Philosophical naturalism, considered in general, is not a unified doctrine but a broad label applied both to methodological stances (e.g., "The methods of philosophy are continuous with those of empirical science") and to substantive positions (e.g., "For a belief to be epistemically warranted is for it to be the product of a certain kind of causal process"). The two are often combined, as when a naturalistic interpretation of a given domain of discourse is justified as "the best explanation" of associated practices. However, the two are in principle independent. In the moral case, for example, it has been argued that a projectivist or noncognitivist interpretation gives a better explanation of moral practice than any substantive naturalism (Blackburn 1984, Gibbard 1990).
But what makes a method or interpretation naturalistic? Attempts to give an explicit definition have largely been abandoned in favor of pointing. Roughly, naturalistic methods are those followed in actual scientific research (including—according to some but not all naturalists—mathematics and social sciences as well as natural sciences). And a naturalistic interpretation of a discourse is one based upon predicates or terms that play a role in the explanatory theories that research has generated.
This characterization of naturalism is informative but incomplete. There are vigorous debates within the philosophy of science over just what the methods, concepts, or posits of contemporary science are. Moreover, interpretation based upon naturalistic terms encompasses some quite different tasks. Some examples follow, but first we should ask, Why stay within naturalistic terms at all? Science is a theoretical, descriptive/explanatory enterprise while morality is held to be essentially practical and normative. One might think, no sooner did morality emerge from the shadow of religion than philosophers began trying to push it into the shadow of science. Is it never to be allowed to stand in its own right as a distinctive domain of inquiry?
An answer of sorts is possible. Morality by its nature cannot stand entirely on its own. Moral discourse is supervenient upon the nonmoral and, specifically, the natural—two actions or agents cannot differ in their moral qualities unless there is some underlying difference in their natural qualities. This and other truisms about morality, such as "Ought" implies "can," tie moral evaluation to the natural world in ways that no ethical theory can altogether ignore. Moreover, morality presents us with various epistemic and metaphysical puzzles. We believe that we have come to possess at least some moral knowledge—but how? (See Harman 1977.) We treat moral statements as if they stated genuine propositions—but can this idea be sustained in light of the normative role of moral judgment? We freely make moral judgments, but do they have presuppositions or make claims that are incompatible with our understanding of the natural world?
Hard determinists, for example, have challenged intuitive attributions of moral responsibility by arguing that the notion of free agency they presuppose is incompatible with the world revealed by physics. And John L. Mackie is led to an "error theory" of morality by his diagnosis that moral evaluation attributes to states of the world an objective "to-be-pursuedness" that cannot be fit with any plausible empirical theory (Mackie 1977).
Immanuel Kant, for one, frankly accepted that he could see no way of reconciling the deliberative standpoint of morality with the causal perspective of science. Rational agents must, he held, postulate the compatibility of moral agency with the natural order, even though this remains inexplicable to them. But few philosophers have been willing to stop there. Empirical science affords the best-developed picture we have of ourselves and our world. Without the special authority of religion to back it up, morality inevitably becomes a focus of practical and theoretical concern.
Substantive moral naturalists in effect propose to overcome some of the mystery and potential conflict surrounding the relation of morality to our empirical self-understanding by showing just how much of morality might be found within the domain of the natural. This could be done by providing a naturalistic account of moral discourse that affords an analysis of moral terms (Lewis 1989), or permits a worthwhile revision of moral language that nonetheless can serve virtually all the same functions (Brandt 1979), or enables us to reduce moral properties to natural properties (Railton, 1993), or shows moral properties to be natural properties in their own right (e.g., thanks to their contribution to empirical explanation; see Boyd 1988, Miller 1985, Sturgeon 1985). Substantive ethical naturalism promises to explain such important features of moral discourse and practice as the applicability of notions of truth and falsity to moral claims, the supervenience of the moral upon the natural, the role of natural properties in justifying moral claims, and the possibility of semantic and epistemic access to moral notions through ordinary experience.
The first half of the twentieth century had not been kind to substantive ethical naturalism (for a brief history, see Darwall et al. 1992). Condemned by G. E. Moore (1903) for committing the "fallacy" of trying to close an "open question" by analytic means and rejected by nonfactualists (emotivists, prescriptivists, etc.) for failing to capture the special relation of moral evaluation to motivation and action, naturalism fell into disuse. But by mid-century naturalism had begun to win its way back. The initial steps were taken, independently, by Philippa Foot (1958–1959) and Geoffrey Warnock (1967), who argued that one could not be competent in moral discourse unless one possessed some substantive, contentful moral concepts. Moral evaluation is distinguished from aesthetic or prudential, for example, in part because it has a certain descriptive, arguably natural content—namely, a concern with the effects of our actions on the well-being of others. If we came upon a society in whose behavioral code the key notion was guleb, a term applied in the paradigm case to warriors who have killed an enemy bare-handed, we would certainly mislead if we translated guleb as "morally good" or "just" rather than "valiant" or "courageous."
Meanwhile, Peter Geach (1965) showed convincingly that existing nonfactualist views could not account for the full grammar of moral discourse, in particular, the logical behavior of unasserted moral claims in conditionals.
Foot (1972) took the next step as well, challenging the "internalist" conception of the relation of moral evaluation to motivation that served as the basis for nonfactualism. She argued that ordinary moral agents are able to see themselves as motivated by a rationally optional concern for others. Those who lack such a concern might lack moral character, but they do not make a linguistic mistake in using the moral vocabulary.
This sort of moral "externalism" offers an alternative explanation of why moral evaluation and motivation are so intimately related, at least in paradigm cases. Concern for others is a very basic part of normal human life. An Aristotelian would say that human nature itself is social; a Darwinian would emphasize the contribution of concern for others to inclusive fitness and to the possibility of benefiting from reciprocal altruism. Speculative biology apart, it is possible to see how social norms involving concern for others, keeping promises, and so forth might emerge and be sustained in virtue of their contribution to solving various serious coordination and collective-action problems. Such norms will function best only if well internalized by a major part of the population. It should therefore be unsurprising that moral judgment is usually accompanied by a positive attitude. Moreover, it should not be forgotten that moral judgment is a species of assertion and that assertion itself involves, not only signaling a cognitive attitude of belief, but also various forms of active endorsement or encouragement, as well as associated claims of authority. Moral externalism, by drawing upon these ingredients (and others) for an alternative explanation of the evidence—such as it is—offered on behalf of internalism, has attracted a number of defenders (see, for example, Boyd 1988, Brink 1989, Railton 1986).
Another sort of naturalism, however, takes the opposite tack. It treats the purported relation to motivation as fundamental but interprets it in a subjectivist rather than nonfactualist manner. Subjectivist interpretations of moral discourse have historically faced difficulties in accommodating all the elements of an interconnected set of features of morality: the critical use of moral assessment, the nonrelativistic character of moral judgment and the possibility of genuine moral disagreement across social or cultural differences, the limits on empirical methods in resolving moral disputes, and the seemingly normative character of the relation between moral judgment and motivation. Can new forms of subjectivism succeed where others have failed?
Consider the simple subjectivist formula:
- Act A is morally good = A is such that one would approve of the performance of A.
Since approval is a positive attitude, (1) establishes a relationship with a source of motivational force. But is it the right relationship?
We do not typically regard our current tendencies to approve or disapprove as morally authoritative—they might, for example, be based upon hasty thinking or false beliefs. This has led naturalists to modify (1) to require that the approval be well informed and reflectively stable. (See, for example, Brandt 1979 and Firth 1952. For criticism, see Velleman 1988.)
Moreover, not all species of approval have a moral flavor. I can approve of an act because of its aesthetic or pious qualities, for example. Some naturalists therefore amend (1) to restrict the object of approval (e.g., to the set of rules one would—reflectively, informedly, etc.—approve for a society in which one is going to live [cf. Brandt 1979]). Others attempt to identify in naturalistic terms a specifically moral sort of attitude of approval or disapproval (e.g., an attitude of impartial praise or anger). Critics have argued that no noncircular characterization of this kind is possible (for a subjectivism without reductive ambitions, see Wiggins 1987).
Formulas like (1) also threaten to yield relativism. Since they introduce a necessary link to facts about motivation, moral attribution becomes tied to contingencies of individual psychology. That seems wrong, since moral evaluation purports to abstract from individual interest and motive and to prescribe universality. If one is not correspondingly motivated, that is a deficiency in oneself rather than an excusing condition or a limit on the reach of moral judgment. Each of us recognizes that he or she can in this sense be motivationally defective from a moral point of view. (But see Harman 1975 for a defense of a naturalistic moral relativism.) This has led naturalists to modify the formula away from the individualistic language of "one" or "I" and in the direction of a more inclusive "we" or "everyone" or even "normal humans" (see, respectively, Lewis 1989, Smith 1994, and Firth 1952). New problems arise. The notion of "normal human" threatens to introduce a term that itself requires naturalization—since we believe that statistically "normal" humans might be motivationally defective from a moral point of view—for example, in lacking sympathy with those from other groups. (Of course, one could at this point also embrace circularity.) If we insist that everyone approve, there is again a risk that contingencies of motivational idiosyncrasies will receive authority—this time, in preventing us from attributing moral value to states of affairs virtually all (but still not quite all) of us approve heartily on reflection. A less ambitious alternative is to replace "one" with "us" and seek moral consensus where we may. This would help explain the "outreach" function of moral discourse without altogether removing the account's relativism.
An alternative approach avoids relativism by "rigidifying" the subjectivist formula (cf. Wiggins 1987). One fixes the truth conditions of moral judgments by reference to the motivations actually prevalent in one's moral community (e.g., "A is such that we, with our actual motives and with full and informed reflection, would approve of it"). This secures the desirable result that changes in our motives will not in themselves change what is morally good. But it undermines some of the critical role of moral assessment in our own society (since, again, we can imagine that our actual motives are morally defective) and will have the result that those brought up in different social environments with different acquired motivations will lack a common subject matter even though they believe they are having a genuine moral disagreement (for discussion, see Johnston 1989).
No ethical naturalism has emerged that meets all the desiderata of an account of moral discourse and practice. Nonnaturalists and nonfactualists attribute this to a mistaken starting point. But no alternative account has met all the desiderata, either. Moral naturalists have often been accused of "changing the subject"—shifting the locus of attention from the position of the agent involved in practical deliberation to that of the scientist engaged in theoretical description. But this criticism begs the question. Naturalists seek to explain, not ignore, moral experience; if they are right, the phenomena they study are the very stuff of moral thought and action.
See also Conditionals; Determinism and Freedom; Foot, Philippa; Kant, Immanuel; Mackie, John Leslie; Metaethics; Moore, George Edward; Naturalism; Philosophy of Science, Problems of; Projectivism; Supervenience.
Blackburn, S. Spreading the Word. Oxford: Clarendon Press, 1984.
Boyd, R. "How to Be a Moral Realist." In Essays on Moral Realism, edited by G. Sayre-McCord. Ithaca, NY: Cornell University Press, 1988.
Brink, D. O. Moral Realism and the Foundations of Ethics. Cambridge, U.K.: Cambridge University Press, 1989.
Darwell, S. et al. "Toward Fin de siècle Ethics." Philosophical Review 101 (1992): 317–345.
Firth, R. "Ethical Absolutism and the Ideal Observer." Philosophy and Phenomenological Research 12 (1952): 317–345.
Foot, P. "Moral Beliefs." Proceedings of the Aristotelian Society 59 (1958–1959): 83–104.
Foot, P. "Morality as a System of Hypothetical Imperatives." Philosophical Review 81 (1972): 305–316.
Geach, P. "Assertion." Philosophical Review 74 (1965): 445–465.
Gibbard, A. Wise Choices, Apt Feelings. Cambridge, MA: Harvard University Press, 1990.
Harman, G. "Moral Relativism Defended." Philosophical Review 84 (1975): 3–22.
Johnston, M. "Dispositional Theories of Value." Proceedings of the Aristotelian Society 63 (1989): suppl., 139–174.
Lewis, D. "Dispositional Theories of Value." Proceedings of the Aristotelian Society 63 (1989): suppl. 113–137.
Mackie, J. L. Ethics: Inventing Right and Wrong. New York: Penguin, 1977.
Miller, R. "Ways of Moral Learning." Philosophical Review 94 (1985): 507–556.
Moore, G. E. Principia Ethica. Cambridge, U.K.: Cambridge University Press, 1903.
Railton, P. "Moral Realism." Philosophical Review 95 (1986): 163–207.
Railton, P. "Reply to David Wiggins." In Reality, Representation, and Projection, edited by J. Haldane and C. Wright. New York: Oxford University Press, 1993.
Smith, M. The Moral Problem. Oxford: Blackwell, 1995.
Sturgeon, N. "Moral Explanations." In Morality, Reason, and Truth, edited by D. Copp and D. Zimmerman. Totowa, NJ: Rowman and Allanheld, 1985).
Velleman, D. "Brandt's Definition of 'Good.'" Philosophical Review 97 (1988): 353–371.
Warnock, G. Contemporary Moral Philosophy. London: Macmillan, 1967.
Wiggins, D. "A Sensible Subjectivism?" In Needs, Values, Truth: Essays in the Philosophy of Value. Oxford: Blackwell, 1987.
Peter Railton (1996)