IN Greece the long process, begun by Homer, of redefining the relations between myth and language could only continue by rejecting him. However sophisticated we may find Homer he makes little or no distinction between “fact” and “fiction,” “abstract” and “concrete,” or even between “general” and “particular.” For the Greece Homer dominated, if one asks “What is history?” the answer is Homer; “What is poetry?” Homer again, and those like him, his followers, the Homeridae. And also, to “What is philosophy?’ the answer is still Homer. Though “religion” remains (but Homer also has a scriptural tinge), archaic Greece notably provided an impetus still unexhausted for redefining myth in relation to language by separating off poetry from philosophy and from history. The possibility of history, and philosophy, as for developing poetry, emerges in an exemplary way as Greece tests for itself the definitions of myth.
The values codified in Homer and his place in Greek tradition make him for all practical purposes the sort of tribal figure against whom Havelock asserts Plato to have had to exercise his major revisionist force. Simultaneous homage and refutation are to be found as early as Xenophanes, who says that all men have learned with Homer (kath’ Homer on) from the beginning (Diels 21 B9) but directs the brunt of his own revising attack against Homer.
Heraclitus’ radical revision of the relation between myth and language in Homer not only alters the approach to myth but recasts the structures of the language which is used to handle the questions he asks, where Xenophanes had stayed within the epic and elegiac meters. Heraclitus recasts literary form, continues the process of releasing formal expression from the tribal bonds of meter, challenges myth, discovers a powerful tool of philosophy in applying the principle of contradiction, and redefines the daimon in myth while vitalizing the bases of language, all in one verbal act. His immediate predecessors had provided a basis for questioning myth—Xenophanes by challenging a blind acceptance of the Homeric gods, and Anaximander by abstracting the cyclic time of a Homeric world-view into a principle of compensatory justice, “justice . . . payment . . . according to the arrangement of time” (“diken . . . tisin . . . kata ten tou chronou taxin” [Diels 12 B.l]).
Philosophy tends to abolish myth from its domain. But in activating itself philosophy also oddly activates myth; it reestablishes and revitalizes the relationship between myth and language in the very act of questioning language or myth. As Pavese said, “The peoples which have had a rich mythology are the peoples who have then avidly philosophized: Indians, Greeks, Germans.1 And the last person in the world to do so had made a similar connection: Aristotle. The principle of wondering inquiry to which he attributes the inception of philosophy is said to serve equally for myth: “Therefore even he who loves myth is sort of a philosopher too; for myth is composed of wonders.” (“dio kai ho philomythos philosophos põs estin, ho gar mythos synkeitai ek thaumasiõn.”Metaphysics 982 В 17-19).
A tribal group-domination, both one phase of myth and a permanent tendency for mythic systems, already loses its grip on the verbal expression of the lyric poets. They are dedicated to the primacy of personal pleasure (Sappho, Alcaeus) and to a private perception that in Aleman could still also lend itself to the cult hymn. The fusion of the private and the cultic achieved a full elaboration in Pindar and Bacchylides, without breaking the hold of the tribal group, whereas the much later artifice hymns of Callimachus are already at another stage. They center on gods as different from the Aphrodite of Empedocles or even the Venus of Lucretius as these are from the Venus of Shakespeare. The discursive lyric stayed close to the tribal code in Solon’s presentation of proverbial principles and Theognis’ more expansive series of advisory aphorisms.
The codification of myth, first in Hesiod, and later, more scientifically, in Pherecydes, indicates anxiety about its stability. By the time of Plato the word theologia has been invented,2 and attention is focussed on what may be deduced directly from the unifying idea of a single divine originator. Much later, in the work of Apollodorus and Plutarch, the impulse to codify myth will have become exhaustive and detached from the force of myth.
Pythagoras certainly did found a cult, one based on correspondences between the most abstract translinguistic systems, mathematics and music. A calculating process, to which Lévi-Strauss has given a quasimathematical form, underlies mythic systems, and music characteristically accompanies public rituals of homage to mythic beings. In Pythagoras all the mythic figures themselves have disappeared, it would seem, but mathematics and music, each defined in terms of the other, in effect replace them and become the basis for a cult. The cult’s practice carried along with it the dietary stipulations that, again according to Lévi-Strauss, characteristically accompany and reflect mythic systems.
In Pythagoras’ new abstract system, the parts of the world are defined and ordered according to an assigned number, itself placed in a system, the Pythagorean triangle of ten. This is formed by placing one dot, then two, then three, then four, adding up to ten. The triangle takes over many of the functions of the Olympian system. The presence of music, and the emphasis on it, attests to the dominant power of the oral-aural.
Mathematics, if applied to music, identifies the oral and the visual. And visible entities are what the mathematics of Pythagoras is taken to plot, actual geometric figures in space to which numbers are assigned, as for the “Pythagorean theorem.” Greek mathematics was not to sever itself from perceptible figures until more than a generation later, and the visual itself remained a puzzle for Greek philosophers after Pythagoras, from Heraclitus through Plato and Aristotle.3
The abstractions of Pythagoras remained close to hearing and sight. Pythagoras is compared to Homer as a cult-guide (Plato, Republic 600 a, b), looking backward; looking ahead, the Pythagorean circle was often compared to Plato’s Academy (Burkert 1, p. 74). Plato speaks of the opposition in Pythagoras, like that in Homer, between the public (demosia) and the private (idia), an opposition which is not, however, schematized in Homer.
A similar tension obtains in Heraclitus, accompanying a similar transposition of mythic systems, one that alters their content while attempting to retain their power. Heraclitus regards a private consciousness (idian phronesin, 22 В 2) as an illusion: but in the very act of asserting the community of the logos he divorces himself from the tribal community of persons: it is just the many (hoi polloi) who make this error about the private consciousness.
The protracted effort to derive a cosmology from what we have of Heraclitus’ statements has resulted, I should like to demonstrate, not so much in error, which would be a strange consequence of so much intricate and intelligent thought.4 Rather, it has resulted in a misplacement of emphasis so considerable as to obscure the philosophical character and direction of utterances that are nothing if not pointed. If Heraclitus is already obscure, skoteinos5 in the attribution of antiquity, it would not help to further obscure the thrust of the kind of statement he makes. To do so would, in fact, obscure the conditions of his obscurity.
First of all, how did Heraclitus see himself with reference to the cosmological and other concerns of his contemporaries and predecessors? We have, as it happens, a number of his statements about them, considerably more than we have from any other pre-Socratic philosopher. That alone should alert us to Heraclitus’ concern to distinguish himself from them, not so much by refining their doctrines, though he may be said to refine Anaximander and Anaximenes if not others anonymously.6
Every one of Heraclitus’ explicit statements about writers involves not the rejection of a specific doctrine but, arguably, the attribution of a total wrongheadedness of approach. Of these statements the most comprehensive is:
Learning many things does not teach thought, for it would have taught Hesiod and Pythagoras, and also Xenophanes and Hecataeus. (Diels В 40)
The sense of manthano (“learning”) in polymathie (“learning many things”) must be active7—the acquisition of concerted bodies of intellectual skill, and not the miscellaneous information which, for one who has noos (“thought”), would be a necessary raw material for philosophoi. The philosophoi are to inquire about many things quite thoroughly (eu mala):
Those men who love wisdom must be thorough inquirers into many things. (B 35)
The four thinkers mentioned in В 40 have not been taught by their polymathie to have noos, presumably because they have stopped inquiry (B 35, historas) short of that point of unity which for Heraclitus is equivalent to wakefulness and to intelligence—to understanding that the logos is xunos (В 1).
Bollack (ad loc) asks what sets these particular four writers into relationship. But first we need to know in what respect each of them may be said to have polymathie and yet to lack noos.
For Hesiod, the broad classification of his extant work into theology and husbandry provides such a split. The distinction between men and gods, so important to Heraclitus, would then not be correctly perceived along Hesiod’s lines because the principle that unifies men and gods would not have been discovered. So that even within the Theogony Hesiod would be in error, as В 57 tells us, the word didaskalos recalling the ou didaskei of В 40:
Hesiod is a teacher of the greatest number. They are convinced this man knew a number of things, who did not understand day and night; for they are ors. (В 57)
This statement must be taken as a criticism of the theological-philosophical adequacy of the Theogony, where we hear of the origins of day and night. Hesiod has not sought the principle of unity underlying what men have taken as entities distinct enough to be given the different names “night” and “day.”
It is just such a doctrine of seeming unity that Pythagoras, the second figure in the list, is reputed to have offered, and there would seem to be no definite areas into which one might divide his polymathie. Music and numbers, however, are obvious candidates, and even though Pythagoras’ system would seem to have defined each in terms of the other, Heraclitus’ assertion would have to be taken as undercutting Pythagoras’ claim of their unity. Such is the force of his other statements about Pythagoras. The longest of these, В 129, begins with seeming praise and descends to a pejorative anticlimax, the penultimate word, again, being polymathie:
Pythagoras, son of Mnesarchos, pursued inquiry most of all men, and having chosen out these writings made them his wisdom, his much-learning, his bad skill. (B 129)
Historie is the desired activity of the philosophoi in В 35. “All men” is a double-edged expression if this context of praise is correlated with Heraclitus’ other statements about collective humanity. The act of choosing (eklexamenos) gave Pythagoras a skill-wisdom which was merely private;heautou sophien, in a pejorative context, is pretty much equivalent to the idian phronesin, “private perception,” that the hoi polloi mistakenly suppose they can rely on (B 2).
It is the very claim to unity, when in fact his particularity is finally to be defined as bad skill, kakotechnie, that must make Pythagoras the “tutelary chief of liars” (B 81, in which the term archegos carries some suggestion of religion and hence of his cult,8 and the rare word kopidon, cognate with a term for knife, perhaps refers to bad distinctions rather than to the unachieved unification of his system).
The third in the list, Xenophanes, is also said to have made an assertion of proto-pantheistic unity, and his demythologization is similar in tone to Heraclitus’ own. All the more reason he should be rejected with the others, since even his remarkable scientific observations, like the deduction of much higher primitive sea levels from his discovery and identification of shell fossils (Diels 21.A 33), do not add up to any unified view and so would not teach him noos.
Hecataeus followed Hesiod in compiling genealogies, which constitute a summary of past time; and he followed Anaximander, the first mapmaker, by writing a geography which located peoples in space. In some sense, taking Heraclitus’ list as cumulative, Hecataeus would most fully exemplify polymathie. But the genealogies and the geography could not possibly be set into unitary relation, and so Hecataeus too is a disappointment: a comedown from the unifier Pythagoras even while he is a scientific advance beyond him.
Heraclitus’ fondness for ratios and proportions should indicate that some sets of relations among the four figures are being posited in В 40, as the term autis (“moreover”) further invites us to see.9 One set, that provided by Bollack and Wismann (pp. 151-152), concentrates on both domains and procedures:
The four names present, in the whole extent of science, four exemplary types. All the forms of knowledge are there: the organization of the divine in Hesiod, the numerical constitution of the world in Pythagoras, the renewal of theological thought in Xenophanes, the systematic inquiry into the life of men in Hecataeus.
A complex network of relationships is established among them, to diversify the nature of the term they have in common: a numerous knowledge (le savoir nombreux).
Like Plato after him (Havelock), Heraclitus rejects the Homeric universe. He not only passes over its content in silence and replaces its network of explanations with a rigorously different one (as I shall argue); he also attacks Homer with vehemence:
Homer deserves to be cast out of the contests and flogged, and Archilochus likewise. (B 42)
If there were an intellectual competition like the athletic games, then it would be apparent that Homer is so unworthy as to be disqualified, and to be struck with a cudgel, just as Homer has Thersites struck (Iliad 2.212-277), and for the same reason, his ideas. The verb Heraclitus uses is the very one Xenophanes uses in having Pythagoras say a dog should not be struck that way (Xenophanes B. 7.4-5). And lest we should think that a Thersites-like simple opposition to the Homeric universe would suffice, Heraclitus includes Archilochus identically in his condemnation: counter-assertion in satiric iambs deserves the same treatment as assertion in heroic hexameters. Homer, like the Pythagoras of В 129 and the Hesiod of В 57, is a compound of skill and folly. He is “the most skilled and wisest of all the Hellenes” (“tōn Hellēnōn sophōteros pantōn,” В 56), but he is easily deceived by mere boys, as men tend to be, with respect to the knowledge of the visible (“tēn gnōsin tōn phanerōn”) when confronted not with a holistic Olympian view of the universe but a simple and specific problem, a riddle about fleas, the point of В 56.
Heraclitus further blames Homer (B 105) for saying that no man can escape his fate (Iliad 6.488), coupling that expression to the “born in a single night” attributed to Hector and Polydamas (18.251) so as to conclude that Homer is an “astrologer,” a term which here is pejorative. And the term must also be pejorative when he applies it to Thales in his sole direct reference to any Milesian philosopher (protos astrologesai, В 38). Plutarch places Heraclitus’ further refutation of Hesiod in a similarly astrological context when he says that Heraclitus castigated Hesiod (epeplexen) for not knowing that the nature of any and every day was the same:
Heraclitus castigated Hesiod for making some days good and others bad, as not knowing the nature of every day to be one.
Taken with Heraclitus’ implied rejection of merely astronomical evidence in Hesiod’s corresponding mistake about the identity of day and night (B 57), the whole force of Heraclitus’ testimony about pre-So-cratic and other astronomical cosmology is negative. And if this were a positive rather than a negative statement of Heraclitus’, it would be his sole positive statement about any named person, with the exception of a single one:
In Priene was Bias son of Teutames whose logos was greater than that of others. (B 39)
This is positive however one reads it, though it may also be doubleedged. If we read logos, as most commentators do, in its rare sense “fame, report,” then Bias surpasses all others in Priene without incurring the unwelcome fate of Hermodorus, killed for his excellence by the Ephesians (B 121). Bias’ preeminence would then have the character of a Homeric arete, which also generates fame. If we read logos in its more usual sense, Heraclitean and general, of “statement” or “coherent conception,”logos unifies into one viewpoint the logoi of this lesser-known figure, who is always included in the somewhat unstable list of the legendary Seven Wise Men. The actual logoi of Bias must have had something of the gnomic character we find in his sayings (probably legendary), as reported by Stobaeus from Demetrius of Phalerum (Diels 10.3). If so, then they were homelier versions of the “tribal” thought whose loftier formulation is to be found in the Homeric poems. In such a Homeric world logos in the sense of fame resolves into logos in the sense of thought. Merit and responsibility are as indissociable from one another, Adkins points out, as abstract formulation would also be from them. Such an identity of these two senses of logos would certainly render impossible any achievement of a true Heraclitean logos. Though common (xunos), this Homeric logos would keep people asleep, whereas Heraclitus elsewhere (B 2) contrasts sleep with the wakeful awareness of philosophical unity.
At this point I am close to reading too much into Heraclitus, the pitfall of all his unitary interpreters (Hegel, Nietzsche, Heidegger), when they are not following the misplaced emphasis of those who give their rigorous attention only to the cosmological statements. But we have no choice, before such a barefaced statement as В 39; we must risk reading too much in, or we shall certainly read in too little.
This exhausts the list of Heraclitus’ references to his predecessors and contemporaries. Brief though it is, however, it is unusual for existing at all. Heraclitus is the only one of the pre-Socratics who refers to other thinkers by name.
Such interreference as we may deduce in Heraclitus is to the ideas rather than to names. Parmenides, for example, is alleged to be refuting Heraclitus’ principle of reconciling opposites,10 a doctrine which for Heraclitus himself finds close parallels in the pairs of opposites attributed to Pythagoras (Diels 58 B5). But he does not name Pythagoras in this connection, or the Thales from whom he may have adapted the doctrine of water-transmutation (B 31). If he takes over notions about thunder and lightning, and even the notion of Dike from Anaximander, the former are too much in the mainstream of cosmological speculation, and the latter is too common in the Greek vocabulary, to be attributed to a particular interdependence of two thinkers. When Heraclitus does mention another thinker, his main concern is to dissociate himself, in both doctrine and practice. This concern, unique in his time, should lead us to seek a corresponding uniqueness in his doctrines and practice. And we should examine all his statements as closely as we have the cosmological ones. Those happen most to resemble in concern the very thinkers from whom he is trying to dissociate himself, naming some quite forcefully and passing over others, above all the Milesians, in silence.
The particular originality of Heraclitus in his historical context should be seen first in his form, the separate utterance. Whereas what has come down to us from antiquity of the other pre-Socratics is in the form of fragments, the utterances of Heraclitus are almost surely entire statements. These single aphorisms are for the most part not fragmentary, even though a few of them are longer than aphorisms. And certainly many are lost. Most writers since Aristotle have looked past the individual aphorisms and applied them one to another to make them yield a doctrine. This procedure, while correct and to some degree unavoidable, has often had the effect, definitely avoidable, of re-directing the thrust of the individual utterance.
Each of his aphorisms stands free. It is Heraclitos’ startling originality to make them do so, a view which has been defended at least since Diels. With a master-stroke Heraclitus liberates philosophical statement from the continuous discourse in which Anaximander, Anaximenes, and very probably Thales, resemble Homer. Heraclitus harnesses what we may presume was by his time the folk tradition of the proverb and makes it do duty, singly and succinctly, for the enunciation of whole philosophical propositions.11
In producing the schematically balanced aphorism Heraclitus may be said to have detached the semantic skeleton from the Homeric hexameter, leaving behind the numbing rhythmic overlay. The sort of counterpoise and ratio, or chiasmus, that Heraclitus characteristically offers in his aphorisms may be seen as a special version of the tendency in Homer to balance epithets off against another, by setting two in one line. Similarly, when Heraclitus asserts the unification of opposites, he in effect takes the elaborate schematic contrasts that Whitman and others have found in the Iliad12 and universalizes them by detaching them from a narrative context.
Binary contrasts are deeply characteristic of mythic thinking generally, as Lévi-Strauss has shown us; and parallelism, whether in meters or not, is deeply characteristic of poetic expression generally, especially in early cultures. Binary schemes without myths, parallelism without poetic lines—Heraclitus takes these over as a usable common denominator from the oral culture that preceded him and transposes them into a form of great simplicity and great power.13
Heraclitus’ statements differ from conventional proverbs in posing puzzles rather than merely offering experiential solutions. They are like riddles; he is called ainiktes,14 and at least once he does quotea partieular riddle, the one about Homer, “wisest of all the Hellenes” (B 56): Homer indeed is wisest in his presentation of Olympian myth, but less clever than the boys who offer him a riddle about fleas.
In those terms Homer’s world of myth is closed off from the boys’ world of riddle. Now, as André Jolles (p. 129) argues, “Myth is an answer in which a question has been comprised; riddle is a question that postulates an answer.” Heraclitus has found not a simple form but a simple trans-form of expression. The answer the aphorism offers comprises a question about what is really being said, and in such a way that the aphorism is also a question whose answer is given as well as postulated. The world of myth is demythologized, the world of riddle is departicularized. Myth here is not an inversion of riddle any longer, and one riddle loses some of its point’s particularity by getting the contextual frame of the hyperbolic and itself paradoxical contrast between wise Homer (more than men) and the boys (less than men) of В 56. В 51 is like a riddle plus its solution:
They do not know how what is borne apart is borne together, a back-turning harmony, as of a bow or a lyre. (B 51)
Taken by themselves, the bow and the lyre have some of the character of a Homeric simile. But taking them as themselves illustrating the principle they enunciate, as palintropos,15 they are solutions to the riddle, “turning back” the expression upon itself. “In what way can a thing that is borne together be at the same time borne apart? Answer: a bow or a lyre.”
In their sometimes explicit insistence on a contextual framework the gnomic statements of Heraclitus resemble that riddle aimed towards the future: the oracle. Oracles became quite prominent during the period of tyranny, colonization, and imperial threat of the Asia Minor into which Heraclitus was born. Heraclitus’ one statement about oracles says, on the face of it, what everybody knows about oracles, that they give their indications in cryptic form:
The lord, he whose oracle is the one in Delphi, does not speak or conceal but gives a sign. (B 93)
But this speaking is referred to as not speaking (as therefore a legein that lacks some essential attribute of a Heraclitean logos), and the cryptic utterance is declared not to be cryptic (ou kryptei, when in accordance with another aphorism, “nature loves to hide itself,” “physis kruptesthai philei,” В 123).
В 93 shares the very characteristics of the oracle it describes: it does not conceal and it does not speak; it indicates. Now it is extremely unlikely that this statement is merely a commonplace about oracles. Any Greek knows that the oracle does not give a plain answer. What happens at Delphi which (contrary to expectation) is not speaking and not concealment is (as everybody knows) an indication about the future. In some sense this model would have to serve for all utterances, including preeminently those of Heraclitus himself, which are inferior to those of the oracle in issuing from a man rather than a god (ho anax—the God as superior), but presumably superior in that they do legein; they comprise the Heraclitean logos, in a way that a mere oracle, always applied to a specific future situation, would not.
The dialectic above, where the act of defining the oracle both does and does not define the utterance in which the definition is made, provides its own self-sufficient context while talking about another context. The oracle context has at least six constituents: the god, the prophetess, Delphi, the consultant, the utterance, the future event. All these constituents must be translated into a single pointing of direction (semainei), as the self-sufficient statement before us, it announces, cannot be. Neither can we translate this statement into the later distinction between syntax (implied by legein) and diction (semainein much as the term comes to imply for Plato [Cratylus 393a]), though again that distinction cannot be declared to be absent. Nor, since the oracle comes from a god, can we exclude the Homeric sense of semainein, “command” (Iliad, 1.289, etc.). Heraclitus has here produced a statement which at once exemplifies and qualifies what it says. Consequently, it is more like an oracle than a riddle; though oracle tends to meld with riddle,16 it is also more like a riddle than any oracle. And it is more bare-faced and commonplace in phrasing than a riddle or an oracle ever are.
The proverb is oriented towards the past; it summarizes common experience, and Heraclitus was used in antiquity as a source of proverbs, some of his aphorisms (B 130-135) coming down to us in a collection of proverbs. The riddle is oriented towards concrete objects that can be named: fleas or whatever. The oracle is oriented towards the future: it hints at a definite result in future time. НегасШш’ statements are undefined as to time, and yet they bring together past, present, and future in their melding of these simple verbal forms, much as the seer Calchas is said to have done in the Iliad (1.70). Oracle, too, tends to meld with proverb: the proverbs meden agan and gnothi seauton are attributed to the Delphic oracle. But they are also attributed to this or that person (including Bias) in the list of the Seven Wise Men (Diels 10), and Heraclitus is their successor.
The riddle, Aristotle says, enables one to make a statement through metaphor which one would not be able to make by a combination of terms (“tēn tōn onomatōn synthesin𦀝).17 He has also defined the essence (idea) of the riddle as putting the statements in question (legonta hyparchonta) into impossible combination. And it is perhaps no accident, since Aristotle is our source in both instances, that the word he uses for “combination,” sunapsai, is the same which Heraclitus himself (presumably) uses in the nominal form for the beginning of В 10: “synapsies hola kai oukh hola,” “combinations are whole entities and not whole entities.”
Heraclitus’ form of the riddle bypasses what could be called metaphor; it tends to pass on to the simile,18 a form that makes logical connections more explicit. Its combinations are whole entities, but, riddlingly, they are at the same time not whole entities: they advertise the fact that there is a process by which they require completion. We must supply a thought-process to the statement about the oracle at Delphi, or else we are left with a useless commonplace on our hands. The plain sense remains apparent, “How could anyone escape what never ducks under (or never sets)” (B 16). But on the other hand, concealment is a law, “Nature loves to hide herself” (B 123).
Now, the thought process which we apply to В 93 is a critical one: to make the statement mean more than a commonplace about the oracle it must pointedly question what a statement is (including itself), what an oracle is, and how an oracle relates to a statement. The process underlying this single sentence already coils within itself the elenchic and contextual process by which Socrates, pretending to be puzzled, asks what arete is.
The development of such an elenchic process begins before Plato. According to Montgomery Furth, it is the only way we can solve Parmenides’ seeming self-contradictions:
First, it is of the essence of Parmenides’ procedure, as I understand it, that he is not at this point putting forth an ontology of his own, but is practising dialectical criticism upon that being put forth by Betathon; his own word for his argument is elenchos (B 7.5), which we must assume means for him, as it presently was to mean for Socrates, the technique of refuting an opponent by reasoning from a premise that the opponent accepts to a conclusion that he must regard as intolerable, such as an explicit self-contradiction or the negation of some proposition that (for whatever reason) he cannot deny, and thus forcing him to abandon the premise, [p. 118]
Applying Furth’s perspective to В 93, we may say that Heraclitus is “practising dialectical criticism” not upon what an interlocutor puts forth, but upon what he himself is putting forth. While not so Platonic in his elenchic procedure as Furth’s Parmenides, Heraclitus does provide an elenchos of the statement upon itself. We test the meaning of “the Delphic oracle does not speak or conceal but gives a sign” by what it says, by asking in what sense it itself does what it describes (a procedure quite different from just seeing how many meanings it has). From the suspended contradictions of Heraclitus’ method we know he does not regard an explicit self-contradiction as impossible.19 In this elenchos he has seen the possibility of coming at explicit contradictions not by elaborate Parmenidean dialectic (if indeed he could have lived long enough to be aware of Parmenides), and not by Pythagorean mathematical fiat, but by a kind of statement where contradictions are bypassed, in the very process of making the statement. Such a statement would not be either an answer (like the myths of his predecessors) nor a question-with-hidden answer (like the riddles of his folk contemporaries), but a logos, a mere statement.
There would be a sort of looseness, of course, in the way such a statement would critically apply to itself. A similar looseness obtains even of the much more fully developed elenchos in Plato, which exhibits a variety of logical functions as Robinson presents them. The very openness of possibility for Plato, as earlier for Heraclitus and Parmenides, would allow the elenchos to develop. And we may apply to Heraclitus what Robinson says (p. 15), commenting on Plato’s use of logos, “the assumption that there are no extra premisses is made easier by the ambiguity of the phrase ‘according to your logos,’ which Socrates frequently uses in refutation, especially in drawing the conclusion.”20
By the time of Heraclitus logos means more than a single word (the Greek word logos implies a statement, as against epos or onoma). Where Heraclitus speaks of a single word, he connects it with the process of an utterance which, in syntactic analysis, would be predicative:
The wise, one, alone does not wish and wishes to be called the name of Zeus. (B 32)
The signifier onoma for a single word connects immediately to a unitary signifier, hen, which implies a further prediction of skill/wisdom, to sophon, which nests a distinction: “one” taken globally as hen and one taken severally as mounon. That the question of identity or difference between hen and mounon may be pressed to yield some of the theoretical foundations of modern number theory would be incidental here. Nor is such theory entirely absent in a philosophical context where the one-many problem was already a live issue, or in a tradition that had produced Pythagoras and would very shortly produce Zeno, Melissus, and Eudoxus. In В 32, however, the quasi-resolved contradiction between hen and mounon is itself nested in a syntactic contradiction on which it depends, ouk ethelei kai ethelei, “does not wish and wishes.’ This contradiction reverses normal Greek syntax; we would usually get the negative ouk second in order, not first. This phrase ouk ethelei also stretches the subcategorization rules (to use modern linguistic terms) by which the verb ethelo is restricted to animate human or superhuman nouns for its subject. “The one” cannot “wish”—unless “the one” is somehow animate, which would be the case of the god Zeus. Other aphorisms of Heraclitus about the difference between 6human and divine perception allow us to include in the sense of В 32 a special, intensive sense of the verb ethelo when it takes a divine subject. “Not wishing in some human sense I mean, but wishing par excellence it is, when the one wishes to be called the name of Zeus.” This slant to the reading would also assign a strong semantic function to the abnormal prior position of the negative ouk.
There are, I have been saying, several slants to the reading of this aphorism. The slants interlock, to produce not what one normally thinks of as ambiguity, a puzzled suspension between one or more mutually exclusive readings or even the rich harnessing of these for “poetic” effect. Rather, the several slants come sharp and clear. Their coexistence acts not so much for enrichment, though Heraclitus is ; incidentally “poetic” in this sense, as many poets are. He is also sternly propositional, as poets tend not to be. The effect of the several slants is elenchic and dialectical. They themselves constitute a logos in the sense of a chain of reasoning (sense iv, 1, Liddell and Scott) as well as a logos in the sense of a voiced statement (sense vi) and a logos in the sense of an ordered explanation (sense III).Legesthai in В 32 means primarily the second of these, but the other two senses cannot be ruled out, particularly in view of the implied quasi-resolved contradiction with onoma (contrasted with logos as “locution” with “statement” in earlier usage), and particularly in view of the centrality of the term logos for Heraclitus.
The vast complexity of this term works even more forcefully as an indicated process describing the aphorisms, logoi in an intensively summarizing sense, than it does for the rich plurisignificance of the inclusive term.21 The elenchic procedure by which one verbal statement tests another (or with Heraclitus tests itself) may be traced in the counter-Homeric act of questioning, an act explicit in Xenophanes and implicit already in Hesiod’s defining separation of truth from falsehood. One final result of such intellectual definition would be a holistic doctrine, like those of Pythagoras and Empedocles. Or the procedure of definition itself can be made the center of speculative activity, at the service of a holistic doctrine. Heraclitus, who may well possess a holistic doctrine like most of his contemporaries, is remarkable and unique for making the elenchic procedure itself become a sort of metalinguistic demonstration:Logos as proposition, logos as gathering-together, logos as affirmation, and logos as mere utterance, become one act.
Take, for example, the two “initial” statements which have often been taken as the axiomatic groundwork of Heraclitos’ doctrine:
The logos being this always men are without understanding both before they hear and when they first hear. For when all things do come to be according to this logos, they are like the untried, trying these words and deeds, such as I search out according to nature distinguishing each and telling how it is. What they do when awake escapes other men just as what they forget when they are asleep.
The logos being common, the many live as if they had private thought. (B 2)
In В 2 a distinction between speaking (logos) and thinking (phronesis) —a distinction still very much alive in philosophy—is superseded by the distinction between “common” and “private,” because the first distinction disappears when the second one is perceived. If men realized that the logos is common they would also then have a common phronesis rather than a private one, a state of affairs which the aphorism presents as philosophically correct. Still, it is an unattained goal rather than a mere fact, very much along the lines of В 18, where one must hope for something to see something:
If he does not hope for it, he will not find the unhoped-for, it being uninvestigated and unapproachable. (B 18)
What does “the logos being common” mean? It surely must refer, as has nearly always been assumed, to Heraclitus’ holistic and unitary doctrine. That is the strong sense of logos. But there is also a weak and casual sense, where “the logos being common” would mean “speech is a shared activity among men” or “we must share speech to understand one another; so in that sense at least our thought is not private.” This casual observation about language, the weak sense, can be made to produce the strong sense, again by the elenchic and metalinguistic operation of the aphorism upon itself.
In В 1 men are always (if aei may be said to carry through the clause)22 trying individual words (epeon) rather than logoi, and particularly this logos (the weak sense) of Heraclitus, which they fail to understand both before they hear it and after. They remain in the semblance of the untried or inexperienced; the peiromenoi resemble a-peiroisin, the latter word usually referring to deeds (ergon) not words, but epeon and ergon are on a par here as the objects of “trying.” Moreover, the words and deeds they are trying are no different from the ones that Heraclitus sets out. They are “just such (toiouton hokoion) as I am setting out according to the way things are (kata phusin).” All this process takes place, again, “when all things come into being according to this logos” —with the strong and weak senses of В 1 again repeated for logos. There is a permanent situation (aei) with an implicit possible contrast (ginomenon could mean “coming into being” or “existing”) and an actual contrast, the permanence of aei set against the punctualmoment-by-moment of the aorists in akousai and akousantes, not likely to be gnomic in so long a passage.
The final statement here serves both as a conclusion and as a transposition to the category waking-sleeping, which functions in other statements of Heraclitus. That category is here adduced to remove the distinction, strangely, between waking and sleeping, by a logical inference based on a casual observation. If men “let escape from their memory” (epilanthanomai is causative) what they do when asleep, then the things they do when they are awake, if those things escape them, have the same status (hokosa, hokosper). In that case their activity (poiousin) and their passivity (akousai, “hear” or “obey”) have the same character, and sleeping cannot yet for them be distinguished from waking when they permanently lack understanding (aei axunetoi). The last statement may also include an acute observation about the sleeper; he tends to forget his own role in dreams more quickly than the actions of those around him.
In this process to misunderstand (a-xunetoi) the logos may carry the pun, as Bollack remarks, of a divorce from the “common/’ (xunos, В 2, etc.). Speech is there accessible to all, but it must be understood in a “deep” sense and it is called deep in В 45:
The bounds of the soul you would not find out as you go, traversing the whole way; so deep a logos does it have. (B 45)
Full humanity is here seen as an unending process, to which one would not have discovered the boundaries even after “traversing the whole way.” They are the boundaries of the soul, in another aphorism, where the logos increases and exalts (both are senses of auxon) when it is perceived as linked to the soul:
Of the soul is a logos increasing itself. (B 115)
Here the predication insists on itself, since we have a copula and it is a stylistic peculiarity of Heraclitus’ shorter statements to omit the copula (В 3, В 33, В 48, В 54, В 60, В 62, В 67, В 96, В 101а, В 103, В 107, В 119). And the predication is strong: “It is the property of the soul to be a logos that increases itself” and also “it is the logos in the soul that is the element that exalts it.”Psyche retains the weak sense “breath of life,” but he has also given to psyche, according to Martha Nussbaum, a new, strong sense as that in a man which organizes all his faculties. This statement can also be made a version of one reading of В 1 and В 2: “Human life is properly seen as an increment (auxon) of speech-acts.” When human life is not so seen, the possibility of the Heraclitean utterance, in its elenchic dynamism (and so, according to Heraclitus, the possibility of any self-conscious utterance) will have disappeared. “It is itself (heauton) that the logos increases.”
So Heraclitus, who tends anyway to the third person in Bo11ack’s reading (pp. 11-15), erases himself from the very logos which of course he is making at every point, as well as “carrying through” (diegeumai, В 1). So he says auditors must go through the logical step of detaching the speaker from the speech in order to perceive the underlying unity:
Hearing not me but the logos it is wise for them to agree they know the one is all. (B 50)
(For einai Bollack reads the manuscript eidenai, and this difficilior lectio gives a stronger sense.) The act of wisdom (sophon) is the act of simple accord or agreement (homologein), and going through the process is equivalent to knowing that “one” is “all”—that, among other senses, one logos can be taken for all. The stupid man, by contrast, will not face a statement: he flutters in fright before any; and it is part of his stupidity that he is actually fond (philei) of doing so. It is a habit or wont (a sense of philei) which he enjoys too much to break:
The stupid man loves to flutter before every logos. (B 87)
More important, as always, than the doctrine hen panta is the process of wakeful attention in which it is perceived—but the separation cannot be made between process and doctrine, as Heidegger (1, p. 18) reminds us of В 50: “Heraclitus teaches neither [pantheism] nor any teaching. As a thinker he offers only thinking.” Actually, Heraclitus’ elenchic process is so comprehensive as not to permit even this restriction. When he induces a thought, the resultant process of thinking provides a center of affirmation as well as a means of restrictive denial.
The variety of emphases and the possible contradictions in the senses of the word logos exemplify the actual process of setting contradiction into coordinated statement. The implied affirmation that contradictions are meaningful, and the process by which they are found so—taken together—form the single subject we can confidently locate as central to Heraclitus’ thought.
Seen in the light of such an elenchic procedure, a doctrine of Hera־ clitus, the resolution of contradictions, is inseparable from a constant linguistic technique, the metalinguistic thrust of his utterances. “Does not wish and wishes” (B 32), by being referred to a linguistic procedure (and in this case further explicitly to an act of saying, legesthai as used in В 32), is qualified in the contradiction it nests. The same is true for the contradiction between horizontal (peirata, bounds) and vertical (bathus, deep), resolved by being stated as unachieved in В 45 (“You would not find out”). So that even the possibility of a plain reading leads to an elenchic series for В 60:
The road up/down is one and the same. (B 60)
The double affirmation claims, as it were, the necessity of duplication, of saying “one” and “the same,” just because if there must be some sense to linguistic signification, the law of contradiction would seem to be violated in this statement. If the road up and the road down are not in some sense distinguishable then the words ano and kato would have no meaning. One can read В 60 as saying that for a man whose house is on top of a hill the same road is the road down when he is leaving home and the road up when he is returning. Or in the “life is a road” metaphor that seems to underlie В 32 (and is found in the later reported form of the Oedipus riddle), one can say that it would be possible to attribute life to the beginning of a man’s existence and to the end. Any road is “one and the same/’ from whatever standpoint. One could go on to produce other plain readings of В 60, and these would tend not to exclude one another. They would merely have in common a résolution of the seeming contradiction (up/down) which the predicate of the aphorism itself (“one and the same”) would seem to resolve.
Aristotle himself, who formulated the law of contradiction, seems to pause over Heraclitus as a puzzling case when stating the law: he has just summarized the law in the Metaphysics and called it the firmest (bebaiotate) of principles. He then states it again:
For it is impossible to suppose that anything whatever that is the same can be and not be, as some men think Heraclitus says. For it is not necessary, as to what one says, also to suppose those things.
(Metaphysics 1005 b 23)
Does Aristotle agree or disagree with the men who think this about Heraclitus? As he often does in his historical statements, Aristotle has stopped just at the point where we would consider the question to be crucial. The very fact that he has done so ought to put us on our guard about separating the manner in which Heraclitus states the reconciliation of opposites from the doctrine that they are to be reconciled. It is fair to say that Aristotle seems here unwilling either to affirm or to deny that Heraclitus contradicted the law of contradiction.
The principle of disambiguation (applied by panton kekhor ismenon, В 108), which would allow for the contradictions to be sorted out, or even to be asserted as reconciled, cannot for Heraclitus be detached from the context in which the statement about anything, including disambiguation, would be made. Contradictions tend to nest other contradictions, or at least distinctions, as ano/kato above nests a more colloquial distinction, mia/heaute. In В 32 there is a whole series of such nestings: hen/sophon, hen/mounon, ouk ethelei/ethelei, legesthai/onoma, Zenos onoma/hen, and combinations of these. But all the nested contradictions and distinctions can be caught up, as the very multiplicity of the nestings implies, in the elenchic thrust of an utterance: “ethos anthropo daimon” (B 119), “habit for man, god/’ has only three words. It is stripped of its copula for maximum economy (though gnomic sayings are the one place where the copula may be omitted without wrenching the Greek language). And it nests as many distinctions as it has words: ethos/anthropo, ethos/daimon, anthropo/daimon, without reference to interaction between the potential contradictions, which I shall discuss below.
Heraclitus’ aphorism about disambiguation, sets up a context and an experience, itself based on a negation:
Of as many as I have heard logoi, no one [either “person” or “logos” but not both] attains to the point of knowing that wisdom is separated from all things. (B 108)
This startlingly proto-Aristotelian act of radical definition23 exemplifies what it asserts, since it is realizing this one principle that separates Heraclitus from everything else that he has heard, all of that (hokoson) being defined negatively—being separated (kekhorismenon)—by its not attaining to the point of realizing just this principle.
In a similar aphorism (B 72) it is the logos that is personified (dioikeo normally takes a human subject) as administering such wholes (hola) where people incessantly (dienekos) traffic, thereby entailing differentiation:
With what they most continually consort, the logos managing the wholes, with that do they differentiate themselves; and what they happen upon in the day, those things appear strange to them.
There is a superlative at the outset here, malista, which certainly goes with “continually,” and may carry on through the clause. The whole process of “consorting” in any case further entails, and also exemplifies, another process, that of finding alien to themselves or strange (xena) whatever they happen upon. Here, though the process may be haphazard, the principle is carried through, however unaware of it the agents might be. Elsewhere, where the same verb for “happen upon” is used, their imperception is the subject in view:
Many do not understand such things, all the persons that happen upon them, and they do not know them when they have learned them, but they appear to themselves to do so. (B 17)
The final reflexive pronoun, heautoisi de dokeousi, is odd, as Bollack points out. There is an identity which is circular (heautoisi, to themselves) and so only apparent (dokeousi). In this case the same principle of disambiguation is asserted by the intentionality of an opposite case, where the lack of understanding even after things have been learned produces a condition where disambiguation is impossible. The state that Heraclitus is here rejecting would seem, from the point of view of modern logic, very close to his own notion of the identity of opposites. But he is clearly stating it here to be an impossible state of affairs when looked at negatively, even though when looked at positively, the same process of ’happening on” (enkurousi, В 72) can activate a situation that allows for the logos to manage wholes (hola) and for a strangeness (xena) to be perceived instead of a mere seeming identity (heautousi de dokeousi).
Such a process is a necessary prior condition to the conscious realization of any fact whatever, Heraclitus asserts, since any fact must be accompanied by an act of attention, defined as an act of expectation or hope. Without this there is no approach (aporon, В 18) to the fact.
Giving a contextual and metalinguistic twist to his expression of the unity of opposites keeps Heraclitus from being a mere stater of paradoxes, a writer of Dissoi Logoi (Diels 90; II, pp. 405-16). Even without regard for the elenchic process, the unity of opposites is often qualified in his aphorisms by the explicit statement of measure or proportion (В 30, В 31, В 94), and by its linguistic equivalent of the simile.24
I leave aside what is, if not Heraclitus’ all-encompassing doctrine, one of his most frequent subjects, the unity of opposites as it applies to the notions of flux and measure. Not only have these questions been discussed so exhaustively that I have nothing to add to them; I feel that we are not able, given the statements we have, to adjudicate, for example, between the case for a constant flux (Guthrie) and the case for a spasmodic one (Kirk). Moreover, the much-discussed river proposition, which Plato sees as a simile (apeikazon),25 is uncertain in its application, beyond the fact that it was understood in antiquity to indicate flux.
What is uncertain is just how far a general principle can be deduced from the specific assertion, and consequently how the explicitly physical statements relate to those that may or may not be physical. “The road up and the road down” has been related to the river-fragments rather unconvincingly, and even more unconvincingly to the fire-process.
The fire of Heraclitus may well have something to do with the notion of logos, but we do not have the verbal evidence to decide just what. Logos can mean proportion, which can imply measure, and measure (metra) as an adverb is applied to both the kindling and the quenching of the fire which is once identified with kosmos:
This kosmos, the same for all, neither one of the gods made nor a man, but it always was and will be an always-living fire kindled in measure and quenched in measure. (B 30)
Does this allow us to make a much closer connection between logos and the fire-process than we could make, given all the logos aphorisms, if there were no fire aphorisms at all? For that matter, is there evidence here by which we can confidently say that “kindling” and “quenching” are successive rather than simultaneous acts? (A case could be made for perspectivai simultaneity, as with “the road up/down is one and the same.”)
“Measure” applied to a linguistic utterance takes for Heraclitus the form of a proportion. A proportion stated in proverbial form in itself contains an elenchic thrust; we are meant to ask what it can mean. Such a proportion or ratio, in the form A:B as C:D, is a thought pattern common in Heraclitus, as Hermann Frankel (1938) tells us:
A man is called childish with regard to a god as a child is with regard to a man. (B 79)
The form of the proportion here is almost as bare as it can be, but not quite. “Ekouse,” strictly “hears,” is used in its sense of “has the reputation of,” but taken together with “childish” or “foolish” which literally means “wordless,” (nepios) there is a nested contradiction between “hear” and “not speak” which is superseded by the main proportion. If the daimon may be defined as a god whose presence is perceived but whose identity is not yet established, then the term already names the side of the man-god relation which makes man as foolish as a child is in the child-man relation. That is what another aphorism says:
So Heraclitus considered the formed opinions of men to be the pastimes of children. (B 70)
This reverses the ratio of В 79 while simplifying it. The aphorism about the boys who tell Homer the riddle about the fleas, on the other hand, gives a more complicated ratio:
Men are deceived in knowledge of things visible similarly to Homer, who was the wisest of all the Hellenes. The boys deceived him killing fleas, saying “All we have seen and caught, we have left behind; all we have not seen and not caught, we carry.’ (B 56)
Homer/Men-Men/boys is what we would expect, given the fact that Homer is “wiser than all the Hellenes.” However, when Homer is equal to other men rather than greater, the men who “are deceived in knowledge of things visible similarly to Homer”; then mere boys can deceive him with the riddle about the fleas. In the strange (but epistemologically central) case when wise Homer is equal to all men: then mere boys, normally inferior to men, can be superior not only to men but to Homer himself. The boys do have the sense Heraclitus exalts: sight [ opsis,В 55], while Homer is proverbially blind. The subversion of the first part of the ratio (Homer’s superiority to other men is mentioned first) entails the subversion of any part. Again, В 56 can be taken as asserting either this metalinguistic principle or else as asserting the epistemological priority of “the knowledge of things visible.” In fact it asserts both, and asserts them in conjunction with one another. The Homer-men-boys ratio provides an elenchic context for the epistemological desideratum. The desideratum provides a hard test case for even the wisest of the Hellenes, though a simple one if his blindness is considered.
Subversion of normal proportion is often used by Heraclitus to show the revelatory nature of an extreme case: “If all existent things were smoke, noses would perceive” (B 7). “The sea is the purest and most polluted water, drinkable and sustaining for fish, undrinkable and deadly for men” (B 61). “Physicians cutting, burning in all ways, testing the sick badly, ask to take pay from the sick they do not deserve, working so as to make good things and illnesses the same” (B 58). “Corpses should be thrown away sooner than dung” (B 96). “Asses would rather eat sweepings than gold; their food is sweeter to asses than gold” (B 9). “Time (aion) sporting, is a boy playing checkers; the kingdom is a boy’s” (B 52). “The most beautiful cosmos is like a heap of things poured together” (B 124). “War is father of all and king of all; he has shown some as gods, others as men; he has made some slaves, others free” (B 53).
In all of these the proportion is confounded at an extreme point so as to indicate its reach as a cognitive tool. These proportion-subverting aphorisms, seen in this light, in no way contradict (even in a unity of opposites) the aphorisms which assert that proportion is to be maintained at all costs: “It is necessary to quench excess even more than a conflagration” (B 43). “It is necessary for the city to fight for regulatory law (nomos) as for a wall” (B 44a). “The sun will not exceed measure. Otherwise the Furies, allies of Justice (Dike), will find it out” (B 94).
Justice, Dike, implies a proportion, a logic of counterbalances, in Homer; and also in Anaximander’s abstract application of the term to cosmological processes. This last aphorism is cosmological, of course, but Heraclitus elsewhere offers the term in an “existential” and also a metalinguistic sense: “They would not know the name of Dike if these things did not exist” (B 23). And again, another use of Dike is at once epistemological and social: “For the most believable man knows, guards the believable; and indeed even Dike will seize the fashioners and witnesses of lies” (B 28).
The thought pattern in В 90 takes the form of a normal ratio, A:B as C:D.
All things are exchanged for fire and fire for all things, just as goods are exchanged for gold and gold for goods. (B 90)
The naked eye can always tell the clear difference between gold and goods. The point is that they are equivalent and reversible only when the exchange process is activated that allows С to be replaced by D or D by С indifferently.
Gold and goods, С and D, are in the subordinate clause. The main clause deals with “fire” and “all things,” whose less perceptible exchange process it explains by a perceptible one. The thought-process, as always, is elenchic, a fact emphasized by the co-existence of three meanings for antameibetai: “exchange,” “repay,” (a notion assimilable to dike), and “answer” (a metalinguistic touch). The act of forming ratios lies under the caveat of another aphorism, “Let us not make connections or accords (sumballometha) at random about the greatest matters” (B 47). And another can be taken either as a caveat or else as a second-order contradiction about contradictions: “That which changes (metaballon) comes to rest” (B 84a), “rest” being the final (and inclusive?) term in a series of contradictory pairs—“Sickness makes health sweet and good, hunger satiety, toil rest” (B 111). The term paired with “rest,” kamatos (“toil, weariness”), also finds definition elsewhere: “It is toil for the same persons to struggle and to be ruled (archesthai)” (В 84b, where the slightly commoner contemporary sense “begin” cannot be ruled out for archesthai).
In these paired contradictions there is a focus on naming things properly. Finally, of course, Heraclitus is coming to terms with, and solving, problems that are not just verbal. He would seem to be interested in “sickness,” for example, a topic to which he returns for its own sake.
“The wisest man with respect to a god appears as an ape in wisdom and beauty and all other ways” (B 83). The formula “Man is to god as ape (is to man)” does not solve any questions about man and god, a relation that is a major preoccupation for Heraclitus. In fact, his tendency is to refer one major preoccupation to another, as the question about the relation of man to god is not only defined by the term ape but also referred to another question he often discusses, that of appearances and perceptions (“appear,”phaneitai). “An invisible (aphanes) harmony is better than a visible (phaneres) one” (B 54) rings changes on the same verb root as “appears.” The aphorism about Homer’s failure to guess the boys’ riddle about fleas turns on the same root (“the visible,”ton phaneron, В 56). But the area of perception is not a firm one, even in its problems, if one sets “Eyes are sharper witnesses than the ears” (B 101a), where the eyes get priority, against “Of as many things as there is sight, hearing, knowledge; these I prefer” (B 55), where ears and eyes are not only put on a par, but these two witnesses of the perceptible are themselves put on a par with the technique for thought, mathesis; then again, “Eyes and ears are bad witnesses for those who have barbarian souls” (B 107) and “Sight lies” (B 46).
The pre-Socratics did not clearly distinguish between physical and mental phenomena, and yet some unstable form of that distinction is present here, as it must underlie rejection of Hesiod (and everyman) “who did not know day and night; for they are one” (B 57), where day and night certainly do not appear as “one’ to the senses. Their identity is an “invisible” harmony (and so preferable by the standards of В 54) rather than a visible one.26
Man may be defined by his perceptions. His perceptions, in turn, are influenced by another distinction important for Heraclitus, the distinction between life and death, one that often involves for him a third term like sleep, or water, or immortality. He defines misunderstanding and imperception of the logos in the last clause of В 1, in terms of a spurious identity between sleeping and waking. “All the things they do when waking escape other men just as they forget such things when sleeping.” “Death is what we see when awake; what we see when asleep, sleep” (B 21). “Man kindles himself a light in the night having quenched his eyes for himself; living he touches the dead when asleep, waking he touches a sleeper” (B 26). “There await men when they die what they do not expect or conjecture” (B 27). “For souls water is death, for water it is death to become earth, but water comes from the earth and the soul from water” (B 36). Whatever physical process is established here with the help of the fire-earth-water series of В 31, there is a logical process which makes the agent or the element or the means of death the same as the cause of coming into being (genesthai). To call “psyche” merely “the breath of life” is to rule out a main distinction; “soul” is radically different from “water” and “earth,” though at the same time subject here to the process of coming-to-be and passing-away, to use the Aristotelian terms which are only a later development of this distinction. On the spiritual side, the connection of souls with water is an old Indo-European notion and a Babylonian one as well. Thales, like Heraclitus an Asiatic, echoes this connection in his notions “All things are full of gods” (Diels IIA 22) and “All is water” (Diels IIA 14), notions that here Heraclitus transposes in В 36 by putting them through a logical process. (He may also be refuting the metempsychosis of that contemporary he mentions most scathingly, Pythagoras.) “Sickness” and “souls” themselves come into conjunction in В 68, where the mysteries serve as a remedy from the circumstances “inherent in coming-to-be” (en te genesei), and perhaps also in В 67a, where the soul acts like a spider to repair lesions in the body.
The principle of contradiction can apply even to such comprehensive terms: “The same thing are living and having died, and having waked and being asleep and young and old; for the latter change into the former and the former change back again into the latter” (B 88). “The god is day night (euphrone), winter summer, war peace, satiety hunger (all things are the opposite; this is mind), and it alters itself just as (fire) when it mixes with incenses, is called according to the pleasure of each” (B 67).
In this last statement the metalinguistic arbitrariness of nomenclatures (onomazetai) refers not to bewilderment, but to the most comprehensive term, god. As the god here subsumes the pairs of opposites, so he serves in В 102 as a vantage from which human judgments have less than finality, “To the god all things are beautiful and good and just, but men suppose that some things are unjust and others just.” Here the term of distinction begins as three (beautiful and good and just) and ends as one, the just (dikaios), a term which includes the dike of still other aphorisms.
“Human habit does not have knowledges, but the divine does” (B 78). Heraclitus also provides a deliberate paradox in the first clause, since the word for “knowledges’’ is the same as for “proverbs” (gnomas) and it is precisely “habit” or character or custom (ethos) that proverbs express. And by another aphorism, the definition of ethos for man is precisely the divine (B 119, daimon, however, rather than theion).
Moreover, in another statement the distinction between life and death is lifted in order to lift another distinction, that between men and gods, a distinction itself made in the language as between the mortal (thnetoi) and the immortal (athanatoi):
Immortals are mortal and mortals immortal, living the death of the former and dying the life of the former. (B 62)
Heraclitus here wrenches the normal Greek of “the former” (ekeinos) and “the latter” (houtos) by omitting the second term and repeating the first instead.Ekeinon, “those,” repeated when it should be complemented, is suspended in the meaning “the former” and might be taken to mean only “those.” The pairing of ekeinon in reference implies the distinction “former-latter,” but when the paired term is the same term, as seldom or never elsewhere in Greek, the grammar of the pronouns emphasizes the arbitrariness of identifications which in any case in the very first clause, with its emphatic suppression of copula, is equally arbitrary in its application. In what sense are immortals mortal? Surely, not finally in the sense of Xenophanes’ cultural relativism. The hardest clause is first; it is more difficult to understand how gods die than how there is an afterlife for human beings, because if gods die in Xenophanes’ sense, they cannot really be called “immortal”—or have attributed to them the powerful comprehensive function of the other statements quoted above.
And whether we read “ekeinon” as “the former” or “those,” we are equally uncertain as to how far back we are to go in the clauses to pick up the pronominal reference. Does “the former” include both members of the first clause (as opposed to both members of the second) “immortals-mortal?” Does it include only the first member of the second clause, “mortals (as immortal)?” In that case should I have added that last parenthesis? Does it include the first member of the first clause and the first member of the second, “immortals (as mortal) and mortals (as immortal)?” Should the parentheses be retained in both instances? Or, when read as “those,” does ekeinon include both members of both clauses in their identifications as in their distinctions? The failure to balance ekeinon with touton opens all these possibilities and turns the simplest form of flat predication, when it identifies men and gods in the charged area of their central distinction and relationship, into a metalinguistic maze.
As for Heraclitus’ direct treatment of the myths of gods, there is a split between his theological aphorisms and his religious ones. When he talks about theology he tends towards an abstract monotheism that is distinctly post-Olympian; his god sounds like the nous of Anaxagoras or the holistically perceiving God of Xenophanes (Diels 21, В 24). But when he speaks of religious observance, he is distinctly pre-Olympian. He mentions the mystery cults of chthonic revival most often, identifying Hades with Dionysus, who has a strong chthonic character.27 The Thracians, legendary for their savagery, are said by Herodotus (5.7) to worship only three gods—Ares, Dionysus, and Artemis—all three of whom are mentioned by Heraclitus, if we may call “Polemos” (B 53, В 67, В 80) a non-Olympian version of Ares. Diogenes does not recognize the distinction when saying (Diels Al) that Heraclitus wrote three books, “On the All, Politics, and Theology,” since the term theology28 in the sense contemporary with Diogenes could equally be applied to Heraclitus’ statement about “the all,” and to his statements about individual figures, the Erinyes, Dike, Dionysus, and the Artemis of common cult at Ephesus on whose altar he is reputed to have lain his book (Diels 22.A1). Empedocles gives his elements the names of gods, but when Heraclitus mentions the chief of the Olympians, Zeus,29 he says that “the one . . . does not wish and wishes to be called the name of Zeus” (B 31), and “name” (onoma) cannot entirely be separated from one of its senses (sense m in Liddell and Scott), “a name and nothing else; a false name.” Even such a seemingly cosmological statement as “The sun is new every day” (B 6) also echoes an ancient Indo-European myth.30 The pigs of В 13 (“Pigs like mud better than pure water”) since the paradox turns on “pure,” are not perhaps to be distinguished from the purification rites of the Eleusinia, where they were used for sacrifice.31 This could bear on the point of “pigs bathe in dung” (B 37). The kukeon, the barley mixture which could only be drunk off after being swirled around so its elements were separate, “the kukeon separates when mixed” (B 125), was also used in the service of Demeter, the chthonic deity of the Eleusinia, as Guthrie points out (p. 200).32 In this aphorism, the Heraclitean unity of opposites is itself applied to something which is neither questioned nor defined, a ritual drink in a chthonic mystery cult.
Plato neutralizes the chthonic element of the Cave in the Republic while utilizing the primitivism of its social circumstances. Heraclitus, on the other hand, centers on the utterances of a priestess in circumstances associable to caves and chthonic cults. The serpent is a chthonic deity, and it is after killing the serpent, Pytho, that Apollo settled on Delphi as the shrine where his priestess would speak oracles from the cave. Heraclitus is the first person in literature to mention the Sibyl, and he does so in order to praise the unvarnished character of her utterance, as distinct, if we are to trust Plutarch, from the charming songs of Sappho:
The Sibyl with mad (mainomeno) mouth uttering what is unlaughable and unadorned and without incense attains to a thousand years with her voice through the god. (B 92)
It is the unexpected side of the god, the awe-filled daimon, to which he attributes the essence of humanity, or else the mere habit of humanity, “habit for man: god,” “ethos anthropo daimon, ’ (B 119). This statement at once theologizes and demythologizes itself (“Not a daimon but character leads a man”). It has three nested contradictions, and they do not really overlap: the known (ethos) versus the unknown (daimon), man (anthropo) versus either his habit or his lair (a Homeric sense of ethos); man versus god.33 In each case we have a logical transfer, “A to В via X,” rather than a true ratio or proportion. The transfer dynamically names, and names as dynamic, the relationship between man and god, who as a daimon keeps alive and revives the chthonic setting at the same moment that it is getting a post-Olympian definition. В 92 refers to the legomena of a rite where the dromena issue in legomena. In the dromena of the Dionysian rite, the pre-Olympian phallic procession, resides a comparable quickening of definition. And it is to them that Heraclitus speaks, according to Clement: so he defines his audience:
Night-wanderers, magi, Bacchants, revellers, mystic initiates; for the mysteries men hold in regard (nomizomena, also “keep”) they celebrate (mueuntai, also “keep secret”) in an unholy way. (B 14)
This list confounds Persian and Greek, surely deliberately, by including magi in the list, and the fact that they were interpreters of dreams (Herodotus 7.37) may correlate with the first item in the list, “nightwanderers,” a term more vaguely inclusive than the others and so perhaps applicable to all of them.
These names suggest both praise and blame, as does his statement about the procession:
If it were not to Dionysus that they performed the procession and sang a hymn to the shameful parts, most shameful things would have been done. Hades and Dionysus are the same, to whichever (hoteo) they rave and revel.34 (B 15)
The connection of shame with the act of worship induces, it would seem, not Xenophanes’ question but a renewed and enlivened worship, or otherwise it would not be to bacchants and other night wanderers that Heraclitus would be prophesying. The sacrifices that bring a wholly inward purity (“apokekatharmenõn pantapasin”), as distinguished from the “material” ones (enula), are those of “few-well-numbered men” (B 69). The purification with which such mysteries are properly carried out is inseparable from the awareness of its conditions, as he also says:
They are purified, polluted with another blood as though one stepping into mud would wash off with mud. He would seem to be mad if someone took notice of him doing this; and they also pray to these images as if someone should converse with buildings, not knowing gods and heroes for what they are. (B 5)
“The gods and heroes” to whom statues stood in the Greek cities of Heraclitus’ time were largely Olympian ones, and the mysteries of purification were largely chthonic. Here it is claimed that purification and images are ineffectual if they are not accompanied by awareness. The split in Heraclitus between abstract theology and formal pre-Olympian secret cults does not abrogate the latter in favor of the former. For this very reason he does not have to mount Xenophanes’ attack on the Olympians. If “immortals are mortal and mortals immortal” there is a continuity between gods and man, questionable but ineluctable, perceived in the very use of the senses by the soul, “souls smell in Hades/ according to Hades” (kath’Haiden, В 98).
The split in Heraclitus’ presentation of religious phenomena is paralleled by the split between himself and the society in which he lived. Unlike the bard of oral culture, who summarizes what are simply the views of the tribe, even if he does so as consummately as Homer, Heraclitus questions the values of the tribe. His goal, he says in an aphorism, was himself, “I inquired after myself” (B 101).
The binary patterns of his statements, as I have said above, are reminiscent of the binary patterns that Whitman finds in the organization of incident in the Iliad. If mythic thought, the Olympian system or another, is understood as organized into series of binary oppositions, then the contradiction-system of Heraclitus would reproduce the functional procedure of mythology without having recourse to mythic figures or even, except rarely, to mythic tales.
This does not happen without stress between the thinker and his society. One of the few tales Heraclitus tells is that of Hermodorus, who we know from other sources to have engaged in an activity also characteristic of newly literate societies; he was a lawgiver. The Ephesians cast out Hermodorus because he was “most useful” (onestos) among them. For doing so they were “fit to be hanged, all but those who had not attained puberty” (B 121). So starkly does Heraclitus view the situation of someone like himself who separates himself from the multitude. And yet such a separation is what he insists on as a prior condition not only of proper religious observance, as above, but of wakeful thought itself.
The transition to literacy brings about such a self-conscious separation, an orientation towards the eye that can read seriatim words in a book that one person writes and one person reads, away from the ear that intimately and instantaneously hears a common metered message sung aloud to the accompaniment of music. Heraclitus, indeed, does stress the superiority of the eye to the ear explicitly, (“Eyes are sharper witnesses than ears,” В 101a) and the term visible or perceptible, phaneros, also stresses the eye. At the same time in his own role he has not lost the marked status that adheres to the scribe in societies where literасу is not yet common. He often uses the word “hear,’akouein, in the sense of “understand a doctrine.”35
The very fact that legends accrued to him, whether or not they were historical, attests to Heraclitos’ “tribal” status. He ate grass on the mountains like Nebuchadnezzar, he died of dropsy lying on a dung heap because of taking his own maxims too seriously, he played knucklebone with children rather than participate in political deliberation. These tales are reminiscent of those told about Pythagoras: that a fold in his garment at Olympia revealed his thigh to be made of ivory; and about Empedocles, that he plunged at night into the flaming crater of Etna but left his sandal on the rim.
At the same time, Heraclitus is sensitized towards the definition of social role, not only in the grim aphorism about Hermodorus. One of his three books according to Diogenes Laertius was politicon. The isolation of mental disposition (B 14) or mental perception (B 5) as the criterion for distinguishing pure sacrifices from impure amounts to an ethical one. And Gigon says, with some plausibility, “His central thought is ethical.” Heraclitus is already beyond the hieratic literacy of such tribal figures as Epimenides, who, according to the legend, slept in a cave fifty-seven years and had letters written all over his body at death.36
Lyric poetry already redirects the Homeric practice to the sensibility of the private consciousness. Heraclitus retains the centered act of consciousness when he strips his utterance of meter. And he goes further than the Xenophanes he criticizes for polymathie when he gets his acts of demythologizing and his acts of abstract theologizing into a single frame. That frame, however, itself includes, and talks about, a relation to society that the Milesians and Xenophanes, whatever their experience, did not discuss. Where they are inescapably but silently contextual, Heraclitus refers dialectically to his context. And further, he contains a reference to context in some of his most comprehensive utterances, as he does in В 1.
Sometimes he merely establishes the context by indicating what characterizes the imperceptive (В 19, В 20, В 29, В 34, В 71, В 73, В 74, В 86, В 87, В 89, В 97, В 110, В 116, В 117). Не can, however, establish the context graphically and forcefully, as in the aphorisms about exiled Hermodorus and the Ephesian adults (B 121) or Homer and the flea riddle (B 56). “One is ten thousands to me, if he be best” (B 49a). Aristos could also mean “possessed of arete,” and the absence of the definite article before aristos allows it to be read in more than an exclusive sense. It could refer to an entire aristocratic class. Heraclitus’ own definition of arete links it to wisdom, “sound thought (sophronein) is the greatest arete, and it is wisdom (sophia) for those who understand (epainontos) to speak the truth and act it according to nature” (B 112). “Best” would then involve a Heraclitean philosophical perception, and so it might be said to anticipate the philosopher-kings of Plato’s Republie; though, unlike Plato, Heraclitus says nothing in this aphorism to indicate that the qualification would not remain open. Any one could conceivably be worth ten thousands.
There is a tension between the common and the private here, which must be maintained for either aspect to be perceived. On the one hand, hoi polloi deludedly suppose that they have private knowledge when the logos is common (B 2). On the other hand, hoi polloi have a tendency to sink back into the common, and that, too, is to be avoided:
What is their mind (noos) or their sense (phren)! They hearken to the singers of the people and they employ the crowd (homilo) as a teacher, not knowing that the many (hoi polloi) are bad and the few are good. (B 104)
Here is the Homeric context of ear-oriented transmission in a nutshell. It would not matter that in one sense a preeminent bard might be one of the few rather than the many. By not comprising the tension between the many and the few, or between private consciousness and common language, his songs could not lead to “mind” or to “sense.” In such a state they could only “seem to themselves” (B 17). Yet “Those who sleep are performers and co-workers of what comes about in the kosmos’’ (В 75).
In focussing on this tension between the consciousness of the aphorist and the society where he finds himself, Heraclitus still falls back on its conventional emblems. “They do not understand that what is borne apart is borne together with itself: a back-turning harmony-fit, as of a bow or a lyre” (B 51). What they do not understand is what they find as common social objects around them. The bow here gets the name toxos rather than the name he puns with the same name for life in В 48. “The name (onoma) of the toxos is bios, but its work is death” (B 48). In the Homeric poems the bow is a chief instrument for performance (ergon, “work”); the lyre is a chief instrument for unifying the act of using words (onoma). The bow and the lyre serve as emblems to make clear the distinction between onoma and ergon, as the word bios in Homer does not. The bow and the lyre also serve as emblems for the Homeric society from which at the same time they are abstracted. To see them as “back-turning,’37 itself a revision of the Homeric epithet “back-sounding” (palintonos), reconceives the conditions of the Homeric society and ultimately transposes them,38 without performing Plato’s act of rejection upon them.
In this verbal act of Heraclitus’, we are not yet able to separate “poetry” from “philosophy” even though his statement itself constitutes a comparable act of separating them. At the same time, this verbal act remains applicable for us in its striking conflation of poetry and philosophy, where Heraclitus’ notions about fire, should we have correctly disentangled them, belong at best to the early history of science, and his aphorisms about flux express what is by now a truism. The doctrine of flux, as a principle of change, however, became a truism only after it had been subjected to elaborate analysis at the hands of both Plato and Aristotle, who would seem to have got very largely from Heraclitus a focus on the principle of change which is central to the epistemology and the ontology of both philosophers. If they departed from the powerful verbal technique in which he had couched it, that technique remained, to be revived perhaps accidentally by Pascal, by Blake, by Nietzsche, and Wittgenstein.39 Its force is a permanent resource.