Linguistics and Information Science: a Postscript

The object of Linguistics and Information Science (Sparck Jones and Kay, 1973) was to show how far the suppposedly natural connection between linguistics and information science existed in practice. We surveyed linguistic theory and computational linguistics to identify approaches potentially applicable to information science, and to information, i.e., document, retrieval in particular; and we investigated the linguistic operations of automatic document retrieval to establish their linguistic sophistication and the extent to which linguistic theories were being, or could be, applied. We also looked for evidence of feedback from automatic information retrieval to linguistics. Our general conclusion was that there was very little actual connection between linguistics and information retrieval. Linguists were preoccupied by concerns rather remote from any practical activity like information retrieval, for example the properties of linguistic theories, and had failed to provide tools of potential utility to retrieval workers. At the same time, in both practice and research in information retrieval, needs which might be met by linguistic theory were not properly specifi-d. In general, the linguistic procedures of automatic information retrieval were found to be very simple, and it was not obvious how useful refined linguistic tools would be, either as aids to automation, or as devices for improving retrieval performance.