Click for abstract

P. S. Greenspan

University of Maryland

I. Introduction

In this paper I want to outline a metaethical view that I think represents an addition to the standard alternatives in metaethics. I shall indicate how it connects to historical approaches and to some leading views in the contemporary literature. There are several recent authors working along similar lines, but the view emerged in my own work from an extended treatment of emotion, which supplies a somewhat different moral- psychological basis.2

The view is meant to contrast with noncognitivist accounts of ethics in terms of emotion (emotivism and more recent versions of expressivism), but I shall not review in detail the arguments that the basis it assigns to ethics allows for moral realism.3 Instead I want mainly to present the view here and to recommend it for further attention, first in more or less the form in which it emerged, as a way of explaining the special motivational force of moral judgments (Section II), and then recast in somewhat artificial terms, for comparison with another (at least arguably) realist approach to metaethics known as "response-dependency" (Section III).4 My own view can be seen as a social or two-level version of response-dependency, but with a less rigid account of motivational force introduced by the move to the social level. I shall go on to acknowledge some difficulties but also to cite further advantages of my suggested approach on issues of justification (Section IV) and on the question of the relevance of emotion to moral judgment (Section V).

In a nutshell: A response-dependent account explains a concept in terms of a disposition to respond in some way under standard conditions. The paradigm case outside ethics is the definition of color concepts--of red, say, as applying to things that would look red to normal observers under normal conditions of color perception. As currently understood, a response-dependent account need not amount to a definition; but in application to ethics it is exemplified in its most familiar form by definitions of moral concepts in terms of the reactions of an "ideal observer"--someone satisfying conditions of accurate moral response, on the order of impartiality and imaginative acquaintance with the facts.5 This is sometimes seen as the way of basing ethics on emotion within a realist or descriptivist framework--though some of the current nondefinitional variants of response- dependency (which do not have to worry about circularity) substitute moral judgment for the relevant response. But despite my own emphasis on emotion, I want to suggest something less subjective. For the concept of moral wrong, which I understand as the fundamental distinctively moral concept, the relevant response on my proposed account will come out in the first instance as something done by society or social practice: a collective response on the order of "forbidding" that s brought to bear on individual motivation by the way we teach moral language and behavior in conjunction with emotion.

II. Emotions and Moral Motivation

Emotion comes into the account I want to propose, not as directly supplying the meaning or content of a particular moral judgment, but rather as an indirect prop of moral meaning to the extent that it supports the overall institution of moral discourse. It is not itself the relevant moral response, that is, but a presupposition of it in general terms. Let me explain my proposal first with reference to the intermediate position it takes on the question of motivational force, or moral internalism versus externalism.

Unfortunately, this is an area in which "isms" and senses of them--and senses of the terms used in defining them--abound, though at the same time they often are too sharply dichotomized, so that our conception of the alternatives still ends up impoverished. On many of these issues my own aim is to work out a "halfway-house" position that satisfies the reasonable aims of each side. Let me just stipulate briefly, to begin with, that I shall be using the word "motivation" to mean something psychological, with "motivational force" understood as a causal or quasi-causal push toward action. There is another sense of this terminology, to denote a kind of rational or justificatory force, that I prefer to cordon off with different terms such as "normativity." In the case of moral language, the question of motivational force in the form that concerns me here is whether moral terms or judgments themselves motivate--so that someone who genuinely holds a moral judgment must have some tendency to act on it.

Those who say "yes" are thought of as internalists, since they take motivational force to be "internal" to moral meaning. Those who say "no"--who hold that motivation is supplied by something further, typically some noncognitive factor such as desire--count as externalists. However, neither of these competing views yields a fully satisfying account of the possibility of remaining totally unmoved by some moral judgment one acknowledges. This is summed up in the figure of the "amoralist," a rational agent who understands moral judgments and accepts at least some of them on an intellectual level but just does not care to act accordingly. On an internalist view amoralism comes out as impossible--not just in such extreme form but even in application to more circumscribed areas of behavior, where it seems quite common, intuitively speaking, to have a few moral opinions one is not prepared to live up to.

Someone might accept, say, that eating factory farm animals is wrong without making any move to revise her behavior accordingly, instead limiting the judgment to an assessment of behavior from an ideal but impracticable moral standpoint. As long as it is limited to isolated areas of morality, this possibility seems to make perfect sense, whether or not it would make sense if generalized to morality as a whole. On the other hand, one wants to say, there is something very odd about it. What could it mean to accept a prohibition on future action if not that one feels some barrier to performing it? On an externalist view even amoralism of the global sort, applied to morality in general, would come out as conceptually unproblematic. But surely some sort of influence over behavior is part of what we mean by calling a judgment "moral."

An intermediate possibility is usually overlooked here, though it is obvious enough once seen: Motivational force may be essential to the meaning of moral language generally--to the social institution of moral discourse, as something whose defining purpose involves getting people to do things--without being required by any particular moral judgment. I think of this as "general internalism," though on standard definitions it comes out as a version of externalism. That classification seems apt to the extent that the view relies on noncognitive factors, factors external to moral belief, to explain moral motivation.

We might take this as a version of Hume's insistence on basing morality on the passions. However, I would stop short of a thoroughgoing Humean or other "sentimentalist" position on these issues, which would treat moral judgments as built upon emotions rather than amounting to beliefs in the sense of strictly cognitive contents.6 In any case, my own roughly Humean approach to the bases of ethics will not depend on the view usually associated with Hume in the contemporary literature, that a particular belief would be incapable of motivating in the absence of emotion.7 What I want to suggest is just that we rely on emotion as a supplement to cognition in general terms in the way we teach morality--and thereby set up the institution of moral discourse.

I take it that we teach moral language, moral emotion, and moral rules more or less simultaneously to children.8 What comes first is basic emotion tendencies such as sadness and anger, along with some sort of higher-order tendency to pick up emotional reactions from others-- something like Humean sympathy (in psychologists' current terms, empathy) or its developmental predecessors, emotional contagion or some form of emotional identification.9 These allow us to inculcate what John Stuart Mill called "internal sanctions" of morality--guilt feelings for a wrong act, say--via social disapproval and various forms of punishment for prohibited behavior.10 A child does something we have forbidden, and we make a show of anger, focusing negative personal attention on her: "Bad girl!" She hurts a sibling, and we also play up our empathy for her victim. But this way of instilling moral emotion is just how we teach moral behavior, and for that matter, what moral terms mean. It is how we invest certain terms like "wrong" with a moral meaning. Moral language is not taught first and applied to the teaching of moral rules or other social norms or expectations; nor are specifically moral emotions available at the outset of the process, or independently of it, to be plugged in as motivators. The shaping of moral behavior simultaneously uses and conveys all three.

Though this fairly commonplace account is framed in terms of individual development and hence presupposes a surrounding institution of moral discourse, I think we can extend it to the origins of the institution, if we bear in mind the biological basis of emotions, along with pre-linguistic possibilities of social shaping, and think of moral emotions as initially formed from a more basic set by something like the process I have indicated. However, the creation of the institution on any reasonable account would have to occur in small increments: The first stage cannot involve judgments expressed in moral language but rather, say, imperatives from some powerful source. Much of our current training of children goes on in such terms anyway: "Mustn't do that" gets going from "Don't!" plus a display of anticipatory anger from a parent or other object of childhood dependency.

The result is a nonaccidental connection between moral meaning and motivation--what externalism of the standard variety denies-- but without the standard internalist claim of necessary connection in a particular case. In adult life it is possible (though not always easy) to undo some of the connections forged by childhood education; and in any case they are only rough or by-and-large connections. Awareness of an unfulfilled obligation, say, typically involves an element of anticipatory guilt or some similar psychological goad toward action--a penumbra of associated anxiety or other discomfort--and the fact that it does is no accident (conceptually as well as causally). But most of us at least some of the time can get beyond this, and there are those whose childhood moral education never had this effect in the first place, at any rate with regard to some particular area of behavior.

So emotion as a subjective state is in this sense essential to the meaning of moral judgments, but not because it supplies their meaning in the direct sense of expressivist or sentimentalist accounts. This allows for a naturalist account of moral meaning as something intrinsically motivational, but without traces of logical positivism--and without taking moral emotions as straightforwardly representing some independent reality. According to what I have proposed thus far, emotion serves to register in individual motivation the social standpoint of assessment that moral judgment, whatever its content, presupposes. Guilt, for instance, as a kind of uncomfortable self-disapproval inculcated according to the scenario just depicted, provides a socially manipulable goad toward conformity to social rules. It is enough for the general institution of morality that this motivational mechanism be effective most of the time. So amoralism comes out as possible, at any rate in isolated cases.

III. Two-Level Response-Dependency

Now I want to fill out my picture of the motivational bases of ethics by analogy with "response-dependency," which in this area amounts to the dispositional theory of value. I should stress, though, that my own view is not meant to extend to the general subject of value but only to what I think of as the fundamental distinctively moral value, or rather disvalue, namely "wrong." For ease of comparison I want to offer a schematic account of "wrong" that amounts to a social version of the response-dependent analysis. But it is important to bear in mind that this is not put forth as an attempt to explain moral value in terms of some sort of nonevaluative or nonnormative notion--to reduce it to something more basic, presumably some sort of psychological or social property or fact. Rather, my suggestion will presuppose notions of group rationality, or arguably even moral assessment, on the order of "social flourishing," or the "viability" or "adequacy" of a moral code, understood in terms of promotion of flourishing. The point is not to explain value or normativity, particularly in the sense of explaining it away, but rather just to locate it correctly--not "out there," as a property of acts (on the model of Mackie's "objective prescriptivity" or "to-be-doneness") but instead in social practices and ends, of a sort that we can argue about without entanglement in metaphysics.11

The notions of social value that the account relies on may or may not themselves be understandable dispositionally, on the same model--that is, in terms of our tendency to react in some specified way under the appropriate conditions. If a dispositional theory of value does apply at this level, then my suggestion for understanding moral wrong could be seen as adding a further stage to standard response-dependency. But the account is also meant to be compatible with an appeal to moral intuition at this level--the level of assessing social arrangements or moral codes as "adequate" or states of society as "flourishing." My own hope is that various strategies of empirical and contractarian argument will let us avoid or minimize such appeal; but we may have to resort to it in order to handle cases of repugnant social practices such as slavery that otherwise might be held to promote group flourishing.12

In any case, some of the proponents of response-dependency have argued that a nonreductive or nondefinitional analysis on this pattern need not be undermined by failure to break out of "the circle of the normative."13 Its essential purpose would just be to exhibit the interrelation of concepts. In the case that concerns us here these would be normative or value concepts-- on the model of Aristotle's account of virtue in terms of (individual) flourishing, and with some of the same acknowledged limitations.

The dispositions I would propose as underlying moral assessment at least in the first instance are dispositions of groups; and the responses in question are practical, on the order of "forbidding." Response-dependency in some of its forms allows for these inter alia.14 But the usual focus is on subjective responses, meaning responses of individual subjects, in the sense of subjects of moral experience. This is currently taken to include cognitive modes of registering experience, such as judgment. Stated schematically, to cover its several varieties, and in a way that makes its possible circularity apparent, a response-dependent analysis might read "x is wrong" as:

As the brackets are meant to indicate, "seems wrong" is meant just as a "dummy" expression, or placeholder, not conveying significant information but rather just standing in for various ways of characterizing the relevant subjective responses, with appropriate conditions spelled out accordingly as normal or ideal for the relevant form of response. I sum these up, following the historical versions, as conditions specifying an "ideal observer," in the sense of an individual subject with a specified range of psychological reactive attitudes--in this case, say, attitudes of aversion or disapproval. By contrast, my own suggestion might be stated as follows: The reference to a "morality" here is meant to apply to a social institution, not (at least in the first instance) a body of thought. Instead of "forbidden" I use "pronounced wrong" in this formula as a parallel expression to "seems wrong" above; but it is misleading in similar ways that reflect some oddities of the account to be discussed later. For the moment, just note that this also is meant as a dummy expression covering different forms of response--not necessarily group judgments that something is wrong, any more than "seems wrong" is meant to imply individual judgments (though some versions of response-dependency allow for these). In general, one should bear in mind that the schema is not meant to convey concrete information but rather just a structure--a pattern of analysis, formally similar to standard response- dependency.

My proposal might be summed up, then, as taking a judgment of moral wrong to assert that some act is dispositionally forbidden-- by society in general, or its moral code (a term I use to cover any system of norms, whether or not literally codifiable as a set of rules), rather than the person who makes the judgment, or any particular moral agent or observer. However, these references to society and its acts and moral code presumably must be explicable in terms of individual reactions. The reactions will not be limited to subjective responses such as feelings or judgments of disapproval; overt acts of censure and even legal punishment also count, though their relevance will ultimately depend on subjective attitudes: responses of individuals to violations of the code.15 So morality on the socially-based account does have an indirect basis in individual emotional and other reactive attitudes just insofar as the existence of institutional norms depends on the responses of people within the relevant group to norm violations.

Thus, for instance, the judgment that stealing is wrong does not in itself ascribe to its speaker, or to someone else, a negative reaction to acts of theft, even under hypothetical conditions of moral perception. Rather, in the first instance, the judgment makes a hypothetical claim about social norms: the prohibitions and practices of a well-ordered society. But what keeps a set of norms in place is our responses--whatever it takes to bear out the claim that we by and large accept them. The tendency of group members to impose some sort of blame for acts of theft is the basis for the social prohibition. So in making the moral judgment we do imply something about how people react or would react to violations.

On the other hand, the responses in question will not be allocated in any fixed way to group members. In some cases such as legal punishment they involve coordinated group acts whose participants may vary; but even for simpler responses such as social disapproval, all that is required is enough participation in the general social practice, not participation on the part of all (or of any particular) individuals. This means that the socially-based view allows for intuitively plausible cases of amoralism of the sort I discussed earlier.

On standard response-dependent accounts, by contrast, we seem to face a choice between internalism and externalism in application to a given moral judgment. Which view is adopted depends on whether the fill-in for the relevant subjective response--for the dummy expression "seems wrong" in our schema--is taken as motivating or as motivationally inert, and on how it is supposed to be related to moral judgment. A halfway-house position might initially seem to be achievable by specifying appropriate conditions of moral judgment. David Lewis' discussion of his own version of response- dependency suggests that he has this in mind.16 But to make a long story short, I think any likely moves in this direction would just relocate the problem.

The amoralist is supposed to be a rational agent, remember--not just in general terms but particularly as concerns his behavior in the relevant cases. This rules out weakness of will and similar lapses from rationality such as temporary depression or accidie. It is important to setting up the problem of amoralism that the agent's insensitivity to moral reasons still leaves him fully functional with regard to the capacity (as opposed to the inclination) to act as morality requires. Such conditions of moral judgment as he might be said not to meet would presumably involve taking a wider perspective than self-interest--undergoing something like a Humean "correction" of emotions for distortions of personal standpoint, but here conceived as supplying missing emotions.

However, would the upshot of such a process necessarily be motivating? Whether or not the missing emotion itself would be motivating, it is the dispositional judgment about it--that it would be motivating under the relevant conditions--whose motivational force is in question on a response-dependent account. This latter is what spells out the content of the moral judgment; but it would seem to benefit from the motivational force of the emotion in question only to the extent that the one who makes the judgment currently happens to fulfill the conditions of appropriate moral response. This either exaggerates or fails to account for the force of moral judgments that we come to as conclusions of "removed" reflection, rather than in contexts of immediate practical choice. Such judgments do not necessarily motivate, but typically they do--and not accidentally, in view of their meaning.

What preserves our sense of the "magnetism" of moral language on my own proposed account is the indirect but foundational role assigned to emotions as motivators. Reference to a wider, social perspective is built into the content of a moral judgment, but in a way that makes it possible to acknowledge that perspective and the judgment it supports without being motivated by it. On the other hand, the reference on my account to the "viability" of a moral code is meant to supply a conceptual connection of moral judgments by and large to emotional and other sources of motivation. A moral code simply would not be viable unless it can rely on some such source as the basis for moral teaching.

IV. Justificatory Issues

It might be possible to build something similar into standard response-dependency, but I think there are further reasons for moving to a two-level version. The socially-based view amounts to "two-level" response-dependency in the sense that the notions of social response it uses to explain moral value ultimately are supposed to be explainable themselves in terms of reactions of (indefinitely specified) individuals. One advantage of this approach has to do with the different sort or level of "correction" it involves: correction of a moral code for viability, in a sense that implies adjustment to an end, or conditions of adequacy to a presumed purpose, rather than representational fitness to an apparent object of apprehension--the sort of perceptual analogy that encourages metaphysics.

Setting up my view as a social version of response-dependency highlights an ambiguity in the notion of a "moral response": Though forbidding (on the part of society, or the moral code) might be conceived as a response to the conditions of social life, or something else of moral significance, it could also be understood as a response to some moral judgment or presumed moral fact. Though my schematic account allows for circularity in order to capture the varieties of response-dependency, what I have in mind is the first reading. In "pronouncing" an act wrong in the sense intended we are passing judgment on it, performing a speech act (in the simplest cases) that under specified conditions would amount to the moral fact in question.17 We are not registering some prior fact about it--declaring it to be wrong in some independent sense--as if the moral code consisted in a set of propositions or moral judgments, like a primitive attempt at a moral theory.

Instead, what I mean to suggest is that sophisticated moral judgments, the sort we discuss as components of moral theories, should be understood as saying something about moral codes-- or more precisely, about acts in relation to corrected versions of them. Codes in turn are to be understood as social institutions meant to achieve or to allow for certain ends--what I sum up as "group flourishing." There might seem to be frequent uses of "wrong" from within society that fail to fit this analysis, judgments that merely report a local norm. But many such judgments will be appropriate utterances in their local contexts and yet not strictly true. Whatever the speaker might have in mind by them, on my account their truth depends on whether the norm in question really is required by group flourishing--whether there are reasonable alternatives to it, and whether it would survive correction of the moral code. In any case, such moral "pronouncements" serve to accomplish something: They use moral language (whether or not with strict accuracy) to effect moral norms.

Moral truth comes out on this account as holding independently of the elements of any particular moral code--including, of course, the actual or surrounding code--or the moral responses of any particular speaker, to the extent that these may or may not adequately promote the end of social flourishing. What a moral judgment records (if it is on the mark) is a requirement of flourishing--including hypothethical corrections of the moral code, ways it needs to be changed to promote flourishing. So the account can be put forth as a version of realism, with the same qualifications as for standard response-dependency. But the two- level version will behave differently in explanatory terms in relation to moral judgment. Let me briefly show how I think the result is an improvement--modifying some objections to response- dependency outlined by Simon Blackburn.18

On a standard sort of response-dependent account of moral judgments, by analogy to judgments of color perception, the "real" property, the one we are analyzing, apparently explains our reaction: It makes sense to say that something seems wrong- -or that we have the more particular responses that might be filled in here--because it is. However, whether we can reverse the order depends on what sort of fill-in we choose for "seems wrong" and how we interpret "because": An act is not wrong because we think it is, presumably; but if we fill in some more particular reaction for "seems wrong"--aversion or disapproval--an act's wrongness might be said to be grounded in its tendency to produce that reaction in ideal (or ideally situated) observers.

In that case, however, a question arises about our practices of justification: It is only from the standpoint of theoretical explanation that personal moral responses like aversion or disapproval can be said to supply grounds for moral judgment. In practical life, they provide our initial access to moral truth, but when challenged as to our reasons for thinking some act wrong, we normally focus outward--not on our community of co-believers but on the implications of the act. Only those who lack any real understanding of ethical argument--or have exhausted the normal means of argument, or are impatient with them to begin with---would respond by citing the reactions of other observers or the conditions shaping their own.

By contrast, though it is meant to satisfy some of the same aims as standard response-dependency, the socially-based view adds some helpful complexity with its addition of a further level of response. Note, first, that we can still say that something seems wrong because it is, at the level of individual response. What we might mean is that our moral education generated the negative reactions it did because it could thereby adequately promote group flourishing; this is the moral fact corresponding to an act's actually being wrong. The way morality is taught via emotion also explains how we can use our reactions as a guide to moral truth: "It just feels wrong" gives a reasonable prima facie ground for abstaining from action, even if not a justification of the corresponding judgment.

Something similar also holds on the social level, about the actual moral code: We can say that the code pronounces something wrong because it is wrong--meaning that any corrected version of the code would do the same. If all goes well, moreover, we can take the code as a rough guide to moral truth. However, its justification--and hence the justification of individual moral reactions--ultimately depends on complex hypothetical statements about the promotion of group flourishing, of the sort that spell out the content of "wrong." At this point, in contrast to standard response-dependency, our order of explanation can unproblematically run in reverse--without violating ordinary justificatory practice, and without demanding at the outset some more specific fill-in for the presupposed conditions on moral codes. Greater specificity can be seen as the outcome of an ongoing interactive process of reflective equilibrium: balancing refinements in the notion of group flourishing against intuitive judgments on particular cases on the individual level, on the assumption that our moral education was roughly on the mark; and on the social level, carving out jointly feasible group projects.

What makes an act wrong, then, on the account I propose, will be the fact that it would be prohibited by any moral code that is corrected for the sake of adequate promotion of group flourishing, taken as fulfillment of the purpose of a moral code--any "viable" code, as I sum this up. The code does not forbid an act because it is wrong in some further sense. So our explanation terminates in self-reference--it exhibits how the concept in question depends on us, on our responses--in roughly the way that response-dependent accounts have in mind. But its collective interpretation of "us" also explains our external justificatory focus: away from our moral feelings, even as a group, and toward the conditions of life in a group--which of course depend on various feelings on the part of individuals. It is in this indirect--and essentially social-- sense that ethics is based on emotion.

V. Emotions and the Content of Ethics

My account refers to emotions rather than desires--the usual fill-in among contemporary Humeans for noncognitive sources of motivation--because I think desires on their contemporary conception have been drained of the element of negative feeling that we need for an explanation of motivational force. They have become mere wants or preferences, without clear distinction from the act tendencies they are supposed to explain, though the term still carries residual associations of emotional "unease" from its earlier uses in British philosophy.

What I have in mind by an emotion, in any case, is partly cognitive, at least in the developed form that supports its role in moral teaching.19 Emotions on my account amount to complexes of affect and associated evaluative thoughts that can be seen as representing the practical significance of some state of affairs by a given emotion's positive or negative aspect. The component of thought extends to moral evaluations, for distinctively moral emotions such as guilt; but it also takes much less sophisticated forms, in a way that is meant to allow for the origin of moral judgment in more basic emotions--and for the origin of developed human emotions in infantile and animal response. The function of emotion is to direct attention--by loading some thought or view of the environment with affective comfort or discomfort. Just because of the way discomfort absorbs attention, I take escape from discomfort to be motivating in a stronger sense than mere avoidance of some object of aversion--for a distinction between behavioral goals and goads that the contemporary reference to desire as a motivator tends to obscure.

On my suggested picture of childhood moral teaching, the content of morality is supplied by specific behavioral precepts on the order of "It s wrong to hit your brother." But what is taught here, or what ultimately is learned from it, of course has to be something beyond a code of rules, to allow for the possibility that the rules taught are defective. If nothing else, it often is necessary to teach moral behavior at an early stage on the basis of oversimple versions of the rules; and the actual social rules in force may in any case need improvement, sometimes even radical rehauling. We teach specific moral rules as an approximation to something more general--something on the order of "the requirements of life in a group" or what I sum up as "group flourishing."

This term, which I originally took from J. L. Mackie, amounts to an extension of the Aristotelian notion of flourishing, applied by Aristotle to the individual case to explain the value we place on various more immediate ends of action.20 Here its intent is just to capture in a simple phrase the general point of moral rules, the aims of morality as a social institution. For metaethical purposes, we would do best to leave its content open, to allow for various positions on normative ethics. But this is compatible with an admission that, in order to support intuitively plausible rules of social justice, it may need to have some normative content.

In a fuller treatment, that is, we might want to impose some structural or other normative constraints on what would count as a "flourishing" society or an "adequate" morality: a requirement of reasonable balance in its distribution of individual flourishing, say, or perhaps some more explicitly contractarian requirement of justificatory reasonableness.21 So despite its teleological formulation, this account of the bases of ethics is not meant to tip the scales in favor of a consequentialist view. Since flourishing on this account may not amount to a strict welfare notion, the appeal to group flourishing does not yield a version of utilitarianism--just as Aristotle s appeal to individual flourishing does not yield egoism. As with virtue on Aristotle's account, moral norms or principles may be at least partly constitutive of the moral end, not just instrumentally valuable as means to it.

Contrary to what Mackie probably had in mind, then, social flourishing or the viability of a society or its moral code is not to be understood simply as a matter of "fitness" in something like the biological sense--the notion involves more than group survival or perpetuation--though it implies this.22 On the other hand, it need not be understood as ideal or maximal promotion. I think of a "viable" moral code as one that adequately promotes group flourishing. Any consequentialist terminology in my account might just be replaced by a reference to conditions of adequacy--applied to a moral code, seen as a set of institutional norms, but measured against a conception of the point of the institution, the aims whose fulfillment determines its success. A viable code in this sense has to meet a certain standard; but there might be several competing codes that meet the standard, at least in principle-- assuming that competition for members under appropriate conditions of choice would not threaten the survival of all but one of them.23

However, I understand the notion of moral wrong in such a way that this element of social relativism need not yield a relativist account of "wrong" in philosophers' terms: one on which moral truth varies with the speaker's (social or other) standpoint. I take the judgment of wrong to imply, not just (or necessarily) that the act in question is forbidden by the actual moral code in force, or presupposed by a certain speaker, but rather something stronger, with a hypothetical element: that the act in question would be forbidden by a corrected version of the code--corrected for the sake of viability. This means any corrected version--in the limiting case, "corrected" simply by replacement by another code-- so that a prohibition that is valid enough from within a particular social standpoint need not yield a true judgment of moral wrong.24 A violation of filial piety, say, may be "locally forbidden," and appropriately so (as part of one adequate scheme for organizing moral life, let us say), yet not wrong--though we often use the language of moral assessment from within our limited social standpoint to sum up the locally applicable rules.

To call an act wrong, then, on my suggested schematic account is really to make a claim, not about the surrounding moral code, but rather about any code that adequately promotes group flourishing-- any "viable" code, in short. This "broad-brush" characterization is intended just to convey a structure--one that reconfigures the standard approaches in ways I have tried to illustrate with response-dependency. It is meant to be open to interpretation, both for purposes of metaethical neutrality and just because a certain amount of indeterminacy is appropriate to capturing what we have in mind.25 However, my proposed account is not intended as a semantical analysis but something more like a structural breakdown, of a sort that can accommodate different attempts to make our ordinary meaning more precise.

The account essentially relocates the standard appeal to "human nature" as fixing the content of morality--often taken as justifying a conservative approach--to a position that allows for social decision and experimentation on important moral matters. Natural human emotion tendencies count among the grounds of morality to the extent that they constrain the content of a viable moral code. We need teachable, workable, and sustainable rules, in short, and cannot be satisfied with codes that either fail to exploit or overestimate the basic resources of human nature: our sense of relatedness, in varying degrees, to similar others, filled out by (and sometimes balanced against) the need for intelligible rationales for what we do and allow.

Can such an account capture our common notion of "moral" value, or does it delimit or distort the common notion by tying it so closely to personal reactions and the requirements of social life? What if human psychological makeup turns out to impose some requirements on a viable morality that do not themselves intuitively count as "moral"?26 For instance, even supposing that a violation of filial piety is merely "locally" forbidden, there might be some more general need for clearly marked symbols of respect or even reverence that holds for any well-functioning social group and in our culture bans such things as flag-burning. Some set of limitations on human sexual interaction might be taken as a requirement of a viable moral code in this sense, even if victimless violations do not in themselves seem "wrong"--and not because they conform to some set of alternative conventions.

Another source of problematic examples is the need for relatively simple general rules: Perhaps morality in order to be widely teachable has to avoid too many fine distinctions among types and degrees of norm violation, so that the question whether a case of very minor shoplifting or tax-evasion is really wrong, say, might well leave us "of two minds." There is a general need for artificial line-drawing in play (in different ways) in both of these examples that may not just be "local" in the sense of being limited to a particular culture or set of property- conventions. We might want to say that there are some things that have to be forbidden "globally"--in all cultures, though under some suitably broad description--but at the same time merely accidentally, as a result of how humans happen to be constituted rather than the content of morality per se.

I think that an answer to objections of this sort would involve spelling out either the criteria for inclusion in a moral code or the conditions of adequacy to allow for results of higher-level intellectual inquiry. There may be presuppositions of adherence to a code or consequences of applying it that do not count strictly as parts of it, or of the rationally satisfying version of it that we seek in moral theory. It may indeed be a fact of human nature that moral and certain nonmoral prohibitions are enforced by the same emotions and will therefore be seen by many people as much the same thing. But the attempt to "rationalize" our moral responses--to mold them into a coherent set of reasons for moral behavior--is no less natural to us.

The perfectability of a moral code in response to later demands- -up to a reasonable standard of adequacy, that is--also has a historical dimension that should help allay some of our worries about relativism. We might impose a requirement on a genuinely viable code that it be so over the long haul--in a sense that entails extensibility to a global culture: the ability to handle contact and overlap (of the sort we now increasingly have) with alternative modes of social organization. My brief suggestion earlier that some moral codes would not survive this sort of cultural competition might be taken as supplying a socially-based alternative to the optimistic assumptions about convergence in our moral responses needed to support standard response-dependency.

The possibility of group members "voting with their feet" on the social norms--of women abandoning more traditional ways of life, say, if offered genuine alternatives--should be taken as a supplement to the more standard appeal to human nature (including reason) as the source of norms. We need not decide at this level of theory just what is "programmed in" and what has to be supplied or can be modified by social learning. In general, the schematic account I offered above for purposes of comparison with response- dependency is subject to fine-tuning, or even basic structural modification, in light of both empirical results and the results of further philosophical inquiry into ethics and moral psychology.

VI. Conclusion

My main concern here has been with the metaethical role of psychological reactions, specifically emotions: These do have a place among the grounds of moral judgment on the socially-based approach I have outlined--in contrast to a Kantian approach, or even to Mill's treatment of "internal sanctions" as useful in enforcing morality but irrelevant to its content. Their influence on the content of ethics is indirect, though--in contrast to Humean "sentimentalism" or some other version of response-dependency. In calling a moral code "viable" we make a claim about its relation to the raw materials of human psychology. It has to be teachable on the basis of natural emotion tendencies; and in later life, presumably, if it pulls against them, there will be social costs to be reckoned with in sustaining that version of the code. So in that sense the view accords moral truth a basis in emotion.

There are other constraints on a viable moral code--a code that adequately fulfills the functions of a moral code--and I do not mean to set up emotions and other moral responses as the sole ground of moral judgment. Insofar as group flourishing involves individual flourishing in a group, and supposing that the arrangements meant to secure it have to be defensible to group members in order to have a claim on their adherence, the standard resources of consequentialism, contractarianism, and other systems of normative ethics still apply. But the social/emotional basis of ethics also provides an important line of argument that on this account is not merely "tacked on" as a source of motivation to an independently derived systematic foundation, as on most current versions of externalism.

The mechanism of emotional identification, most notably, as the basis of moral teaching would seem to impose natural limits on a viable moral code. Consider issues involving the narrowing of moral concern to an inner circle--whether oneself or one's own subgroup or the human species. Infants early on exhibit a crying response to the sound of another infant crying. This is is the bottom-level foundation of ethics, via moral teaching, on the account I have suggested; and it is relatively unselective, though different societies mold it in different ways. Arguably, however, it would be difficult--not at all impossible, particularly in isolated states of society, but self-undermining for society in general over the long haul--to maintain a code of rules based on some artificial cut-off point in the extension of concern. Which lines are in fact tenable is a complicated question; they may or may not have to include other species, for instance. Sometimes, moreover, our natural reactions will constitute barriers to social flourishing and will have to be altered or simply accommodated with the least effect. My point for purposes of this paper is just to indicate briefly how argument from emotions figures at the most basic level of moral theory.


1. This paper is forthcoming in The Journal of Ethics (1998). An earlier version, entitled "Social Bases of Ethics," was written while I was Visiting Fellow in the Philosophy Program of the Research School of Social Sciences at the Australian National University in 1995. I owe thanks, for comments, to Philosophy Department audiences there and at La Trobe University, Monash University, and the University of Queensland. Let me also thank Michael Smith and Susan Dwyer for further helpful suggestions.

2. Cf. David Wong, Moral Relativity (Berkeley: University of California Press, 1984), David Copp, Morality, Normativity, and Society (New York: Oxford University Press, 1995), and my own Practical Guilt: Moral Dilemmas, Emotions and Social Norms (New York: Oxford University Press, 1995), esp. chs. 3 and 6. For my earlier work on emotion see esp. Emotions and Reasons: An Inquiry into Emotional Justification (New York: Routledge, Chapman and Hall, 1988).

3. For an emotion-based version of expressivism in the contemporary literature see esp. Allan Gibbard, Wise Choices, Apt Feelings: A Theory of Normative Judgment (Cambridge, Mass.: Harvard University Press, 1990). My defense of what I call "social artifact realism" appears in Practical Guilt, chs. 3 and 6.

4. See esp. Phillip Pettit, "Realism and Response-Dependence," Mind, C (1991), 587-626, pp. 606-21, for an argument that the view counts as realist, albeit with qualifications; cf. also David Lewis, "Dispositional Theories of Value," Proceedings of the Aristotelian Society, Suppl. LXIII (1989), 113-74, pp. 132- 37. The term "response-dependency" was introduced in Mark Johnston, ibid., 139-74, pp. 144-55. The general approach in the contemporary literature stems from John McDowell, "Values and Secondary Qualities," in T. Honderich (ed.), Morality and Objectivity: A Tribute to J. L. Mackie (London: Routledge, 1985); cf. also David Wiggins' version of "subjectivism" in Needs, Values, Truth (Oxford: Blackwell, 1987). I shall take "realism" very simply here, to cover views according to which moral judgments describe some independent reality--meaning independent of any particular subjects' assent to the judgments.

5. For a contemporary formulation of the position see Roderick Firth, "Ethical Absolutism and the Ideal Observer," Philosophy and Phenomenological Research, 12 (1952), 317-45. The historical view is sometimes attributed to Hume; a clear statement occurs in Adam Smith, Theory of the Moral Sentiments. "Moral sense" theories in earlier eighteenth-century British moralists also count as variants of response-dependency, with the relevant response taken as an assessment of one's own states of feeling; cf. Lewis, "Dispositional Theories of Value," pp. 115- 16ff., on second-order desires.

6. See David Hume, A Treatise of Human Nature (Oxford: Clarendon Press, 1964), esp. Bk. III, Pt. I, Sect. 2. For a version of the view I put forward here, see Philippa Foot's treatment of Hume on practical force in Virtues and Vices (Berkeley: University of California Press, 1977), pp. 74-80.

7. Gerald Postema (personal communication, 8/95) points out, however, that Hume might be taken to allow for judgments that do not provide overriding motivation--and possibly provide none at all, as in cases of amoralism--since the sentiments Hume takes as the bases of moral judgment, love and hatred, are not on his view essentially motivating but only motivate by virtue of their causal tie to other passions, e.g. anger in the case of hatred.

8. For a fuller version of the account I sketch here see Practical Guilt, ch. 3, sec. 1. A "moral emotion" on the use I intend is one with a specifically moral content--on the model of guilt, taken as ascribing moral responsibility. There may be moral reactions in a broader sense, of reactions keyed to morally significant states of affairs (e.g., disapproval of some act as harmful to others rather than as an infraction of rules of politeness or similar nonmoral conventions); and some of these reactions may precede any moral learning. My account is not meant to deny this possibility--or particularly to represent punishment as the source of all moral awareness in children, though it figures in the illustration that follows.

For that matter, I think the account I sketch here could be reframed to allow even for some innate moral emotions in my sense (as suggested, say, by Melanie Klein's view of guilt as arising in infancy for aggressive wishes towards the mother). The point is just that the content of such feelings--and our stock of moral emotions--would be modifiable by social learning in the way I outline.

9. For important distinctions in this area cf. esp. Lauren Wisp‚, "History of the Concept of Empathy," pp. 17-37, and Ross A. Thompson, "Empathy and Emotional Understanding: The Early Development of Empathy," pp. 119-145, in N. Eisenberg and J. Strayer (eds.), Empathy and its Development (Cambridge: Cambridge University Press, 1987). I use "emotional identification"--e.g., in "A Case of Mixed Feelings: Ambivalence and the Logic of Emotion," pp. 223-50, in A. O. Rorty (ed.), Explaining Emotions (Berkeley: University of California Press, 1980); cf. also Emotions and Reasons, esp. ch. 5)--to cover cases of situation-specific empathy without identity- confusion; but cf. John Deigh, "Empathy and Universalizability," pp. 199-220, in L. May, M. Friedman, and A. Clark (eds.), Mind and Morals: Essays on Ethics and Cognitive Science (Cambridge, Mass.: M.I.T. Press, 1996), pp. 213-14.

10. See Utilitarianism, ch. III.

11. See J. L. Mackie, Ethics: Inventing Right and Wrong (Harmondsworth: Pelican, 1977).

12. For some suggestions on these issues see Practical Guilt, ch. 3, sec. 3, and ch. 6, sec. 2.

13. See Johnston, "Dispositional Theories of Value," p. 147, pp. 155ff. This is in contrast with traditional "ideal observer" definitions, which attempt to work out a nonnormative specification of the conditions of observation; cf. esp. Firth, "Ethical Absolutism and the Ideal Observer."

14. See Pettit, "Realism and Response-Dependence," p. 617.

15. Cf. the discussion of Foot and Hare on insititutional norms in Michael Smith, The Moral Problem (Oxford: Blackwell, 1995), pp. 77-84; cf. Simon Blackburn, "The Flight to Reality," in R. Hursthouse et al (eds.), Virtues and Reasons: Philippa Foot and Moral Theory (Oxford: Clarendon Press, 1995), pp. 35- 75, esp. pp. 51-53.

16. Cf. Lewis' "iffy" connection between value judgment and desire in "Dispositional Theories of Value," pp. 116-17.

17. This might provide grounds for classifying the present view as a version of "constructivism," which essentially makes out moral facts as a product of minds or mental activity; cf. John Rawls, "Kantian Constructivism in Moral Theory," Journal of Philosophy, 77 (1980), 515-72. However, my move to the social level makes a difference as to whether a view thus classified counts as a version of idealism rather than realism; cf. David Brink, Moral Realism and the Foundations of Ethics (Cambridge: Cambridge University Press, 1989), p. 15.

18. See Simon Blackburn, "Circles, Finks, Smells and Biconditionals," in J. E. Tomberlin (ed.), Philosophical Perspectives 7: Language and Logic (Atascadero, Calif.: Ridgeview Publishing Co., 1993), pp. 259-79.

19. Cf. Emotions and Reasons, esp. ch. 2; note that I make out emotional affect as having a propositional content, not as directed toward a separable element of propositional thought (in the sense of an episode of thinking).

20. See Aristotle, Nicomachean Ethics, trans. David Ross (Oxford: Oxford University Press, 1980), Bk. I, esp. sec. 5; cf. my treatment of Mackie in Practical Guilt, ch. 3, sec. 2.

21. Cf. esp. T. M. Scanlon, "Contractualism and Utilitarianism," in A. Sen and B. Williams, Utilitarianism and Beyond (Cambridge: Cambridge University Press, 1982), pp. 103-127. My approach essentially relocates the issue between utilitarians and contractarians to a stage after initial agreement on the fundamental subject matter of ethics.

22. Cf. J. L. Mackie, Persons and Values: Selected Papers, Vol. II (Oxford: Clarendon Press, 1985), esp. pp. 152-69; see Scanlon, "Contractualism and Utilitarianism," p. 128.

23. For a bit more on these issues see Practical Guilt, ch. 6. The preference for adequacy over maximization criteria also may be invoked to explain why my schematic account of "wrong" is not based on direct estimation of the effects of forbidden acts (as opposed to the moral codes that forbid them) on group flourishing: Many of those acts we intuitively consider wrong may have only negligible effects in themselves, though the general practice of allowing them would be another story.

24. Cf. Scanlon, "Contractualism and Utilitarianism," p. 112.

25. Cf. James Dreier, "Internalism and Speaker Relativism," Ethics, 101 (1990), 6-26, esp. pp. 23ff., for a nice way of building in indeterminacy that might be used to modify my suggestions here. Note that Dreier also defends an intermediate position on internalism ("modest internalism"; see esp. p. 14), though he resists a social version (p. 21) as a consequence of his central argument, that "speaker relativism" yields the best explanation of the position.

26. André Galois brought this up as a problem with any response- dependent or other counterfactual account of moral judgment. His made-up example--of a psychological need for a ban on stepping on cracks on the sidewalk--would make the same general point.

Last revised, 8/1/97, by P. S. Greenspan