Notwithstanding complaints about incoherence in Establishment Clause doctrine, courts by and large administer the Clause responsibly. They do so by mediating between a number of powerful considerations, none of which can ever be entirely disregarded. These considerations include, but are not limited to, separation of church and state, the value of religiosity, the imperative of affording equal treatment to religious and similarly situated nonreligious entities, and the proper role of courts in a democratic political system. This is not to say that courts cannot overstep their bounds and provoke an adverse reaction from other powerful elements within the polity. It is only to say that courts, being sensitive to important political considerations, tend to avoid the kind of provocation that would undermine their role in the political system.

In Part II of this article, I briefly summarize the apparent incoherence of Establishment Clause doctrine. In particular, I note that the various "rules" of nonestablishment are logically irreconcilable and tend to undermine the significance of the Free Exercise Clause, itself part of the Constitution. I then offer four possible explanations for this incoherence, the first two of which are based on indirect arguments and the second two of which are based on direct arguments. First, I suggest that any rule, such as the Establishment Clause, that contains the seeds of incoherence and that was originally intended to apply in limited circumstances may well appear profoundly incoherent when applied on a massive scale. Second, I observe that the decision by most nations of the world not to disestablish religion may be circumstantial evidence of the difficulty of the task. Third, I argue that any government that expects to act rationally in all instances must predicate its actions on some conception of the "good life," thereby implicating establishment. Finally, I note that people cannot subsist as a matter of psychology without religion of some sort; therefore, positive government must make religious choices in order to function. These last two arguments partake of the "republican" approach to lawmaking, which expects government and citizenry to support each other in fostering individual and collective virtue.

In Part III, I argue that courts nevertheless make sense of the Establishment Clause by mediating between such logically irreconcilable values as separation of church and state and republicanism. To help illustrate this point, I use the arguments set forth by James Boyd White in his monograph When Words Lose Their Meaning. In one chapter of this work, White described a rhetorical community in which players addressed each other by reference to a set of important considerations, remembering always to include each consideration at some level in their argument. As White noted, if any player in this community dared to eliminate one consideration from the mix, as one did, that player risked outlawry and destruction at the hands of the others. I argue that the same risks apply to Establishment Clause jurisprudence. If the courts systematically ignored an important consideration underlying the Clause, such as separation of church and state, the polity would react strongly and somehow force the courts to rectify the error. I then list some powerful considerations that inform Establishment Clause doctrine.

In Part IV, I bring my observations to bear on a specific rule promulgated pursuant to the Establishment Clause. In particular, I note that the prevailing test for implementing the Establishment Clause, the so-called "Lemon test," calls for an empirical analysis of the effect of legislation. Specifically, the test asks whether the "primary effect" of a law under review is to promote or inhibit religion. I then note that the Supreme Court typically honors this rule in the breach by expressly disavowing any willingness to quantify effects. In light of my earlier analysis, however, I conclude that the Supreme Court should not engage in an overly technical measurement of a law's effect, but instead should take the measure of the law's effect as appreciated by the polity in context in making its ultimate decision.