Reflective equilibrium is a coherence method of philosophical justification or inquiry. Nelson Goodman (1955) introduced reflective equilibrium, although not under that name, to contemporary philosophy in a discussion of deductive and inductive logic. It is arguable, however, that philosophers have employed something such as reflective equilibrium to inquire into a wide range of topics since ancient times.
Goodman maintained that we justify an inference by showing that it conforms to the rules of either deduction or induction. But for the inferences to be justified, these rules must be valid. Goodman held that we justify rules of inference by showing that they accord with judgments we make about which particular inferences are acceptable and which are unacceptable.
According to the epistemological reading, when we complete the process of mutual adjustment Goodman describes, thereby bringing our judgments regarding the particular inferences and the rules of inference we accept into a state of reflective equilibrium, these rules and particular judgments are by definition justified.
According to this reading, being justified consists in being part of a system of beliefs, including rules and particular judgments, that has the sort of coherence that reflective equilibrium represents.
The methodological understanding of reflective equilibrium accepts that a philosophical inquiry into inductive or deductive inference is properly conducted by a process of mutual adjustment of the kind Goodman describes; it agrees that this is the best we can do in an attempt to justify the inferences we make and the rules of inference we accept.
But the methodological reading is not definite about the epistemic status of the particular and general judgments we manage to bring into reflective equilibrium. It leaves open what positive epistemic status, if any, principles and judgments that are in reflective equilibrium might have.
In addition, whereas the epistemological reading is committed to a coherentist account of justification, the methodological reading leaves open how best to account for the precise epistemic status (or statuses) attained by judgments that are in reflective equilibrium. Although Goodman probably intended an epistemological reading, let us adopt a methodological understanding of reflective equilibrium in order to keep as many epistemological options open as possible.
We owe the term reflective equilibrium to John Rawls (1971), who developed the method further and applied it to moral inquiry. If we think of the method as something to be applied in a stepwise fashion, which is useful even if not entirely accurate, then an inquirer, S, begins with a large set of initial moral judgments.
These judgments will be intuitive for S in the sense that they are cognitively spontaneous; they might concern propositions that are either particular or general. The first step on the road to reflective equilibrium, according to Rawls, is for S to eliminate certain initial moral judgments.
For example, judgments that are not stable over time or in which S has little confidence should be dropped, as should judgments formed when S is emotionally distraught. In general, S eliminates those judgments formed in circumstances where there is some obvious reason for suspecting error. The remaining judgments will be S's considered moral judgments.
S's next task is to formulate a moral theory, that is, a set of moral principles that accounts for S's considered moral judgments. The coherence element of the method comes into play at this stage because S will grant neither considered moral judgments nor moral theory a privileged status. S will make revisions on both sides in the attempt to forge a coherent system of moral beliefs.
When the emerging theory is found to conflict with central, very confidently made considered judgments, S must revise the theory. But if a well-confirmed element of the theory that is independently plausible is found to conflict with less firmly held considered judgments, then S will revise these judgments.
S's decision regarding what to revise is made for each case on the basis of what seems most likely to be true or correct to S upon due consideration (there are various other ways of expressing this idea: We might, for example, say that S is to decide on the basis of S's degrees of belief or commitment or on the basis of what seems most plausible or acceptable to S).
To this point, Rawls's method corresponds with Goodman's, but Rawls does not allow S to stop here, at a point of narrow reflective equilibrium. According to Rawls, S must next consider alternatives to the moral theory that S accepts in narrow equilibrium along with philosophical arguments for and against S's own theory and the various alternatives S is considering.
In his important work on reflective equilibrium, Norman Daniels (1979) argues that we can think of this as an attempt to attain coherence between the considered moral judgments and theory that S accepts in narrow reflective equilibrium and the background theories S accepts.
The idea is that the philosophical arguments that S constructs will use premises drawn from among S's broader background beliefs, which might include such things as sociological views regarding the role of morality in society and philosophical or psychological theories regarding rational decision or the nature of persons.
An argument in favor of an alternative to the moral theory that S accepts in narrow reflective equilibrium that is successful in the sense that S finds it compelling would, in effect, show that S's moral theory, considered moral judgments, and background beliefs are not coherent.
As before, S is not bound to favor any type of belief when responding to such an argument; S must decide whether to revise considered moral judgments and moral theory or background beliefs on the basis of what seems most likely to be true to S after thorough reflection. When S attains a coherent system of considered moral judgments, moral theory, and background beliefs, S will have reached a state of wide reflective equilibrium.
An inquirer can certainly move from narrow to wide reflective equilibrium in the way that Daniels maintains, but this is not the only or most interesting way things can go. Consideration of alternative moral views and the relevant philosophical arguments can provide an occasion for a more radical type of revision of belief.
Daniels seems to suppose that the only way in which considering alternatives to one's own view can force one to revise beliefs is by revealing that something else one believes, and believes more strongly, conflicts with one's prior view but coheres with the alternative.
But it is quite clear that consideration of alternative moral and broader philosophical views can also lead one to revise beliefs in a way that is not dictated by one's prior beliefs and degrees of commitment.
It is possible for an inquirer to find an alternative view attractive in its own right, even though it conflicts with everything the inquirer previously thought; and if upon reflection the inquirer finds the new alternative sufficiently attractive, he or she might well respond by accepting the alternative and revising his or her previous views.
Wide reflective equilibrium is best understood in a way that allows for this radical type of belief revision. According to this understanding, achieving wide reflective equilibrium is not simply a matter of rooting out conflicts among the beliefs one already holds and forging general principles that coherently account for one's considered moral judgments.
It crucially involves exposing one's self to alternative moral and philosophical views with the knowledge that reflection upon such alternatives might lead one to make a radical break with one's previous views.
On this understanding, the ideal of wide reflective equilibrium is not defined merely as achieving coherence among all of one's beliefs: considered moral judgments, a moral theory, and background beliefs. The ideal crucially involves attaining a kind of reflective stability. Inquirers who have attained reflective equilibrium are, in effect, immune to threats from the inside and the outside.
There will be no conflicts within such inquirers' systems of belief, and in addition they can be confident that there are no alternatives to their own systems of belief that they would find more compelling than their own upon due reflection. The first sort of reflective stability is provided by reflective equilibrium on either understanding, the second only if reflective equilibrium is understood in a more radical way.
It is important to recognize that the essential feature of a belief revision that is radical, in the strict sense here at issue, is not the number or range of beliefs that are altered but rather the fact that the alteration is not continuous with the things that one previously believed.
When a belief is revised in a way that is not strictly radical, the change is required in order to attain coherence among one's beliefs, and the alteration is dictated by other things that one believes more firmly than the belief that is revised.
Particularly if the belief that is revised in this way concerns a general principle, the change can require revisions to a large number of other beliefs; such a revision would likely be called radical in common parlance, but it would not be radical in the strict sense.
When a change is one that counts as radical in the strict sense, the new beliefs come to seem compelling to one on their own, apart from their logical or evidential relations to one's previous beliefs. Indeed, the new belief will likely contradict things that one previously believed very strongly. Such a change may involve many beliefs or only a few.
Philosophers, who might be guilty of considering such matters only abstractly, may find it difficult to accept the possibility of such radical changes in belief, but it is easy to find descriptions in novels, biographies, and autobiographies of people altering their views in ways that seem to be radical in the strict sense.
Because reflective equilibrium grants the inquirer's considered moral judgments a crucial role in inquiry, it has been widely criticized as a sophisticated version of intuitionism, making it an unreliable and extremely conservative method.
Daniels (1979) sought to rebut the charge of intuitionism by arguing that reflective equilibrium is not compatible with foundationalism, which is a characteristic of intuitionism. Daniels's basic idea is that the inquirer's considered moral judgments do not function as intuitions because reflective equilibrium allows for such extensive revision of these judgments.
This response seems to rely upon too narrow a conception of foundationalism; in particular, it seems to suppose that the beliefs that serve the foundational role must be identifiable in advance of inquiry and also be unrevisable, or at least relatively unrevisable.
Given these suppositions it is natural to think that if reflective equilibrium is a foundationalist method, the inquirer's initial considered moral judgments must be the foundations, and then conclude that it cannot be a foundationalist method because these judgments are subject to way too much revision.
However, if one supposes instead that the foundations might emerge through a course of inquiry, it is unclear that reflective equilibrium does not constitute a version of foundationalism. During a person's inquiry there will be various relatively strongly held judgments that determine the course of the inquiry and the views the person comes to hold in reflective equilibrium.
It is unlikely that all these judgments will be drawn from among the inquirer's considered moral judgments, but almost certainly many will. Others might come from among the person's background beliefs, some might concern moral principles, and perhaps some will be about which member of a conflicting set of beliefs should be revised.
But that does not really matter. The fact remains that at the end of inquiry, it will be possible to identify a set of judgments that provide a psychological basis for the rest of the beliefs the person holds in reflective equilibrium. Many, although perhaps not all, of these judgments will probably be intuitive in the sense that they are cognitively spontaneous.
It is possible, therefore, that these intuitive judgments serve as an epistemological foundation for the rest of the beliefs the person holds in reflective equilibrium. It remains to be seen whether this constitutes a ground for objection to reflective equilibrium.
Reflective equilibrium fares better when it comes to the charge that it is extremely conservative. The extensive revisability of considered moral judgments may show that it can be construed as a version of intuitionism, but it surely shows that the method is not guaranteed to produce nothing more than a cleaned up, systematized version of conventional morality.
The method has the potential to, and indeed is likely to, lead many inquirers to make extensive changes to their moral views. A particular inquirer might, of course, end up holding very conventional views in reflective equilibrium.
Indeed, this is just what will happen if the inquirer is more strongly committed to enough elements of conventional morality than he or she is to anything that conflicts with them, and retains these commitments through the course of reflection upon alternatives to and criticisms of conventional morality. But it is not clear that a method of moral inquiry is inadequate unless it absolutely excludes this possibility.
More worrisome is a general fact illustrated by the possibility just considered: Given the right (or perhaps one should say wrong!) moral judgments and background beliefs held strongly and tenaciously enough, it would seem to be possible for an inquirer to end up holding virtually any moral view, even a bizarre or repugnant view, in reflective equilibrium. The worry is not confined to reflective equilibrium when used as a method of moral inquiry.
No matter what a person might use this method to inquire about, given the right intuitive beliefs held with sufficient strength and tenacity, the person could end up holding virtually any view one could imagine in reflective equilibrium: extreme skepticism, solipsism, nihilism, anarchism, totalitarianism, atheism, or theism—you pick whatever views you think are beyond the pale.
How then could anyone take reflective equilibrium to be an acceptable approach to moral inquiry, or philosophical inquiry more generally? As various critics have put the point: The method clearly leads an inquirer to the coherent position he or she finds most acceptable, the position that best preserves beliefs to which he or she is most strongly committed, but why think this position is anything more than that, in particular, why think it is true or likely to be true?
One might have once thought that this fundamental objection is really pressing only against reflective equilibrium when used to inquire into morality and other such things, where it is all too obvious that different people can hold, and hold very strongly, very different and incompatible considered judgments.
One might have presumed that for such purposes as working out valid rules for deductive or inductive inferences, which is what Goodman originally proposed that the method be used for, there is no real problem because there just is not the same sort of diversity and conflict between the strongly held considered judgments of different people.
But as Stephen Stich most particularly has stressed, empirical work has shown that many of the inferences ordinary people find intuitively acceptable are in fact fallacious. Hence, we can foresee that the rules of inference these people accept in reflective equilibrium will not be valid. So we cannot even trust reflective equilibrium to be an acceptable method of inquiry for those areas where it was originally proposed.
It might seem, therefore, that Rawls's early critics were right to argue that unless we can find some reason for trusting the reliability of the intuitive judgments that play such a crucial role in the method of reflective equilibrium, this method of inquiry cannot be acceptable.
Daniels (1979) was perhaps right when he claimed, in response, that it is unreasonable to expect such a reason to be provided before we begin our inquiries and that such a reason should only be expected to emerge as part of the overall system of beliefs accepted in reflective equilibrium.
But surely it seems unreasonable to hope that it will ever be possible to offer even such an internal defense of reflective equilibrium as a method that is reliable for all who might employ it. Such a defense would require that all, or nearly all, inquirers employing the method converge on the same theory, and this seems rather unlikely.
In addition, this sort of defense seems to underestimate the obstacles facing reflective equilibrium: It is not just that there is no reason to think the method is reliable, but also none for suspecting that it is unreliable—there are fairly strong reasons for believing that the method is unreliable, that it is not the case that a very high proportion of those who employ it will be led to accept a system of beliefs that is largely correct.
Nevertheless, it is possible to offer a defense, albeit a modest defense, of reflective equilibrium. The first step is to recognize that there are a number of different positive epistemic statuses. For simplicity, let's distinguish only two.
The first is the positive epistemic status that plays the major role in distinguishing knowledge from mere true belief. This status is most commonly referred to as justification or warrant. Attempts to account for justification in terms of the reliable formation of belief have been popular and influential.
Even if such attempts fail, the majority of epistemologists would still maintain that there is some sort of strong connection between justification and truth: justified beliefs must, in some sense, be likely to be true. The second positive epistemic status is the sort of subjective rationality that Richard Foley has stressed.
A belief is rational in this sense when it satisfies the believer's own epistemic standards, that is, when the believer would consider the belief likely to be true after due reflection. Unlike justification, there seems no reason to suppose that beliefs that are rational in this sense are likely to be true.
Having distinguished these two positive epistemic statuses, it should be fairly clear that reflective equilibrium can, in fact, be guaranteed to lead the inquirer to hold rational beliefs.
It should also be easy to see that an inquirer who deviated from reflective equilibrium would be led to hold some beliefs that are not rational because, in order to deviate, the inquirer would have to resolve some conflict by rejecting a belief that, upon reflection, the inquirer considers more likely to be true than the belief being retained.
What reflective equilibrium cannot guarantee every inquirer is justified beliefs. If it followed that by employing the method of reflective equilibrium an inquirer was sure to form rational beliefs but equally sure to form unjustified beliefs, and hence fail to attain knowledge, the method would indeed be unacceptable.
But it would be hasty to infer that no inquirer employing reflective equilibrium can be led to hold justified beliefs simply because the method cannot guarantee justified beliefs to all inquirers.
In an influential paper on the method of moral inquiry written years before he advocated reflective equilibrium, Rawls (1951) argued that we should construct a moral theory by formulating principles that account for the considered moral judgments of competent moral judges.
Whereas the notion of considered moral judgments is used in describing reflective equilibrium, the notion of the competent moral judge has fallen by the wayside. But suppose, as we clearly do in our ordinary lives, that some people are competent moral judges, whereas others are not.
We ordinarily suppose that the intuitive moral judgments of competent moral judges are reliable. We might be wrong, and of course we might also be right, that those people who have the characteristics we commonly associate with moral competence, in fact, make a person a reliable moral judge.
So let us understand competent moral judges as those whose intuitive moral judgments are reliable. The beliefs competent moral judges would hold in reflective equilibrium obviously would be reliable.
If one condition for being justified is that a person must not only be reliable, but be able to prove that he or she is, then perhaps even the beliefs held by competent moral judges in reflective equilibrium are not justified.
For competent judges will not be able to prove to incompetent judges that they are reliable. But this condition for being justified is almost certainly too strong: If we were to apply it across the board, we would know little or nothing.
If what is necessary for justification is only that one is reliable, not that one be able to prove that one is, then the beliefs competent moral judges hold in reflective equilibrium may, for all that has been said so far, be justified.
This is not, of course, all that we might have wanted in the way of a defense of reflective equilibrium. We cannot prove that anyone is or is not a competent judge. We cannot prove which characteristics make for competent judges and which make for incompetent judges. Perhaps there are no competent judges. Perhaps there are, but even they do not know that they are.
But it is not outlandish to think that there are competent moral judges and that most of us know something about who they are and what they are like. But for the sake of this argument, suppose only that it is possible that there are competent moral judges. If there are any competent judges, then the beliefs they hold in reflective equilibrium are justified. And this suggests that these beliefs, or at least many of them, might count as knowledge.
So we can say this much about reflective equilibrium. It is the only rational method of inquiry and it is possible that, by employing this method, a person will be lead to hold justified beliefs and to attain knowledge. This is certainly less than one would like to be able to say in support of a method of philosophical inquiry, but it is sufficient to show that reflective equilibrium is an acceptable method for ethics, and philosophy more generally.