Moral by definition? (Some slightly technical philosophy.)

Introduction

The recent TED talk by Sam Harris brings important metaethical issues into the popular arena. Is there a way to establish the objectivity of morality, and in particular the objective bindingness of utilitarian morality?

I'm afraid not. A good place to start with this is Peter Singer's concluding chapter of Practical Ethics, immediately prior to an appendix about his unfortunate experiences of being silenced in Germany. Here, Singer presents a searching inquiry into the question of why we should act "morally" in a sense that he defines (which amounts to acting like a preference utilitarian). This chapter, which is actually entitled, "Why Act Morally?", evidences how difficult - nay, impossible - it is to establish that anyone has a reason to be moral, or act morally in the stipulated sense.

Morality by tautology

Singer makes clear that he is not using the word "morally" and its cognates, such as "moral", to refer to action in accordance with whatever principles we happen to find overridingly important. This must be stressed, because it means that he cannot use the argument: acting morally is acting is acting in accordance with the principles that are overidingily important; we ought to act in accordance with the principles that are overridingly important; therefore we ought to act morally. That is all just tautological, and gets us nowhere.

Nor, evidently, do these words ("morality", etc.) refer to actions that we might have best reason for, all things considered, in which case a similarly tautological argument could show that we ought to act morally. Again, that gets us nowhere.

Such definitions do not tell us what actions are those we have best reason for, all things considered, or what actions are in accordance with principles that are overridingly important. There seems no room for doubt that we should act morally in these senses of the word - it's pretty much true by definition - but that tells us nothing at all about the ways in which we actually should act.

Morality as impartiality

Rather, Singer defines "morality" as action that is impartial with respect to the interests of all affected sentient beings. To act morally in this sense, then, I must first accept that my own interests do not count more than the interests of any other being that actually has interests. I then respond by acting in a way that maximises - or, perhaps, is objectively likely to maximise - the summed interests of all. I can then be said, according to Singer's stipulated usage, to be acting in accordance with "the ethical point of view".

To illustrate Singer's conception of moral action, if I wish to act in accordance with the so-called ethical point of view, and if I see that Φ-ing (say, selling my house and donating the proceeds to Community Aid Abroad) is the unique way for me to do so in my current circumstances, then it can be said that Φ-ing is what I ought to do.

Notice, however, that I expressed this as a hypothetical imperative. It is what I have reason to do if I already wish to act from the ethical point of view. At this stage, no good reason (some kind of non-moral, or pre-moral, "ought") has been given as to why I should, or might, wish to act in accordance with the ethical point of view. It's no good saying that my interests are not objectively more important than anyone else's. So what? They are still my interests, and I may desire to further them. How have I made any error if I set out to do so? My desire to further my own interests is not the sort of thing that can entail any truth-claims that might be in error. I simply have desires ... and they motivate me.

Thus, Singer's question is actually a question about what reasons (of a non-moral, or pre-moral, kind) there are to adopt the ethical point of view or to act morally, as he defines these expressions.

Universalisation

Singer insists that his definition of morality is more than an arbitrary stipulation. The supporting claim is that morality, even as understood before we get to his theoretical account, takes a universal point of view. This claim, in turn, is supported first by an argument that when we justify our acts in a way that we recognise as "moral" or "ethical", and not just any justification will do. For example, it is not a moral justification if I seek to explain my commission of a murder in terms of its expediency for pursuing, like Macbeth, my "vaulting ambition".

Note, though, that this immediately creates a problem that Singer appears to pass over: surely I can adequately explain my decision to study philosophy simply by stating that (1) I am interested in the subject (which is not, by itself, a justification) and adding (2) studying philosophy breaches no moral constraint that applies to me. Singer's example of Macbeth does not rule out a conception of morality, such as favoured by Kantians, that allows for a wide range of permissible - but not obligatory - actions within certain deontic constraints.

Singer then looks to the historical teachings of a range philosophers and moralists, who have all agreed "that ethics is in some sense universal." However, it is not obvious that all these teachers have considered morality to be universal in the same sense. While Jesus of Nazareth is reported as teaching that we should love our neighbours as ourselves, which may mean giving their interests equal weight, other interpretations of morality's universal component might be much weaker, such as the idea that the same moral constraints, if any, apply to everyone, given relevantly similar circumstances.

J.L. Mackie has suggested that a system of norms need not reach such an advanced level of universalisation as that described by Singer to conform to the concept of a moral system, while Neil Levy also refers to a more basic level when he describes the social need for rules that ensure we are responsive to each other's actions in predictable ways. Elsewhere, Steven Pinker notes that no one could argue with pragmatic success that moral restrictions in a society should apply to everyone but not herself, so the price of having any sort of moral code, with its benefits, is that it also apply to oneself. Similarly, Sober and Wilson emphasise that moral principles are, like other principles, general in application. They specify "general criteria or relevant considerations for deciding what one ought to do." In short, the idea is widely acknowledged in the literature.

Pace Singer, all these considerations, moral principles, rules, or maxims may take many different forms: for example they may prescribe that we act in certain ways to people within the group, or to all humans, or to all sentient beings. They need be universalisable only in the sense of saying that "anyone with such-and-such feature is to be treated thusly." I.e., they involve rules, principles, criteria, and so on, that are of general application.

From his claim that morality has a universal aspect "in some sense", Singer moves to the position that acting morally involves taking into account everyone's interests and give equal weight to them. Hence, it involves giving no more weight to our own interests than those of anyone else.

However, this is a non sequitur. Nothing forces us to adopt such a specific, and arguably extreme, sense of "universal" and related words as Singer actually uses. There are forms of universality that do not require giving equal weight to all interests in all circumstances, but merely that the rules be of general application.

Most obviously, a system such as that favoured by Kant can allow for a wide range of merely permissible actions that might be chosen on some other basis than the one recommended by Singer, perhaps prudential or perhaps eroscentric (favouring those one loves). No such system may be correct, but that is not the issue; the point is, it may well be recognisable as a moral system in our ordinary understanding of what a moral system is like.

Singer himself is aware that the universal aspect of morality can be described in a thin, formal way involving general applicability of norms of conduct, and that this would embrace many theories of correct action. Conversely, as he mentions, there is a danger of giving the universal aspect so much substance as to smuggle in "our own ethical beliefs into our definition of the ethical". Unfortunately, he appears to err in the latter way. As a result, we should be aware that his definition of what it is to "act morally" is not the only one available. But all this, and other considerations adduced in the last few paragraphs, can establish is that we are not compelled to adopt Singer's terminology.

We are not compelled - on danger of being illogical or making a mistake about the world - to define morality in terms of impartially maximising preference-atisfaction while also thinking of acting morally as acting in the way that we ought to act in all the circumstances (or acting in accordance with principles of overriding importance).

Conclusion

We can define "moral", "morality", and related words such as "ethical" and "ethically" and so on, however we choose - but choose we must! If we define these words in terms of how we ought to act in all the circumstances (or something similar), then, sure enough, the definition entails that we ought to act morally, but it cannot tell us what this amounts to in practical detail. The latter is still an open question.

Conversely, if we define the words in Singer's way, we know what is involved, but it becomes an open question whether we ought, all things considered, to act morally (rather than, say, egocentrically, or eroscentrically, or in accordance with some compromise approach).

Utilitarians can't have it both ways. If they define morality in terms of acting in the way that we ought to act in all the circumstances, then they still need to demonstrate that this involves acting like a utilitarian. They cannot succeed in convincing us of this if our most basic (or highest order) desires are to the contrary (e.g. if they contain egocentric or eroscentric elements).

Conversely, if they define acting morally in terms of acting like a utilitarian, then they still need to demonstrate that this is the way that we always ought to act in all the circumstances. Again, they cannot succeed in convincing us of this if our most basic (or highest order) desires are to the contrary (e.g if they contain egocentric or eroscentric elements).

Thus, even if we defined such words as "morality" or "ethics" in accordance with Singer's usage, this could not be relied upon to derive an acceptable system of action-guiding norms, as R.M. Hare arguably attempts to do in such works as Freedom and Reason. As Simon Blackburn has pointed out, if a word such as "ethically" refers to reasoning of this kind, "we may still prefer and campaign for other ways of reasoning." I.e., if we think of the word "ethically" in this way, we may quite rationally decide not to act ethically! We are not making a mistake if we so decide.

There is no prospect of defining words such as "moral" and "morality" in such a way as to compel us to act as a utilitarian would wish, on pain of making a mistake about the world or being irrational (at least in a sense of "irrational" that we need care about, since this word can also be defined in more than one way). As Singer himself ultimately concedes in Practical Ethics, it is not possible to compel someone to accept utilitarianism, or "the ethical point of view", by anything like a brute exercise in logic. His own eventual approach is to try to sell us the attractions (the sense of meaning it can give us, and so on) of living a "moral" life.

This entails that we do not necessarily have reasons to be moral. Morality can be objective if we define it in a way that refers to something naturalistic and does not include its power to give reasons (as with "what you ought to do in all the circumstances"). But it is then not necessarily mistaken to act immorally, or even to reject morality. You can't retain both of these aspects (the strict objectivity of morality and the rational requirement to act morally) at once.

A better approach to morality is to point out that almost all of us wish, all things being equal, that other lives go well and that suffering be ameliorated. Accordingly, from within our own value systems we (almost) all have reasons to act in ways that have these effects. At the same time, there is always the possibility of a clash with other values, such as our own survival and the happiness of loved ones. At least for beings like us, who are not omnipotent and cannot deal benevolently with all interests at once, morality is not about perfect altruism. It is about constraining ourselves and living within the constraints, particularly in such respects as accepting the strong prima facie requirement to act honestly and non-violently.

This is the kind of thing that actual moral systems tend to demand, and it is the kind of demand that others around us can reasonably expect us to accept. People who want us to support a particular system are stuck with appealing to the structures of desires that we actually have, though they can, of course, ask us to subject those desires to rational reflection (they can ask us to consider which of our desires are really most important to us). They cannot, however, compel us to accept and abide by a system of morality by the clever use of definitions.

In this field, definitions always have their price.