Linguistic Corpus Search

Searching corpora with linguistic questions requires both additional information encoded in the corpus and efficiency as in “traditional” search engines. We describe a search engine-like approach to querying plain as well as part-of-speech-tagged monolingual corpora. This approach makes use of a ‘minimalist’ query language which nevertheless allows powerful searches by optionally ignoring positional as well as inflectional features in the corpus sentences. Many queries can be formulated without detailed training via a simple web-based front-end. Relevant applications of this search tool in knowledge extraction are discussed as well.