by Tim O'Keefe
1.
If atoms did not swerve, there would not be
“free volition” (libera voluntas).
2.
There is free volition.
* * *
3.
Therefore, atoms swerve.
The argument follows the same familiar Epicurean pattern as the argument for void and the first argument for the swerve’s existence. Each argument starts from something evident in perception, such as the existence of motion or of compound bodies, and on its basis infers a conclusion about what is not directly accessible to perception, such as void or the swerve. In this passage, Lucretius says that we cannot directly perceive individual atomic swerves, but that animals act freely, which we can see, allows us to infer that they exist.
Lucretius writes that an occasional random atomic swerve initiates new motion, which prevents the existence of an endless chain of atomic causation, of new motion unalterably arising out of old. This swerve annuls the decrees of fate and allows us to have free volition.
But how is that supposed to work? Lucretius does not tell us, but a natural line of interpretation is the following:2 Epicurus appropriates most of his metaphysics and physics from Democritus, including the identification of the mind with a bodily organ, an atomic aggregate. However, he recognizes that Democritus’ atomism has unacceptable deterministic consequences. If all atomic motions are causally necessitated, then our decisions, which are identical to atomic motions in our minds, would likewise be causally necessary, as would the actions that flow from those decisions.
On this interpretation, what is wrong with this possibility is that we must have the ability to do otherwise than we do in order to be free and morally responsible. For example, imagine that a student insults my fashion sense, and I angrily punch him in the face. If Democritus were right, then given the disposition of my mind at the time, plus the (atomic) stimulus of hearing the student insulting me, my punching him in the face would be causally necessary. But then, I would not be able to do anything other than punch him in the face and, if that were so, I did not act freely, and I am not morally responsible for punching him.
To avoid this problem, Epicurus introduces the swerve, an indeterministic atomic motion, which allows him to deny that all atomic motions are causally determined. The decisions that produce our actions are identical to swerves that occur in our minds. After I hear the student insult me, an atomic swerve – or maybe a series of atomic swerves – occurs in my mind and initiates my action of punching him in the face. Thus, I did not have to punch him in the face; if the decision/swerve producing that action had not occurred, or had occurred at a different time, I would have acted differently, and so I acted freely and can be held responsible for what I do.
If this interpretation is right, Epicurus would be rightly hailed (or derided) as one of the first people to formulate the problem of free will and determinism, and to offer a libertarian response to it.3 The problem of free will and determinism concerns the charge that causal determinism threatens our ability to do otherwise than we do, and that we must have this ability in order to be morally responsible for our actions. A libertarian believes that causal determinism and the ability to do other are incompatible with one another, while affirming that we do have free will, and so he rejects causal determinism.
Some of Lucretius’ language suggests this interpretation. He declares that libera voluntas is incompatible with determinism. While I have been rendering this phrase by “free volition”, a more common translation is “free will”. Furthermore, Lucretius draws an analogy between our free will and the atomic swerve; just as the atoms have the ability to initiate new motion by swerving, so too we can swerve off our course at no fixed time or place, wherever we wish. This analogy is easy to understand if our decisions are atomic swerves: we “swerve” like the atoms do because our decisions are atomic swerves.
Despite these attractions, the preceding line of interpretation suffers from some serious problems, both philosophical and textual. The philosophical problem is that a random atomic swerving in one’s mind is an unpromising basis for the production of free and responsible actions, instead of random and blameless twitches.4 Let us imagine that I am an Epicurean sage, and when the student insults my beloved but ratty T-shirts, I realize that a concern for fashion would be a vain and empty desire, and catering to the opinions of the many on this topic leads to unnecessary disturbance. As I prepare to walk away calmly, however, a random atomic swerve occurs, and I punch him in the face instead. Having random, uncaused swerves intervene between my desires and beliefs and the actions prompted by them would undercut rather than preserve my control over what I do. And within Epicureanism, it is difficult to see why being able to randomly swerve off in my actions occasionally in this way would be regarded as a valuable ability to be preserved rather than a crippling disability to be guarded against. There may be cases in which the results of the swerve happen to align with my beliefs and desires, but this would just be a lucky coincidence, it seems. It is hard to construct any story about the role of the swerve in which it actually helps makes actions more under the control of the agent, instead of merely sometimes not undercutting that control too much.
The textual problems are twofold. Less serious is an argument from silence: if swerves played a central role in the production of free actions, we would expect Lucretius to mention them in his account of how volition arises and moves the body, but he does not.5
More seriously, the “free volition” that Lucretius describes as preserved by the swerve bears little resemblance to the “ability to do otherwise than one does” that figures prominently in the problem of free will and determinism described above.
Lucretius spends most of his argument (DRN II 261–83) illustrating why its second premise (“there is free volition”) is supposed to be obviously true, and in so doing, he shows what sort of “free volition” determinism threatens. Free volition is what allows creatures throughout the earth, both human and non-human, to do what they want to do and to advance wherever pleasure leads them. Lucretius establishes that free volition exists by showing that the body follows the mind’s desire. He gives two examples. Both are meant to show that animals have an internal capacity to initiate or resist motion, and that this capacity distinguishes animal motion from the way in which inanimate objects are shoved around by external blows. Voluntary motion has an “internal source” in the literal sense of being produced by the animal’s mind, an organ in its chest.
The first example is of racehorses eager to burst from the gates (DRN II 263–71). Lucretius claims that we see a slight delay between the external stimulus of the gates’ opening and the resultant motion of the horses surging forward. This delay supposedly shows that motion initiated by the mind exists, as it takes some time for the mind’s decision to move all of the matter of the horse in a coordinated manner. Motion caused by external blows, on the other hand, does not require time for internal processing: a horse struck from behind by another horse is immediately pushed forwards.
The second example (DRN II 272–83) appeals to our own experience of situations such as being in a jostling crowd: we are not always helplessly shoved around by these outside forces but can sometimes fight against them to go where we wish. Imagine being carried down a river by its swift current unwillingly, sharp rocks looming downstream. Unlike an inanimate object, such as a log, we need not allow ourselves to be carried along but can fight against the current and swim for the shore in order to avoid danger.
The sort of freedom at stake here may be dubbed “effective agency”. Two differences between it and “free will” (as the phrase is often used) are worth underlining. First, effective agency is possessed by all animals that can do what they wish, including many that do not have the rational capacities needed to be rightly praised or blamed; many animals possess effective agency that do not have “free will”. Secondly, “effective agency” need not involve the ability to do otherwise than one does. The horses Lu
cretius describes at the starting gates are not trying to decide whether or not to break from the gates, and a man caught in a current is not concerned with whether or not to swim for the shore. Instead, as Lucretius portrays it, free volition is what allows them to move around in the world in order to obtain what they desire.
So even though Lucretius may seem initially to be saying that the swerve is needed to allow us to be able to do otherwise than we do, with our decisions being identified with atomic swerves, there are good reasons to doubt this. Instead, swerves are somehow needed for all animals to pursue pleasure.
“Effective agency” and the principle of bivalence
As Lucretius portrays it, then, libera voluntas is what allows us to be effective agents, to act as we wish to in order to get what we desire. However, he does little to explain how causal determinism threatens its existence, or how introducing a random atomic swerve overcomes this threat. For that, we need to turn to Cicero’s On Fate, which describes a debate between the Epicureans, Stoics and Academic Sceptics on issues of fate and freedom.
One of its central topics is the “lazy argument”, one member of a family of arguments, including the argument concerning tomorrow’s sea battle in Aristotle’s De Interpretatione 9, that try to show that accepting the universal applicability of the principle of bivalence would have unacceptable consequences on our agency. The principle of bivalence is the thesis that every proposition either is true or is false, including propositions about what will occur in the future. The type of determinism at issue here we might dub “logical” determinism”.6 Here is a sketch of how the “lazy argument” from bivalence goes. You are sick, and you are trying to decide whether or not to call a doctor. However, if you accept the principle of bivalence, then either it is true, and has always been true, that you are going to recover from the disease, or false (and always false) that you will recover, and hence always true that you will not recover (Fat. 29). But if either of two alternatives has been true from all eternity, that alternative is also necessary (Fat. 21), because the past is immutable (Fat. 19–20, 21, 28, 29). And because there is no point in deliberating about what is necessary, then it is pointless for me to worry now about whether or not to call the doctor, as if my present actions could change the outcome one way or the other (Fat. 28–9).
So, logical determinism is apparently incompatible with the contingency of the future, where this contingency is necessary for the effectiveness of deliberation and action. Epicurus asserts that it is obvious we engage in effective action and deliberation, that the future is therefore contingent, and accordingly he rejects “logical” determinism (i.e. he rejects the principle of bivalence). The “lazy argument” is quite different from the problem of free will and determinism described above, but if we keep the differences between the two in mind, we would not go too wrong in describing Epicurus as a “lazy argument libertarian”.
Causal considerations are not present in the lazy argument as I have described it. However, in order to escape the “necessity of fate” that this argument would establish, Epicurus posits the atomic swerve (Fat. 22; see also 18, 48). He does so because he thinks that logical and causal determinism are inter-entailing; let us call this the “inter-entailment thesis”. Both Epicurus and the Stoics say that things that will be true must have causes of their future being (Fat. 26; see also 19). The point is that, since the future is not yet – it has not obtained – there is not yet anything there in virtue of which a statement about the future can be true, unless there presently obtain conditions to bring about the state of affairs described by the statement. (Likewise, for a statement about the future to be false now, there must presently obtain conditions to preclude the state of affairs described by the statement.)
Consider a statement such as “Tim will die from his disease”. My death cannot make it true, as it has not yet occurred. For the statement to be true right now, there must obtain conditions at present sufficient to bring it about; for example, my skin cancer has metastasized throughout my body and made my death from the disease inevitable. If you are a “lazy argument libertarian” like Epicurus and accept the inter-entailment thesis, you need some sort of physical mechanism – like the swerve – to underwrite the rejection of the principle of bivalence. The swerve allows him to admit that statements such as “Tim will die” are already true (because, as a human being, there at present obtain conditions sufficient to ensure that I will die) whereas many statements like “Tim will die of such-and-such a disease” are for now neither true nor false, as their truth is not yet settled.
If it were already true either that “Tim will die from his disease” or “Tim will not die from the disease”, then it would be pointless to deliberate over which of the two to make true. But the swerve allows me to escape the necessity of fate. What will occur is not yet settled, and because the future is open in this way, I am free to decide to do as I wish in my pursuit of pleasure.
Academic criticisms of Epicurus
Carneades (214–129 BCE), the most prominent head of the sceptical Academy, makes two astute criticisms of the Epicurean position described above. The first is to deny the inter-entailment thesis. Carneades unequivocally accepts the principle of bivalence for all propositions, at all times: statements such as “Jimi Hendrix dies of a drug overdose” are eternally true (Fat. 37). He denies, however, that any deterministic consequences follow from the principle of bivalence, because the truth of a statement does not imply that there are “immutable eternal causes” that make it true (Fat. 28). Instead, if somebody were to say in 1965, “Jimi Hendrix will die of an overdose”, it is simply the fact that Jimi Hendrix will actually die of an overdose that makes the statement true. And now we know that a person who said that in 1965 did say something true, because things turned out as the person said they would.7
This, however, would not satisfy the Stoics and Epicurus, who would ask in virtue of what it was a fact that Hendrix would die, if there were not any causes obtaining at that time to make it a fact that this would occur. In order to defend his position, Carneades appeals to the symmetry of the past and future: just as something being true in the past does not depend on its having certain effects now, something being true in the future does not depend on its having certain causes now (Fat. 27). If somebody were to say in 2010, “Jimi Hendrix did die of an overdose”, it is simply the fact that Jimi Hendrix did actually die of an overdose that makes the statement true.
We can imagine, especially in an indeterministic world, certain past events whose effects are “washed out” over time: for example some atom decayed far in the past, at time t, and the present state of the world is compatible with that atom decaying either at that moment or at time t + 0.001s. We would then say not that there used to be a determinate truth about what time the atom decayed, but that since then it has become indeterminate whether the atom decayed at time t or not.
Therefore, Carneades thinks, logical determinism does not imply causal determinism. Carneades rejects the lazy argument, when it is stated in terms of truth, because the fact that it has always been true that something will occur does not make its occurrence necessary or inevitable in any way. If Carneades is right, then what I do now can have an effect on the past, in an attenuated sense. If the necessity of the past is tied to its irrevocability, then there is an asymmetry between the past and the future, in so far as I can affect the future, and not the past. I can kill myself today or tomorrow; I cannot kill myself yesterday. However, by my present actions, I can make it to have been true that something was going to occur. If I commit suicide, I make the statement “Tim O’Keefe is going to kill himself” to have been true in the past. If it is simply my freely deciding to blow my brains out that makes it to have always been the case that I would kill myself at that time, and thus for the statement “Tim O’Keefe will kill himself” to have been true for an eternity before I pulled the trigger, the eternal truth of that statement does nothing to threaten my freedom.
This brings us to Car
neades’ second criticism. Like Epicurus, Carneades rejects causal determinism, because he thinks it is incompatible with there being a voluntary movement of the mind and with anything being in our power (Fat. 19, 28, 38). Carneades denies that even Apollo can foreknow events such as 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 inevitably bring them about, such that Apollo could inspect the conditions at the time in order to tell that the actions are going to occur (Fat. 32–3).
Carneades agrees with Epicurus that if there were pre-existing causes determining what will occur, then no actions would be in our power, and we would be helpless. So when restated in terms of causal determinism instead of logical determinism, the lazy argument is powerful (Fat. 31). But he then rightly says that positing a motion without a cause, like the swerve, would be beside the point in solving the problem. Carneades’ solution is to say that 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 that has an intrinsic nature of being in our power and of obeying us (Fat. 24–5). For example, let us suppose that I am diagnosed with early-stage skin cancer and seek treatment. This action has a cause: me, and my deciding to seek the treatment. But that I engaged in that action was not itself causally necessitated; instead, it was entirely up to me.8 But if we have such a power to engage in voluntary action, actions that are under our control in this strong sense, there is no reason to posit, in addition, a fundamental physical indeterminism like the swerve.9
