Show Summary Details

Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE ( © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

date: 27 February 2021

Aristotle on Practical Reason

Abstract and Keywords

For Aristotle,phronēsis, the excellence of the practical intellect, is two-fold, consisting of a true conception of the end to be achieved by action and correct deliberation about the means to achieve that end. Three accounts have been given as to how that true conception of the end is acquired: i) by virtue of character, ii) by dialectic, i.e. critical reasoning concerning authoritative beliefs, and iii) by induction from data of experience. Virtue of character is the proper responsiveness of the appetitive element in the soul to reason; it is itself a rational state, presupposing a prior grasp of the end by the intellect. Dialectic and experience are each required for the attainment of that grasp, the role of the former being apparently to formulate more or less indeterminate principles that it is the task of moral experience to make determinate.

Keywords: phronēsis, intellect, theoretical, practical, excellence, virtue, character, dialectic, experience

I have discussed this topic in Taylor (2008), with particular emphasis on the question of how the intellect is related to the agent’s long-term goals. Does the intellect itself lay down the goals that are to be pursued, or is it restricted to working out means to the achievement of goals that it takes as given by some non-intellectual faculty, for example, by the passions, as in Hume’s theory? I here present a revised version of that discussion, which takes account of some work that has appeared subsequently to it.

1. The task of practical reason in general

The first point to note is that there is not just one activity or kind of activity that is the practical exercise of the intellect, but many: just as there are many theoretical sciences and kinds of sciences, so there are many practical enquiries and kinds thereof. We have first the distinction (EN VI.4)1 between productive and practical crafts, the former (e.g., housebuilding and medicine) directed to the production of things separate from their own exercise, the latter (e.g., musical performance) directed only toward their own excellent performance. The product of a productive craft need not be a substance, such as a house or pot, but may be a state (e.g., the healthy state of a patient), but in either case the craftsman aims to produce something beyond his own excellent performance (indeed, excellent performance of a productive craft presupposes such an external object, being defined as the kind of performance that generally produces the desired object). In the case of a practical craft, however, “the excellent practice is itself the goal” (1140b7). Even more central to our enquiry is a distinction relating specifically to practical crafts. In the opening chapter of the Nicomachean Ethics Aristotle points out that different crafts, both practical and productive, may be subordinated to one another in a hierarchy, the goals of the more specific crafts being subsumed under the goals of the more general, e.g., various ancillary crafts such as those of bridle-making are subsumed under the craft of horsemanship, which is in turn subsumed under, e.g., the art of warfare. The supreme practical craft, he argues in chapter 2, is that of politics, which subsumes the ends of all other crafts under the highest end, that of promoting the good life for all the members of the community. Excellent practice, then (eupraxia), which is the end of praxis, may be confined to a special area of life, or may be non-specific, that is, identical with the excellent organization of one’s life as a whole; the latter is the mark of the phronimos, who “is able to deliberate well about the things which are good and beneficial to him, not in a restricted sphere, e.g. with a view to health or strength, but such as promote living well as a whole” (1140a25–8). Here we see Aristotle implying, rather than stating in so many words, the distinction between the first-order pursuit of intellectual goals and the second-order activity of organizing the pursuit of those goals into a satisfactory life. The clearest indication that Aristotle has seen this distinction is the fact that he assigns a particular name, phronēsis, to the specific excellence of that second-order activity, distinguished alike from theoretical excellence (sophia) and from first-order practical excellence, i.e., skill at any particular craft.

We have, then, narrowed down our enquiry to focus on that activity of the intellect whose object is the securing for the individual of the overall good life (to eu zēn holōs 1140a28 = eudaimonia 1095a19-20), and whose specific excellence is phronēsis. In investigating that activity we shall, of course, have to take into account both those features that it shares with other manifestations of practical intelligence (the various sorts of poiēsis and praxis), and those that are peculiar to it.

In general, the task of the practical intellect, whether exercised in poiēsis or in praxis, is the initiation of change. The activity by which the intellect brings about change is deliberation (bouleusis), whose field of operation is things that are a) capable of change, b) within the agent’s power to affect. Yet at the same time the ergon, the specific function of the intellect as such, is to attain to truth (1139a28). Hence, in its action-initiating function, the practical intellect must still be seen as aiming at the truth; specifically, its object is practical truth, which Aristotle defines as “truth in agreement with right desire” (1139a26–31). The concept of practical truth is highly problematic, and detailed discussion is beyond the scope of this paper.2 I must be content to assert dogmatically that what Aristotle appears to mean is that if an agent is to reach a correct decision on any practical matter, two things must be the case: first, his/her practical beliefs must be true; and second, he/she must be motivated to act in accordance with those beliefs, i.e., in Aristotle’s terms, his/her desires must say the same thing as his/her intelligence (VI.2). Suppose my aim is to be healthy; then if I am to reach correct decisions on how to act, say in deciding what, how much, etc., to eat, I must have true beliefs, e.g., that too much animal fat is bad for you, and that such and such foods are high (low) in animal fat, and I must want to abstain from these foods I truly believe to be high in fat and to eat appropriate quantities of those I truly believe to be low in fat. The aim of the practical intellect, then, is to reach decisions to act; and decisions are reasoned judgements (prohairesis presupposes bouleusis 1112a15–16), expressive of desire. Aristotle’s formula “desiderative intellect or intellectual desire” seems happily to express that complex situation, which is also reflected by his insistence that the practical intellect both issues in action (via decision) and has truth as its object (in that decisions are reasoned judgements that involve truth as described in this paragraph).

This direction toward practical truth is common both to the departmental, first-order function of the practical intellect and to its general, second-order search for the good life; this is shown by the parallelism between the definition of technē, productive excellence, as “a productive disposition with a true logos” (i.e., a true conception of what is to be produced, 1140a9–10) and of phronēsis, overall practical excellence, as “a true practical disposition, with a logos, concerned with things good and bad for man” (1140b5–6). In its search for practical truth, i.e., in deliberation, the practical intellect operates under an important restriction, viz. that deliberation is not concerned with ends but with the choice of actions conducive to the attainment of ends already given. So a doctor doesn’t deliberate about whether to cure someone, or a statesman about whether to create eunomia; they take these as the aims of their respective undertakings, as already given, and busy themselves with the enquiry into how those aims are to be realized (112b11–16). We desire things, e.g., health or eudaimonia, as ends, and select the means to them, but it is inappropriate to speak of selecting those ends themselves (1111b26–9). We have to be cautious here. First, Aristotle does not say that the desire for a given end lacks propositional content; the theory of desire expressed in VI.2, MA 7 and De an. III.9–10 requires that the desire for, e.g., health is expressed in a thought such as “Health is something good.” Second, it is unspecified in EN III how determinate is the conception of the end that enters into the desire for it; the example of health suggests a fairly full conception of what health is, which can be a starting point for practical medical studies, whereas the example of eudaimonia is compatible with a thin conception, e.g., “I want a worthwhile life, but have no clear conception of what kind of life that would be.” Third, the account of deliberation is silent on the question of how the conception of the end and the belief that the end is something good are acquired: that account does not, therefore, exclude the possibility that that conception and/or belief are acquired as the result of some process of intellectual enquiry. All that Aristotle is committed to is the negative thesis that that conception and that belief are not the result of deliberation. And that negative thesis has the air of a tautology; for “deliberation” seems just to be the name for that species of practical thinking that is concerned with the selection of actions conducive to the achievement of a given end. For any particular process of deliberation, therefore, it is trivially true that the end assumed in that process was not arrived at via that process. But there is no reason why the end assumed in one piece of deliberation should not have been arrived at via another; e.g., strategic deliberation might lead to the result that the cavalry ought to be equipped with a certain sort of bridle, and the smith might consequently take that conception as the starting point for his deliberation as to how to make it. Aristotle’s theory must allow for such chains of deliberation, but any deliberation or chain of deliberations must start from the conception of some end that is not arrived at by deliberation; otherwise the chain of deliberation could have had no beginning.

2. Deliberation toward given ends and grasp of the right ends

Deliberation, then, cannot be autonomous; necessarily, it takes its starting points from elsewhere. This restriction on deliberation is applied by Hume to reason in general, in his famous insistence that reason is, and ought only to be, the slave of the passions (Treatise II.iii.3). The function of reason in Hume’s theory is the discerning of the relations between ideas, which in the practical sphere amounts to deliberation; reason can determine that if a certain thing is done, a certain outcome will result, or that if such and such is to be achieved, such and such must be done, but reason itself is incapable of determining that a given end is to be pursued. Ends are dictated by desires, i.e., by the passions; reason is restricted to working out the best means to what we happen to want, i.e., it is the slave of the passions. An influential tradition of exegesis, deriving from Walter (1874),3 has attempted to fit Aristotle also into this mould. On this view, the functioning of the practical intellect alike in its second-order activity and in its departmental role is identified with deliberation. The excellence of the practical intellect, phronēsis, is therefore identified with skill in deliberation to the attainment of ends determined not by the intellect but by ēthikē aretē, the excellence of the non-intellectual, appetitive side of the personality. This interpretation fits Aristotle’s insistence on the necessity of proper upbringing for the good man (e.g., 1095b2–5), and is supported above all by two famous passages from VI.12–13, 1144a7–9, and 1145a4–6, both of which say that eudaimonia requires phronēsis and excellence of character, the latter making one’s aim right, the former enabling one to do what promotes the achievement of that aim (ta pros ton skopon, ta pros to telos).4

Some of the defects of what we might call the “Humean” interpretation of Aristotle were well brought out by Allan,5 and his work has been taken further by others, notably Richard Sorabji, David Wiggins, and John McDowell.6 Allan’s principal contribution is to cite the abundant textual evidence to show that Aristotle’s conception of phronēsis is not restricted to skill in deliberation toward a pre-determined end, but at least includes a correct conception of that end itself. A single instance will suffice here: at 1140b11–21, Aristotle argues that the conception of the good that is the starting point of the practical reasoning of the phronimos will not be available to someone who has been corrupted by excessive desire for pleasure or aversion from distress; e.g., someone who can’t endure any distress will simply not appreciate that courage is a necessary component of the good life. It is in virtue of preserving this conception that sōphrosunē is so called, hōs sōizousan tēn phronēsin (“as preserving phronēsis” (b12)). We should not indeed read this passage as stating or implying that the conception of the good as the starting point (archē) of practical reasoning that is preserved by sōphrosunē is exclusively the over-arching conception of the supreme good: it is not every conception, Aristotle says, that is corrupted and distorted by pleasure and pain (and therefore preserved by sōphrosunē), e.g., the conception that the angles of a triangle are equal to two right angles, but the conceptions[plural] about what is to be done. That is, the conception that is preserved by sōphrosunē is simply the conception of whatever good it is that the agent proposes as his or her end, which of course includes the conception of a single supreme end, but is neither restricted to nor presupposes that conception. But central to this passage is the thesis that the conception of the end or ends that is preserved by sōphrosunē is the conception of a starting point for practical reasoning. Starting points are mentioned three times in five lines: i) “the starting-points [plura]) about things to be done are that [singular] for the sake of which things are to be done” (b16–17); ii) “to the person corrupted by pleasure and distress the starting-point will simply not be apparent, nor that one should choose and do everything for the sake of this and because of this” (b17–19); iii) “for vice is destructive of the starting-point” (b[19–20).

This conclusion is strengthened by consideration of the structural parallelism between the respective excellences of the theoretical intellect (sophia) and the practical (phronēsis). Both consist of two parts: a) a correct grasp of first principles, and b) the ability to reason correctly from principles thus grasped. Aristotle’s term for the former, certainly in the theoretical sphere and perhaps also in the practical,7 is nous (1141a7–8, 1142a25–6, 1143a35–b5); the latter he calls epistēmē in the theoretical sphere (1139b31–2), euboulia in the practical (1142b16, 31–3). Sophia is nous + epistēmē (1141a18–20); phronēsis is nous + euboulia. Aristotle does not, indeed, give the latter formula in so many words; but he comes very close to it in 1142b31–3, and in fact deviates from it in the direction of an even stronger emphasis on the grasp of the end, which he not only distinguishes from euboulia, but appears actually to identify with phronēsis itself; euboulia is “correctness concerning what conduces to the end, of which phronēsis is the true conception.” In common with the majority of scholars who have discussed this passage, I think that the antecedent for “which” is “the end,” not “what conduces to the end.”8 Grammatically, either is possible, but the contrast between a true conception of the end and correctness in the choice of means to that end is at least clearer than whatever contrast might be drawn between a true conception of the means to an end and correctness with respect to those means. Perhaps the contrast might be that between correctly conceiving the means and actually putting those conceptions into effect, but in that case what is contrasted with phronēsis is practical effectiveness, for which euboulia, defined by Aristotle at 1142b16 as orthotēs tis … boulēs, i.e., soundness in judgement, prudence, is an inappropriate term.

One component, then, of the excellence of the practical intellect is having a correct conception of what the good for man is. Aristotle’s language here is uncompromisingly cognitivist; phronēsis is a true conception of the end (1142b33); it is a true practical disposition concerned with things good and bad for man (1140b4–6, 20–21); excellence of character requires “excellence either natural or acquired in having correct beliefs (orthodoxein) about the first principle” (1151a18–19). Just as someone who possesses scientific knowledge must deduce what he/she knows from first principles that are themselves known (An Post I.1, 71a1–2), so the phronimos must derive his/her reasoned selection of action from first principles that he/she grasps as true. How does the phronimos come to grasp those true principles?

3. Who or what grasps ethical principles? (Three interpretations)

Here we find ourselves facing a problem that has taxed the ingenuity of some of Aristotle’s ablest commentators. Basically, the problem is posed by the fact that Aristotle suggests at various points three different accounts of how the possessor of practical wisdom acquires a reliable grasp of principles9; these accounts are, moreover, a) prima facie incompatible with one another, b) such that two of them appear to threaten the possibility of knowledge of principles.10 On the first, “Humean” account, one comes to a grasp of principles, not via the intellect at all, but through the habituation (ethismos) of the appetites, resulting, where successful, in excellence of character.11 Having a reliable grasp of the correct principles of conduct is simply a matter of having been brought up to like and enjoy good kinds of action and to dislike bad ones and find them unpleasant (EN II.1–3, especially 1104b3–13); and the good and bad are simply those kinds of action that the person of practical wisdom likes and dislikes. On the second account, the principles of conduct are “reputable” (endoxa) or well-grounded opinions, identified as those accepted by everyone, or by most people or by the wise, i.e., theorists of repute12. A reliable grasp of such principles is achieved by a method of critical enquiry, exemplified by Aristotle’s own procedure, which seeks to identify the most plausible principles and to establish their credentials by, as far as possible, removing objections and eliminating apparent inconsistencies between them. On the third, the account of the grasp of practical principles is identical with that of theoretical: like the latter, the former are grasped by nous, which is the name for a grasp of general principles arrived at by induction from sensible particulars.13 The first two accounts are prima facie inimical to claims that the rational agent possesses knowledge of principles; on the first, the basic attitudes are apparently not cognitive at all, but rather affective, while on the second, it might appear, principles are not known, but accepted on the ground that they, or propositions supporting them, are believed by everyone, or by most people, or by some eminent theorist or other. But on that account, surely, nothing more can be claimed for them than plausibility.

Let us take the first account, according to which the first principles are grasped by habituation of character. This is supported by 1098b3–4, where Aristotle lists various ways in which principles are grasped. “Some,” he says <are grasped> “by induction, some by a sort of habituation and some in other ways,” and, given his insistence (1095b4–7) on proper upbringing for a grasp of the starting points of moral reasoning, it is an extremely plausible inference that he regards habituation as the process by which moral principles are grasped.14

Jessica Moss has recently given a sustained defense of this position (Moss [2011 and 2012]) against the consensus initiated by Allan (see Section 2, second paragraph). Her central claim is that the ends of action are set, not by any exercise of intellect, but by moral character, which, being the excellence of the appetitive, in contrast with the intellectual element of the soul, is a “wholly non-intellectual state” (Moss [2011], abstract, p. 204). But though virtue of character is non-intellectual (or, equivalently, non-rational), it is not, as critics of the “Humean” interpretation have assumed, non-cognitive. Virtue of character does indeed embody a conception of an end or ends to be achieved, but that conception is itself non-rational, in that it is not a belief nor grounded in any prior beliefs, but is an exercise of phantasia, which is a non-doxastic state in which things appear or seem a certain way to a perceiver, independent of whether the perceiver believes them to be as they seem (e.g., the sun appears [i.e., looks] a foot across, whether or not one believes that that is how big it actually is).15 The role of phantasia in virtue of character is as follows: to think of a certain kind of action as good is to have a pleasurable phantasia of an instance of that kind of action, which is caused by previous experience of the pleasure of doing or imagining something of that kind. Someone properly brought up has experienced many such pleasures in the course of his/her upbringing, which produce the appropriate phantasia, which in turn motivates the doing of the appropriate action. The function of practical reason is to work out the best means of achieving the “phantasised” end.

This theory is developed in the works mentioned with a wealth of textual support and considerable argumentative ingenuity and inventiveness, to which it impossible to do justice in a brief discussion. It does, however, seem to me founded on a basic misunderstanding of Aristotle’s view of the appetitive element of the soul, and consequently of what kind of state virtue of character is. Moreover, Moss’s own discussion brings out only too clearly the lack of fit between her basic conception of the appetitive soul as non-rational and the role that, correctly representing Aristotle, she ascribes to the virtue of that part of the soul in the direction of action. The basic misunderstanding is simply the characterization of the appetitive soul as, unqualifiedly, non-rational, and hence of the habituation of the appetites as a non-rational process (e.g., Moss [2011], p. 205: “we each reach our view of what happiness consists in—virtuous activity, for example, or the life of pleasure or of honour—not by any intellectual process, but instead through the non-rational habituation of the non-rational part of the soul”; and [2012], p. 70: “In the ethical works Aristotle attributes the passions to a part of the human soul which exercises perception and phantasia but is not capable of belief … which implies that if they are based on cognition at all it must be non-rational cognition, perception or phantasia”). In EN I.13 Aristotle does not characterize the appetitive soul as simply non-rational, but precisely contrasts it in respect of rationality with a part of the soul that is simply non-rational, namely the vegtative soul, i.e., the non-rational capacities for nourishment, growth, and reproduction. The latter, he says (1102b29–1103a3), “does not share in reason [logos] in any way, but the appetitive and in general desiderative does share in it in a way, in that it is obedient and subject to it. That is the way we speak of taking account [echein logo] [See preceding author’s note.]of one’s father and friends, not as in mathematics. And that the non-rational is in a way persuaded by the rational is indicated by admonition and all reproof and encouragement. And if we must say that his too possesses reason, then [not only the non-rational (1102b28–9) but] the rational also will be two-fold, the one strictly and in itself, the other as something which listens as to its father.” As the last sentence makes clear, the contrast is not that between the rational and the non-rational, but between on the one hand what is strictly, essentially, or per se rational (I use those three terms equivalently), and on the other what is secondarily or derivatively rational, though strictly, essentially, or per se non-rational. I take Aristotle to be saying that, whereas the intellect is essentially the capacity to reason and understand, appetite (primarily bodily appetite) and wanting more generally are not essentially capacities to reason, nor necessarily accompanied by or grounded in reason, since they are possessed by creatures that lack reason (non-human animals) and even in humans may occur apart from the exercise of reason. But in human beings appetite is responsive to reason, in that humans are capable of modifying their appetites in various ways in response to rational considerations, e.g., by coming to recognize that certain ways of getting what one wants are unfair, greedy, dishonest, likely to frustrate one’s aims in the longer run, etc. Human appetite is, then, capable of listening to reason; it “is something which listens as to its father,” in the case of the growing child no doubt to the words of its actual father and of others in positions of authority, in the case of adults to reasons given either by others or by the agent’s own deliberation. A major element in habituation (ethismos) is developing in the child or young adult the capacity to listen to reason. And listening to reason involves, not merely understanding what you are told, but recognizing that what are presented to you as reasons are in fact good, or sufficient, reasons for acting. Listening to reason therefore essentially involves having beliefs, and understanding the rational grounds for those beliefs. Habituation of course involves affective responses, enthusiasm for what is presented as fine or noble, disgust at what is presented as dishonest or demeaning, etc., but such responses are inseparable from evaluative beliefs.16

It seems to me that Aristotle sees the relation between the essentially rational intellect and the derivatively rational appetites as analogous to that between an expert mathematician and a pupil. The former can devise a proof, whereas the latter cannot, but is nonetheless capable of following, and of being convinced by, a proof presented to him/her. And being convinced by it involves assenting to the propositions composing the proof and the inferential steps linking them.

The description of the appetites as listening to reason is, then, inconsistent with Moss’s basic assumption that the appetites are unqualifiedly non-rational. But that description is so deeply embedded within Aristotle’s theory that she herself in several places describes the relation between appetite and reason in those very terms, without any indication of tension with her own assumption. Thus in (2012), p. 168, she writes: “[T]he passions and actions of a strictly virtuous person do not merely happen to coincide with what the well-functioning intellect would prescribe, but they are such as to wait upon the right prescription (the orthos logos) before becoming active.” By analogy, “a servant who receives no instructions, or no good instructions, from his master, might nevertheless tend to do the right thing, but will be in a state very different from that of a servant practiced in obedience to an excellent master. The former acts on his own impulses, the latter takes his lead from his superior.” But plainly the servant of the excellent master not merely understands the latter’s commands, but believes that the right thing for him to do is what the master commands, that the reasons that the master gives are good reasons, etc.17 These passages seem to me correctly to represent Aristotle’s view that appetite is responsive to reason, but in so doing to contradict Moss’s own assumption that appetite is unqualifiedly non-rational. In some other passages she departs even further from that assumption, by reversing the relation between non-rational appetite and reason, so that appetite is no longer responsive to reason, but reason to appetite. Thus at (2012) p. 224, she writes: “[P]art of phronēsis, then, will be having an explicit intellectual cognition that x is the end. But this in no way entails that intellect adds anything to the view of the end supplied by character. Instead, it suggests that intellect’s grasp of the end consists merely in agreeing with—assenting to—what character supplies.” Here an allegedly non-rational part of the soul, supposedly incapable of reason or belief, supplies something that reason assents to; but that must surely be a belief, and since reason assents to it, reason accepts it as true.18

Habituation, then, which everyone agrees to be the process by which virtue of character is acquired, is not the non-rational conditioning of a non-rational capacity or set of capacities. It is not, e.g., the acquisition of the habit of taking pleasure in the non-doxastic appearance of certain kinds of action as good. Rather, it is the process by which the derivatively rational appetites learn to listen to reason, and thereby to develop the secondary form of rationality that is appropriate to them.19 We must conceive of the process of habituation, not as a conditioning of “blind” appetites, but as a “twin-track” process in which the appetitive responses are progressively refined under the guidance of the intellect, which is itself undergoing a parallel process of refinement or, rather, enlightenment; the clearer the insight the intellect has of ethical principles, the more precise the instructions it can issue to the desires.20 But the desiderative component is essential for the grasp of principles to do its motivational work. This interpretation has the advantage of eliminating the prima facie incompatibility between the thesis that practical principles are acquired by habituation and the other two. If habituation itself presupposes a method by which the intellectual element is developed so as to grasp the truth, then, consistent with the thesis that principles are grasped by habituation, either of the other theses might be a correct account of that method.

The thesis that the principles of practical thought are “reputable opinions” (endoxa) raises a number of complex problems. First, what is the scientific status of endoxa and hence of arguments from them? Second, does Aristotle claim that all or only some moral principles are endoxa? At Top. I 1–2 he distinguishes scientific reasoning, whose principles are necessary truths, from dialectical reasoning, whose principles are endoxa. The latter is not itself scientific, but is a critical method among whose uses Aristotle particularly mentions the role of dialectical reasoning in examining the principles of the sciences; the principles of any science cannot be discussed within that science, since they are primitive with respect to it, but dialectic “being a technique of examination provides a way towards the principles of all the sciences” (101b3–4). It is not totally clear what Aristotle means by “a way towards” the principles. Does he mean that dialectic proves the principles? It is hard to see how arguments from received opinion are supposed to issue in knowledge of necessary truths. A possible response to these difficulties is to suggest that the role of dialectic in general is not to prove principles, but a) to find supporting arguments for principles already grasped inductively21 and b) to provide arguments against putative principles. If that general account were applied to the special case of practical principles, the second and third accounts listed above cease to be rivals, not merely to the “habituation” thesis, but also to one another. The phronimos grasps first principles inductively, thereby exercising nous, and is able to support them by arguments from endoxa. Do the texts provide any support for this appealingly eirenic suggestion?

Disappointingly, the texts are indecisive. Some may indeed be read as supporting this suggestion: thus EE 1216b26–8 says that the aim of ethical enquiry must be “to seek conviction through the arguments, using ta phainomena [i.e. ta endoxa] as pieces of evidence and examples,” which suggests the supportive role described in the immediately preceding paragraph. EN 1098b9–12 is also naturally read in the same way. Aristotle has argued for his identification of human good as excellent actualization of soul from general principles about the function of a thing, viz. that in the case of things that have a function, their good or bad state is determined by the performance of their function, and that the function of a human being is rational activity. He then supports this conclusion by arguments from common opinion, designed to show that excellent rational activity fits generally accepted beliefs about human good, prefacing these with the statement, “We must examine it not only from the conclusion and the premisses of the argument but also from what is said about it; for everything which is the case is consistent with what is true, but what is true soon disagrees with what is false.” Elsewhere, however, the phainomena seem to provide the principles of proofs themselves, though in these passages it is not clear that the conception of proof is uniform. EE 1215a7–8 says that refutations of opinions are demonstrations of the theories (logoi) opposed to them, which suggests a formal proof by reductio ad absurdum. But at EN 1145b2–7 Aristotle introduces his discussion of akrasia by saying that, as in the other cases, we must begin by setting out the phainomena and asking questions, and so prove (deiknunai) all the reputable opinions, or if not all, as many as possible and the most trustworthy, “for if the difficulties are resolved and the reputable opinions are left in place, a sufficient proof will have been given” (cf. Phys. 211a7–11). If we may take the phrase “as in the other cases” as indicating that this method is to be employed in every case (as is perhaps also suggested by EN 1095a28–b7), then the notion of proof in ethics is radically different from that suggested by the axiomatic model. According to the latter, proving a proposition is showing that it has to be true, since it follows from one or more premisses that, independently, have to be true. On the ethical model, proving a proposition consists in showing that it is (or follows from) a reputable opinion that has survived the process of setting out the phainomena and raising difficulties about them. Instead of a system of proof, Aristotle appears to be offering a coherentist scheme of justification that issues at best in defeasible judgements of the form “Since p forms part of the best available scheme of ethical beliefs, it may be held as true until good reasons are discovered for revising the scheme.”

4. The role of phronēsis in the choice of ends

This assumes, however, that all ethical principles either are or are derived from endoxa, and it is not clear that that is true. In Aristotle’s ethics, general philosophical principles, especially those from his metaphysics and psychology, play a fundamental role.22 We have already cited his use of principles about the function of things in identifying the good for man; other instances are the metaphysical arguments in EN X for the thesis that theoretical excellence is the highest good, and his frequent employment of the principle that, for any subject matter where things appear to different observers, the way things really are is the way they appear to the observer in good or proper condition (e.g., the way things taste to the healthy person is how they really taste).23 He gives no sign of thinking of these as merely reputable opinions, or as derived from such opinions. Rather they are principles fundamental to his whole scheme of thought, and as such they seem to be regarded as unassailably true. But how are they known? He gives us no explicit guidance. It remains a possibility that they are supposed to be grasped inductively, by nous.

The crucial text is EN 1143a35–b14.24 There Aristotle distinguishes two objects of nous, viz. undemonstrated principles on the one hand and particular instances falling under them on the other; these are the “extremes in either direction,” i.e., the starting points and finishing points of reasoning, and they share the property of being undemonstrated (since there is no demonstration of singular propositions in Aristotle’s syllogistic). This distinction is then complicated by the introduction of the distinction between theoretical and practical reasoning, giving a distinction between nous of undemonstrated theoretical principles and nous of particular practical instances (b1–3). Is practical nous here then just a faculty of moral perception, by the exercise of which we see instances of conduct as falling under moral characteristics, without any reference to principles? That appears to contradict 1142a23–30, where that very faculty is called a special sort of perception and is contrasted with nous, precisely on the ground that the latter is concerned with undemonstrated principles, which must, given the context, be practical principles. A resolution of this apparent contradiction is suggested by 1143b4–5, where Aristotle says that particular instances of conduct are “principles of that for the sake of which <sc. we act>, for universals come from particulars: so we must have perception of the latter, which we call nous”; this appears to describe the inductive procedure familiar from An. Post. II.19.25 Aristotle would then have a tidy and unified doctrine were he to hold that there is nous of moral as well as of theoretical principles, that both kinds of nous are acquired inductively, and that perception of morally significant instances is (another kind of) nous, which provides the perceptual data from which moral principles are inductively derived.26

He does not, however, set things out thus neatly. Rather, the emphasis in this discussion of nous is on its perceptual role; it is a natural endowment that develops with experience (b6–9), which is why we must pay attention to the undemonstrated pronouncements of people of age and practical sense, “for, having an eye from experience, they see aright” (b11–14). The sort of eye one has from experience would naturally seem to be the trained perception of particular instances as falling under some significant characteristic, exemplified by a general’s eye for terrain; what the untrained observer sees purely as a pleasant rural scene, the trained military eye simply sees as the place for the enemy to concentrate his armored reserve. That type of perception is indeed trained by experience, but is nonetheless exercised in particular instances, not in grasping any general proposition, as in standard cases of induction. Yet a little later (1144a29–36), Aristotle attributes the grasp of first principles of conduct (represented schematically as “Since such and such [let it be whatever you like] [Round parentheses, not square brackets: “let it be whatever you like” is Aristotle’s insertion.]is the best thing and the good …”) to “the eye of the soul,” pointing out that that eye does not see properly without excellence of character. It therefore appears that Aristotle may have thought that the possession of the trained eye enables its possessor to formulate universal principles. But he also emphasizes that correct conduct cannot be formulated in exceptionless generalizations, but has to be determined by particular circumstances discerned by perception (1104a6–10, 1109b20–3, 1126b2–4; cf. IX.2). On that conception of phronēsis, general principles specify the end only indeterminately, e.g., one has a general conception of a good life that embraces virtues of character such as courage, but what it is to be courageous cannot be exhaustively specified in any formula, but has to be recognized by the trained judgement (i.e., perception) of the courageous person.27 That is certainly Aristotle’s dominant account of excellence of character. It therefore seems that the view of practical nous that that account requires fits awkwardly with the texts that suggest that nous ought to be a grasp of universal principles, reached by induction from particular instances. His actual doctrine of goodness of character tends rather to assimilate it to the quasi-perceptual grasp of instances of moral concepts.28

A solution to this problem is possible, if we assume a) that it is the task of induction to lead from sensible particulars to the mastery of a general concept, and b) that this mastery consists precisely in the ability to recognize an indefinite number of particular instances falling under the concept, without the necessity of subsuming these particular instances under a single general formula.29 According to this account, habituation itself is understood as the acquisition of this perceptual capacity, and furthermore the acquisition of that capacity may itself be fostered by the reputable opinions of people of practical sense on whose unproved sayings and opinions we are to rely no less than on proofs, since they have the eye of experience and so see aright (1143b11–14). Reputable opinions and perception are thus mutually reinforcing elements in moral education.30 Aristotle’s account, though obscure in detail, is unambiguous in its central contention that cognitive capacities are necessary for the correct orientation of the appetitive responses.

This result presents two problems. First, what are we to make of those passages that seem to support the Humean interpretation? The second, more substantial problem is this: does the exercise of moral nous allow any genuine role to deliberation?

In fact, the passages from VI.12 and 13 fit our interpretation quite well. At 1144a6ff., Aristotle is making the point that the achievement of eudaimonia requires excellence both of the practical intellect (i.e., phronēsis) and of character (i.e., ēthikē aretē). The latter, he says, makes one’s aim right, the former makes right the things one does to achieve that aim. On the interpretation that I have proposed, both these things are true, though admittedly they do not amount to a complete account of phronēsis. Virtue of character makes one’s aim right, in that it consists in both a rationally conditioned cognition of the end to be pursued and the motivation to pursue that end, while phronēsis identifies the means by which that end is to be achieved. Of course, as we have already seen, that is not the whole task of phronēsis, since the appetites have developed their motivationally active cognition of the end precisely by becoming properly responsive to the practical intellect, the excellence of which is phronēsis.31 1145a2–6 makes the same point: for prohairesis to be right we need both phronēsis and ēthikē aretē; hē men gar to telos hē de ta pros to telos poiei prattein. The verb to be supplied after telos is presumably poiei, and the sentence is to be read as “the one [i.e., aretē] makes the telos [i.e., makes something the telos], the other [i.e., phronēsis] makes one do the things conducive to the telos,32 where “makes the telos” is understood in the same sense as “makes the aim right” in the previous passage. Finally, it is worth mentioning two passages that might appear to support the Humean interpretation, though in fact they tell firmly against it, in favor of the interpretation I have defended. The first is the assertion at 1139a35–6 that “thought by itself moves nothing,” which defenders of the Humean interpretation have interpreted in the sense of “reason is the slave of the passions.” But the sentence runs ”thought by itself moves nothing, but practical thought, i.e., thought directed with a view to something, does; for this [i.e., practical thought] is also the first principle of productive thinking.” Aristotle agrees with Hume that the initiation of action requires desire as well as thought: he rejects Hume’s central thesis that the function of reason is simply to devise ways of carrying out the dictates of desire. The second passage is one mentioned earlier, 1151a15–20, where Aristotle is reiterating his point that goodness of character preserves one’s grasp of the first principles of morals, while wickedness destroys it (cf. 1140b11–21). In practical reasoning the starting point is the goal, in mathematics it is the axioms; “for neither in one nor the other is reason demonstrative of the first principle, but excellence, whether natural or acquired, of having correct beliefs about the first principle.” Plainly this cannot mean that our grasp of first principles in general is not given by the intellect, but somehow emanates from a non-rational element, since that is obviously inconsistent with Aristotle’s general doctrine of how we come to grasp first principles.33 Rather the point is that in neither form of reasoning does reason demonstrate those principles, which have therefore to be known non-demonstratively.34

It is, then, clearly Aristotle’s doctrine that excellence in deliberation (euboulia) is only one element of phronēsis, the other being the grasp of first principles. But that schematic account can be realized in either of two ways. In the first, set out in the doctrine of the practical syllogism, it is nous (understood as the grasp of general principles) that formulates the major premiss, thereby providing a complete specification of the end to be achieved, while the premiss or premisses setting out the steps by which this end is to be achieved are worked out in a quasi-deductive way by euboulia. Thus suppose that shelter is necessary for survival, and hence a good. It is possible to specify completely what shelter is, and, given our knowledge of the world, specifically of the nature and availability of building materials, and our knowledge of techniques of building, we can work out that if shelter is to be obtained then such and such materials must be organized in such and such ways.

5. Two tasks of the practical intellect

This model certainly applies readily to technical deliberation, from which Aristotle normally takes his examples of practical reasoning. But on the other hand Aristotle’s doctrine of excellence of character, and the moral nous that that excellence presupposes, does not allow the end to be specified in advance of the process of “deliberation.” Hence that doctrine requires the second way in which the schematic account is realized. In this way, the end is specified indeterminately, e.g., “the good life requires that one act courageously,” and that indeterminate conception is made determinate by the exercise of moral insight in concrete situations. Excellence of character is a disposition to select actions, a disposition that is in a mean, i.e., that consists in being neither too given to a certain motivation nor insufficiently sensitive to it. And what counts as being neither excessively nor insufficiently motivated cannot be exhaustively specified by any formula, but is determined by the judgement of the phronimos. E.g., being courageous is being neither excessively timorous nor being insufficiently motivated by fear, and similarly being neither excessively nor insufficiently bold (thrasus). But what counts as being courageous in any actual situation has to be discerned by the phronimos via his/her informed judgement of the appropriate action in those circumstances; there is no way in which he/she can read off the requirements of courage from his/her antecedent grasp of the nature of courage, “for it does not fall under any craft or precept [parangelia], but those who act must always look to the particular occasion, as is also the case in medicine and steersmanship” (1104a6–10).

That is why the phronimos needs nous; he/she has to have an eye for the requirements of courage, a sense of what is appropriate for the courageous person, which eludes capture in any formula. Lacking that, his/her grasp of the end is indeterminate; the inexperienced individual who believes that courage is a good and wants to be courageous, but lacks that insight into what courage requires in this or that situation, does not yet have a determinate conception of what courage is. The acquisition of that determinate conception does not consist in learning rules or formulae; it consists in the development of a style of behavior, of the ability to respond flexibly and appropriately to an indefinite variety of situations, like the expertise of a master helmsman.35 But now the role of deliberation in phronēsis appears to be reduced to a vanishing point; for the phronimos does not deliberate how to achieve a pre-determined goal, any more than the helmsman deliberates how to steer the ship. Deliberation would seem to have at best a subsidiary, partly technical role; e.g., having decided that courage requires me to stand and fight now, I may have to set about improvising a weapon, a task calling for technical deliberation.

It seems to me, then, that so far from tending to identify phronēsis with excellence in deliberation, Aristotle’s theory has in fact the opposite tendency; phronēsis tends towards identification with moral nous, understood as the capacity for moral perception, leaving deliberation with a role of uncertain and at best marginal scope. Perhaps that is Aristotle’s final position; we conceive the ends of action only indeterminately, e.g., as living temperately, courageously, etc., and the task of practical thinking is to make those conceptions determinate.36 Perhaps that is what deliberation is, and we are misled by those passages that suggest that deliberation requires a determinately specified end.37 We must, however, consider another possibility, which also has some textual support. The practical skills by which Aristotle illustrates the activity of the phronimos, medicine and navigation, have independently specifiable goals, respectively the maintenance or restoration of health and a safe and speedy voyage. What counts as doing well in these activities may not be formulable in any finite set of directions, but it is determined by the formulae “doing what best promotes health” and “doing what best assures a safe and speedy voyage.” The concluding sentences of EN VI suggest that an independent goal for phronēsis, answering to health in the case of medicine, is provided by the promotion of theoretical activity. That is fairly clearly the case in the final chapter of the Eudemian Ethics, and the position would fit well with the supreme value assigned to theoretical activity in EN X.38

On that view the criterion of right action is always that this action best promotes theōria, whether directly, by furthering the agent’s own theorizing, or that of others, or indirectly, e.g., by promoting conditions, such as a stable and prosperous society, in which theōria can flourish. It may, however, be, as in the case of medicine, that it is impossible to give a finite formulation of all the kinds of action that have that character, and that the insight of the phronimos will not necessarily refer to the promotion of theōria, but will simply be the immediate insight that this or that is the right thing to do here and now. On this model the grasp of the end is simply the grasp of the truth that theōria is the supreme good, while the task of phronēsis is to make determinate in action the indeterminate conception of “character and conduct such as best promotes theōria”. On the other model the grasp of the end is the grasp of the truth that the good is living excellently, while the task of phronēsis is simply to make that indeterminate conception determinate. Either model allows that moral nous has the twofold task of first grasping the end and then making determinate this (more or less) indeterminate grasp. Either model thus allows Aristotle to hold both that phronēsis is the supreme virtue of the practical intellect and that it is concerned with ta pros to telos. The fact that the text of the Nicomachean Ethics contains indications of support for both these models may suggest that Aristotle may have found it difficult to choose between them, or even that he had not clearly distinguished them from one another. Either would be symptomatic of a tension that seems to me to pervade the Nicomachean Ethics, between the demands on the one hand to see morality as autonomous and, on the other, to elevate the theoretical intellect as the bearer of supreme value.39 I do not believe that Aristotle ever satisfactorily resolved that tension, but that is another and much longer story.


Ackrill, J.L. and J.O. Urmson. [1980], Aristotle, The Nicomachean Ethics (Ross’s translation (see Ross [1925]), revised by J.L. Ackrill and J.O. Urmson), Oxford.Find this resource:

Allan, D.J. [1953], “Aristotle on the Origin of Moral Principles,” Actes du XIe Congrès Internationale de Philosophie, vol. xii, 120–127, repr. in J. Barnes, M. Schofield, and R. Sorabji, eds., Articles on Aristotle 2: Ethics and Politics, London, 1977.Find this resource:

Anscombe, G.E.M. [1965], “Thought and Action in Aristotle,” in R. Bambrough, ed., New Essays on Plato and Aristotle, London and New York,. 143–158, repr. in J. Barnes, M. Schofield, and R. Sorabji, eds., Articles on Aristotle 2: Ethics and Politics, London, 1977.Find this resource:

Barnes, J. [1994], Aristotle, Posterior Analytics, 2nd rev. edn., Oxford.Find this resource:

Bostock, D. [2000], Aristotle’s Ethics, Oxford.Find this resource:

Broadie, S. [1991], Ethics with Aristotle, New York.Find this resource:

Burnet, J. [1900], The Ethics of Aristotle, Oxford.Find this resource:

Burnyeat, M. [1980], “Aristotle on Learning to Be Good,” in Rorty [1980], 69–92.Find this resource:

Cooper, J. [1975], Reason and Human Good in Aristotle, Cambridge, MA.Find this resource:

Dahl, N.O. [1984], Practical Reason, Aristotle, and Weakness of Will, Minneapolis.Find this resource:

Dirlmeier, F. [1969], Aristoteles, Nikomachische Ethik (Aristoteles, Werke in Deutscher Ṻbersetzung, vol. vii), Berlin.Find this resource:

Dow. J. [2009], “Feeling Fantastic? Emotions and Appearances in Aristotle,” Oxford Studies in Ancient Philosophy 37, 143–175.Find this resource:

Engberg-Pedersen, T. [1983], Aristotle’s Theory of Moral Insight, Oxford.Find this resource:

Fortenbaugh, W.W. [2002], Aristotle on Emotion, 2nd edn., London.Find this resource:

Gauthier, R.A. and J.Y. Jolif. [1958/9], Aristote, L’Éthique à Nicomaque: Introduction, Traduction et Commentaire, 2 vols. in 3 parts (vol. i 1958, vol. ii 1959), Louvain.Find this resource:

Gosling, J.C.B. and C.C.W. Taylor. [1982], The Greeks on Pleasure, Oxford.Find this resource:

Hardie, W.F.R. [1980], Aristotle’s Ethical Theory, 2nd edn., Oxford.Find this resource:

Irwin, T. [1978], “First Principles in Aristotle’s Ethics,” in P. French, T. Uehling, and H. Wettstein, eds., Midwest Studies in Philosophy III: Studies in Ethical Theory, 252–272.Find this resource:

Irwin, T. [1988], Aristotle’s First Principles, Oxford.Find this resource:

Irwin, T. [1999], Aristotle, Nicomachean Ethics, Translated with introduction, notes and glossary, 2nd edn., Indianapolis.Find this resource:

Kenny, A. [1978], The Aristotelian Ethics, Oxford.Find this resource:

Kenny, A. [1979], Aristotle’s Theory of the Will, London.Find this resource:

Loening, R. [1903], Die Zurechnungslehre des Aristoteles, Jena, repr. Hildesheim, 1967.Find this resource:

Lorenz, H. [2006], The Brute Within: Appetitive Desire in Plato and Aristotle, Oxford.Find this resource:

Lorenz, H. [2009], “Virtues of Character in Aristotle’s Nicomachean Ethics,” Oxford Studies in Ancient Philosophy 37, 177–212.Find this resource:

McDowell, J. [1979], “Virtue and Reason,” The Monist 62, 331–350, repr. in R. Crisp and M. Slote, eds., Virtue Ethics, Oxford, 1997, and in Sherman [1999].Find this resource:

McDowell, J. [1980], “The Role of Eudaimonia in Aristotle’s Ethics,” Proceedings of the African Classical Association 15, repr. in Rorty [1980], 259–276.Find this resource:

Monan, J.D. [1968], Moral Knowledge and its Methodology in Aristotle, Oxford.Find this resource:

Moss, J. [2011], “‘Virtue Makes the Goal Right’: Virtue and Phronesis in Aristotle’s Ethics,” Phronesis 56, 204–261.Find this resource:

Moss, J. [2012], Aristotle on the Apparent Good: Perception, Phantasia, Thought, and Desire, Oxford.Find this resource:

Müller, A.W. [1982], Praktisches Folgern und Selbstgestaltung bei Aristoteles, Freiburg.Find this resource:

Price, A.W. [2011], Virtue and Reason, Oxford.Find this resource:

Reeve, C.D.C. [1992], Practices of Reason, Oxford.Find this resource:

Rorty, A.O., ed., [1980], Essays on Aristotle’s Ethics, Berkeley, Los Angeles, and London.Find this resource:

Ross, W.D. [1925], Ethica Nicomachea, in The Works of Aristotle Translated into English, vol. vii, Oxford.Find this resource:

Rowe, C.J. [2002], Aristotle, Nicomachean Ethics, Translation (with commentary by S. Broadie), Oxford.Find this resource:

Sherman, N. [1989], The Fabric of Character, Oxford.Find this resource:

Sherman, N. ed., [1999], Aristotle’s Ethics: Critical Essays, Lanham MD, Boulder CO, New York, and Oxford.Find this resource:

Smith, A.D. [1996], “Character and Intellect in Aristotle’s Ethics,” Phronesis 41, 56–74.Find this resource:

Sorabjl, R. [1973–1974], “Aristotle on the Role of Intellect in Virtue,” Proceedings of the Aristotelian Society 74, 107–129, repr. in Rorty [1980].Find this resource:

Striker, G. [2006], “Aristotle’s Ethics as Political Science,” in B. Reis, ed., The Virtuous Life in Greek Ethics, Cambridge, 125–142.Find this resource:

Taylor, C.C.W. [2006], Aristotle, Nicomachean Ethics Books II-IV: Translated with a Commentary, Oxford.Find this resource:

Taylor, C.C.W. [2008], “Aristotle on the Practical Intellect,” in Taylor, Pleasure, Mind, and Soul, Oxford, 204–222, originally published (in German) in T. Buchheim, H. Flashar, and R.A.H. King, eds., Kann man heute noch etwas anfangen mit Aristoteles?, Hamburg, 2003.Find this resource:

Urmson, J.O. [1984], Nicomachean Ethics (Ross’s translation (see Ross [1925]) revised by J.O. Urmson) in J. Barnes, ed., The Complete Works of Aristotle: The Revised Oxford Translation, 2 vols., Princeton.Find this resource:

Walter, J. [1874], Die Lehre der praktischen Vernunft in der griechischen Philosophie, Jena.Find this resource:

Wiggins, D. [1975–6], “Deliberation and Practical Reason,” Proceedings of the Aristotelian Society 76, 29–51, rev.version in Rorty [1980].Find this resource:

Woods, M.J. [1986], “Intuition and Perception in Aristotle’s Ethics,” Oxford Studies in Ancient Philosophy 4, 146–155, Woods, Four Prague Lectures and Other Texts, Prague, 2001.Find this resource:


(1) Except where otherwise specified, references to Aristotle are to the Nicomachean Ethics (EN).

(2) See Anscombe [1965], Müller [1982], ch. 6, esp. pp. 231–236 and 258–259, and Broadie [1991], pp. 219–225.

(3) For details see Allan [1953].

(4) See also 1178a16–19.

(5) See n. 3.

(7) It is disputed whether (as maintained by Kenny [1978], pp. 171–172) the word nous designates the correct grasp of practical, as well as theoretical principles. On practical nous see main text, sections 4–5 and n. 13 and 29.

(8) So Gauthier and Jolif [1958/59], vol. 2, pp. 518–519, Kenny [1979], pp. 106–107, Müller [1982], pp. 267–268 and Bostock [2000], p. 85. The opposite view is defended by Burnet [1900] and by Moss [2011].

(9) Though I initially set these out separately for the sake of clarity in exposition, an adequate theory (as I later suggest, in common with most commentators) must combine elements of these different proposals.

(10) Cf. Irwin [1978], p. 258: “Aristotle seems to recognize three incompatible ways to first principles; habituated virtue without wisdom [= my first alternative], virtue including wisdom [= the third alternative], and dialectic [= the second alternative].”

(11) Bostock [2000], pp. 88–96 maintains this position tentatively. Cf. Müller [1982], p. 304.

(12) Irwin [1988] defends a version of this position. In his view the first principles of ethics are discovered, not by “pure dialectic,” but by “strong dialectic.” In strong dialectic, principles based on well-founded opinions are themselves justified by fundamental truths of metaphysics. Irwin writes: “Aristotle needs strong dialectic, to make up for the weakness of pure dialectic. He thinks he can evaluate the common beliefs from a sufficiently independent view, because he has metaphysical arguments to justify some of his leading principles, and especially those principles that support his view of substance, matter, nature and essence. His metaphysical arguments are not non-dialectical, since they do not stray outside common beliefs altogether. But they are not merely dialectical, since they rest on premisses that we must believe if we are to believe in an objective world at all, or if we are to understand or to explain nature. The arguments of strong dialectic resting on those principles underlie Aristotle’s conception of the soul and of human nature, which in turn underlies the main claims of the Ethics” (p. 358).

(13) So Reeve [1992], ch. 1. On Aristotle’s doctrine of the grasp of theoretical principles see Barnes [1994].

Cooper [1975] agrees with the thesis that the first principles of ethics are grasped by nous, but rejects the inductive account of nous. In his view the principles that are grasped by nous are established by dialectical thinking: “Aristotle holds both that the principles of the sciences are known intuitively, by nous, and that they can be established by discursive dialectical argument” (p. 67). “A similar condition obtains in his moral philosophy” (p. 69).

Since Reeve too insists on the indispensable role of dialectic in the clarification of principles that have already been justified by experience (p. 40), his position does not differ widely from that of Cooper.

(14) For a sensitive account of habituation (ethismos) see Burnyeat [1980].

(15) Dow [2009] argues that emotional phantasia (which is the kind of phantasia involved in training of character) consists not, as Moss maintains, in an object’s appearing non-doxastically thus and so to the subject, but in the subject’s taking (i.e., believing) the object to be thus and so. His article contains numerous references to the extensive literature on the topic of phantasia.

Fortenbaugh [2002] maintains the same account of emotional phantasia (Epilogue, pp. 96–103). I am therefore unable to understand why Dow asserts ([2009], p. 144, n. 2) that Fortenbaugh denies that Aristotelian emotions have a cognitive component; on the contrary, the central thesis of Fortenbaugh’s book is precisely that emotions have a cognitive component. See p. 94, “In the main body of this monograph I argued that Aristotle analyses emotions as complex phenomena involving inter alia thoughts or beliefs.” See also the passages quoted in n. 19.

(16) Hendrik Lorenz [2009], pp. 177–179 rejects my account of virtue of character as the responsiveness of the appetites to reason, as expressed in Taylor [2006], intr. p.xv, and p. 106; he cites me as holding the view which he opposes, namely that “the virtues of character are wholly constituted by excellent conditions of a part or aspect of the human soul that is not itself capable of reasoning, deliberating, or thinking, but that is responsible for the person’s dispositions or tendencies to desire pleasures and to be averse to pains and to feel emotions such as anger and shame” (p. 178). Against that, he contends that for Aristotle virtue of character is a state not merely in accordance with (kata) correct reason but with (meta) correct reason (EN 1144b24–8), and consequently that “the virtues of character are states that are constituted, in part, by a certain correct state of reason, namely by correct reason about what is good for humans” (p. 211). I hope that it is clear from the evidence which I have presented (specifically in section 3, fourth paragraph) that I do not think that virtue of character is an excellence of a part of the soul that is incapable of reasoning, deliberating, or thinking (and indeed I do not believe that I maintained or was committed to that view in my [2006] commentary).

Lorenz and I agree that to have virtue of character is to be in a state of perfect responsiveness of the appetites (and spirit) to reason; see Lorenz [2006], p. 190: “In the virtuous person, appetite and spirit have come to be in perfect harmony with reason (1102b28).” Where we disagree is that, whereas I think that appetite is capable of being persuaded by reason, as the child and the novice mathematician are, his view is that “the non-rational part cannot strictly be reasoned with, because it is unable to grasp inferential connections” (Lorenz [2006], p. 189). That claim rests ultimately on the contrast which Aristotle draws at 1098a3–5 between what “has reason” in the sense of being persuaded by (or obedient to) reason (hōs epipeithes logōi) and “what has it [i.e., reason] and thinks” (hōs echon kai dianooumenon), which Lorenz understands as showing that appetite is strictly speaking incapable of thought. To me it seems that this passage states in a preliminary way the contrast that is spelled out more fully and precisely in I.13, between what is essentially rational on the one hand and what is incidentally and derivatively so on the other (see section 3, fourth paragraph).

Lorenz writes: “The virtuous person’s appetitive desires are as they are not because reason has managed to persuade the non-rational part to participate fully in the person’s pursuit of a flourishing life through activity that expresses the best and most complete virtue. They are as they are because the virtuous person has learned to take pleasure in those things, and only in those things, that one should take pleasure in, and in those ways, and only in those ways, that one should take pleasure in them” ([2006], p. 190). I concur totally with the second quoted sentence, which seems to me an admirable description of how reason persuades the non-rational part to participate fully in the agent’s pursuit of the flourishing life. The appetites are schooled to see certain pleasures as worth pursuing because they are in one way or another fine, and to reject other pleasures because they are in one way or another base; the understanding of those reasons, i.e., the grasp of those inferential connections, is integral to the virtuous person’s pleasure.

The thesis that the appetitive soul is incapable of thought appears prima facie to be suppported by EE 1235b 26–9. (I am grateful to Anthony Price for drawing my attention to this passage. See Price [2011], p. 119.) There, in the context of his discussion of akrasia, Aristotle says that pleasure is an apparent good, since some people think that it is good, while to others it appears (phainetai) good even though they do not think that it is. (I take it that in the last clause Aristotle is describing being attracted to some pleasure that one does not believe to be good, i.e., having what Moss calls the non-doxastic phantasia of that pleasure as good, analogous to the non-doxastic phantasia of the sun as being a foot in breadth. Aristotle explains the occurrence of that latter state in the words “for phantasia (or ‘the phantasia’) and belief (or ‘the belief’) are not in the same part of the soul” (ou gar en tautōi tēs psuchēs hē phantasia kai hē doxa). Those who take this passage to show that the appetitive soul is incapable of belief must take the clause as making the universal claim that phantasia and doxa are never in the same part of the soul. While that is indeed a possible construal of the Greek, there are at least two other possible construals, neither of which supports the universal claim. These are i) “When a pleasure appears good, though one does not believe that it is, appearance and belief are not in the same part of the soul,” and ii) “When a pleasure appears good, though one does not believe that it is, the appearance (viz. the appearance that it is good) and the belief (viz. the belief that it is not good) are not in the same part of the soul.” Each of these is at least as plausible in the context as the construal supporting the universal claim (indeed both are in my view more plausible). This passage therefore offers only weak support for the thesis that the appetitive soul is incapable of thought.

(17) See also p. 210: “[P]roper habituation teaches the non-rational part to obey and wait on the rational part. Thus it trains one to aim at fine things in a special way—namely, in such a way that one’s non-rational impulses will wait on the deliverances of deliberation before pursuing them.”

(18) Even more mysterious is p. 231: “[C]haracter supplies the content of the goal, and intellect merely conceptualizes it”; prima facie that suggests that what character supplies is unconceptualized content, but it is not at all clear what that might be.

(19) Cf. Fortenbaugh [2002], p. 17 “[E]motional response is intelligent behaviour open to rational persuasion”; p. 30 “Emotion … involves judgement and therefore is open to reasoned persuasion and properly classified among cognitive phenomena”; p. 31 n. [In the Eudemian Ethics] “Both reasoning and emotional response are cognitive and therefore human activities.”

(20) So Woods [1986], p. 148, Reeve [1992], pp. 60–61. For fuller discussion see Sherman [1989], ch. 5.

(21) So Reeve [1992], p. 40: “Dialectical arguments are not proofs of first principles…. Instead, dialectic takes principles already justified by experience and made true by reality and ‘clarifies’ them. And this is as true in ethics, where the object of enquiry is the good, as it is in science.”

(22) See Irwin [1978] and [1988], esp. ch. 16.

(23) For fuller discussion of this principle see Gosling and Taylor [1982], ch. 17.2, and Taylor, Pleasure, Mind, and Soul (Oxford, 2008), ch. 6.

(24) For a full discussion of this difficult passage see Dahl [1984], pp. 227–236.

(25) Cf. EN 1142a19, where Aristotle suggests tentatively that the principles <sc.of practical reasoning> come from experience.

(26) So Kenny [1978], pp. 170–171, [1979], pp. 151–152, Engberg-Pedersen [1983], pp. 211–219, and Dahl [1984], pp. 41–45.

(27) See Engberg-Pedersen [1983], pp. 198–199.

(28) For a helpful discussion of this issue, see Striker [2006], pp. 132ff. She argues that the role of practical experience is not to enable the phronimos to act independently of general moral principles, but to supplement (perhaps it would be better to say “to make determinate”) primary moral principles, conceived by Aristotle as identical with principles of general justice.

(29) Dahl [1984], p. 44 expresses this idea exactly: “What nous grasps are particular actions that are good or to be done. Its grasp of these is based on experience and hence on a kind of induction. Implicit in this grasp are universal ends that one should be aiming to promote. One sees the universal ends that one should be aiming at by seeing them in the actions that one recognizes one should perform. One sees them in these actions because, in recognizing that one should perform them, one recognizes that actions like them should be performed. Whether or not one can articulate the content of what one has grasped, one has inductively acquired a universal end.” Cf. Engberg-Pedersen [1983], pp. 218–221.

For a fuller exposition of this conception of the mastery of a concept and its connection with Wittgenstein’s remarks (in Philosophical Investigations I, paras. 143 ff.) on following a rule, see McDowell [1979].

(30) So Sherman [1989], pp. 43–44.

(31) In my [2008] I restrict the role of character to the motivational, ascribing the selection of the end to phronēsis; in that interpretation I follow Allan [1953], who himself acknowledges his debt to Loening [1903], and Gauthier and Jolif [1958/9], vol. 2, p. 557. Cf. Sherman [1989], pp. 80–81. That now seems to me a mistake, in that it fails to do justice to the cognitive character of rationally conditioned desire that I have emphasised in this paper, especially in section 3 and in n. 16. But rationally conditioned desire is none the less conditioned, and ultimately what conditions it is phronēsis.

(32) Some assume that what is supplied after telos is poiei prattein rather than poiei; thus Irwin [1999] translates “[virtue] makes us achieve the end, whereas [prudence] makes us achieve the things that promote the end”; Rowe [2002] “the one causes us to act in relation to the end, the other in relation to what forwards the end”; Moss [2012], p. 157, “the one makes us do the end and the other the things towards it.” Ross rev. Ackrill and Urmson [1980] has “the one determines the end and the other makes us do the things that lead to the end,” which appears to agree with my rendering.

(33) Most commentators (e.g. Engberg-Pedersen [1983], p. 185, Reeve [1992], p. 49, Irwin [1999], p. 111, Bostock [2000], pp. 90 and 224, Moss [2012], pp. 157 etc.) interpret the passage in precisely the sense that I reject. Bostock is typical; he translates a17–19 as follows: “In neither case is it reasoning (logos) which teaches us the first principles, but [in the case of actions] it is virtue, either natural or habitual (ethistē), that teaches us right opinion about the first principle.” The supplementation in square brackets is essential for this interpretation; but this supplementation seems to me unnecessary, and therefore unjustified. Without supplementation it is quite clear that the phrase all’ aretē ē phusikē ē ethistē tou orthodoxein peri tēn archēn applies both to theoretical and to practical thought, not merely to the latter, as Bostock et al. suppose. Aristotle is saying here that in neither kind of thought are the principles proved by reasoning, but are acquired by another kind of excellence. The theme of the surrounding context is certainly the role of ethical virtue and vice in the preservation or corruption of the grasp of ethical principles, but it does not follow that this particular sentence refers to the specifically ethical sphere rather than to both the ethical and the theoretical.

Ross’s original translation (Ross [1925]) agrees with Bostock et al.: “in actions the final cause is the first principle, as the hypotheses are in mathematics; neither in that case is it argument that teaches the first principles, nor is it so here—virtue either natural or produced by habituation is what teaches right opinion about the first principle.” So Gauthier and Jolif [1958/59], vol. 1, p. 208, and Dirlmeier [1969], p. 108. But the revision of Ross by Urmson (Urmson [1984]) favors the reading that I propose: “neither in that case is it reason that teaches the first principles, nor is it so here—excellence either natural or produced by habituation is what teaches right opinion about the first principle.” Müller [1982], p. 271 also favors that reading: his paraphrase runs “In mathematics no less than in practical thought [Praxis] a natural or acquired proficiency [Qualifiziertheit] brings about correct assessment in the area of starting-points.”

(34) “acquired” renders ethistē a19, literally “acquired by habituation.” Here is another passage in which Aristotle says that ethismos is a method of acquiring correct beliefs; but since it is cited as a possible source of such beliefs in the theoretical as well as in the practical sphere, it is likely that it has the more general connotation of “training” than the specific connotation “training of the appetitive soul,” which it has in EN II.

(35) See Sherman [1989], esp. ch. 2.

(36) For a version of this view, see Smith [1996].

(37) So Woods [1986], p. 159, and Broadie [1991], ch. 4.

(38) See Gauthier and Jolif [1958/59], vol. 1, p. 29*, vol. 2, pp. 560–563.

(39) Cf. Monan [1968]. chs. 5 (esp. pp. 111–115) and 8.