A new dialogue is beginning between students of international law and international relations scholars concerning compliance with international agreements. This article advances some basic propositions to frame that dialogue. First, it proposes that the level of compliance with international agreements in general is inherently unverifiable by empirical procedures. That nations generally comply with their international agreements, on the one hand, or that they violate them whenever it is in their interest to do so, on the other, are not statements of fact or even hypotheses to be tested. Instead, they are competing heuristic assumptions. Some reasons why the background assumption of a propensity to comply is plausible and useful are given. Second, compliance problems very often do not reflect a deliberate decision to violate an international undertaking on the basis of a calculation of advantage. The article proposes a variety of other reasons why states may deviate from treaty obligations and why in many circumstances those reasons are properly accepted by others as justifying apparent departures from treaty norms. Third, the treaty regime as a whole need not and should not be held to a standard of strict compliance but to a level of overall compliance that is "acceptable" in the light of the interests and concerns the treaty is designed to safeguard. How the acceptable level is determined and adjusted is considered.