Abstract

Knowledge of collocations is essential for English academic writing. However, there are few academic collocation referencing tools available and there is a pressing need to develop more. In this paper, we will introduce the ACOP (Academic Collocations and Phrases Search Engine), a newly developed corpus-based tool to search large academic corpora. Using the AntCorGen corpus crawling tool, 100,000 articles from 10 academic areas were downloaded from the PLOS One journal database. A multidisciplinary 500-million-word academic corpus was thus created. Our team further developed the ACOP to allow users to search for collocations and phrases in the academic corpus that we compiled. Additionally, this new tool will allow learners to search for not only bigrams but also for three- to five-word strings. Learners can further search using parts of speech. To validate whether the new tool can assist EFL learners in finding appropriate academic collocations, we compared this tool with BYU’s state-of-the-art COCA corpus search tool. The 35 students who participated in the study were randomly divided into two groups. One group used BYU’s COCA tool and the other used the ACOP tool. Both groups were asked to identify suitable collocates for 25 gap-fill questions. Paired t-tests and independent t-tests were used to analyze the data. The results indicated that both groups showed a significant improvement in the posttest when they had access to the collocation retrieval systems. However, statistically, the results showed no significant difference between the two tools, as they were equally useful in helping the students find appropriate collocates. According to the users’ surveys, compared with the COCA group, the participants in the ACOP group found the ACOP easier to use and the interface was clearer and simpler. The Mann-Whitney U tests of the two groups showed significant differences in the aspect of the interface. The preliminary results indicated that the ACOP can assist EFL learners in finding suitable academic collocations and phrases, and the users’ surveys showed positive results.

摘要

搭配知识对于英语学术写作至关重要。然而，可用的学术搭配参考工具很少，迫切需要开发更多。在本文中，我们将介绍 ACOP（学术搭配和短语搜索引擎），这是一种新开发的基于语料库的工具，用于搜索大型学术语料库。使用 AntCorGen 语料库爬取工具，从PLOS One下载了 10 个学术领域的 100,000 篇文章期刊数据库。一个多学科的 5 亿词学术语料库由此创建。我们的团队进一步开发了 ACOP，允许用户在我们编制的学术语料库中搜索搭配和短语。此外，这个新工具将使学习者不仅可以搜索二元字母，还可以搜索三到五个单词的字符串。学习者可以使用词性进一步搜索。为了验证新工具是否可以帮助 EFL 学习者找到合适的学术搭配，我们将此工具与 BYU 最先进的 COCA 语料库搜索工具进行了比较。参与研究的 35 名学生被随机分为两组。一组使用 BYU 的 COCA 工具，另一组使用 ACOP 工具。两组都被要求为 25 个填空问题确定合适的搭配。配对t-测试和独立t-测试用于分析数据。结果表明，两组在访问搭配检索系统时在后测中均表现出显着改善。然而，从统计上看，结果显示这两种工具之间没有显着差异，因为它们在帮助学生找到合适的搭配方面同样有用。根据用户调查，与COCA组相比，ACOP组的参与者认为ACOP更易于使用，界面更清晰、更简单。两组的 Mann-Whitney U 检验显示界面方面存在显着差异。初步结果表明，ACOP 可以帮助 EFL 学习者找到合适的学术搭配和短语，用户调查显示了积极的结果。