One of the benefits of working in applied settings is that problems tend to be well defined. It is relatively easy to measure progress, and analytical techniques can be objectively evaluated. For the laboratory researcher, however, problems are often ill defined. There can be no clear starting point, well-stated goal, or simple way of evaluating or marking progress. Consider the study of pure memory, an enterprise widely pursued by researchers in cognitive psychology. Most memory researchers focus on a particular memory system or process, such as episodic, semantic, or implicit memory, and they conduct experiments to (a) isolate its underlying mechanisms and (b) determine its parameters of operation. The goal is to analyze the system's structure and its component parts in much the same way that a chemist might analyze a chemical compound by breaking it down into simpler elements. Attention is rarely given to the system's function at this point, largely because understanding function is presumed to depend on understanding structure. After all, how can we determine the function of a system unless we first understand the system itself? It is necessary to isolate the critical components, along with some rules for their interaction, and then--perhaps--the adaptive role that the system plays in cognition can be specified. In this chapter, I discuss some of the implications of this widely practiced structuralist, or nonfunctional, approach to the analysis of memory. To begin, as noted above, problems crafted within a structuralist framework tend to be ill defined. When we set out to study implicit memory, it is difficult to gauge progress, or measure success, because the objective is unclear. The components of an implicit memory system are unknown; consequently, there is no way of determining when, or if, the system has been fully described. Researchers often end up studying tasks as a result, such as paired-associate learning or word-fragment completion, because task performance is easy to evaluate objectively. The trouble with this focus, however, is that the link between the studied tasks and the true memory system of interest can be tenuous or, more likely, inadequately specified. (PsycINFO Database Record (c) 2012 APA, all rights reserved)