abstract Because we are more comfortable with judgements of conceptual conceivability than with judgements of practical possibility, we content ourselves with imaginary cases, which are useless for making many decisions that practical people most need to make, notably all-things-considered decisions about when to follow an admitted general principle and when to make an exception. The diverse cases of climate change, preventive attack, and torture all illustrate how the avoidance of the difficult task of integrating empirical judgements with conceptual judgements through (...) the flight into the sanitized abstraction of imaginary cases undermines attempts at practical ethics. All three cases involve allegedly exceptional, or emergency, situations, although climate change seems to require more than the usual compliance with general principle while preventive attack and torture supposedly require less than the usual compliance. By my lights the proposed exceptions are, respectively, fully, sometimes, and never justified. But the fundamental point is that one unfortunately cannot decide any of the cases without assessing what is in fact likely as well as what is conceptually possible. (shrink)