La Nature et le Bien:
Une exploration du naturalisme éthique antique
dans De Finibus de Cicéron

Fecha de recepción: 2011 - 10 - 12
Fecha de aceptación: 2011 - 11 - 17

For nothing in life is more worth investigating than philosophy in general, and the question raised in this work in particular: what is the end, what is the ultimate and final goal, to which all our deliberations on living well and acting rightly should be directed? What does nature pursue as the highest good to be sought, what does she shun as the greatest evil? (Fin. I.11)¹

In this paper I want to examine the relation between nature and the good, and compare the overall attitudes to this issue in ancient and modern philosophy. This is no doubt a gigantic task that exceeds by far an essay's limits, so I will focus on De finibus, which results particularly relevant as a window into the big problem, since it contains detailed discussions of the major ancient philosophical schools (Epicureanism, Stoicism, the Platonic Academy and the Peripatetics), which seem to some extent to accurately reflect the main ethical views and debates of the ancient tradition up to Cicero's time.

After [1] an attempt to formulate the paper's key issue ('what is the relation between nature and the good?') in more precise terms, [2-3] I will try to specify the ancients' shared assumptions on the topic, and contrast them with what seems to be the general modern trend (very schematically described). This will hopefully allow us to [4] draw some preliminary conclusions and further questions.

The following is this paper's deep motivation: Perhaps the two major goals of ancient philosophy scholarship are, on the one hand, shedding new light on our understanding of the classical world (a mainly historical concern), and on the other, bringing ancient philosophical ideas into new currency by reintroducing them into current debates, thereby shedding on them the refreshing light of an ancient-based perspective (mainly a philosophical endeavour). In David Charles' terms (1992, ix-x), ancient philosophy scholarship can be traditional classical or philosophical in character. I write this paper as a (necessarily partial and superficial, but hopefully suggestive) case-study in the requirements and challenges in accomplishing the properly philosophical goal.

(A couple of methodological caveats: Since the paper's aim is so big, I will focus on trying to find regularities and general trends, which for reasons of space implies that I will have to leave many questions of detail at least partly unanswered. I therefore ask the reader both to try to suspend judgement on the issues of detail and to see whether the broad connections
I will try to make can actually hold in general. Also, as a general principle and an interpretive exercise, I will try to constrain this analysis as much as possible to the views expressed in De finibus, being fully aware that my conclusions apply to the ancient tradition as a whole only to the extent that this work reflects it properly. How much it does is a question that remains open to discussion.)

1. Clarifying the question and its terms

In this section I will attempt to translate our question about nature and the good to more precise terms. I begin with a very brief history of the problems related to ethical naturalism, which stemmed from a famous passage in Hume's Treatise on Human Nature which seems to be a wider statement of the problem found in his previously quoted passage:

In every system of morality, which I have hitherto met with, I have always remark'd, that the author proceeds for some time in the ordinary ways of reasoning, and establishes the being of a God, or makes observations concerning human affairs; when of a sudden I am surpriz'd to find, that instead of the usual copulations of propositions, is, and is not, I meet with no proposition that is not connected with an ought, or an ought not. This change is imperceptible; but is however, of the last consequence. For as this ought, or ought not, expresses some new relation or affirmation, 'tis necessary that it shou'd be observ'd and explain'd; and at the same time that a reason should be given; for what seems altogether inconceivable, how this new relation can be a deduction from others, which are entirely different from it (Hume 1975, III.1.1.469-470).

Scholars argue about what exactly Hume is attacking here.² But, however large the interpretative discordance, some things about the passage are nevertheless clear: (i) Hume considers the relation expressed by 'is' to be radically different from the relation expressed by 'ought.' (ii) This difference implies that a justification is required to derive an an 'ought' sentence from 'is' sentences. (iii) The claim that the 'ought' relation is different from the 'is' relation receives no justification here.

In the twentieth century G. E. Moore (1993, I.10-13) appropriated Hume's intuition that 'isought' derivations require a justification, andcriticised what he called 'naturalistic fallacies,' i.e. arguments that, given an object X, infer from premises of the kind 'X is N' a conclusion of the kind 'X is M,' where N is a natural predicate and M is a moral predicate. (He focused on arguments that concluded 'X is good,' but his argument can be extended to arguments that infer any moral property from natural properties.)

Many philosophers have criticized Moore's view that arguments of this kind constitute a fallacy, and now it is fairly agreed that they do not (see Frankena 1939 and Williams 1985, 121- 122). Nevertheless, the original intuition that seems to have unsettled Hume and Moore remains unanswered: Is it not required to justify a derivation of an 'X is M' conclusion from 'X is N' premises? And how could such derivation be properly justified? —That this issue remains relevant is visible through the fact that the philosophical landscape today is divided between ethical naturalists (those who contend that such derivation is possible and justifiable) and ethical non-naturalists (those who deny that it is).

Thus, the same intuition seems to underlie questions about the relations between nature and virtue; between is and ought; or between natural properties and moral properties. We can phrase this intuition in terms of the —perhaps clearer to contemporary thinkers— opposition between describing something and evaluating something. I will use the notions 'descriptive' and 'factual sentences' as synonyms, as well as the notions 'normative,' 'evaluative' and 'prescriptive sentences.'³

Now we can give a clearer translation of the original question ('What is the relation between nature and the good?') by replacing the terms 'nature' and 'the good' for the terms 'descriptive sentences' and 'normative sentences.' We can also specify the kind of relation that concerns us, namely logical derivation. Thus, a full translation of the main question would be: 'How can we justifiably derive an evaluative sentence from descriptive sentences?'

This translation replaces the obscure notions 'nature' and 'the good' for the intuitively clearer 'describing' and 'evaluating.' Describing means stating how a thing is (or seems to be), its factual properties; whereas in evaluating we state the thing's validity condition with respect to a given normative polarity (good or bad, right or wrong, virtuous or vicious…). We can also explain this distinction by claiming that describing something consists in predicating of it properties that are true or false about it, whereasevaluating something implies predicating of it properties that are right or wrong, good or bad, virtuous or vicious, about it.

2. Description-Evaluation Derivations in De finibus

In the passages quoted above, Hume states that fact-norm derivations should be justified, but does not make it clear whether he accepts some type of derivation; conversely in the Ciceronian passage above Piso certainly admits a derivation of normative claims about the human good from a description of human nature, but he says nothing about whether such derivation requires justification. Do ancient ethical theories generally perform such derivations? If so, do they offer some sort of justification? In part 2.1 I will look for evidence in De finibus as to how widespread this derivation was in ancient ethical theories. In part 2.2 I will explore whether, and if so how, the ancients justified such derivation.

2.1. Fact-Norm Derivations and Justifications in Ancient Ethics

Did all ancient ethical theories perform a derivation, or did only a few of them do it? Judging by De finibus, the fact-norm derivation was a common feature of all major ethical schools. This is suggested by (1) Carneades' division, and by (2) the presentations of each school's account.

2.1.1. Fact-norm derivations in Carneades' division

In his 'division,' the Academic skeptic Carneades classifies and compares the ethical views of his tradition. It is of interest here be cause the criteria chosen to classify all ethical theories might reveal some conceptual assumptions common to all of them.

It might be objected that Carneades' division is not a neutral classification, but rather a characterization developed with singular theoretical interests in mind and from a particular interpretive stance, which should thus not be taken as a pure reflection of ancient theories. I respond for now that my argument's success is certainly partially dependent on the division's
general exhaustiveness and adequacy, but that the division seems sensible enough to make this investigation at least interesting. (More on this below, in 2.1.3.)

Among the passages that mention Carneades' division, V.15-22 is particularly relevant, because there Piso mentions a part of its conceptual framework.⁴ He introduces practical reason (prudentia) as the art that deals with the good life (Fin. V.16), and then goes on:

It is almost universally agreed that what practical reason is concerned with and wants to attain must be something that is well suited and adapted to our nature, something that is attractive in itself and capable of arousing our desire (what the Greeks call horme). […] The origin of the whole dispute about the highest goods and evils, and the question of what among them is ultimate and final, is to be found by asking what the basic natural attachments [prima invitamenta naturae] are. Discover these, and you have the source from which the rest of the debate about the supreme good and evil can be traced (V.17).

So (i) ancient philosophers agree that the good life consists in things we desire or pursue naturally; and (ii) they disagree concerning what it is exactly that we naturally desire or pursue. Unsurprisingly, then, Carneades builds up his classification around the diverse conceptions of the primary natural attachments, since, as it seems, there is an agreement that they are the root of the ethical discourse, but disagreement about what things count as such. Accordingly, Carneades' division classifies theories in the first place according to whether they consider pleasure, the absence of pain or the primary natural goods (like health, intelligence, etc.) as the primordial objects of natural attachment. The second criterion, as we find out a bit further ahead in Piso's account, is whether they add morality, or virtue, to the chosen primary attachment or not. These are the division's main classificatory criteria.

What does this tell us about the presence of fact-norm derivations in ancient ethical theory? In short, it suggests that all major theories perform it. For if, according to Piso's statements, the accounts of the natural attachments are the "source" of the debate about the highest good, this is because it is generally admitted that the starting-point for a theory of the good is this description of natural attachments, from which what is good can be derived. Practical reason's task is, first, to locate our primary objects of of natural desire, and then to develop a full conception of the happy life in accordance with them.

But since Piso says "it is almost universally agreed," one wonders whether some important ethical views, that did not start from an account of nature, were left out by Carneades in his division. This would invalidate the generality of the trend I am describing. Piso confesses that "the theories of Pyrrho, Aristo and Erillus" were left out of the categorization because "they were never worthy of application" (V.23). This is a recurrent theme in De finibus: their theories are disregarded because they
give no account of what it is that moves us toward the good things. According to the text, Pyrrho and Aristo claimed that morality is the only good (IV.43), and Erillus held the same about knowledge (V.23). Cicero criticizes these views on the grounds that everything else is left undetermined: nothing else is treated as conducive to the good or natural or particularly desirable. According to Cicero (the character), these theories are impracticable because they "abandon nature" (IV.40-43) in the sense that they disconnect the good from the sources of natural motivation. According to Piso, Erillus "has removed the whole basis of rational planning and of the discovery of right action" (V.23) by providing no connection between the good and all the other things. For Cicero (who apparently reflects Carneades here), an ethical theory becomes impracticable, and therefore uninteresting, if it does not link the good with the motors of action. This is what the other theories do by presenting a natural object of attachment: they build a bridge between the things that attract us and moral correctness by calling some of the former 'natural.'

There did, then, seem to exist some ancient views that did not relate nature and the good, but it seems equally clear that they were a minority, and that the main debate took it for granted that normative claims should be grounded on a description of nature in order for it to have relevance in practice.

2.1.2. Fact-norm derivations in the accounts of ethical theories in De finibus

This is confirmed by the way each ethical theory is presented in De finibus: in all cases, the first argumentative move is to state human being's original natural tendencies, from which a certain conception of the good is then derived.

Torquatus' account (I.29-71) starts his defence of Epicureanism by claiming that "every animal as soon as it is born seeks pleasure and rejoices in it" in accordance with "uncorrupted nature." From the universal attractiveness of pleasure he concludes that is the highest good (I.29). At the end of his argument he claims that his account is "drawn from the fount of nature" (I.71).

Additionally Cato, representing Stoicism, states firstly that every animal's primary natural tendency is to seek its own preservation, and that humans later on discover that consistently following this original tendency leads to an orderly conduct of life, which then is called 'virvirtue' or 'morality,' and identified with the good. Cato thus concludes that the good consists in a life of performing actions which "are aimed at attaining the natural principles" (III.22), i.e. the things conducive to self-preservation. Stoics consider the orderly life, and not the primary attachments, to be the really valuable thing, but this orderly life is still composed of activities aimed at the primary things, which means their account of the good is built upon their interpretation of primary nature: the good cannot be conceived without the natural tendency toward self-preservation. Cicero (the character) accuses Stoicism of abandoning nature, just like Pyrrho and Aristo do (IV.43), but this is a different point: even if the Stoic account does consider some things natural and not good (which is Cicero's point), it is still true that their notion of the good depends on their notion of nature and their description of natural attachments (which is what I want to argue).

Finally, Piso's account of Antiochus' Academic Peripatetic theory starts, just like Cato's Stoic account, by establishing that animals are originally and naturally moved by self-love and towards self-preservation. His account narrates how each organism gradually realizes its particular natural features and capacities as it grows, and this leads it to direct its actions towards the activities that more precisely correspond to its natural properties with increasing accuracy (V.41-42). This results in an Aristotelian conclusion that each organism's nature is realized in its proper activity, and that "for human beings […] the ultimate good is […] to live in accordance with human nature as fully realized and needing nothing" (V.24-26). Thus he starts from a description of natural tendencies and builds from it an account of the good. He accepts that this is the strategy taken by his tradition by claiming that "our greatest authorities have been quite right to seek the foundation of the
supreme good in nature" (V.33).

Thus, according to the presentations in De finibus, all major ancient ethical theories (Epicurean, Stoic, and Peripatetic-Academic) derive their normative claims about the good from descriptions of the primary natural attachments, in harmony with Carneades' division.

It remains to investigate whether the ancients offer a justification for this derivation.

2.1.3. On the Accuracy of the Carneadean Division

Just before we do that, though, this seems a moment to ask: Are ancient theories properly represented by Carneades' division? Julia Annas (2001: xxv-xxvii) mentions three concerns: (i) Many positions in the classification lack actual proponents, or are held by dark characters of whom we know nearly nothing. (ii) By only accepting theories grounded on human nature, the classification rules out Platonic and Aristotelian 'becoming like God' theories. (iii) Epicurean and Stoic theories are presented in ways that make them seem odd: the former as a clumsy combination of two theories, the latter as considering the goal to be merely trying to get, instead of actually getting, natural things.

Perhaps Annas considers i to count against the accuracy of Carneades' division because a classification that leaves so many blank spots does not seem to be accurately capturing the main aspects of the classified things. However, the elements that the division seeks to classify are not only the existing theories, but rather all possible theories, either famously defended, not famously defended, or even those not defended at all before, but still possible (Fin. V.16). However, the division is built with the elements that the main theories present (pleasure, absence of pain or natural objects as candidates for the primary natural things; morality added or not), without introducing any foreign features. All the other spots (those defended by dark figures or not held by anyone) merely are possible combinations of these elements, hypothetical extensions of the main theories. In this sense, the classification's still focuses on the major theories, so the presence of the other views in it is no reason to be suspicious about its accuracy.

Now, Annas' second point assumes that 'becoming like God' implies leaving human nature behind. This is certainly true about human nature taken as a whole, but there must be some feature in us that allows us to grasp the divine in the first place, and this feature is a part (probably the best one) of human nature. This idea is present Aristotle's conception in NicomacheanEthics X, and Piso's account reflects this giving pride of place to the human capacity for contemplation (Fin. V.58). And in Plato's Theaetetus Socrates claims that a man should become not as just as God is, but rather "as just as it lies in human nature to be" (176d [tr. M. J. Levett & M. Burnyeat]), This is also compatible with Piso's Antiochean account, so I do not think Carneades' focus on nature excludes 'becoming like God' theories.

Finally, regarding Annas' last point, the division does present Epicurus weakly, since it is built from the assumption that pleasure and the absence of pain are different things. Although this is indeed a serious problem of interpretation, it does not alter Carneades' accuracy with respect to our concern, i.e. that Epicurus based his conception of the good on his conception of of the original natural tendencies, as both the division and Torquatus' account suggest that he did. Regarding the Stoics, it is true that in book V (20) they are depicted as claiming that the highest good is aiming at, not attaining, the natural things, while in book II (35) the Stoic goal is identified with morality alone. However, although this emphasis on merely aiming and not getting is indeed odd, V.20 also claims that aiming consistently at the primary natural things is identical to morality. So, even if they only try to get the natural things, in doing so they do get the true highest good (III.22-23).

To conclude, I do not think any of the concerns Annas raises seriously affects the adequacy of Carneades' division as a reflection of the conceptual commonalities of ancient ethical theories.

2.2. Justifying the Derivation

Since Hume's is-ought problem gained notoriety, many alternatives to ground the derivation have been put forward. I will mention here three that seem relevant, and then try to find evidence in De finibus as to whether any of these was the path taken by the ancients.

1. Disregard - A possible attitude towards the justification problem is simply ignoring it. This is what Hume says he finds in all the moral systems he knew. Perhaps Hume has some ancient theories in mind, so it would be worth examining whether they simply ignore the problem, i.e. provide no justification whatsoever.

2. Reduction - A way to bridge (or more precisely, destroy) the gap between the descriptive and the normative is claiming that they are not two separate realms, but that actually moral evaluations are a kind of factual statement. According to some interpreters (e.g. Scott-Taggart 1961), this is what Hume does in his own moral theory, by claiming that moral judgements are not a particular type of judgement, but rather a description of one's emotional reactions of approval or disapproval towards a particular action.

3. Bridging - In his influential article on Hume's is-ought passage, Alasdair MacIntyre (1959) suggests that the fact-norm derivation can be justified by employing "bridge notions" that (like desire, hope or intention) are able to link the realms of fact and norm. He claims that Aristotle did this, so, if he is right, perhaps other ancient philosophers did as well.

3. Bridging - In his influential article on Hume's is-ought passage, Alasdair MacIntyre (1959) suggests that the fact-norm derivation can be justified by employing "bridge notions" that (like desire, hope or intention) are able to link the realms of fact and norm. He claims that Aristotle did this, so, if he is right, perhaps other ancient philosophers did as well.

2.2.1. Disregard

It is perfectly possible that the ancients simply took the derivation for granted and, perhaps unawarely, dismissed the task of justification, as Hume suggests everyone had done before him. The best way to know whether they did this is to try to find evidence against it, and see if it shows a particular justificatory strategy in support of the derivation. If this kind of evidence cannot be found in De finibus, then we will have to accept that the ancients simply disregarded the issue. Let us then examine the other possibilities first.

2.2.2. Reduction

A part of Torquatus' exposition suggests that his argument might be a reduction of normative claims about the good to merely sensory information. In this passage, he claims that the nature of pleasure is directly perceivable, and needs no further argument to support it:

He [Epicurus] denies that there is any need for justification or debate as to why pleasure should be sought, and pain shunned. He thinks that this truth is perceived by the senses, as fire is perceived
to be hot, snow white, and honey sweet. In none of these examples is there any call for proof by sophisticated reasoning; it is enough simply to point them out (I.30).

Torquatus seems to be grounding the whole Epicurean conception of the good on allegedly perceptual knowledge, without any need for further argument (so that pleasure's normative character would be as perceivable as snow's whiteness). But there are two reasons to deny that this is a reduction of the normative to the perceptual. First, what Torquatus claims to be perceivable is not necessarily that pleasure should be sought (as Woolf translates), but perhaps that it is worthy of being sought, or simply that it is to be sought (expetenda), i.e. that animals have a tendency to seek it. According to a deflated translation, what is perceivable is pleasure's desirability, not its normative character (i.e. its being good).⁵ Secondly, the softer translation seems to be more consistent with further evidence, i.e. that the Epicurean ethical theory implies a distinction between two kinds of pleasure: static and kinetic (I.37-38), only the former of which is identified with the good. Torquatus does not hold that we directly perceive either this distinction or the identity between static pleasure and the highest good. Torquatus' perceptual argument could suggest only that pleasure (without qualification) is good; the more refined identification of the good withstatic pleasure requires further evidence.⁶ These elements of Epicurean theory, vital as they are, are not reducible to, or derivable from, perceptual information or merely descriptive features of human nature. This implies that the Epicurean account of the highest good is not reducible to descriptive claims, nor does it seem to have been intended to be so, despite Torquatus' emphasis on perceptual evidence.

2.2.3. Bridge notions

The possibility remains that the ancients connected nature and the good using bridge notions. McIntyre:

If anyone says that we cannot make valid inferences from an "is" to an "ought," I should be disposed to offer him the following counterexample: "If I stick a knife in Smith, they will send me to jail; but I do not want to go to jail; so I ought not to (had better not) stick a knife in him." […] The transition from "is" to "ought" is made in this inference by the notion of "wanting." And this is no accident. Aristotle's examples of practical syllogisms typically have a premise which includes some such terms as "suits" or "pleases." We could give a long list of the concepts which can form such bridge notions between "is" and "ought": wanting, needing, desiring, pleasure, happiness, health—and these are only a few. I think there is a strong case for saying that moral notions are unintelligible apart from concepts such as these (1959: 462-463).

In order to link descriptions and prescriptions, bridge notions must be "factual in nature but refer to aspects of human nature that are inextricably tied to our moral experience" (Arrington 1998: 258).⁷

I will argue that all the theories depicted in De finibus reach their conception of the highest good by means of bridge notions. I will present first a general argument and then one argument for each theory. The general argument is also based on Piso's description of Carneades' division (which I had quoted earlier):

It is almost universally agreed that what practical reason is concerned with and wants to attain must be something that is well suited and adapted to our nature, something that is attractive in itself and capable of arousing our desire (what the Greeks call horme) (V.17 again).

"What practical reason is concerned with" is another name for 'the ultimate end,' since that is practical reason's object of study. A condition for any candidate to the ultimate end, therefore, is that it must be "well adapted to our nature," i.e. it must be able to attract us, or in other words, to motivate us, to arouse our desire. This suggests that desire and attraction are used as bridge notions that link the primary objects of attachment (pleasure, absence of pain, natural objects) to the good: the primary objects of attachment are proper sources of the good because of our tendency to desire them. This is why practical reasoning —in Carneades' framework— starts from something that is naturally motivating: motivation itself seems to be a bridge notion, i.e. a notion that connects a merely factual feature of our nature with normative claims.

Now, do particular theories actually work like that? Do they use motivation as bridge notions? Torquatus's second argument to the effect that pleasure is the ultimate end seems to employ one such bridge notion:

The impulse to seek and and to avoid and to act in general derives either from pleasure or from pain. This being so, it is evident that a thing is rendered right and praiseworthy just to the extent that it is conductive to a life of pleasure. Now since the highest or greatest or ultimate good —what the Greeks call the telos— is that which is a means to no other end, but rather is itself the end of all other things, then it must be admitted that the highest good is to live pleasantly
(I.42).

How does this argument work? Pleasure and pain, the primary natural objects according to Epicurean theory, are the objects of all our impulses to act, both in pursuance and in avoidance. Since pleasure is the universal object of pursuance, and pain is the universal object of avoidance, Torquatus concludes that the rightness of an act depends on whether it contributes to a pleasurable life. Moreover, since pleasure is the universal object of attraction, Torquatus also concludes that the telos is a life of pleasure. This shows how attraction works as a bridge between pleasure and the happy life.

Epicureans, moreover, conceive of desire as a natural phenomenon studied by physics; this means that it is not a normative or ethical, but a natural concept —which, still, is linked to our ethical experience. This makes it a perfect candidate for a bridge notion. Indeed, natural science or physics "provides […] self-control, by explaining the nature and varieties of desire" (I.64). Physics determines which desires are natural, and which are not. Natural desires lead to static pleasures, and are easy to fulfil; but the others lead to 'empty pleasures,' have no limit, and thus condemn the person who follows them to dissatisfaction (I.45). It is physics, then, that shows which desires are in accordance with our nature (and can thus be connected to the happy life), and advises us to resist the other, unnatural desires.

The Stoic story about human nature is a more complex case, because it includes a transformation (Fin. III.21): We humans begin our lives just like any other animal, following the natural impulse towards self-preservation and the objects that promote it. As we learn to make the choices that effectively lead us to the attainment of these primary natural objects, our behaviour and decisions increasingly adopt an orderly pattern. The key event in our moral growth is the intellectual grasp of this pattern: when we become aware of the order produced by choosing correctly in a consistent manner (i.e. consistently aiming at the primary objects with accuracy), we realize that this order has such value that, in the face of it, the actual attainment of the primary objects becomes irrelevant, and the rational pattern of choices becomes the goal of our actions. This disposition is called 'virtue,' and a life of virtue is the Stoic good.

How do the Stoics justify the passage from the primary natural objects to virtue? Cicero, in his reply to Torquatus, claims that "for the Stoics […the highest good is] a life that understands what it is that happens by nature, and selects those things that are in accordance with nature, their contraries being rejected" (II.34). I want to emphasize that the Stoic good life is a life of understanding nature. The Stoic virtuous person rationally grasps nature's principles, i.e. the principles of a choice that correctly leads to the attainment of things in accordance with nature. This encourages me to think that in Stoicism, the rational capacity to perceive orderly patterns of choice works as a bridge notion. Cato actually recognizes this explicitly: "By the process of rational inference our mind ascends from those things which are in accordance with nature to a conception of the good" (III.33). Rationality works here as a bridge notion because it can surmount the factual particularity of each natural impulse to grasp the general orderly pattern of correct choices. Grasping this pattern generates an intellectual transformation in the agent's motivational structure, thus directing her actions to the rational order rather than to the particular instances of self-preserving things that motivated her before. In accordance with Carneades' framework, then, Stoic reason plays a motivational role: it the rationally informed will of the wise recognizes virtue as the best thing (actually the only good thing), and is thus motivated by reason to pursue it. Thus reason, or understanding, acts as a bridge notion in Stoicism.

Someone might contend that this view should no longer be called 'naturalistic,' since the Stoics leave the primary objects of attraction behind to focus on the (no longer natural) orderliness of choices. This, however, misses the point that the orderliness is identical to living in harmony with nature, and that this life is still based on choosing the primary natural things.
Stoic virtue does not abandon the pursuit of primary natural objects; it only systematizes it.

Now, in Piso's Academic-Peripatetic account there seems to be a bridge notion that combines both desire and reason. The original natural object is self-preservation, and nature leads us to it through desire:

Every living creature loves itself, and as soon as it is born strives to preserve itself. For the purpose of its life-long protection, nature bestows on it from the beginning a desire for selfpreservation and for maintaining itself in the best possible state according to nature (V.24).

But this desire is initially inaccurate, since the organism ignores its specific nature. Consequently,

our earliest desires have no other aim than to keep us safe and sound. Then, however, we begin to look around us and become aware of what we are and how we differ from the other animals. At this point we start to pursue the real objectives for which we were born (V.41).

Antiochus' theory thus seems to have a bridge notion as well, viz. a sort of adaptive desire. As each organism becomes aware of its own nature, its desires also become adapted to this nature and guide the organism to desire activities that cohere with its highest capacities. Adaptive desire guides the organism from general self-love to the particular activities that fulfil the organism's proper potentialities. Thus this desire acts as a bridge between the descriptively natural tendency to preserve oneself and the normative account of happiness.

In sum: De finibus (in Carneades' division and in the exposition of each theory) offers evidence that all major ancient ethical accounts are naturalistic, in the sense that they derive normative claims about the good from factual statements about human nature.⁸ De finibus also presents evidence for the claim that a justification for this derivation can be found in the use of concepts that play the role of "bridge notions" which tie the claims about the natural objects of attraction to evaluative notions of rightness and the good. The ancients, contrarily to what Hume suggested, seem to have developed a way to justify their is-ought derivations.

3. Guiding Nature, Mute Nature and the Disenchantment of the World

In this part I will argue that everything I have said so far is deeply mistaken. The argument up to this point fails to capture a profound difference between modern and ancient worldviews, which might vitiate it entirely. First [3.1] I will endeavour to specify a difference between ancient and modern notions of nature, and its relevance to our argument. Then [3.2] I will further argue that this difference is a consequence of the great process that Max Weber calls the rationalization of worldviews, which sharply separates the Classical conceptions of nature from our own.

3.1. Guiding Nature vs. Mute Nature

For in all natural things there is something wonderful.
Aristotle: De partibus animalium
(I.5.645a17-18)

At first sight, the upshot of Hume's passage quoted at the outset of this paper seems to be that there is no relation between nature and virtue. This is not quite precise: the conclusion is, rather, that nature is not linked with virtue any more than it is linked with vice. Why would nature be more inclined towards the excellent character than towards the wicked one? Why would it be closer to good than to evil? —This is Hume's question, and if we lack a proper answer, as he does, we must admit that nature is normatively mute, and that, as he says, it would be unphilosophical to assert that the natural marks the
boundaries of virtue and vice.

This point has a deep philosophical significance, and, as far as I can see, was altogether absent from ancient philosophical reflection.⁹ In the writings of the ancients there is a kind of intuitive, immediate, perhaps even unconscious tendency to place nature closer to the good than to its opposite. This is visible in Cicero's double statement of De finibus' main question:

What is the end, what is the ultimate and final goal, to which all our deliberations on living well and acting rightly should be directed? What does nature pursue as the highest good to be sought, what does she shun as the greatest evil? (Fin. I.11)

Cicero repeatedly personifies nature attributing preferences and desires to it,¹⁰ but this seems to be just a literary resource. Nevertheless, a key philosophical point of this passage is that Cicero freely rephrases a question about the good life and the rightness of actions in terms of natural tendencies: the evident necessary presupposition is that nature tends towards the good, and shuns from evil. Here Hume would ask: Why is nature closer to goodness rather than to evil? —Cicero offers no answer. Nor does any other ancient philosopher, as far as I know.¹¹ Yet they do not perceive nature as a mute entity: nature (i.e. the regular tendencies of organisms) is to them a guide with respect to normative issues.

Ancient derivations from natural properties to normative conceptions of the good should not be interpreted as description-evaluation derivations, because ancient natural properties interweave descriptive elements with a normative intuition: the natural tendency toward the good.

The upshot is not that ancient philosophy accepts nature's proximity to the good while modern philosophy does not. The point I want to make here is that it is typically ancient to take this proximity as given, while it is typically modern to ask for a justification to either its proximity or its neutrality. This reflects the opposition between a modern mute nature that has to be forced (or at least convinced) to speak prescriptively, and an ancient guiding nature that is always pointing in the right direction.

This difference between guiding and mute nature modifies the validity of my previous arguments, because it breaks an important analogy. I drew a parallel between Cicero's notions of nature and the good, Hume's remarks about 'is' and 'ought,' and Moore's concerns about natural predicates entailing moral predicates. I summarized all of these concerns in the question: 'How can we justifiably derive evaluative claims from prescriptive claims?,' suggesting that a strict analogy holds between the dualities naturegood, is-ought, natural predicate-moral predicate and description-evaluation. However, ancient nature's
role as a normative guide shows that this analogy does not hold, because for the ancients there is really no radical distinction between description and evaluation, between 'is' and 'ought.' Natural traits are not merely descriptive, since they already contain the evaluative seeds from which a normative account is ready to stem.

It is absolutely clear to Hume that the relation between subject and predicate expressed by 'ought' is different from the one expressed by 'is' (see above, 1.1); this clarity seems to be absent from ancient thought. We might see this like a failure of ancient systems of thought, since we— moderns after all—tend to share Hume's intuition. But it should be noted that, at least in the examined passages, Hume offers the same justification for the is-ought distinction as the ancients offer for their continuity: none. Both conceptions seem to rely equally on pre-theoretical intuitions, and to provide us with no definitive reason to prefer one over the other.

Thus the problem in our question's phrasing is that it erroneously presupposed that the ancients accepted the description/evaluation distinction, or a similar one. This is why what I argued in parts 1 and 2 is fundamentally wrong. It still seems true that (according to De finibus at least) the ancients derive normative claims from 'natural' properties, and that they do so by relying on bridge notions related to motivation. But these 'natural' properties are no longer natural to modern eyes. A rather different question jumps now to the spotlight: What accounts for the shift between the ancient guiding nature and the modern mute nature? Or: Why does the description/evaluation distinction seem intuitive and clear in the modern worldview, but is absent from the ancient Classical sworld?

3.2. Nature and the Rationalization of Worldviews

3.2.1. Guiding nature and natural teleology

Since many ancient philosophers explained natural phenomena in terms of means, ends and perfection, they allowed a guiding element to enter their conception of nature. Modernity destroyed this view of nature when it rejected natural-teleological explanations. —This seems like a convincing account of the difference between ancient and modern nature, but it is not adequate for two reasons. In the first place, Epicureans rejected teleological explanations of nature (they appealed instead to the rather mechanical notions of chance and atomic swerve (Fin. I.17-19)), but their nature was still a normative guide (the fact that all animals naturally pursue pleasure provides the foundations for their theory of the good). Secondly, Darwinian evolutionism is not anti-teleological; it rather has a peculiar kind of teleological explanation which, though certainly avoiding notions like intention, design or perfection, does resort to random variation and other teleological concepts like selection, fitness, and adaptation to explain natural phenomena. So Darwinism is teleological, but nevertheless remains mute regarding normative questions.¹²

Thus the presence or absence of natural-teleological explanations does not account for the difference between guiding nature and mute nature. The right account for this transformation in the conception of nature must lie somewhere else.

3.2.2. Guiding nature and rationalization

I will now introduce Weber's account of the disenchantment of worldviews (and its reception by Habermas (1981, 143-271) because this account, which describes the development of modern societies, narrates the process of separation of multiple spheres of value, like truth and rightness. Since the description/evaluation distinction is related to the truth/rightness distinction (see above, section 1), the theory of rationalization might illuminate the origin of the former distinction, and why it was absent in the ancient world.

Weber noticed that many aspects of modern Western society had originated solely in the West. Some of these aspects are, in the social context, the independent development of public administrative institutions, formal law, and the capitalistic market, each one according to its own autonomous principles; and in the cultural context, the independent development of modern science, normative ethics, and autonomous art, each one likewise according to its own independent principles of validity. These features add up to form a mode of life with particularly methodical traits. Weber seeks to explain how this mode of life came about, and why it did so exclusively in the West. He calls the process that led to this lifestyle and this kind of society rationalization, by which he means a process of increasing disenchantment (i.e. the gradual transformation of the world, from a locus of magical significance and mystical meaning that behaves in particular, non-universalizable ways, into a
set of quantifiable and systematically treatable entities usable for rational-strategic purposes) and systematization (i.e. the organization of worldviews according to universal principles and under the constraint of logical consistency). Weber's question, then, is: Why was the process of rationalization of worldviews and modes of life achieved to this extent only in the West?

Weber describes the process of rationalization through the development of world religions. ('Religion' here is understood as a
worldview that comprises a particular way to interact with the world and the consciousness structures that shape and reflect such interactions.) To him, the birth of world religions already implied a progress in rationalization, because they separate from traditional tribal worldviews by introducing values that break the limits of family and clan, and approach universal application and validity. Religions are classified according to two categories: (i) the representation of the divine (either as a transcendent and personal god, or as the immanent and impersonal cosmic order); and (ii) the orientation towards salvation (either an affirmation of the world as harmonious with the divine principle, or a rejection of the world as governed in ways different from the divine).

Some of the preponderant world religions are thus classified like this (Habermas 1981: 204):

Among world-rejecting religions, Weber distinguishes two attitudes towards the world: the believer can adopt an active life of ascetic mastery of the world (like in the Judaeo-Christian tradition) or a passive life of mystical flight from the world (as in the Indian traditions). He considers world-rejecting religions to have a greater rationalization potential than the (Chinese) world-affirming ones, because the lack of tension between the divine principle and the world-order leads to a practical adjustment to the events and their stochastic ordering, rather than to a tendency to make events fit into transcendent universal principles whose validity is independent from them (Weber 1964 229). Further, of the two world-rejecting attitudes, the ascetic life of mastery of the world happens to have the larger potential for rationalization, since here the devout perceives herself as a tool for God's action in the world, so that, when the world's order contradicts the divine principles, she has a motivation to transform it according to the divine law. Weber thinks that this attitude, together with the Protestant idea of predestination, gave rise to the methodical-ascetic, active order of life that characterizes modern Western societies.

Now, back to our question: What can this theory tell us about the description-evaluation distinction? Habermas seems to be right, at any rate, when he places the Greek philosophical tradition within the world-affirming orientation (1981: 203-204); this is precisely what a guiding notion of nature suggests: most ancient thinkers conceived the cosmos to be ethically meaningful, closer to goodness than to evil; conceived the natural order to be continuous with the moral order. Does this mean that ancient philosophical thought retains an element of magical, still enchanted thought? In Weber's account it does, and this element is visible in the collision that takes place whenever a worldaffirming religious orientation enters in conflict
with science:

The tension between religion and intellectual knowledge definitely comes to the fore wherever rational, empirical knowledge has consistently worked through to the disenchantment of the world and its transformation into a causal mechanism. For then science encounters the claims of the ethical postulate that the world is a God-ordained, and hence somehow meaningfully and ethically oriented, cosmos. In principle, the empirical as well as the mathematically oriented view of the world develops refutations of every intellectual approach which in any way asks for a 'meaning' of inner-worldly occurrences. Every increase of rationalism in empirical science increasingly pushes religion from the rational into the irrational realm; but only today does religion become the irrational or anti-rational supra-human power (Weber 1958: 350-351).

At the theoretical level, the separation between science and normative ethics as separate realms of value —a product of the rationalization process— is equivalent to the separation between spheres of value: empirical truth becomes independent from moral rightness; each is explored and developed according to its own inner logic by autonomous disciplines. This separation was possible only in a world-rejecting tradition, because it located the divine-ethical principle totally outside of the natural world, thus making it possible to sever the value sphere related to nature and the value sphere related to goodness and right action. Since the Classical philosophers had a different orientation, they never totally severed the links between natural and practical philosophy.¹³

Thus, the notion of guiding nature, that allows the Greeks to make ethical sense of the natural world, is a magical, still enchanted element of their worldview that sets a limit on the process of rationalization that would allow a radical description-evaluation distinction to appear. It is an enchanted element because it stops classical culture from seeing nature as a quantifiable material devoid of intrinsic ethical significance.¹⁴

3.2.3. Enchantment and repetition

A sign that ancient nature is still enchanted (i.e. that it cannot be seen as completely quantifiable, mute material) is provided by the ancient theories that read natural, or cosmic, order as a product of the benevolence or rationality of a God or demiurge (see above, note 15), and even in an undesigned, uncreated cosmos like Aristotle's. This is manifest in his defence of natural teleology against mechanistic views:

All the things that are by nature occur either always or for the most part, but none of the things that occur by hance or spontaneously [occur always or for the most part]. […] Therefore, if things occur either by chance or for the sake of something, and it is impossible for these [regular] things to occur by chance or spontaneously, they will be for the sake of something. But all of these [regular] things are by nature […]. Thus the 'for the sake of something' is resent
among the things that come to be and are by nature (Physics II.8.198b36-199a8).

This argument hangs on (i) the dichotomy between chance and regularity, and (ii) the intuition that regular events have an end. From these two it follows that things either occur by chance (and thus rarely) or they occur teleologically. Aristotle did not seem to consider the (for us) obvious alternative: that phenomena can be regular and lack any connection whatsoever to a goal. This is still an enchanted argument to the extent that it relies on the intuition that the regularity of an event implies its goal-directedness: natural regularity marks the presence of normativity. The trait that makes this argument enchanted is that the merely descriptive claim that a phenomenon is regular directly means (not even 'implies,' because there seems to be no inference process here) that the phenomenon has the normative character of goal-directedness.¹⁵

An analogously enchanted argument can be found at the basis of all the ethical theories laid out in De finibus: the exponents of Epicureanism, Stoicism and the Academic-Peripatetic synthesis take what happens always or for the most part (i.e. the original natural tendencies toward pleasure, self-preservation or self-love) as meaning something in relation to the good life and the rightness of action. They find in these descriptive regularities normative seeds of goodness and rightness. The direction toward which organic nature frequently points is taken to be the path that leads to the good. Sheer repetition becomes the root of normative validity. Frede claims that "in Greek thought regularity of behaviour as a rule is associated with design by an intellect" (2011, 13). For the previous considerations, and in order to include Aristotelianism in the rule he signals, I would rather say that regularity is interpreted as the presence of normatively-oriented rules (whether coming from a divine intellect or from the cosmos' intrinsic order).¹⁶

Hume was skeptical about the idea that virtue or goodness is more usual (and in this sense more natural) than vice or evil. We, as moderns, have an intuitive tendency to separate the two things: regularity is one thing, and the evaluative validity of such regularity is another (at least as far as our conscious, rational awareness goes; long-standing habits tend to become norms even for us). The Greeks seem to have had the opposite intuitive tendency to perceive the two things as one.

To conclude: (3.2.1) The difference between ancient and modern worldviews, which explains why the description-evaluation distinction is absent in one case and evident in the other, cannot be identified with the presence or absence of teleological explanations.17(3.2.2) It should rather be sought in the process of rationalization and disenchantment of worldviews, one of whose consequences is the separation of the value spheres of empirical truth and normative rightness, of regularity and rule. The Greeks did not reach this level of rationalization because, by maintaining a world-affirming orientation, they held on to a notion of nature that allowed regularities to express rules. (3.2.3) This tendency to equate regularity and normativity is a trait of the paradigmatically ancient mix of descriptive and normative elements in the notion of nature.

4. Conclusions

In this paper I tried to cover at a great speed a huge span of materials and discussions: I used Cicero's De finibus to reflect on the widespread naturalism of the whole Greek-Roman ethical tradition, and a few passages from Hume's Treatise on Human Nature to represent the general modern concern about the justification of naturalism. The conclusions (whose validity depends on how well these two texts represent the main lines of their traditions, and on the accuracy of Weber's theory of rationalization) are the following:

I mentioned in the introduction that this research's main goal was to assess ancient scholarship's ability to revitalize ancient concepts or theories within contemporary philosophical discussions. The recent resurgence of virtue ethics is more than enough proof that this is possible. I want to point out, though, that there is a profound distance between ancient and modern worldviews, which complicates the task of reviving the currency of ancient philosophical theories. This distance is not the one between an ethical naturalist and a non-naturalist view, as we understand it. Rather, the distance lies in that the connection between regularity and rule was a structural element of the ancient pre-theoretical worldview —buried so deep in the structure that it was not even explicitly considered or questioned at all—, whereas for a modern consciousness it is a theoretical alternative that requires justification. Nature spoke to the ancients in a normative tongue; moderns have to force it to speak normatively by justifying the derivation between descriptive statements and normative claims. This entails that, if an ancient view is to re-emerge from its original context into ours, the modern philosopher must attempt to reenchant the world, i.e. re-ligate spheres of value that have been separated by the relentless processes of modern rationalization.

The ancient scholar's philosophical task is therefore double: If she seeks to bring an ancient conception back to philosophical currency, first she has to generate a translation of it into the terms of contemporary discussions, and try to open a conceptual space of relevance in them —a matter that already requires considerable theoretical prowess. And second, she has to prove the existence of a link between a natural regularity of human life and the normative rules by which human life should abide. In facing the second task, the philosophically-minded ancient scholar can expect no aid coming from her ancient sources; for, from their worldaffirming orientation, this was pre-reflectively given to them, and shaped the background before which philosophical reflection and inquiry took place.

_____________________________________________________________________________________________________

1 Unless otherwise noted, I employ Woolf's translation of De finibus (Annas 2001).

2 For a summary discussion see Schmidt (2004, 227-229). Tweyman (1994, 485-577) compiled a series of influential studies on the topic.

3 A brief note about vocabulary: It is not rare to find a distinction between deontic normativity (right-wrong) and evaluative normativity (good-bad). I will use the terms evaluation, prescription and normativity to simultaneously refer to both. Words like happiness and virtue also fall within this category, since they are related to the notion of the good: happiness is the good (or the best) human life, and virtue can generally be considered as the disposition of character that corresponds to the good life.

4 It also seems to be closest to Carneades' original (Algra 1997, 120-121). See also II.33-43, III.30-31, IV.49-50.

5 It must be admitted that there is an ambiguity in Latin terms like 'expetendum', analogous to that of Greek terms like 'haireton,' that makes these two words translatable both as 'that should be sought' and 'seekable'; and that this ambiguity allows us to interpret Epicurus' perceptual knowledge of pleasure as descriptive and not yet normative, so this doesn't provide definitive evidence for a reductive strategy, but need not be interpreted in this direction. Much remains to be said about these immensely interesting ambiguities, but suffice it to claim for now that it is a sign of the relation nature and the good had in the Greek and Latin worlds, a relation which I will try to portray in the third section.

6 Cicero's response to Torquatus tackles precisely this point: his argument does not show that a newborn animal seeks static pleasure, nor does he offer a separate argument for this claim (Fin. II.31-32).

7 Happiness should perhaps not be included in MacIntyre's list because, as mentioned earlier (note 3), it is a normative term, due to its connection with 'good.' Likewise, although MacIntyre's list of bridge notions mentions some of De finibus' primary natural attachments, we cannot assume that they work as bridge notions in the De finibus, because they are the natural, descriptive traits that need to be linked to the normative features. Thus, if the ancients used a bridge notion, it should be different from whatever they regard as the primary object of natural attachment. Regarding health, Irwin (2003: 355-356) employs it against Pufendorf's attack to ancient and Scholastic naturalism: health (a natural and descriptive, but also morally relevant concept) links the world of natural science to that of morality, so that moral properties need not be an imposition on nature if they are continuous with health.

8 To see how widespread this naturalistic tendency was, it might be useful to compare the examined theories to the seemingly diametrically opposed views of some sophists, because they are said to defend a distinction between nomos and phusis, thus breaking from the naturalistic tradition. That is just an appearance: these thinkers do not reject naturalism. On the contrary, a sophist like Antiphon (probably Plato's inspiration for building his immoralist characters Callicles and Thrasymachus (Barney 2004)) considers social norms to be antinatural— thus unjust—, and nature to have different, valid, norms of its own. This view stands against the idea that social norms are in harmony with nature's laws, not against the idea that normativity is based on nature. Along these lines, Callicles claims in the Gorgias that "trying to get a greater share than most is said to be unjust and shameful by law [...] but I believe that nature itself reveals that it's a just thing" (Gorgias 483c-d [tr. D. J. Zeyl]). This view is not an exception to ethical naturalism, but rather a corroboration of its ubiquity in the ancient world.

9 The closest ancient text would perhaps be Physics II. 8, where Aristotle argues against some of his predecessors that nature acts teleologically rather than by chance. (More on this below (3.2.2).)

10 Interestingly, Cicero (the character) personifies nature attributing it desire, language, and choice (See e.g. Fin. II.42, 45-46, 110; III.40; IV.16, 47, 55), and seems to do it more often than any other character in the dialogues.

11 It could be argued that Plato does answer this question in the Timaeus. However, his answer there (29a) would be that nature tends towards the good because the benevolent Demiurge designed it so. Something similar occurs in Stoic cosmology: we can say that nature tends towards the good because it is determined by God's rational plan. This implies that nature's tendency towards the good is due to a consciousness that purposefully directs it. The modern question about nature's direction is then not so much as answered, but rather moved to the field of theology—and multiplied: How do we know that there is a consciousness acting on (or in) nature, and how do we know that its desings are benevolent?

12 Lennox (1993) argues that Darwin defended a specific kind of natural teleology, which tended to be seen as anti-teleological because it did not match any previously existent kind of final account. There are many discussions about how proper natural-teleological explanations should be like, but it is clear that there is a teleological element at the core of adaptationist explanations, namely the idea that a given trait exists in an organism because it was selected due to the consequent increase in fitness that it brought about. (For further bibliography see Lennox 1993, 420n13.)

13 Contra Weber's view that world-affirming orientations have less rationalization potential, Habermas argues that if Weber had considered the Greek philosophical tradition he would have found in it a worldview where world-affirmation and "radical disenchantment" coexisted (1981, 203-204). Habermas, however, thinks that the Greek world was radically disenchanted because it reached a high level of theoretical systematization. But, first, this not immediately imply disenchantment (the development of social institutions that quantify and manipulate the world toward practical strategic purposes is also a key element); and, second, the evidence that the Greek notion of nature is normatively guiding speaks against Habermas' interpretation of the Greek worldview as radically disenchanted.

14 Weber's account has been read as a theory of social progress or evolution, as if modern society were more advanced in a normative sense, but this is not right: Weber was skeptical of 'modern progress,' and viewed rationalization as a process which brings forth many undesirable, negative consequences.

15 Aristotle's passage does not present a full case of natural guidance, because it does not link the teleological notions like 'end' with explicitly normative notions like 'good.' He does this, however, in many other passages. Stoic cosmology presents a similar connection between regularity and the presence of normativity (see an analysis in Striker 1991, 11-13). Epicurean atomism might be a controversial case, but I have argued (3.2.1) that it still posits a guiding nature because the regularity of animal pursuance of pleasure means its goodness. Many other ancient cosmologies should be examined to see whether, and to what extent, they are still enchanted, but I take it that this is the general trend.

16 The tendency to elevate repetition to a normative level contrasts radically with the tendency (also present in classical philosophy) to reject the habits of the many when it comes to human affairs. Cicero presents a solution to this tension in De finibus by focusing on cradle arguments, i.e. arguments based on the first impulses that organisms (or humans) display as they are born. These accounts can bypass the acquired habits of the masses while retaining the reference to regular phenomena.

17 Irwin (2003, 358) has claimed that the change from an ancient to a modern concept of nature does not undermine ancient moral naturalism, relying on Butler's claim that modern science does not underwrite teleological reasoning: Even in the modern scientific framework, in order to understand an organic system, we need to grasp its overall functions, its overall ends; this implies a sort of teleology. To link teleology to morality, Irwin relies on health as a bridge notion between natural science and morality. This is a sound argument, but it does not count as a defense of ancient naturalism, for, as we have seen, ancient naturalism implies an entwinement of nature and normativity, not a derivation of normative claims from a purely descriptive natural claims. In other words, Irwin's naturalism is inspired by ancient philosophy, but is typically modern.

_____________________________________________________________________________________________________

Bibliografía

Algra, Keimpe. 1997. "Chrysippus, Carneades, Cicero: The ethical divisiones in Cicero's Lucullus." In Assent & Argument: Studies in Cicero's Academic Books, edited by Brad Inwood and Jaap Mansfeld. Leiden; New York: Brill.

Arrington, Robert L. 1998. Western Ethics: a historical introduction. Malden, Massachusetts: Wiley-Blackwell.

Barney, Rachel. 2004. "Callicles and Thrasymachus." In Stanford Encyclopedia of Philosophy. Stanford: Stanford University Press. http:// plato.stanford.edu/entries/callicles-thrasymachus/

Cicero. 2001. On Moral Ends, edited by Julia Annas; translated by Raphael Woolf. Cambridge: Cambridge University Press.

Charles, David. 1992. Aristotle's Philosophy of Action. London: Duckworth & Co.

Frankena, William. 1939. "The Naturalistic Fallacy." Mind 48: 464-477.

Habermas, Jürgen. 1984. The Theory of Communicative Action, translated by Thomas McCarthy. Boston: Beacon Press.

Hume, David. 1994. Critical Assessments, vol. 4, edited by Stanley Tweyman. London; New York: Routledge.

Hume, David. 1975. A Treatise of Human Nature, edited by Lewis Amhers Selby-Bigge. Oxford: Oxford University Press.

Irwin, Terence. 2003. "Stoic Naturalism and its Critics." In Cambridge Companion to the Stoics, edited by Brad Inwood. Cambridge: Cambridge University Press.

Lennox, James G. 1993. "Darwin Was a Teleologist."Biology and Philosophy 8: 409-421.

Lévy, Carlos. 1992. Cicero Academicus: Recherches sur les Académiques et sur la philosophie cicéronienne. Rome: École Française de Rome.

MacIntyre, Alasdair. 1959. "Hume on 'Is' and 'Ought'." Philosophical Review 68, 4: 451-468.

Moore, George Edward. 1993. Principia Ethica. Cambridge: Cambridge University Press.

Scott-Taggart, M. J. 1961. "MacIntyre's Hume." Philosophical Review 70: 239-244.

Striker, Gisela. 1991. "Following Nature: A study in Stoic ethics." Oxford Studies in Ancient Philosophy 9: 1-73.

Weber, Max. 1958. "Religious Rejections of the World and their Directions." In From Max Weber: essays in sociology, edited by H. H. Gerth and C. Wright Mills. New York: Oxford University Press.

Weber, Max. 1964. The Religion of China: Confucianism and Taoism. New York: Free Press.

Williams, Bernard. 1985. Ethics and the Limits of Philosophy. Cambridge, Massachusetts: Harvard University Press.