Insofar as the material residues of interest to archaeologists are cultural and, as such, have specifically symbolic significance, it is argued that archaeology must employ some form of structuralist analysis (i.e. as specifically concerned with this aspect of the material). Wylie examines the prevalent notion that such analysis is inevitably ‘unscientific’ because it deals with a dimension of material culture which is inaccessible of any direct, empirical investigation, and argues that this rests on an entrenched misconception of science; it assumes that scientific enquiry must be restricted to observables. It is clear, as realist critics of this view have argued, that scientific (explanatory) understanding depends fundamentally on theoretical extensions beyond observables; extensions which bring into view underlying and inaccessible causal structures or mechanisms responsible for the manifest phenomena through a procedure of analogical model construction. In consideration of realist models of these procedures and of the potential of linguistic modes of analysis for archaeology, it is proposed that archaeologists might (and, in fact, often do) effectively grasp the symbolic, structural order of surviving material culture through analysis governed by a rigorous and controlled use of ethnographic analogy. It is claimed, moreover, that the archaeological record can provide empirical bases for evaluating these theoretical constructs if a procedure of recursive and systematic testing is adopted in research, but the standard hypothetico-deductive model is seriously flawed as an account of an ideal for this procedure. Glassie's analysis of Middle Virginian folk housing is an example of research along these lines which illustrates the potential for a rigorous structuralist alternative.