Following this interpretative tradition, an indexer will ask questions such as, “Who will use this document?”, “To what domain do the users of the document belong?”, and “Who wrote this text?” to determine the subject matter of a document. Only if the indexer were instructed to, say, derive nouns from the title––or perform some similarly mechanical sort of indexing––would it be possible for the indexer to base indexing on the document attributes alone, without any interpretation or judgment about the potential use of the document. Any indexing that allows the indexer to make choices will force the indexer to make an interpretation of and judgment about the use of the document.
In sum, the indexer cannot determine and represent the subject matter of a document without some understanding of the future use of the document. Furthermore, the indexer cannot understand potential usage of the document without an understanding of the context in which the document is used. The determination and representation of the subject matter of documents is connected to discourse and activities in a context, and the indexer needs to have an understanding of this discourse and these activities.