There is growing interest in the use of semantic collections in order to identify and analyse domain knowledge. This paper describes some technical issues to consider when contemplating research which incorporates small-to-medium domain-specific word sets. The purpose of the corpus construction described was to provide an external word collection which could be transformed to a numeric frequency scale which could take the place of an “expert ” in order to evaluate the lexical content of aircraft Visual Landing Approach concept maps. Although this paper is based on research in the field of aviation education, the underlying principles are more widely applicable. General Corpora Definitions and Uses The study of naturally occurring word frequencies has been a focus of computational linguistics, and in particular of the field of corpus linguistics. A word collection known as a corpus is constructed from some set of texts in order to determine what is characteristic of that text set through the identification of vocabulary patterns that either differ or conform to a norm (Ide & Walker, 1993). In contrast to Chomskyan generative linguistics, which focuses on internal knowledge of language structures (Chomsky, 1957), empirical corpus linguistics seeks to describe