大数据视域下的汉英词典兼类表征策略研究——以《汉英词典》(第三版)为例

A Big Data Perspective on the Representation Strategy of Heterosemy in A Chinese-English Dictionary(3rd Edition)

词类标注在汉英词典编纂中可谓牵一发而动全身,其中兼类表征尤甚。《汉英词典》(第三版)最大修订莫过于词类标注及其对微观结构所做的相应调整,此举在《汉英词典》编纂历史上具有重要的里程碑意义。文章以双层词类范畴化理论为指导,基于自建的"《汉英词典》(第三版)词类标注数据库",对该词典兼类词表征策略进行深入研究。研究发现,该词典在兼类表征问题上得失并存:与《现代汉语词典》(第5版)相比,其表征的汉语兼类概括词数量大增,微观结构亦更趋合理,但仍未能全面反映现代汉语兼类现状。研究表明,该词典仍然受制于"简约原则"的影响,而"简约原则"是唯科学主义在语言学中的映射,用来指导抽象语言系统中概括词的兼类表征是一种僭越;概括词的词类属性是社群语言系统词库中的涌现特征,需要基于语言大数据的使用模式调查方能准确发现。

Word class labeling is a major concern in Chinese-English lexicography,in which heterosemy is a particularly key issue.The most noteworthy revision in the third edition of A Chinese-English Dutionary(henceforth called CED3)is word class labeling together with the necessary adjustments to its micro structure,which is a milestone in the history of CED.From the perspective of the Two-level Word Class C ategorization Theory,this paper is intended to study the representation strategy of heterosemy in CED3 based on the DIY D atabase of the Word C lass Labeling Information of the dictionary.It is found that both achievements and problems exist:In terms of achievements,the number of heterosemous lexemes in CED3 has greatly increased together with some improvements in its microstructure.Compared with the fifth edition of the Contemporary Chinese Dictionary,the number of heterosemo us lexemes in CED3 has increased by 49.72%.The reason why the number of heterosemo us lexemes greatly increased in CED3 is that,in addition to the fact that the lexicographers are more receptive to heterosemous lexemes in C hinese,taking English as the reference language helps to reveal similarities of word classes in C hinese and English as well as particularities in Chinese word classes.When it comes to the problems,there are some confusions in the micro structure of CED 3,which highlights the fact that the lexicographers have different criteria for word class identification,and the representation of C hinese heterosemous lexemes is still incomplete,which fails to reflect fully the status quo of heterosemy in Modern Chinese.T aking the disyllabic entries of X-bian in CED3 for example,some noteworthy problems remain:(1)Some lowfrequency multifunctional lexemes like jibian and shanbian are labeled as multi-category ones,whereas some higher frequency multifunctional lexemes like gaibian and zhuanbian are represented as single category ones;(2)The translation of some lexemes or their senses violates the Principle of Translating by Word Class in bilingual lexicography,and even worse there are some mixed translations;(3)For those examples in which the lexemes are used as subjects or objects,some are regarded as noun usages,whereas some others as verb usages.It is shown that,to a large extent,the bewildering word class labeling in CED3 mirrors the dilemma in the word class study of Modern Chinese:Scholars who advocate the Principle of Parsimony on the heterosemy issue adhere to the scientistic research paradigm,and categorize second-order entities like lexemes in the communal language as first-order natural products,and do not distinguish the different categorization processes of word types and word tokens.Although the overall compilation of CED3 is excellent,it is still constrained by the Principle of Parsimony,which mirrors scientism in linguistics and which is an invalid guiding principle for heterosemy representation of lexemes in the abstract communal language system.In fact,the class membership of lexemes are emerging properties in the communal language system,which can only be discovered by Big Data usage-based surveys.In view of this,the Two-level Word Class Categorization Theory has both epistemological value and methodological significance.