Cross-script Masked Translation Priming With Chinese-English-Japanese Trilingual Cognates

With the development of globalization and multilingual education,trilingualism is becoming a common social phenomenon.Trilinguals need to build a complex language system in their brains so that they can not only use three languages independently,but can also communicate with people across languages,which means that the three language systems need to interact to some extent in cross-language processing.This then leads to some questions:How do the three language systems organize lexicon and semantics?And in what ways do the lexical representations relate to their underlying linguistic information?Using interlingual cognates,these questions can be investigated.Since interlingual cognates have overlaps in phonology,semantics and/or orthography(i.e.,spelling),they provide a direct way to reveal the cross-linguistic influence of trilingual representations.In the present literature,researches on the processing of trilingual cognates mostly take place in North America and Europe,involving mainly alphabetic languages,such as,Dutch,English,German,French,Polish and Spanish,etc.However,the most obvious feature of interlingual cognates in alphabetic languages is their high similarity in orthography,which makes the representation and processing models of trilingualism more applicable to the same-script languages.Currently,very little research exists on the processing of trilingual cognates in different-script languages.Therefore,it is of great importance to study trilingual cognates of this kind to address this gap in the literature.For this reason,the present study aimed to examine masked translation priming of trilingual cognates with three different cross-script languages.The participants in this study were native Chinese speakers,who acquired English as a second language and Japanese a third.In terms of the order of language acquisition,these participants were late unbalanced Chinese-English-Japanese trilinguals.Using Chinese-English-Japanese trilingual cognates to test masked translation priming effects across languages,this study would further reveal the mechanism underlying word processing of cognates in the three different scripts.The stimuli in this study bore no visual similarities(e.g.,咖啡-Coffee-コーヒー),but they overlapped both phonologically and semantically.Chinese and Japanese cognates in this study were borrowed from English(i.e.,loanwords),and they were translation equivalents.The three languages have their own characteristics:Chinese is a logographic language,belonging to the Sino-Tibetan language family;English is a syllabic language,belonging to the Indo-European language family;and Japanese katakana is a moraic language,also belonging to the Indo-European language family.The three languages have different spelling-sound mappings,and these distinctions can provide new perspectives in understanding trilingual cognate processing.The research questions are as follows:(1)What priming effects do Chinese-English-Japanese trilingual cognates have in Chinese-English cross-script masked translation priming,and does English proficiency affect the processing of Chinese-English cognates?(2)What priming effects do Chinese-English-Japanese trilingual cognates have in Chinese-Japanese cross-script masked translation priming?(3)What priming effects do Chinese-English-Japanese trilingual cognates have in English-Japanese cross-script masked translation priming?(4)What overall patterns do Chinese-English-Japanese trilingual cognates have in cross-script masked translation priming?In eight masked translation priming and lexical decision tasks,the cross-script masked translation priming effects from six priming directions were investigated,i.e.,where Chinese primed English,English primed Chinese,Chinese primed Japanese,Japanese primed Chinese,English primed Japanese,and Japanese primed English,respectively.The priming effects of participants with high and low English proficiency were also compared in the Chinese-English experiments in order to understand the role of second language proficiency.After the data were analyzed,the findings of the study are summarized as follows:First,the cross-script masked translation priming effects of Chinese-English cognates are asymmetrical.Chinese cognates significantly facilitated the recognition of English cognates,but not vice versa.English proficiency did not play a role in the ChineseEnglish cognate masked translation priming effects.Second,the cross-script masked translation priming effects of Chinese-Japanese cognates are asymmetrical.Chinese cognates significantly facilitated the recognition of Japanese cognates,but not vice versa.Third,the cross-script masked translation priming effects of English-Japanese cognates are asymmetrical.English cognates significantly facilitated the recognition of Japanese cognates,but not vice versa.Overall,the findings from the cross-script masked translation priming experiments with Chinese-English-Japanese trilingual cognates exhibit a number of patterns.Firstly,there are no bidirectional cross-script priming effects across Chinese-English-Japanese trilingual cognates;rather,there is significant cross-script masked translation priming asymmetry.The data from eight experiments showed that dominant languages facilitated the processing of weak languages,but weak languages did not facilitate the processing of dominant languages.Specifically,Chinese facilitated the recognition of English(Exp.1a and 2a)and Japanese cognates(Exp.3a),while English(Exp.1b and 2b)and Japanese(Exp.3b)failed to facilitate the recognition of Chinese cognates.English accelerated the recognition of Japanese cognates(Exp.4a),while Japanese did not affect the recognition of English cognates(Exp.4b).Moreover,in the cross-script priming experiments of Chinese and English,English proficiency did not play a significant role(Exp.1 and Exp.2).Secondly,the role of phonological similarity in the cross-script masked translation priming of ChineseEnglish-Japanese trilingual cognates is limited.The results of the experiments showed that when the weak language primed the dominant language,phonological similarity failed to exert an influence.Even when the phonological distance of two languages are close,for example,when they belong to an alphabetic language(e.g.English)and a moraic language(e.g.Katakana),phonological similarity does not result in a priming advantage in the reverse translation priming direction.Thirdly,the pattern of masked translation priming effects mirrors the pattern of language dominance in participants.While the mother tongue(i.e.,Chinese)showed significant priming effects on two foreign languages(i.e.,English and Japanese),the foreign languages(i.e.,English and Japanese)failed to facilitate the processing of the mother tongue.Furthermore,in comparing the two foreign languages,English was acquired earlier and had been studied over a longer period of time than Japanese for participants.English cognate translation primes could facilitate the recognition of Japanese targets,whereas Japanese primes did not aid in the recognition of English targets.As further evidence,participants completed a self-assessment of language ability,the scores from which could better predict the pattern of masked translation priming effect of cognates when compared with objective language proficiency tests.Fourth,the decoding and recognition processes of cognates are influenced by script characteristics.Orthographic features,especially orthographic-phonological mappings,play an important role and seem to regulate the cognate priming effects.Fifth,later-learned languages exhibit a dependent/parasitic relationship with more dominant languages among Chinese-EnglishJapanese trilingual cognates.With regard to the masking translation priming effects,the medium(i.e.study)language of foreign language learning is more important than phonological similarity as an influencing factor.Based on the above research findings,a parasitic model of the mental representation of Chinese-English-Japanese trilingual cognates is proposed in order to answer the fourth research question.In this model,the semantic links from host(i.e.,dominant)languages to parasite(i.e.,weak)languages are strongest.According to the model,host languages can produce masked translation priming effects,whereas masked translation priming effects from parasitic languages to host languages would be more unlikely.In addition,the priming effect of cognates are influenced not only by the orthographic-phonological mappings of the foreign language,but also by the phonological decoding of the native language.However,with the improvement of relative language proficiency,the interaction between phonological and semantic similarity will become greater,eventually strengthening the links between host and parasitic languages and increasing the probability of weak languages priming dominant languages.The present study contributes to the literature in the following ways:To begin with,the present study is original in that it is the first of its kind in the literature,both in China and abroad,expanding the research of trilingual cognates to three different-script languages,which introduces a new perspective from which the mental representation and processing of trilingual lexicon can be studied and understood.Moreover,this study systematically investigated cross-script activation and processing in six priming directions;that is,not only investigating priming effects from dominant languages to weak languages,but also investigating priming effects from weak languages to dominant languages.Additionally,priming effects between two foreign languages were also investigated and given special attention.Most of trilingual lexical processing studies in China adopt the long-term repetition priming paradigm,but this study is the first one to use the masked translation priming paradigm and the lexical decision task when investigating trilingual lexicon.It has improved the limitation of traditional research methods and enriched the current research methods.Secondly,the present study is innovative in the employed research method.Most trilingual lexical processing studies in China adopt the long-term repetition priming paradigm,but this study is the first to use the masked translation priming paradigm and the lexical decision task in investigating Chinese-English-Japanese trilingual lexicon.It has improved the limitation of traditional research methods and enriched the current research methods.Thirdly,the present study is also advanced in the employed data analysis method.While most domestic studies focus on variance analysis,this study makes use of a new data analysis method,the Generalized Linear Mixed Effects model in R,which has been widely recognized in recent years in international psycholinguistics research.Finally,this study contribute theoretically to the literature by proposing a parasitic model of the mental representation of different-script trilingual cognates,which outlines the potential mechanism underlying the priming asymmetry across Chinese-English-Japanese trilingual cognates,and has discussed the roles of phonological similarity,language dominance and second language proficiency,and also have illustrated the cross-script influence of phonological and semantic interaction across Chinese,English and Japanese in masked translation priming.The theory proposed in this study will improve overall understanding of the processing of trilingual cognates with different-script languages and provide a valuable analytical and interpretive framework for research on the trilingual lexicon.