3.3 The argument from reXective equilibriumIf we ask what might be the source of our judgements concerning rational-
ity and what gives them any authority they have, we may begin to see the
force of the argument from reXective equilibrium (Cohen, 1981, 1982). In a
debate on the philosophy of science Kuhn once remarked that if we
suppose we have an authority for assessments of rationality independent
of that supplied by reXection on the best scientiWc practice, then we are
opening the door to cloud-cuckoo-land (1970, p.264). More generally, one
might say that any theory of rationality must ultimately be answerable to
human practices of reasoning. If that is so, then it must be doubtful
whether we can bring in any overall verdict of irrationality on the way that
people think.
It will help to bring matters into focus if we compare and contrast the
issue concerning rationality, on the one hand, with ethics, on the other – an
area in which a method of reXective equilibrium is widely accepted as the
appropriate way to develop and reWne moral theory. The idea behind the
methodology of reXective equilibrium in ethics is that we should aim to
Wnd principles which systematise moral judgements which are held to be
intuitively correct after due and appropriate consideration. What the
ethical theorist is doing is dependent upon what ordinary thinkers are
doing when at the top of their form. The theorist wants to arrive at
principles which deliver, in general, just the same verdicts as ordinary
thinkers, trying to minimise cases where the judgements of ordinary thin-
kers have to be rejected as erroneous.
There are morally diYcult situations in which, with the best will in the
world, one might be uncertain as to what one ought to do. We also use
reXective equilibrium to assess principles against received judgement in less
problematic cases, and then extrapolate the principles in the hope of
resolving the diYcult and contentious issues – rather as if we were building
astatute lawof moral theory on the basis of thecase lawof normal moral
judgements. It is quite possible that we may not feel happy about the
resolution of some moral dilemma when subsumed under a proposed
principle. But in that case it is up to us to articulate the reasons for our
unease – and these reasons for rejecting the proposed solution then be-
come, in their turn, part of the data which a more reWned moral theory
needs to take into account. That is what the process of reXective equilib-
rium between general principles and particular judgements involves.
People are by no means always certain to make the right judgements, of
course, even when they are trying to do the right thing. They may make
mistakes because they areXustered, upset, confused, involved in too
partisan a way with some of the people aVected by a moral decision, and so
Philosophical arguments in defence of rationality 117