Why do I keep going in circles? On the impossibility of constructing rational preferences

In standard models of rational choice it is typically taken for granted that preferences are given and defined over the alternatives alone, and the possibility of making a rational choice is simply a matter of assumption. In this paper I generalise this aspect of the economic model so that preferences over alternatives are constructed from given preferences defined over various characteristics of the alternatives under consideration. I characterise the decision problem before investigating what conditions a procedure for aggregating preferences over attributes into preferences over alternatives must satisfy in order for the latter to be rational. I then consider what the implications of these conditions for the procedural rationality of the aggregation process.