Ethics and the vision of value
Part One: Metaethics
Chapter 1: Moral perception
It is obvious that phronesis is not scientific knowledge. As has been said, phronesis is of the ultimate particular; for the thing to be done is an ultimate particular. So it stands opposite to nous; for nous is of definitions, of which there is no logos, whereas [phronesis] is of the ultimate particular, of which there is not scientific knowledge, but perception—not sensory perception, but like the perception whereby we perceive that the triangle is the ultimate particular in geometry.
We should give no less credit to the unproved remarks and opinions of those who are practically wise, or old and experienced, than we give to [their] proved views. For since they have the eye that experience gives them, they see right.
Aristotle, Nicomachean Ethics 1142a24-29, 1143b14[1]
1.1. Moral perceptions: sight and feeling
I want to begin with the idea of moral perception. That idea is old, widespread, and hardy. Over the twenty-seven or more centuries of Western culture, it has never entirely gone away, either within or outside philosophy. It is there in Aristotle as quoted in the epigraphs, and elsewhere. It is there in the Greek tradition before Aristotle: the same image of moral knowledge as sight is memorably present in the Republic’s famous myths of Sun, Line and Cave. It is there outside the Greek tradition too, for instance in the Gospel of John: “I am come into this world that they which see not might see; and that they which see might be made blind” (John 9.39). Given that they stood in a tradition that combined Greek and Judaeo-Christian influences, it is no surprise that moral theorists in the eight hundred years from Jerome to Aquinas who developed theories of conscience (synderesis, conscientia) evidently often had something like a perceptual ability in mind.[2] So too—outside philosophy—did Shakespeare, though he seems, like Socrates (Apology 39e-40c), to have thought of conscience less as a kind of sight, and more as a different sort of perception—a feeling of unease caused by the agent’s belief that he is on the point of doing wrong. (“A man cannot steal, but it accuseth him; a man cannot swear, but it checks him; a man cannot lie with his neighbour’s wife, but it detects him… it fills a man full of obstacles”: Richard III 1.4.132-5.)
In the Enlightenment (notice the term, by the way) we find the same tropes of moral perception as sight and as feeling—usually some sort of unease such as “the pangs of conscience”.[3] The notion of moral perception as feeling is developed by that “exalter of conscience”[4] Bishop Butler—and then at once treated as analogous to sight as well:
Since, then, our inward feelings, and the perceptions we receive, from our external senses, are equally real; to argue from the former to life and conduct, is as little liable to exception, as to argue from the latter to absolute speculative truth. A man can as little doubt whether his eyes were given him to see with, as he can doubt of the truth of the science of optics... And allowing the inward feeling, shame; a man can as little doubt whether it was given him to prevent his doing shameful actions, as he can doubt whether his eyes were given him to guide his steps. And as to these inward feelings themselves; that they are real; that man has in his nature passions and affections, can no more be questioned, than that he has external senses. Neither can the former be wholly mistaken, though to a certain degree liable to greater mistakes than the latter. (Butler, 15 Sermons preached at the Rolls Chapel, 2)
For the archetypal “moral sense” theorist, Lord Shaftesbury, moral perception can be both sight and feeling:
No sooner are actions viewed, no sooner the human affections and passions discerned (and they are most of them as soon discerned as felt), than straight an inward eye distinguishes and sees the fair and shapely, the amiable and admirable, apart from the deformed, the foul, the odious, or the despicable. (Shaftesbury, The Moralists, Part 3, Section 2)
Talk of moral perception is common in twentieth-century ethics too, in a variety of authors including David Wiggins 1987, John McDowell 1979, Thomas Nagel 1991, Jonathan Dancy 1992, David McNaughton 1988, and Iris Murdoch 1970. Wiggins develops an account of moral perception as sensitivity—as something like a feeling:
…some <property, response> pairs will and some will not prove susceptible of refinement, amplification and extension… Some pairs are such that refinement of response leads to refinement of perception and vice versa… Some are and some are not capable of serving in the process of interpersonal education, instruction and mutual enlightenment. Those pairs that do have this sort of advantage, we may expect to catch on and survive…the properties that figure in these <property, response> pairs will be primitive, sui generis, and incurably anthropocentric, and as unmysterious as any properties ever will be to us. (Wiggins 1987: 196-197)
While Iris Murdoch is emphatically visual in her imagery:
Attention… a just and loving gaze directed upon an individual human reality… I believe to be the characteristic and proper mark of the active moral agent. (Murdoch 1970: 34)
What philosophical sense can we make of this image of moral awareness as perception (whether sight or feeling)? Is the image only an image, or is there some literal truth in it? I shall argue for the latter alternative. I shall develop my own version of the claim that there is a viable account of moral awareness which takes talk about moral perception pretty well completely straight. The notion of moral perception can be understood without what Lewis 2005: 315-6 calls a “disowning preface”: literally, and without apology. In my treatment, the leading analogy will be with sight rather than with feeling; but that point is not essential to the argument.
1.2. The mysteries of moral perception; and some other mysteries
The claim that we can make out a credible notion of moral perception will need careful defence. Contemporary philosophy is not short of moral theorists who find talk of moral perception entirely mysterious. A common first gambit (see e.g. Blackburn 1993: 160) is to ask “If moral perception is really perception, then what is its organ?” The force of the question is obvious; the dissection-table reveals no parts of the human body that are devoted to moral perception, as there are parts devoted to visual and to auditory perception. Writers like Butler apparently mean to reject this question by talking of conscience as a faculty—faculties being things that presumably do not always correspond directly to particular organs. Aristotle too is apparently rejecting the question by his denial, in my first epigraph, that the aisthesis involved in phronesis is sensory perception, the perception of the special senses (oukh hê tôn idiôn).
Behind this first and perhaps relatively resistible challenge lurk other challenges that a moral realist may find harder to repel. The moral irrealist—J.L.Mackie and Simon Blackburn are, in different ways, modern paradigms of the type—sees mysteries both about the realist’s moral epistemology, and about his moral metaphysics. The irrealist is dubious not only about how we could possibly know about moral properties, but also about what moral properties could possibly be (Blackburn 1984: 182):
[An irrealist ethical theory] asks no more than this: a natural world[5], and patterns of reaction to it. By contrast a theory assimilating moral understanding to perception demands more of the world. Perception is a causal process; we perceive those features of things which are responsible for our experiences. It is uneconomical to postulate both a feature of things (the values they have) and a mechanism (intuition) by which we are happily aware of it.
Where would moral perceptions fit into a scientific world-picture? How would they relate to the other properties that science tells us about? If moral perceptions are really perceptions, how can they also have the “magnetism”, the power of motivational attraction, that they are supposed to have? Then there are the well-known phenomena about cultural differences that prompt cultural relativism. If different societies all claim to be perceiving moral properties, and yet perceive them so differently, that casts doubt (according to the moral irrealist) not only on the idea that some one of these societies is seeing things right, but on the idea that any of them are.
One possible response is surrender: we could simply agree that these challenges cannot be met. If there is a phenomenon of moral perception or observation—and moral sceptics need not deny that there is such a phenomenon—then it must be a large-scale illusion. So some—the “error theorists”—have concluded: Mackie 1977, Harman 1977, Ruse 1993, Joyce 2002, Carter 2004[6]. More recently a new variant of error theory, moral fictionalism, has developed the thought that ethics may be one of a number of discourses that are admitted to be systematically false, without this being in any way an objection to them, since it is not their truth that makes them useful to us (Kalderon 2005).
Error theory commits us to the view that every first-order moral claim is strictly false—and so, for instance, to the view that it’s false to say that Hitler was evil, or that torturing babies is wrong. Such a result inevitably creates an alienating tension between what the error-theorist believes qua error-theorist, and what (one hopes) she believes qua morally decent person (5.3). It also prompts a subtler objection to error theory, which Crispin Wright 1995 has urged. There self-evidently is some distinction between “Hitler was an evil man” and “Hitler was a saint” that we may reasonably call a distinction between truth and falsehood. So it is pointless to deny, with the error-theorist, that moral discourse sustains some sort of truth-predicate. The interesting question is not about that, but about whether the truth-predicate of moral discourse is a realist one. In other words, expressivists locate the right question, and error-theorists do not—except perhaps for fictionalists, who can treat “true” and “false” as “true-in-the-moral-fiction” and “false-in-the-moral-fiction” respectively. (But while this gives the fictionalists some account of what truth and falsity might be in ethics, it does not seem to equip them against the charge of alienation. The idea that “Hitler was evil” is on a par, as to its truth-aptness, with “Hamlet was Danish”, seems entirely unattractive.)
Another response to the irrealist’s challenge is to redefine what counts as meeting it. Maybe you can do enough to count as meeting it without accepting any form of error theory, but without going all the way to moral realism either. This is the aim of those “projectivists” who, like Simon Blackburn, Alan Gibbard, Mark Timmons, Terry Horgan, and others, wish to show that moral properties are intelligible as no more (and also no less) than projections of our sentiments, our reactive dispositions, onto the world. Those who take this line hope to show how (to home in on the usual contrast) scientific talk and moral talk relate to each other, without denying either the basic intuition of the irrealist—that moral discourse is not rooted in the fundamental reality of the world in the same way as scientific discourse is; nor the basic intuition of the user of moral discourse—that moral discourse admits of truth and falsity. “Quasi-realism is most easily thought of as the enterprise of showing why projectivism needs no truck with an error theory” (Blackburn 1998: 175).
One popular way for quasi-realists to accommodate the idea of truth and falsity in moral discourse is by adopting minimalism about truth—the view, roughly, that all a discourse needs to be truth-apt is a way of assigning “true” and “false” to its sentences that allows the discourse to be logically well-regimented. At the same time, they can accommodate the irrealist’s basic intuition by saying that, while moral discourse is truth-apt, there are different reasons why a discourse can be truth-apt. Scientific discourse is truth-apt because it is logically well-regimented and describes the world; moral discourse is truth-apt simply because it is logically well-regimented, even though it doesn’t describe the world.[7] Thus the quasi-realist aims to avoid error theory and other strong forms of moral irrealism, by keeping hold of the idea that moral discourse is truth-apt; while simultaneously keeping the contrasts between the different ways in which various discourses can be truth-apt, and so also avoiding quietism—the view that just any old discourse you like can be just as truth-apt as any other, so that there is really no difference between the kinds of truth you get in ethics and in science (see Smith 1994b).
This sounds an attractive combination. Once we see the difference between the claim that a discourse is minimally truth-apt, and the various possible claims about what makes it minimally truth-apt, we will not be inclined to say that this expressivist position fails straight away because it both needs there to be a contrast between scientific and moral discourse, and needs there not to be such a contrast. The position does indeed have both needs, but has them in different places: what it needs is no contrast as to truth-aptness, but contrast as to grounds of truth-aptness.
The position raises a number of deep and interesting problems, some of which will be focal to my entire discussion up to chapter 4 of this book. One of these problems—to be investigated in chapter 2—is the question whether an ethical discourse that wears its expressivist nature on its sleeve can in fact be even minimally truth-apt. (If an ethical discourse is only covertly expressivist, then of course there need be no difficulty about whether it is truth-apt. But there will be this difficulty about the expressivism that lurks beneath the surface of the discourse, and a further difficulty, of the sort that troubles error theory, about the relation between the surface and what lies beneath it.)