Revealed preference theory is not a specific theory; it is a broad programmatic framework for analyzing choice behavior. Within this broad framework there are a number of different revealed preference theories (different versions of the program)—they all share common features, but there are also sharp differences. The diversity of revealed preference theory is not well understood, and one purpose of this article is to improve our historical understanding of the field by examining this historical diversity. This history is valuable for its own sake, but also because it is relevant to recent debates about the methodological foundations of rational choice theory among experimental psychologists, behavioral economists, neuroeconomists, and others. The second purpose of this article is to use material from the Paul Samuelson archives to help us understand how Samuelson, the originator of revealed preference theory, viewed his contribution to the program and how he evaluated the different versions of revealed preference theory. The article will examine Das Paul Samuelson Problem: the question of whether Paul Samuelson changed his mind about the foundations (the nature, significance, and purpose) of revealed preference theory over time.
