There are a few structural requirements that are necessarily common to any tax system. In Part I of this article I examine those requirements. I show that by isolating the necessary structure of a tax system from its particular content or goals, we can better understand the role played within the federal system of certain of its most characteristic features. In particular, I trace the function of the realization requirement for the recognition of income and distinguish it from another sort of function that could lead to the recognition of income within an income tax system. This second function is a corrective one, and it is achieved in the federal system by a mechanism I call "reconciliation.

" Reconciliation has considerable descriptive power within the federal system. I show this in Part II by examining, as examples, three general circumstances that are well described as instances of reconciliation. Two of the three lead to the recognition of income, yet neither can be properly described in terms of realization.

Finally, in Part III, I argue that the notion of reconciliation can play an important role in normative discussions about the federal income tax system. This role has two parts. First, reconciliation describes actual phenomena within the federal system with more coherence and rationality than traditional vocabulary and concepts allow. Those who accept coherence within a tax system as a value will attach some value to that description. Second, describing those phenomena in that way implies certain further descriptions of other parts of the tax system. Sometimes, however, these further descriptions do not fit the law as it actually is, and so the overall coherence of the system will turn out to be less than it would have been if the law had been otherwise. In such cases, a desire for coherence above all else would indicate that the law ought to be changed. However, other, sometimes competing, values such as administrability or equity, can be relevant to determining whether the law really should be changed. When values compete, one will have to choose between particular aims on the one hand and the coherence and rationality of the system on the other. My purpose here is not to recommend a particular choice among these options, but rather to establish that self-conscious articulation and exploration of those values and their importance should be a goal of normative discussion about our income tax system. In Part III, I illustrate this argument by showing that, in contrast to the usual description, reconciliation allows us to describe the tax treatment of simple borrowing in a way that is internally consistent and coherent. I discuss the implications of this coherent description for the analysis of purchase money mortgages and for the central issue in Tufts. Ironically, although my analysis does suggest the kind of transactional bifurcation advocated by Barnett, the requirements of rationality tum out to be somewhat different from what he claims, and to come closer to the current system than we might expect. There are important respects, however, in which the analysis is not consistent with current practice. Instead of arguing that the law ought either to be reformed or differently interpreted, I explore one significant administrative cost of implementing the analysis and suggest that a full articulation of the issue in Tufts would be more complicated than indicated by either Barnett or the Supreme Court.