SOURCE: Malcolm Schofield, "Heraclitus' Theory of Soul and Its Antecedents," in Psychology, edited by Stephen Everson, Cambridge University Press, 1991, pp. 13-34.

[In the essay below, Scholfield explores Heraclitus's conception of the soul and psychology, concluding that the philosopher held the soul to be, like the universe itself, "a physical substance subject to the unity of opposites and to opposite sequences of transformations."]

I

For souls it is death to become water, for water it is death to become earth; from earth water comesto-be, and from water, soul.

A dry soul is wisest and best.

A man when he is drunk is led by an unfledged boy, stumbling and not knowing where he goes, having his soul moist.

You would not find out the boundaries of soul, even by travelling along every path: so deep a measure does it have.

(KRS 229-32 (=nos. of Frr. in [61]), frr. 36, 118, 117, 45)

We have now given a general answer to the question, What is soul? It is substance in the sense which corresponds to the account of a thing. That means that it is what it is to be for a body of the character just assigned. Suppose that a tool, e.g. an axe, were a natural body, then being an axe would have been its essence, and so its soul; if this disappeared from it, it would have ceased to be an axe, except in name. As it is, it is an axe; for it is not of a body of that sort that what it is to be, i.e. its account, is a soul, but of a natural body of a particular kind, viz. one having in itself the power of setting itself in movement and arresting itself. Next, apply this doctrine in the case of the parts of the living body. Suppose that the eye were an animal—sight would have been its soul, for sight is the substance of the eye which corresponds to the account, the eye being merely the matter of seeing; when seeing is removed the eye is no longer an eye, except in name—no more than the eye of a statue or of a painted figure. We must now extend our consideration from the parts to the whole living body; for what the part is to the part, that the whole faculty of sense is to the whole sensitive body as such. (Aristotle, de Anima II, 1.412b 10-25 (Revised Oxford Translation))

Four aphorisms, one paragraph (or what one might reasonably deem to be a paragraph) of technical prose: the aphorisms selected by G. S. Kirk to begin his account of Heraclitus' psychology,¹ the paragraph composed by Aristotle to sum up and clarify his general theory of what soul is.

We know that Aristotle's text is part of the statement of a theory because we have the theoretical treatise from which it is an extract. But its metaphysical vocabulary and argumentative style, its preoccupation with abstract definition and conceptual distinctions, in any case betray its theoretical cast. Did Heraclitus' aphorisms also originally help to articulate or at least intimate a psychological theory?

Few would hold that the explicit articulation of a theory can have been Heraclitus' object. Although these and similar aphorisms, complete in themselves, are only fragments of a whole we no longer possess, they simply do not look much like bits of theory. Frr. 117 and 45 hint at arguments, even if they contain no logical connectives; fr. 36 seems to state a general law or pattern; fr. 45 is concerned with the logos, 'account', of the soul no less than is the Aristotle passage, and the other aphorisms implicitly address the question of its nature. But they remain aphorisms, sayings designed to puzzle and provoke and in some measure enlighten; it is hard to think that they were ever woven into a discursive sequence which would aptly be described as 'setting out' or 'arguing' Heraclitus' 'view' of the soul. A further aphorism is generally interpreted as a comment by Heraclitus on the non-declarative functioning of language in general and his own use of it in particular (KRS 244, fr. 93):

The lord whose oracle is in Delphi neither speaks out nor conceals, but gives a sign.

Virtually all interpreters, however, see Heraclitus' sayings as designed to elicit from the reader or hearer an understanding of the logos, the true account of things expressed in his own logos or discourse. And where the logos of the physical universe or of the soul is concerned, most believe that our task is to decipher Heraclitus' Delphic code and uncover what can properly be called his cosmological and psychological theories. Thus among recent authors Jonathan Barnes refers to his 'idiosyncratic theory of man and of the human soul',² and Charles Kahn writes as follows³:

I agree with Kirk on the need to reconstruct for Heraclitus a kind of 'identity theory' of body and mind, in which stages of physical change and states of moral psychology are not merely put in one-toone correspondence but are conceived as aspects of a single reality: wisdom and excellence simply are the dry condition of the psyche. Furthermore, we agree that this psychophysical theory must, as in the case of Empedocles, take account of different destinies after death for the noble and the base, the wise and the foolish. We differ, however, in regard to the physical constitution of the psyche in a living man, and hence in regard to the 'greater destiny' for a noble soul after death.

At any rate they agree that such a soul has a continuing existence, whereas Martha Nussbaum comes to the opposite conclusion⁴: 'Heraclitus' psuchē theory recognises death as necessary and denies posthumous survival.' None of these scholars explains the notion of theory being employed here. The Shorter Oxford English Dictionary (s.v. 3(a)) offers a definition which more or less serves: 'a scheme or system of ideas or statements held as an explanation or account of a group of facts or phenomena'—although if the posthumous fate of the soul is to fall within the scope of theory, we had better add: 'or as an answer to questions they prompt'.

I shall not argue that the logos which Heraclitus' aphorisms on soul reflect or express does not constitute a theory so defined. What needs attention which it has seldom received is the question what sort of theory it is.

II

Consider KRS 229, fr. 36:

For souls it is death to become water, for water it is death to become earth; from earth water comesto-be, and from water, soul.

In his major edition of the fragments of Heraclitus, Marcovich takes this saying to refer 'to the normal constant and regulated physiological processes within the living human organism, and not to the destiny of soul after the physiological death' (which is the usual interpretation).⁵ For the processes in the microcosm are presented as parallel to those in the macrocosm described in KRS 218, fr. 31:

Fire's turnings: first sea, and of sea the half is earth, the half 'burner' [i.e. lightning or fire] … is dispersed as sea, and is measured so as to form the same proportion as existed before it became earth.

And these large-scale physical changes, as would generally be agreed, are 'the normal, natural processes which are every day going on', sustaining the cosmic order.⁶ What physiological processes did Heraclitus have in mind? The constant transmutations of soul into blood and other bodily humours, and of these in turn into flesh, ligaments and bones, and vice versa (i.e. flesh etc. into blood etc. into soul). Marcovich concedes that such 'metaphorical' uses of'water' and 'earth' are strange, but argues that they are not surprising given that Heraclitus' aim was to assimilate the processes of the microcosm and the macrocosm as closely as possible. He thinks they might have been supposed by Heraclitus to be the more intelligible in the light of 'such folkloric commonplaces' as for example, Xenophanes' remark (KRS 182, fr. 37):

For we all came forth from earth and water

and the line from Homer (Iliad VII.99):

May all of you become water and earth.⁷

Marcovich's reading of fr. 36 seems to presuppose a policy of interpretation:

1. Formulate the proposition p said or implied in the fragment [in the case of fr. 36 Heraclitus represents the soul as subject to a pattern of physical changes parallel to those which sustain the cosmic order].

2. Find the hypothesis which most plausibly explains Heraclitus' saying or implying p [in fr. 36 his saying that the pattern specified in (1) is true of the soul].

3. Take the fragment to be asserting the hypothesis (2) directs us to find [in fr. 36 the system of physiological transmutations Marcovich hypothesises].

Success in implementing this interpretative policy will result in the attribution to Heraclitus of a hypothesis which is presumably then to be considered as part of his theory of soul.

To my mind it is quite incredible that fr. 36 was meant to communicate the physiological thesis Marcovich thinks it was designed to convey. (Homer and Xenophanes were no doubt talking in the same vein as people do nowadays when they say that human beings are just oxygen and hydrogen and carbon and …) Where has Marcovich gone wrong?

Not in the broad guidelines of his policy of interpretation, which might indeed be challenged, particularly perhaps on precept (3), but does not seem to be responsible for what is unbelievable about the hypothesis he extracts from fr. 36—which is unbelievable as an account of Heraclitus' meaning because of its specific content and indeed (as I shall suggest) precisely because its content is specific.

Marcovich's mistake lies in the direction he has looked in following out precept (2) of the policy. There are in fact two mistakes. First, he assumes that the parallel with fr. 31 is no true parallel unless fr. 36 specifies processes within the ordinary life of the microcosm—as though only so could it refer to 'normal, constant' processes. The supposition is false. Biological processes of birth and death happen to humans and other animals normally and constantly, but are not everyday occurrences within the microcosm. The second error is independent of the first, and more interesting because it involves an assumption which, if generalised, embodies a widespread understanding of Heraclitus' method. Marcovich has assumed that what lies behind Heraclitus' assertion of the pattern of changes indicated in fr. 36 must be primarily a specific idea about human physiology: as though he could not have made the assertion if he did not have some detailed biological evidence to support it.

The supposition is in a way an attractive one. For it suggests a scientific Heraclitus: the Marcovichian author of fr. 36 holds a proper view about the relation of evidence to general law. And if this kind of supposition is generally applied within the context of Marcovich's interpretative policy to the relevant fragments, it promises to yield a detailed picture of the soul as conceived by Heraclitus. Each of his Delphic pronouncements on the subject will be made intelligible by attributing to him specific beliefs which support them. Thus, for example, fr. 24, 'Gods and men honour those slain in battle', might be explained (as by Kirk:⁸ Marcovich himself does not go so far) by belief in the posthumous reunion of pure fiery souls with fire in the celestial regions. The resulting fabric of belief, rather than the aphorisms themselves, will be what constitutes Heraclitus' theory of the soul. 'Reconstruction' of Heraclitus' belief-system by this method, and so of his theory of the soul, is a project not peculiar to Marcovich, but one taken to be the appointed task of interpretation by many of the most distinguished and influential Heraclitus scholars of recent times….

But despite the attractions of Marcovich's supposition that Heraclitus would not have asserted what he says in fr. 36 except on the basis of some specific physiological evidence, there is strong reason to reject it. We are faced with the difficulty that, if it were true, it might be—probably would be—impossible for us to identify with any security what evidence Heraclitus would have been relying on; and that his original readers or hearers were not likely to have been in a significantly better position. This objection to the supposition is particularly awkward in view of Heraclitus' epistemology. He is insistent that his logos is not his private voice or idiosyncratic viewpoint, but something common; and he is critical of the many for living as though they had a private or special understanding when it is commonly accessible (KRS 194-6, frr.1,2,50):

Of the Logos which is as I describe it men always prove to be uncomprehending, both before they have heard it and when once they have heard it. For although all things happen according to this Logos men are like people of no experience, even when they experience such words and deeds as I explain, when I distinguish each thing according to its constitution and declare how it is; but the rest of men fail to notice what they do after they wake up just as they forget what they do when asleep.

Therefore it is necessary to follow the common; but although the Logos is common the many live as though they had a private understanding.

Listening not to me but to the Logos it is wise to agree that all things are one.

Nothing could be much more private or less commonly accessible than the belief about physiological transmutations which Marcovich finds in fr. 36.

KRS 194-6 do not stand isolated from Heraclitus' other sayings. Many of his paradoxes seek to shock us into recognising facts which are obvious enough when pointed out, but might never have struck us otherwise (KRS 199-201, frr.61,60,111):

Sea is the most pure and the most polluted water; for fishes it is drinkable and salutary, but for men it is undrinkable and deleterious.

The path up and down is one and the same.

Disease makes health pleasant and good, hunger satiety, weariness rest.

The message implicit in all this is that to understand the world we need to redescribe or find the meaning in what we know already; to grasp the initially unobvious structure in what is familiar; and to see the opposition in unity and unity of opposites which is its fundamen tal logos (KRS 203, 207, frr. 10,54):

Graspings: wholes and not wholes, convergent divergent, consonant dissonant, from all things one and from one thing all.

An unapparent connexion is stronger than an apparent one.

Moreover, when we turn from Heraclitus' epistemology to his cosmology, we find confronting us in the key texts precisely the kind of redescription or reordering of what is familiar we have been led to expect. From the fragmentary and not perlucid fr. 31 it was already apparent that Heraclitus interprets the meteorological and other physical interaction of earth, sea and fire (presumably the fire of the heavenly bodies) in terms of a sequence of changes from fire into non-fire and (presumably) back again. The interpretation is made yet more explicit in two further sayings (KRS 217, 219, frr. 30,90):

This world-order [the same of all] did none of gods or men make, but it always was and is and shall be: an everliving fire, kindling in measures and going out in measures.

All things are an equal exchange for fire and fire for all things, as goods are for gold and gold for goods.

Like everything else the kosmos is itself a dynamic unity of opposites: fire being quenched and rekindled; and so exchanged with what is not fire.

Precept (2) of Marcovich's policy of interpretation enjoins us to find a suitable hypothesis to explain a Heraclitean assertion such as fr. 36. Consideration of the core fragments on epistemology, the logos and the kosmos has indicated a constraint on suitability: the hypothesis must be not a private belief, special to the believer, but something common—something we would all be bound to accept once we really understood it. Can reflection on this constraint suggest a better way of trying to fulfil precept (2) with respect to fr. 36 than Marcovich's attribution to Heraclitus of a specific physiological thesis?

An obvious option to explore is the possibility that some much more general proposition might be suitable, as more likely to constitute something common or commonly accessible. In fact such a Heraclitean proposition is easily identified. As commentators on fr. 36 standardly point out, soul is allotted the place in the cycle of elemental transformations which is assigned to fire in the cosmic cycle of fr. 31 : as fire turns into sea which turns into earth, so soul becomes water which becomes earth. I propose that the main reason why Heraclitus asserts fr. 36 is that he believes in the general pattern of elemental change specified in fr. 31. In other words, fr. 36 is not to be explained (as Marcovich thinks) as a more general and abstract formulation of a specific biological thesis, but as itself a specific application of a more general thesis about physical processes.

One advantage to this proposal is that it enables us to give Heraclitus the credit of allowing the soul to be a difficult topic—which I take to be implied by his remark (fr. 45) that it has a 'deep logos (account)', i.e. (I suppose) one hard to fathom. For if what lies behind fr. 36 is mainly fr. 31, we may suppose Heraclitus to have approached the job of understanding the soul in a properly modest and agnostic frame of mind, tempered by one or two convictions: first, that it is reasonable to suppose that human nature will be subject to the fundamental logos of opposition in unity like everything else in the universe; second, that since man is part of the natural world, his soul will be subject to the same physical laws embodying the universal logos as obtain elsewhere in it. These ideas amount to the thought that the same general form of explanation will apply to soul as to other things.

Of course, Heraclitus still needs some clue as to how that general form of explanation is true of soul. It seems plausible that the vital hint is the one yielded by the drunkenness aphorism (fr. 117): the simultaneous loss of physical and mental control exhibited by the drunkard not only confirms that the soul is a material entity but also indicates that it is weakened—and could no doubt be destroyed—by wet substances. This points to fire or something analogous as constituting the nature of soul, for in the kosmos fire is the controlling element (cf. KRS 220, fr. 64: 'Thunderbolt steers all things'), and it is something quenched by water.

Once Heraclitus can rely on such a commonly accessible clue as is supplied by fr. 117, he—and we—know where to fit soul into the cycle of elemental changes. We can go on to assert fr. 36 even though we have no story to tell of how exactly soul by becoming water dies, or how soul comes to be from water. Nevertheless our reasonable conviction in the applicability to the special case of soul of the common logos and of the physical patterns which exemplify it entitles us to assert that these changes occur—must occur—although we quite lack the sort of specific physiological evidence for the assertion which Marcovich thought Heraclitus would have to have believed he had in order to be in a position to make it. In short, we can be clear about the general shape of the truth concerning the soul even if its logos is too deep for much else to be clear. Most interpreters venture some guesses about what Heraclitus is likely to have thought about the detailed how. If my analysis is correct, we need not suppose that he necessarily held any particular views on how. It is worth stressing that this position is not an unreasonable one to be attributing to Heraclitus. Our confidence in a scientific conjecture may often rightly outrun the evidence for its holding in particular instances supposed to fall within its explanatory scope (think of black holes). Perhaps it would have been better if Aristotle had not spoiled his brilliant insight that there must be some mechanism whereby the genetic code controls the development of the embryo with his false theory about how the process works.

So far as concerns fr. 36, then, psychological theory in Heraclitus is to be located not in any hypothesis about the detailed physiological realisation of the pattern of processes it describes, but in the bare fact that the common logos of things, or more precisely the version of that logos true of fire, applies to soul. As we shall see, this is a focus which is characteristic of Heraclitus' sayings on man, mortality and the soul. But before we pursue our investigation on these lines further, it is appropriate to reflect further on fr. 36 and 117 and particularly on the conception of soul they indicate.

III

A man when he is drunk is led by an unfledged boy, stumbling and not knowing where he goes, having his soul moist. (KRS 231, fr. 117)

Drunkenness deprives a man of the use of his active powers: he is led passively by a child whose powers, though still undeveloped, are yet adequate to the simple task in hand. The idea and its expression would be entirely commonplace, if wry, were it not for the final phrase: 'having his soul moist'. It is not clear just how unexpected a punch these words pack. Kahn seems to suggest that a fifth-century Greek reader might have seen in them 'a joke about the man who has drenched his soul with drink'⁹—and then the shock Heraclitus administers is the implicit directive to take this metaphorical drenching quite literally as a physical soaking. But I know of no appropriately early evidence for the Greek usage thus apparently postulated. What we do find, for instance in the fifth-century tragedians, is hugros (translated literally 'moist' above) employed, usually of the body or its parts, to mean 'supple' or 'loose', sometimes with the implication 'slack', 'languid'. So there may indeed be a typically Heraclitean play on language: if we attend to the metaphors of everyday or at least poetic speech (which is something 'common', as Heraclitus would put it), we shall see the unforseen literal physical truth disclosed by them. Even if hugros is not entirely surprising, however, psuchē, 'soul', surely is. Stumbling might generally be agreed to be a sign of slack limbs: ordinary Greek did not talk of 'soul' in connection with either stumbling or slackness, or indeed soaking. So its introduction as the penultimate word of the saying was no doubt calculated to create maximum shock.

Something of the force of the shock can be gauged by considering Homer's use of the word psuchē, important not only because Homer represents (if partially) the cultural ambience in which sixth- and fifth-century Ionian Greeks moved, but because it is clear from other fragments that one of Heraclitus' objects was to mount a critique of the authority of Homer (fr. 56; cf. fr. 42). The salient features of Homeric usage, pointed out many years ago in an influential article by Bruno Snell,¹⁰ have been conveniently summarised by Martha Nussbaum.¹¹ First is a tendency to mention psuchē only in 'negative contexts': its functioning is noticed only when it goes wrong.

When one's life breath leaves, it is called psuchē. But when a hero wishes to say 'as long as the breath of life remains in me [lit. 'my chest']', he says: Εἰ ς ὂ άϋτμή [aütm, 'breath'] / εν στήθ/ε/σσς μέή (Iliad 10.89-90). A hero may fight for psuchē (22.161) or risk his psuchē in battle (9.321) or discourse about the irrecoverability of the psuchē once it is lost (9.408). But he is never aware of doing anything by means of it in life; only once is it mentioned as being present in a living man at all. And the sole point of mentioning its presence there is to declare the man mortal and vulnerable.

Second is the absence of any sense of a central faculty, such as psuchē represents in, for example, Aristotle, connecting the disparate faculties of the living man.

In speaking of his faculties, the Homeric man distinguishes a number of 'Organs' with separate functions and locations: thúmos, kēr, ētor, phrenes, noos, etc. He does not refer explicitly to anything which connects them, or in virtue of which he is a single being.

Against the background of Homeric usage Heraclitus' choice of the word psuchē in fr. 117 is strategically radical. The role of psuchē as the controlling agent of the normal activity of the living man is stressed by the sequence of participles in the second part of the aphorism, which as Bollack and Wismann have pointed out demands a causal reading: the drunkard has no control over his body, because he has lost perception, because his soul is wet.¹² All might be well with both his cognitive and his motor capacities if it were dry.

It is not easy to judge how original Heraclitus was in his radicalism. There is a lot to be said for the view that, from the very first, philosophy conceived of psuchē as the central governing faculty it becomes in Heraclitus and in Aristotle. One of the most striking features of the thought of Heraclitus' Ionian predecessors Anaximander and Anaximenes is their determination to see the kosmos as a unified system, explicable in terms of a small number of basic physical (and in Anaximander's case numerological) concepts. Man and his psychology were not, so far as we can tell, a main preoccupation of the Milesians, but one would guess that to the extent that they did concern themselves with human nature the same kind of unifying explanatory strategy would have been employed. The evidence, such as it is, bears out this conjecture.

We begin with some texts relating to the founding father of philosophy, Thales, teacher of Anaximander (KRS 89-91):

It seems, from what they relate, that Thales too supposed the soul (psuchē) to be a sort of mover, given that he said that the magnet has soul because it moves iron.

(Aristotle, de Anima II.2, 405a19)

Aristotle and Hippias say that he gave a share of soul even to inanimate [lit. soulless] objects, taking his evidence from the magnetic stone and from amber.

(Diogenes Laertius 1.24)

And some say that it [soul] is intermingled in the universe, which is perhaps the reason why Thales also thought that all things are full of gods.

(Aristotle, de Anima 1.5,41 Ia7)

If we may believe the first two of these reports, psuchē already in Thales has a positive, not a Homerically negative, context, namely the explanation of movement. They suggest that he sees the capacity to move things as invariably a function of soul—hence his ascription of soul to stones, normally conceived of as inanimate. Another way of putting the point would be to say that magnets can cause movement because they are alive, for psuchē is evidently being treated here as the principle of life: from which we may infer that Thales was well on the way to conceiving of psuchē—again unhomerically—as the central essence which connects our various faculties. In animals the power of movement is intimately related to cognitive powers, to the extent that it is hard to imagine Thales not ascribing cognition also to psuchē, had the question occurred to him. Unfortunately we have no information as to whether he considered it or not.

The obscure KRS 91 may indicate some pantheistic role for psuchē. More important in a comparable vein is a famous text relating to Anaximenes (KRS 160):

Anaximenes son of Eurystratus, of Miletus, declared that air is the principle of existing things; for from it all things come to be and into it they are again dissolved. As our soul, he says, being air holds us together and controls us, so does wind [or breath] and air enclose the whole world. (Air and wind are synonymous here.) He, too, is in error in thinking that living creatures consist of simple and homogeneous air and wind. (Aetius 1.3.4)

Although the vocabulary employed is probably not Anaximenean throughout, this evidence confirms the Milesian association of human soul with coherence and control (presumably control of movement at least). And by now that urge to find unifying explanations takes the form we have already noticed in Heraclitus, of making soul the physical principle of the microcosm in such a way as to be the analogue of the basic physical principle of the macrocosm.

Was there, then, no important anti-Homeric innovation left for Heraclitus to make? I think there was. First, in Heraclitus the soul figures explicitly in epistemological and ethical aphorisms as much as in contexts to do with life and movement (e.g. KRS 198 and 240, frr. 107 and 85):

Evil witnesses are eyes and ears for men, if they have souls that do not understand their language.

It is hard to fight with anger; for what it wants it buys at the price of soul.

There is no reason to think that previous philosophy had spoken of soul in these kinds of context, although Ionian lyric poetry of the sixth century talks of psuchē as the seat of the emotions and as what enjoys emotional satisfactions. Heraclitus is in any case something of a pioneer in making human nature as important a subject of philosophy as the kosmos, so the suggestion of originality at this point chimes with our other evidence about him. Second, there is a philosophically important literary difference between Anaximenes and Heraclitus. Anaximenes made his comparison of air and soul in a book setting out his general theory of nature. We possess nothing of it, but it was presumably a primitive version of what became the sort of technical treatise perfected by Aristotle. It seems unlikely that it engaged directly or even indirectly with Homer or with ordinary language. Anaximenes will simply have invited his reader or hearer into his own conceptual world, trusting to the power of his explanatory scheme to persuade. Heraclitus, by contrast, brings the new language of physics and the new talk of the soul into ordinary discourse and everyday sentence rhythms. The jarring effect of psuchē at the end of the sentence which constitutes fr. 117 engineers a collision between philosophy and Homer, while KRS 198 and 240 begin to naturalise psuchē (philosophically conceived as the central faculty of understanding and control) within the common language. After all, it has to be part of Heraclitus' project to get us to accept his use of the word as (in his sense) common.

It may be that one of his fragments on soul is specifically designed to mock the Homeric conception of psuchē. This is the puzzling and amusing remark (fr. 98):

Souls sniff in Hades.

No doubt Heraclitus intends us to try out a variety of different hypotheses to make sense of this saying. But one attractive line of approach bids us note the unusual use of the definite article (hai psuchai), and suggests that its point is to refer to the souls Homer talks about, i.e. the shades of the underworld. Then if we exploit the common etymology of Hades (aidēs, aspirated) as the sightless place (aidēs, unaspirated), we have a joke: the shades have to sniff their way around Hades—since use of the eyes would be profitless. The remark so construed would presumably be meant to constitute a reductio ad absurdum of the Homeric notion of psuchē: if that's what it commits us to, we'd better abandon it.¹³

Fr. 36 is naturally read as a rather more straightforward challenge to the Homeric conception. 'For souls it is death to become water.' But the Homeric psuchē was precisely that part of us which survives death, or what replaces us when we die—a sort of pale self-substitute.

Nussbaum (who brings out well this anti-Homeric dimension of the saying) also makes a point which is not quite right. The expression thanatos psuchais, she says,¹⁴ 'is clearly a word-play, a juxtaposition of opposites: "It is death to the life-faculty … '" The Greek for that would be thanatos psuchēi (singular). Just as in Homer, where psuchai in the plural are shades or self-substitutes so called because the life-breath they are constituted of is all that is left of the persons they were, so here psuchai (plural count noun) must refer to the selves whose being depends on the central life faculty (psuchē, singular abstract noun): 'For selves it is death to become water.' Thus interpreted, Heraclitus' words must have a further target besides Homer and his view of the soul.

The Milesians were not the only early philosophers to introduce talk of psuchē in 'positive' contexts, giving it a central explanatory role relative to behaviour. Pythagoras followed suit and went further still. His name is and was from the first associated above all with the doctrine of reincarnation. There seems excellent reason to believe that he expressed this as a claim about psuchē; and what he must thereby have been claiming is that the self—the true or essential self—is condemned to a cycle of incarnations until its sins are expiated. The evidence is compelling although necessarily indirect, since Pythagoras wrote nothing. All our earliest witnesses—Xenophanes in the sixth century, Herodotus and Ion of Chios in the fifth—talk as though Pythagoras was notorious not only for teaching reincarnation but for propounding it as a teaching about psuchē, the soul or self (KRS 258, 260, 261):

Ion of Chios says about him [Pherecydes]: 'Thus did he excel in manhood and honour, and now that he is dead has a delightful existence for his soul—if indeed Pythagoras was truly wise, who above all others knew and learned the opinions of men.'

(Diogenes Laertius 1.120)

On the subject of reincarnation Xenophanes bears witness in an elegy which begins: 'Now I will turn to another tale and show the way.' What he says about Pythagoras runs thus: 'Once they say that he was passing by when a puppy was being whipped, and he took pity and said: "Stop, do not beat it; for it is the soul of a friend that I recognized when I heard it giving tongue."'

(Diogenes Laertius VIII.36)

Moreover, the Egyptians are the first to have maintained the doctrine that the soul of man is immortal, and that, when the body perishes, it enters into another animal that is being born at the time, and when it has been the complete round of the creatures of the dry land and of the sea and of the air it enters again into the body of a man at birth; and its cycle is completed in 3,000 years. There are some Greeks who have adopted this doctrine, some in former times, and some in later, as if it were their own invention; their names I know but refrain from writing down.

(Herodotus II.123)

It needs little argument that the self is what psuchē must mean in this context, for the point of the belief that my friend's soul inhabits this dog is precisely that my friend continues as just that—himself—in the dog. And it seems not unlikely that it was Pythagoras' teaching which encouraged the development of the use of psuchē to mean 'self', even on the lips of those who rejected or had no particular commitment to his doctrine.

Heraclitus thought Pythagoras enough of a nuisance and a danger to right-thinking to insult him in a couple of aphorisms (KRS 255 and 256, frr. 40,129):

The learning of many things does not teach understanding; if it did, it would have taught Hesiod and Pythagoras, and again Xenophanes and Hecataeus.

Pythagoras, son of Mnesarchus, practised enquiry beyond all other men and selecting these made them his own—wisdom, the learning of many things, artful knavery.

He could have been expected to dislike above all Pythagoras' cabbalistic attitude to truth and wisdom; but given that Pythagoras was chiefly famous for his doctrine about the soul, it would be hard to avoid concluding that the stark assertion of the death of the soul in fr. 36 is meant to contradict the basic tenet of Pythagoreanism—which makes our selves, as Herodotus puts it, immortal (essentially immortal). The inference is only reinforced when we notice that, like Pythagoras, Heraclitus here sets the soul within the pattern of a cycle of changes. His cycle is surely designed to challenge and displace the Pythagorean cycle. As we shall see in the next two sections, the question of human mortality perhaps engages Heraclitus more deeply than any other. How could he fail to have wanted to be perceived as at odds with the great charlatan Pythagoras on the issue which preoccupied each of them most?

IV

At one point in his commentary Kahn says¹⁵ : 'it is the phenomena of sleep and dreaming that may initiate us into these mysteries' (he is speaking of the way Heraclitus invites us to draw together cosmology and anthropology into 'a unified vision of life and death'). The observation is made à propos of a saying sometimes judged rather sourly as 'obscure'¹⁶ or 'one of the most puzzling',¹⁷ but to my mind a marvellous specimen of Heraclitus' philosophical method and more particularly of what Hussey calls his 'cryptic'¹⁸ or riddle style (KRS 233,fr.26):

[A] A man in the night kindles [haptetai] a light for himself when his vision is extinguished; [B] living he is in contact [haptetai] with the dead, when asleep, [C] and in contact [haptetai] with the sleeper, when awake.

As Hussey's term implies, the fragment needs deciphering. The process has three stages, representing successively deeper levels of meaning.

First is the superficial riddle-solving level. Fr. 26 is obviously a puzzle about dreams: obviously when you see it, as is the way with riddles. [A] employs the metaphor of lamp-lighting to describe the phenomenon: although his eyes are closed in sleep, the dreamer creates an effect as of illumination. [B] requires more input from the reader. I guess that Heraclitus is reminding us of how (at least in ancient Greek experience) the departed present themselves to us in dreams to advise or issue warnings. [C] might only be noting that there is physical or psychological continuity between the sleeping and the waking person, but it would be more pointed if making an observation about experience: the waking man is in contact with the sleeping inasmuch as he remembers his dreams. Then [B] and [C] together entail that man at his most alive—when awake—may be in touch with the dead, given the transitivity of contact. In short, the riddle has no arcane or private solution: to crack it we need only to read the lamp-lighting metaphor, and all else is fairly easily understood simply by reflection on Heraclitus' words in the light of common usage and experience.

Second is the level at which the universal logos is disclosed. Heraclitus' general pronouncements about the logos as well as the method of paradox prepare an expectation that, after the elementary decoding of level (1), a saying like fr. 26 which introduces oppositions (living versus dead, awake versus asleep) will be found to point to their unity. And so it does, via the word haptetai, translated 'kindle' in [A], 'is in contact with' in [B] and [C]. If we take haptetai now in a single sense, 'kindle', we obtain a more tightly unitary as well as a more pregnant meaning. In kindling the light of his dream [A], the living man kindles himself, although 'dead to the world' as we say nowadays [B], and the waking self kindles the sleeping self - for a dream is a sort of wakeful sleep [C]. What is described in [B] and [C] turns out to be just the same as what is described in [A]. And it constitutes a highly dramatic example of the unity of opposites. A dream is not merely a death-like state converted into a life-like one, sleep transformed into something approaching wakefulness. It is a state which is simultaneously life and death, sleep and waking, for it exhibits at one and the same time features characteristic of each. The verb haptetai therefore still carries at this level of interpretation the sense 'be in contact with' also, but now conveys thereby something stronger than at level (1).

Third is the level of physical truth. Study of fr. 36 and fr. 117 leads us to expect that the psychological truth discovered at level (2) is a function of an underlying physical truth. Heraclitus' vocabulary is designed to confirm the expectation. Compare the principal cosmological fragment (KRS 217, fr. 30):

This world-order did none of gods or men make, but it always was and is and shall be: an everliving fire, kindling in measures and going out in measures.

Here 'kindling' (haptomenon) and 'going out' (aposbennumenon) are the same verbs as are used in fr. 26 ('extinguished', aposbestheis). The repetition in fr. 26 must be deliberate. We are to infer that although the fire of the sleeper's psuchē is in a measure quenched, he rekindles it in a measure in his dream. Once again the metaphors we employ to talk of the soul and its activities ultimately demand a literal reading, and when given it yields the physical logos of psuchē.

It is a curiosity of Kahn's discussion of fr. 26 that, having drawn attention to the connection with fr. 30, he refuses in the end to exploit it:¹⁹ 'no physical doctrine is stated in Fr. 26; and there is no clear basis in the fragments for a stage-by-stage correlation of waking and sleeping with the elemental cycle of Fr. 36'. The refusal is the more striking given Kahn's willingness elsewhere to see intimations of a physical logos where other scholars find none, for instance in KRS 200, fr. 60, and in the doubtfully Heraclitean fr. 115:²⁰

The path up and down is one and the same.

The soul has a logos which increases itself.

Of the reasons he puts forward for this refusal, the first seems sheer petitio principii, while the second makes the Marcovichian mistake of supposing we would have to be looking for a detailed physiological story, when Heraclitus wants us as usual to concentrate on the big generalities of the logos. So Kahn's stance seems distinctly arbitrary. It is hard to resist the suspicion that what lies behind it is his conviction, contrary to what most readers have supposed, that Heraclitus takes psuchē to be not fire but (like Anaximenes) air.²¹ Yet the clearest positive evidence we have speaks for the usual interpretation (fr. 26 and fr. 36, where Heraclitus writes psuchē at the point where, on the analogy of fr. 31, we would have expected 'fire'). Kahn objects that air, wind or vapour may be either dry (cf. fr. 118) or moist (cf. fr. 117); whereas the contrast 'makes no sense if applied to fire or flame'.²² To this there is an easy reply. When soul becomes moist or wet its fire is put out—it dies (fr. 36) either entirely or in part, as in drunkenness, fr. 117, or in sleep, fr. 26. On the interpretation of psuchē as fire Heraclitus does not need to be seen as committed to the absurd concept of wet fire.

Kahn is in fact forced by the textual reading he accepts in fr. 118 to concede that soul in its finest state is conceived by Heraclitus in terms of light. My translation of his text follows Bollack and Wismann's:²³

Gleam of light: dry soul, wisest and best.

Light would naturally be taken to be a function of fire. But Kahn seems strangely reluctant to accept this, despite Heraclitus' palpable determination to make fire the key explanatory notion in his cosmic logos. There its chief exemplars are apparently the heavenly bodies which are the principal sources of light in the kosmos. At one point Kahn even suggests²⁴ that 'we think of celestial fire … simply as atmospheric air or exhalation become dry and luminous'. This puts the cart before the horse with a vengeance. Happily it does not represent Kahn's considered interpretation of the cosmology. His general position, surely correct, is that in cosmology Heraclitus is a radical anti-Anaximenean.²⁵ The decision to read him as following Anaximenes in his psychology, when the evidence falls far short of compelling the reading, is hard to understand.

V

We ordinarily think that life and death, sleeping and waking are discrete and mutually exclusive conditions: that is what their very opposition seems to imply. Fr. 26 begins the task of undermining the belief, by showing us that common usage and experience themselves, if probed a little, indicate contact between them and indeed simultaneity of a sort. Numerous further fragments directly or indirectly continue the process.

Here, for example, is a famous saying in the cryptic style (fr. 48):

The name of the bow is life (bios), but its work is death.

A trickier cryptic remark subverts an assumption fr. 26 itself may have induced (fr. 21):

Death is all we see when we awake: all we see asleep is sleep

Fr. 26 had encouraged us to think that the dreamer to a degree lights the flame of life, a life unequivocally associated with waking. But what (Heraclitus now asks) do we see about us when we wake from our sleep? Death. We begin to flounder—but also to sense that fr. 21 is a riddle about the nature of life, which fr. 26 had allowed us to treat as unproblematical.

If death is what we see awake, are we to infer that life is what we see in our sleep? Our unspoken question meets a brusque response from Heraclitus: that would be absurd—sleep is just sleep, no matter how good our dreams are.

So the fragment seems to rule out two options on what life is, and in so doing sets us a puzzle. Seems to rule out: for I take the solution to be that life, also, is what we see when we awake. The idea that fr. 21 excludes this obvious solution would be valid only if death and life are mutually exclusive. If we drop the assumption that they are, the way lies open to the inference that in fact they are the same.

The thought that life and death are the same is highly paradoxical. It is explained in the following aphorism, not intended as cryptic (KRS 202, fr. 88):

The same things in and among us are living and dead, what is awake and what sleeps, young and old: for these changed around are those, and those changed around once more are these.

Heraclitus here puts his argumentative cards on the table with an explicitness and clarity seldom matched elsewhere in the fragments. The ground he offers for his identity claim is the reciprocal change of life and death and other opposites into each other (cf. fr. 57, on the unity of night and day); and the repetition of the language—'these changed are those', 'those changed are these'—reinforces the point that it is the pattern of transformation that makes them the same.

So sameness is not Leibnizian identity. These opposites count as the same because they share the same form of change. Of each member of each pair one and the same thing is true: that it changes into the other. To a modern—or indeed to Aristotle—this might seem a weak brand of identity for Heraclitus to be defending. But that would be to misunderstand the importance of process in Heraclitus' scheme of things—nowhere more memorably evidenced than in KRS 214, fr. 12:

Upon those that step into the same rivers different and different waters flow.

Identity is identity of process. Hence, as many readers have sensed, his decision to describe the kosmos as a fire, something continually subject to the processes of kindling and quenching.²⁶

What remains puzzling in fr. 88 is not the notion of sameness involved but the idea that the living come from the dead and the young from the old as well as vice versa. I do not think Heraclitus can have meant anything very private or counter-intuitive by this. The structure of the saying indicates where the surprise is meant to come: in the first clause and its claim that the opposites are the same, the great central Heraclitean message which he would not wish obscured or jeopardised by a fiercely paradoxical supporting statement. Compare the form of KRS 199, fr. 61:

Sea is the most pure and the most polluted water; for fishes it is drinkable and salutary, but for men it is undrinkable and deleterious.

Moreover the reciprocity of the changes between waking and sleeping is quite uncontroversial. There is nothing in the way he writes to suggest that Heraclitus thinks the other reciprocities are any more dubious or remote from common experience. I therefore settle for a banal hypothesis to explain what he had in mind in these cases: the younger generation not only becomes old in due course, but takes the place of the older generation before it; and as his grandfather dies the grandson is born to take his place in the family, destined to die himself in due course. It is not to be doubted that Heraclitus has deeper things to tell us about life and death than that, but we are not being invited to think of them here in the first instance.

The deepest of all Heraclitus' sayings about life and death is KRS 239, fr. 62:

Immortals mortals, mortals immortals, living their death and dying their life.

Kahn rightly says:²⁷ 'This is in point of form Heraclitus' masterpiece, the most perfectly symmetrical of all the fragments.' The core idea of the paradox, as he shows, is:

Mortals live the death of immortals.
Immortals are dead in the life of mortals.

Clearly it is a riddle which can be solved only by the reader who has thought his or her way through the whole Heraclitean logos already. If one takes the soul to be the reference of 'mortal', and earth and water, as physical elements, to be that of 'immortal', then reflection on fr. 36 will yield a coherent interpretation. Our life as mortals depends on the temporary extinction of the physical elements from which it comes to be and into which it is destroyed. And the processes of living and dying will be not only (as in fr. 88) the same in form but numerically identical: the birth of the soul is the death of water, and vice versa. Whether this is a proper interpretation or not, the saying penetrates further than any other in its suggestion, brilliantly conveyed by the multiple ambiguities of its syntax, that the very concepts of mortality and immortality, life and death, collapse into each other. The resonances of this implication reach further than the confines of an interpretation in terms of physical truth. It threatens both traditional religion, with its confident separation of mortals and immortals, and Pythagoreanism, with its consoling promise that there is really no such thing as mortality—it is merely a form of immortality.

VI

Heraclitus hints that he has a distinctive project in philosophy (KRS 246, fr. 101: a project, incidentally, not a method—there is thankfully no sign of introspection as a method in the fragments):

I searched out myself.

The outcome of the project is, I have argued, the idea that the same logos which is true of all other things, and the particular version of it which is realised in the physical kosmos, are true of the soul also. It is a physical substance subject to the unity of opposites and to opposite sequences of transformations.

Heraclitus intimates without stating this theory of the soul. For its status in his eyes is not that of private doctrine but something publicly available to all if we will only reflect intelligently on our common usage and experience. Such reflection is, therefore, what we have to be got to engage in. His aphorisms and riddles are designed to prompt us to it.

It might be objected to this account of Heraclitus' psychology that the theory I have elicited from the fragments is idiosyncratic and unbelievable (now if not in the fifth century B.C.). Certainly the talk of fire, water and earth will no longer do. But otherwise, I submit, Heraclitus is near enough right and (as important) obviously right. Our selves, if not bodies, are very much functions of our bodies; and the crucial fact about them is that they are continually subject to opposite processes of psychological and (like everything else in the physical world) physical transformation.

Commentators ancient and modern are constantly tempted to ascribe a much more elaborate psychology to Heraclitus than I have done. In succumbing to temptation they have created batteries of private doctrines, common to no one. Partly in order to avoid doing likewise I have steered clear of the eschatological fragments—sayings in which Heraclitus tells us very little, but invites us to make hypotheses about why he might have formulated them, for example KRS 235, fr. 25, and fr. 27:

Greater deaths (moroi) are allotted greater portions (moirai).

What awaits men at death they do not expect or imagine.

In fr. 25 the word play on moros and moira finds him once again extracting a moral from common language. It hints at a good destiny for those who die heroic deaths. Reasonable guesses are possible about the guesses Heraclitus intimates he would make about the meaning of such hints.²⁸ But he could not see what specific form that destiny will take. We should recall his own warnings about our universal human limitations in the face of death (fr. 27), and about the difficulty of pinning down the soul (fr. 45).²⁹ If we do, we will not assert on his behalf either an eschatological theory presented as comparable in status to the general theory of the soul or the denial that he would have countenanced eschatology at all.³⁰ The signs are signs, but not strong or specific enough for dogmatism or a determinate theory to be in order: the unexpected cannot be searched out (fr. 18). A readiness to be surprised is more appropriate.

This [essay] makes a number of assumptions it does not attempt to justify: for instance that we can securely identify a number of sayings as authentically Heraclitean; that these are mostly self-contained aphorisms or riddles or bursts of invective best understood by reflection on their mutual resonances and against the background of what we know of earlier and contemporary philosophy and literature; that commentary on Heraclitus by later classical authors in antiquity is always to be viewed with some suspicion, often justified. The meaning of every sentence, clause and indeed word in Heraclitus is fiercely disputed by scholars; I have made few attempts to show why or how or even that this is so in presenting individual fragments. Footnotes are with one or two exceptions simply references to well-known treatments of Heraclitus in the modern literature, almost exclusively in English, that have been mentioned in the text.

Notes

¹ Kirk, Raven, Schofield [61], 203ff.

² Barnes [58], 1.61.

³ Kahn [67], 249.

⁴ Nussbaum [74], 169.

⁵ Marcovich [66], 364 (his italics).

⁶Ibid., 289.

⁷Ibid., 363.

⁸ Kirk [73]; cf. KRS [61], 207-8.

⁹ Kahn [67], 244.

¹⁰ Snell [42], ch. 1.

¹¹ Nussbaum [74], 1-3.

¹² Bollack and Wismann [70], 323-4.

¹³ Cf. Nussbaum [74], 156-7.

¹⁴Ibid., 153.

¹⁵ Kahn [67], 214.

¹⁶ Marcovich [66], 244.

¹⁷ Robinson [68], 93.

¹⁸ Hussey [72], 54.

¹⁹ Kahn [67], 216.

²⁰ Cf. Hussey [60], 57-9, for an attractive alternative: no account of the soul can be final, reaching its limits (cf. fr. 45), because the fact and the significance of the fact that the soul is itself author of the account is not included within it; include them within it, but then recognise that the fact that the soul is author of this enlarged account is not included in it; and so ad infinitum.

²¹Ibid, 238-40, 247-52.

²²Ibid, 239.

²³ Bollack and Wismann [70], 325. The textual problem is a difficult one. But if Heraclitus had himself conceived of both alternatives (auge psuchē, augē xērē psuchē), he would surely have preferred the former.

²⁴ [67], 250.

²⁵Ibid, 22-3, 136-8.

²⁶ Cf. Wiggins [76].

²⁷ [67], 218.

²⁸ So perhaps Kirk [73]; KRS [61], 207-8.

²⁹ Cf. Burnyeat [71]: 'The logos is language speaking about itself, giving signs to show us things that cannot be said, embracing opposites that cannot be reconciled, pointing us to alternative perspectives which we cannot take up. No wonder it is difficult. The difficulty is an irreducible part of the message. After explanation and elucidation have done their best, there is nothing for it but to let these memorable sayings take effect in the psyche in their own way.'

³⁰ Nussbaum [74], 169.