Most of what Epicurus and Lucretius say about our freedom is compatible with causal determinism. This should not surprise us. Epicurus is not concerned with the freedom required for genuine moral responsibility, but with securing a sort of freedom of action — a rational self-rule that allows us to control our actions and shape our character, so that we can attain a tranquil life. This type of freedom is compatible with causal determinism. However, Epicurus was led, by a series of philosophical mistakes — understandable mistakes, but mistakes nonetheless — to posit an indeterministic atomic motion to help defuse a threat to our freedom. The swerve plays only the subsidiary role of defusing the fatalist implications of the Master Argument and similar arguments based on the universal applicability of the Principle of Bivalence. Despite the swerve's relative unimportance in Epicurus' own theory of freedom, it is this element of his theory that in the end has had the greatest philosophical impact!

If I am right that it would be anachronistic to saddle Epicurus with a libertarian response to the 'traditional' problem of the apparent incompatibility of determinism and the 'ability to do otherwise' necessary for genuine moral responsibility, this raises the following questions: how did the 'traditional' problem arise, who was the first person to put forward a libertarian response to this problem, and how was this transmitted into the western philosophical tradition? Also, what role did Epicurus' positing of the swerve play?

To answer these questions fully, and to justify those answers, would go well beyond the bounds of the present study. However, now that I have laid out how I think Epicurus' theory works, I would like to close this book by briefly telling the story, as I see it, of the way this theory helped lead to the creation by later philosophers of a libertarian solution to the problem of free will and determinism. I plan to fill out and defend in future work the claims sketched out here.

Epicurus helped shape the subsequent problem of free will and determinism through his influence on Carneades, and Carneades should be credited as the first person to come up with a libertarian position vis-a-vis the 'traditional' problem. This view requires arguing for three distinct theses: that the problem Carneades is confronting is fairly described as similar to the 'traditional' problem, that his solution to this problem is libertarian, and that Epicurus was the inspiration for Carneades' solution. It is often tricky to ascertain the exact influence of one philosopher on another, but this is not one of those cases.

When Cicero describes Carneades' position, he says quite explicitly that Carneades developed it as a modification of the Epicurean position: Carneades' doctrine is that the Epicureans could maintain that there is a 'voluntary movement of the mind,' against the Stoic Chrysippus, without positing the swerve or any other sort of motion without a cause. (De fato xi 23)

To spell out what sort of problem Carneades is confronting, it will be helpful to refer back to the schema of different 'free will and determinism' problems, as involving a combination of three factors, detailed earlier (section 1.1):

Type of determinism X threatens (at least potentially) Y, and Y is a necessary condition for Z

I have argued that Epicurus' primary concern is with fatalist arguments that hinge on logical determinism — the position that there is already a fact of the matter, and always has been, about what is going to occur in the future — but that he also opposed causal determinism because of the Interentailment Thesis [logical and causal determinisms entail each other]. As described above (sections 6.3 and 6.4), Carneades uncouples logical and causal determinism. He maintains that logical determinism need not imply causal determinism, and that logical determinism has no unwelcome implications for our freedom: if it is simply my freely deciding to raise my arm at 12:15 EST on July 28, 2004 (as I just have) that makes it to have always been the case that I raise my arm at that time, and thus for the statement "Tim O'Keefe will raise his arm at 12:15 EST on July z8, 2004" to have been true for an eternity before I raised my arm, the eternal truth of that statement does nothing to threaten my freedom.

However, Carneades does deny causal determinism — the thesis that the "laws of nature" along with past states of the universe determine exactly one unique future — because it is incompatible with there being a voluntary movement of the mind and with anything being in our power. Carneades does not put the thesis of causal determinism in these terms, but he is clearly targeting this sort of doctrine. The Stoics, against whom Carneades is arguing, conceive of fate as an everlasting ordering and interconnection of causes, flowing from all eternity like the unwinding of a rope. And so, nature has contained in it already (and always has) the causes which will bring about what is going to occur. When Carneades attempts to pry apart logical and causal determinism, he gives examples of free action (like his going down to the Academy) and says that the statements describing these actions are true from eternity, but not in virtue of "an eternal stream of natural and necessary causation," or "foreordained causes" (De fato ix 19), or "immutable causes, eternally existing" (De fato xii 28), or a "nexus of eternal causes" (De fato xvi 38).

Carneades' denial of causal determinism, however, is most clear in his discussion of epistemic determinism. Carneades thinks that epistemic determinism entails causal determinism, and since causal determinism is false, so too is epistemic determinism. Carneades' argument proceeds from the assumption that to know what is going to occur in the future, one must know what presently obtaining causes will bring about that future event. For instance, in order for me to know that a major earthquake will occur in California a year hence, I would need to have information about the present disposition of California's various faults, the pressure they are exerting on one another, etc., along with the facts about how faults, rocks and dirt of various kinds, etc., behave, which all together will bring about the future quake. Carneades says that not even Apollo, however, can foretell events like Oedipus killing his father (although it has always been true that he would do so). That is because such actions, before they occurred, had no pre-existing causes that would bring them about, that Apollo could inspect in order to tell that they are going to occur (De fato xiv 32-33).

The parallel with a Laplacean omniscient demon is striking. Laplace claims that such a demon, knowing the present disposition of all of the matter in the universe, along with the laws of nature, would on this basis be able to predict everything that was going to occur (as well as retrodict all that had occurred). Although Laplace affirms determinism and Carneades denies it, their epistemic examples serve similar functions: the demon's abilities illustrate Laplace's thesis of universal causal determinism, and the limitations facing even Apollo drive home what is involved in Carneades' denial of that thesis.

If causal determinism were true, according to Carneades, then nothing would be "in our power" (De fato xiv 31-32). What this amounts to can be seen more clearly by looking at how Carneades objects to the Stoic position. The Stoic Chrysippus replies to the 'Idle Argument' as follows: just because it is fated that you will recover from a snake-bite does not make your taking anti-venom in order to recover from the bite pointless. (See sections 6.3 and 6.4.) Chrysippus says that certain events are 'co-fated': for instance, it is fated (and causally determined) both that I will recover from the snake-bite and that I will take the anti-venom; it is through my fated action of taking the anti-venom that my fated recovery will occur (De fato xiii 3o). The Stoics say that motions like my taking the anti-venom, which are brought about by fate through me, are "in my power," and they deny that we are free to choose between opposite actions (Alexander On Fate 181,13-182,20 L&S 62G).

Carneades, however, rejects the whole class of 'co-fated' events. He restates the 'Idle Argument' in terms of causes, instead of truth; Cicero believes that this restated argument is as tight as can be. Carneades argues that if everything is fated, everything occurs because of antecedent causes. This would make everything take place in a closely knit web of natural interconnection, which would make all things occur of necessity, and nothing would be in our power (De fato xiv 31).

Given this context, and Carneades' repeated railings against causal determinism, for an action to be in our power most likely means that it is in our power either to perform it or not to perform it. And so, Carneades thinks that causal determinism threatens one's ability to do otherwise than one does, and for this reason he rejects causal determinism.

This may seem wrong, since Cicero derides Epicurus for positing 'motion without a cause,' whereas Carneades (via Cicero) affirms that every event, including human action, does have a cause. But even though every event has a cause, Carneades' position is still incompatibilist. We can best understand Carneades' position and the arguments he works through by comparing what he says to Roderick Chisholm's views on 'agent causation.' Like Carneades, Chisholm thinks that both causal determinism and causal indeterminism are incompatible with freedom.

Like Carneades, Chisholm thinks that merely uncaused events cannot be free, because freedom is not the same as randomness. So Chisholm proposes that every event has a cause, but not all events are caused by other events in accordance with exceptionless laws of nature. Instead, some events, such as voluntary human actions, are caused by the agent — by the person — yet how this 'agent causation' itself acts is not causally determined by previous events or states of affairs.

This is quite similar to Carneades' description of the 'voluntary motion of the mind' that is in our power. All events, including human actions, have causes. However, voluntary actions do not have antecedent causes stretching back eternally to past events and states of affairs. Instead, these actions are simply the result of a 'voluntary motion of the mind,' a motion which has an intrinsic nature of being in our power and of obeying us (De fato xi 24-25).

Having this sort of ability to do otherwise than one does is a necessary condition both for the rationality of deliberation and action, and for moral responsibility. As noted above, Carneades' rejection of Chrysippus' 'co-fated' events occurs in his restatement of the Idle Argument, which concerns the question of whether determinism renders us helpless and makes action pointless. Cicero, who acts as Carneades' spokesman through the dialogue, also thinks that causal determinism would fetter the human mind in the chains of a fated necessity (De fato ix 20).

The case for attributing to Carneades a libertarian position on the incompatibility of determinism and the sort of free will necessary for moral responsibility is less direct, but still quite plausible. The key text is De fato xvii—xix 39-45, where Cicero takes up the question of whether fate is consistent with justified praise, blame, and punishment, and where he describes Chrysippus' attempt to make them compatible.

Some people, according to Cicero, assert that not everything takes place by fate, because fate and freedom are incompatible. Their argument goes as follows: a necessary condition on justified praise, blame and punishment is that the assents, which produce our actions, be in our power. But if everything is fated, then everything takes place with an antecedent cause. Our assents are caused by our desires. But our desires would also then have antecedent causes, causes which are outside us and not in our power. (For instance, my vicious decision to eat a chicken patty made from factory farm-raised chickens is prompted by my hunger, along with a sense-impression of a chicken patty. But the cause of this sense-impression is an external object, and that I had this sense-impression was not something in my power.) But if the causes of our assents are not in our power, then our assents themselves are not in our power.

However, Chrysippus wants everything to be fated and have antecedent causes, and also assent to be in our power. He tries to maintain both by distinguishing types of causes. He agrees that assents must be prompted by sense-impressions, e.g., I would not reach for the chicken patty without the appropriate sense-impression. These sorts of auxiliary causes, however, which act as triggers to action, do not place assent out of our power. That is because they are not sufficient conditions on our actions, and what sort of action they prompt depends on the person, so that the cause of action is still internal to him, and the action in his power. The character of the person himself is the principal cause of his actions. So, for instance, the same sort of sense-impression which triggers my eating the chicken patty would not cause somebody to eat it who cares more about the suffering of de-beaked and overcrowded chickens than about his gustatory pleasure.

So far, so good. Cicero presents Chrysippus' distinctions approvingly, and says that once they are accepted, the differences between the disputing parties over fate might seem merely verbal (De fato xix 44-45). But then he adds,

This distinction [between antecedent and principal causes] is approved by both sides, but one of the two schools holds that although fate does govern those matters in which, when antecedent causes have occurred, it is not in our power to make the results turn out otherwise [non sit in nostra potestate ut aliter ilia eveniant], yet fate is not present in the case of matters which are in our power...

Unfortunately, the text breaks off at this point. But it is not hard to see that the same Carneadean point is being made here, about in what sense actions must be 'in our power' in order for praise, blame, and punishments to be justified, as was being made earlier about in what sense actions must be in our power in order to escape the Idle Argument. Once an antecedent cause has occurred as a trigger to action, do I then have the ability to make results turn out otherwise, or not? Once the chicken-patty impression has struck me, do I have the ability either to eat the patty or not to? If I have this ability to do otherwise than I do, then my action is not fated. But if how I respond is 'up to me' merely in the Chrysippean sense that it causally depends on my present character, which is itself 'co-fated' and the way in which fate works its will through me, then my action is both fated and not truly in my power, so that praise, blame, and punishment would not be justified.

Ironically, the first person to put forward a libertarian theory of the freedom of the will may not himself have been a libertarian. As head of the skeptical academy, Carneades was in the business of refuting the arguments of others in order to show that they did not have the knowledge they thought they did. He did not confine himself to criticizing others' arguments, however. He also retailed arguments of his own in support of positive theses — not in order to show that they are true, but because doing so would help counterbalance the arguments for contradictory theses and bring his audience to realize that none of the opposing arguments establish their conclusions.

Beyond Cicero, tracing the paths of historical influence becomes even murkier. But one line by which this 'traditional' problem of free will, along with the libertarian solution to it, is likely then transmitted into the Western philosophical tradition is via St. Augustine, particularly in the position he lays out in On free choice of the will (De libero arbitrio). In addition to his deep familiarity with Cicero's corpus overall, Augustine is in particular familiar with what Cicero has to say about the incompatibility of free choice and epistemic determinism (i.e., foreknowledge), as contained in works like De fato and De divinatione. Part of his purpose in writing On free choice of the will is to reconcile divine foreknowledge and freedom. And so, even though he does not explicitly credit Cicero, the many points of overlap between Augustine's conception of freedom and the libertarian position reported by Cicero are likely not just coincidental. However, Augustine's purposes are in many ways importantly different from Carneades'; let me note briefly the content of Augustine's doctrine and its goals.

Like Carneades, Augustine asserts that voluntas has a cause: it is a movement of the soul which is under our control (De lib. arb. 2.20). Also like Carneades (as reported in De fato xi 24-25), he makes this point by likening this movement of the soul to the movement of a body: just as it is the nature of a stone to move downwards, so to it is the nature of the soul to be able to move itself by the will, the difference being that how the soul exercises this power is up to it (De lib. arb. 3.1). He accepts the doctrine that sense-impressions function as triggering causes to action but are not sufficient to determine one's will. Augustine says that only something that is seen (in a sufficiently wide sense of 'seen') can incite the will to act. He adds that we control whether we accept or reject what we see, but we do not control what we see (De lib. arb. 3.25). This is supposed to help make all acts of will at least somewhat explicable: they have a motive, which is an object that the agent perceives as desirable.

Augustine emphasizes even more strongly than Cicero that will is a self-movement of the mind that has no cause beyond itself. He notes at many places that what we will is, in a very strong sense, within our power (e.g., De lib. arb. 3.3). And so, the will, as a higher part of soul, cannot strictly speaking be overcome by passion — instead, a mind that sins of its own volition turns away from virtue and chooses to be a slave of desire (De lib. arb. 1.10-11). But Augustine also notes, as Cicero does not, that if this is so, the will also can ignore the verdict of right reason. Augustine maintains that the will is the cause of sin, and that there is no cause for the will beyond itself, because if there were a cause of the will beyond itself, there would be no sin (De lib. arb. 3.17). And so, the will can simply turn from a greater perceived good to a lesser one — a movement which is sin — for no reason whatsoever beyond its own decision. As Mann notes, "[One] aspect of the power of the human will is to reject the verdict of reason." This autonomy of the will from both desire and from reason would not be welcomed by Epicurus; we will briefly explore below why Augustine insists on it.

In his discussion of the freedom of the will, Augustine does not bring up the 'Idle Argument' at all. Instead, just like the foes of fate that Cicero describes in De fato 39-45, Augustine insists that our actions being in our power, in the sense of our being able to choose one way or the other, is necessary for them to be justifiably subject to moral appraisal. When asked why God would allow people to have a free will, since free will is what allows evil into the world, one of Augustine's replies is that we cannot do right without having free will, and no action would be either a sin or a good deed if it were not performed by the will (De lib. arb. 2.1). Part of the nature of free will, however, is that it can be misused. Without this built-in liability, which allows us to sin, we would not be able to choose to live rightly. But Augustine, in particular, wants to vindicate the justification of punishment, especially God's eternal punishment of sinners: because sinners turn away from virtue of their own volition, and this turning away from higher goods to lower ones is caused by the sinner's own will and not by God, God is just when He punishes sinners for it (e.g., De lib. arb. 2.19).

Augustine's other reason for insisting on this libertarian free will is in order to defuse the problem of evil. On free choice of the will opens with Augustine's interlocutor Evodius asking whether God is the cause of evil, and the main task of the book is to show that He is not. Since God is wholly good, He cannot create anything that is evil. In sin, the will turns from eternal higher goods (such as justice, conceived of by Augustine as akin to Platonic ideas in the mind of God) to changeable, lower ones (such as bodily pleasure). The will itself is good, the higher goods are good, the lower goods are good, and all of these are from God. The turning from higher to lower goods is evil, but this movement is not caused by God, but simply by the will itself and nothing else (De lib. arb. 2.19; see also 1.15).

Augustine's concern to construct a theodicy also helps to explain why he insists on the autonomy of will from both desire and from reason. As MacDonald notes, if free choice is supposed to explain the origin of evil in a world that is created entirely good, with no defect or corruption, Augustine realizes that this "requires him to maintain that the first sinners are not created defective in any morally relevant way — that is, that the moral defects constituted by their primal sin are not preceded by any other morally relevant flaw in creation." Choices like Satan's defection from the ranks of the angels and Adam and Eve's decision to eat the forbidden fruit illustrate the "sheer willfulness of sin."

Near the end of On free choice of the will, Augustine brings together his concerns to provide a theodicy and to justify divine punishment, saying that God deserves no blame when somebody fails to do what he ought but deserves praise when somebody suffers what he ought (De lib. arb. 3.16), and that people repay their debt to God, their creator from whom all good things come, either by doing what they ought or by suffering what they ought (De lib. arb. 3.15).

By this point, we have traveled quite far from Epicurus' concerns. Carneades' modifications of Epicurus' position, in order to strengthen it as a counterweight to Stoic compatibilism, are quite understandable. So too are the ways in which Augustine appropriates for his own theological purposes what he found in Cicero of Carneades' theory. But if Epicurus had foreknown what the intellectual descendant of his theory was going to be like — a notion of a will, whereby we can perversely reject the verdicts of right reason, a notion which is used to defend the existence of a providential creator-god against the problem of evil and to justify eternal afterlife punishments at the hands of that god — he would have turned away in horror and dismay.