One possibility is the following. Let us suppose a situation in Fig. 1, in which S1 is the field Computing, and that it has four facets (and no subfacets). Let each of these facets include a term “Null” (never shown to the user), and that this is the initial default value in every classification assignment. When a classifier or searcher selects another term in a facet, this replaces Null. If a searcher only enters Computing, the faceted search query is actually Computing-Null-Null-Null-Null. This only matches with document representations that have been similarly classified, i.e. simply as Computing, the facets retaining their Null defaults. The system could then report two counts of retrievable document sets: (1) those classified as Computing-Null-Null-Null-Null and (2) those classified as Computing, irrespective of the term assignments in the facets. Only the first set would be displayed to the searcher. Does this sound feasible?