The paper discusses the interpretation of data that users of the Gigafida corpus obtain when trying to solve their language dilemmas through Gigafida%s web concordancer. Illustrated by four examples, we emphasize the perils of making hasty generalizations, especially when dealing with frequency information and data presented in the Text type, Source and Year filters (within basic search), as well the need for understanding the characteristics and accuracy of automated corpus annotation (within advanced search). In the discussion, we emphasize that the interpretation of query results should be made by taking the corpus content and structure into account, regardless of the search query or search type.
|
