Free Will in Aristotle?

A recent large book by Olivier Boulnois, Généalogie de la liberté [Genealogy of Freedom] (2021), provides great detail on the emergence of the notion of free will (liberum arbitrium in Latin). I have previously mentioned his Être et représentation (1999), which gives fascinating documentation of the role of Duns Scotus in the formation of modern notions of representation, and metaphysics as ontology. Boulnois has written extensively on medieval philosophy. Like Alain de Libera, he turns a broadly Foucaultian “archaeological” method in the direction of showing the large and largely unknown role of medieval philosophy in the development of common Western philosophical concepts. This post will focus on a part of his discussion of how Aristotle stands largely apart from later views.

Early on in the present work, Boulnois cites the first sentence of G. E. Moore’s Principia Ethica:

“It appears to me that in Ethics, as in all other philosophical studies, the difficulties and disagreements, of which its history is full, are mainly due to a very simple cause: namely to the attempt to answer questions, without first discovering precisely what question it is which you desire to answer” (quoted in Boulnois, p. 34).

Boulnois notes how Latin translations extensively transformed the meanings of Aristotelian concepts.

“Is free will proper to the human? All of medieval and modern reflection on action rests on this thesis…. The concept of proairesis is analyzed for the first time in book III of the Nicomachean Ethics” (Boulnois, p. 133, my translation throughout). “To begin with, must we follow the [12th century Latin] translation of Burgundio? Does proairesis really mean free will?” (p. 134).

“With the translation of proairesis by ‘choice’, the key concept of Aristotelian ethics is integrated into the semantic field of will. The Burgundian point of intersection, which connects Aristotle and [the Church Fathers] Nemesius and Damascene in the same bundle of translations, constitutes a second origin of proairesis as free will in the 12th century” (p. 137). “Aristotle defines proairesis as a kind of desire joined with deliberation (with discourse)” (p. 139). “With Aristotle, we are indeed far from the medieval and modern interpretation, which speaks of choice (electio)” (p. 141).

Until now I have followed Joe Sachs’s translation in using “choice” myself. Sachs uses the English word with the same root as the French that Boulnois is questioning, but he explains it in terms of Aristotle’s definition just mentioned. But as we will see in an upcoming post, the Latin tradition gave it a voluntaristic coloring that is foreign to Aristotle.

“The most rigorous translation seems to be ‘resolution’…. Resolution introduces logos and time” (p. 140). “The object of our resolution is first of all the object of our desire, insofar as this results from a deliberation” (pp. 140-141). “Our moral character depends not on our theoretical attitude but on the ensemble of our resolutions, the repetition of which constitutes our disposition to act, indeed our aptitude for beautiful actions” (p. 142).

This ensemble of our resolutions sounds like the commitments of which Brandom and other contemporary writers speak.

“Since Cicero, the boulesis of the Stoics has been translated to Latin by voluntas [will]. But the Stoics affirm that such a mastery of impulse is found only in the sage…. It is only later, with Augustine, that this prerogative of the sage becomes a faculty accessible to all. Our concept of will, as capacity to consent or not to our representations, carries all this history. But in Aristotle, boulesis always designates the desire that pertains to that part of the soul that is capable of speech. It is a form of desire that is enunciable, intelligible…. Boulesis indeed is a wish: a wish does not require the existence of a will; like all desire, it is moved by its object” (p. 143).

“Among our desires, resolution results from a deliberation, which deploys itself in language…. The true principle of our resolution is of the order of language and of thought. Like all living beings, the human is necessarily moved by what appears to her as desirable; but in distinction from the other living beings, it is through speech that the desirable appears to her…. Resolution bears on action that can be accomplished by us…. There is only a place for deliberation when several lines of action are possible…. We deliberate on that which no one can do in our place” (p. 145, emphasis in original).

Here the resolution of deliberation is explicitly grounded in language and thought, and not in an anachronistic non-Aristotelian notion of a faculty of will.

“It is necessary to underline: Aristotle does not exactly say that the resolution depends on us. What depends on us is the action that is the object of that resolution…. Aristotle does propose a theory of decision, or of resolution, but not a theory of freedom of choice, or of free will” (pp. 145-146, emphasis in original).

Unlike Boulnois here, some 20th century writers have used “decision” in the sense that Boulnois gives to “choice”.

“Such actions are ontologically contingent: they could [also] not be. The metaphysical tradition deduced from this the existence of a power of choice, of a will or a free will” (pp. 146-147).

The important point here is that this “metaphysical tradition” does not include Aristotle.

“Nothing obliges us to understand contingency in an absolute manner, and as subject to our power. Aristotle refers to the ensemble of actions that depend on us (eph’ hemin) in general (in the sense of a collectivity: the humans, the agents, the citizens), but not to those that depend on me (here and now)…. But Aristotle never claims that an agent has, other things being equal, at a given instant, the capacity to do one thing or its contrary, to act and not to act. ‘What depends on us’ is a generic capacity, proper to humankind, and not to some particular individual, here and now. Reciprocally, in insisting on the idea that deliberation is the cause of our resolution, Aristotle holds that at the interior of this bivalence, we more often see what the logos presents as desirable” (p. 147, emphasis in original).

“In this context, what is the principle of action?” (p. 150). He quotes Aristotle, “and the principle of resolution is desire and reason…. What affirmation and negation are in thought, pursuit and flight are to desire…. For resolution to be good, it is necessary at the same time that the discourse be true and the desire just” (p. 150).

Boulnois summarizes, “Thus the properly human principle of action is resolution, that is to say the conjunction of desire and of logos in the pursuit of an end” (p. 151).

He turns to the background that Aristotle is responding to.

“The sophists affirm that every desire is good; Plato, that we can only truly will the true or absolute good” (ibid). “While the sophists identify freedom with the capacity to do what we want, and Plato with sole adherence to the Good, Aristotle proposes a third way: ethical freedom is the art of deciding well, of arriving at the resolution the circumstances require…. The resolution taken by the homme de bien [literally, “man of the good”] is indeed a point of coincidence between the apparent good (which all seek) and the true good (that she discerns). In leaving behind the confrontation between sophists and Platonists, Aristotle has recourse neither to a relativism of appearances nor to an objective norm; in the element of virtue (and of desire), he aims at a coincidence between the phenomenon of good and its truth” (p. 152).

This Aristotelian idea of a point of coincidence between appearance and truth was later taken up by Hegel.

“Is it necessary to say that resolution is free? Does it consist in a choice?… Aristotle never makes the concept of freedom intervene in relation to the question of choice, nor even that of willingness” (p. 153). “Nevertheless, Aristotle knows and uses a concept of freedom (eleutheria). But this pertains to politics and not to the theory of action” (ibid, emphasis in original).

“The Latin translation not withstanding, ‘hekousion’ (willing) does not mean ‘voluntary’ (voluntarium); it is rather the mode of our action, except in cases of constraint or ignorance. Proairesis does not mean ‘free will’ (liberum arbitrium), but decision, resolution… Freedom is the ethical horizon of our action, and not the metaphysical attribute of a will that Aristotle did not envisage” (p. 154).

“It is indeed possible to analyze human action without postulating in the agent a central instance of arbitration, and without inscribing her in a physical determinism…. Ethics is indeed thinkable without a theory of will” (p. 155).

As I have noted many times, Plato and Aristotle founded ethics without the later notion of a faculty of will.

“Ethical action does not necessarily require a freedom of choice. It depends on a resolution (proairesis), which mainly refers to a dimension of anticipation, and does not always imply a choice. Ethics is indeed thinkable without the doctrine of free will” (ibid).

Here again the emphasis is on something like commitment.

“Resolution presupposes a process of deliberation in the agent: she at least implicitly evaluates the reasons to do x rather than nothing; that which she does is not imposed on her from the outset. That is to say that for her, there are a multitude of options and reasons to act” (ibid, emphasis in original).

Deliberation is concerned with the goodness of reasons.

“A stranger to the metaphysical problem of free will, freedom is ethical in essence; it consists in the ability to act well; it is only acquired at the completion of an education in virtue” (ibid).

“All thought about action situates itself in the horizon of the good, and freedom is nothing other than liberation from the bad” (p. 156).

“It is ethics that founds freedom, and not freedom that founds ethics” (ibid, emphasis in original).

“On the plane of finite existence, no one knows if they have absolutely accomplished the best action. In the same way in their reflections on our moral lack of power, Aristotle and Ovid speak of that which is ‘better’ (comparative) and not of that which is ‘best’ (superlative)” (ibid).

Being and Representation

In L’Être et représentation (1999), Olivier Boulnois documents the emergence of “metaphysics” in its distinctively non-Aristotelian modern sense among various 13th century authors, including Roger Bacon, Henry of Ghent, and Siger of Brabant, leading to its decisive formulation by the Franciscan theologian John Duns Scotus in the 14th century. Avicenna had already claimed that metaphysics is about “being in general”, whereas Aristotle himself had emphasized that “being is said in many ways”, which implies that there is no “being in general”.

Boulnois suggests that the 13th century authors just mentioned paved the way for Scotus’ innovations by already treating being as a concept. We are so used to that, that it is hard for us to grasp what Aristotle means in suggesting that “being in general” is not a proper concept at all.

Scotus argued against Aristotle that there is a unifying, logically minimal criterion of being, and it is representability. To be representable is to be “not nothing”. Unicorns and other imaginary creatures are representable, whereas Aristotle would not have called them beings. Scotus’ concept of representation seems to be purely logical; to have a representation of something is not necessarily to have understanding of it. For Scotus, God and creatures are equally representable, even though creatures, as finite, can be properly understood by the human mind and God, as infinite, cannot. Whereas Aristotle never speaks of an infinite being — only of a perfect one — Scotus’ generic concept of being is very explicitly indifferent to distinctions between finite and infinite.

It is one thing to acknowledge representation as a logical concept among others, and quite another to give it the kind of special first place status that Scotus does in his ontology, and that Locke does in his epistemology.

Boulnois says it is with Scotus that metaphysics became linked to what Kant later called ontotheology. While separating metaphysics as the account of being from theology as the separate account of God, Scotus also made God indifferently one of the objects of metaphysics, along with all the other beings. The combination of these changes actually brought metaphysics closer to revealed theology, and helped it to be perceived as the safe handmaiden of the later Latin tradition, rather than as independent philosophical theology that some found threatening.

If one speaks of a subject of representation, it could be — in a sense of “subject” closer to that of Aristotle — that in which something stands for something, or it could be — in a modern sense — the one who represents. “In the context of representation, the soul is not the content of its thought, but rather has a representation, distinct from itself” (Boulnois, op. cit., p. 152, my translation).

It seems that for both Scotus and Locke, the mind has representations. The soul in Aristotle is thoughts and feelings and capabilities, not something standing behind them. (See also Repraesentatio; Ontology; Being, Existence; Being and Representation Revisited.)

Johnston’s Pippin

Adrian Johnston’s A New German Idealism just arrived, and I’m taking a quick look. It is mainly concerned with Slavoj Žižek’s work. But for now, I’m just concerned with chapter 2 — where Johnston launches a broadside against “deflationary” readings of Hegel, particularly the one he attributes to Robert Pippin — and the preface.

Johnston can be forgiven for not addressing Pippin’s 2018 work on the Logic, but I do not understand why he ignores my favorite book by Pippin, Hegel’s Practical Philosophy (2008).

There, Pippin dwells extensively on Hegel’s Aristotelian side. Much of interest could be said on what it means to be Aristotelian in a post-Kantian context. Many received views will be challenged by such an examination. (For a beginning, see Aristotle and Kant.) As I have said, I read Hegel as both Kantian and Aristotelian (as well as original).

In any case, Johnston seems to think Pippin in Hegel’s Idealism (1989) was intent on reducing Hegel to Kant. That book was indeed concerned to show a strong Kantian element in Hegel. But I did not think of it as reductive. If anything, I read Pippin’s book as a salutary response to those who want to reduce Hegel to a pre-Kantian, and to read Hegel as rolling back from Kant rather than moving forward from Kant. Because he assumes a bad old subjectivist reading of Kant, Johnston seems to think Pippin’s reading of Hegel necessarily rules out the possibility of seeing a realist side to Hegel.

The whole challenge of Hegel is to understand how it it is possible in his terms to be both Critical and realist, without engaging in logical nonsense. (But see Realism, Idealism.) This sort of thing typically requires significant semantic labor, but the achievement of such semantic elaboration is the whole point. Here I worry where Johnston intends to go with his defense of “undialectical” distinctions in the preface. It is one thing to recognize that Hegel does not intend to just do away with Understanding and its distinctions, and quite another to treat those distinctions as final. (See also Univocity.)

Johnston’s lengthy discussion of the positive value of Understanding in the preface does not address how it relates to dialectical transitions. He mainly wants to defend Žižek’s tactic of presenting forced binary choices at particular moments. In particular cases and circumstances this conceivably can be good pedagogy, but it is the details that matter, and Johnston offers no advice on how we are to distinguish a pedagogically good forced choice from a bad one.

(I suspect Žižek’s tactic may be related to his friend Badiou’s defense of the Maoist “One divides into Two” line, which always seemed like blustering nonsense to me. There have been some very rational strands within Marxism; I do not comprehend why someone as intelligent as Badiou would prefer to apologize for the coarsest and most anti-intellectual, but to a lesser extent Althusser did as well. See also Democracy and Social Justice.)

(Worlds away from this, Brandom has a wonderfully clear account of the nonfinality of Understanding’s particular conclusions, illustrated precisely by its very important positive role in the recognition and resolution of error, in which the operations of the Understanding on its own terms give rise to dialectical transitions at the level of Reason, understood in terms of the revision of commitments and possibly of concepts.)

Johnston also seems to assume there is something necessarily reductive about a non-ontological (or not primarily ontological) reading of Hegel. Again, I don’t see why.

I think Aristotle’s metaphysics was basically a semantic investigation, just like his physics. It is the historic forcing of this inquiry back from the wide universe of meaning onto narrow registers of being and existence that I see as reductive.

Based on the work of Olivier Boulnois on the role of the medieval theologian Duns Scotus in the reinterpretation of metaphysics as ontology, I have come to think that in general, modern emphasis on ontology tends to reflect what I take to be historically a medieval Scotist mystification of things Aristotle approached in clearer terms we should recognize today as mainly semantic. (For what it’s worth, the homonymous use of “ontology” in computer science is also mainly semantic.)

Metaphysics or “first philosophy” or “wisdom” was supposed to help us with higher-order understanding, not to be a place where strange existence claims are made.