It is a truism in information science that indexing is associated with ‘aboutness’, and that index terms that accurately represent what a document is about will serve the needs of the user/searcher well. It is contended in this paper that indexing which is limited to the representation of aboutncss serves the novice in a discipline adequately, but docs not serve the scholar or researcher, who is concerned with highly specific aspects of or points-of-view on a subject. The linguistic analogs of ‘aboutness’ and ‘aspect’ are topic and comment respectively. Serial indexing services deal with topics at varying levels of specificity, but neglect comment almost entirely. This may explain the underutilization of secondary information services by scholars, as has been repeatedly demonstrated in user studies. It may also account for the incomplete lists of bibliographic references in many research papers. Natural language searching of fulltext databases docs not solve this problem, because the aspect of a topic of interest to researchers is often inexpressible in concrete terms. The thesis is illustrated with examples of indexing failure in research projects the author has conducted on a range of linguistic and library-information science topics. Finally, the question of whether indexing can be improved to meet the needs of researchers is examined.