In this contribution relations between scientific articles, conference proceedings and patents relating to telephone switching are analyzed. The state-of-the-art in scientometrics and science and technology studies leads one to expect relations among these documents. Empirical findings show the opposite. To interpret these findings I focus on two key issues in scientometrics, namely the frequency of linkages among documents and the nature of communication. The resulting conclusion is that scientometrics should be informed by theories concerning the evolutionary dynamics of science and technology.