiety individual’s features and attentional bias, are still unclear. In the present study, we will focus on the attention bias of the anxiety group, and try to figure out the possible reasons underlying IOR features by using the cue-target paradigm with different cues. Traits Anxiety Inventory was used to test the undergraduates before the experiment, and finally twenty-nine undergraduates with high trait anxiety and twenty-eight with low trait anxiety were chosen as participants. In two experiments, participants were instructed to perform a cue-target task as accurately and quickly as possible, which included two kinds of uninformative cues: neutral cognitive cues and emotional face cues. Experiment 1 suggested that participants of high anxiety reacted more slowly than that of low anxiety under the condition of neutral cognitive cues. Although IOR did not appear in this experiment. Experiment 2 indicated that IOR effect occurred under all SOA levels for both groups, and IOR magnitudes of the two groups were modulated by emotional cues. The two groups showed no significant difference on IOR effect under the condition of positive cues, while IOR effect of the high anxiety group turned out to be much smaller than that of the low anxiety group under the condition of negative cues. These experiments suggested that only under negative emotional circumstances, groups of different trait anxiety showed different IOR abilities, and particularly, high anxiety group’s IOR ability was impaired, which provided specific information on possible improvement of the anxiety symptoms.
    Related Articles | Metrics
    The Influence of Leader on Group Member’s Cooperation in Common Resource Dilemmas
    ZHANG Zheming, JIN Shenghua, WU Song, ZHOU Xiang
    2013, 45 (4):  453-465.  doi: 10.3724/SP.J.1041.2013.00453
    Abstract ( 808 )   PDF (591KB) ( 2277 )  
    In past leadership research using common resource dilemmas, most researchers have studied the behavior of participants who identified as the leader. Individuals labeled “leaders” tended to harvest more common resources and impede the implementation of the public interest (De Cremer, 2003; De Cremer & Van Dijk, 2005; Van Dijk & De Cremer, 2006). Less is known, however, about the effects of establishing a leader on the behavior of the participants. If the betrayal of the leader could stimulate the self-restraint behavior of the group members, then the establishing a leader might promote the realization of the public interest. Three experiments were used to examine the impact of a leader’s behavior on the group member’s decision making. Experiments 1 manipulated the presence or absence of a leader and tested how the different behavior of a leader or common group member would influence the group member’s decision. Participants showed more self-restraint behavior when a leader showed the betrayal than when a common group member showed the betrayal. The participants also found the leader’s betrayal fairer, and regarded the leader as out-group member. Experiment 1 didn’t find the explanatory differences between Equity Theory and Social Identity Theory. Experiment 2 was conducted to evaluate the group member’s reaction to appointed and elected leaders. Participants showed more self-restraint behavior when an appointed leader showed the betrayal than when an elected leader did so. The participants also found the betrayal of the elected leader fairer, and regarded the appointed leader as out-group member. Experiment 2 confirmed the explanatory power of Social Identity Theory. Experiment 3 tested the impact of the betrayal of the in-group and out-group elected leader to the group member, in order to examine the Social Identity Theory. Participants showed more self-restraint behavior when out-group elected leader showed the betrayal. In contrast, when the in-group elected leader showed the betrayal, participants did not show self-restraint behavior. The results from all three experiments supported the hypotheses and suggested implications for organization management. Elected leaders should display justice and try to avoid the pursuit of personal interest. Leaders also should keep distance from their subordinates in some circumstances. Future research on leadership using the social dilemmas could evaluate the effects of personal variables.
    Related Articles | Metrics
    Comparison of MIRT Linking Methods for Different Common Item Designs
    LIU Yue;LIU Hongyun
    2013, 45 (4):  466-480 .  doi: 10.3724/SP.J.1041.2013.00466
    Abstract ( 702 )   PDF (1219KB) ( 962 )  
    A great number of educational assessments usually measure more than one trait (Ackerman, 1992; DeMars, 2006; Reckase, 1985). In order to adjust scores on these different test forms, multidimensional item response theory (MIRT) and its linking procedures should be developed. So far, some researchers have already extended UIRT linking methods to the multidimensional structure (Davey et al., 1996; Hirsch, 1989; Li & Lissitz, 2000; Min, 2003; Yon, 2006). There were numerous studies comparing MIRT linking methods in the literature. However, although choosing anchor items was of great importance in common item designs, a few of studies compared MIRT linking methods under different common item designs. It was still in doubt that, how we could choose the common items across different MIRT linking methods. The purpose of this study was to compare five MIRT linking methods under two kinds of common item choosing strategies in various situations. The study was a mixed measure design of simulation conditions (between-factors) and linking methods (within-factor). There were six between-factors: (1) 2 test lengths (40 items and 80 items); (2) 2 levels of the proportion of the number of items in one dimension to another (1:1 and 1:3); (3) 3 anchor lengths (1/20, 1/5 to 1/3 of total test); (4) 2 strategies of choosing common items (averagely choosing the items in all dimensions or choosing according to the proportions of items in every dimension); (5) 3 correlations between two ability dimensions (r=0, 0.5, 0.9); (6) 2 levels of equivalent/non-equivalent ability levels between two populations. The five MIRT linking methods we investigated were: Mean/Mean (MM) method, Mean/Sigma (MS) method, Stoking-Lord’s (SL) method, Haebara’s (HB) method and Least Square (LS) method. Under each condition, the number of examinees was fixed as I =2000, and 30 replications were generated. BMIRT (Yao, 2003) was applied to estimate item and ability parameters using an MCMC method. Based on the previous studies about equating (Kim & Cohen, 1998; Kim & Cohen, 2002), a two-step of linking was applied. The first step was to transform the scale of parameters in the new test onto the base test, and the second step was to transform the scale of all the simulated items onto the generating scale. In each step, the transformation matrices were produced by LinkMIRT (Yao, 2004) and the R package called “Plink” (Weeks, 2010). Finally, the recovery of parameters was evaluated by four criteria: bias, mean absolute error, root mean square error, correlation between the parameters after equating and true values. To compare the five MIRT equating methods, the results showed that: the RMSE for parameters under SL, HB and LS methods were smaller and more stable in different situations; however, the RMSE for parameters in MM and MS methods were significantly large, especially in non-equivalent group conditions. Therefore, the latter results were displayed for the SL, HB and LS methods. It was found that these methods were not affected by the common item design. It meant that in multidimensional linking, if the number of common items was more than 5% of the total test, the RMSE became acceptable. Meanwhile, the strategy of choosing common items didn’t have significant influence on the linking results of the three methods across different conditions of test structure. Moreover, for other simulation factors: as test length increased, the RMSE of these methods decreased; as the correlations between two ability dimensions increased, the RMSE of ability parameter decreased; the difference of ability levels between two populations had smaller effect on these methods, that only for intercept parameter, the non-equivalent group condition produced larger error. In conclusion, SL, HB and LS methods generally performed better than the other two methods across all conditions, so it was highly recommended to use these methods in practical. The performances of SL, HB and LS methods were similar under different common item designs, which was amazing for MIRT linking. Once an appropriate method was chosen, shorter anchor set could be applied, as developing good common items for multidimensional tests was quite time-consuming. Meanwhile, the common items could be chosen either according to the proportions of items in every dimension or averagely in all dimensions. This might be more convenient for practitioners as well. Lastly, as test length had significant effect on the accuracy of equated parameters, it wss suggested to make sure the test was comprised of enough items in every dimension before conducting an MIRT linking.
    Related Articles | Metrics
    Cognition and Body: A Perspective from Theoretical Psychology
    YE Haosheng
    2013, 45 (4):  481-488.  doi: 10.3724/SP.J.1041.2013.00481
    Abstract ( 1638 )   PDF (308KB) ( 4298 )  
    With the approach of theoretical psychology, this paper is to explore the possible relations between cognition and body in light of embodiment thesis. According to present author, Dualist viewpoints of mind-body relationship, until recently, have been in dominant position in western culture. The denial of the body in consideration of human mind has been a heritage of the western intellectual zeitgeist since the time of the ancient Greeks. Plato put the body in the position of distraction in intellectual life, and separated the human soul from the body. He is the earliest representative of dualism. In the 17th century, a philosopher of France, Rene Descartes, epistemologically, demonstrated the existence of the dualist world, and distinguished between physical substances and thinking substance. According to Descartes, physical substances (“res extensa”)could be measured and divided, and occupied a physical space, while the thinking substance (“res cogitans”) could not be divided, and was unextended into physical space. Following in Descartes’ footsteps, traditional cognitive science took the metaphor, the MIND IS A COMPUTER. Human mind has been modeled as a digital computer, and cognition was considered as an autonomous, logical, and disembodied process. There was a sharp line between human physical capacities and its intellectual abilities. This kind of dualist viewpoints has been challenged, however, in recent decades in cognitive science. One of the most recent developments that has implications for understanding the relationship between cognition and body is that of embodied mind. According to proponents of theoretical models of embodiment, cognition is body’s cognition, and body is the subject rather than the object of cognition. In author’s opinion, the dependence of cognition on body can be in the following aspects: (1) an agent’s body in action is a powerful constraint on how the agent conceive their environments. Because of its bodily shape and structure, some forms of cognition tend to be easier, and some kinds of cognition tend to be difficult even impossible. (2) an agent’s body can function to distribute cognitive tasks between brain and body, and between body and environment. This means that the mind extends beyond the skull and skin into the outer physical environment, and cognitive systems may loop into the world and hence constitute hybrid systems consisting of both neural and non-neural parts of the body. (3) an agent’s body can regulate cognitive activity and influence mental processes. Nodding our heads result in more positive evaluations, while shaking our heads result in more negative evaluations. This indicate our body’s influence on our mind. Generally speaking, body and mind are not separated but unified. The interaction between body and the world made the mind become possible.
    Related Articles | Metrics