MELD-SCH: A megastudy of lexical decision in simplified Chinese

A generative approach to extrapolate word concreteness ratings

A wealth of psycholinguistic and clinical research is supported by normative ratings of lexicosemantic properties, e.g., word concreteness, word valence, age-of-acquisition, etc. Collecting such ratings for a sufficiently large number of words is, however, notoriously labour-intensive. This study utilised the mixture density network (MDN), a generative approach, to implement a computational expansion of the concreteness ratings for simplified Chinese words. Based on different word embeddings, the MDN was trained to generate the probability density of a word's trial-level ratings, allowing us to predict not only the word's mean concreteness rating (con.mean), but also the potential variability (con.var) in people's perceptions about the word's concreteness. The resulting estimates were shown to largely converge with human ratings in both central tendency and variability, and to precisely reflect the important representational features of the construct. Apart from these internal validations, we also examined the contributions of con.mean to Chinese lexical processing. The results revealed the concreteness effect on event-related potentials associated with visual word recognition. To assist and enhance future research, we released the extrapolated concreteness ratings, along with degrees of variability, for over 78,000 Chinese words in the Open Science Framework (https://osf.io/gwr5h/).

Composition as Nonlinear Combination in Semantic Space: A Computational Characterization of Compound Processing

Most Chinese words are compounds formed through the combination of meaningful characters. Yet, due to compositional complexity, it is poorly understood how this combinatorial process affects the access to the whole-word meaning. In the present study, we turned to the recent development in compositional distributional semantics, and employed a deep neural network to learn the less-than-systematic relationship between the constituent characters and the compound words. Based on the compositional representations derived from the computational model, we investigated the combinatorial process in terms of the degree of overlap between the compositional and the lexicalized representations as well as the degree of distinctness of the compositional representation. Analyses of lexical decision and eye-tracking data revealed the effects of both compositional attributes over and above the effects of constituent character features and compound features, indicating an active engagement of the combinatorial process in compound processing. Moreover, with the increase of compound frequency, and thus the increased likelihood that the holistic route prevails, these compositional effects appeared to be dampened. These findings, therefore, provided a computational characterization for the dual-route framework, which sheds light on the universal process of compound comprehension across different languages.

Coregistration of eye movements and EEG reveals frequency effects of words and their constituent characters in natural silent Chinese reading

We conducted two experiments to examine the lexical and sub-lexical processing of Chinese two-character words in reading. We used a co-registration electroencephalogram (EEG) for the first fixation on target words. In Experiment 1, whole-word occurrence frequency and initial constituent character frequency were orthogonally manipulated, while in Experiment 2, whole-word occurrence frequency and end constituent character frequency were orthogonally manipulated. Results showed that word frequency facilitated eye-tracking measures, while initial and end character frequencies inhibited them. Classical word frequency effects on N170 and N400 in the posterior region and reversed word frequency effects over the anterior region were consistently observed in both experiments. Experiment 1 revealed an inhibiting effect of initial character frequency on anterior N170. In Experiment 2, interaction between end-character frequency and word frequency showed reliable effects on anterior N170 and N400. These results demonstrate both facilitating and inhibiting word frequency effects, along with inhibiting effects of character frequency and that word frequency moderates the inhibiting effects of end constituent character frequency during natural silent Chinese reading.

Perception of emotion across cultures: Norms of valence, arousal, and sensory experience for 4923 Chinese words translated from English in Warriner et al. (2013)

Perception of emotion conveyed through language is influenced by embodied experiences obtained from social interactions, which may vary across different cultures. To explore cross-cultural differences in the perception of emotion between Chinese and English speakers, this study collected norms of valence and arousal from 322 native Mandarin speakers for 4923 Chinese words translated from Warriner et al., (Behavior Research Methods, 45, 1191-1207, 2013). Additionally, sensory experience ratings for each word were collected. Analysis demonstrated that the reliability of this dataset is satisfactory, as indicated by comparisons with previous datasets. We examined the distributions of valence and arousal for the entire dataset, as well as for positive and negative emotion categories. Further analysis suggested that valence, arousal, and sensory experience correlated with various psycholinguistic variables, including the number of syllables, number of strokes, imageability, familiarity, concreteness, frequency, and age of acquisition. Cross-language comparison indicated that native speakers of Chinese and English differ in their perception of emotional valence and arousal, largely due to cross-cultural variations associated with ecological, sociopolitical, and religious factors. This dataset will be a valuable resource for research examining the impact of emotional and sensory information on Chinese lexical processing, as well as for bilingual research investigating the interplay between language and emotion across different cultural contexts.

A large-scale database of Mandarin Chinese word associations from the Small World of Words Project

Word associations are among the most direct ways to measure word meaning in human minds, capturing various relationships, even those formed by non-linguistic experiences. Although large-scale word associations exist for Dutch, English, and Spanish, there is a lack of data for Mandarin Chinese, the most widely spoken language from a distinct language family. Here we present the Small World of Words-Zhongwen (Chinese) (SWOW-ZH), a word association dataset of Mandarin Chinese derived from a three-response word association task. This dataset covers responses for over 10,000 cue words from more than 40,000 participants. We constructed a semantic network based on this dataset and evaluated concurrent validity of association-based measures by predicting human processing latencies and comparing them with text-based measures and word embeddings. Our results show that word centrality significantly predicts lexical decision and word naming speed. Furthermore, SWOW-ZH notably outperforms text-based embeddings and transformer-based large language models in predicting human-rated word relationships across varying sample sizes. We also highlight the unique characteristics of Chinese word associations, particularly focusing on word formation. Combined, our findings underscore the critical importance of large-scale human experimental data and its unique contribution to understanding the complexity and richness of language.

A corpus of Chinese word segmentation agreement

The absence of explicit word boundaries is a distinctive characteristic of Chinese script, setting it apart from most alphabetic scripts, leading to word boundary disagreement among readers. Previous studies have examined how this feature may influence reading performance. However, further investigations are required to generate more ecologically valid and generalizable findings. In order to advance our understanding of the impact of word boundaries in Chinese reading, we introduce the Chinese Word Segmentation Agreement (CWSA) corpus. This corpus consists of 500 sentences, comprising 9813 character tokens and 1590 character types, and provides data on word segmentation agreement at each character position. The data revealed a high level of overall segmentation agreement (92%). However, participants disagreed on the position of word boundaries in 8.96% of the cases. Moreover, about 85% of the sentences contained at least one ambiguous word boundary. The character strings with high levels of disagreement were tentatively classified into three categories, namely the morphosyntactic type (e.g., "-"), modifier-head type (e.g., "-"), and others (e.g., "-"). Finally, the agreement scores also significantly influenced reading behaviors, as evidenced by analyses with published eye movement data. Specifically, a high level of disagreement was associated with longer single fixation durations. We discuss the implications of these results and highlight how the CWSA corpus can facilitate future research on word segmentation in Chinese reading.

Exploring the differences in processing between Chinese emotion and emotion-laden words: A cross-task comparison study

Affective words can be classified into two types: emotion words (EM words, e.g., "happy") and emotion-laden words (EL words, e.g., "wedding"). Several studies have shown differences in processing between EM and EL words, although results are inconclusive. These two types of words may have representational differences because affective content is an inherent part of the semantic features of EM words (i.e., denotative meaning) but not of EL words, whose affective content is part of their connotative meaning (i.e., these words do not name emotions, but are associated with emotions). In this study, we tested a set of Chinese EM and EL words. Both conditions included positive and negative words. The study involved two tasks, an implicit task, in which emotional content was not relevant (lexical decision task, LDT), and an explicit task, in which the emotional content was relevant (affective categorisation task, ACT). Our results showed that participants responded faster to EM words than to EL words. This advantage was mostly observed in the ACT and with negative words. These results reveal differences in processing between EM and EL words which can be related to the greater relevance of affective content in the meaning of EM words compared with EL words.

HeLP: The Hebrew Lexicon project

Lexicon projects (LPs) are large-scale data resources in different languages that present behavioral results from visual word recognition tasks. Analyses using LP data in multiple languages provide evidence regarding cross-linguistic differences as well as similarities in visual word recognition. Here we present the first LP in a Semitic language-the Hebrew Lexicon Project (HeLP). HeLP assembled lexical decision (LD) responses to 10,000 Hebrew words and nonwords, and naming responses to a subset of 5000 Hebrew words. We used the large-scale HeLP data to estimate the impact of general predictors (lexicality, frequency, word length, orthographic neighborhood density), and Hebrew-specific predictors (Semitic structure, presence of clitics, phonological entropy) of visual word recognition performance. Our results revealed the typical effects of lexicality and frequency obtained in many languages, but more complex impact of word length and neighborhood density. Considering Hebrew-specific characteristics, HeLP data revealed better recognition of words with a Semitic structure than words that do not conform to it, and a drop in performance for words comprising clitics. These effects varied, however, across LD and naming tasks. Lastly, a significant inhibitory effect of phonological ambiguity was found in both naming and LD. The implications of these findings for understanding reading in a Semitic language are discussed.

Revealing hidden interactions in mean performance through distributional analyses: Evidence from Chinese lexical decision performance

Important insights in visual word recognition have been provided by studies examining the combined influence of multiple factors on participants' mean response times to English words in the lexical decision task. However, to make progress towards a complete understanding of how meaning is activated by print, researchers need to conduct more detailed analyses of behavioral patterns beyond mean response latencies and accuracies, particularly how variables influence different components of response time distributions. Moreover, it is critical to extend patterns found in English to the diverse scripts encountered by readers across the world. The present study is the first to explore the theoretically important effects of stimulus quality and word frequency on lexical decisions involving two-character Mandarin Chinese and Cantonese Chinese words, using participants from Singapore and Hong Kong, respectively. Despite the profound differences between the English and Chinese writing systems, we observed remarkably similar trade-offs in the stimulus quality × word frequency interaction across different portions of the response time distribution for both orthographies, indicating that the optimization of lexical processing by leveraging available codes in response to task demands extends across multiple and highly diverse writing systems.

The good, the bad, and the ambivalent: Extrapolating affective values for 38,000+ Chinese words via a computational model

Word affective ratings are important tools in psycholinguistic research, natural language processing, and many other fields. However, even for well-studied languages, such norms are usually limited in scale. To extrapolate affective (i.e., valence and arousal) values for words in the SUBTLEX-CH database (Cai & Brysbaert, 2010, PLoS ONE, 5(6):e10729), we implemented a computational neural network which captured how words' vector-based semantic representations corresponded to the probability densities of their valence and arousal. Based on these probability density functions, we predicted not only a word's affective values, but also their respective degrees of variability that could characterize individual differences in human affective ratings. The resulting estimates of affective values largely converged with human ratings for both valence and arousal, and the estimated degrees of variability also captured important features of the variability in human ratings. We released the extrapolated affective values, together with their corresponding degrees of variability, for over 38,000 Chinese words in the Open Science Framework ( https://osf.io/s9zmd/ ). We also discussed how the view of embodied cognition could be illuminated by this computational model.

Phonological Well-Formedness Constraints in Mandarin Phonotactics: Evidence From Lexical Decision

This article investigates the role of phonological well-formedness constraints in Mandarin speakers' phonotactic grammar and how they affect online speech processing. Mandarin non-words can be categorized into systematic gaps and accidental gaps, depending on whether they violate principled phonotactic constraints based on the Obligatory Contour Principle (OCP). Non-word acceptability judgment experiments have shown that systematic gaps received lower wordlikeness ratings than accidental gaps. Using a lexical decision task, this study found that systematic gaps were rejected significantly faster than accidental gaps, even after lexical statistics were taken into account. These findings thus provide converging evidence for the essential status of the OCP-based phonotactic constraints in Mandarin speakers' phonological knowledge.

A large-scale database of Chinese characters and words collected from elementary school textbooks

Lexical databases are essential tools for studies on language processing and acquisition. Most previous Chinese lexical databases have focused on materials for adults, yet little is known about reading materials for children and how lexical properties from these materials affect children's reading comprehension. In the present study, we provided the first large database of 2999 Chinese characters and 2182 words collected from the official textbooks recently issued by the Ministry of Education (MOE) of the People's Republic of China for most elementary schools in Mainland China, as well as norms from both school-aged children and adults. The database incorporates key orthographic, phonological, and semantic factors from these lexical units. A word-naming task was used to investigate the effects of these factors in character and word processing in both adults and children. The results suggest that: (1) as the grade level increases, visual complexity of those characters and words increases whereas semantic richness and frequency decreases; (2) the effects of lexical predictors on processing both characters and words vary across children and adults; (3) the effect of age of acquisition shows different patterns on character and word-naming performance. The database is available on Open Science Framework (OSF) ( https://osf.io/ynk8c/?view_only=5186bd68549340bd923e9b6531d2c820 ) for future studies on Chinese language development.

Rethinking orthographic neighbor in Chinese two-character word recognition: Insights from a megastudy

The influence of orthographic neighbors on visual word recognition is well established in alphabetic scripts. To determine the universality of this effect across writing systems, researchers have been keen on exploring its presence and nature in Chinese word recognition. Given that Chinese is logographic, it necessitates a different definition for orthographic neighbors from the ones used in alphabetic scripts. One popular approach is to consider words that share characters as orthographic neighbors. Adopting this definition, a facilitative effect has been observed for characters that can create more words. However, as characters are also morphemes in Chinese, the facilitation found might actually come from a larger morphological family size. This possibility was tested in the present study by analyzing data from the Chinese Lexicon Project (CLP; Tse et al., Behavior Research Methods, 49, 1503-1519, 2017, Behavior Research Methods, 49, 1503-1519, 2022), a megastudy of two-character word recognition in traditional Chinese. If the effects of character-sharing are indeed morphological in nature, the facilitation should be smaller for ambiguous characters because the words formed are distributed over several morphological families. The results of the analyses were consistent with this hypothesis, revealing interactions between the number of words formed by a character and the number of meanings of the character. The implications of these findings were discussed in the context of definitions of orthographic neighbors and theories of word recognition in Chinese.

Effects of Lexical Properties in L2 Chinese Compound Processing: A Multivariate Approach

Previous studies on L2 (i.e., second language) Chinese compound processing have focused on the relative efficiency of two routes: holistic processing versus combinatorial processing. However, it is still unclear whether Chinese compounds are processed with multilevel representations among L2 learners due to the hierarchical structure of the characters. Therefore, taking a multivariate approach, the present study evaluated the relative influence and importance of different grain sizes of lexical information in an L2 Chinese two-character compound decision task. Results of supervised component generalized linear regression models with random forests analysis revealed that the orthographic, phonological and semantic information all contributed to L2 compound processing, but the L2 learners used more orthographic processing strategies and fewer phonological processing strategies compared to the native speakers. Specifically, the orthographic information was activated at the whole-word, the character and the radical levels in orthographic processing, and the phonological information at the whole-word, the syllable, and the phoneme levels all exerted contributions in phonological processing. Furthermore, the semantic information of the whole words and the constituents was accessed in semantic processing. These findings together suggest that the L2 learners are able to use cues at all levels simultaneously to process Chinese compound words, supporting a multi-route model with a hierarchical morphological structure in such processing.

Assessing lexical ambiguity of simplified Chinese characters: Plurality and relatedness of character meanings

Lexical ambiguity is pervasive among Chinese characters as many of them are polysemantic, with one orthographic form carrying unrelated meanings, related meanings, or sometimes both unrelated and related meanings. A large-scale database with ambiguity measures for simplified Chinese characters has yet to be developed, which could greatly benefit psycholinguistic research on the Chinese language or cross-language comparisons. This article reports two sets of ratings by native speakers, the perceived number of meanings (pNoM) for 4,363 characters and the perceived relatedness of meanings (pRoM) for a subset of 1,053 characters. These rating-based ambiguity measures capture the representational nuance about a character's meanings stored in average native speakers' mental lexicon, which tends to be obscured by dictionary- and corpus-based ambiguity measures. Consequently, they each account for a reliable portion of variance in the efficiency of character processing, above and beyond the effects of character frequency, age of acquisition, and other types of ambiguity measures. Theoretical and empirical implications with regard to the plurality and the relatedness of character meanings, the two focal aspects of debate on lexical ambiguity, are discussed.

LexCHI: A quick lexical test for estimating language proficiency in Chinese

A prominent methodological issue in cognitive research on bilingualism is the lack of consistency in measuring second language (L2) proficiency. To reduce the inconsistency in L2 proficiency measurements, brief and valid vocabulary tests have been developed as an objective measure of proficiency in a variety of languages (e.g., English, French, Spanish). Here, we present LexCHI, a valid lexical test to measure Chinese proficiency. This freely available short test consists of 60 two-character items presented in simplified Chinese. Although it only takes a few minutes to complete LexCHI, the LexCHI scores in two studies correlated significantly with L2 participants' performance in a translation task and a cloze test. We believe that LexCHI is a useful tool for researchers who need to objectively measure Chinese proficiency as part of their investigations.

Comparing word recognition in simplified and traditional Chinese: A megastudy approach

Although it is well established that the visual complexity of a written word can influence processing, it is far less clear from a cross-script perspective, whether the overall visual complexity of the entire written lexicon also affects word recognition. This question can be answered with the data in megastudy of lexical decision in Chinese (MELD-CH), which was developed with over 800 participants responding to 12,587 simplified and traditional Chinese words. The results showed that lexical decision was slower but more accurate in simplified Chinese, which has about 22.5% less strokes, than traditional Chinese. This pattern could not be explained by a speed-accuracy trade-off. Moderate correlations were found in response times and error rates between the two scripts, indicating considerable overlap in processing despite the script difference. In addition, (generalised) linear mixed-effects modelling was used to examine whether the simplified and traditional Chinese groups differed in sensitivity towards linguistic variables. The results showed that the effects of word frequency, word length, and number of strokes were stronger in recognising simplified Chinese words, whereas the effects of number of words formed and number of meanings of the constituent characters were stronger in traditional Chinese. These results suggested that the visual-perceptual processing demands of simplified Chinese might force readers to focus more on local properties of the words, making them less sensitive to global properties that are defined over the entire lexicon. Finally, limitations and alternative explanations of the results were discussed.

CCLOOW: Chinese children's lexicon of oral words

In this article, we introduce the Chinese Children's Lexicon of Oral Words (CCLOOW), the first lexical database based on animated movies and TV series for 3-to-9-year-old Chinese children. The database computes from 2.7 million character tokens and 1.8 million word tokens. It contains 3920 unique character and 22,229 word types. CCLOOW reports frequency and contextual diversity metrics of the characters and words, as well as length and syntactic categories of the words. CCLOOW frequency and contextual diversity measures correlated well with other Chinese lexical databases, particularly well with that computed from children's books. The predictive validity of CCLOOW measures were confirmed with Grade 2 children's naming and lexical decision experiments. Further, we found that CCLOOW frequencies could explain a considerable proportion in adults' written word recognition, indicating that early language experience might have lasting impacts on the mature lexicon. CCLOOW provides validated frequency and contextual diversity estimates that complements current children's lexical database based on written language samples. It is freely accessible online at https://www.learn2read.cn/ccloow .

Sentiment annotations for 3827 simplified Chinese characters

Sentiment analysis in Chinese natural language processing has been largely based on words annotated with sentiment categories or scores. Characters, however, are the basic orthographic, phonological, and in most cases, semantic units in the Chinese language. This study collected sentiment annotations for 3827 characters. The ratings demonstrated high levels of reliability, and were validated through a comparison with the ratings of some characters' word equivalents reported in a previous norming study. Relations with other lexico-semantic variables and character processing efficiency were investigated. Furthermore, analyses of the association between constituent character valence and word valence revealed semantic compositionality and sentiment fusion characteristic of larger Chinese linguistic units. These ratings for characters, expanding current Chinese sentiment lexicons, can be utilized for the purposes of more precise stimuli assessment in research on Chinese character processing and more efficient sentiment analysis equipped with annotations of single-character words.

Age of acquisition effects in Chinese two-character compound words: a megastudy of eye movements during reading

Words learned earlier in life are processed faster than words learned later in life. In Chinese, the age of acquisition (AoA) effect has been extensively studied in one-character words, yet no studies have explored this effect in compound words. Given that compound words are a widespread word type and are processed differently than one-character words, it is unclear how AoA affects compound word processing. The present study examined the role of multiple variables in the processing of Chinese two-character compound words based on a database of eye-movement measures. We focused on the AoA effect from the whole compound word and the first and second characters. We found that whole-word AoA and first-character AoA have an early and long-lasting influence, with no AoA effect from the second character. The results indicated that two-character compound words are accessed via the character and whole word simultaneously. These findings are discussed with respect to theories of the AoA effect.

The Chinese Lexicon Project II: A megastudy of speeded naming performance for 25,000+ traditional Chinese two-character words

Using a megastudy approach, (Tse et al., 2017 Behavior Research Methods, 49, 1503-1519) established a large-scale repository of lexical variables and lexical decision responses for more than 25,000 traditional Chinese two-character words. In the current study, we expand their database by collecting norms for speeded naming reaction times (RTs) and accuracy rates, and compiling more lexical variables (e.g., phonological consistency and semantic neighborhood size). Following Tse et al.'s procedure, about 33 college-aged native Cantonese speakers in Hong Kong read aloud each word. We conducted item-level regression analyses to test the relative predictive power of orthographic variables (e.g., stroke count), phonological variables (e.g., phonological consistency), and semantic variables (e.g., semantic transparency) in naming performance. We also compared the effects of lexical variables on naming performance and Tse et al.'s lexical decision performance to examine the extent to which effects are task-specific or task-general. Freely accessible to the research community, this resource provides a valuable addition to other influential mega-databases, such as the English Lexicon Project (Balota et al., 2004 Journal of Experimental Psychology: General, 133, 283-316), and furthers our understanding of Chinese word recognition processes.

Probing Lexical Ambiguity in Chinese Characters via Their Word Formations: Convergence of Perceived and Computed Metrics

Lexical ambiguity is pervasive in language, and the nature of the representations of an ambiguous word's multiple meanings is yet to be fully understood. With a special focus on Chinese characters, the present study first established that native speaker's perception about a character's number of meanings was heavily influenced by the availability of its distinct word formations, while whether these meanings would be perceived to be closely related was driven by further conceptual analysis. These notions were operationalized as two computed metrics, which assessed the degree of dispersion across individual word formations and the degree of propinquity across clusters of word formations, respectively, in a distributional semantic space. The observed correlations between the computed and the perceived metrics indicated that the utility of word formations to tap into meaning representations of Chinese characters was indeed cognitively plausible. The results have demonstrated the extent to which distributional semantics could inform about meaning representations of Chinese characters, which has theoretical implications for the representation of ambiguous words more generally.

Hong Kong Chinese character psycholinguistic norms: ratings of 4376 single Chinese characters on semantic radical transparency, age-of-acquisition, familiarity, imageability, and concreteness

Several norms of psycholinguistic features of Chinese characters exist in Mandarin Chinese, but only a few are available in Cantonese or in the traditional script, and none includes semantic radical transparency ratings. This study presents subjective ratings of age-of-acquisition (AoA), familiarity, imageability, concreteness, and semantic radical transparency in 4376 Chinese characters. The single Chinese characters were rated individually on the five dimensions by 20 native Cantonese speakers in Hong Kong to form the Hong Kong Chinese Character Psycholinguistic Norms (HKCCPN). The split-half reliability and intra-class correlations testified to the high internal reliability of the ratings. Their convergent and discriminant patterns in relations to other psycholinguistic measures echoed previous findings reported on Chinese. There were high correlations for semantic radical transparency, imageability and concreteness, and moderate-to-high correlations for AoA and familiarity among subsets of items that had been collected in previous studies. Concurrent validity analyses showed convergence in predicting behavioral response times in various tasks (lexical decision, naming, and writing-to-dictation) when compared with other Chinese character databases. High predictive validity was shown in writing-to-dictation data from an independent sample of 20 native Cantonese speakers. Several objective psycholinguistic measures (character frequency, stroke number, number of words formed, number of homophones and number of meanings) were included in this database to facilitate its use. These new ratings extend the currently available norms in language and reading research in Cantonese Chinese for researchers, clinicians, and educators, as well as provide them with a wider choice of stimuli.

Morphological decomposition in Chinese compound word recognition: Electrophysiological evidence

The present study examined the effect of both morphological complexity and semantic transparency in Chinese compound word recognition. Using a visual lexical decision task, our electrophysiological results showed that transparent and opaque compounds induced stronger Left Anterior Negativity (LAN) than monomorphemic words. This result suggests that Chinese compounds might be decomposed into their constituent morphemes at the lemma level, whereas monomorphemic words are accessed as a whole-word lemma directly from the form level. In addition, transparent and opaque compounds produced a similar N400 as each other, suggesting that transparency did not show an effect on the involvement of constituent morphemes during access to the whole-word lemma. Two behavioral experiments additionally showed similar patterns to the EEG results. These findings support morphological decomposition for compounds at the lemma level as proposed by the full-parsing model, and no evidence is found to support the role of transparency during Chinese compound word recognition.

Full-form vs. combinatorial processing of Chinese compound words: Evidence from mismatch negativity

This study examined whether Chinese spoken compound words are processed via full-form access or combination through morphemes by recording mismatch negativity (MMN). MMN has been shown to be larger for linguistic units that involves full-form access (lexical MMN enhancement) and smaller for separate but combinable units (combinatorial MMN reduction). Chinse compound words were compared against pseudocompounds, which do not have full-form representations in the long-term memory and are "illegal" combinations. All stimuli were disyllabic (bimorphemic). Word frequency was manipulated with the prediction that low-frequency compounds are more likely processed combinatorially, while high-frequency ones are more likely accessed in full forms. The results showed that low-frequency words elicited smaller MMNs than pseudocompounds, which supported the prediction of combinatorial processing. However, neither MMN enhancement nor reduction was found for high-frequency words. These results were interpreted within the dual-route model framework that assumes simultaneous access to words and morphemes.

Familiarity ratings for 24,325 simplified Chinese words

The present work collected familiarity norms for 20,275 two-character, 1231 three-character, and 2819 four-character simplified Chinese words from 1300 native speakers of Mandarin Chinese. The familiarity of each word was rated on a 7-point scale by at least 21 participants. The results supported the reliability and validity of the present familiarity ratings, which is the first large familiarity database for Chinese in the field. These familiarity norms can be downloaded from the supplemental materials. Furthermore, the contribution of familiarity to Chinese lexical processing was investigated using the present familiarity ratings and previous data (lexical features and visual lexical decision), mainly from two major Chinese lexicon projects, MELD-SCH and CLP. Regression analysis suggests that familiarity explained a substantial percentage of the variance in lexical processing of both simplified and traditional Chinese words, over and above the effects of word frequency and other lexical features, including age of acquisition (AoA). Further analysis identified a significantly greater familiarity effect for lower-frequency words than that for higher-frequency words. Together, among the first, our findings support the important contribution of familiarity with Chinese words to lexical processing, especially for low-frequency words.

Word Structure Tunes Electrophysiological and Hemodynamic Responses in the Frontal Cortex

To date, it is still unclear how word structure might impact lexical processing in the brain for languages with an impoverished system of grammatical morphology such as Chinese. In this study, concurrent electroencephalogram (EEG) and functional near-infrared spectroscopy (fNIRS) recordings were performed to inspect the temporal and spatial brain activities that are related to Chinese word structure (compound vs. derivation vs. non-morphological) effects. A masked priming paradigm was utilized on three lexical conditions (compound constitute priming, derivation constitute priming, and non-morphological priming) to tap Chinese native speakers' structural sensitivity to differing word structures. The compound vs. derivation structure effect was revealed by the behavioral data as well as the temporal and spatial brain activation patterns. In the masked priming task, Chinese derivations exhibited significantly enhanced brain activation in the frontal cortex and involved broader brain networks as compared with lexicalized compounds. The results were interpreted by the differing connection patterns between constitute morphemes within a given word structure from a spreading activation perspective. More importantly, we demonstrated that the Chinese word structure effect showed a distinct brain activation pattern from that of the dual-route mechanism in alphabetic languages. Therefore, this work paved a new avenue for comprehensively understanding the underlying cognitive neural mechanisms associated with Chinese derivations and coordinate compounds.

CCLOWW: A grade-level Chinese children's lexicon of written words

In this article, we present the Chinese Children's Lexicon of Written Words (CCLOWW), the first grade-level database that provides frequency statistics of simplified Chinese characters and words for children. The database computes from a corpus of 34,671,424 character tokens and 22,427,010 word tokens (including single- and multicharacter words), extracted from 2131 books. It contains 6746 different character types and 153,079 different word types. CCLOWW provides several frequency indices of simplified Chinese for three grade levels (grade 2 and below, grades 3-4, grades 5-6) to profile children's experience with written Chinese in and outside of school. We describe in this article the distributions of frequency and contextual diversity of the characters and words, as well as word length and syntactic categories of the words in the corpus and the subcorpora. We also report results of correlation analyses with other written corpora and of several naming and lexicon decision experiments. The findings suggest that CCLOWW frequency measures correlate well with other corpora. Importantly, they could reliably predict children's and adults' naming and lexical decision performances. They could also explain variance in adults' visual word recognition, in addition to frequency measures computed in an adult corpus, indicating that early print exposure might influence readers' lexical processing later on beyond an age of acquisition effect. CCLOWW will help researchers in language processing and development as well as educators with selecting language materials appropriate for children's developmental stages. The database is freely available online at https://www.learn2read.cn/database/ .

An ERP megastudy of Chinese word recognition

This article reports the first ERP (event-related potential) megastudy in traditional Chinese word recognition. Fifty-one native Chinese undergraduates in Hong Kong, who were native Cantonese speakers, provided ERP data to 1020 two-character words and 204 two-character pseudowords in a go/no-go lexical decision task (go trials: pseudowords). The item list and the ERP data were compiled into a database called "E-MELD" (ERP MEgastudy of Lexical Decision). To illustrate how E-MELD can be utilized in research of traditional Chinese word recognition, a series of linear mixed-effects (LME) models were conducted to examine how properties at word (contextual diversity, number of strokes, and concreteness) and character (contextual diversity, number of homophones, and semantic transparency ratings) levels influenced the ERP amplitudes in different time windows. The results showed that in all time windows, both word and character variables influenced the amplitudes of ERP signals, which argued against the proposal that Chinese two-character words are recognized holistically. At the same time, there was no evidence that character effects preceded word effects (i.e., no evidence of character-mediated word recognition). Overall, the pattern suggests that characters and words are accessed simultaneously in Chinese word recognition. E-MELD is made available online, such that interested researchers can download it and use the data innovatively for their research purpose.

Linking norms, ratings, and relations of words and concepts across multiple language varieties

Psychologists and linguists collect various data on word and concept properties. In psychology, scholars have accumulated norms and ratings for a large number of words in languages with many speakers. In linguistics, scholars have accumulated cross-linguistic information about the relations between words and concepts. Until now, however, there have been no efforts to combine information from the two fields, which would allow comparison of psychological and linguistic properties across different languages. The Database of Cross-Linguistic Norms, Ratings, and Relations for Words and Concepts (NoRaRe) is the first attempt to close this gap. Building on a reference catalog that offers standardization of concepts used in historical and typological language comparison, it integrates data from psychology and linguistics, collected from 98 data sets, covering 65 unique properties for 40 languages. The database is curated with the help of manual, automated, semi-automated workflows and uses a software API to control and access the data. The database is accessible via a web application, the software API, or using scripting languages. In this study, we present how the database is structured, how it can be extended, and how we control the quality of the data curation process. To illustrate its application, we present three case studies that test the validity of our approach, the accuracy of our workflows, and the integrative potential of the database. Due to regular version updates, the NoRaRe database has the potential to advance research in psychology and linguistics by offering researchers an integrated perspective on both fields.

Database of word-level statistics for Mandarin Chinese (DoWLS-MAN)

In this article we present the Database of Word-Level Statistics for Mandarin Chinese (DoWLS-MAN). The database addresses the lack of agreement in phonological syllable segmentation specific to Mandarin by offering phonological features for each lexical item according to 16 schematic representations of the syllable (8 with tone and 8 without tone). Those lexical statistics that differ per phonological word and nonword due to changes in syllable segmentation are of the variant category and include subtitle lexical frequency, phonological neighborhood density measures, homophone density, and network science measures. The invariant characteristics consist of each items' lexical tone, phonological transcription, and syllable structure among others. The goal of DoWLS-MAN is to provide researchers both the ability to choose stimuli that are derived from a segmentation schema that supports an existing model of Mandarin speech processing, and the ability to choose stimuli that allow for the testing of hypotheses on phonological segmentation according to multiple schemas. In an exploratory analysis we illustrate how multiple schematic representations of the phonological mental lexicon can aid in hypothesis generation, specifically in terms of phonological processing when reading Chinese orthography. Users of the database can search among over 92,000 words, over 1600 out-of-vocabulary Chinese characters, and 4300 phonological nonwords according to either Chinese orthography, pinyin, or ASCII phonetic script. Users can also generate a list of phonological words and nonwords according to user-defined ranges and categories of lexical characteristics. DoWLS-MAN is available to the public for search or download at https://dowls.site .

The Effects of Orthographic Neighborhood Size and the Influence of Individual Differences in Linguistic Skills During the Recognition of Chinese Words

There are still inconsistencies as to whether frequency and orthographic neighborhood size affect the reading and recognition of Chinese words. In addition, research on Chinese reading still adheres to the view that "all skilled readers read in the same way" and pays little attention to the influence of individual differences in linguistic skills on word recognition. In this research, we studied the recognition of Chinese two-character words in a lexical decision task (LDT) by manipulating neighborhood size and word frequency and controlling the frequency of the initial constituent character. Individual differences in linguistic skills were assessed through tests of spelling and reading comprehension. The results showed that: (1) A larger orthographic neighborhood size of the initial character had a facilitative effect on Chinese word recognition. The orthographic neighborhood size effect is modulated by word frequency, but this modulation effect was not stable. (2) Spelling and reading comprehension skills are good indicators to assess individual differences in Chinese linguistic skills, and they are significantly correlated. (3) Individual differences in linguistic skills influence the neighborhood size effect, which is moderated by word frequency.

Can we use the internet to study speech production? Yes we can! Evidence contrasting online versus laboratory naming latencies and errors

The closure of cognitive psychology labs around the world due to the COVID-19 pandemic has prevented in-person testing. This has caused a particular challenge for speech production researchers, as before the pandemic there were no studies demonstrating that reliable overt speech production data could be collected via the internet. Here, we present evidence that both accurate and reliable overt articulation data can be collected from internet-based speech production experiments. We tested 100 participants in a picture naming paradigm, where we manipulated the word and phonotactic frequency of the picture names. We compared our results to a lab-based study conducted on different participants which used the same materials and design. We found a significant word frequency effect but no phonotactic frequency effect, fully replicating the lab-based results. Effect sizes were similar between experiments, but with significantly longer latencies in the internet-collected data. We found no evidence that internet upload or download speed affected either naming latencies or errors. In addition, we carried out a permutation-style analysis which recommends a minimum sample size of 40 participants for online production paradigms. In sum, our study demonstrates that internet-based testing of speech production is a feasible and promising endeavour, with less challenges than many researchers (anecdotally) assumed.

Objective ages of acquisition for 3300+ simplified Chinese characters

We report the construction of two age-of-acquisition (AoA) norms for 3300+ characters in simplified Chinese, which make up about 99% of the texts used in daily life. We determined a character's AoA according to the time in which the character is formally learned in two sets of leading textbooks of Chinese in compulsory education, published respectively on the basis of the 2001 and 2011 national curriculum. Apart from having a significantly larger coverage of characters than previous norms, the current norms also outperformed them in explaining accuracy and reaction times in four large-scale databases for character decision, character naming, or character handwriting, even after controlling for the effects of frequency, number of meanings, and number of strokes. The explanatory advantage of the current norms suggests that, compared to earlier norms, the current norms capture more up-to-date character AoAs; these findings also highlight the diachronic nature of some lexical variables such as AoA and frequency. The developed objective AoA norms can be used for subsequent research on Chinese character recognition or production.

Age of acquisition ratings for 19,716 simplified Chinese words

This study collected age-of-acquisition (AoA) norms for 19,716 simplified Chinese words provided by 1765 native speakers of Mandarin Chinese. Analysis of demographic factors revealed both gender and cohort differences in ratings and thus differences in trajectory of vocabulary development, suggesting the utility of AoA ratings to investigate individual differences and societal changes in levels of academic achievement. Moreover, consistent with past analysis conducted with English words, AoA ratings accounted for an extra portion of variance in lexical processing above and beyond the most prominent predictor, word frequency. Further analysis suggested that AoA ratings might be more predictive of processing low frequency words relative to high frequency words. Cross-language comparison also indicated that future research should explore the unique factors implicated in the processing of Chinese words.

Valence and arousal ratings for 11,310 simplified Chinese words

This study reports valence and arousal ratings for 11,310 simplified Chinese words, including 9774 two-character words, 949 three-character words, and 587 four-character words. These affective ratings are validated through comparisons with prior ratings of smaller word samples. All but four words included in this study are from the MEgastudy of Lexical Decision in Simplified CHinese (MELD-SCH) database. As age-of-acquisition ratings and concreteness ratings have recently become available for large portions of words in the MELD-SCH, the affective ratings not only further enrich the database as a valuable research tool, but also allow us to gain insight into a range of psycholinguistic constructs based on normative ratings of a large set of Chinese words. Cross-language comparisons of the valence ratings between Chinese words and English words appear to indicate cultural and sociopolitical influences reflected in affect representations.

The Graded Priming Effect of Semantic Radical on Chinese Character Recognition

This study used priming paradigm with lexical decision task to examine the effects of different levels of semantic relatedness on the identification of Chinese phonetic-semantic compound characters. Unlike previous studies that simply classify Chinese compound characters as semantically transparent or opaque, we categorize the semantic relatedness between semantic radicals (i.e., prime) and the target characters containing them into five levels: highly related (i.e., high condition; e.g., prime ± vs. target ), moderately related (i.e., moderate condition; e.g., prime ± vs. target ), minimally related (i.e., minimal condition; e.g., prime ± vs. target ), unrelated but sharing the semantic radical (i.e., form-only condition; e.g., prime ± vs. target ), and unrelated without sharing the semantic radical (i.e., control condition; e.g., prime ± vs. target ). Moreover, three stimulus-onset asynchrony (SOA)s (i.e., 57, 140, and 243 ms) were used in this study to dissociate the radical- and character-level semantic priming effects. Results revealed a graded priming effect of the semantic radical on character recognition in Chinese readers for all SOAs. More specifically, the facilitative effect of the semantic radical on character processing was most evident for the high condition, followed by the minimal, form-only, and control conditions. This suggests a graded priming effect of the semantic radical on character identification.

Recognition times for 62 thousand English words: Data from the English Crowdsourcing Project

We present a new dataset of English word recognition times for a total of 62 thousand words, called the English Crowdsourcing Project. The data were collected via an internet vocabulary test in which more than one million people participated. The present dataset is limited to native English speakers. Participants were asked to indicate which words they knew. Their response times were registered, although at no point were the participants asked to respond as quickly as possible. Still, the response times correlate around .75 with the response times of the English Lexicon Project for the shared words. Also, the results of virtual experiments indicate that the new response times are a valid addition to the English Lexicon Project. This not only means that we have useful response times for some 35 thousand extra words, but we now also have data on differences in response latencies as a function of education and age.

Age of acquisition effects on traditional Chinese character naming and lexical decision

Across languages, age of acquisition (AoA) is a critical psycholinguistic factor in lexical processing, reflecting the influence of learning experience. Early-acquired words tend to be processed more quickly and accurately than late-acquired words. Recently, an integrated view proposed that both the mappings between representations and the construction of semantic representations contribute to AoA effects, thus, predicting larger AoA effects for words with arbitrary mappings between representations as well as for tasks requiring greater semantic processing. We investigated how these predictions generalize to the Chinese language system that differs from alphabetic languages regarding the ease of mappings and semantic involvement in lexical processing. A cross-task investigation of differential psycholinguistic effects was conducted with large character naming and lexical decision datasets to establish the extent to which semantics is involved in the two tasks. We focused on examining the effect sizes of lexical-semantic variables and AoA, and the interaction between AoA and consistency. The results demonstrated that semantics influenced Chinese character naming more than lexical decision, which is in contrast with the findings related to English language, though, critically, AoA effects were more pronounced for character naming than for lexical decision. Additionally, an interaction between AoA and consistency was found in character naming. Our findings provide cross-linguistic evidence supporting the view of multiple origins of AoA effects in the language-processing system.

Compound word frequency modifies the effect of character frequency in reading Chinese

In two eye-tracking studies, reading of two-character Chinese compound words was examined. First and second character frequency were orthogonally manipulated to examine the extent to which Chinese compound words are processed via the component characters. In Experiment 1, first and second character frequency were manipulated for frequent compound words, whereas in Experiment 2 it was done for infrequent compound words. Fixation time and skipping probability for the first and second character were affected by its frequency in neither experiment, nor in their pooled analysis. Yet, in Experiment 2 fixations on the second character were longer when a high-frequency character was presented as the first character compared with when a low-frequency character was presented as the first character. This reversed character frequency effect reflects a morphological family size effect and is explained by the constraint hypothesis, according to which fixation time on the second component of two-component compound words is shorter when its identity is constrained by the first component. It is concluded that frequent Chinese compound words are processed holistically, whereas with infrequent compound words there is some room for the characters to play a role in the identification process.

Morphosemantic activation of opaque Chinese words in sentence comprehension

Two cross-modal priming experiments were conducted to investigate morphological processing in Chinese spoken word recognition during sentence comprehension. Participants heard sentences that contained opaque prime words and performed lexical decisions on visual targets that were related to second morpheme meanings of opaque words or whole-word meanings. The targets were presented at the auditory onset of the second morphemes or the subsequent syllables after the opaque primes to examine the time course of effects. In a neutral sentence context (Experiment 1), opaque word morpheme meanings produced morphological priming on target word recognition, which preceded lexical priming. When context biased toward whole opaque words (Experiment 2), morphological priming disappeared, while the effect of lexical meanings remained significant and emerged earlier than the effect of lexical meanings in the neutral context. These findings suggest that morphemes play a role in Chinese spoken word recognition, but their effects depend on the prior context during sentence comprehension.

Concreteness/abstractness ratings for two-character Chinese words in MELD-SCH

The concreteness-abstractness continuum is considered a primary dimension in the representation of semantic networks. Its theoretical importance and clinical significance are widely acknowledged. To assist and enhance future research, this study collected and evaluated concreteness/abstractness ratings for 9,877 two-character Chinese words retrieved from the MEga study of Lexical Decision in Simplified CHinese (MELD-SCH, Tsang et al, 2018). The ratings were validated through comparisons with previous rating studies on concreteness and imageability of smaller word samples. Relations of word concreteness with word frequency, age-of-acquisition, and efficiency of lexical processing were also examined. These ratings provide an additional dimension of information to two-character words in the database MELD-SCH, permitting not only more comprehensive research on the Chinese language, but also cross-language investigation of the concreteness effect between Chinese and other languages such as English and Dutch where a large database of concreteness ratings is also available.

Individual differences in Cantonese Chinese word recognition: Insights from the Chinese Lexicon Project

The Chinese Lexicon Project is a repository of lexical decision data for 25,286 Cantonese Chinese two-character compound words. To create that repository, 594 participants responded to approximately 1,404 words and 1,404 nonwords over three sessions. Using the data in this repository, the present study examines the variability and reliability of Chinese lexical decision performance, along with the moderating influence of individual differences on lexical processing. We generally found high to very high within- and between-session reliabilities for mean response times, ex-Gaussian parameters, accuracy rates, and a composite proficiency measure tapping lexical processing fluency. Using linear mixed effects models, we also found reliable interactions between fluency and two lexical effects. Specifically, more fluent readers showed larger effects of word frequency and semantic transparency. These results attest to the stability of Chinese word recognition performance, and are most consistent with a flexible lexical processing system that adapts optimally to task demands.

Word prevalence norms for 62,000 English lemmas

We present word prevalence data for 61,858 English words. Word prevalence refers to the number of people who know the word. The measure was obtained on the basis of an online crowdsourcing study involving over 220,000 people. Word prevalence data are useful for gauging the difficulty of words and, as such, for matching stimulus materials in experimental conditions or selecting stimulus materials for vocabulary tests. Word prevalence also predicts word processing times, over and above the effects of word frequency, word length, similarity to other words, and age of acquisition, in line with previous findings in the Dutch language.

Chinese lexical database (CLD) : A large-scale lexical database for simplified Mandarin Chinese

We present the Chinese Lexical Database (CLD): a large-scale lexical database for simplified Chinese. The CLD provides a wealth of lexical information for 3913 one-character words, 34,233 two-character words, 7143 three-character words, and 3355 four-character words, and is publicly available through http://www.chineselexicaldatabase.com . For each of the 48,644 words in the CLD, we provide a wide range of categorical predictors, as well as an extensive set of frequency measures, complexity measures, neighborhood density measures, orthography-phonology consistency measures, and information-theoretic measures. We evaluate the explanatory power of the lexical variables in the CLD in the context of experimental data through analyses of lexical decision latencies for one-character, two-character, three-character and four-character words, as well as word naming latencies for one-character and two-character words. The results of these analyses are discussed.

Syllable retrieval precedes sub-syllabic encoding in Cantonese spoken word production

Two experiments were conducted to investigate the time course of syllabic and sub-syllabic processing in Cantonese spoken word production by using the picture-word interference task. Cantonese-speaking participants were asked to name individually presented pictures aloud and ignore an auditory word distractor. The targets and distractors were either phonologically related (i.e., sharing two identical word-initial phonemes) or unrelated. In Experiment 1, the target syllables were all consonant-vowel (CV)-structured. The phonological distractor was either a CV syllable (i.e., Full Syllable Overlap) or a CVC (consonant-vowel-consonant) syllable (i.e., Sub-syllable Overlap). Relative to the unrelated control, Full Syllable Overlap distractors facilitated naming in all stimuli onset asynchronies (SOAs) (-175, 0, or +175 ms) whereas Sub-syllable Overlap distractors exhibited facilitation only at 0-ms and +175-ms SOAs. Experiment 2 adopted a similar design to examine the possible influence of syllabic structure similarity on the results of Experiment 1. The target syllables were all CVC-structured. The phonological distractor was either a CVC (i.e., Syllable-structure Consistent) or CV (i.e., Syllable-structure Inconsistent) syllable. Comparable priming was observed between the two distractor conditions across the three SOAs. These results indicated that an earlier priming effect was observed with full syllable overlap than sub-syllabic overlap when the degree of segmental overlap was held constant (Experiment 1). The earlier syllable priming observed in Experiment 1 could not be attributed to the effect of syllabic-structure (Experiment 2), thereby suggesting that the syllable unit is important in Cantonese and is retrieved earlier than sub-syllabic components during phonological encoding.