All of these documents (books, pages, entities) are indexed by terms, the features of documents that indicate their relevance to particular queries. In addition to basic terms such as the words that occur in a document, we index lemmas, standardized versions of those terms. While English has a relatively small amount of inflectional morphology—adding "s" for plurals and "ed" for preterites—searching languages such as Latin is much more effective when a query for "ars" (art) retrieves documents with all of its various forms, such as ars, artis, arti, artem, artes, etc. We also index pairs of terms in syntactic dependencies, to enable linguists and others to search for particular constructions.

During retrieval, the user can combine these index terms into complex queries: for instance, searching for fixed phrases such as "base ball" or terms within some proximity to each other such as "public library".