Heraclitus' Fragments: Kahn

Thought I'd present a fragment everyday and assoc. commentary by Charles Kahn. Its been years and ages since I've touched him, thought I'd recall him here with all. If someone would like to break the fragments etymologically, word by word in greek as in the original; I'm all eyes and ears. Left to everyone's convenience.


I (D. 1). 1 Although this account (logos) holds forever, men ever fail to comprehend, both before hearing it and once they have heard (or 'when they hear it for the first time').

"Both Aristotle and Sextus tell us this passage came at the beginning of the book. As we have seen, the initial ambiguity in the syntax of aiei 'forever' hints at the deeper ambivalence in the status of Heraclitus' logos: it is both his discourse and something more: some- thing universal (all things occur in agreement with it), even eternal and divine (eon aiei), precisely in virtue of the fact that it is 'common' or 'shared' by all (xynos in III, D. 2).

This first sentence sounds the twin themes of hearing and comprehending that will recur with increasing significance throughout the fragments (II, D. 34; XIV, D. 55; XV, D. 101a; XVI, D. 107; XVII, D. 19; XXVII, D. 108; XXXVI, D. 50, etc.).

The complaint that his auditors are unable to comprehend is a natural one on the part of an author who has chosen the language of enigma and equivocation. What is more puzzling is the insistence that men prove uncomprehending not only 'once they have heard my discourse' but even before. How can they be expected to understand it in advance? This will make sense only if Heraclitus' logos represents a truth that has been there all along: if, like Fire, it always was and is and will be.

Thus the logos here cannot be just 'what Heraclitus says', not merely the words he utters or even the meaning of what he has to
say, if meaning is understood subjectively as what the speaker has in mind or his intentions in speaking. The logos can be his 'meaning' only in the objective sense: the structure which his words intend or point at, which is the structure of the world itself (and not the intensional structure of his thought about the world). Only such an objective structure can be 'forever', available for comprehension before any words are uttered. Which is not to say that we can translate logos by 'structure' or by 'the objective content of my discourse'. The tension between word and content is essential here, for without it we do not have the instructive paradox of men who are expected to understand a logos they have not heard."

I. 2 Although all things come to pass in accordance with this account (logos), men are like the untried when they try such words and works as I set forth, dis- tinguishing each according to its nature (physis) and telling how it is.


"The second sentence begins with a clause in the genitive case that echoes not only the syntax but also the vocabulary of the beginning. The verb ginesthai 'become, come to pass', which first expressed men's lack of comprehension, is here applied to all the things that there are (ginomenon panton) and that men fail to comprehend. The genitival construction, which first depended upon axynetoi 'unable to comprehend', is now connected with apeiroi 'lacking in experience'.

The formal parallels between these two sentences suggest that Heraclitus is developing a single point. In both cases the announcement of the logos and its universal truth (in the genitival clause) is contrasted with the incompetent response of mankind (in the principal clause). The tension between the two aspects of logos — the actual words of Heraclitus and their everlasting content — is stretched still further here, where the logos as universal law is juxtaposed with Heraclitus' reference to his own exposition, in the emphatic first person (hokoion ego diegeumai 'such as I set forth').

'They are like the untried', 'they resemble men without experience' is a surprising phrase; for it suggests that in fact men do have the experience in question. And well they might: since it is experience of things that occur according to the 'logos', and these are all things, no one can be without such experience. But it is difficult for men to grasp this truth, even when Heraclitus announces it to them directly. They can make nothing of his words (epea), nor of the facts (erga) which he points out, although he 'tells it like it is' and puts each thing in its place, 'according to its physis'.

The word for 'facts', erga ('works', 'deeds') has epic overtones: it may refer to heroic exploits and also to more humble labor, as in Hesiod's Works and Days. The term physis, on the other hand, for the genuine nature or structure of a thing, is the watchword of the new natural philosophy that radiates from Miletus. By the use of this characteristic word, which recurs in X (D. 123) and XXXII (D. 112), as by his use of historie ('inquiry') and kosmos (in the sense of 'world order'), Heraclitus expressly claims affinity with the new scientific tradition, and thus offers his own truth as a supplement or as a rival to that of the natural philosophers."

Isn't Heraclitus presenting himself as the one who sets forth the logos, judging others by how well they've understood?

This sounds more like a preparatory speech for what is to come.
A clever preemptive accusation.

Yes.
Lets read further.

- - -

I. 3 But other men are oblivious of what they do awake, just as they are forgetful of what they do asleep.

What is striking here is not so much the self-assurance (not to say arrogance) of the thinker who regards 'other men' as sleepwalkers, but the almost pathetic epistemic isolation of a man trying to convey the vision of an obvious and immediate truth to men who stagger past, unable to notice what they are doing all day long, as if it were a dream they cannot grasp or hold on to. The image of sleep (which will elsewhere provide a kind of link between life and death) serves here to give a more drastic expression to the idea of cognitive alienation. Coming as it does at the very beginning, this paradoxical conception of the human condition as a state of deepest ignorance in the face of an immediately accessible truth serves to define the basic framework within which the specific doctrines must be understood.

In particular, it warns us against an over-hasty interpretation of his relationship to Ionian science. The historie of men like Anaximander or Hecataeus, or even Pythagoras and Xenophanes, represents new and technical knowledge, the product of special research, whether derived by watching the stars or by visiting the Thracians and Persians. Such historie is certainly not the apprehension of a universal truth of immediate experience, as accessible to men as 'what they do awake'. Hence the attitude of Heraclitus to such science will turn out to be profoundly ambivalent. His own philosophic vision is inspired by the new scientific study of the world, but it is directed towards a truth of an entirely different kind.

Perhaps one may compare Heraclitus' sense of epistemic isolation, and this ambivalent relationship to contemporary scientific knowledge, with the position of Bishop Berkeley in regard to Newtonian physics and optics. Berkeley's sensory idealism depends historically and psychologically upon the new science of his day, and upon care- ful studies in the geometry of vision. But his philosophical position as such does not logically depend upon any technical knowledge whatsoever. On the contrary, it involves a complete reinterpretation of what all scientific knowledge is about. Berkeley's teaching too involves a truth (from his point of view) closer than hands and feet, which ought to be obvious to every person, but which is in fact devilishly difficult to communicate even to a benevolent reader.

The most crucial disanalogy here — apart from the contrast between the state of the sciences in 500 B.C. and in A.D. 1700 — is the fact that Berkeley's doctrine is concerned with issues that are purely theoretical or cognitive: with the nature of knowledge and what is known. For Heraclitus, questions of cognition are inseparable from questions of action and intention, questions of life and death. The blindness he denounces is that of men who 'do not know what they are doing'. It is the life of mankind that is the subject of his discourse, not the theory of knowledge and perception.

With this proviso, we can say that his proem does characterize human life in epistemic terms, in terms of a well-nigh universal failure to make sense of one's experience. His initial concern is less with the structure of reality than with the extreme difficulty of grasping this structure."

To me, the phrase is much simpler.
I do not read so much into it.
I think these aphorisms are to be read in combination...one building upon the other, all in conjunction offering an insight.

Heraclitus simply says that some men are morons; that they are as lucid as a sleeper.
"Other" men, meaning besides he and his kind, are "oblivious of what they do awake": they act intuitively, instinctively, emotionally, like animals, without "knowing" what they do....or what governs their actions.
They are actors with no knowledge...and knowledge is the past.

A sleeping man has no control over his dreams...he sees visions which are confusing to him, full of meanings and sensations with no congruity.
He witnesses his own actions like a detached Will, helpless to intervene and condemned to just watch himself act and be acted upon.

When asleep the Will is disconnected from the mind. Sensuality comes from inward, not from the world.
A sleeper is a solipsist.

He's trying to set Heraclitus in an overall historical context and philosophical continuity.
Its just an aside. To me, it means, Heraclitus thinks men seem to be living mindlessly whether they are awake or asleep. They live de-natured; unaware of the logos of their own be-ing. Because life itself is unconsciously lived, obliviously, the dreams too offer no wealth, no weal. Dreams are where the unmanageable/unmanaged excess of living tries to sort itself out - nourishing/depleting the person acc. to his will; i.e. if one absorbs nothing willfully, is mindful of nothing when awake itself, then dreams too are like dry manure. One takes back nothing from it. One retains nothing. Mind-full-ness when awake is active Re membering when asleep. You Re member yourself together - the bits that escaped the awake conscious.
Thank you.

- - -

II (D. 34) Not comprehending, they hear like the deaf. The saying bears wit- ness to them: absent while present.
Although the account (logos) is shared, most men live as though their thinking (phronesis) were a private possession.

Since II and III contain elaborate echoes of the proem in diction and in thought, it is natural to suppose they followed on it rather closely. (In the case of III Sextus tells us as much.) Three of the first four words in II are formal repetitions from I: axynetoi, akousantes, eoikasi. The new element is the comparison to the deaf. The paradox 'absent while present' confirms the sense of epistemic isolation. There seems to be an audience there, men listening, but no communication is possible, nothing gets through. These pathetic listeners, who include most men and most of Heraclitus' illustrious predecessors, must be somewhere else, 'off on their own trip'.

Ill (D. 2) also opens with a phrase from the proem. But instead of 'forever' the logos now has as its predicate xynos 'common', 'shared'. In its first occurrence this key term is marked by stylistic emphasis, where the mechanism of repetition leads us to expect aiei, the epithet of divinity.

On first reading, we understand the logos as 'common' because it
is shared by all things or events, which take place in accordance with it (1.2). But common logos also means 'common consent', 'common cause', as when several powers combine in an agreement or alliance. This is brought out in a later echo of 'listening to the logos', when Heraclitus speaks of wisdom on the listeners' part as 'speaking in agreement', homologein (XXXVI, D. 50). The notion of 'consensus' or 'agreement' is only latent here; more unmistakable is the notion of the 'common' as the public, what belongs to the community as a whole, in contrast to what is private.

Another overtone of xynos logos will emerge in the word play of XXX (D. 114), where toi xynoi panton 'what is common to all' presents a phonetic echo of xyn nodi legontes 'speaking with understanding'. That 'common' (xynos) in III may also suggest the word for understanding or intelligence (noos) is made likely by the contrast here between the common logos and the private phronesis which men have or claim. I translate phronesis as 'thinking' to preserve the resonance with the cognate verb phronein, translated 'think' in IV and elsewhere. The term also means 'intelli- gence', 'good sense', 'wisdom' in practical decisions. Hence the verb zoousin here: it is a question of how men 'live their lives'.

In sum, the logos is 'common' because it is (or expresses) a structure that characterizes all things, and is therefore a public possession in principle available to all men, since it is 'given' in the immanent structure of their shared experience. The logos is also shared as a principle of agreement between diverse powers, of understanding between speaker and hearer, of public unity and joint action among the members of a political community. The logos is all these things because the term signifies not only meaningful speech, but the exercise of intelligence as such, the activity of nous or phronesis. The deepest thought of xynos logos, more fully expressed in XXX, is that what unites men is their rationality, itself the reflection of the underlying unity of all nature.

I assume that logos means not simply language but rational dis- cussion, calculation, and choice: rationality as expressed in speech, in thought, and in action. (All these ideas are connected with the classic use of logos, logizesthai, epilegein, etc., e.g. in Herodotus.) This is rationality as a phenomenal property manifested in intelligent behav- ior, not Reason as some kind of theoretical entity posited 'behind the phenomena' as cause of rational behavior. The conception of logos as a self-sub sistent power or principle is foreign to the usage of Heraclitus, but essential in the Stoic conception of the divine Reason that rules
the universe. (See on V Below).

IV (D. 17) Most men do not think (phroneousi) things in the way that they encounter them, nor do they recognize what they experience, but believe their own opinions.

"The first two clauses of IV reformulate the theme of failure to under- stand; the third introduces the topic of seeming knowledge or guess- work (dokein). So far the cognitive condition of mankind has been characterized negatively, as deafness, absence, lack of understanding. The description in more positive terms begins with the comparison to sleep and hence dreaming in 1.3, and with the suggestion in III that error involves treating phronesis as if it were something private. (The convergence of these two ideas, privacy and sleep, will come in VI.) The last clause of IV points to a connection between this description of error and an older, half-technical concept of 'appearance' or 'opinion', expressed by cognates of the verb dokein 'to seem'. Heraclitus' own use of the concept of 'opinion' is discussed below,
on LXXXIV (D. 27)-LXXXV (D. 28A).

The notion of incomprehension is developed by a complex literary device: each clause takes a traditional topos or commonplace of Greek poetic wisdom and stands it on its head. The effect is to present Heraclitus' thesis not simply as a challenge to contemporary savants, but as an implicit correction of the wisdom of Homer, Hesiod, and Archilochus.

One of the most famous passages on 'thinking' in Homer is Odyssey XVIII.136: The mind (noos) of men upon the earth is such as the day which the father of gods and men brings upon them.' In context, this means that the fortune and happiness of men is dependent upon the decision of the gods; but the dependence is stated in terms of men's understanding of their situation. (And this statement of Odysseus finds its immediate application in the folly of the hearer, Amphinomus, who fails to heed the warning of which these words
are part, and which would have saved his life.)

Almost the same phrase occurs, with roughly the same meaning, in a fragment of Archilochus (68 Diehl): T h e heart (thymos) of mortal men, Glaucos son of Leptines, becomes such as the day which Zeus brings upon them, and their thoughts (phroneusin) are such as the deeds (ergmata) that they encounter (enkyreosin)\ that is, their thought is determined by their situation.

The first clause of IV contains, in its two verbs and its comparative structure (toiauta . . . hokoia 'in the way that' or 'such . . . as') a clear echo of Archilochus' own words. But Heraclitus echoes Archilochus only to deny what the latter affirms. Men's thinking is not in conformity with the reality they encounter — would that it were!

The second clause denies another familiar lesson: learning the hard way, pathei mathos. Before becoming the theme of Aeschylus' Oresteia, 'knowledge through suffering' was the standard character- ization of the fool who learns too late. In Hesiod's parable of the two roads, one of Justice (Dike) and one of Crime (Hybris), the way of Crime is said to be easier at the beginning but disastrous at the end: 'and the fool recognizes this (egno) when he has suffered his punish- ment (pathon), (Works and Days 218).

Again, Heraclitus alludes to Hesiod's commonplace only to contradict it: experience, even suffer- ing, teaches men nothing at all. They do not have the wit to recognize (ginoskein) what is happening to them. The insights that men find in the poetry of Homer, Hesiod, and Archilochus are nothing more than 'believing their own opinions': enough wisdom to satisfy their own folly.

This is our first encounter with the verb gignosko 'to recognize, perceive, know', which occurs in Hesiod's proverb as in Archilochus' own claim to insight.67 Heraclitus has taken over this verb, with its cognate nouns gnome and gnosis, as his own term for 'cognition' in a privileged sense, for the insight which men lack and which his own discourse attempts to communicate."

V (D. 71—3) Marcus Aurelius: [Always bear in mind what Heraclitus said . . . And bear in mind the man who forgets where the way leads, and that 'they are at odds with (diapherontai) that with which they most constantly associate' (the logos which controls all things), and that what they meet with every day seems strange to them, and that one should not act and speak like men asleep.]

(Here I render the fuller context in Marcus; in the translation I gave only what might be a citation from Heraclitus.)
Although this text contains an echo of several fragments (notably of IV), there is no clear trace of literal quotation and hence no basis for detailed commentary. They are at odds with that with which they most constantly associate' represents something Heraclitus might
have said, and the implicit word play on diapherontai is attested else- where (CXXIV, D. 10). But Marcus Aurelius is citing from memory, and since his memory is not very accurate there is no way to tell whether he is preserving a genuine fragment or simply developing his own recollection of IV.
In any case, the words in parenthesis, 'the logos which controls the universe', represent (even in terminology: dioikein ta hola) the Stoic rather than the Heraclitean conception of logos.

VI (D. 89) Plutarch: [Heraclitus says that the world (kosmos) of the waking is one and shared but that the sleeping turn aside each into his private world.]

Plutarch is paraphrasing, not quoting verbatim: he uses the ordinary word koinos for 'common', where Heraclitus always has xynos (5 times in the extant fragments). The absence of any archaic overtones in the sense of kosmos here also suggests a post-Platonic phrasing. In general, the rhetoric of the passage is that of Plutarch, not Heraclitus. But I see no reason to doubt that Plutarch is rendering Heraclitus' thought correctly.

VII (D. 18) He who does not expect will not find out the unexpected, for it is trackless and unexplored.
VIII (D. 22) Seekers of gold dig up much earth and find little.
IX (D. 35) Men who love wisdom (philosophoi andres) must be good inquirers
(histores) into many things indeed.
X (D. 123) Nature (physis) loves to hide.

I have grouped these four quotations on the basis of their common imagery of searching, finding, being hard to find. Only in one case (IX, D. 35) is there any doubt as to the authenticity of the wording. If genuine, this would be the earliest occurrence of the term philoso- phos in Greek. Since the word here has its etymological force ('lover of wisdom' or 'eager for learning new and clever things'), and is paired with the term histores that suggests Ionian inquiry (cf. historie in XXV, D. 129), I have ranked this as a quotation. (Clement is gener- ally one of our best sources for literal citations.)

If Heraclitus used the term philosophos, as I suppose, he may have intended an allusion to the other masters of wisdom, the seven sophoi or sages, two of whom (Thales and Bias) are mentioned in the fragments. Thus philosophoi andres admits a secondary reading: 'men who want to become sages'. Heraclitus himself makes a special use of the concept of to
sophon 'what is wise'. It would be in character for him to introduce the theme of wisdom in the compound form philo-sophos, as the object of ardent desire.

As grouped here, these four quotations deal with the difficulty of cognition from the side of the object. Whereas I—IV describe human incomprehension as a kind of perverse blindness in regard to what is staring us in the face, VII—X recognize that the truth, the character- istic nature of things (physis), the prize of wisdom hunted by phil- osophical goldseekers, is not simply there for the taking. Even if the logos is common to all, so that the structure of reality is 'given' in everyday experience, recognition comes hard. It requires the right kind of openness on the part of the percipient — what Heraclitus calls 'hope' or 'expectation' (elpesthai in VII; compare LXXXIV, D. 27).
And it requires inquiry and reflection — digging up a lot of earth and
judging it with discretion. The 'gnosis' which Heraclitus has in mind is rational knowledge, and it has to be gained by hard work; it is not the miraculous revelation of a moment of grace.

XI (D. 47) Let us not concur casually about the most important matters.
XII (D. A23) Polybius: [It would no longer be fitting to take poets and story- tellers as witnesses for things unknown, as our ancestors did in most cases, citing untrustworthy authorities on disputed points as Heraclitus says.]
XIII (D. 74) Marcus Aurelius: [We should not act and speak like 'children of our parents', in other words, in the way that has been handed down to us.]

These three texts express a critical attitude towards traditional or current practice and belief; they add little to the more authentic frag- ments. I print XI as quotation rather than paraphrase only because it occurs in a context (D.L. IX.71—3) with verbatim citations from other authors.

XIV (D. 55) Whatever comes from sight, hearing, learning from experience: this I prefer.
XV (D. 101a) Polybius: [According to Heraclitus, eyes are surer witnesses than ears.]
XV1 (D. 107) Eyes and ears are poor witnesses for men if their souls (psychai) do not understand the language (literally, 'if they have barbarian souls').
XVII (D. 19) Not knowing how to listen, neither can they speak.

These four statements develop and clarify the thought of VII—X: the usual sources of information are necessary but not sufficient for the cognition that Heraclitus is trying to describe.

XIV seems to insist upon the fact that the truth is not so recondite as it might seem, or as some will pretend: far from requiring any special revelation or abstruse theory, what Heraclitus treasures is something that can be grasped by sight, hearing, and ordinary experience of life (mathesis: compare mathontes in IV , D. 17). The preference for eyes over ears has a proverbial ring (as in the Gyges story, Herodotus 1.8.2). It expresses not so much an epistemic ranking of the senses as the reliance upon direct experience rather than upon hearsay. This is basic both for Ionian historie (inquiry conceived as finding out for oneself) and for Heraclitus' own attack upon the traditional wisdom of the poets.

But learning from experience involves more than keeping one's eyes and ears open: in order for thinking (phronein) to become recognition (ginoskein) the soul of the observer must not be 'barbarian', that is, it must know the relevant language. Heraclitus thus develops the double aspect of logos indicated in the proem: spoken words, and a universal pattern of experience.

The world order speaks to men as a kind of language they must learn to comprehend. Just as the meaning of what is said is actually 'given' in the sounds which the foreigner hears, but cannot understand, so the direct experience of the nature of things will be like the babbling of an unknown tongue for the soul that does not know how to listen.
This is apparently the first time in extant literature that the word psyche 'soul' is used for the power of rational thought. (Contrast the 'irrational' psyche described as waking when the body sleeps, in the Pindar fragment cited below, p. 127.)

The new concept of the psyche is expressed in terms of the power of articulate speech: rationality is understood as the capacity to participate in the life of language, 'knowing how to listen and how to speak'.

These intimate connections between speech, intelligence, and the logos which is 'common to all' appear again in XXX (D. 114), as in III (D. 2), and once more in XXXVI (D. 50).

XVIII (D. 40) Much learning does not teach understanding (noos). For it would have taught Hesiod and Pythagoras, and also Xenophanes and Hecataeus.
XIX (D. 57) The teacher of most is Hesiod. It is him they know as knowing most, who did not recognize day and night: they are one.

XVIII continues the thought of XVI ('eyes and ears are poor wit- nesses'), applying it not to the ordinary perception of the man in the street but to the claims of experts: such learning will not give the relevant kind of understanding. That requires something more, some- thing there conceived as speaking the right language and here called noos or 'intelligence'. Noos (Attic nous) is, among other things, the capacity to speak and to listen well.

The word can even denote the meaning or intention of what is said. (See Powell's Lexicon s.v. noos, 3.) This is a specialized development from the more general sense of noos as what someone 'has in mind', the character and intentions of men which underlie their words and actions. Thus Odysseus in his travels 'came to recognize (egno) the noos of many men', as the Odyssey tells us. But like phronein, which has the neutral sense of 'thinking' but the laudatory sense of 'thinking well, using good sense', the term noos also has a normative use. Thus doing something 'with noos9 means doing it in a reasonable way; acting 'without noos9 means acting foolishly. The word has its positive sense here, as in XXX (D. 114); cf. LIX(D. 104).

There is no inconsistency between this depreciation of 'learning many things' and the claim in VIII—X that a great deal of knowledge and experience is required in the pursuit of wisdom. Heraclitus does not say that the poly mathie is a waste of time, only that it is not enough: that the mere accumulation of information will not yield understanding, unless it is accompanied by some fundamental insight.

Four exemplars of learning are named here. Hesiod is the poet of the remote past, almost two centuries earlier than Heraclitus, whose didactic poems had come to enjoy the status of revered handbooks familiar to every educated Greek: the Theogony as the authoritative account of origins, dynasties, and family connections among the gods, the Works and Days as a summa of practical lore, from farming and astronomy to instruction on unlucky days. As an established expert on all matters human and divine, Hesiod is a natural target. Pythagoras may be named next to Hesiod because he, alone among Heraclitus' recent predecessors, had achieved a kind of legendary prestige within his own lifetime. Xenophanes and Hecataeus have a narrower claim to fame: they represent the diffusion of Milesian historie in literary form.

From our perspective on the history of early Greek philosophy, gained from Aristotle and the Theophrastean doxography, it is natural to wonder why we find no mention here of the three Milesians whom we think of as the creators of the new science:
Thales, Anaxi- mander, and Anaximenes. Thales' name will in fact occur in XXIV. Heraclitus' silence concerning Anaximander and Anaximenes is prob- ably neither the result of ignorance nor a sign of special respect. Heraclitus (like Plato after him) simply did not think of himself as addressing the narrow audience that would be familiar with the work of these men. (They are not mentioned by any extant writer before Aristotle, and by scarcely anyone later outside of the technical doxo- graphic tradition.) Xenophanes and Hecataeus were influential authors in their own right and would be known to a wider public.

The attack on Hesiod in XIX provides the first explicit riddle, and the first clue as to the content of Heraclitus' logos: day and night are one. The point of the riddle is sharpened by the ironical use of three different verbs for 'to know': epistantai for the popular intelligence which selects this teacher; eidenai for the knowledge they ascribe to him, and Heraclitus' favored term ginoskein for the cognition which is denied to Hesiod.

The content of such cognition is expressed here by an unexplained assertion of unity between night and day. This is one Heraclitean riddle about whose solution we can be reasonably certain. It stands first as a kind of emblematic statement for the doctrine of unity or interdependence between opposing powers that constitutes the gen- eral pattern and the formal theme for Heraclitus' teaching. It also represents a quite specific insight of an almost technical sort.

Hesiod had conceived Night (Nyx) in the manner of early mythic thought, as a positive force which blots out the light of day and the vision of men, as death blots out human life. In Hesiod's view, Night is naturally paired with her opposite number Day, whom she meets at the door of the underworld, one leaving as the other returns from the trip above ground (Theogony 748—57). No pattern of opposition seems more firmly grounded in the basic experience of mankind than that between daylight and nocturnal darkness. So no affirmation of the unity of opposites could be more striking nor, at first glance, more paradoxical. But the paradox is only apparent. The natural unit of time, constant throughout the year, is not the day but what the Greeks later called the nychthemeron, the day-and-night interval of twenty-four hours. The astronomers of early Greece had no clocks, but their interest in the calendar made them pay close attention to the changing length of daylight through the year.

The discovery of the nychthemeron is just the discovery that the nights get shorter as the days get longer, though the time from noon to noon does not change. The doxography reports, and the fragments of Heraclitus confirm, that considerable attention was paid to the measurement of solstices and equinoxes, and thus to the regular correlation between the visible course of the sun and the relative length of day and night. The knowledge of solstices or tropai, as the place where and the time when the sun annually reaches its farthest point to the north and to the south, is as old as Homer and Hesiod; but such matters came under more intensive study in Miletus in the sixth century, probably under Babylonian influence.71 I have no doubt that these are the 'measures' of the sun referred to below in XLIV (D. 94), where they are said to be enforced by Justice herself.

Something similar must be intended by the 'limits of dawn and evening' in XLV (D. 120), over
which Zeus has set his guardian. The word for solstices, tropai, occurs as a cosmological term in XXXVIII (D. 31 A). The insight that night has no positive force but is merely the absence of sunlight, and there- fore varies directly with the changing course of the sun, is clearly expressed in XLVI (D. 99): without the sun it would be night. In this apparently childish remark lies the solution to our riddle and the real point of this attack on Hesiod.72 Far from being separate and irrecon- cilable powers, day and night are complementary aspects of a single unit.
This riddle and its solution is a paradigm for the relation between the logos of Heraclitus and the poly mathie of Ionian science.

The new science has provided Heraclitus with a much clearer understand- ing of the relationship between solar motion and length of night-time than was available to Hesiod. But no amount of historie or astro- nomical knowledge alone could situate this relationship within a uni- versal pattern of interdependent or co-variant opposition — within the general framework that we call Heraclitus' 'doctrine of opposites'. And this doctrine, as expressed in the identification of night and day, is itself a symbol and a clue for the understanding of human life and death as a unity, which forms the central insight in what Heraclitus means by 'wisdom'.

XX (D. 106) Plutarch: [Heraclitus attacked Hesiod for counting some days as good, others as bad, because he did not recognize that the nature (physis) of every day is one.]

It is not clear whether this is a separate criticism of Hesiod or a variant on XIX (D. 57).73 It seems unlikely that Heraclitus would have diluted his attack on Hesiod's doctrine of opposites — of which Night and Day are splendid examples — by turning to the conceptually less interesting contrast between lucky and unlucky days; but the exist- ence of this saying in two independent forms (Plutarch and Seneca) makes it difficult to reject out of hand. Perhaps the reference is not to the annual course of the sun and its effect upon the balance of night and day but rather to the nature of the sun itself as a mani- festation of cosmic fire; or to the equivalence between one day and its successor.

XXI (D. 42) [He said that] Homer deserves to be expelled from the competition and beaten with a staff — and Archilochus too!

This remark presupposes that the poems of Archilochus, like those of Homer, were recited by rhapsodes in public competitions at the fes- tival games (agones), in connection with athletic contests. (Compare the situation a century later, when Plato names Archilochus and Hesiod with Homer as the standard authors for rhapsodic compe- tition, Ion 531 A.) Heraclitus takes the wise words of all three 'official' poets — Homer, Archilochus, and Hesiod — as targets for his attack.

As Bollack-Wismann have pointed out (p. 157), XXI can be read as beginning with an invective against Homer in the salty style of Archi- lochus himself, and then turning the new weapon against its inventor. The patron of the rhapsodes, whose poems were the original subject of recitation, deserves to be punished with his own symbol, the rhapsode's staff or rhabdos. But the competitor who would replace him is expelled in turn by the same staff

XXII (D. 56) Men are deceived in the recognition of what is obvious, like Homer who was wisest of all the Greeks. For he was deceived by boys killing lice, who said: what we see and catch we leave behind; what we neither see nor catch, we carry away.

This fragment gives the solution to what seems to be a traditional riddle. But it poses its own puzzle: why does Heraclitus find the story significant?
In the Lives of Homer that circulated in late antiquity, the riddle of the lice is regularly cited in connection with Homer's death: the poet is usually said to have died of grief at not being able to guess the answer.76 The later versions emphasize Homer's blindness, and that is probably presupposed by XXII. If so, we have a parallel to II (D. 34): as deafness there represents men's hearing without comprehen- sion, so here the failure to recognize what is clearly visible (ta phanera) is compared to Homer's blindness.
Now the story does serve Heraclitus' purpose of showing the
wisest of the Greeks to be a fool: he cannot guess a riddle known to children.77 But the riddle itself should be meaningful.

Bollack-Wismann propose an etymological play here on the word for lice, phtheir, phonetically identical with the first syllable of phtheir 6 'to destroy': what the boys are doing is literally killing the killers', and this is what Homer fails to understand. Furthermore, the riddle has
the symmetrical form of a double paradox: denying the expected consequence of 'seeing and catching', and then affirming this conse- quence for not seeing and catching. If the play on killing phtheires is a clue, we might well be reminded of Euripides' paradox which has
the same double form: 'Who knows if life be death, and death be regarded as life below?' (fr. 638 N). Heraclitus himself will offer a paradox of this form, in which mortals and immortals exchange positions with regard to life and death (XCII, D. 62).

Of course no one could possibly guess this meaning on a first reading of XXII. But then the device of proleptic statement, or more generally of resonance, implies that many dimensions of meaning will not be immediately accessible on our first contact with the text. ('Men fail to comprehend this logos when they hear it for the first time.')
The proposed solution becomes even more plausible if we suppose that the riddle of the lice was already associated with the story of Homer's death. For then it is natural to understand the 'recognition of what is apparent', in which men are deceived, by reference to Homer's perplexity in the face of his own death. The etymology of lice (phtheires) as 'destroyers' thus comes into full play.

Heraclitus' detailed presentation of this riddle, in one of the longest of the fragments, suggests that genuine wisdom is like an enigma, dif- ficult and obvious at the same time, or difficult at first and obvious later. It also suggests that wisdom has something to do with what we see and grasp, but more to do with what we ordinarily do not see and do not grasp.

XXIII (D. 105) Scholia on Iliad XVIII.251: [Heraclitus calls Homer an astron- omer.]
XXIV (D. 38) Diogenes Laertius: [According to some sources, Thales seems to have been the first to practise astronomy and predict solar eclipses and solstices
. . . Heraclitus also bears witness to him.]

There is not much to be made of these bits of doxographical infor- mation, except that the names of both Homer and Thales were men- tioned in connection with astronomy, perhaps in the same context (as Marcovich has suggested), just as Heraclitus mentions Hesiod and Hecataeus together in XVIII (D. 40). Exactly what Heraclitus said about Thales we can scarcely guess. But the omission of Thales' name in XVIII (D. 40), where it might easily have stood between Hesiod and Pythagoras, suggests that the tone was not one of violent hostility; and that is on the whole confirmed by the context in Diogenes.

Probably the reference to Thales reflected his legendary fame as a star- gazer, as in the story of the eclipse in Herodotus 1.74.2: a similar story seems to have been mentioned by Xenophanes (DK 21.B 19). The reference to Homer in XXIII is equally uninformative. Its authenticity has been questioned on the grounds that the scholiasts who report it assume that astrology is meant, as we can see from the verses they cite.

Now there is no trace of such astrology in Greece in the time of Heraclitus or for a century or two thereafter. But it does not follow that we should reject the whole reference. The astrological interpretation with its selection of Homeric verses must be the work of a Hellenistic commentator. But there would have been nothing for him to comment on if he had not found some connection between Homer and astronomy in Heraclitus' own text

XXV (D. 129) Pythagoras son of Mnesarchus pursued inquiry (historie) further than all other men and, choosing what he liked from these compositions, he made a wisdom of his own: much learning, artful knavery.

Diels believed XXV was a forgery, because of the reference to written 'compositions' or treatises (syngraphat) earlier than Pythagoras, apparently dealing with scientific research or inquiry. Since its defence by Reinhardt and Wilamowitz, XXV is now universally recognized as authentic, but there is still room for doubt as to exactly what 'com- positions' are referred to.79 The word syngraphai, like its Attic equiv- alent syngrammata, normally refers to works in prose. The earliest prose treatises known to us are those of Anaximander and Anaxi- menes of Miletus and Pherecydes of Syros.

As it happens, there was an ancient tradition associating Pythagoras with Pherecydes, the author of a strange mythic cosmogony of which some portions have been preserved.80 The treatises of Hecataeus, known to Heraclitus as embodiments of Milesianpolymathie (see XVIII), were probably too recent for him to have in mind as sources for Pythagoras. But there were no doubt other prose treatises circulating in the neighbourhood of Miletus, Samos, and Ephesus, of which we know little or nothing. We can only guess what teachings or exploits of Pythagoras would have provoked Heraclitus' particular indignation.

He would probably have been shocked by Pythagoras' claim to recall his previous incar- nations, which might have seemed like a caricature of the notion of pre-existence and survival of the psyche which plays a part in Heraclitus' own thought. For all his 'scientific' Ionian background, there was something of the sorcerer and wonder-worker about Pythagoras which was probably inseparable from his enormous pres- tige in the Greek colonies of South Italy and his considerable repu- tation throughout the Greek world. Though he was certainly no shaman in the sense of being a primitive witch-doctor, he does seem to have cultivated the role of charismatic leader with superior powers, like bilocation — a man who would be venerated by his followers as more than human, but regarded by outsiders as something of a fraud. That Heraclitus shared the latter view is clear from XXVI, if indeed it refers to Pythagoras.

XXVI (D. 81) Philodemus: [According to Heraclitus 'the prince of imposters.']

In Philodemus' text the picturesque phrase kopidon archegos refers to rhetoric, not to Pythagoras. For its application to Pythagoras we have only indirect evidence.82 The invective certainly fits Pythagoras. There may be a play on archegos (translated here as 'prince') in the sense of 'founder', 'the one who starts things going'. Pythagoras was not only an imposter himself, he founded a school! This corporate relationship between initiator and followers seems to be reflected again in the warning of LXXXVII (D. 28B): Justice will catch up not only with those who invent lies (pseudon tektones) but also with those who swear to them (martyres).

XXVII (D. 108) Of all those whose accounts (logoi) I have heard, none has got so far as this: to recognize what is wise, set apart from all.

Here for the first time (in my arrangement) Heraclitus begins to dis- close the positive content of his doctrine, so far hinted at only by the mysterious assertion of the unity of night and day (in XIX, D. 57). The imagery of the logos is now reversed: at first it was Heraclitus' account, or the logos as such, to which others must attend (and so again in XXXVI, D. 50); now it is others who do the talking and Heraclitus is cast in the role of listener. But he listens
in vain, for it is not in their words that he can find a recognition (ginoskein) of what he is seeking. This missing content is designated by the neuter form of the word for wise man or sage: sophon. Those who passed for wise did not deserve the name: they did not know what wisdom is, and hence were separated from it.

The last words of the sentence present an ambiguity that recurs frequently: the form panton 'of all' may be read either as masculine (i.e. animate) or neuter (inanimate). The difference is considerable: is the wise separate 'from all men' or 'from all things'? The first read- ing suggests wisdom as ordinarily conceived, a property men wish to possess, though in fact they fail to attain it. The second seems to posit the 'wise' as a cosmic or divine principle, separate or trans- cendent like the Intelligence (nous) of Anaxagoras (fr. 12) which is 'not mixed with anything else'.

Heraclitus could easily have specified his meaning by adding the word 'men' or 'things' after 'all'. Since he did not choose to eliminate the ambiguity, it is not up to us to do so: the principle of hermen- eutical generosity requires us to keep both options open. In this case there is good evidence to support both readings of panton:
(1) that wisdom is inaccessible to men is, in effect, stated by LV—LVIII (D. 78, 82—3, 79, 70); whereas (2) 'the wise' is asserted as a unique divine principle of the universe in CXVIII (D. 32). Since the same ambiguity between masculine and neuter readings of pronouns and quantifier- words will be found again and again, the possibility of two interpret- ations is not likely to be accidental.

It is significant that, when Heraclitus begins to specify the content of that insight which perception and inquiry alone cannot provide, his characterization is strictly ambivalent between a human property that is rare and difficult and a cosmic power which is unique and remote from all others. For it is precisely this ambiguity or duality between the life of man and the life of the cosmos that structures Heraclitus' entire discourse, as in the duality already noted between logos as the utterance of a man and logos as the pattern of cosmic process. It is just this duality, this interlocking of man and cosmos expressed by the double value of logos and sophon, as by the double reading of the pronoun pan ton, which constitutes the 'recognition'
that other men have missed.

XXVIII (D. 101) I went in search of myself.

This is as straightforward a paradox as any in Heraclitus. Normally one goes looking for someone else. How can I be the object of my own search? This will make sense only if my self is somehow absent, hidden, or difficult to find. Thus XXVIII states, or presupposes, what one might have thought was a distinctly modern reading of the Delphignothi sauton: self-knowledge is difficult because a man is divided from himself; he presents a problem for himself to resolve. We are surprisingly close here to the modern or Christian idea that a person may be alienated from his own (true) self.

XXIX (D. 116) It belongs to all men to know (ginoskein) themselves and think
well (sophronein, keep their thinking sound).

The connection between self-knowledge and sophronein is taken for granted in the archaic conception of wisdom preserved in the sayings of the Sages and inscribed at the entrance to the sixth-century temple of Apollo at Delphi, where meden agan 'nothing too much' was written next to gnothi sauton 'know (recognize) thyself. Both maxims might reasonably be paraphrased as sophronei 'be of sound mind'.

The pairing of self-recognition and sound thinking is thus tra- ditional; what is unusual in XXIX is the claim that all men have a share in this excellence. The Delphic injunction suggests that self- knowledge is difficult, and Heraclitus himself implies as much in XXVIII. So XXIX presents an apparent contradiction both with the Delphic motto and with Heraclitus' own words elsewhere.

The contradiction is not unresolvable; it renews the initial paradox of the logos which is there before us, but which we are unable to grasp. Self-knowledge and world-knowledge will in the end converge in this comprehension of the common logos. If we interpret XXVIII
'I went in search of myself as meaning 'I found in myself the universal logos, the cosmic law', we ignore the literal sense and the enigmatic technique by which Heraclitus presents his thought. But this is indeed the final implication of what he is saying. Self-knowledge, like under- standing the logos, belongs to all men by right. But in fact precious few will achieve it.

The formal challenge to the Delphic proverb, the apparent self- contradiction, and the etymological overtones of so-phronein brought out by the parallel to ginoskein, should suffice to guarantee the auth- enticity of XXIX (D. 116) against persistent doubts. These doubts rest only on the questionable judgment that the thought and wording of XXIX are too banal for Heraclitus.

XXX (D. 114) Speaking with understanding (xyn nodi) they must hold fast to what is shared (toi xynoi) by all, as a city holds to its law (toi nomoi) and even more firmly. For all human laws are nourished by a divine one. It prevails as it will and suffices for all and is more than enough.

The juxtaposition with XXIX (D. 116) is meant to suggest that the subject of XXX is precisely 'all men', whose right to self-knowledge or sophronein is asserted in XXIX. XXX thus specifies what conditions must be satisfied for that right to be realized in fact.

The text contains two examples of the animate-inanimate (mascu- line-neuter) ambiguity already noted. In the phrase 'common to all', panton can mean either 'all human beings' or 'all things', precisely as in XXVII (D. 108). Similarly, in 'it suffices for all' past can be 'for all men' or 'for all things'.87 There is a comparable ambiguity in the phrase '(nourished) by the divine one', which can be construed either in the neuter, with 'the divine' (theion) as a term for the supreme cosmic principle, or as the masculine form agreeing with nomos: 'the one divine law'. I suspect that this duality is also deliberate: the single divine principle, 'what is common to all', is and is not willing to be designated as nomos 'law'. (For this ambivalence with regard to a par- ticular name, cf. CXVIII, D. 32.)

The community of logos is expressed here by the phrase 'holding fast to (or 'strengthening oneself with' ischyrizesthai) what is common to all (toi xynoipanton)\ after the initial word play of 'speaking (together) with understanding' (xyn nodi legontas) with its double allusion to the logos ('speech') and to 'what is common' (xyn). But what is it that can teach understanding (noos) in speaking and listen- ing? What cognition can keep one's thinking sound (so-phronein) and prevent it from sinking into the private phronesis (III, D. 2) that has lost contact with what is shared by all? XXIX (D. 116) located the salvaging of one's thought in self-recognition, in what might seem to be the most private of cognitions. Yet Heraclitus everywhere insists that soundness and safety lies in what is public: the civic law is as essential as the city wall to the security of those who dwell within (LXV, D. 44). What is the connection between ^//-recognition and this stable reliance upon what is common?

The clue lies in the notion of self-alienation implied in XXVIII: 'I went in search of myself.' The lost or hidden self must be precisely the common, in search of which the private self or T must go, in order for one's thinking to be safe and sound. Heraclitus son of Bloson, an individual of Ephesus, went seeking for his own self, and what he found was something common not only to all the citizens of Ephesus but to all men, and not only to all men but to all things whatsoever. Thus did his self-knowing become so-phronein, the sal- vaging of thinking through understanding (xyn nodi), through hold- ing on to 'what is common {xynoi) to all'.

We can see why the nomos, the public law and moral tradition of a city like Ephesus, should be chosen as an illustration of what pro- duces sound thinking, but why it is only an illustration.88 It is com- mon, but not common enough: common to all Ephesians but not common to all men and all things. As a result, in the search for soundness of mind one must hold on to something stronger, the source not only of this but of 'all human laws'. The Stoics and other later writers had no hesitation in speaking here of 'divine law' or 'natural law' (as in Lucretius' lex natural). But in the archaic language of Heraclitus the term nomos has too human and too social a sound to lend itself to such a locution. Hence Heraclitus hints at, but does not express the notion of a 'divine law' theios nomos. Instead he leaves us with a characterization of the common as 'the divine one'.

We can see the syntactic ambiguity between 'all men' and 'all things' in XXX as a formal expression of the view that an individual man, searching for his own self, will come to rest and to rely upon a deeper identity which is that of the universe as a whole. The terms in which this unity is described recall the traditional references to the supreme god, and prepare us for Heraclitus' own introduction of the name of Zeus (in XLV, D. 120; CXVIII, D. 32)

XXXI (D. 113) Thinking (to phronein) is shared by all.

We meet once more the ambiguous pronoun pasi: shared by all men? or all things? The first reading simply restates in a weaker form the thought of XXIX (D. 116: so-phronein belongs to all men), adding the notion of what is common to all (as in XXX); 'What is "common to all" is not what all, or almost all, happen to think, but what all should think, and would, if they had sense' (Vlastos, 347). Yet this leaves XXXI essentially repetitious, and tends to justify the doubts of those who would condemn it as inauthentic.

But the ambiguity of pasi (in contrast with the unambiguous anthropoisi pasi in XXIX) permits a stronger reading, which has been ignored only because it is perplexing: 'Thinking is common to all things.' Taken literally, this implies something like panpsychism: if all things have some thought or perception, all things must be alive. This view may seem strange to some, but it is seriously held by Spinoza and Leibniz in modern times, and unmistakably held by Empedocles in the fifth century B.C.: 'For know that all things have thought (phronesis) and a share of mind (noma = noema).' (DK 31.B 110, 10; ci.panta 'all things' in B 102—3 and 107.)

This stronger meaning saves XXXI from banality by reinterpreting 'what is shared by all' as a kind of thought or intelligence. And this makes it easier for us to understand how self- knowledge can lead to the knowledge of what is common to all, since the universal principle is understood precisely as thinking, the activity of an intelligent psyche. Such panpsychism gives a new dimension to the ambivalence of Heraclitus' concept of 'the wise', oriented both towards the human and towards the cosmic. And the strong reading of XXXI fits well with the doctrine of the limitless psyche in XXXV (D. 45).

XXXII (D. 112) Thinking well (sophronein) is the greatest excellence and wisdom: to act and speak what is true, perceiving things according to their nature (physis).

This is the most important of the three fragments dealing with thinking (phronein) and thinking well (or sound thinking, sophronein) that are preserved only in the Anthology of Ethical Sayings composed by John Stobaeus, the other two being XXIX (D. 116) and XXXI (D. 113). The authenticity of one or more of these has been questioned.

The burden of proof properly falls on those who would deny authenticity, since all three citations are found together with other frag- ments that are indisputably genuine, and together with no fragments that are clearly inauthentic.9 l Since Stobaeus' work is a kind of scrapbook, the documentary value of his testimony varies greatly from case to case, and his reliability at any given point is only as good as that of the source he happens to be following. Now in the two con- texts that interest us, in his chapters On Virtue and On Temperance (sophrosyne) in Book III, we find quotations from Heraclitus of great importance and unquestionable authenticity, which in some instances are preserved in no other source. That is the case for XXVII (D. 108), XXX (D. 114), the important double fragment LXVII (D. 110—11), and the description of the drunken man with the wet soul (CVI, D. 117). Since neither context contains any manifest forgeries, the natural presumption (followed by Bywater, Diels, and Bollack- Wismann) is that all of the quotations from Heraclitus in these two brief sections of Stobaeus should be authentic.92 It would really be remarkable if an inferior paraphrase had been handed down to us sandwiched in between XXVII (D. 108), LXVII (D. 110-11) and XXX (D. 114).

To outweigh these circumstantial probabilities, a critic would have to show that the language and thought of XXXII cannot belong to Heraclitus. The interpretation which follows is therefore an implicit defense of the fragment. The question of authenticity is of some importance, since XXXII is the only fragment in which the word arete 'excellence' (or 'virtue') occurs, and the only one to character- ize sophronein in a way that goes beyond the Delphic assimilation to self-knowledge in XXIX (D. 116). If XXXII is not his, Heraclitus has nothing really original to say on sophrosyne, the paramount virtue of his age.93 But if it belongs to Heraclitus, XXXII is his most interest- ing utterance as a moral philosopher.

The interpretation depends upon the punctuation, which either falls after the third word, megiste 'greatest', or after the fifth, sophie 'wisdom'. The former punctuation, given in Stobaeus and most modern editions, allows only two words in praise of sophronein and passes on to speak of sophie. The second punctuation, proposed by Bollack-Wismann, turns the whole fragment into an explication of sophronein.

I accept the latter construal, which gives a richer sense. The first clause, then, states that thinking well or soundness of mind (so-phronein) is the greatest excellence (arete) which a man can have and also the greatest wisdom.94 We are a long way from the Homeric conception of arete as the warrior's virtue of valor and skill in combat, but we are not far from the Delphic parallel between sophronein and self-knowledge encountered in XXIX (D. 116).

The conjunction of sophronein with arete, although post-Homeric, is a familiar theme from early Attic grave inscriptions, which regularly speak of the dead as 'brave (agathos) and sensible (sdphron)' or mention his 'courage (arete) and good sense (sophrosyne)'. In the formula for archaic verse epitaphs arete and sophrosyne preserve their Homeric values, the former designating manly courage and the latter prudence or good sense. But Heraclitus has taken over the tra- ditional formula only to transform its meaning. He has in effect rendered the Homeric notion of saophrosyne or good judgment by sophie, so that arete and sophie here represent the two virtues of valor and discretion conjoined in the grave inscriptions; and he has equated both of these with his own conception of sound thinking {so-phronein) modelled on Delphic self-knowledge. Thus Heraclitus does not dissent from praise of valor and good sense; he only insists that these virtues in their highest form are united in thinking well, in salvaging one's thought in the self-knowledge that is also a recognition of what is common to all.

The conception of 'sound thinking' articulated in the first clause by the nouns 'excellence' and 'wisdom', is expressed in the second clause by two infinitives: 'acting and speaking what is true'.96 Speak- ing the truth is not a virtue which the Greeks admired without qualification, nor would it always appear to be the course of prudence. One Homeric example of saophrosyne describes Telemachus' discretion in not revealing his father's true identity but deliberately accepting Odysseus' disguise as an old beggar (Od. XXIII.30). To identify truthful speech as the highest excellence is to take up an uncompromising position on an issue where the traditional Greek attitude was (and is) ambivalent. But this emphasis on 'speaking the true' is what we might expect from the spokesman for a logos 'which is (true) forever'; and it echoes Heraclitus' own promise in his pro- logue to 'tell it like it is'. The terminology of the proem is again suggested by the phrase 'perceiving (or 'noticing' epaiontes) things according to their physis\ (Cf. 1.2: 'distinguishing each according to its nature'.) Accepting the Delphic view of sound thinking as self- knowledge, Heraclitus thus reinterprets it as speech and perception in accordance with the universal logos.

On the punctuation adopted here, the two infinitives 'to speak and to act' belong together: in chiastic order 'acting' will echo arete ('man- liness') just as 'speaking' echoes sophie ('wisdom').97 The difficulty with this natural reading is that it implies the phrase 'to act what is true' (alethea poiein), which is not so easy to understand.98 I have not found a parallel to this expression in Greek. In rhetorical con- texts, a fifth-century author may speak of the 'truth of deeds' in contrast to a distorted appearance or a false and partial report, but what he means by 'truth' here is what really happened; the contrast is conceived from the point of view of the hearer or spectator (who might be deceived), and not from the point of view of the agent.

Since it is not clear what would be meant by doing the true, the reader may be tempted to construe poiein with what follows rather than with what precedes. Heraclitus will then speak not of 'acting the true' but of 'acting according to the nature of things (kata physin), perceptively (epaiontes)\ or perhaps 'perceiving things according to their nature, and acting accordingly'. (So roughly DK and Marcovich, p. 96.) These readings are defensible but not obviously correct, as can be seen from the disagreement among the commentators.

Here again I suggest that Heraclitus has cunningly left us in suspense between two constructions (as in the first clause in I.I), to provoke us into some reflection on what 'acting in accordance with (one's knowledge of) the nature of things' might mean, and how that might relate to the other reading 'to speak and act what is true'.

The solution I propose is to give alethes 'true' its etymological value: 'not concealed', 'not hidden in one's heart'.100 Soundness of thinking will then mean to speak and act the true in the sense of communicating the logos in 'words and deeds' (epea kai erga, 1.2), sharing with others one's perception of how things hang together in unity and are also distinguished in their own nature (physis). The man whose thinking is sound will not hide the truth but signify it in his actions as in his words. In this he will imitate the lord of Delphi, who does not hide the truth but shows it with a sign (XXXIII), and whose lesson to mankind is sophronein.

What is distinctive here is the meaning of self-knowledge as recog-nition of one's true or hidden self, and the connection of this with knowledge of a universal logos which distinguishes things 'according to their nature'. More revolutionary still is the cosmic dimension given to sophronein by the claim (in XXXI) that 'thinking' (phronein) is common to all, if this is taken to mean that all things think. For if phronein characterizes the whole universe of things, then sound thinking means more than keeping one's wits whole; it means to salvage what is common, and hence precious, throughout the universe.

XXXIII (D. 93) The lord whose oracle is in Delphi neither declares nor conceals but gives a sign.

The description does not name Apollo, but identifies the god by rank (anax), function, and locality. (Bollack-Wismann, p. 273, suggest that this circumlocution is itself an instance of the Pythian style.) The god's mode of utterance is said to be neither direct statement (oute legei) nor concealment (oute kryptei) — which would mean falsehood, on the interpretation oi alethea 'truth' suggested above — but 'signifying', giving a sign (alia semainei). There is no doubt that Heraclitus is referring to the Delphic practice of giving advice in indirect form, by imagery, riddle, and ambiguity, so that it was obvious to a man of sense that an oracle required an interpretation.

Even when the surface meaning is clear, it may be necessary to look for a second meaning underneath — as Oedipus discovered to his hurt, when he forgot to ask which man was his father; and likewise Croesus, when he for- got to ask whose kingdom he would destroy. Socrates, on the other hand, shows his wisdom by refusing to take an apparently unambiguous oracle at face value. Instead he asked 'What is the god saying, and what meaning does he intend by this riddle (tipote ainittetai)?'101 'Giving a sign', then, means uttering one thing that in turn signifies another — what the Greeks called hyponoia, a 'hint' or 'allegory'. The sign may be of different types: image, ambiguous wording, or the like.

The Delphic mode of utterance presents a plurality or complex- ity of meaning, so that reflection is required, and unusual insight, if the proper interpretation is to be discovered.
Is the Delphic mode a paradigm for Heraclitus' own riddling style, as readers since antiquity have supposed? Or is the complexity of meaning to be located in the nature of things, in the structure of appearance understood as a logos, a kind of meaningful language?102 There is much to be said for either view, and we need not choose between them.

One can scarcely miss the Delphic elements in Heraclitus' own style. But the notion of men who 'listen without comprehension', who fail to understand because they have 'barbarian souls', is presented as a characterization of the human condition, not merely a description of Heraclitus' own readers or auditors. It is reality itself, or the nature of things, that requires close investigation and readiness to discover the unexpected, 'for it is trackless and un- explored' (VII, D. 18). These two interpretations form the foreground and the background meanings of XXXIII, the obvious sense (applying to Heraclitus' own style) and the hyponoia that emerges only upon reflection (applying to the nature of things understood as logos).

In this way the semantic complexity that is described by XXXIII in reference to Apollo is also illustrated by these words in their reference to our double logos: the discourse of Heraclitus and the structure of reality. And this parallel between Heraclitus' style and the obscurity of the nature of things, between the difficulty of understanding him and the difficulty in human perception, is not arbitrary: to speak plainly about such a subject would be to falsify it in the telling, for no genuine understanding would be communicated. The only hope of 'getting through' to the audience is to puzzle and provoke them into reflection. Hence the only appropriate mode of explanation is allusive and indirect: Heraclitus is consciously and unavoidably 'obscure'.

The point is not that Heraclitus' paradoxical style is designed to mirror the nature of reality, if this is thought of in the Tractatus sense of 'the world', as a structure independent of human under- standing. The paradox lies in any attempt to comprehend and formu- late this structure in human terms: 'opposites are one, and conflict is justice'. It is not that reality as such is contradictory; what is reflected in the semantic difficulty of interpreting these utterances is the epistemic difficulty of grasping such a structure, the cosmic logos, as the underlying unity for our own experience of opposition and contrast.

XXXIV (D. 92) Plutarch; [The Sibyl with raving mouth', according to Heraclitus 'utters things mirthless' and unadorned and unperfumed, and her voice carries through a thousand years because of the god .]

Here again the problem is to identify just what Heraclitus said about the Sibyl, and to specify what her mode of prophecy is a sign for. In both respects we are at a disadvantage by comparison with XXXIII (D. 93), for in that case we can be reasonably certain that Heraclitus' own words have been preserved. In XXXIV Plutarch has so blended the citation into his own text that it is scarcely possible to tell how much is accurate quotation. Some editors would give the whole sentence to Heraclitus; Reinhardt denied him even the mention of the Sibyl! No doubt the truth lies somewhere in between. Heraclitus must have referred to the Sibyl, for otherwise there would be no reason for Plutarch to mention her here, where he is discussing the Delphic Pythia. On the other hand, the calculation of a thousand years is made from Plutarch's point of view, not from that of Heraclitus. Recent editors have agreed that only the words placed within quotation marks above belong to Heraclitus, and the rest to Plutarch.105 But Plutarch's paraphrase may contain further traces of Heraclitus' thought.

Heraclitus is likely to have intended some reference to divine possession or seizure, since the Sibyl is precisely the type of prophetess who, like Cassandra in Aeschylus' Agamemnon, bursts into agonized utterance of a vision in her state as entheos, invaded by Apollo.106 If her utterances are 'mirthless' or 'gloomy' (agelasta), that is not only because her typical vision is of a coming disaster, but because the Sibylline experience of enthousiasmos is itself a form of suffering, a kind of spiritual rape.

This is the earliest reference to the Sibyl in extant literature. It is not clear whether the Sibyl was originally thought of as a unique individual or as a type, whether the word 'Sibyl' is a proper name or a common noun.107 Later authors spoke of an eighth-century Sibyl named Herophile, from the Ionian city of Erythrae not far from Ephesus, whose oracular wanderings had left a clear trace at Delphi.108 There is no reason to suppose that the story of Herophile at Delphi need be as old as Heraclitus; but long before his time the institution of ecstatic prophecy had been established at Delphi in the person of a regular priestess, the Pythia. We may think of the Pythia as a kind of sedentary Sibyl, a holy woman who has become a fixed appendage to a particular shrine.

Whatever the historical connections may have been between the Pythia and the Sibyl, they represent prophetic phenomena of the same type. In contrast to oracles by signs (from rustling oak leaves, the flight of birds, the liver of the sacrificial vic- tim) and in contrast also to oracle by lots, both the Pythia and the Sibyl are women who serve as 'mediums', who see the future in a trance experience, in the superhuman vision of the god. It is the woman's voice that speaks, but it is Apollo's word that is uttered.

Because of this religious parallel between Pythia and Sibyl, as the same kind of mouthpiece for the same god, it is natural to suppose that Heraclitus himself intended some connection between the oracles of the Delphic god in XXXIII and those of his Ionian prophetess in XXXIV. The meaning of XXXIV would then be that just as the Sibyl's power comes not from herself but from the god, so Heraclitus' authority is derived not from his own person or opinions but from the cosmic logos in whose name he speaks: 'listen not to me but to the logos9 (XXXVI, D. 50).
This view of XXXIV previously seemed attractive to me, and might just possibly be correct. But the text which is most unmistakably genuine is the description of the Sibyl as prophesying 'with raving mouth'.

And from what we know about Heraclitus' attitude to madness in CXVI (D. 15) and CXVII (D. 5), as well as his contempt for drunkenness in CVI (D. 117), it is most unlikely that he would have cast himself in the role of a prophet whose insight was to be com- pared to the ecstasy of a woman possessed. It was this correct per- ception of Heraclitus' deep distaste for Sibylline frenzy which led Reinhardt to doubt the authenticity of the whole fragment, since Plutarch presents the quotation from Heraclitus as part of a positive evaluation.110 But we can accept Plutarch's quotation while rejecting his evaluation. In its original context, the contrast between the Sibyl's madness and her social prestige as spokesman for Apollo may have been part of Heraclitus' critique of current religious practices, along the lines of CXV-CXVII (D. 14, 15, D. 5): if one reflects upon the fact that people accept wisdom from raving lips, one is likely to judge them as mad as the Sibyl herself.

XXXV (D. 45) You will not find out the limits of the soul (psyche) by going, even if you travel over every way, so deep is its report (logos).

'Heraclitus is the first to have given serious thought to, and had some- thing to say about, the soul in man' (Der Glaube der Hellenen I (Berlin, 1931), 375). Wilamowitz's point was that the psyche in Homer is mentioned only when it leaves the body, in death or in syn- cope: it is the life-breath or animating 'spirit' which departs as a ghost. So even for Pythagoras the psyche, although deathless, belongs only temporarily and as it were on loan to the individual man. But with the fuller development of the notion of personality and a more intimate sense of identity between a man (or woman) and his (or her) emotional life that found expression in the lyric poetry of the seventh and sixth centuries, the time was ripe for a new conception of the
psyche as source or center of the human personality. And so we find in Heraclitus, as Reinhardt said, 'for the first time a psychology worthy of the name'.

But if there is wide agreement on the originality of Heraclitus' view of the psyche, there is less agreement on just what that view is. Some authors give physical accounts of Heraclitus' doctrine which equate the psyche with cosmic fire, or one of its privileged manifestations; but Bruno Snell has proposed an interpretation which emphasizes the psychic as such, the direct intensity of thought or feeling denoted by 'Inner lichke it' (inwardness) or 'self-awareness' (Entdeckung des Geistes, p. 37).

There is a grain of truth here: the philosopher who said 'I went in search of myself, is one who has reflected upon the nature of his psyche in terms of his living experience. But the fragments show no trace of a Cartesian or Platonic contrast between mind and matter, soul and body.
We have seen that the psyche in XVI is identified as the cognitive or rational element in human beings, and this intellectual conception of the psyche must be emphasized here, since it has been overlooked in several influential studies where the originality of Socrates in this respect has been grossly overstated.

Socrates' position in the history of philosophy is secure enough without attributing to him a revol- utionary new concept of the psyche as a cognitive principle — a concept which he might have got directly from Heraclitus, but which was probably 'in the air' in fifth-century thought and usage. The concept was a new one, and only after Plato did it come to dominate the earlier view of the psyche as essentially biological, emotional, or non- rational. (The older view survives in Aristotle's conception of the vegetative or nutritive soul, as in Plato's view of the 'irrational' parts of the psyche.)

Heraclitus' original readers might have expected an account of the psyche rather more like Pindar's description of 'the image of vitality' which 'sleeps when the limbs are active, but shows to sleeping men in many dreams' a sign of future events (fr. 116 Bowra), or the more mundane view of Semonides, who urges his friend to endure the brevity of life 'by indulging your psyche with whatever good things you can get' (fr. 29.13 Diehl, assigned by some to Simonides). In the former case the psyche is thought of as a power of establishing contact with the supernatural in dreams; in the latter case it takes delight in sensual pleasures and 'the good things of life'. Both ideas will be exploited by Heraclitus elsewhere.
(For the dream experience see XC, D. 26; for the soul in debauchery, see CVI, D. 117 and the more dubious CVIII, D. 77.) But in his own view the psyche is primarily a principle of rational cognition.

Since Heraclitus is a monist, however, the psyche is also a physical principle. We shall see in commenting on CII (D. 36)-CIX (D. 118) that he identifies the soul not with fire (as is often supposed) but with atmospheric vapor, air or 'exhalation', in the traditional Greek manner.

In the text of XXXV four points call for comment: (1) the 'limits' of the soul, (2) the reference to 'finding out', (3) the 'way' (hodos) to be travelled, and (4) the logos which the soul has.


(1) On first reading, the mention of 'limits' (peirata) to which one travels might recall the Homeric formula for a voyage 'to the ends of the earth' (//. XIV.200 =301, etc.). So Hesiod speaks of a place where 'the sources and limits (peirataY of earth, Tartarus, sea and heaven are located together (Theogony 738 = 809). But Heraclitus denies that anyone can reach the limits of the psyche, no matter how far they travel. On second reading, then, this will suggest that the soul is limitless, apeiron in a truer sense than the 'boundless earth'. A reader acquainted with Milesian philosophy may recognize an echo of the Limitless {apeiron) of Anaximander or the limitless Air of Anaximenes: the denial of limits involves an allusion to the supreme principle of cosmic structure.

Thus the psyche for Heraclitus plays the role of a 'first principle', a Milesian arche; and hence the natural tendency of modern interpret- ers to identify it with cosmic fire. They are right in principle though wrong in fact. Although the psyche as such is not composed of fire, psyche and fire are alike in playing a double role in Heraclitus' thought. Each one represents a specific physical phenomenon, one elemental form among others. (Thus XXXVII, D. 30 will speak of
fire 'going out' — and hence ceasing to be fire —just as CII, D. 36 will speak of 'the death of psyche' in the birth of water.)

At the same time, both psyche and fire are in some sense universal, all-inclusive. Thus Fire is exchanged for all things, and all things for fire (XL, D. 90). The universality of psyche is expressed here in the denial of limits: wherever you travel, there psyche will be. This is what we must expect, since XXXI (D. 113) was interpreted as a statement of panpsychism. If all things think, then all things are alive. But if all things are alive (empsycha), then all must have a share in psyche, the life-principle. The soul can travel everywhere, since everything is psychic territory. Perhaps the underlying assumption is that funda- mental discontinuity, the emergence of life from non-life, is irrational and unnatural. The doctrine of Thales, that 'all things are full of gods', which Aristotle takes to mean that psyche is mingled throughout the universe (De Anima 1.5, 411a7 = DK 11.A 22), could well apply to Heraclitus: all things are full of soul.


(2) The limitlessness of the psyche is expressed by a phrase that has echoes elsewhere in the fragments: 'you will not find them out'.114 The verb 'find out' (exheurisko) occurs also in VII (D. 18): he who does not expect the unexpected will not find it out (cf. VIII, D. 22, on 'finding little' gold). In LXXXIV (D. 27) 'the unexpected' will be the fate of men after death. In XLIV (D. 94) the Furies as ministers of Justice (Dike) 'will find out' the Sun, if he oversteps his measures. In LXXXVII (D. 28B) it is Dike herself who will lay hands on liars, as in CXXI (D. 66) fire will 'lay hands on all things'. In these last three fragments, the notions of finding out and catching hold of pretty nearly coincide. Thus the usage of this verb elsewhere in Heraclitus suggests that the limitlessness of the soul is closely connected both with the mystery of death and with notions of cosmic order and personal punishment for wrongdoing. These suggestions are too elusive for us to understand or explicate within the present context. The verbal echoes simply mark this as an instance of proleptic statement, a hint of themes more fully developed elsewhere.


(3) 'By going (on foot), by travelling over every way.' It is natural to think here of 'the way up and down, one and the same' mentioned in CM (D. 60), and perhaps also of the carding instrument in LXXIV (D. 59), whose path is 'straight and crooked', since the term hodos ('way') appears in each case. But the fragment on the carding wheels is too problematic to help us here. I see only two plausible readings for the way travelled in seeking the limits of soul, (i) It can be a very general image for a search that proceeds in vain: 'I looked every- where', (ii) It can allude to the course of elemental transformation, with 'the way down' of CM interpreted as a change to water and earth, 'the way up' as a change to fire.115 The first reading corresponds to the surface meaning of the text; (ii) suggests a cosmic hyponoia.


(4) Finally, what is the deep logos which the soul 'has', and which explains our inability to find its end points? On first reading, logon echei should mean something like 'it has something to say', 'it has the right (or the capacity) to speak'.116 Now the possession of rational speech may be a significant overtone here, but this idea can- not explain the adjective 'deep'.
Most authors have rightly assumed that logos in XXXV must mean 'measure', as in XXXIX (D. 31B). The limitlessness of the psyche is then to be understood in terms of impenetrable depth. Soundings are in vain, for no plummet-line will get to the bottom of it. Cf. Democritus fr. 117: 'we know nothing truly, for truth lies at the bottom (en bythoi)\ It takes a deep diver to discover it, as Socrates is supposed to have said about the meaning of Heraclitus' book.117 So for the soul's logos: it is vast, subtle and deep; and intelligence — not travelling about — is required to find it out.


A logos so profound and limitless can scarcely be distinct from the universal logos, according to which all things come to pass. The sol- ution to XXXV thus gives us the fuller explanation for XXVIII (D. 101): by seeking for his own self Heraclitus could find the identity of the universe, for the logos of the soul goes so deep that it coincides with the logos that structures everything in the world. Hence the error of those men who treat thinking as private, in the face of the
fact that 'the logos is common' — common to them and to everything else (III, D. 2).

XXXVI (D. 50) It is wise, listening not to me but to the report (logos), to agree that all things are one.

It seems likely that this sentence, one of the weightiest of all, came at the end of the introductory section when Heraclitus returns (by a kind of ring composition) to the theme of the logos with which he began. The following points call for comment:

(1) the contrast between 'listening to me' and 'listening to the logos,
(2) the word play on homologein 'say in agreement',
(3) the definition of wisdom, and,
(4) the final formula 'all things are one'.

(1) The reference to a logos somehow independent of Heraclitus will be immediately clear if he has just spoken of the 'deep logos' of the soul. The thought will be: listen not to me but to the discourse within your soul, and it will tell you all. If we set aside XXXV (D. 45) and consider only 1.2 and III (D. 2), listening to the logos will imply the conception of the world order as a meaningful language which one hears with more or less comprehension. Heraclitus is after all a prophetes ('spokesman') for the logos: his words are an attempt to make this larger discourse audible to a few, at least, among the many who seem deaf.

(2) This thought is developed in homo-logein 'to agree', which carries a phonetic echo of logos and an etymological sense of 'speak- ing together with, saying the same thing'. Knowing how to listen will enable one to speak intelligently (XVII, D. 19), to hold fast to what is common (XXX, D. 114), by speaking in agreement with the uni- versal logos of 1.2.

(3) Wisdom or what is wise (sophon) consists in just this fitting of the private to the public, the personal to the universal. (See above on xynos logos in III.) By its rational structure and its public function in bringing men into a community, language becomes a symbol for the unifying structure of the world which wisdom apprehends.

(4) The sentence which began with a paradoxical contrast between two things that seem identical — listening to Heraclitus and listening to (his) logos — culminates in a greater paradox of identification: all things are one. The sense of tension, if not contradiction, is all the greater because the unity of all things is here the content that wisdom will put itself in agreement with, whereas in XXVII (D. 108) 'what is wise' is said to be 'set apart from all (things)\panton kechorismenon.

This tension between the themes of isolation and community will be fully resolved only in the context of other fragments. (See on LIV, D. 41 and CXVIII, D. 32.) The unity of opposites and the community of the logos (with its triple application to discourse, soul, and uni- verse) provide the initial clues for interpreting this extraordinary claim, whose full meaning requires an understanding of Heraclitus' thought as a whole. In that sense, the rest of our commentary will be an exegesis of this proposition: hen panta einai.

This is the earliest extant statement of systematic monism, and probably the first such statement ever made in Greece. In textbooks on the history of philosophy, the Milesian thinkers are represented as monists who reduce all things to water, to air, or to the Unlimited. But this view, which ultimately rests upon Aristotle's interpretation of their doctrine in terms of his own concept of the material cause in Metaphysics A 3, is anachronistic and misleading, insofar as it imputes to the earliest Greek naturalists a post-Parmenidean notion of some true unity underlying the apparent plurality of things, even when it does not actually assign to them Aristotle's own characteristic con- cept of the underlying material substratum of change (hypokeimenon). There is no good evidence for assigning monism to the Milesians, if this is understood as the claim that 'all things are water' or 'all things are air', that (as Diogenes of Apollonia puts it) earth and sky and sea and air are not really 'different in their own physis' but are at bottom one and the same thing.

But the Milesian philosophers were monists in a different sense, and their theories must have provided the background for Heraclitus' thesis. What he could draw from them was the double claim

(i) that all things are derived from a single arche or starting point, and,

(ii) that as now constituted all things are organized within a single world structure or kosmos. And we may add,

(iii) that Anaximander surely, and Anaximenes probably, thought of the initial principle not only as encompassing (periechein), and thus physically unifying the world, but also as 'steering' and governing it by imposing a rational structure.

Aspects (i) and (ii) of Milesian monism seem to be reflected in XXXVII (D. 30); the aspect of cosmic guidance (iii) will emerge in LIV (D. 41)."

XXXVII (D. 30) The ordering (kosmos), the same for all, no god nor man has made, but it ever was and is and will be: fire everliving, kindled in measures and in measures going out.

This text, with the two that follow (D. 31A and 31B), gives us our primary information on Heraclitus' cosmological thought and the most natural interpretation of his claim that all things are one.

But all three texts are surrounded by thorny problems. What is the relationship between Heraclitus' doctrine and the cosmological theories of his Milesian predecessors? In what sense can a world order be identified with 'everliving fire'? And is Heraclitus here denying the general assumption of a development of the world from some more primitive source?
In the fragments as arranged here, we have had no reference to physical theory (though there is some hint of this in the repeated occurrence of the term physis), no mention of the Milesian cosmologists as such (since Thales was apparently paired with Homer, as Pythagoras, Xenophanes and Hecataeus are named with Hesiod), and no reference to cosmology. (The only occurrence so far of the term kosmos is in the dubious 'fragment' VI, D. 89.) A reader encountering XXXVII for the first time might well begin by taking kosmos in its normal literary sense of 'good order', 'adornment', as in the brilliant toilette of a woman of fashion or in the impressive array of disciplined troops.

A 'kosmos, the same for all' may then be understood either as

(1) a moral or political order, applicable to all men, like the 'divine one' by which all human laws are nourished in XXX (D. 114); or,

(2) an ornament like jewelry, fine clothing, or a work of art. In the long run this naive reading of XXXVII cannot succeed. But the older, literary sense of kosmos is a natural starting point, since Heraclitus does not write in a technical language and the new 'cosmological' sense of kosmos is itself the heir to all these older nuances of the word. The technical notion of 'world order' will emerge only as a hyponoia brought to the surface by reflection, once it is seen that no good sense can be made of the idea of 'an adornment the same for all'; and if 'a social order the same for all (men)' does make sense, it is not easy to understand how such an order could be identified with eternal fire. It is the formula for eternity ('ever was and is and will be') and the mention of fire that will force us onto the terrain of natural phil- osophy.

Before this move to the technical sense of kosmos is required, we are told that 'no god nor man has made the kosmos, the same for all'. Scholars have scratched their heads over this denial that any human being has made the kosmos (why should anybody suppose that?) and have generally dismissed these words as a so-called polar expression, as if 'neither man nor god' meant simply 'no one at all'. But even the notion that a. god has made the world is poorly attested in Greece before Plato's Timaeus. The whole problem here is an artificial one, created by the mistaken assumption (in turn supported by the mis- taken reading tonde) that kosmos must from the beginning have its technical sense of 'world order'. For of course if kosmos means 'moral (or social) order', or if it means 'ornament', then we naturally suppose it to be the work of an orderer, whether human or divine.

Thus the sons of Atreus are regularly referred to in the Iliad as 'orderers of the host' (kosmetore laon). On our first, naive reading of XXXVII we see very clearly what Heraclitus is denying, though we cannot immediately understand why. We are faced with two para- doxes: an array that is not local or particular but 'the same for all'; and an instance of order without an orderer, like a disciplined host without a commander, a law without a lawgiver, or a work of art without an artisan. The first paradox is resolved by a shift to the phil- osophical sense of kosmos. The world order is naturally 'common', the same for all men and for all things: that is just what is meant by a world order. But how can we have an ordered world without a power to set it in order?

By denying that this order is a work of art, Heraclitus implies that it is a work of nature: self-made or self-grown. Thus the cosmological idea begins to emerge, and becomes explicit when the kosmos itself is invested with the attributes of divinity: 'it ever was and is and will be'. This is the thought echoed in a famous fragment of Euripides:

the student of Ionian science (historia) beholds 'the ageless order (kosmos) of undying nature (physis)' (fr. 910 Nauck = DK 59.A 30). The new philosophical paradox of XXXVII is a denial of any funda- mental duality between a generated world order and the eternal source from which it arises or the ruling intelligence by which it is organized. Insofar as the kosmos is made, it is self-made; insofar as it is organized, it is self-organized; insofar as it is generated, it is identical with its own eternal source, everliving fire.

XXXVII is built up by wave upon wave of paradox. If the initial problems are resolved by taking kosmos in the sense of 'world order' and by the identification of this order with fire, that identification is itself paradoxical, whether we think of fire as an element within the world — for in that case the whole is identified with one of its parts — or whether we think of it as some primordial or transcendent power, as the emphasis on its eternal being would suggest. (For then the world is identified with something transmundane!) But the culminating paradox is provided by the last two participles, when this principle of cosmic eternity is said to be regularly rekindled and regularly going out. What sense can we make of an eternal bonfire going out? And what are the measures according to which it is kindled and extinguished?

So far I have avoided the shoals of controversy, but that is no longer possible. For the question of the measures of fire going out is just the question of whether or not Heraclitus envisaged the world as gradually taking shape from (and eventually reabsorbed into) primor- dial fire, as the Stoics did after him and in his name. And this double question — the issues of cosmogony and cosmic conflagration, which hang together — stands at the storm center of scholarly dispute. If the Stoics were not actually following Heraclitus, they were certainly following the view indicated by Aristotle and presented in detail by Theophrastus:

'The kosmos is generated from fire and is ignited again according to certain periods alternating through all eternity' (D.L. IX.8 = DK 22.A1; see Appendix IIA).

Theophrastus' account was accepted in antiquity by all later writers, and by Zeller, Diels, and most modern scholars (though not by Burnet) until the publication of Reinhardt's book on Parmenides in 1916. Since then the tide has turned, and I find myself almost alone today in suggesting that, after all, Theophrastus and the Stoics understood Heraclitus correctly on this point. Although a strong basis was laid by Burnet's presentation of the case, the great success of Reinhardt's argument is due to a new and important insight: not only do the extant fragments not present any detailed statement of cosmogony, but there is good reason to doubt that any lost fragments were more explicit. The doxography of Theophrastus, and the Stoic interpretation that is built upon it, show every sign of relying precisely upon those fragments whose original text has been preserved. So the ancient interpretation has no inde- pendent authority.On this point modern scholarship is unanimous. Whatever interpretation we offer must be based upon Heraclitus' words alone, together with whatever we can know of their historical context.

Once this point has been admitted, I believe that the recent denial of cosmogony for Heraclitus will turn out to be a temporary over-reaction, an exaggerated by-product of our emancipation from the authority of the Stoic and doxographical interpretations. If we read Heraclitus' words with an ear for their rich allusiveness, we will find that they do not contain a dogmatic denial of cosmogony any more than they contain a full statement of that process. On the contrary, we will find that XXXVII and a dozen other texts are best understood as presupposing rather than as denying a genetic account of the order of nature, and as playing fruitfully with the notion that the world will one day go up in smoke. On this point — the primary or pre-eminent sense of 'kindled in measures and in measures going out' — the Stoics did not misunderstand Heraclitus; they distorted his cosmic speculation only by transforming a subtle, poetic vision of the cosmic pro- cess into a rigid orthodoxy.

This argument will be pursued in the next sections of the com- mentary. But I do not share the common view that the question of cosmogony and ecpyrosis is decisive for the understanding of Heraclitus. If we eliminate this cycle of world formation and destruc- tion from his system, the vision of nature will be lacking in complete- ness and in symmetry, but it will still be essentially the same vision. For the pattern of natural law is the same for macrocosm and for microcosm, for the origins of heaven and earth and their present pat- tern of transformation: 'kindled in measures and in measures going out' applies to all of these.
The great cosmic cycle is only the ordinary cycle of natural change and human life writ large. What is crucial is not the debate about a particular doctrine but the recog-nition of the kind of discourse which Heraclitus presents, and which separates him from the natural philosophers like Anaxagoras and Diogenes. This is what Theophrastus and (to a lesser extent) the Stoics failed to understand, and what Heraclitus himself realized was so difficult to express. Once we understand his ironical self-distancing from technical cosmology, and his reinterpretation of all other con- ceptions in terms of an understanding of human life and death, his lofty acceptance-but-also-denial of a cosmic cycle will appear as a natural consequence of his general attitude to Ionian physics.

Before returning to the exegesis of XXXVII, I want to repeat one very general argument against the currently predominant view, which holds that by saying the kosmos 'forever was and is and will be, ever- living fire' Heraclitus meant literally and unambiguously to deny that the world had emerged from some earlier and simpler state of affairs. It seems that little thought has been given to how strange, almost unintelligible, would be the dogmatic rejection of cosmo- genesis by an archaic thinker. The instinct to explain things by telling how they began and how they developed is not only at the basis of all mythic thought; it also dominates all scientific or philosophic speculation down to and including Plato's Timaeus.

'The principle of cosmogony was rejected by no one before Aristotle, not even by Parmenides, and it has perhaps been rejected by no one since, except under Aristotelian influence. The scientists who write on "the birth of the solar system" are only giving us the latest version of the cre- ation story.' Aristotle alone broke with this millenial tradition, and he had a strong motive for doing so. He had abandoned Plato's realm of imperishable Forms but not Plato's belief that scientific knowledge requires a fixed and unchanging object. Hence it was
of the greatest importance for him to find an equivalent pattern of eternal stability within the structure of the natural world. What is lacking in the case of Heraclitus is any comparable philosophical motive for espousing such a rare and radical heresy, more than a century before Plato's Timaeus.

To return to the identification of kosmos and fire: what does this mean, and why has Heraclitus chosen fire? As indicated in the Intro- duction, the nearly contemporary theory of Anaximenes provides the historical background against which Heraclitus' own cosmo- logical monism is to be understood.125 Our best account of the doctrine of Anaximenes comes from Theophrastus, and it seems to con- tain a fairly close paraphrase of Anaximenes' own words:

He said that the first principle was limitless Air (aer), from which arises what comes into being, what has become, and what will be, and gods and things divine; but other things arise from its off- spring. The form of the air is as follows: when it is most uniform it is invisible [sc. as atmospheric air] ; but it is made manifest by cold and heat and moisture and motion. It moves continually; for it would not change as much as it does if it were not in motion. As it thickens or rarefies it appears as different. For when it spreads out into rarer form it becomes fire; winds on the other hand are air as it thickens; from air cloud is produced by compression; and water by still more compression; when further thickened it be- comes earth and in its thickest form stones. (Theophrastus in the excerpt of Hippolytus, DK 13.A 7)

The rest of the doxography, and parallels from Anaximander and from Anaxagoras (especially frs. 15 and 16), make clear that this sixth-century monism must be understood in the context of a cos- mogony: if the boundless Air is taken as arche or starting point, that is just because Anaximenes believes he can explain how, beginning with the nature of air alone, the whole diversity of the world and its structure has evolved, by thickening and thinning in connection with cooling and heating. In this way all things arise from the Air — either directly or 'from its offspring', as the doxography reports.

This genealogical derivation of the world from a single ancestor adapts a pattern that is as old as Hesiod's Theogony and will remain standard in the tradition of natural philosophy down to Plato's Timaeus (where the Demiurge is called 'father' of the created world at 37C7, but the Forms are father at 50D3, where the Receptacle is mother and the world of becoming is itself 'offspring', engonos).
In XXXVIII and XXXIX (D. 31A-B) Heraclitus will exploit some aspects of this pattern — and again in LXXXIII (D. 53), where War is called 'father of all things'. But in XXXVII he directly rejects this pattern by insisting upon the eternal pre-existence of the world order as everliving fire.127 This break with the Milesian scheme has the effect of identifying the world with its eternal source or arche, the cosmic order with its divine helmsman or regulator. This is monism with a vengeance. But why is it fire that is selected to represent the 'one'?

A recent writer on the history of science has spoken of 'the air of magic that boils out of the fire: the alchemical feeling that substances can be changed in unpredictable ways. This is the numinous quality that seems to make fire a source of life and a living thing to carry us into a hidden underworld within the material world.' Fire is indeed a mysterious symbol of life, of superhuman life — despite or because of the fact that it is the one element in which no animal can live, and a power that in ancient Greece (as in modern India) often served to receive human bodies at death.

Thus in representing life and creativity it also represents death and destruction. As an altar flame consuming the sacrifice, it represents the gods. As fire for cooking and for warmth in winter it sustains human life. As instru- ment of the arts, the stolen gift of Prometheus, it points to the divine element in human activity, the techniques and industry that separate us from the animals. Fire has many qualities. But it is a most unlikely choice for a starting point in a literal account of the development of the world in material terms, since it is not itself a kind of matter, not a body at all, but a process of transition from one state to another, a symbol of life and death at once, the very element of paradox.

These are some of the thoughts which Heraclitus' choice of fire has imposed upon the pattern of cosmic transformation taken over from Anaximander and Anaximenes. He takes a physical theory as the background against which his words are to be understood (once we have been led to interpret kosmos as 'world order'); but his utterance is not itself the statement of a physical theory. Instead, the paradoxical denial that the kosmos has any origin or history at all is redoubled by the description of an everliving fire that is always going out.

The error of recent interpreters has been to deprive this second paradox of its sting by refusing to take the words literally, reading them as a poetic reference to elemental transformations, while con- struing the first paradox as a literal statement of doctrine. If we take both statements at face value they indicate that the everliving fire could equally well be described as ever dying, that it is wholly transi- tory and always changing, while remaining eternally the same for all.

In order to unravel these puzzles, we must know more about the measures by which fire is kindled and put out. One clue is provided in the 'turnings' of fire and the measures of sea in XXXVIII—XXXIX; another will be found in the measures of the sun in XLIV (D. 94).