SOURCE: Julia Annas, "Atomism and Agents," in Hellenistic Philosophy of Mind, University of California Press, 1992, pp. 123-56.

[Below, Annas examines Epicurus's physical theories in order to determine the Epicurean idea of the relationship of the human subject to the physical world, particularly to atoms, the universe, and the body.]

a) Physicalism and Reductivism

Epicurean and Stoic theories of the soul are often structurally very similar and sometimes also similar in detail. The two theories have very different metaphysical backing: the Stoics have a continuum theory of matter and hold that the universe is animate and runs by laws which reveal the workings of providence, while the Epicureans have an atomic theory of matter and reject all appeal to providence and any kind of teleology. They also have different ethical contexts: the Stoics think that rationality is what is crucially important in our ethical development, while Epicurus holds that our final end is pleasure, and that this is revealed to us directly by our feelings. However, the two theories share a common physicalist framework of thinking about the soul and in many ways have far more in common with each other than either does with a theory like Aristotle's. The chief differences are due to the fact that the Stoics are heavily influenced by contemporary medical and scientific theories, whereas Epicurus is less impressed by scientific results and more reliant on a combination of commonsense folk psychology and straightforward philosophical argument.¹

"The soul is a body of fine parts, spread over the whole assemblage."² Epicurus has one brisk argument for the soul's physicality, which appeals to the principles of Epicurean physics: everything in the world must be explained in terms only of Epicurus' meager ontology of atoms and void.³ This applies to the soul in just the same way that it applies to everything else. Thus, he says, we cannot conceive of anything existent that is not body, except the void. The void can neither act nor be acted on. But soul clearly does both. The soul therefore must be body. The crucial premise here is that only body (ultimately, atoms) can act or be acted on. Lucretius develops a different argument, from interaction: the soul moves the body, and what happens to the body affects the soul.⁴ Hence the soul is a body. Lucretius relies on the more round-about premise that interaction requires touch and that this requires body.

We have seen the Stoics use versions of these arguments;⁵ they are part of shared Hellenistic philosophical currency. They make it clear that Epicurus is a physicalist, as defined in part 1. Study of the soul is part of phusik, enquiry into the natural world; Epicurean phusike recognizes only two kinds of basic item, atoms and void, so the soul must be accounted for, in some way, in terms of atoms and void.

Epicurus is sometimes thought to have abandoned physicalism (or, alternatively, to have made his version of physicalism untenable) because he modified Democritean atomism by allowing a random "swerve" among the atoms; the swerve is connected in our chief sources with our having freedom of action,⁶ and it is often concluded that free human actions are, for Epicurus, due to events which breach regular Epicurean physical laws….

The swerve introduces an indeterminate element into physics, but this is a thesis within physics,⁷ not an abandonment of physicalism. It complicates the physical picture but is not a breach of it. It is not defined in terms of solving a problem of free agency; it is just a factor in the physical world, which operates in us, and also enables the production of worlds to get going.

It is often assumed that Epicurus was not only a physicalist but a reductivist. Possibly the tone of some of his writings may have given a handle to this: Epicurus is sometimes aggressively philistine, and we find Sextus asking the Epicureans how pleasure can exist in "the heap of atoms" they call the soul.⁸ But it is clear that Epicurus is not a reductivist from a striking passage in book 25 of On Nature, fragments of which we pos sess.⁹ Epicurus argues in this passage against the thesis that all events in the world, including our actions and thoughts, come about "of necessity." He allows that the truth about the physical world is given by atomic theory, which is (apart from the swerve, which is not mentioned in this text) determinist. So at the atomic level, events do happen "of necessity." But it is a mistake to conclude from this that, at the macro-level, my actions happen of necessity. It seems as though they must; for my action of arguing, say, is atoms moving in various ways which come about of necessity: so how can it not come about of necessity that I am arguing? Against Democritus Epicurus argues that this involves one in a blindness to oneself that lands one in a self-refuting position.¹⁰

He begins from the fact that we have practices of praise, blame, and the like, which make sense only on the assumption that we are agents capable of choice between perceived alternatives, and not just nodes in causal chains. He contrasts these practices with the way we treat wild animals, which we do not treat as agents but merely handle as best we can.¹¹ Epicurus then addresses the reductivist opponent who claims that these beliefs and practices are undermined by the truth of deterministic atomism, since all our actions are "merely" movements of atoms, so that even our very praising and blaming are "of necessity." Epicurus maintains that

such an argument refutes itself [lit. turns itself upside down] and never can establish that everything is such as the things which are said to happen according to necessity. Rather, he combats a person on this very point as though it were because of himself that the person were being silly. And even if he goes on ad infinitum saying that the person is doing that according to necessity, always from arguments, he is failing to reason in that he ascribes to himself the cause of having reasoned correctly and to his opponent the cause of having reasoned incorrectly.¹² (Arrighetti [34.28]; Sedley 1983, 19; Long and Sedley 1987, 20 C [5]-[6])

Epicurus is using a standard ancient "overturning" argument.¹³ It involves what we would call pragmatic self-refutation. There is no formal self-contradiction, but what the person says or puts forward is, it is claimed, undermined by her way of saying it or putting it forward. An example would be proving that there is no such thing as proof. If I prove to you that there is no such thing as proof, then what I prove (that there is no such thing as proof) is pragmatically refuted by the fact that I prove it. If I win, I lose. Epicurus claims here that reductivists like Democritus fall into a trap of this form and that they are "blind to themselves" because they fail to notice this point.

The reductivist holds that because all human actions are movements of atoms, human actions are "necessi tated"; thus they are not really up to us, as we suppose them to be, and there is really no such thing as free human acting. If so, of course, there is really no such thing as arguing, criticizing, and so on; what we think to be such is "nothing but" atoms moving in the void in ways that they have to move. However, the reductivist argues against Epicurus to this effect, states and defends his view, criticizes Epicurus for getting it wrong, and so on. And all this undermines his thesis, since it presupposes that the thesis is mistaken. Hence he is landed with a conflict between his thesis and what he is doing in stating and defending his thesis. He can of course retreat, admitting of what he says that it was necessitated. But the retreat can never be quite fast enough; in admitting this he is presupposing the falsity of his thesis. As Epicurus points out, at every stage of his retreat what he does is in conflict with the thesis he holds.

Epicurus does not here meet the more sophisticated determinist who claims that the necessitation appropriate to atomic motions is not in conflict with human agency because it is compatible with it. Epicurus is here concerned only with the opponent who tries to undermine our everyday concept or prol psis of agency. The opponent, he says, is trying to change our concept of what it is to act. But he has not succeeded in doing this unless he can evade the selfrefutation argument, and otherwise he is in effect just changing a word by calling "necessitated" what we call free agency, and this is futile, since it makes no real difference.¹⁴

Epicurus' is the first in a long line of arguments to establish nonreductive physicalism by showing that reductivism (at least in a determinist version) cannot be consistently stated. So, there are facts about atoms and facts about human agency, and each set of facts will be real; it will be wrong to treat the latter as a mere appearance of the former.¹⁵ We should note that this argument does not show that Epicurus is not a determinist. It shows that he thinks that, properly understood, determinism must be compatible with our commonsense understanding of ourselves and of the world. The argument is thus more properly antireductivist than antideterminist.¹⁶

Epicurus is thus justified in distinguishing between what happens by necessity or by chance from what depends on us (par' h mas).¹¹ It is noteworthy that Epicurus does not claim that things are up to us (eph ' h min) but that they depend on us (par' h mas). He is defending the intuitive idea that we are agents, and seems not to want to defend a very strong and possibly unintuitive version of the idea.¹⁸

In particular, moral development is real; Lucretius insists that by reasoning the individual can overcome handicaps of inherited temperament.¹⁹ In sev eral unfortunately fragmentary and difficult parts of On Nature 25 Epicurus insists that our atomic constitution is to be distinguished from our "development" (apogegennēmenon), which depends on us. It depends on me, not just my atoms, how I develop and what kind of a person I become; even though it is a truth of physics that I am atoms. This is a defense of common sense: my physical makeup and the experience I have put some constraints on what I can become, but still how I develop depends on me. As Epicurus explains,

from the first beginning we have seeds directing us, some toward some things, others toward others, others toward both—in every case seeds, which may be many or few, of actions, thoughts, and dispositions.²⁰ Thus it depends on us at first absolutely what becomes of what is already a development, whether of one or another kind, and the things which of necessity flow in from the environment through the pores depend on us when they come about at some time, and depend on our beliefs that come from ourselves. (Arrighetti [34.26]; Sedley 1983, 36-37; Long and Sedley 1987, 20 C [1])

Epicurus clearly has great reliance on our commonsense view of ourselves as free, developing agents.²¹

In this passage, which is unfortunately both highly technical and very fragmentary, Epicurus talks of, on the one hand, the self ("we") and the development and, on the other, of the atoms, the nature, and the constitution. Sometimes the development seems to depend on the self (as in the above passage), but the text as a whole supports the view that Epicurus is simply talking about an agent who develops. Sometimes the agent is identified with her development, sometimes the development is discussed separately, as being an aspect of the agent as a whole. On the other side, the constitution (sometimes the original constitution) is distinguished from the development. The development changes the original constitution and gets it to change or "grow" in some respect. The nature is simply the nature of the constitution; and likewise the atoms are the atoms of the constitution. There is no implication that the self or the development are nonatomic, but the atoms of the constitution can be contrasted with the atoms which impinge from the outside and help to produce the development.²²

Epicurus thus sees us, commonsensically enough despite the jargon, as developing agents, indeed as agents who develop ourselves. Although humans are atomic compounds like any other, they differ from other kinds of atomic compounds in that their growth and functioning is not to be explained solely in terms of automatic response to stimuli from outside. How they develop depends to some extent, though not totally, on themselves, on what they do with the information they take in, how they decide to react selectively to it, and what kinds of character and dispositions they build up.²³ Thus Epicurus is concerned to do justice to folk psychology's belief that we are agents who move and develop ourselves. This point on its own does not determine either the outline or the detail of any metaphysical conception of the self; it is a minimal basis compatible with a number of different kinds of theoretical explanation. As to how such self-development is possible we find in the fragments only the point that we have from the very start "seeds," potentials for developing one way rather than the other. This is taken for granted and not further defended.

The only hint we find in the remains of this book as to how we develop ourselves lies in the reference to the information we take in from the environment depending on the beliefs we have. It is because we have reason and can form beliefs that we develop as agents; this is clear already from the passage from Lucretius which tells us that by developing our reason we can order the rest of our nature and overcome the tendencies we are born with. Reason, however, takes many forms; we shall see in the next section that they are not all limited to humans.

There is one passage of book 25 which seems to suggest something stronger than the commonsense picture:

Many [developments?]²⁴ which have a nature which is capable of becoming productive of both this and that through themselves do not become productive, and it is not because of the same cause in the atoms and in themselves.²⁵ These in particular we combat and rebuke…in accordance with²⁶ their nature, which is disturbed from the beginning, as is true of all animals; for in their case the nature of the atoms has contributed nothing to some of their actions, and to the extent of their actions and dispositions, but the developments themselves contain all or most of the cause of some of these things. As a result of that nature some of the atoms' motions are moved in a disturbed way, not in every way through the atoms, but through what enters…from the environment into the natural…combatting and advising many people together, which is opposed to the necessary cause of the same kind. Thus when something develops which has some distinctness among the atoms²⁷ in a differential way which is not like that from a different distance,²⁸ it acquires a cause from itself, then transmits it at once to the primary natures and in some way makes all of it one.²⁹ (Arrighetti [34.21-22]; Sedley 1983, 36-38; Long and Sedley 1987, 20 Β [1]-[6]; Laursen 1988, 17-18)

This passage has been made the basis for claims that Epicurus holds that we are not just agents in the commonsense understanding of self-developers, but However, selves in a it way is important altogether. that transcends to note atomism that this passage³⁰ con cerns disordered people, in whom something, though it is not clear just what, has gone wrong. Furthermore, even in these people it is only some of their actions to which the atoms contribute nothing,³¹ and even then we are told that the development accounts for all or most of the cause of what they do. This is, therefore, not an account of normal agency, and so cannot give us Epicurus' own view of human agency.

However, it is interesting to us even as an account of deviant agency, and it is a pity that it is so hard to see just what has gone wrong. Epicurus says that the condition of "all animals" is disturbed from the beginning; this may well include humans, and seems to embody the idea that we achieve the desirable Epicurean ethical end of untroubledness (ataraxia) as we mature, by imposing order on our initially disorderly nature. What is wrong with the agents here is that they are like immature, disorderly agents. However, this does not seem to be what they are themselves, since the passage suggests that they are perverse or deviant, rather than immature. Perhaps the nearest we can get to a general interpretation of what is going on is that two things are true of these agents. The way they are developing is at odds with their constitution. They are developing, or trying to develop, in ways that do not fit the way they have developed hitherto. Secondly, their constitution is, as a result, disorderly, like the initial state of immature agents.

However we interpret the details of this passage, the overall picture which emerges is that of an Epicurus who is impressed by the fact that humans are a self-moving kind of thing, with potential to develop in diverse ways, but who does not react by abandoning physicalism. We shall see further that he is quite ingenious in working out details of a thoroughly physicalist account of the soul. He is aware that there is a tension between reductive physicalism and our commonsense view of ourselves, which he wants to preserve, and he responds with an argument to show that reductivism cannot, in principle, be true. It is clear even by this point that Epicurus' account is answering to a number of constraints. His philosophy of mind must be physicalist, and in particular must be developed within his version of atomism. It is also, as we shall see, an empiricist account. But it also takes very seriously what we believe about ourselves, and where this conflicts with a possible way for his theory to develop, it is the theory he rejects and not our intuitive picture of ourselves as developing agents. We have to wait until later to find out what makes us free agents.

b) Soul in the World

The Stoic cosmos is animate, designed by providence, and permeated by reason. The Epicurean cosmos is none of these things, and this makes a big difference to the place in it of human beings.

The motions of atoms in the void give rise to compounds among which are animate, sensing, beings; but the atoms themselves are inanimate;³² Epicurus rejects the panpsychist demand that life be present in the ultimate constituents of what has life. Hence there is no Epicurean world soul; living beings do not display in themselves the workings of principles of life also at work in the universe as a whole.

Further, Epicurus rejects providence and any teleology, both for the world as a whole and for its parts.³³ Since the world is not an ordered whole, humans have no particular place in it; there is no scale of beings. "The soul is a peculiar kind of thing, like nothing else."³⁴ For Epicurus there is also nothing like Stoic rationality which permeates the world and gives it (in many senses) significance. Rationality does not cut humans sharply off from animals as it does for the Stoics. Animals as well as humans act freely; at least Lucretius illustrates the existence of free impulse (libera voluntas) from the example of horses in a race.³⁵ Further, a passage in On Nature 25 distinguishes between wild and tame animals by the extent to which their reaction is straightforwardly caused by input from the environment, or depends on the animal itself.³⁶

The Epicureans in fact have a position strikingly different from the Stoics' on "the reason of animals." Not only does Lucretius talk of horses having libera voluntas, he talks of horses and deer as having a mind.³⁷ However, Hermarchus, an early Epicurean, denies, in a discussion of justice and animals, that animals have logos or reasoning. That is, he says; why we can make no contracts with them.³⁸ Humans, as opposed to animals, have advanced in civilization because they can reason about what is in their interests, whereas animals have only "irrational memory."³⁹

Are these mutually contradictory views on the part of different branches of the school? They seem rather to be a matter of differences of emphasis. We find a more nuanced view in a later head of the school, Polystratus.⁴⁰ Animals, he says, share broad general features with us but are importantly different. They take in, but do not understand as we do,⁴¹ certain things: prudential concepts (healthy, expedient), ethical concepts (fine, base), religious concepts (sacred, profane), and signs (sēmeia). The last amounts to the claim that animals have no inferential reasoning; this explains, for Polystratus, why they cannot foresee problems, learn from the past, assess their own interests, or reflect on their lives as wholes.⁴² So animals "do not share in reasoning, or not one like ours."⁴³ Yet Polystratus thinks it ridiculous to deny that we are in general ways like animals, as the Stoics do.

The obvious way to make all this consistent is to recognize that animals have some reasoning capacities, but not others; in particular, not the ones that distinguish humans, which we inevitably call the higher ones. This is a commonsense conclusion, but it has an important, though overlooked, consequence: for the Epicureans rationality is not a single kind of thing but a cluster of capacities, some of which animals share with us and some not. We shall see that Epicurus frequently falls into difficulties over the status of the rational part of the soul, and much of his philosophy of mind would have benefited from taking more to heart this consequence of denying a sharp cutoff between humans and other animals.⁴⁴

Kuria doxa 32 encapsulates the Epicurean attitude to what divides us from animals: "As for those animals that cannot make contracts about not harming one another or being harmed—toward these there is no just or unjust; and similarly with those nations that cannot or will not make contracts about not harming or being harmed."⁴⁵ We do not owe duties of justice to animals; but this is merely because they do not have enough reasoning capacity to make and keep contracts—something true of some humans also. And we can see from book 5 of Lucretius how deeply ambivalent the Epicureans are about the "progress" of civilization and the ways in which we have used our reasoning to differentiate ourselves from animals.

c) The Nature and Structure of the Soul

"The soul consists of the smoothest and roundest atoms, greatly superior [sc. in these respects] to those of fire."⁴⁶ The soul animates the entire body without depending on bulk or brute force, merely because of the nature of its composition. "The soul provides nature with the reason for the [presence or absence of] life. For even though it does not possess the same number of atoms as the body, being placed in it with its rational and irrational elements, nevertheless it encompasses the whole body and, being bound by it, binds it in its turn, just as the shortest dash of acid juice curdles a vast quantity of milk."⁴⁷

The soul is a combination of four kinds of soul atom. It is puzzling that Epicurus' own Letter to Herodotus 63 so understates the doctrine as to seriously misleading,⁴⁸ but we know from other sources that the soul is constituted of atoms of four kinds: firelike, airlike, pneuma-like, and nameless.⁴⁹ The claim that the first three kinds of atoms are like the atoms of fire, air, and so on presumably amounts to something like the following. The soul does not contain just the kind of fire that we find in fireplaces, but something which is like that in basic respects, but more refined (it does not burn the rest of the soul, for example). The idea we have of it comes from our idea of fire, indeed for Epicurus it has to, since he is an empiricist and holds that our concepts are built up from what we encounter in experience. Thus our concept of it is simply something firelike, since we have no direct access to it in experience; all we can do is simply extend the experiential conception that we do have.

The basis for this account of the soul's composition is just the commonsense observation that "a certain thin breath mixed with heat leaves the dying, and heat, further, brings air along with it."⁵⁰ It is notable here that pneuma has retained its commonsense meaning of "breath," in contrast to its dramatic theoretical development in Aristotle and the medical writers. So far is Epicurus from what was to become the scientific mainstream, in which pneuma is essentially warm, that his pneuma is characteristically cold.⁵¹

Lucretius develops a theory about differing contributions made by the first three elements.⁵² Fierce lions have a preponderance of heat; timid stags illustrate the dominance of cold pneuma, and placid cows that of stable air. He goes on to apply the idea to explain differences of temperaments between individual humans; it is not clear whether this is his own contribution, or how it is to be extended from the idea of type differences.

The fourth, nameless element has a privileged position. It greatly exceeds the other elements in the fineness of its parts (leptomereid) and "thus is more sensitive to (sumpathēs) the entire assemblage."⁵³ According to Plutarch it is from this nameless element that there comes about "that by which the agent judges and remembers and loves and hates, and in general the intelligence and reasoning."⁵⁴ According to Aëtius the fourth element is the only one that can produce sensation.⁵⁵ So the fourth atom type seems to be responsible for sensation, thought, emotion, and memory.

Why is the fourth element nameless? Epicurus is hardly reluctant to coin new jargon elsewhere. Here he is constrained by his empiricism about concepts and language. We have some idea of what the firelike atoms in the soul are like from our experiences with fire, which have led us to produce the word "fire"; our concept of the ingredient in the soul works outward from this. But in the case of the fourth kind of atom there is nothing in our experience capable of giving us any, even partial, idea of what it is like. Not only do we never encounter it, we never even encounter anything that stands to it the way fire in fireplaces, stands to the firelike atoms in the soul. The nameless kind of atom is the only purely theoretical entity in Epicureanism. Even in the case of atoms and void we can conceive both by extension from things in our experience which are indivisible and empty.

Many have found the anomalous nature of the nameless atoms an embarrassment. Critics ancient and modern have claimed that here Epicurus is driven back upon a something he knows not what, and that this really amounts to an abandonment of physicalism; for the nameless kind of atom is physical, but, in appealing to something that has no experiential basis whatever, Epicurus is just providing a stand-in for everything that is hard to explain, given a physicalist position.⁵⁶ It is undeniable that Epicurus is weakening his empiricism here to a great extent; we have to rely on there being a theoretical entity which does a great deal of work in the theory but of which we have no idea at all from experience. But this need not be seen as objectionable; indeed it can be seen as merely realistic. Our idea of the soul goes far beyond what we can readily extrapolate from the natures of fire, air, and pneuma. Nor is there any reason to think that Epicurus is abandoning physicalism, thinking of the fourth kind of atom as in effect a magic addition which will bring to life something that physicalist principles cannot account for.⁵⁷

The role of the fourth element emerges from Lucretius.⁵⁸ The motions of the atoms, he says, so interpenetrate that they cannot be separated, nor can their properties be divided off. The atoms (the kinds of atom, presumably) are like the many powers of a single body. A living creature is one thing, although it has many properties like smell, heat, and taste; similarly, the kinds of atom form "a single nature." We are reminded of the comparison of the Stoic unified soul with its different powers to an apple with its different properties. The fourth element is the power that makes the soul into a unity—without it, Lucretius says, the other three kinds of atom would not hold together and be enabled to function as they in fact do in an animate body.⁵⁹

The fourth element is "hidden deepest" in the soul, as the soul is in the body; it is "the soul of the soul" and "runs things in the entire body." Clearly the fourth element is not spatially farthest inside, boxed in by the other three.⁶⁰ The soul, after all, is not boxed in by the body. Rather the soul is "hidden" in the sense that we do not encounter it in experience. We see clearly enough the effects of having a soul: it animates and directs the body. But the soul itself is not open to observation. Similarly, the fourth element is what "animates" the soul. Although we cannot observe the soul, we can make inferences as to its nature, and in particular infer the existence of a kind of atom which gets the soul to function as a whole, and which is distinct from the other soul elements whose nature we can partially describe from experience.

Soul and body, as Lucretius says,⁶¹ are mutually dependent: soul is like the scent in a perfume which you cannot remove without destroying the substance. And the fourth element stands to the soul as the soul stands to the body; it and the other soul elements are mutually dependent in that without them it would have nothing to "animate," and without it they would not hold together as a single kind of thing. How does the fourth element do this? It cannot be by operating, in a seemingly magical fashion, on its own. Rather, it must, by its particularly fine nature, enable the other elements to come together in a new sort of compound. It makes the soul a unity in the straightforward sense that its nature forms the necessary basis for the other atom kinds to cohere in a compound that has the properties of a soul. The introduction of the fourth element marks an insistence that there is a physical difference between souls and other kinds of body.

In one way this fits well into Epicurean theory: the soul's operations are supposed to involve particularly fine, invisible processes, and the fourth element serves to explain how the soul, though physical, can have a peculiarly fine structure enabling these to occur. But in other ways the move seems undermotivated. Epicurean physics and cosmology operate with atoms and void: atomic motions and the resulting compounds they give rise to are all we have to explain the varied phenomenal world. Faced by a complex and self-reproducing kind of thing like a tree, an Epicurean has to admit that the way it grows and reproduces is accounted for by its pattern of functional organization, which is stable enough to establish trees as things with persisting natures. Given an ontology as meager as that of atoms and void, and a rejection of teleology, patterns of functional organization are required to explain a world where things fall into species with stable behavior. But why will the approach deemed adequate to explain the species-specific behavior of trees not suffice to explain the behavior of people? To reverse the point, if we need a special kind of nameless atom to explain what souls are, why do we not need another kind of atom to explain what trees are?

It may be that Epicurus simply thought that animals and humans are so different in their complexity from things like trees that the same type of explanation would leave something out in their case. More likely, he may have thought that appealing merely to patterns of functional organization in the case of humans to explain what is characteristic of them was problematic from the point of view of atomist methodology. It is all right to say that a tree is the kind of thing it is because its atoms are organized in a particular stably functioning way. But to say this of humans might sound dangerously close to Aristotle, and would verge on recognizing a metaphysical principle like form as being as basic for explanation as matter. If it really provides an explanation to say that I perceive and act because there are stable perceptive, reactive, and so on patterns of functioning which my soul enables my body to carry out, these patterns seem to have a large explanatory role. And we can see why Epicurus would find this problematic; large differences of explanatory role ought, in a physicalist system, to have a physical basis. Thus the nameless atom type, far from signaling a retreat from physicalism, reveals confidence in the adequacy of physicalism as a theory of the soul. There is a physical difference between souls and other kinds of thing; so we do not need anything like Aristotelian forms to explain the way the soul functions.

Is this move successful? Aristotle argues that ignoring the role of form leaves us unable to explain functioning. Is the postulation of a physical difference, a new kind of ingredient, adequate to meet this kind of challenge? We might feel unhappy when we recall that the ingredient is nameless, since theory postulates something of which experience gives us no idea. A successful challenge to Aristotle would rely on achieved science and point to acknowledge complexity of structure to do the explanatory work assigned to form. But not only is Epicurus not in a position to appeal to such science, he is in general not very interested in low-level, working science. He accepts atomism as the best available scientific theory and tends to assume that what is needed can be worked out within atomism, without waiting for actual research. Thus in his appeal to nameless atoms there is a considerable element of faith—the kind of faith in science which philosophers often have who do not do any actual science.

Epicurus' account of the soul tries to interpret common sense in terms of atomic theory. Unlike the Stoics, he does not try to push the interpretation of soul in the direction of the mental. He accounts for much of what we call the mental by the rational soul, but the rational soul is merely a part of the whole soul, and that is clearly taken to be the physical basis of all the functionings of a living thing. We can see from a fragment of Diogenes of Oenonda how closely Epicurus stays to common sense:

Often when the body has been brought to surrender by a long illness, and reduced to such thinness and wasting that the dry skin is almost adhering to the bones whilst the nature of the inward parts seems empty and bloodless, nevertheless the soul stands its ground, and does not permit the creature to die. And this is not the only indication of supremacy: the severing of hands, and often the removal of whole arms or feet by fire and steel cannot undo the bonds of life. So great is the sway of life held by that part of us which is soul. (Frag. 37, cols. 2-3)⁶²

Soul is what makes us alive, and so functioning. This is a commonplace, but Epicureanism stresses the importance of rightly understanding the commonplace.

d) Parts of the Soul

The soul is not uniform; "the rational part" (to logikon) is located in the chest, while the rest, "the irrational part," is diffused through the whole body. This part of the theory, surprisingly absent from the Letter to Herodotus, is well attested in a scholium on the letter and, in the same words, in Diogenes of Oenonasa.⁶³ Lucretius makes much of it; he calls the parts animus and anima, elegant Latin which unfortunately loses the point that the animus is the rational part and the anima the irrational one.

The rational part is responsible not only for reasoning and cognition but for emotions such as "fears and joys." Lucretius says that in it are located both the understanding (consilium) and the governing (regimen) of life.⁶⁴ In fact the irrational soul tends to be thought of as responsible solely for perception, in which role it has some independence: the eyes themselves see, rather than being windows through which the rational soul sees.⁶⁵

There is a clear contrast with the Stoics, who put perception and impulse together as characterizing the whole soul, and who take thinking to be involved in all the soul's activities. In fact, while the Epicurean rational soul is bound to remind us in some ways of the Stoic hēgemonikon—it centralizes all the soul's activities, for example—there are striking differences. For the Epicureans sensation is registered in the sensing organ; for the Stoics the sensation is registered in the hēgemonikon. Thus the Epicurean rational soul is not involved in all events in the soul in the way that is true of the hēgemonikon. And while the Stoics come to use "the hēgemonikon" to refer to the soul as a whole, this is not the case with the Epicurean rational soul. In fact Lucretius says explicitly that when he refers to the soul as a whole he will use anima, the word for the irrational soul.⁶⁶ This is surprising, and in many ways unfortunate. One wonders whether he would have done so as readily had he been using words which reflected the fact that the parts are introduced by Epicurus in ways that make clear their relation to rationality.

The rational soul is located in the chest, because this is the region of emotions.⁶⁷ This is reminiscent of Chrysippus' insistence that the hēgemonikon is in the chest and not the head. Two interesting fragments of Demetrius Lacon show that later Epicureans had to contend, much as Chrysippus did, with the discovery of the function of the brain and the nervous system⁶⁸ Demetrius mentions Epicurus' view that the location of the soul's reasoning part allows of enquiry that is both practical (pragmatikē) and rational (kata logon). There is a claim that it is obvious that movement and emotion "drag" toward the chest. "Many doctors" are mentioned, who use some inductive reasoning (sēmeiō sis) to establish that reasoning is in the head. We seem to have a fragment of a confrontation very similar to the Stoic one.⁶⁹ Scientific research points to the role of the brain; but the philosophers refuse to abandon folk psychology.

However, the Epicureans' response differs in two ways. Firstly, they are in general not much impressed by the lower levels of science. Secondly, they do not have Chrysippus' reluctance to divide the soul; for the Epicurean rational soul is not the rational aspect of the whole soul. The whole soul is not rational; the rational soul is a part of the whole, as much a part as is a hand or an eye.⁷⁰ It organizes and so dominates the soul's activities, so that it can function as a relatively independent part, while the irrational soul depends on it.⁷¹ It is located in a specific part of the body, and damage there is more destructive to life than damage to other parts.⁷²

Two questions suggest themselves. Is the rational soul itself a unity? We have already seen that the conclusion suggested by Epicurean views of humans and animals is that it is more like a cluster capacities.⁷³ Rationality is shown in a variety of ways and comes in different kinds.⁷⁴ It cannot be said, however, that the Epicureans recognize this point explicitly, as one might expect them to do.

Secondly, how does the division of the soul affect the thesis that it is the fourth, nameless kind of element that makes it function as a unity? This thesis has to be rendered consistent with the partially independent workings of the animus and anima, and the obvious solution is that the effects of having the fourth element must be differentiated. Since it is associated with the soul's exceptionally fine structure, it is tempting to take it as located primarily in the animus; its most prominent activity is thinking, the activity most likely to require fine, rapid processes.

If the fourth element were located only in the animus, however, it would be the working of the animus that accounted for the unity of the whole soul. This Stoic kind of picture is arguably what Epicurus needs, and what he implicitly assumes much of the time. But it sorts ill with the relative independence of the anima, and the state of our sources makes the safest conclusion the disappointing one that Epicurus had not thought the point through. In fact Epicurus faces a difficulty over the unity of the soul. The whole soul is a functional unity; but the only part competent to unify it is not involved in all the soul's activities.

Does this matter? Epicurus is concerned to do justice to common sense; does that take for granted that our whole soul, rational and irrational, is strongly unified? One might think that common sense is actually inclined to deny the unity of the soul; the Stoic theory of the emotions, for example, is commonly taken to be highly counterintuitive, and more generally the Stoics might be taken to flout common sense in holding that information reaching the eyes and damage reaching the foot are registered in the hēgemonikon rather than in the eyes or in the foot. So perhaps in making animus and anima partially independent of each other Epicurus is deliberately answering to folk psychology.⁷⁵ This may well be true; certainly the Epicurean soul is much more weakly unified than the Stoic soul, and this may be due to a conscious desire to conform with common sense. However, we also find Epicurean claims that the soul is a unity, rather than two linked systems: Lucretius claims, for example, that rational and irrational soul together form "a single nature," and cannot be separated without mutual destruction.⁷⁶ The very fact that Lucretius is content to use anima to cover the whole soul suggests that he is not taking really seriously the partial independence of the irrational soul. And in his account of the soul-body relation, and in his arguments about death and its importance, Epicurus seems to be presupposing a unified soul and failing to take due account of the differences between its parts and the ways they function.

e) The Soul-Body Relation

There is a tension in Epicureanism over the soul-body relation. On the one hand the body is emphatically said to be the container or vessel of the soul. Epicurus uses such language repeatedly; Lucretius even bases his first argument for the soul's mortality on the comparison of the body with a vessel.⁷⁷ It is the body that holds together the soul and thus enables unified animate functioning—a reversal of the Aristotelian and Stoic view that what makes the agent alive and functioning is the soul's holding the body together.

Such language is, however, surprisingly inappropriate for the soul-body relation as Epicureanism actually develops that idea. Lucretius adds that the soul is the body's "guardian and cause of preservation; for they cling together like common roots and it is seen that they cannot be sundered without destruction." The soul is "in" the body like scent in perfume; it cannot be removed without destroying the substance. "So inter-woven are their elements between them from their first beginning; of they are endowed with a mutual life."⁷⁸ Diogenes of Oenonanda insists that the soul, which is "bound" by the body, "binds" it in turn.⁷⁹ Soul and body are two bodies which, in a living thing, are mutually dependent.

Epicurus presses the point for sentience: strictly speaking, it does not belong to the soul alone but is a joint product of soul and body. Lucretius puts this point more elegantly,⁸⁰ but here Epicurus' famously rugged Greek reveals an interesting conceptual struggle:

We should keep in mind that soul has the greatest share in causing (aitia) sensation (aisthēsis). However, it would not have had this if it had not been enclosed in a way by the rest of the assemblage. The rest of the assemblage, which provides it with this causality, itself has, derived from the soul, a share in just such a property—though not in everything the soul possesses. Hence when the soul departs it lacks sensation. For it did not itself possess this power in itself; something else connate with it provided it, and this, through the power brought about in connection with it depending on movement, at once achieved for itself a property of sentience and supplied it to the other also, depending on juxtaposition and mutual sensitivity, as I said. Therefore while the soul is indwelling it never lacks sensation through the removal of any other part—whatever of it perishes along with the breaking up of the enclosure, in whole or in part, if it remains, it will have sensation. The rest of the assemblage, whether it survives in whole or in part, will not have sensation when it is gone—that is, whatever quantity of atoms is needed to hold together to constitute the soul's nature. Further, when the whole assemblage is broken up the soul is scattered and no longer has the same powers, or moves; so it does not possess sensation either, for we cannot think of it as sentient unless in this composite and using these movements, when the enclosing and surrounding parts are not such as these in which [the soul] now is and has these movements. (Ep. Herod. 64-66)

The point which Epicurus has such trouble getting across is not that the soul requires the body for sensation, nor that sensation is the product of soul and body interacting, nor even that this is necessarily so. All these claims are quite compatible with dualism. Rather, Epicurean soul and body need each other to exist and to function as soul and as body. Without the body, the soul no longer exists or functions as soul, but is just scattered atoms; without the soul, the body no longer exists or functions as a body, but is a mere corpse. Sentience brings this out: it is the product of the mutually dependent soul and body, for the soul needs the body to exist as the soul of a sentient agent, and the body needs the soul to exist as the body of a sentient agent.

Why does Epicurus have such a struggle to express this? The problems are due largely to his clinging to the inappropriate conception of body and soul as vessel and contents.⁸¹ Epicurus often states a thesis in unnecessarily and sometimes misleadingly polemical and crude form; when we examine the thesis we find the crude formulations fail to do it justice. We can only put this down to an imperfect fit between Epicurus' philosophical activity and his pedagogical approach. The latter sometimes requires shock tactics to shake people out of their set views and prejudices. If and when they get involved in studying Epicurean philosophy, they may find that the initially controversial appearance was misleading; but by that time it will probably no longer matter, at least to the convinced Epicurean. Sometimes, however, Epicurus' cruder statements turn out to make trouble for his more sophisticated thoughts.⁸²

What brings out the closeness of the soul-body relation is sentience, which characterizes the irrational soul. We find elsewhere, however, that the Epicureans tend to contrast soul and body, and that when they do they have a different contrast in mind, namely, that between the body plus the irrational soul on the one hand and the rational soul on the other. "The pains of the soul," for example, "are worse than those of the body; for the flesh suffers only for the present moment, but the soul for past, present, and future. Similarly, the pleasures of the soul are greater."⁸³ Here "the body" clearly refers to the sentient body, closely linked to the irrational soul, and "the soul" clearly refers to the rational soul.

Further, many themes in Epicurean ethics stress not only this distinction, but the superiority of the soul, which by drawing on past, present, and future experiences can more than counterbalance what happens to the body. The star example here is Epicurus' dying letter to his friends, where he says that his present agonizing pains are more than counterbalanced by the joy in his soul from memories of philosophical activity.⁸⁴ Epicureans from Polystratus to Lucretius tirelessly urge on us that only the rational activity of philosophy will make us happy, for we need the exercise of the rational soul in order to organize our lives and make sense of the products of the irrational soul.

There is potentially a tension here. For Epicurus it is crucial that I think of my soul as something dependent for its existence and functioning on the existence and functioning of my body. He has shown this for the irrational soul, the source of sentience. But, given the stress on the importance of our identifying with the rational soul, and the contrast between the rational soul on the one hand and the body with the irrational soul on the other, the question is bound to arise whether Epicurus has adequately shown that the soul as a whole is indissolubly linked with the body. It could be objected, of course, that all he needs to show is that the sentient, irrational soul is indissolubly linked to the body, and the rational soul in turn indissolubly linked to the irrational soul; if the soul's unity is weak anyway, we would not expect an argument to show directly that the rational soul was linked indissolubly to the workings of the body. But, while that is arguably what Epicurus needs, we do not find explicitly either any acknowledgment that this is what is to be shown or any arguments to show it.

f) Survival

A famous and fundamental Epicurean teaching is that "death is nothing to us; for what is broken up has no sensation, and what has no sensation is nothing to us."⁸⁵ At greater length:

Get used to the idea that death is nothing to us, since every good and evil lies in sensation, and death is the deprivation of sensation…. So death, the most fearful of all evils, is nothing to us, since when we are, death is not present, and when death is present, then we are not. It is therefore nothing to the living, nor to the dead; for the former it is not, and the latter are no longer. (Epicurus Ep. Men. 124-25)

Lucretius puts this point forcefully: what happens after I am dead will be of no concern to me, since there will be no me, just as the Punic Wars were of no concern to me when they happened, since there was then no me to be concerned. "And even if the nature of the rational soul and the power of the irrational soul go on having sensation after being torn from our body, still it is nothing to us, who are made into one united compound by the mating and marriage of body and soul."⁸⁶ It is possible, he adds, that in the past my soul and body atoms came together in just the way they do now; but any such union was not me. I could not be around before the conception which brought me into being as an ensouled body, and in the same way I cannot be around after the death that breaks up the mutually dependent functioning of soul and body. So what happens after my death is like what happened before my birth—nothing to me.

The argument has raised controversy, ancient and modern. The important point here is the need for the premise that all good and evil lie in sensation. For sensation is, of course, characteristic of the irrational soul; and we have seen that in sentience the body and irrational soul are indeed mutually dependent. But the claim that for us all good and evil lie in sentience seems to neglect the role of the rational soul. This comes out in at least two ways. It is because of the activity of the rational soul that we are able to identify our good with projects whose content goes beyond our own personal pleasures, and which may be fulfilled only after our death. Epicurus himself stresses the value of friendship, and concern for friends and their activities for their own sake. But this will involve an agent in perfectly rational concern for projects and activities whose fruition does not depend on her being alive. It is hard to see how death is nothing to such a person just because she knows she will not be aware of these projects: her concern for the projects did not depend on her being aware of them.

Secondly, because the rational soul can, as Epicurus puts it, compare past, present, and future, it is what gives an agent a sense of himself as an agent continuing through time, a being with a whole life. And this means that though death is nothing to me when it arrives, since it removes the agent in question, it is not necessarily irrational to worry about its happening in the future. Both Epicurus and Lucretius deny this: since death will not concern me when it comes, it is irrational for me to worry about it now.⁸⁷ But why are they entitled to this? That death is not an evil when it comes does not imply that it is not an evil in someone's life as a whole (by coming sooner rather than later, for example).

It might be urged that these objections come from unfairly pressing Epicurus' language of good and evil lying in sensation. Surely he did not mean to limit sensation in this connection to the activity of the irrational soul. Is he not more fairly understood as claiming that nothing is a good or evil for an agent unless that agent can experience it—where "experience" is taken to refer to the activity of the whole soul, rational and irrational? It may be that something like this is what Epicurus did mean. But however generously we interpret "sensation" here, we shall not get out of the problem. Death is not an evil at the time it occurs, but this does not show that it is not an evil in one's life as a whole. But the rational soul is what gives the Epicurean a notion of her life as a whole. Nor does it show that death is not an evil in frustrating concerns that go beyond one's life and do not depend on one's experiencing the results. But the rational soul is what gives the Epicurean her concern for projects and activities that go beyond her life and matter whether she experiences the results or not—for example, the concerns and activities of friends.

Epicurus has an answer to these objections. They all involve in some way the claim that death, while it may not be an evil when it occurs, is nonetheless an evil by depriving us of goods which we would otherwise have; for thanks to our rational soul, we have a conception of our lives as wholes, and of projects that go beyond the reach of our own sentience; yet it is just this which enables us to commit ourselves to there being goods which, so it seems, death can deprive us of. Epicurus can say that only a non-Epicurean will be concerned by this, because she has a faulty conception of what these goods are. An Epicurean will realize that our highest good is pleasure and that all the goods that we can reasonably recognize in our lives are means to, or ways of, achieving this pleasure. We even seek friendship, and goals that extend beyond our own lives, for the sake of pleasure. Epicurus' own theory of what this pleasure, rightly conceived, actually is, is complicated, and the evidence difficult, but some things are clear. It is not to be identified with good feeling: it is a condition of "untroubledness" or ataraxia, which one achieves by following only natural desires and avoiding courses of action which will predictably lead to worry and trouble.

An important aspect of this is that the Epicurean will have achieved equanimity about goods that can be lost; for what she is after, in seeking ataraxia, is not the external results of action, but the inner result, the pleasure that lies in having the right attitude toward things that make other people upset. The pleasure that is our goal of life is radically internalized. One remarkable result of this is the thesis that pleasure is not increased by duration: once you have achieved Epicurean happiness, you have all that you need for happiness, and further time spent doing actions can merely vary what you have, not add to it.⁸⁸ Hence death does not deprive the good Epicurean of goods after all.⁸⁹

The limitation of this response is clear: it works only for the committed Epicurean, who already accepts Epicurean ethics in full. It will not convince the non-Epicurean, who has a different idea of what it is rational to consider good. And it seems a weakness that a thesis about the soul should depend so directly on a very controversial ethical thesis. Unless one is an Epicurean on other grounds, therefore, the death arguments contain a gap, one that makes more obvious Epicurus' lack of an explicit discussion of the nature of the soul's unity, and the relations of the rational and irrational parts of the soul.

Epicurus' argument does not depend on the premise that my soul is mortal; as Lucretius makes clear, it would hold even if my soul did survive the breakup of its union with the body—and even if it were immortal, since death is nothing to me even if there will be a qualitatively identical Doppelgänger constituted of the very same atoms that constitute me. In principle, the Epicurean soul could be immortal without endangering the survival arguments.

It is therefore surprising that Lucretius prefaces his great declaration that death is nothing to us with nearly thirty Lucretius arguments may be to mortal.⁹⁰ Lucretius may be confused.⁹¹ But possibly these arguments are meant to play an important subsidiary role; for the belief that my soul is immortal, while not as crucial as the belief that death is a bad thing for me, does play an important role in the various beliefs that make up fear of death and a negative attitude toward dying. Lucretius' flood of arguments is best understood as a sustained attempt to remove mistakes and to enable us to have a correct prolēpsis or conception of the soul. Their unsophisticated nature is deliberate—Lucretius hammers home simple and undeniable facts, such as that disease affects us psychologically as well as physiologically, and brings them into direct conflict with the notion of the soul as immortal. In Epicureanism it is important that we start from a right conception of what we are investigating;⁹² in the case of the soul this involves the removal of confusions, and an effective way to do this is to appeal repeatedly to our basic intuitions. For Epicurus our awareness of something clear and concrete in our lives is not likely to be corrupted by bad theory. Diogenes of Oenoanda ridicules belief in survival after death in the same downto-earth and unsophisticated way.⁹³

For Epicurus both the belief that my soul is immortal and the belief that death is a bad thing for me are not just false but unhealthy, pathological. As long as we hold them, we will not be happy, for they deeply corrupt our conception of what we are and how we should live. We will be tempted, for example, to locate the significance of our lives in something supposedly waiting for us after death.⁹⁴ Lucretius devotes passionate rhetoric to showing us that this is perverse; the desire to survive death is based on failure to face reality. Those who regret dying are merely rebuked for clinging to immature fantasies.⁹⁵

Notes

¹ Epicurus is clearly influenced in his physics by Aristotle's criticisms of atomism, and we know from a papyrus fragment that he read the Physics and Analytics (Philodemus Pap. Herc. 1005; Arrighetti 1973, 473). But claims that his philosophy of mind and particularly action are heavily influenced by Aristotle seem to me exaggerated (see Diano 1974a; Furley 1967; Englert 1988).

²Ep. Herod. 63.

³ For his argument for the soul's physicality see Ep. Herod. 67. For some texts on the basic principles of Epicurean physics, see Long and Sedley (1987, 4-13). Texts on the soul can be found on pp. 14 and 15. Some passages discussed in this section, on free agency, are on p. 20.

⁴ 3.161-76.

⁵ In part 2, chapter 2, section a.

⁶ Diogenes of Oenoanda frag. 32; Lucr. 2. 256-60….

⁷ And is so introduced by Lucretius (2. 216-93).

⁸Pyr. 3. 187.

⁹ That this is the number of this book has been argued by Laursen (1987). See also Diano (1946); Sedley (1974, 1983, 1989). The text can be found in Arrighetti [34]; there are sections with translation in Sedley (1983); in Long and Sedley (1987, 2:20 Β and C; 1:20 j); and in Laursen (1988). I have had the benefit of seeing Laursen's new readings for much of this book; I am very grateful. A new edition by Laursen of the entire text (which exists in fragments from three papyri) is forthcoming.

¹⁰ Democritus is not named explicitly, but there are Epicurean precedents for seeing him referred to here as "the great man"; see Sedley (1983). For Epicurus' argument see Arrighetti [34.30]; Sedley (1983, 20, 29-30 with n. 28); Long and Sedley (1987, 20 C [13]). This passage has been extensively discussed; see Laks (1981); Gigante (1981, 56-62).

¹¹ Arrighetti [34.30]; Sedley (1983, 24 n. 18); Long and Sedley (1987, 20 j); Laursen (1988, 17). In contrast to admonition (nouthetikos…tropos), we exonerate wild animals instead of admonishing them or trying to reform them, or indeed regarding ourselves as retaliating against them. We treat them not as agents whose developments are up to them, but "conflate their developments (apogegennēmena) and their makeup (sustasis) alike into a single thing." Elsewhere in the book Epicurus denies that we do this in the case of responsible agents.

¹²Cf. Vatican Sentence (hereafter VS) 40: "The person who says that everything comes about according to necessity cannot criticize the person who denies it—for he says that this too comes about according to necessity."

¹³ For this kind of argument see Burnyeat (1976).

¹⁴ Arrighetti [34.(28)19-(30)7]; Long and Sedley (1987, 20 C [8][12]).

¹⁵ There is also a specifically Epicurean argument; see Arrighetti [34.28-30]; Sedley (1983, 20; cf. 27-28); Long and Sedley (1987, 20 C [8]). If the reductivist claims that talk of "necessity" does not conflict with, but rather has proper application to, what we do "through ourselves," by our own agency, then he is merely changing the word; our "conception" (prolēpsis) of our own agency precisely contrasts with being necessitated. This argument also has modern analogues.

¹⁶ As Sedley (1983, 1989) recognizes, though he conflates reductivism with eliminativism and takes the argument to be stronger than it is, claiming that it shows that Epicurus was not only not a determinist but not a physicalist either. Given our total evidence about Epicurus, it is impossible that a breach in his physicalism should have gone unnoticed in the ancient world (see the end of this section), and the argument does not even show that Epicurus is not a determinist; it shows only that if he is (as he seems to be) he must be a compatibilist.

¹⁷ As he does at Ep. Men. 133.

¹⁸ Epicurus does not use epi with the dative with the meaning "up to the person," an idiom common among other philosophers. He prefers para, which has the force of "depending"; there is a parallel in the fourth sceptical Mode (see Annas and Barnes 1985, chap. 7).

¹⁹ 3. 307-22.

²⁰ The Greek is syntactically ambiguous, and Sedley translates with actions, thoughts, and so on being what the seeds direct us toward, not what they are seeds of.

²¹ On the argument of these difficult texts see Sedley (1983, 1989); Laursen (1988); Annas (1991 and forthcoming).

²² Sedley (1983, 1989) claims that the self is distinct from all atoms, and that Epicurus is thus not a physicalist (see above n. 16). For arguments against this as a reading of this text see Annas (forthcoming).

²³ The last clause goes beyond anything explicitly in the papyrus, but the opening and concluding fragments make it fairly clear that the book was concerned with ethically right development….

²⁴ Long and Sedley (1987) take the subject here to be zōia, that is, the agents or selves themselves. Laursen (1988) argues that the subjects must be the developments themselves.

²⁵ The Greek is syntactically ambiguous; Long and Sedley take "through themselves" with "do not become" rather than with "becoming productive."

²⁶ There is a participle here, but the verb is uncertain. Long and Sedley read "hating them" (misountes).

²⁷ Long and Sedley (1987) read "distinctness from the atoms." See Sedley (1983). For the present reading see Laursen (1988, 12-13).

²⁸ See Sedley (1989, n. 45); on "differential" see Laursen (1988, 13-14).

²⁹ See Laursen (1988, 14-15) for difficulties in identifying the subject here. Long and Sedley (1987) translate "he," importing a hitherto unmarked subject.

³⁰ See Sedley (1983, 1989) and the commentary on Long and Sedley (1987, 20 B, C, and j).

³¹ The atoms of the constitution, that is….

³² As Lucretius argues at length (2. 865-990).

³³ Lucr. 4. 823-57.

³⁴ Philodemus On Signs 25. 3-4.

³⁵ 2. 263-71. Huby (1969) finds this problematic and contrasts the On Nature passage about wild and tame animals. But the problem is greatly lessened once we realize that libera voluntas is not "free will" but the capacity for free action….

³⁶ Arrighetti [34.25]; Sedley (1983, 24 n. 18); Long and Sedley (1987, 20 j); Laursen (1988, 17).

³⁷ Horses: 2. 265, 268 (mens), 270 (animus). Deer have a mens at 3. 299.

³⁸ Frag. 34 Longo Auricchio (= Porph. Abst. 1. 7-12, 26, 4) 12. 5-6.

³⁹ Hence the rather bleak conclusion that the rise of human society is at the expense of animals, who are "expelled" from it, and to whom we owe no duties of justice.

⁴⁰ In the opening columns (1-8) of his On Irrational Contempt for Popular Opinions.

⁴¹Sunoran: they cannot "see them together." It is tempting, though speculative, to connect this with passages in On Nature, 25 that talk of thinking of oneself, and seem to be discussing the idea of holding together different experiences as experiences of the same self: Arrighetti [34.14-16].

⁴² Polystratus seems to be denying animals some kind of memory; possibly he would allow, with Hermarchus, that they have "irrational" memory.

⁴³ Polystr. 7. 6-8.

⁴⁴ Epicureans are often willing to see specifically human capacities as more developed forms of what we can see in other animals. They hold this for human language (Lucr. 5. 1056-90), sexual desire (4. 1192-1207), and dreams (4. 986-1010).

⁴⁵ We find the content of this and associated Doxai expanded by Hermarchus, Epicurus' successor, in the work paraphrased by Porphyry in De abstinentia 1. 7-12. Cf. Clay (1983).

⁴⁶ Schol. in Ep. Herod. 66 (= Usener 311); cf. Lucr. 3. 177-230.

⁴⁷ Diogenes of Oenoanda frag. 37, col. 1; trans. Chilton, with slight alterations.

⁴⁸ Kerferd (1971) ingeniously avoids the problem by denying that the relevant sentence of Epicurus refers to the composition of the soul at all.

⁴⁹ See Plut. Adv. Col. 1118d-e; Aët. 4. 3, 11, p. 388 Diels (hereafter D) (= Usener 315); Lucr. 3. 231-322. Sharpies (1980) argues that Lucretius is talking about ordinary fire, air, and wind, not atoms that are firelike and so on. But our other sources give the more cautious view. Plutarch has ek tinos thermou kai pneumatikou; Aëtius has ek poiou purōdous, ek poiou aērōdous, ek poiou pneumatikou.

⁵⁰ Lucr. 3. 232-33.

⁵¹ Lucretius translates it as aura, "breeze," or ventus, "wind." In Diogenes of Oenoanda new frag. 82 pneuma is ordinary wind, which is "cold and high" when there is hail.

⁵² Lucr. 3. 288-322.

⁵³ Epicurus Ep. Herod. 63.

⁵⁴Adv. Col. 1118e.

⁵⁵ Aët. 4. 3, 11, p. 388D (= Usener 315); cf. Lucr. 3. 237-42.

⁵⁶ Cf. Bailey (1928, 392), who sees a "thin disguise for the abandonment of the materialist position."

⁵⁷ This rules out theories that treat the fourth element alone as responsible for the soul's activity or that of its rational part (for a survey of theories on these lines, see Kerferd 1971, 84-87). One persistent version of this point is that the fourth nature is a transformation of Aristotle's "fifth element"; there is no reason to think this, and it is equally misguided to think of either as "wholly spiritual and non-material" (Bailey 1928, 392).

⁵⁸ 3. 258-87.

⁵⁹ 3. 285-87.

⁶⁰ As Diano (1974a) sees.

⁶¹ 3. 232-32.

⁶² Trans. Chilton; cf. new frags. 20, 94.

⁶³ School, in Ep. Herod. 66 (= Usener 311); Diogenes of Oenoanda frag. 37, col. 1.

⁶⁴ Schol. in Ep. Herod. 66 (= Usener 311); Lucr. 3. 95, 140-42.

⁶⁵ Lucr. 3. 359-69. Cicero refers to the "window" theory at Tusc. 1. 46.

⁶⁶ 3. 421-24.

⁶⁷ Schol. in Ep. Herod. 66 (= Usener 311); Lucr. 3. 140-42.

⁶⁸ Pap. 1012, cols. 29-30, pp. 38-39 de Falco (= Usener 313). See Croenert (1906, 117). De Falco thinks that some followers of Herophilus and Erasistratus may be meant, such as Demetrius of Apamea. Croenert identifies the doctors only as Empiricists.

⁶⁹ See part 2, chapter 2, section f.

⁷⁰ Diogenes of Oenoanda frag. 37, col. 1; Lucr. 3. 94-97.

⁷¹ Lucr. 3. 147-60.

⁷² Lucr. 3. 396-416.

⁷³ See section b of this chapter.

⁷⁴ Compare the distinction between wild and tame animals (in section a of this chapter).

⁷⁵ On this issue I am indebted to comments by Rob Cummins.

⁷⁶ 3. 136-37.

⁷⁷ Epicurus Ep. Herod. 63-66 contains three uses of forms of to stegazon for the body, and one of ta periechonta. Cf. 65: "When the whole assemblage is broken up the soul scatters." Lucr. 3. 425-44 is the passage in question; cf. 555. Cf. also Usener 337, where the soul is said to be in the body like wind (pneuma) in a wineskin (and thus to scatter at death).

⁷⁸ Lucr. 3. 323-32; cf. 337-49: "A body is never born by itself."

⁷⁹ Frag. 37, col. 1.

⁸⁰ 3. 331-36, 350-58.

⁸¹ Diano (1974, 146ff.) suggests that this may be an inheritance from Democritus, who calls the body a skēnos (frags. A152, B37, B223 DK).

⁸² This is particularly the case with his account of pleasure, where his crude and shocking slogans are quite misleading.

⁸³ D. L. 10. 137.

⁸⁴ D. L. 10. 22.

⁸⁵Kuria doxai 2.

⁸⁶ Lucr. 3. 830-69, esp. 843-46.

⁸⁷ Epicurus Ep. Men. 124-25; Lucr. 3. 870-977.

⁸⁸ This idea, that one's happiness is "complete," embracing everything worth having, in a way that takes no account of the natural contours of a human life, is highly controversial. See Nagel (1979); Furley (1986); Mitsis (1988a, 1988b); Striker (1988); Rosenbaum (1990).

⁸⁹ Of course this would still leave imperfect Epicureans rationally wanting to live longer, so as to get nearer the goal of ataraxia. It is only the fully wise Epicurean who has no more to gain from another forty years than from another four minutes.

⁹⁰ 3. 425-829.

⁹¹ He says that death is nothing to us because the (rational) soul's nature is mortal, and this just misstates the argument.

⁹² See Asmis (1984, pt. 1).

⁹³ New frag. 2; frags. 34, 35.

⁹⁴ See Konstan (1973) for the claim that when Lucretius says that the terrors of hell are in our own lives, he means that hell is a projection onto the supposed afterlife of false beliefs about this life.

⁹⁵ Cf. the end of Philodemus On Death. See Gigante (1969a).

Works Cited

Annas, J., and Barnes, J. [1986] The Modes of Scepticism, Cambridge.

Annas, J. [1991] "Epicurean philosophy of mind." In Ancient Philosophy of Mind, ed. S. Everson, Cambridge.

Arrighetti, G. [1973] Epicuro: Opere, Turin.

Asmis, E. [1984] Epicurus' Scientific Method. Ithaca, NY.

Bailey, Cyril. [1928] The Greek Atomists and Epicurus, Oxford.

Burnyeat, M. [1976] "Protagoras and self-refutation in later Greek philosophy." Philosophical Review 75: 44-69.

Clay, D. [1983] Lucretius and Epicurus, Ithaca, NY.

Diano, C. [1946] Epicuri ethica, Florence.

Diano, C. [1974] Scritti epicurei, Florence.

Englert, W. [1988] Epicurus on the Swerve and Free Action, Atlanta.

Furley, D. J. [1967] Two Studies in the Greek Atomists, Princeton.

——. [1986] "Nothing to Us?" In The Norms of Nature, ed. M. Schofield and G. Striker, pp. 75-92, Cambridge and Paris.

Huby, P. [1969] "The Epicureans, animals, and free will." Apeiron 3: 17-19.

Gigante, M. [1969] Ricerche Filodemee, Naples.

——. [1981] Scetticismo e epicureismo, Naples.

Kerferd, G. [1971] "Epicurus' doctrine of the soul." Phronesis 116: 80-96.

Konstan, D. [1973] Some Aspects of Epicurean Psychology, Leiden.

Laks, A. [1981] "Un legerete de Democrite." Cronache Ercolanesi 11: 19-23.

Laursen, S. [1988] "Epicurus On Nature Book XXV." Cronache Ercolanesi 18: 7-18.

Long, A. A., and Sedley, D. N. [1987] The Hellenistic Philosophers, Cambridge.

Mitsis, P. [1988a] Epicuruss Ethical Theory, Ithaca, Ν.Υ.

——. [1988b] "Epicurus on death and the duration of life." Proceedings of the Boston Area Colloquium in Ancient Philosophy, vol. 4, ed. J. J. Cleary, pp. 295-314.

Nagel, T. [1979] Mortal Questions, Cambridge.

Rosenbaum, S. [1990] "Epicurus on pleasure and the complete life." The Monist 73, no. 1 (January): 21-41.

Sedley, D. [1974] "The Structure of Epicurus' On Nature." Cronache Ercolanesi 4: 89-92.

——. [1983] "Epicurus' refutation of determinism." SUZETESISi 11-51, Naples.

——. [1989] "Epicurean anti-reductionism." In Matter and Metaphysics, ed. J. Barnes and M. Mignucci, pp. 295-327, Naples.

Sharpies, R. [1980] "Lucretius' account of the composition of the soul." Liverpool Classical Monthly 5: 11-Y7-20.

Striker, G. [1988] "Commentary on Mitsis." Proceedings of the Boston Area Colloquium in Philosophy, vol. 4, ed. J. J. Cleary, 315-20.