Irwin, Terence H. (2000). Ethics as an inexact science: Aristotle's ambitions for moral theory. In Brad Hooker & Margaret Olivia Little (eds.), Moral particularism. New York: Oxford University Press. pp. 100-29.
7. Particulars and Inexactness
In Book II Aristotle tells us more about the practical aims of ethics, and adds a new claim about inexactness. Before presenting his general account of virtue, he reminds us that ethical inquiry aims at practice rather than theory, and that we must be satisfied with inexact accounts (1103b26 1104a5). He adds that further inexactness results when we try to deal with particulars:
And when our general account is so inexact, the account about particular cases is all the more inexact. For these fall under no craft or profession, and the agents themselves must consider in each case what the opportune action is, as doctors and navigators do. The account we offer, then, in our present inquiry is of this inexact sort; still, we must try to offer help. (1104a5-11)
The parallel with medicine and navigation shows how we must understand the claim that 'every account of the actions we must do 20° has to be stated in outline, not exactly'. It does not exclude unrestricted generalizations from ethics. Aristotle suggests only that medicine and navigation must sometimes resort to inexact generalizations, to be applied by trained judge ment in particular cases, because they seek to give practical advice. The same is true of ethics.
This warning about the inexactness of ethical generalizations comes between Aristotle's claim that virtuous action is in accordance with right reason and his claim that virtue is a mean. Indeed, he begins his argument for the latter claim immediately after saying that the moral philosopher must 'try to help' despite the inexactness of ethics (1104a10-11). The doc trine of the mean certainly does not provide a precise quantitative guide to finding the virtues, since we cannot eliminate 'at the right time, to the right people, and so on, from accounts of virtue as an intermediate state. The doctrine is not at all empty, however; it conflicts with the view that virtue is simply continence, and with the view that virtue requires the elimination of all non-rational desires. It is difficult to see how such a doctrine could be represented as a summary of what virtuous people have found in particu lar cases; and Aristotle does not represent it in this way.
Having stated the doctrine of the mean, Aristotle considers how it might be specified in suitable detail. 21 In order to give more detail, he specifies the mean with reference to different kinds of feelings. The different means are specific virtues of character, of which he promises a more 'exact' account later (1107b14-16).
Aristotle does not suggest that generalizations are useless; for here he introduces the accounts of the specific virtues of character, which evidently include many generalizations about the characteristics of different virtuous and vicious people. He is coming closer to 'particulars' by providing more specific generalizations. Instead of simply saying that virtue consists in a certain kind of mean, he adds that bravery is this kind of mean, temperance is that kind, and so on. When Aristotle tells us to attend to particulars, he does not abandon generalizations; he describes the generalizations we should look for. Nothing in these remarks comes close to acceptance of particularism.
Some of Aristotle's directions for approaching the mean give general advice about compensating for the extreme that we are more prone to, though he qualifies the advice with a warning about inexactness (1109a30—b23). We need to recognize inexactness not because we regard generalizations simply as summaries, but because we recognize them as giv ing normative guidance. Once we have accepted the doctrine of the mean, and seen the desirability of trying to achieve the mean, we see that we need to attend to particular cases; but we do not know what particular cases we ought to attend to until we have accepted the doctrine of the mean.
The distinction between the statement of the doctrine of the mean and its application to particular cases corresponds to Aristotle's distinction between the 'general account' and 'the account about particular cases. The claim 'Virtue results in happiness' is a generalization with exceptions, but it does not try to give us advice about any specific circumstances we may face. The closer we come to giving advice about very specific circum stances, the more inexact our advice must be.
This second sort of inexactness cannot result simply from the fact that we mention a particular situation. However particular it might be, we can still advise agents to do the finest action open to them, or to do what is required, all things considered, by the virtues. This advice may, for all Aristotle says, be entirely correct in all circumstances, but it is not of much practical use in helping us to decide precisely what to do. To say that months in the year have fewer than thirty-five days is an unrestrictedly true generalization, but it is not much practical use if we want to pay our bills on the last day of each month and we need to know which day that is.
For this practical purpose the generalization that the months in the year usually have thirty-one days is better than the unrestricted generalization we began with. But it seems that it would be even more useful to formulate a new generalization that includes the appropriate qualifications, instead of merely suggesting, by the use of 'usually, that some qualifications are needed. The most useful generalization says that all the months have thirty one days, except for April, June . . . and so on. Similarly, knowing that Latin prepositions usually take the accusative is fairly useful, but it is more useful to know that they take the accusative, except for ab, ex, clam, coram . . . etc., which take the ablative. Once we have built in all the exceptions, we have replaced our inexact usual generalization with an exact generalization.
Why does Aristotle not advise us to do this in ethics? He suggests his answer when he discusses the cases in which deliberation is necessary. 22 He describes a series of increasing degrees of inexactness that require larger contributions from deliberation. His 'exact and self-sufficient' cases are most like my examples of the months and the Latin prepositions. In the other cases we have fewer and fewer simple rules that guide us effectively in particular cases. The last sentence quoted sums up the situations in which deliberation is needed.
What does Aristotle mean by saying that deliberation is needed when things are 'undefined'? He need not rely on any belief in some real indeterminacy in things. The claim that something is undefined may simply refer back to his remark that we are in doubt about what to do. We need deliberation in cases where the considerations are various and com plex enough to make it inappropriate to appeal to generalizations that are sufficient by themselves to tell us what to do. If I want to write the French word for a cat, and I know that this word is spelt `c-h-a-t', I need not delib erate further about what letters to write. When more considerations bear on the appropriateness of one or another action, it is less reasonable to appeal to generalizations that can be applied without further deliberation.
These remarks about deliberation suggest that choices in particular sit uations involve inexactness because we cannot expect to reach the right choices by appeal to exact generalizations that are immediately applicable, without further reflexion, to particular cases. This does not mean that gen eralizations are inappropriate, but that they must be qualified in certain ways, and that their limitations must be recognized. Recognition of the limits that we have mentioned falls far short of particularism.
8. Particulars and Perception
Some of Aristotle's remarks about particulars also mention perception. These may seem a more promising basis for attributing particularism to him.
After stating the doctrine of the mean, and listing the specific means constituting different virtues, Aristotle gives us the general advice to come as close as we can to the mean, by trying to avoid the extreme to which we are more inclined. He warns that we cannot expect this general advice to give precise guidance in every case. The difficulty in determining questions about particulars seems to result from two things: (a) the differences between different people, and the consequent differences in the training that each person needs to approach the mean; and (b) the different partic ular conditions that each person faces and the consequent difficulty in cov ering all the circumstances relevant to a decision about what is needed to approach the mean (1109b12-23). Since we cannot readily define these points in a general account, Aristotle concludes that 'the judgement about them depends on perception'.
Aristotle suggests a similar role for perception in his remarks about the mean constituted by gentleness. When we try to distinguish the gentle (praos) person both from the people who are too prone to anger and from the people who are not angry when they should be, the boundaries are dif ficult to mark in general terms. Perception is needed to supply the defi ciency of general accounts. We make the relevant judgement in particular cases by using perception (1126a31—b4).
The connection between particulars and perception is marked again when Aristotle denies that particulars are objects of deliberation. Perception is introduced here both for non-evaluative judgments (`this is a loaf') and for evaluative judgments (`it is cooked the right amount') (1112b34-1113a2).
These remarks do not imply acceptance of particularism, for two inde pendent reasons: (1) Aristotle does not rely on a claim that fully qualified generalizations are in principle impossible. (2) Even if he did believe they are impossible, he would not be committed to particularism, because he does not attribute to perception of particulars the sort of priority that is necessary for particularism.
To explain the first reason, I refer back to Aristotle's views about the role of perception in relation to deliberation. If we are not content with usual rules, but we try to qualify them so much that they provide definite advice for every particular case that we meet, we will have to provide numerous qualifications. If these qualifications are extremely numerous, it may be better to equip the learner with some other means of finding the right answer. If generalizations become at all complicated, with many qualifica tions, the different qualifications will refer to different aspects of a situa tion, and the agent applying the generalization will have to recognize these different aspects. If agents equipped with unqualified generalizations and the capacity to recognize ethically relevant aspects of particular cases can reach the right answer, it is better not to burden them with extremely com plicated qualified generalizations.
One of Aristotle's examples may help to illustrate this point. The advice to bake bread until it looks done relies on some trained capacity to estimate when it looks done. Conceivably we could avoid reliance on any such capacity by providing the baker with a thermometer to insert carefully in the bread, a chart to find the appropriate temperature for different kinds of bread, and a colour chart to determine the colour that it ought to be when it should be removed from the oven. Using all this apparatus is a nuisance, however; baking is normally simpler and more efficient if we simply learn to recognize when bread looks done.
Similarly, we can see why an agent needs some capacity to recognize several ethically relevant aspects of a situation, if we consider some of Aristotle's own contributions to casuistry. In Book ix Aristotle refers back to his remark in Book II that it is difficult to give definite answers about particular cases. Still, he makes some suggestions (1164b25-1165a12). He urges, for instance, that usually we ought to pay our debts, but this general rule has exceptions, when something especially fine or necessary is involved.
Aristotle has good reasons for not spelling out all the relevant exceptions and for not even trying to incorporate them into modified rules. Suppose that we could formulate general rules, with the appropriate qualifications built into them, to cover all these cases. In order to apply these rules, agents would have to recognize that a situation involves (say) an obligation to one's father, an obligation to someone who has already paid one's own ran som, an obligation to pay a debt, a request to lend money in return for hav ing been given a loan, and a permission to refrain from lending in cases where one will probably lose one's money. If I can recognize these different aspects of situations, and I also have some idea of which of these consider ations are more important, I will probably do quite well at getting the answers that would be given by the fully qualified generalizations. I need not be able, however, to formulate rules that express my views about the relevant considerations.
Aristotle does not say why the practical purposes of ethics are better served if we learn usual generalizations than they would be if we set out to learn fully qualified generalizations. But his view is reasonable. Fully qual ified generalizations would be so complicated that they would be difficult to learn and difficult to apply; moreover, the skills that we would need in order to apply them, once we had learnt them, will lead us to the right answers without having learnt the fully qualified generalizations. Instead of trying to learn fully qualified generalizations, it is better to try to learn to recognize and to compare the considerations that ought to guide us in assessing the different claims of different usual generalizations.
These points support my first reason for rejecting a particularist inter pretation. Aristotle's emphasis on perception is intelligible whether or not he believes that fully qualified generalizations are possible.
Even if this first reason is mistaken, the second reason still stands. Aristotle's casuistical discussions do not suggest that we can resolve the dif ficulties raised by these situations, unless we understand some generaliza tions. Generally we must follow the rule of reciprocity, because reciprocity takes priority over disposing of our resources as we feel inclined; payment of a debt, therefore, takes priority over a generous impulse to give a present to a friend (1164b30-3). But the priority of reciprocity over generous impulses does not imply the priority of reciprocity over everything. If we also recognize that special obligations to our parents are prior to the gen eral obligation of reciprocity, we will acknowledge the proper place for the rule of reciprocity. To acknowledge this, we must already have understood the moral basis for the presumption in favour of reciprocity. We are not really violating the rule of reciprocity; for the rule holds only on a pre sumption that (in this case) is not satisfied.
This use of the basis of a rule to explain exceptions, or apparent excep tions, is especially clear in Aristotle's discussion of equity. He suggests that the provisions of written law need to be violated in some cases, but that violations do not violate the point of the law (1137b11-32). Once we see what the legislator aimed at in formulating the general rule, we can see that this very same aim requires violation of the rule in some cases.
Ethical rules are different from the laws that need to be violated in order to fulfil their aim; for, in contrast to laws, ethical rules acknowledge their limitations by being expressed as usual generalizations. If we fail to do what they tell us to do usually, we do not violate them, if the principles underly ing them justify our claim that this is not one of the usual cases.
Consideration of Aristotle's examples (friendship and equity) do not suggest that exceptions to general rules are irreducibly particular. We might argue, for instance, that a general rule about the priority of obligation to one's parents over the obligation of reciprocity explains exceptions to the rule of reciprocity. If Aristotle were a particularist, he would have to reject this explanation of the exceptions. Since he does not express any doubts about it, his treatment of these examples does not suggest that he is a particularist.
In these cases, then, even though general rules do not prescribe precisely what we ought to do in a particular situation, their normative guidance contributes essentially to a correct decision about what to do. Aristotle's conception of their role, therefore, conflicts with particularism, which asserts the priority of perception over general rules. The only way to defend a particularist interpretation of Aristotle's view in these cases is to argue that the generalizations that are brought to bear on particular cases are themselves simply summaries of what we have found by perception in par ticular cases. But we have seen no reason to believe that Aristotle accepts this account of ethical generalizations.