BMCR 2007.03.02

The Stoics on Determinism and Compatibilism

, The Stoics on determinism and compatibilism. Ashgate new critical thinking in philosophy. Burlington, VT: Ashgate Pub, 2005. xxii, 132 pages ; 24 cm.. ISBN 0754639762. $79.95.

Table of Contents

For the last seven years or so, Ricardo Salles has been publishing (in Spanish and English) a series of stimulating papers focused on distinct aspects of Stoicism. This well-argued and fine book contains his views on Stoic determinism and compatibilism. Salles examines the topic both in the Older Stoics (mainly the theses and arguments attributed to Chrysippus) and in Epictetus. Although for moments the discussion is highly technical, it seems to me that at least some sections of the book may work as reading material for undergraduates studying key ideas of the Stoics (the introduction is particularly helpful in this respect; the same thing can be said of the introductory remarks to each chapter as well as of abstracts supplied by Salles at the end of some chapters), though the book is obviously aimed for scholars interested in Stoicism and endowed with a reasonable knowledge of the technical issues under discussion. Although Greek and Latin original texts are abundantly quoted, an English translation is always provided. The work is divided into two parts: Part I (“Determinism”) includes two chapters, Part II (“Compatibilism”) consists of the other four chapters of the book, which also contains a “Select Bibliography”, an Index of Names, and an Index Locorum.

Although disagreement is the order of the day in this kind of reviews, I shall not exclusively concentrate on points of disagreement but on trying to offer the reader a clear picture of this densely argued book. However, sometimes I will present my doubts with regard to some particular approach or conclusion in Salles’ argument. The Introduction starts with the guiding question of the whole work: “Is responsibility compatible with determinism?” Before trying to answer it, Salles accounts for the terms “determinism” and “compatibilism” when applied to the Stoics. The Stoics were determinists insofar as they maintained that every state or event is necessitated by prior causes; but, at the same time, they were compatibilists since they were willing to defend the thesis that prior necessitation does not make impossible that we deserve praise or blame for the actions we perform. So, against the so-called “hard determinists”, the Stoics are intent on proving that, despite determinism, the humans are genuinely responsible for their actions.

Salles is particularly concerned with distinguishing three forms of determinism that are not Stoic in character, and which he labels (i) “general determinism”, (ii) “crude fatalism”, and (iii) “external determinism”. The Stoics did not endorse general determinism because, according to them and unlike general determinism, “most objects do admit different states and patterns of behavior at different times” (xv). On the other hand, “crude fatalism” shares with Stoic determinism, Salles argues, the fact that it admits that determinism is compatible with the possibility of change. Besides, crude fatalism maintains that the future is already fixed (this is well-illustrated by the so-called “idle argument”, discussed by Salles on 10-11; he argues that after Aristotle crude fatalism was wrongly associated with the Stoics; I will return to this point later). However, unlike crude fatalism, Stoic determinism contends that the future is determined by the present in the sense that everything in the future is connected to prior necessitating causes on which its occurrence is dependent. In connection with this, Salles favors (following R. Sharples) the thesis that Chrysippus maintained that what was fated to happen, even if unnecessary from one point of view, was necessary from another, and, following Bobzien, he proceeds to distinguish two senses of necessity (although Salles agrees with Bobzien —responsible for one of the most important recent interpretations of Stoic determinism— in a number of points, he also rejects some of her major views. While Bobzien maintains that Stoic determinism is not to be understood as stating that the same effect always follows from the same causes —see her Determinism and Freedom in Stoic Philosophy, Oxford 1998, 33, 43), Salles is worried about establishing that the thesis that the same effects regularly follow from the same causes belongs to Early Stoicism, even though that tenet, he argues, should be considered within the context of eternal recurrence; see chapter 2). Finally, “external determinism” shares with the Stoics the thesis that any event or state is contingent upon earlier events or states, but unlike Stoic determinism, external determinism contends that the prior causes of what we do and of what we are go back to factors that are external to us. In other words, everything is determined just by external causes. These distinctions are relevant due to the fact that the Stoics put forward two theses that apparently look contradictory: on the one hand, they maintained that the world was governed by providence and, accordingly, that all the events were fated to happen. On the other hand, they were willing to assert that individuals are responsible for their own actions. At first sight, these two theses appear to be incompatible, but the Stoics, within their compatibilism, also defend the view that one’s responsibility can be reasonably preserved within a deterministic framework.

In his attempt to show how the Stoics make compatible determinism with moral responsibility Salles starts by developing Stoic arguments in favor of determinism. When describing what he calls “crude fatalism” Salles recognizes that the Stoics were fatalists to some extent; in fact, the Stoics share with crude fatalism the view that the future is fixed and that events and states that will take place are already determined (this coincides with the Stoic characterization of “heimarmene” as an invincible and unstoppable sequence of causes) or, to put it differently: according to the fatalistic ingredient that the Stoics take for granted in their determinism, if a future event Y is fated to occur, even in the case of a preceding event X which necessitates Y were not to happen, Y will occur. Reactions to this argument are well depicted in the so-called “idle argument” (discussed at length by Salles, on 10-12), which can be quickly paraphrased this way: “no matter what you do or not, if it is fated for you to recover from your disease, then you will recover. Conversely, if it is fated for you not to recover from your disease, then you will not recover, whether you take your medicine or not”.

Now Salles maintains that crude fatalism (an example of which is exemplified in the idle argument) was wrongly associated with the Stoics, but at the same time he recognizes that Chrysippus replies to this objection (which he labels as an “Aristotelian” objection, having in mind Aristotle’s, De interpretatione, 18b26-33; cf. 9-11). What’s more, he also implies that the Chrysippean theory of the confatalia was proposed to respond the inconvenient consequences of the argos logos. This being so, it seems pretty obvious to me that Chrysippus himself felt reached by the idle argument objection. Salles surely can reply that this is valid just for the fatalistic ingredient of Stoic determinism, which should not be confused with “crude fatalism”; but what is important at this point is that, as a matter of fact, Chrysippus felt compelled to answer to the argos logos objection and proceeded to do that through his theory of co-fated events (as remarked by Salles, my recovery from my disease is the kind of event that Chrysippus regarded as co-fated; 12). I agree with Salles that the kind of fatalism described in the idle argument was not probably endorsed by the Stoics, but we still need to explain why Chrysippus thought it necessary to rebut it if he thought that his own fatalism could not be identified with that fatalism. (By the way, it is worth noting that Salles’ treatment of co-fated events,12-15, is particularly clear. To my knowledge, he is the only scholar who has explained the boxer Hegesarchus’ case by resorting to the rules of Greek boxing, which clarifies the Stoic argument; see 14, n.28).

In Chapter 2 Salles turns to consider the thesis that causation is necessitating by analyzing the Stoic doctrine of everlasting recurrence, which, in his view, illustrates what he calls “the regularity-based” account (19). For the sake of brevity, I would just say that the discussion on the difficult issue of everlasting recurrence and conflagration is quite clear and that Salles’ reconstruction of this matter is generally quite persuasive (see, for instance, his discussion on the thesis of the reconstitution of the world on p. 20). Salles briefly speculates about the difficult issue of what god, the active principle in accordance with the Stoics, is doing during the conflagration, and suggests that, due to the fact that god is identified with active principle, there must be no time where god does nothing. But once this is granted, one wonders why Salles immediately asserts that god “has to be active once the conflagration subsides” (20-21). This assertion, to some extent, contradicts what has been said earlier, because if god has to be active once the conflagration subsides, one should wonder what god was doing before the conflagration subsides. The solution to this matter, it seems to me, can be found in a passage of Plutarch (not quoted by Salles), where Chrysippus is said to have argued that the cosmic fire is not necessarily different from god, but rather a phase in god’s life. The cosmic order is not everlasting but recurrently passes away; this, though, is not destruction in the strict sense: death is thought to be the separation of the soul from body. However, the world soul is not separated from the world body, but it goes on growing until it becomes co-extensive with the whole matter. The world, therefore, neither dies nor is destroyed in a strict sense. As indicated in the text, what is co-extensive with matter is Zeus himself. So god is operating all the time, and conflagration may be understood as god himself at its fiery stage (Plutarch, Stoic. repug., 1052c, quoting Chrysippus).

The discussion dealt with in chapters 1-2 justifies the Stoic determinism, but the Stoics also defended responsibility within their deterministic framework. According to Salles’ account, they did so first by making a distinction between external and internal causes. Responsibility turns out to be explained as a sort of internal cause. Indeed there is a quite obvious objection to this Stoic maneuver, i.e. that our own choices are still determined, no matter how internal they are. Nevertheless, the Stoics typically claimed that humans are responsible for those choices that, even being internally necessitated, are the result of a previous rational analysis. For some action F to take place, the agent should assent to the motivating impression ( ὁρμητικὴ φαντασία), whose propositional content is of the form “I ought to F” (34-37). Now giving or withholding assent depends upon the agent; then, what really counts is that the choice to give or withhold assent to an impression is the result of a rational process of assessment of such impressions, such process depending on the agent himself. The issue this account raises is whether or not it is possible to talk about an internal cause as depending on oneself in a deterministic world. This is what Salles intends to do in Chapter 3 (especially 40-49). To do this he cites in full the famous passage by Cicero, On fate 43, where the distinction between perfect and proximate causes is done. Salles suggests that the argument of the cylinder (labeled by Salles T 1) is addressed against the externalist objection in order to demonstrate that, contrary to what such an objection claims, the thesis that everything occurs by fate is compatible with the thesis that the acts of assent presuppose external factors but they are not fully determined by them alone. This helps prove that the Stoic psychology of action admits the “internality requirement”. Now the anti-determinist objector might argue that it is hard to see the way the Stoics took the cylinder argument to preserve both the determinism and moral responsibility, since, if the Stoics are fatalists (and thereby they assume that the future is previously determined), no matter that the external factors do not cause the agent to decide this or that manner (on account of the fact that assent depends on oneself), both to give and withhold an assent would have been already prefixed. However, if Salles is right in his suggestion that the Stoic determinism cannot be confused with crude fatalism, we should admit that the Stoics did not accept that the future is prefixed in the same sense crude fatalism thinks it is and, therefore, the anti-determinist objection cannot work (as Salles contends, for the Stoics the future is fixed by the present, just as the present is determined by the past; xv-xvi; 11-12). What the cylinder argument (T 1) actually shows is that what determines the event is the combination of the external factor and the internal assent. This is indeed what the confatalia —the Stoic maneuver to escape crude fatalism— indicate, too. Salles favors Huby’s thesis that the one who presents the externalist objection is Epicurus (43). But maybe there are reasons to believe that Epicurus did not regard the swerve as being the crucial point of human freedom: if we can rely on the reconstruction of Epicurus’ On nature 34, 21-22 (cf. LS 20β we have some reasons (with Sedley) to believe that the swerve allows for acts of volition, and that such acts are not to be taken as uncaused.

Now allow me to briefly focus on some points of Salles’ translation of Cicero, On fate 43: at the beginning of the passage he does not respect the verbal time when translating protrusit and dedit, which could be considered as a minor point that does not affect the argument. On the other hand, it is worth mentioning his interesting interpretation of the Latin visum obiectum, which he translates with “an impressor”. He justifies his interpretation (44, n.35) by arguing that (i) if visum were the impression, Cicero’s distinction between visum and species would become obscure, and (ii) that impressions themselves do not impress the soul’s pneuma. To be sure, Salles recognizes that at Academica 1.40 Cicero himself says that he is translating the Greek φαντασία as visum. Maybe a more accurate translation of visum obiectum would be “what is seen of the object” (as far as I can see, in Salles’ version obiectum is not translated; he modifies this on p. 44, where visum is paraphrased by saying “a thing seen”).

In Chapter 5 Salles tackles what he takes to be the three compatibilist theories of Chrysippus. T 1 (mainly based on On Fate, 43) is complemented, Salles contends, by T 3 (i.e. the thesis that the actions of living beings depend on them, as far as such actions involve impulse and judgment, no matter that impulse is determined. T 3 is reported by Nemesius and Alexander; 52-54; 74) as part of a single strategy against Aristotelian incompatibilism. Salles devotes an important part of the chapter to discuss the attribution of T 3. The question is whether T 3 is Chrysippean or late; Salles argues for the view that it is a theory designed by Chrysippus himself. The issue of attribution of T 3 to Chrysippus is interesting in itself, but I shall leave it aside. The interesting point that it is worth discussing is Salles’ suggestion that T 1, complemented by T 3, were addressed against Aristotelian incompatibilism (69). In accordance with him, T 1 and T 3 were addressed to respond different objections: while T 1 is designed to respond to the externalist objection, T 3 is set out to respond to the libertarian claim that responsibility requires the dual capacity to do otherwise at the time of action (76-77). I am convinced that Salles is right in maintaining that for Aristotle responsibility does require the specific capacity to act otherwise, and that this turns him into the candidate for being the target of T 3 (80-81). Nevertheless, I fail to see in what sense T 1 can be understood as addressed against Aristotelian incompatibilism: if T 1 is set out to face the externalist objection (this objection being the thesis that actions and events are determined by external factors alone), and if it is granted that the Aristotelian view presupposes that a real human action (i.e. a voluntary or responsible action) is that which depends on the agent to some degree, I do not see why one should believe that T 1 is addressed against Aristotelian incompatibilism. For all what T 1 intends to show is that it is untrue that actions are solely determined by external factors; but an Aristotelian view shares with T 1 that it is untrue that actions are determined by external factors alone. Maybe I am emphasizing to much a point that should not be emphasized, since the novelty of Salles’ argument here is focused on T 3, and consists of showing that responsibility does not need to assume the premise that the agent has the capacity to do otherwise. But the way in which the problem is presented in this chapter seems to suggest that T 1, along with T 3, is designed against Aristotelian incompatibilism.1

Finally, the bibliography cited and discussed is really impressive, especially if one takes into consideration that this is a small book. I am aware that a bibliography never can be exhaustive; however, I feel the absence of the excellent commented translation of Alexander of Aphrodisias, On Fate by C. Natali.2 Although this volume is in general impeccably produced, I have found a couple of misprints (on 71, n.7 it should say Griechen; on 86, it should say Corinth) and some mistakes in the Greek (I noted over 15), which will surely be corrected in a subsequent edition.

This is a lucid book; the views defended in it are well-supported by accurate readings of the texts as well as by sound arguments. Both professional scholars and advanced students of ancient philosophy will enjoy it and find much food for thought and discussion.


1. To give or withhold assent to a presentation ( φαντασία) depends, to some extent, on one’s character. One problem that arises in connection with this is whether an agent can be considered responsible for choices that are the result of the mistaken analysis of a presentation before assenting to it. To clarify this point Salles turns to Epictetus (in chapter 6), and argues that “acting from a fully rational impulse is just a sufficient condition for moral responsibility” (91). In other words, it does not matter whether or not one has acted from a fully rational impulse; anyway, one is responsible for his action. Epictetus explains the case by arguing that the agent is responsible for his own “precipitancy” or lack of reflection.

2. C. Natali, Alessandro di Afrodisia. Il destino (Prefazione, introduzione, commento. Bibliografia e indici di Carlo Natali. Traduzione di Carlo Natali ed Elisa Tetamo), Milano, Rusconi, 1996. Natali’s commentary is very valuable in many respects.