In this paper, we propose a model/method for observing the potential lexical productivity of head elements in nominal compounds, and compare the productivity of a few high-frequency head elements in a technical corpus. Much work has been done on various aspects of nominal compounds. Most of this work, however, has been devoted to “syntagmatic” aspects at various levels, such as semantic compositionality, possible variations, lexical cohesiveness, etc. while the “paradigmatic” aspects of nominal compounds have received relatively little attention. By providing a “paradigmatic” perspective for analysing nominal compounds, this paper aims to help build a truly integrative approach to analysing and processing nominal compounds.