UNPLEASANT DIFFICULTIES

What can be so common and so well known to us as pleasure? Do we not seek it throughout the day? Does it not frequently guide or even sway our actions? Many people apparently even live for pleasures of a certain sort.

And yet it seems almost impossible to say what pleasure is. Is it a sensation, like seeing a patch of blue sky? Or is it a feeling, like joy and delight? If it is feeling, can it be enduring, like a “mood,” or is pleasure something passing, more like an emotion? But perhaps pleasure is not a “thing” that we are related to at all, but rather our being related to something in a certain way, so that to be pleased simply is to “take pleasure” in something that is not itself a pleasure. Yet is it not also pleasant to take pleasure in something? (How could it not be?) But then, if it is pleasant to be pleased, a pleasure could serve as an object of pleasure.

Moreover, what is the relationship between pleasure and attraction? Might we be repelled by a pleasure, qua pleasure, or is that suggestion nonsense? But if we are necessarily attracted to a pleasure, must we not in some sense inevitably take it to be good? It would be, if a good is a goal. Or at least this much seems true: the fact that something is pleasant seems to be some kind of sign that it is good.

Let us review the argument to this point. In book 1 Aristotle argues that the ultimate goal of human life is some “activity in accordance with virtue,” and he sets down criteria that he thinks the ultimate goal must satisfy. In 1.13 and 2.1 he analyzes human virtue into thinking-related and character-related virtue, and in the remainder of book 2 he constructs a list, which he regards as exhaustive, of particular character-related virtues. One would think that his next step would be to examine these virtues individually. In fact he puts this off until 3.6. Instead, in 3.1–5 he undertakes a discussion of human action and the psychology underlying human action.

This is not a digression, as it might seem, because Aristotle's discussion of human action follows naturally from, and is even required by, various claims that he had put forward in book 2. Recall that, as Aristotle sees it, the discipline of ethics is practical: its aim is that we become good and live well. But, as he has just argued in book 2, we cannot become good unless we acquire the virtues of character. As we saw, Aristotle thinks that we generally acquire the virtues through being directed, by those in authority over us (by our parents when we are immature, and by legislators, even when we are mature citizens), to do actions characteristic of those virtues and to refrain from actions characteristic of the contrary vices.

I vividly remember my first encounter with Aristotle's Nicomachean Ethics in my first semester of college. I was assigned the text as part of an introductory course in the history of philosophy. My professor, Ed McCann, had said in lecture that it was widely accepted that Aristotle and Kant towered above all other philosophers, on account of their depth and comprehensiveness. So I had the highest expectations as I went off to the library, the text of the Nicomachean Ethics in hand, to grapple with Aristotle's thought.

But lulled perhaps by the soft hum of the heating system in the library, or by the plush comfort of the leather chair into which I had sunk, I simply could not stay awake while reading. I would read a chapter or two of the Ethics; then nod off to sleep; then wake up and read another chapter; and then fall asleep again; and so on. During my brief periods of wakefulness, it was my impression that I was following the argument, and that what Aristotle was saying was, after all, commonsensical – a very common first impression of the Ethics, as it turns out. And yet really I was hardly understanding the text. What was happening was that the seeming obviousness of Aristotle's claims allowed me to run my eyes over the text fairly quickly, and yet the density and concentration of the underlying argument, to the extent that I did grasp it, caused a kind of intellectual overload, from which I would then escape by falling asleep.

One of the most startling and distinctive aspects of Epicurean philosophy is the atomic motion known as the ‘swerve.’ The Epicureans are materialists, holding that the only things that exist per se are bodies and ‘void,’ which is just empty space. Bodies are simply conglomerations of atoms, which are uncuttable, extended bits of ‘full’ space flying through the void as a result of their weight, past motions, and collisions with other atoms. But the Epicurean poet Lucretius writes that if all atomic motion were the deterministic result of past motions and weight, we would not have the ‘free volition’ (libera voluntas) which allows each of us to move ourselves as we wish. Since we evidently do have the power to move ourselves as we wish, there must be a third, indeterministic cause of atomic motion, in addition to weight and past motions – a ‘swerving’ of the atoms to the side at uncertain times and places, which saves us from fate.

In part, this book is an attempt to discern the role the swerve plays in preserving human freedom. However, the swerve cannot be studied in isolation; it must be understood in the context of Epicurus' ethics, philosophy of mind, and metaphysics in general. So the subject of this book is Epicurus' overall theory of human freedom.

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?

My argumentation thus far has been partly positive and partly negative. As far as how the swerve is supposed to help secure our freedom is concerned, it has been largely destructive: the swerve is not involved directly in the production of free action; it is not supposed to secure the agent as the archê of either his character or his actions; it is not needed to protect the emergent self from the threat of reductionism. Neither De rerum natura, nor supposed Aristotelian antecedents in the Nicomachean Ethics, nor On Nature 25 allow us to determine the role the swerve is supposed to play in protecting human freedom.

However, many important positive results about Epicurus' views on our freedom can be gleaned from these texts, as well as Epicurean ethics and psychology overall. Lucretius' discussion shows that libera voluntas is what allows us to act as we wish to act, and that it is a sort of intentional impulse. Somehow, if sequences of cause and effect stretched back infinitely in the past, we would not have this libera voluntas, and the swerve saves us from this. Insofar as Epicurus is picking up on Aristotelian antecedents in the Nicomachean Ethics, he would agree with Aristotle that it is crucial that an agent's character and actions have their origin in the agent himself.

The natural place to look when trying to figure out the role of the swerve in free action is Lucretius' description in DRN 2 251–293 of why the swerve is needed to preserve our voluntas, and a great deal of attention has been lavished on this passage. So, in this chapter I will consider what we can learn about the Epicurean position from Lucretius. As far as the swerve itself is concerned, my conclusions will be mostly negative: Lucretius' description of the swerve tells us remarkably little about the role it is supposed to play in preserving our freedom. However, his descriptions in DRN 4 877–896 of voluntary action and in DRN 2 251–293 of the voluntas the swerve preserves both show what sort of freedom the Epicureans are concerned to defend against the threat of determinism. I will first examine DRN 2 251–293, and then DRN 4 877–896, and show that the De rerum natura undercuts the thesis that Epicurus' concerns are much like those of modern libertarians. Finally, I will turn explicitly to a consideration of the traditional interpretation, that every volition is constituted by a swerve, since the main textual support for this family of interpretations is DRN 2 251–293, and I will argue that De rerum natura actually gives us good reason to reject this interpretation of the Epicurean position.

DRN 2 251–293

DRN 2 251–293 is the longest passage we have by an Epicurean that describes the connection of the swerve to free action.

If our sources can be trusted, Epicurus asserts an incompatibility between libera voluntas (sometimes translated as ‘free will’) and determinism, and he denies that determinism is true, positing a mechanism, the swerve, by which it is rendered false, saying that the swerve is needed in order for us to be free. Because of this, Epicurus has been hailed as the first person to discover the free will problem. But this is too hasty. Much of the recent discussion of Epicurus has been muddled by assimilating his position and his concerns to those of modern libertarians. Before we answer the question of how the swerve is supposed to secure our freedom, at least two other questions need to be addressed: What type of freedom is the swerve supposed to secure? And why should we care about having freedom in this sense – that is, why does it matter? Only if Epicurus is concerned to secure the same type of freedom as are modern libertarians, and for the same sort of ethical concerns, are we justified in calling Epicurus a ‘libertarian.’ Before considering in detail the passages in which Epicurus' theory of freedom and the role of the swerve are discussed, in this chapter I will lay some conceptual groundwork that I hope will clarify the following discussion. First, I will set forth (rather schematically) some varieties of ‘problems of free will,’ and I will situate interpretations of Epicurus, including my own, in terms of these varieties.

Before I turn to the role the swerve plays in preserving human freedom, I wish to consider the other reason Lucretius gives for why there must be an atomic swerve: without the swerve, there would be no atomic collisions, and thus no macroscopic bodies, as there evidently are. Compared to the extensive treatment of the anti‐fatalist function of the swerve, the ‘cosmogonic’ argument for the swerve has garnered relatively little attention. Most treatments of Epicureanism either paraphrase Lucretius' argument in De rerum natura without giving any extensive analysis of whether the argument is cogent or simply dismiss it as inadequate.

This disparity of attention is understandable. In connection with determinism, the swerve may ultimately be a mistake, but at least it opens up interesting questions about the relationship between causal determinism and free will. Lucretius' argument that the swerve is needed for atomic collisions, however, appears to rest on a simple misunderstanding – i.e., that there needs to be a start for collisions – and once this misunderstanding is pointed out, the argument has, at best, some minor historical interest. In this chapter, however, I want to rehabilitate the idea that Epicurus had good reasons to think that the swerve was needed as an archê of collisions, and that the swerve – in this connection, at least – was not simply a blunder or a misguided oversight.

This chapter contains five main sections.

A natural strategy to pursue when trying to understand Epicurus' position is to find a precedent in the works of previous philosophers in the light of which we can better understand how Epicurus thinks the swerve preserves our freedom. After all, it is more likely that the swerve was introduced by Epicurus as a reaction to a problem already placed on the philosophical agenda by his predecessors, instead of it being a radically new innovation without precedent. The predecessor most commonly enlisted to play this role since David Furley's influential study is Aristotle.

I am quite sympathetic to this way of approaching Epicurus' thought. In fact, I will be arguing (in chapter 6) that a key to understanding Epicurus' use of the swerve is to look back to Aristotle's arguments in de Int. 9, in which he contends that the unrestricted application of the Principle of Bivalence to all statements (including ones concerning contingent matters in the future) would have unacceptable fatalist consequences. In this chapter, however, I will be examining a different set of Aristotelian texts, particularly those dealing with voluntary action in Nicomachean Ethics book 3. In NE 3, Aristotle says that, in order for an agent to be responsible for what he does, the ‘origin,’ or archê, of his action and character must be within him.

Other than Aristotle, Democritus is most often cited as the person Epicurus is responding to when formulating his position on freedom. Surely this is correct. For the most part, Epicurus appropriates his atomist metaphysics from Democritus. However, the swerve is one of his major modifications to Democritus. The Epicurean Diogenes of Oinoanda writes that Epicurus ‘discovered’ the swerve, an atomic motion that Democritus did not discern, because Democritus failed to recognize that a swerving motion was needed in order for us to be free, and Epicurus himself writes that Democritus' metaphysics, if true, would have disastrous consequences for our agency – consequences that Democritus himself did not see.

Of course, it is one thing to say that Epicurus feels the need to modify Democritus' metaphysics because it has unacceptable consequences as far as our freedom is concerned, quite another to specify what those consequences are, why Epicurus thinks they follow, and how exactly he modifies Democritus' metaphysics to avoid them. I will make two related claims about Epicurus' response to Democritus. The first is that Epicurus objects to the consequences of Democritean eliminativism; in response, Epicurus tries to show how a reductionist atomism need not discount the reality and causal efficacy of phenomena like minds and reason. The second is that Epicurus primarily worries about the fatalist consequences of Democritus' eliminativism, not determinism as such. Let me briefly expand on these two cryptic claims.

In Western Europe during the twelfth and thirteenth centuries, interesting theories concerning various aspects of language were prolific. In intellectual energy, these centuries could compete with those of Plato, Aristotle, Theophrastus, Epicurus and Chrysippus. One might add the first half of the fourteenth century, but I shall leave it out of account, mainly because the new developments in semantics (due to Burley, Ockham, Buridan and others) were accompanied by an astonishing loss of interest in grammar.

In trying to compare and connect some pieces of scholastic theory with Hellenistic theory I use a totally conventional limitation of the Hellenistic age to the years between 323 bc and 31 bc. I take it for granted, though, that linguistic doctrine attested in Quintilian and/or Sextus Empiricus is almost invariably of Hellenistic origin.

At the beginning of the twenty-first century we do not possess a single major work on linguistic issues from the time between Aristotle and Apollonius Dyscolus. Our scholastic predecessors in academe had even less access to the theories of that age than we have. Inter alia, most of the Greek texts from which we collect our fragments of the theories of the various schools had not been translated into Latin, and nobody read Greek. Besides, some of our Latin ‘classics’ were virtually unknown. Quintilian, for instance, was not a household name to medieval scholars.

For all that, the influence of Hellenistic theorising on developments more than a thousand years later cannot be doubted.