Why Aristotle Isn’t a Virtue Ethicist. Living Well and Virtuously in Aristotelian and Contemporary Aretaic Ethics

Abstract

Drawing on Anscombe, in this essay I argue that we should not take Aristotle to be a moral philosopher, nor a virtue ethicist. This is because contemporary virtue ethics has little to do with Aristotelian ethics. While contemporary virtue ethics (or aretaic moral theory, as one may call it) operates on the level of moral and thus categorical norms, Aristotelian ethics—an aretaic life ethics—is primarily concerned with pragmatic norms. The main question for Aristotle is what a good general conduct of life is. The major concern of aretaic moral theory, on the other hand, is to provide a criterion of morally right action and hence to define the concepts of the morally right, the impermissible and moral duty in aretaic terms. This shows that contemporary authors assume a primacy of virtue, while Aristotle assumes a primacy of eudaimonia. I illustrate this distinction by addressing the question of how the virtues benefit their possessor.

Let me begin by making two bold claims, which I shall defend (in a qualified way) in this paper. First, Aristotelian ethics isn’t a moral theory. Second, it’s no virtue ethics either. These claims are bold because today it is considered uncontroversial that virtue ethics is the third branch of moral theory, and that Aristotle is the ultimate representative of virtue ethics.

Naturally, the truth of both of my assertions depends heavily on what I mean by moral theory and by virtue ethics. Both terms will become clearer throughout this essay. The short answer, however, is this: A moral theory is concerned with moral norms and moral reasons. This, of course, in turn raises the question of what I mean by moral norms and moral reasons. Somewhat more precisely, a moral theory is concerned with categorical norms and categorical reasons of a specific kind, namely, those that are neither legal nor conventional in nature.¹ By virtue ethics, on the other hand, I mean ethical theories such as those developed by authors like Michael Slote, Christine Swanton, Linda Zagzebski, and Rosalind Hursthouse in the course of the last decades. (In what follows, I will focus primarily on Slote, Zagzebski, and Hursthouse because they strike me as paradigmatic for contemporary virtue ethics. Moreover, Hursthouse herself explicitly describes her theory as neo-Aristotelian).²

To justify my bold claims, I shall first say a few words about the development of contemporary virtue ethics. Then, taking Slote, Zagzebski and Hursthouse as examples, I will briefly outline the basic characteristics of contemporary virtue ethics. Next, I will demonstrate how it differs from Aristotelian ethics. Unlike contemporary virtue ethics, the latter is not concerned with moral norms and reasons (i.e., categorical norms and reasons of a special kind). This is evident from the fact that contemporary virtue ethics assumes a primacy of virtue over eudaimonia, whereas Aristotle assumes, conversely, a primacy of eudaimonia over virtue. In the final part, I will discuss this difference on the basis of the question of how the virtues benefit their possessor.

1 The Aretaic Turn

In her seminal 1958 essay ‘Modern Moral Philosophy’, G.E.M. Anscombe advances three theses: She claims, first, that doing moral philosophy is a futile endeavor until we sufficiently understand the nature of actions, i.e., until we have an adequately elaborated theory of action. Anscombe’s second claim is that we should do away with notions of moral duty and moral right and wrong, as well as the moral ought, to the extent that this is psychologically possible. Third, Anscombe argues that the differences between the well-known English moral philosophers from Sidgwick to the present day—meaning, of course, the present day of 1958, but one could easily continue the line to our present day—are negligible (1958, p. 1). I shall now turn primarily to her second thesis.

The reason why, according to Anscombe, we should drop the notions of moral duty, moral right and wrong, and moral ought is that these notions are remnants of conceptions of ethics from an earlier time. This is because they presuppose a ‘law conception of ethics’ according to which moral commands and prohibitions are enacted by God (1958, p. 6). However, these divine moral commandments—and the related metaethical conceptions—have lost their justification in a society like ours, which has long since given up the belief in laws given by God (1958, p. 8).

Anscombe notes that the notions ‘should’, ‘ought’, and ‘needs’ do in themselves refer to some good. However, they do not necessarily refer to a moral good, nor do they in themselves have any moral meaning (1958, p. 5). For example, mechanical machines need to be oiled in order to function well, so we should oil them. In contrast, the notions ‘should’, ‘ought’, and ‘need’, when used in a genuinely moral sense, are roughly equivalent to terms such as ‘being obligated’ or ‘being required to do something’—in the sense in which one can be legally obligated or required to do something (1958, p. 5). By contrast, Anscombe notes that in Aristotle notions such as ‘ought’ and ‘should’ aren’t present in this genuinely moral sense.

How is it, then, that the moral ‘ought’ and ‘should’ have entered our ethical vocabulary? The reason for this is Christianity, as Anscombe explains. Christian ethics, she argues, holds a ‘law conception of ethics’ (1958, p. 5), which is rooted in the fact that it draws its ethical concepts from the Torah and the Bible.³ According to Anscombe, the reason that we still hold to these notions today is that while we may have lost faith in God as a moral legislator, we continue to conceive of our moral judgments as having legal force because our practice of moral judgment has been exposed to the influence of Christianity for so long. Thus, the law conception of ethics is deeply embedded in our speech and thought (1958, p. 5).⁴

However, this poses a problem: a law conception of ethics requires a legislating and law-enforcing authority. Because ‘[n]aturally it is not possible to have such a conception unless you believe in God as a law-giver’ (1958, p. 6), without reference to such a legislating authority we lack the justification to conceive of moral norms as having legal force and hence to maintain the moral meanings of ‘ought’, ‘should’, and ‘need’. These notions, Anscombe argues, run the risk of degenerating into mere verbiage without any actual content. In order to prevent the moral ought from degenerating into a notion that does not contain any ‘intelligible thought: a word retaining the suggestion of force, and apt to have a strong psychological effect, but which no longer signifies a real concept at all’ (1958, p. 8), various attempts—futile in Anscombe’s view—were made in the course of the Enlightenment in order to substitute God as a law-giving authority.

In light of this history of failure to maintain a law conception of ethics without belief in God as moral legislator, Anscombe argues for dispensing with the concepts (or non-concepts, that is) of moral ought and should altogether (1958, p. 8). Because we have no substantial concept of the morally good and bad, we should seek to do ethics without making use of such meaningless verbiage. The possibility of such an ethics can be seen in Aristotle. A closer look at Aristotle’s ethics reveals how different it is from our contemporary conception of ethics. Hence, Anscombe (1958, p. 1) states:

Anyone who has read Aristotle’s Ethics and has also read modern moral philosophy must have been struck by the great contrasts between them. The concepts which are prominent among the moderns seem to be lacking, or at any rate buried or far in the background, in Aristotle. Most noticeably, the term ‘moral’ itself, which we have by direct inheritance from Aristotle, just doesn’t seem to fit, in its modern sense, into an account of Aristotelian ethics.⁵

With this in mind, Anscombe argues that ethics should dispense with the notion of moral ought (as well as the notions of permissibility and duty) and instead make use of so-called thick concepts, such as ‘just’, ‘benevolent’, ‘brave’, ‘honest’, and so on. (Sometimes it is said that the deontic notions Anscombe is criticizing are, by contrast, thin concepts. But Anscombe is far more radical. According to her, they aren’t concepts at all, but merely meaningless conceptual shells, non-concepts, that is. Because they themselves have no meaning, but merely a normative component, they gave rise to the developments in ethics in the first half of the twentieth century, especially to Ayer’s emotivism and Hare’s prescriptivism).

This reconstruction of Anscombe’s call to dispense with the moral ought and instead to pursue an ethics free of deontic notions is anything but trivial. This can be seen, for example, by the inferences Robert B. Louden draws from Anscombe’s essay in his ‘On Some Vices of Virtue Ethics’. Here is how he interprets Anscombe (1984, p. 228):

On the Anscombe model, strong, irreducible duty and obligation notions drop out of the picture, and are to be replaced by vices such as unchasteness and untruthfulness.

So far, so good. But in the very next sentence, Louden continues as follows:

But are we to take the assertion literally, and actually attempt to do moral theory without any duty whatsoever? On my reading, Anscombe is not really proposing that we entirely dispose of moral oughts. Suppose one follows her advice, and replaces “morally wrong” with “untruthful,” “unchaste,” etc.. Isn’t this merely shorthand for saying that agents ought to be truthful and chaste, and that untruthful and unchaste acts are morally wrong because good agents don’t perform such acts? The concept of the moral ought, in other words, seems now to be explicated in terms of what the good person would do.

This leads to two interpretations of Anscombe’s claim: According to one interpretation, we should take Anscombe to mean that we are advised—as I have already sketched—to engage in ethics apart from moral theory, that is turning to aretaic concepts instead of deontic ones, in order to engage in ethics in the Aristotelian way. In doing so, we get along without the concepts of the moral ought, duty, the morally right and the like.

According to the second—Louden’s—reading, the call is to pursue ethics without renouncing these very concepts of moral ought etc., but to ground them on the basis of aretaic concepts. Within the framework of such moral theory, we hence have to reduce deontic concepts to aretaic ones, whereas within the framework of an ethics in the sense of the first interpretation, we try to eliminate deontic concepts in favor of aretaic ones. In what follows, I shall call the first interpretation the elimination thesis, and the second, Louden’s, the reduction thesis.

2 Contemporary Virtue Ethics

When Louden assesses contemporary virtue ethics, he does so in the context of the developments in virtue ethics that had taken place over the course of almost thirty years by the time his text was published in 1984. Apparently, these developments had given rise to a rather different understanding of the aspirations of virtue ethics than Anscombe originally had in mind. Now, let me outline in the following how the majority of contemporary virtue ethicists depart from Anscombe’s view that we should do away with the notion of moral ought, as well as moral right and wrong (i.e., the elimination thesis), and instead develop a moral theory in which they attempt to ground the deontic notions in terms of aretaic concepts. A theory of this sort would be a virtue ethics in line of the reduction thesis.

As John Hacker-Wright points out in his (as yet under-appreciated) essay ‘Virtue Ethics without Right Action: Anscombe, Foot, and Contemporary Virtue Ethics’, this departure becomes apparent when one considers that almost all canonical contemporary virtue ethicists have sought to provide a criterion of right action on a virtue-ethical basis (2010, p. 209).⁶ The first systematic attempt to provide a contemporary account of virtue ethics is made by Michael Slote in his From Morality to Virtue. While at first it seems as if he is taking up Anscombe’s call to eliminate deontic concepts from ethics (for example, when he writes that he wants to eliminate ‘specifically moral aretaic concepts in favor of ‘neutral’ aretaic concepts’ (1992, p. xvi)), it quickly becomes clear that his project is nonetheless one in which he does not want to dispense with deontic notions (‘moral ought’, ‘morally right’, ‘morally wrong’), but wants to reduce them to aretaic concepts. Considering what an ethics nowadays has to provide, he notes (1992, p. xvi):

However, since all the ethical views under serious consideration today give an important or central role to deontic notions, it seems almost unimaginable that a proper ethics should make no use of them. We might grant that an ethics without deontic concepts could present an important fragment of the ethics we ultimately seek, but we can’t imagine a complete ethics doing without such notions altogether.

In order to derive deontic concepts in the form of ‘v-rules’ (virtue-rules) from aretaic ones, Slote devotes an entire chapter to this project.

The way deontic notions are derived from aretaic concepts is remarkably alike among virtue ethicists. They all try to define the concept of the morally right by the concept of the virtuous agent (or other aretaic concepts). All of these definitions ring surprisingly similar. For example, Rosalind Hursthouse (1999, p. 28) defines right action in analogy to the definitions used by deontologists and utilitarians:

An action is right iff it is what a virtuous agent would characteristically (i.e. acting in character) do in the circumstances.

In Virtues of the Mind, Linda Zagzebski sets out in a very similar vein (1996, p. 233):

[A] right act is, roughly, what a virtuous person would or might do in a certain situation.

Discussing the ambiguity of the concept of the morally right (which can mean both that something is morally obligatory and that something is merely morally permissible), she comes to a definition of the right action. From that she also derives the concepts of moral obligation and prohibition (1996, pp. 239–240):

A right [i.e., permissible] act, all things considered, is what a person with phronesis might do in like circumstances. A wrong act, all things considered, is what a person with phronesis would not do in like circumstances. A moral duty, all things considered, is what a person with phronesis would do in like circumstances.⁷

In doing so, Zagzebski highlights that she even goes beyond the claim to merely give a criterion of right action. For, according to Zagzebski, a pure virtue theory like hers ‘treats the rightness of an act as strictly dependent upon virtue’ (1996, p. 232; my italics).

Such aretaic reductions are notoriously problematic. One such problem is the explanation of moral dilemmas. (In this respect, both Hursthouse and Zagzebski feel compelled to defend their definitions against the objection that they are unable to explain such dilemmas). Suppose a virtuous agent finds herself in an irresolvable dilemma, that is, a dilemma in which both courses of action are wrong and equally bad. Because in such a situation the agent is forced to choose one of the two wrong actions, he will necessarily perform a wrong action. By definition, however, those actions are right which the virtuous agent might choose. This means, however, that the course of action which the virtuous agent chooses would be both (ex hypothesi) wrong and (by definition) right at the same time. If such irresolvable dilemmas can occur, aretaic reductions are in trouble. Hursthouse attempts to solve the problem by showing that a virtuous agent cannot emerge from such a situation unscathed. Her life will inevitably be marred by her horrific act (1999, p. 74). And Zagzebski likewise admits that in the case of tragic dilemmas, her aretaic definitions of deontic notions would need to be modified to account for the virtuous person’s emotional reaction (1996, p. 241). Dilemmas, however, are not the only problem for aretaic reductions.

One of the alleged merits of virtue ethics is that it gives an adequate account of supererogatory action (Horn 1998, p. 120). However, upon closer examination, this too gives rise to problems. First, one may note that Hursthouse’s definition of right action also leads to the definitions of the morally wrong and moral duty. (This is because the notions of ‘right’, ‘wrong’, and ‘obliged’ may all be substituted by the notion of permissibility: ‘φ-ing is permissible’, ‘φ-ing is impermissible’, and ‘non-φ-ing is impermissible’.) Given their being substitutable, if a problem occurs for any one of the notions, the problem occurs for all of them.

Here is the problem: while virtue terms are perfectly capable of accounting for supererogatory action, they fail to do so in conjunction with the proposed aretaic definitions of deontic notions. Take, for example, the concept of mercy, which is a typical aretaic concept. A virtuous agent will be merciful and will (typically) perform merciful acts (insofar as they are adequate). However, we also had defined that those actions are obligatory that the virtuous person would do in the given circumstances. Hence, if the virtuous person were to be merciful under the given circumstances, the respective action would be obligatory. But this cannot be. For the very point of the concept of mercy is that merciful acts are invariably supererogatory, that is, they can never be obligatory. Yet the problem is even more severe. Supererogatory actions are (by definition) those that go beyond what is obligatory and thus deserve special praise and, in some cases, even admiration. Virtuous agents are such agents whose actions often merit this form of praise. However, because the actions performed by the virtuous are (by definition) obligatory, and supererogatory actions (by definition) can never be obligatory, the virtuous agent can never perform a supererogatory action. This strikes me as absurd.⁸

The problem is due to the fact that the aretaic definitions of contemporary virtue ethicists establish a conceptual relation between virtue and moral rightness (and also wrongness and duty). This is because they approach virtue in the context of a moral theory (that is, as a system of categorical norms). One requirement of a sound moral theory is to provide action guidance. However, according to contemporary authors, this can only be done by means of a criterion of right action (Hursthouse 1999, p. 26). Once we have an aretaic criterion of right action, the other deontic notions can be derived from it, and moral theory may then be pursued in the usual way, like utilitarians and deontologists do. In this respect, it comes as no surprise that contemporary virtue ethicists conceive of their theories primarily as rivals to deontology or utilitarianism—that is, as moral theories.⁹ Given this, it is quite apt for Louden (1984, p. 228) to state, with regard to the contemporary project of virtue ethics:

[C]onceptual reductionism is at work in virtue ethics too. Just as its utilitarian and deontological competitors begin with primitive concepts of the good state of affairs and the intrinsically right action respectively and then drive secondary concepts out of their starting points, so virtue ethics, beginning with a root conception of the morally good person, proceeds to introduce a different set of secondary concepts which are defined in terms of their relationship to the primitive element. Though the ordering of primitive and derivatives differs in each case, the overall strategy remains the same. Viewed from this perspective, virtue ethics is not unique at all.

It is for this reason that I believe the term ‘virtue ethics’ has been morally contaminated by contemporary authors. (I will return to this later in explaining why we should take Aristotle to be an aretaic ethicist rather than a virtue ethicist.)

In my view, this development was a mistake.¹⁰ First, I don’t think that the study of virtue to which Anscombe has appealed must necessarily take place within the framework of a moral theory. Second, given this, I also don’t think that the conceptual link between virtue and moral rightness that is often assumed really exists. In fact, virtue in Aristotelian ethics had nothing to do with what is morally right (or with morality in our modern sense) whatsoever. This, after all, is the very thesis of Anscombe and Williams. Hence, in the following section I want to illustrate how we (or rather Aristotle) can do ethics without deontic terms, that is, without being a moral philosopher or even a virtue ethicist.

3 Aristotle’s Aretaic Life Ethics

If Anscombe and Williams are right, the deontic or moral notions of rightness, permissibility, and duty do not matter in ancient ethics, particularly in the ethics of Aristotle, and the concept of morality ‘just doesn’t seem to fit, in its modern sense, into an account of Aristotelian ethics’. Why should we think they are right?

First of all, let us consider the role of ethics within scientific inquiry as a whole for Aristotle. After all, Aristotle was not only concerned with philosophy, but also with questions of physics, medicine, biology, and many other fields. For Aristotle, everything—philosophical as well as non-philosophical—is ultimately an explanation of the natural world in which we find ourselves as human beings. This is the context within which his ethics must also be conceived. Ethics, for Aristotle, is the explanation of man’s place in the world of which he is a part. Hence, ethics has the same naturalistic-explanatory status as physics or biology.

In this respect, it is hardly surprising that in his Nicomachean Ethics Aristotle opens with an observation of action theory, namely that all action aims at some good, which is why the aim of all activity rightly is called ‘good’ (1094^a 1–5).¹¹ Some things we do for their own sake, others to achieve some further end. This results in a chain of teleological explanations of action. However, this chain cannot be continued indefinitely, because all endeavors would become empty and vain if it did (1094^a 17–23). The last link in this chain of teleological explanations of action, which we try to realize with all our actions, will thus be the final good (1094^a 17). It is the final cause (τέλος) of all our actions.

Aristotle already makes clear at this point the relevance this observation has for ethics: If in all our actions we are always oriented toward some good, then it will be of the utmost importance for us to identify this good and to figure out how to attain it (1094^a 22–25). With this, the underlying mission of ethics, as Aristotle conceives it, is sketched out: Explanatorily, that is, as a theory of action, ethics tells us towards which end our actions ultimately aim, thus simultaneously contributing practically to a flourishing life. (This already highlights the difference between Aristotelian ethics and contemporary virtue ethics. Unlike the latter, Aristotle starts with the question of the flourishing life).

Aristotle then moves on to what the ultimate good actually consists of, namely happiness or eudaimonia.¹² Thus, we can state:

[i] Man’s ultimate good is eudaimonia.

As Aristotle correctly notes, however, identifying of the greatest good with eudaimonia does not shed much light upon what the greatest good really is: ‘to say that happiness [εύδαιμονία] is the chief good seems a platitude, and a clearer account of what it is is still desired’ (1097^b 22–23). What becomes clear in this context is that ‘happiness’ might not be a particularly good translation for the concept of eudaimonia. For by happiness today we all too often mean something emotional and subjective.¹³ The Aristotelian concept of eudaimonia, however, is something objective. While I may feel and consider myself happy, I may still be mistaken about whether my life is ‘eudaimon’. In this respect, a more apt translation possibly would be ‘flourishing’. However, this term too has disadvantages. Hence, I will simply keep using the term eudaimonia and sometimes, although it might be somewhat cumbersome, I will translate eudaimonia as ‘good life’.

Out of this difficulty, Aristotle proceeds to the question of what, according to the people, eudaimonia consists in: in the pursuit of pleasure, of honor, or in contemplation (1095^b 15–20). This digression into different ways of life, which at first may seem odd, becomes intelligible when we realize once again that Aristotle's ethics is not a moral theory interested in action, but an (aretaic) ethics of life. (In this respect, contemporary virtue ethicists are indeed correct in saying that Aristotelian ethics does not focus primarily on actions. However, contrary to them, Aristotle also does not focus primarily on character, but on life as a whole).¹⁴ If what ultimately matters in ethics is the good life, then it would be futile to deal with isolated assessments of action or with any criterion of right action. In short, the project of contemporary virtue ethics does not serve any purpose at all if we are concerned with life rather than action.

From this point of view, Aristotle’s remarks become quite clear, which, from the point of view of a moral theory, must seem at least peculiar. One example is the idea that in order to lead a good life, external goods are required. In the context of a moral theory which is concerned only with the permissibility and impermissibility of certain actions, the matter of external goods serves no purpose whatsoever. So does the question Aristotle raises about the extent to which ill-bred children or poor ancestry are detrimental to the good life and how beauty is beneficial to it (1099^b 1–10). All of these are matters for which (apart from some special circumstances) one cannot be held responsible. For the moral theorist, considerations of this kind are futile. After all, insofar as all of these matters are beyond the control of the agent and for which he thus cannot be held responsible, we cannot blame him morally in this regard. However, within the framework of an ethics of good life, as developed by Aristotle, those questions regarding external goods—having good children, being of decent birth, making bad friends, dealing with difficult social circumstances, going to war, being in love, …—are of utmost importance. After all, the prosperity of our lives does not only rely on what we may be held responsible for and what we can be blamed for. Sometimes it is the greatest fortunes that enrich our lives, and quite often those matters we cannot control fall upon us as calamities. For all this tragedy of human life the moral theorist remains blind, for he only knows his rules of conduct, on the basis of which right action is assessed. For the life ethicist, however, right action can only ever be one element among many that make for a good life.

Even before discussing the importance of external goods regarding the good life, Aristotle makes a point that links his ethics directly to his metaphysics. For according to him, ‘human good turns out to be activity of the soul according to its virtue [ἐνέργεια ψυχῆς κατ’ ἀρετήν]’ (1098^a 16–18). We can note:

[ii] The ultimate good for man is activity of the soul according to its virtue [ἐνέργεια ψυχῆς κατ’ ἀρετήν].

In conjunction with [i], we get:

[iii] Eudaimonia consists in the activity of the soul according to its virtue.¹⁵

At first, this sounds rather cryptic. So we need to ask what exactly Aristotle means by activity (ἐνέργειᾰ), soul (ψῡχή) and virtue (ἀρετή). This brings us directly to his metaphysics. Given the Aristotelian metaphysics and philosophy of nature, however, it quickly becomes clear what Aristotle means. Activity (ἐνέργειᾰ) is the realization of a potential to be such-and-such, that is, the actual exercise of those functions which make, say, a human being a human being. This is a direct reference to his concept of ἔργον and his account of the four causes (1097^b 23–34). After all, for any given being, its goodness consists in the exercise (ἐνέργειᾰ) of its specific function (ἔργον). Consequently, Aristotle, in the context of the quoted passage, asks whether man has a specific function (1097^b 28–33):

Have the carpenter, then, and the tanner certain functions or activities, and has man none? Is he born without a function? Or as eyes, hand, foot, and in general each of the parts evidently has a function, may one lay it down that man similarly has a function apart from all these?

Because man is an animate being, we must look more closely at the concept of the soul to determine its proper function. With the concept of the soul (ψῡχή), Aristotle points us to his theory of the soul, which he elaborates primarily in his De anima, but which he briefly outlines again in the context of the quoted passage. What distinguishes man from the other animate beings is that he possesses reason. While he shares the nutritive faculty with plants and animals and the desiderative faculty with animals, he additionally has reason, with one part (or aspect) of his soul merely obeying reason and the other actually possessing it proper (1098^a 1–7).

However, we must keep in mind that for Aristotle the concept of the soul simply means being alive. In this sense, we can say that man, unlike plants and animals, lives a life in light of reason. This is to say that the function (ἔργον) of man lies in a rational way of life.

The concept of virtue (ἀρετή) is once again metaphysically charged. For the concept of virtue refers primarily to the suitability of a thing as such a thing.¹⁶ A virtuous person, therefore, is one who is fit as a human being. This suitability is linked to a being’s function (ἔργον). Hence, a knife is good as a knife if it serves its function—cutting—in a particularly well manner. The metaphysical aspect of this thought becomes particularly clear in the Eudemian Ethics (EE 1219^a 1–5):

For example, a cloak has a virtue, since it has a function and use, and its best state is its virtue. The same applies to a boat and a house, and so on, and hence to the soul, since it has some function.

In other words, a being is good insofar as it fulfills the function proper to it. Alternatively, one could also say: A being is good insofar as it fulfills its formal cause, that is, its causa formalis, which makes it this very specific being. This is the context to which the concept of ἐνέργειᾰ refers in the quoted passage. In summary, we can state: Eudaimonia for man is the rational conduct of life. That is to say:

[iv] Eudaimonia is the rational exercise of those activities of life typical for human beings.

What, then, is the rational exercise of the activities of human life? To illuminate this, we must take a closer look at the virtues. For with regard to the two faculties of the soul in man that have a share in reason (or possess it proper), one can see that the human ἀρετή, i.e. human suitability, breaks down into two domains: the suitabilities (or virtues) of character and the suitabilities (virtues) of thought.¹⁷

While both domains are essential for the realization of eudaimonia, with regard to the central issue, I shall focus only on the virtues of character. What is their role in Aristotelian ethics? As said, the ethical virtues are a manifestation of man’s rationality. They always relate to some emotion which humans frequently face.¹⁸

At this point one could argue that only the virtues of thought are manifestations of reason. The virtues of character, on the other hand, are allocated to the part of reason that is not itself rational.¹⁹ And yet I think that the ethical virtues are equally manifestations of reason. Unlike humans, animals, which also possess the desiderative part of the soul, lack the virtues of character. Humans have the virtues of character only because reason has an influence on the desiderative part of their soul. In light of a transformative model of the soul, one can say that the emotional make-up of man is not the same as that of the animal, but is modified by his possession of reason—i.e. the rational part of the soul (Martin 2011). His desire is not the desire of the animal, but a rational one.

In this respect, the virtues consist in (a disposition to) a successful handling of certain domains of human life, the grounding experiences, as Martha Nussbaum (1988) calls them. That’s why I think Foot is right when she says that there would be no virtue of bravery or temperance for humans if they did not tend to run away too quickly in dangerous situations, or if they did not regularly tend to give in to their desires, and that the virtues of character are thus correctives. (Foot 1977, pp. 8–9). Once again, a close examination of human nature is in order to figure out with regard to which emotions humans are in need of these correctives and which items are hence included in our list of virtues. It turns out, for example, that people are often too heavily (and less often too weakly) determined in their actions by fear (the corresponding corrective is bravery), are intemperate or obtuse with regard to pleasure (temperance), are stingy or lavish with money (generosity), are too easily or hardly angered (gentleness), and so on. The exact elaboration of what the appropriate, that is, rational, conduct with regard to the emotions pertinent to human life consists of takes up the greatest portion of the Nicomachean Ethics, following the identification of the chief good as eudaimonia. What is striking, however, is that it does not at any point involve moral obligations.²⁰

A common objection is that Aristotle frequently uses the ancient Greek term δεῖ throughout the Nicomachean Ethics, which denotes something that should or ought to be done or which is valid. And he also repeatedly stresses the rightness of reasons and desires. So, does he after all have a concept of the moral ought and of the morally right? I think we should not be too hasty in assuming that Aristotle is thus concerned with the concepts of moral ought and moral duty, or that Aristotle provides an aretaic criterion of morally right action, as contemporary virtue ethicists do. Remember Anscombe’s second claim: she didn’t believe the concepts of ‘should’ and ‘ought’ to be problematic per se. However, she did point out that in ethics we use them in a special, moral sense. It is this particular employment of the moral ought that she rejected. Thus, Aristotle is perfectly capable of asserting that certain actions ought to be done and that it is right to have certain desires, without this implying that he is speaking of a moral duty to act or feel in a certain way. ‘Should is simply should’, as Williams (1985, p. 6) says. The same applies to the ancient Greek term δεῖ. This is commonly translated as ‘(moral) duty’. But here, too, we shouldn’t be too hasty. After all, its literal meaning is ‘to be bound’ or ‘to be fitting’. It is Anscombe and Williams who urge us to be cautious about translating terms from Aristotelian ethics directly into concepts of modern moral philosophy. Boundness doesn’t imply moral boundness. For it to be some moral obligation, it would need to be, as I will argue below, a non-purposive (or categorical) norm. That is, an obligation that holds independently of the end the agent is pursuing. In the context of Aristotelian ethics, however, we are concerned merely with a boundness with regard to eudaimonia as the ultimate end of human action, that is with pragmatic norms.²¹

In order to illustrate the difference between what ancient ethics as life ethics requires of us and what modern virtue ethics does, we need to consider the difference between various types of norms. From this we can see that ancient life ethics operates at the level of pragmatic (ethical) norms and contemporary virtue ethics operates at the level of categorical (moral) norms.

Kant’s Grundlegung zur Metaphysik der Sitten (GMS, AA 04: BA 39–44) provides a helpful discussion of these types of norms. Hypothetical norms are such norms whose validity (i.e., first-personal normative relevance) depends on the will of the agent. For example, a typical hypothetical norm is: ‘If you want to build a sturdy tower of building blocks, you should stack them horizontally.’ In contrast, the validity (first-personal normative relevance) of categorical norms is independent of the agent’s will. For example, a typical categorical norm is: ‘Thou shalt not take anything that does not belong to thee.’ Notice that hypothetical norms usually take the form of conditionals, whereas categorical norms do not. ‘Usually’ because one cannot tell what kind of norm it is just from its syntax (Mackie 1977, p. 28). This is because categorical norms can be expressed in the form of conditionals (‘If something doesn’t belong to you, don’t take it.’) and hypothetical norms will often occur to us in the form of ellipses (‘You should stack the blocks horizontally.’).

Pragmatic norms are a particular kind of hypothetical norms since they relate to an end the pursuit of which can usually be presumed: the good life. Hence, according to Kant, the validity of technical norms (hypothetical norms in the narrow sense) is problematic—the agent’s will cannot be taken for granted—whereas the validity of pragmatic norms is assertoric. However, because they are categorical norms, the validity of moral norms is completely independent of the agent’s will. One can also see this from the fact that we hold them against someone especially when they do not intend to abide by them (Luckner 2005, p. 41).

As already mentioned, it is not possible to tell with certainty based on the syntax which type of norm we are concerned with. This presents us with a challenge, particularly with regard to pragmatic norms. Whereas technical norms (i.e., hypothetical norms in the narrow sense) only sometimes appear to be categorical norms, that is, moral or legal commands, pragmatic norms do so rather commonly. For since the agent’s will to live a good life can almost universally be presumed, we may likewise almost universally omit the antecedent of the conditional (‘If you want to live a good life, …’). Thus, in most cases, ‘If you want to live a good life, be brave, generous, etc.’ turns into ‘Be brave, generous, etc.’ This injunction often looks like a moral imperative, when in fact it is merely a pragmatic imperative, in other words, an advice of prudence. But just because we cannot see that it’s a pragmatic norm doesn’t mean it’s not.

The key point with regard to Aristotle is that we can (and, I think, should) conceive of his ethics in terms of a system of pragmatic norms. With exception of Book V—in which, however, he is concerned with legal norms rather than moral ones—one does not find any moral precepts in Aristotle’s Ethics. Moreover, Aristotle does not give any justification for why we should care about the virtues. The orientation towards the good life is a precondition for the requirements of the virtues (‘Be brave, temperate, gentle, …’) to bear any normative relevance to me. If I didn’t care about the good life, I wouldn’t have to care about what virtue demands.²²

This also highlights the difference between the moral norms of moral theory and the pragmatic norms of Aristotle’s life ethics. Remember, we hold moral norms against people especially when they don’t want to abide by them. Especially in case I intend to betray my best friend, I will be reminded that it is morally wrong to do so. If I do it anyway, I must expect the appropriate reactions from my surroundings, above all social disdain. But it is different in the case of pragmatic norms, i.e. prudent advice. If a friend tells me to be more moderate and hence to smoke less, and at the very next occasion I light up a cigarette nonetheless, he is in no position to accuse me of having done something wrong morally. He may accuse me of acting foolishly, and say that it would only be for my own good if I followed his advice. But advice, unlike moral precepts, does not bear unconditional validity. If I want to live a good life, I should follow his advice and smoke less. But if (for whatever reason) I don’t care about living a good life, I may not care about his advice either. However, even if I don’t care whether I wrong my friend by betraying him, it still is morally wrong to do so, and social disdain is in order.

Yet we do not even need to believe that one can seriously dismiss the will to live a good life. Kant, after all, believes that pragmatic norms hold assertorily because the will to live a good life is given a priori. But even if this will were always given, the demands of virtue would still be pragmatic ones, not categorical ones. Just because the agent’s will toward which a norm is directed is invariably given does not mean that the norm holds independently of that will.

This interpretation of Aristotelian ethics is in line with Williams’s remark that the ancient Greeks had no concept of morality in our modern sense, that is, as a class of reasons distinct from all other reasons. And it also is consistent with Anscombe’s claim that in Aristotelian ethics we can see how to do ethics without the notions of moral ought and obligation. If we take Aristotle’s ethics in the sense I have indicated, we can also eschew the problem of the moral legislator. However, as soon as we take the demands of the virtues to be categorical norms and hence moral precepts, we enter into a law conception of ethics and must face the question of where the unconditional, that is, categorical, validity of these precepts stems from.

Once we acknowledge that Aristotelian ethics is a system of pragmatic norms, however, we must concede that Aristotelian ethics and contemporary virtue ethics are ultimately two different kinds of aretaic ethics. I say aretaic because with the rise of contemporary approaches, the term virtue ethics has grown to designate specifically those contemporary approaches. As Hursthouse states at the beginning of her On Virtue Ethics, ‘‘Virtue ethics’ is a term of art’ (1999, p. 1). If contemporary virtue ethics differs so radically in its metanormative foundations from Aristotelian ethics, perhaps we should not refer to Aristotle as a virtue ethicist. However, because virtue (ἀρετή) figures so prominently in his ethics, I propose to refer to it as an aretaic life ethics. The project of contemporary virtue ethics, on the other hand, is that of an aretaic moral theory.

I shall now close by illustrating this distinction on the basis of a heavily debated example of contemporary virtue ethics: the question of how the virtues benefit their possessor.

4 The Virtues Benefit Their Possessor

Within the discourse of virtue ethics, it is basically uncontested that the virtues benefit their possessor. But how exactly this claim can be construed without running into serious difficulties is not entirely clear. This is due to what Philippa Foot describes as the problem of the tight corner: There may be situations in which virtue requires us to sacrifice our welfare or even to give our lives. Then how can it be that in such cases the virtues benefit their possessor?

I argue that the problem arises because of a misunderstanding of the claim that the virtues benefit their possessor. Most authors adopt a modern, subjective conception of happiness. For them, the question of how the virtues benefit their possessor is equivalent to asking, ‘Does it pay to be moral?’ This reveals an instrumental and moral conception of the virtues. ‘Paying off’ refers to something external that in some sense the virtues can ‘produce’. This represents a dominant conception of happiness, according to which happiness is a state that can be attained by virtue. Yet, the question is, who is paying if being good is to be rewarding? This moral conception of the benefit-claim takes us back to Anscombe’s problem of the moral legislator. If we conceive of the benefits of the virtues in this way, then there seems to be the underlying assumption—whether conscious or not—that God or the universe or whoever cares that we behave well and that they will reward us for it by making our lives pleasant. According to this view, the relation between virtue and the good life is thus a causal one: if I behave well, it will come about that my life will be pleasant.

This, however, immediately raises the problem that (1) virtue sometimes requires of us to sacrifice our own well-being (Foot’s problem of the tight corner) and (2) even the wicked can enjoy a pleasant life. How can this be? Various solutions have been offered to address this issue. Hursthouse, for example, suggested that the virtues only benefit their possessor ‘for the most part’. Like following your doctor’s advice to quit smoking, it is the same with the virtues. Even people who smoke can live a long, healthy life. And refraining from smoking does not guarantee that you will enjoy a long and healthy life. But refraining from smoking is the only reliable way to maintain good health (1999, p. 173). And even though the wicked sometimes lead pleasant lives and the virtuous sometimes must sacrifice their well-being for the sake of virtue, virtue is the only reliable way to happiness. I believe this solution to be absurd. There are at least three problems with it:

• (i) Those who give up smoking do so primarily for the sake of health. But the genuinely virtuous person isn’t virtuous for the sake of happiness. Those who refrain from stealing, while always eyeing their personal gain, aren’t genuinely virtuous; they merely act like the virtuous person (1105^a 5–10). Hursthouse does point out that she doesn’t take the benefit-claim to provide a motivational reason for acting well (1999, p. 170). But by conceiving of eudaimonia as a good that is to be produced, it nonetheless takes on a motivational role.

• (ii) The alleged solution does not meet the benefit-claim. This is because in the case of non-smoking, both relata are conceptually independent of each other. Non-smoking is not defined by the concept of health. Virtue, however, according to the benefit-claim, is defined by its contribution to eudaimonia. The concept of virtue is thus conceptually dependent on the concept of eudaimonia (Copp and Sobel 2004, pp. 529–530; Halbig 2013, p. 41). Hence, it is a conceptual relation rather than a causal one, as suggested by Hursthouse.

• (iii) Even if we grant that the connection is to be taken as a causal one: Why should we believe that (‘for the most part’) everything will turn out fine just because we act well? After all, it is precisely the problems mentioned—the wicked living well and the virtuous perishing because of their virtue—that challenge this assumption. Claiming that the virtues are the only reliable way to achieve happiness is merely an assertion without any justification. Why would the universe (or God or whoever) care to pay for our good deeds—be it only for the most part?

The problem stems from a different view of what role virtue plays in aretaic ethics. Within aretaic moral theory, contemporary authors assume a primacy of virtue over eudaimonia. The basic concept is that of virtue. The concept of the good life either plays no role at all or is secondary in terms of reasoning—as, for instance, in the case of Hursthouse. For her, the concept of eudaimonia only serves to clarify which items are to be on the list of virtues. The concept of eudaimonia thus further specifies the concept of virtue. Aristotle, on the other hand, is primarily concerned with the concept of eudaimonia as the chief human good. That is the fundamental concept, which in turn is to be further elucidated by the virtues (in the sense that eudaimonia expresses itself in the virtuous—i.e., suitable—life). So, while contemporary authors assume a primacy of virtue (or, as it is more often said, a primacy of character), Aristotle is interested in character and virtue only indirectly. Primarily what he is concerned with is the good life. It is, so to speak, mere coincidence that in his ethics the concept of virtue ultimately becomes so vitally important.

With that being said, what help is this for gaining a better understanding of how the virtues benefit their possessor? On the basis of what I have just said, instead of a moral interpretation, I argue for a metaphysical reading of the benefit-claim. First of all, we must take seriously that the benefit-claim indeed establishes a conceptual relation between virtue and eudaimonia. A person’s suitability as a human being expresses itself in the virtues. The virtuous (suitable) person lives up to his human nature (his ἔργον). This is what qualifies him as a good human being, that is, a good specimen of the human species. Moreover, we stated that for man, eudaimonia consists in the fulfillment of his function, namely, in the rational (virtuous) execution of the activities of human life. If we take seriously the conceptual relation between virtue and eudaimonia, then the virtuous person is ipso facto one who leads a good life. This, however, does not mean that he feels happy. That would be a modern (subjective) conception of happiness.

What about the wicked, then? Because we distinguish a modern conception of happiness from the ancient—metaphysical—conception of eudaimonia and adopt a conceptual rather than a causal relation between virtue and eudaimonia, the wicked are unable to live a ‘eudaimon’ life. Because eudaimonia refers to life according to the human function, and only those who possesses the virtues fulfill this function (because rationality, which is man’s function, manifests itself in the virtues), the wicked ex hypothesi can’t live a good life. They can feel happy but their lives can’t be ‘eudaimon’.

Yet virtue sometimes requires us to sacrifice our happiness. Then how can it be beneficial to its possessor? This is where we also must keep in mind the life-ethical interpretation that I have proposed. First, remember the distinction between a person’s subjective happiness and her objectively good life (eudaimonia). Then, indeed, virtue may sometimes require us to sacrifice our subjective happiness for the sake of the good (suitable) human life. (Think, by analogy, of the soldier ant, which may have to sacrifice its life to fulfill its function). This demand of virtue, however, is not a moral imperative—as such, it would be absurd. Rather, it is a pragmatic norm: if you want to live a good human life, you may need to sacrifice your life in a given situation. Read: Given this situation, no genuinely human life is possible afterwards. One can of course refuse to comply with this demand by prioritizing one’s will for subjective happiness over the will to lead a flourishing human life. (If the ant were rational, it might decide not to sacrifice itself. But then its life would no longer be that of a good—suitable—ant. Likewise, mine would no longer be that of a good—suitable—human being). Those who choose to withdraw from the requirement of virtue may not be blamed morally. Giving one’s life is a supererogatory act. But they opted for a life that a (metaphysically, not morally speaking) good specimen of the human species would not have chosen. If they had wanted to live a ‘eudaimon’ life, they would have had to choose to forego their subjective happiness. Note: if they had wanted to live such a life. Such is the crux of pragmatic imperatives.