by Daniel A. Kaufman
When speaking of morals, the overwhelming view among philosophers is that they are criterial. Something – some action or a state of affairs – is right or wrong, obligatory or prohibited, because it satisfies certain criteria; criteria that typically come from a larger moral theory. Thus, for a Utilitarian, actions are right or wrong, obligatory or prohibited, because they either succeed in or fail in meeting the criterion of promoting the general welfare. And for a Kantian, actions are right or wrong, obligatory or prohibited, because they succeed or fail in meeting the criterion of rational universalizability. Indeed, the very idea of a moral theory is defined, in part, in terms of identifying the general characteristics of moral rightness and wrongness, obligation and prohibition and this is tantamount to providing a criterion for the application of the words ‘right’, ‘wrong’, ‘obligation’, and ‘prohibition’.
That morals are criterial also seems embedded in the very way we speak about them; specifically, in the giving of reasons for our ethical judgments and actions. People ask us why they ought to do something or why we have a certain moral attitude towards a given state of affairs, and in providing what we think is a moral answer, we offer reasons that mention various features of the action or of the state of affairs in question. In doing so, we are effectively appealing to criteria.
That this view is so prevalent among professional philosophers is interesting, in part because the criterial view of terms and their application has faced a number of very tough challenges, over the course of the last century, to the point that it is largely in tatters.
Wittgenstein showed that the substantial and open-ended heterogeneity that one finds in the extensions of common terms like ‘game’ and ‘art’ renders it impossible that they might be applied by appeal to criteria. (1) Keith Donnellan and Hilary Putnam argued that natural kind terms like ‘mammal’ or ‘lemon’ cannot be criterial, in light of atypical and borderline cases. (2) Saul Kripke maintained that proper names cannot be criterial, because the person to whom the name refers might not have had the characteristics associated with him, and in speaking of that possibility, we would still be speaking about him and not someone else. (3) And philosophers like Frank Sibley have pointed out that aesthetic terms like ‘delicate’ and ‘vibrant’ cannot be applied by appeal to non-aesthetic criteria, because the very same characteristics may be found in something to which these terms do not apply or even worse, to which the opposite terms may apply. The very same non-aesthetic characteristics that make one thing “delicate,” for example, may make another “insipid.” (4)
One might wonder, then, about the prevalence of the criterial view in moral philosophy. After all, if scientific terms like ‘mammal’ and common words like ‘game’ cannot be applied by way of criteria, what chance is there that terms like ‘moral’ or ‘immoral’ or ‘obligatory’ can? This seems a fair question to ask, before we even get to the further question of the merits or faults of any particular moral theory. And when we do turn to moral theories, what we find is a mess of wildly different views, many of them mutually exclusive, invoked and applied in a haphazard, inconsistent manner, often by the same person, in a single day. An encounter in the morning may provoke me to think and act with regard to the general welfare, and another, in the afternoon, may elicit concern for another person’s dignity and rights. Of course it is possible that our words, ‘right’, ‘wrong’, ‘obligatory’ and the like have different meanings on each of these occasions, so that when I say “this is the right thing to do” in the morning and “making you do that wouldn’t be right” in the afternoon, I mean entirely different things by ‘right’, but the simpler and in my view, much more plausible way of interpreting these facts is that our applications of moral terms are not criterial.
Of course, not everyone who has worked in moral philosophy over the last century has embraced the criterial view. The largely ignored and in my view, criminally underrated Intuitionists – notably, H.A. Prichard and W.D. Ross – rejected the idea that our moral judgments and actions are the result of applying criteria. “The sense of obligation to do, or of the rightness of, an action of a particular kind is absolutely underivative or immediate,” Prichard wrote. “[W]e do not come to appreciate an obligation by an argument, i.e. by a process of non-moral thinking…,” which means that we do not determine whether the word ‘right’ or ‘wrong’ applies to an action, by seeing if the action meets a set of required, “moral-making” criteria. (5) We know this, because whenever we try to derive an obligation, say from the claim that some action is intrinsically good or has a good result, we wind up with an infinite regress (for such an inference to go through, we must presume that what’s good ought to be the case, which is simply to invoke another obligation). And when we consider what we actually do when we wonder whether we were right in thinking something is a duty, we don’t consult a theory or look to criteria, but rather place ourselves back in the situation and see whether the sense of obligation arises again. (6)
It is important to be clear that Intuitionists like Prichard are not suggesting that perceiving some fact about an action or a state of affairs cannot get a person to the point that they feel obligated to act in a certain way or find some action or state of affairs to be wrong. Seeing a person suffering from cancer may lead to my feeling obligated to donate to cancer research and I might feel that I ought to reciprocate, when I remember that my neighbor has done me a kindness. But this is not the criterial view of morals, according to which it follows from the fact that an action or state of affairs has certain characteristics that the word ‘right’ or ‘wrong’ applies to it and therefore, that I ought to do or accept it or not.
Another non-criterial approach to obligation, associated with the Wittgensteinian philosopher Cora Diamond, sees obligations as often arising from the application of concepts that are, themselves, morally “thick,” meaning that their moral content is implicit, rather than inferred, on the basis of criteria. We are all familiar with morally thick concepts, as they are what one typically finds in moral codes like the Ten Commandments. ‘Stealing’, ‘Murder’, and the like are morally thick, insofar as they have both descriptive and moral content. To steal is to take someone’s property wrongfully, and to murder is to kill someone wrongfully, and thus, the statements “Stealing is wrong” and “Murder is wrong” are analytic and necessarily true, whereas their morally neutral, “thin” counterparts, “Taking someone’s property is wrong” and “Killing someone is wrong” are synthetic and only contingently true (or false).
Diamond maintains that concepts like ‘person’, ‘neighbor’, ‘friend’, ‘pet’, and others are also morally thick in this way and that looking at things from this perspective far better explains our behavior than the criterial approach. In her famous paper, “Eating Meat and Eating People,” Diamond dismantles the criterial approach to our treatment of both people and non-human animals, and specifically with regard to so-called “ethical eating” practices, like vegetarianism and veganism. (7)
Ethical vegans want to say that we shouldn’t eat non-human animals for the same reason that we shouldn’t eat people: we satisfy the same moral criteria. The reason why we shouldn’t eat people – so the story goes — is because they have a certain morally relevant characteristic, namely the capacity for suffering, and since non-human animals also have this characteristic, we shouldn’t eat them either.
As Diamond points out, however, if this is the reason why we shouldn’t eat people, then there is no reason not to eat our dead, so long as the person in question was not unjustly killed and so long as the meat is good to eat. Likewise, there is no reason why we shouldn’t eat amputated limbs, so long as the same caveats hold. And yet, most of us in the modern world don’t think that we should do either of these things, other than in the direst of circumstances (like if we find ourselves in a Donner Party-style scenario), which means that whatever the basis of the prohibition against eating people, it does not lie in the fact that they meet certain criteria. Diamond maintains, instead, that it is because the concept ‘people’ is morally thick and as a result, people are not things to eat.
The same is true on the non-human animal side of things. If the reason why we shouldn’t eat beef or chicken is because animals like these are capable of suffering, then there is no reason why a vegan shouldn’t eat an animal that has died of natural causes or been struck by lightning or lost a limb, in an accident. That vegans eschew all meat eating shows, therefore, that whatever the reason, it is not because cows or chickens have certain characteristics and thereby meet certain criteria. And consider, even, the non-vegan or non-vegetarian person. I might be more than willing to eat Gaegogi (dog), in a restaurant in Seoul, while at the same time being appalled at even the suggestion that I might eat my pet Bichon Frise. Is this because the dog used to make my Bosintang lacked some morally relevant characteristic that my dog possesses? Or is it like this: I think of my Bichon Frise as a pet; ‘pet’ is a morally thick concept; and pets are not things to eat.
One might wonder how certain things manage to fall under morally thick concepts while others don’t. Certainly, there will be a story in each case – one that explains, for example, why this Bichon Frise and not some other dog became my pet – but there may be as many such stories as there are cases, and it seems highly unlikely that there will be some general principle, unifying them all, which could serve as some kind of meta-criterion, by which the criterial view can be saved.
The criterial view of ethics is part of a broader rationalism in philosophy, about which I have been quite critical. (8) It renders ethics procedural and thus transparent and scrutable, and this is both comforting and flattering to us. In contrast, the anti-criterial approaches to ethics that I have described treats ethics as a matter of perception, feeling, conception, and naming, which are not procedural and which render morals opaque, ultimately unreasoned, and thus, somewhat inscrutable. This is far less comforting and flattering, but it does have one tremendous advantage over its criterial competitor – it may actually be true.
- Ludwig Wittgenstein, Philosophical Investigations ( 1953), §66.
- Keith Donnellan, “Necessity and Criteria.” (1962)
Hilary Putnam, “Is Semantics Possible?” (1970 )
- Frank Sibley, “Aesthetic Concepts.” (1959)
- Saul Kripke, Naming and Necessity (1970)
- H.A. Prichard, “Does Moral Philosophy Rest on a Mistake?” (1912), pp. 7 & 9.
- Ibid., pp. 16-17.
- Cora Diamond, “Eating Meat and Eating People.” (1978)
- Daniel A. Kaufman, “Excessive Reason”; “Rationalism in Mainline Philosophy: A Discussion.”