Scholarlink

<p>This study presents a corpus-internal validation of the Academic Spoken Collocation List (ASCL), originally developed from the British Academic Spoken English corpus (BASE) through a mixed-method approach combining corpus-driven procedures and expert judgment. Although the ASCL was constructed using BASE, its internal distributional stability within the source corpus has not been systematically examined, a necessary step before extending validation to independent corpora. The present study therefore examines the 416 ASCL entries within BASE in terms of frequency, coverage, dispersion, and distributional stability across disciplines and genres, with correspondence analysis employed as a complementary exploratory method to visualize patterns across the four disciplinary subcorpora and the two spoken academic genres. The results indicate that the ASCL entries achieve substantial coverage and show generally balanced dispersion across disciplines. Although some entries occur more frequently in particular disciplines and overall coverage is marginally higher in Arts and Humanities than in the other disciplines, the degree of disciplinary variation is limited and of very small practical magnitude. The few cases of highly uneven dispersion appear to arise from two distinct sources: concentration in particular disciplinary contexts and instability resulting from very low overall frequency. Correspondence analysis further suggests that the ASCL represents a common core of spoken academic collocations within BASE, with only a small number of discipline-associated distributional tendencies; genre-based differences between lectures and seminars are similarly constrained. Taken together, these findings indicate that the ASCL demonstrates generally consistent internal distributional stability within its source corpus, lending preliminary support to its potential use as a general resource for academic spoken English learning and teaching, and laying the groundwork for future cross-corpus validation. By combining correspondence analysis with established coverage and dispersion measures, this study enhances the methodological rigor of vocabulary list validation through a complementary multivariate approach to examining disciplinary and genre-based distributional patterns.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Academic spoken collocations across disciplines and genres: a corpus-based validation study

  • 期刊论文

  • 发表日期：

    2026年7月31日

查看全文
  • Yawen Zhong

摘要

This study presents a corpus-internal validation of the Academic Spoken Collocation List (ASCL), originally developed from the British Academic Spoken English corpus (BASE) through a mixed-method approach combining corpus-driven procedures and expert judgment. Although the ASCL was constructed using BASE, its internal distributional stability within the source corpus has not been systematically examined, a necessary step before extending validation to independent corpora. The present study therefore examines the 416 ASCL entries within BASE in terms of frequency, coverage, dispersion, and distributional stability across disciplines and genres, with correspondence analysis employed as a complementary exploratory method to visualize patterns across the four disciplinary subcorpora and the two spoken academic genres. The results indicate that the ASCL entries achieve substantial coverage and show generally balanced dispersion across disciplines. Although some entries occur more frequently in particular disciplines and overall coverage is marginally higher in Arts and Humanities than in the other disciplines, the degree of disciplinary variation is limited and of very small practical magnitude. The few cases of highly uneven dispersion appear to arise from two distinct sources: concentration in particular disciplinary contexts and instability resulting from very low overall frequency. Correspondence analysis further suggests that the ASCL represents a common core of spoken academic collocations within BASE, with only a small number of discipline-associated distributional tendencies; genre-based differences between lectures and seminars are similarly constrained. Taken together, these findings indicate that the ASCL demonstrates generally consistent internal distributional stability within its source corpus, lending preliminary support to its potential use as a general resource for academic spoken English learning and teaching, and laying the groundwork for future cross-corpus validation. By combining correspondence analysis with established coverage and dispersion measures, this study enhances the methodological rigor of vocabulary list validation through a complementary multivariate approach to examining disciplinary and genre-based distributional patterns.