In the previous chapter, I explained that we may need to reconsider familiar formulations of fundamental principles when we apply them in new contexts. In this chapter, I ask how we might even embark on such evaluations. A traditional and commendable scholarly reflex is that we must ‘ground’ our analysis in certain and self-evident bedrock. I will show the infeasibility of this search for secure moral foundations.

I suggest a non-foundational approach, using a coherentist method: we do the best we can do with the available clues and arguments. The clues include patterns of practice, normative arguments, and casuistically-tested considered judgments. We can work with ‘mid-level principles’ to carry out fruitful analytical and normative work.

The coherentist approach accepts that our principles are human constructs, that our starting points are contingent, and that we have no guarantees of ‘correctness’. Discussion of fundamental principles is not a matter of ethical computations; it is a conversation. It is a human conversation, a fallible conversation, and nonetheless an important conversation. I also argue that coherentism offers the best explanatory and justificatory account of the method used in most criminal law theory. In other words, it is the best theory of criminal law theory.