On the Use of Citations in Studying Scientific Achievements and Communication

Abstract

Rather than trying to reply in kind to some recent, slightly polemical, criticisms of the use of citations, we will discuss the assumptions underlying the use of citations in the study of science. We shall attempt to explicate the principles underlying our work, with certain technical problems, and end with a brief panegyric on research programs and approaches to study of science and scientific communication. Any value of citations and the restrictions on such value stems from the combined operation of several assumptions, three massive qualifications, and one critically important conjecture. The operation of these principles d~(fers widely from field to field and from application to application. In discussing these principles we would note that the assumptions are very weak ones, and that their power derives from the stochastic nature of the }vorld, laying the groundwork for true quantification only with very large tiles. For example, we can see a possible restriction in mathematics, where the literature base is small and each article contains few references. In such a field citations must be far more robust than in molecular biology, where the reverse holds on both dimensions. Let’s look now at four assumptions: 1. A document x cited by document y is more likely to be judged as related in content to document y than one not cited. If we were to eponymize, this might be called the Garfield2-Kessler3 Assumption, since it underlies both the remarkable Science Citation Index” service and the benchmark research at MIT. it’s a hard assumption to gainsay.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 14 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.