eophil_R 10/25/05 8:40 AM Page 290 REFLECTIVE EQUILIBRIUM MINIMALISM. Traditional theorizing about reference is Bibliography ambitious; the possibility of a broad and deep theory Almog, J., J. Perry, and H. Wettstein, eds. Themes from Kaplan. such as it seeks has been questioned by Richard Rorty, New York: Oxford University Press, 1989. Robert Brandom, Paul Horwich, and others. The follow- Brandom, R. “Reference Explained Away: Anaphoric Reference and Indirect Description.” Journal of Philosophy 81 (1984): ing schema constitutes a minimal account of reference 469–492. (“a” is replaceable by singular terms): Devitt, M. Designation. New York: Columbia University Press, 1981. (R) For any x, “a” refers to x if and only if x = a. Evans, G. The Varieties of Reference. Oxford: Clarendon Press, 1982. “London” refers to London and nothing else. A minimal- Fodor, J. Psychosemantics: The Problem of Meaning in the ist account adds to (R) the claim that (R) exhausts the Philosophy of Mind. Cambridge, MA: MIT Press, 1987. nature of reference. French, P., T. Uehling, and H. Wettstein, eds. Contemporary Perspectives in the Philosophy of Language. Midwest Studies Some qualifications are necessary. First, if anything in Philosophy, Vol. 5. Minneapolis: University of Minnesota but a singular term replaces “a” in (R), the result is ill Press, 1979. formed, for only singular terms should flank the identity Horwich, P. Truth. Oxford: Blackwell, 1990. sign. If expressions of other syntactic categories refer, Kripke, S. Naming and Necessity. Cambridge, MA: Harvard those categories will require their own schemas. The University Press, 1980. schema for predicates might be: Lewis, D. “Putnam’s Paradox.” Australasian Journal of Philosophy 62 (1984): 221–236. (R') For any x, “F” refers to x if and only if x = Fness. Neale, S. Descriptions. Cambridge, MA: MIT Press, 1990. Putnam, H. Philosophical Papers, Vol. 2: Mind, Language, and Second, the notion of a singular term must be explained Reality. Cambridge, U.K.: Cambridge University Press, 1975. (can “my sake” replace “a”?). Third, (R) does not say Récanati, F. Direct Reference: From Language to Thought. which singular terms refer. When “a” does not refer, (R) Oxford: Blackwell, 1993. Rorty, R. Philosophy and the Mirror of Nature. Princeton, NJ: may not express a proposition. Fourth, (R) cannot be Princeton University Press, 1979. generalized by the prefix “In all contexts”: “today” used Schwartz, S., ed. Naming, Necessity, and Natural Kinds. Ithaca, tomorrow does not refer to today. Rather, (R) should be NY: Cornell University Press, 1977. understood as instantiated by sentences in different con- Timothy Williamson (1996) texts (for instance, uttered tomorrow with “today” for “a”). Fifth, when one cannot understand the term “a,” one cannot understand (R). Thus, one will find many instances of (R) unintelligible. reflective equilibrium One’s grasp of the minimal theory is not a grasp of Reflective equilibrium is a coherence method of philo- each of many propositions; it is more like one’s grasp of a sophical justification or inquiry. Nelson Goodman (1955) general pattern of inference. For (R) the pattern is in the introduced reflective equilibrium, although not under sentences that express the propositions, not in the propo- that name, to contemporary philosophy in a discussion of sitions themselves (it is not preserved when a synonym deductive and inductive logic. It is arguable, however, that replaces the unquoted occurrence of “a”). This generality philosophers have employed something such as reflective does not satisfy all philosophers. Many accept the mini- equilibrium to inquire into a wide range of topics since mal theory but reject minimalism, because they postulate ancient times. a deeper (for instance, causal) theory of reference that explains (R) and (R'). Although the reductionist demand Goodman maintained that we justify an inference by for strictly necessary and sufficient conditions for refer- showing that it conforms to the rules of either deduction ence in more fundamental terms may be overambitious, or induction. But for the inferences to be justified, these a good picture of reference might still reveal more than rules must be valid. Goodman held that we justify rules of (R) and (R') without meeting that demand. inference by showing that they accord with judgments we make about which particular inferences are acceptable See also Frege, Gottlob; Indexicals; Kaplan, David; and which are unacceptable. Goodman addressed the Kripke, Saul; Marcus, Ruth Barcan; Philosophy of Lan- obvious objection to such a procedure as follows: guage; Proper Names and Descriptions; Putnam, This looks flagrantly circular. I have said that Hilary; Rorty, Richard; Russell, Bertrand Arthur deductive inferences are justified by their con- William; Sense; Strawson, Peter Frederick. formity to valid general rules, and that general ENCYCLOPEDIA OF PHILOSOPHY 290 • 2nd edition eophil_R 10/25/05 8:40 AM Page 291 REFLECTIVE EQUILIBRIUM rules are justified by their conformity to valid cognitively spontaneous; they might concern proposi- inferences. But this circle is a virtuous one. A tions that are either particular or general. The first step on rule is amended if it yields an inference we are the road to reflective equilibrium, according to Rawls, is unwilling to accept; an inference is rejected if it for S to eliminate certain initial moral judgments. For violates a rule we are unwilling to amend. The example, judgments that are not stable over time or in process of justification is the delicate one of which S has little confidence should be dropped, as making mutual adjustments between rules and should judgments formed when S is emotionally dis- accepted inferences; and in the agreement thus traught. In general, S eliminates those judgments formed achieved lies the only justification needed for in circumstances where there is some obvious reason for either. suspecting error. The remaining judgments will be S’s (GOODMAN 1955, P. 67) considered moral judgments. It is possible to read Goodman’s proposal as purely S’s next task is to formulate a moral theory, that is, a methodological or as more epistemological. According to set of moral principles that accounts for S’s considered the epistemological reading, when we complete the moral judgments. The coherence element of the method process of mutual adjustment Goodman describes, comes into play at this stage because S will grant neither thereby bringing our judgments regarding the particular considered moral judgments nor moral theory a privi- inferences and the rules of inference we accept into a state leged status. S will make revisions on both sides in the of reflective equilibrium, these rules and particular judg- attempt to forge a coherent system of moral beliefs. When ments are by definition justified. According to this read- the emerging theory is found to conflict with central, very ing, being justified consists in being part of a system of confidently made considered judgments, S must revise beliefs, including rules and particular judgments, that has the theory. But if a well-confirmed element of the theory the sort of coherence that reflective equilibrium repre- that is independently plausible is found to conflict with sents. less firmly held considered judgments, then S will revise these judgments. S’s decision regarding what to revise is The methodological understanding of reflective made for each case on the basis of what seems most likely equilibrium accepts that a philosophical inquiry into to be true or correct to S upon due consideration (there inductive or deductive inference is properly conducted by are various other ways of expressing this idea: We might, a process of mutual adjustment of the kind Goodman for example, say that S is to decide on the basis of S’s describes; it agrees that this is the best we can do in an degrees of belief or commitment or on the basis of what attempt to justify the inferences we make and the rules of seems most plausible or acceptable to S). inference we accept. But the methodological reading is not definite about the epistemic status of the particular To this point, Rawls’s method corresponds with and general judgments we manage to bring into reflective Goodman’s, but Rawls does not allow S to stop here, at a equilibrium. It leaves open what positive epistemic status, point of narrow reflective equilibrium. According to if any, principles and judgments that are in reflective Rawls, S must next consider alternatives to the moral the- equilibrium might have. In addition, whereas the episte- ory that S accepts in narrow equilibrium along with mological reading is committed to a coherentist account philosophical arguments for and against S’s own theory of justification, the methodological reading leaves open and the various alternatives S is considering. In his how best to account for the precise epistemic status (or important work on reflective equilibrium, Norman statuses) attained by judgments that are in reflective equi- Daniels (1979) argues that we can think of this as an librium. Although Goodman probably intended an epis- attempt to attain coherence between the considered temological reading, let us adopt a methodological moral judgments and theory that S accepts in narrow understanding of reflective equilibrium in order to keep reflective equilibrium and the background theories S as many epistemological options open as possible. accepts. The idea is that the philosophical arguments that We owe the term reflective equilibrium to John S constructs will use premises drawn from among S’s Rawls (1971), who developed the method further and broader background beliefs, which might include such applied it to moral inquiry. If we think of the method as things as sociological views regarding the role of morality something to be applied in a stepwise fashion, which is in society and philosophical or psychological theories useful even if not entirely accurate, then an inquirer, S, regarding rational decision or the nature of persons.