Development of a WoridWideWeb System to Help Researchers Write Scientific Papers in English

这个研究项目的目的是开发一个新颖的万维网系统，以帮助日本研究人员撰写英文研究论文。为了实现这一目标，我们首先编制了一个英语语料库，其中包括主要与生命科学有关的科学论文，主要由英语母语人士撰写。该语料库现在包含超过3000万个运行词。对语料库的统计分析使我们能够获得哪些特定单词或短语频繁出现以及哪些单词通常与给定单词搭配的信息。日本科学家长期以来一直希望获得这类信息，以避免偏离母语人士在写作英语时所遵循的语言惯例。在我们的项目中开发的系统为用户提供了有用的信息，几乎每个单词的一致性和搭配，只需点击鼠标按钮，只要检索到的单词包含在语料库中。大多数日本研究人员在使用英语时缺乏像母语人士那样的直觉，这导致了诸如“高可能性”或“可能”之类的尴尬表达。然而，现在这个系统在互联网上广泛地向公众开放，他们可以在需要确认特定英语单词组合的自然性时参考它，而不是简单地依靠自己的非母语直觉。这将有助于减轻非英语母语人士（如日本研究人员）在用英语撰写研究论文时的负担。

The aim of this research project was to develop a novel World_Wide_Web system to assist Japanese researchers in writing English research papers. To accomplish this, we first compiled an English corpus consisting of scientific papers relating mainly to life sciences and principally written by native speakers of English. The corpus now contains more than 30,000,000 running words.A statistical analysis of the corpus enabled us to obtain information on which particular words or phrases occur frequently and which words commonly collocate with a given word. Japanese scientists have long wanted to have access to this kind of information so as to avoid deviating from the linguistic conventions observed by native speakers in writing English. The system developed in our project provides users with useful information on concordances and collocations for almost each and every word with just a click of a mouse button, as long as the word retrieved is contained in the corpus.Most Japanese researchers lack native-like intuition in using English, which has resulted in awkward expressions such as 'high possibility' or 'may probably.' However, now that the system is widely available to the public on the Internet, they can refer to it whenever they need to confirm the naturalness of a particular combination of English words instead of simply relying on their own non-native-like intuitions. This should thereby help to lessen the burden on non-native speakers of English (such as Japanese researchers) when writing research papers in English.