by A. A. Long
28 There are uncertainties about the text. The first word may be “fittingstogether” (synapsies); it is not certain that the other clauses all belong together.
29 If we may set aside the putative solution offered by B102:
To God all things are fine and good and just: but human beings have supposed some things to be just, others to be unjust.
There are philological grounds for doubting the authenticity of this remark, which is also out of line with Heraclitus’ treatment of opposites (see section 3).
30 The translation “draughts” is conventional; the board game in question (pessoi) was closer to backgammon.
31 B124 (on the interdependence of large-scale order and small-scale chaos?) may also be relevant.
32 “Those who have lived with the logos are Christians, even though reputed godless, such as were, among the Greeks, Socrates and Heraclitus and those like them” (Justin, Apol. 46.3).
33 Schleiermacher [260]; Hegel [22] (vol.1, 279: “There is no proposition of Heraclitus which I have not adopted in my Logic”). The next substantial contributions were Jakob Bernays’ early work (1848-54 = Bernays [237] 1–106), and Ferdinand Lassalle’s monograph of 1858 (Lassalle [249]).
34 There is still, for example, no comprehensive study of Heraclitus and the Stoics (but see Long [251]; Dilcher [239] 177-200). On Heraclitus in the Christian writer Hippolytus (an important source), see especially Mansfeld [51]; also Mueller [53] (a review of, and corrective to, Osborne [52]).
35 On the new evidence in the papyrus found at Derveni in 1962, see now Sider [262], Tsantsanoglou [263], which contain the best available readings of the relevant part of the text.
36 This context, besides Homer, Hesiod, and the Ionian natural philosophers, may include the Ancient Near East, Judaism of the exile period, and early Zoroastrianism.
37 I am indebted to all those who over the years have helped me in understanding Heraclitus, and in particular to Mantas Adomenas, Roman Dilcher, and David Wiggins.
DAVID SEDLEY
6 Parmenides and Melissus
Parmenides and Melissus were bracketed in antiquity as the two great exponents of the Eleatic world-view which denies change and plurality.1 In modern times their treatment has been curiously unequal. Too much has been written on Parmenides – albeit the greater thinker of the two – too little on Melissus. Too much has been said about Parmenides’ use of the verb “be,” while too little has been said about his detailed arguments for the individual characteristics of what-is. However, neither these nor other anomalies should disguise the immense wealth of scholarship that has furthered the reconstruction of their Eleaticism.
PARMENIDES
Around 150 lines of Parmenides’ hexameter poem, written in the early- to mid-fifth century, have been recovered, most belonging to its first part. His densely metaphorical diction is replete with Homeric echoes, and presents the further difficulty of having to use the very language of change and plurality that it aims ultimately to outlaw. These are among the many aspects to which it will be impossible to do justice in the present chapter.
The poem opens with an allegorical description of Parmenides’ journey to the House of Night, mythologically located where the paths of day and night join.2 This symbolizes Parmenides’ intellectual journey of distancing himself from a phenomenal world in which (as the second half of his poem will explain) light and night alternate to produce the illusion of plurality and change.3
There a goddess addresses him, promising to expound “the unshaken mind of well-rounded truth,” and the unreliable “opinions of mortals.” These correspond to the two halves of the poem, respectively the “Way of Truth” and the “Way of Seeming.” The entire philosophical exposition is delivered by the goddess herself. She may be taken to represent the god’s-eye view of being that Parmenides’ arguments have enabled him to attain for himself. There is no question of her discourse being mere divine revelation: every step towards the truth is hard won by argument.
The Way of Truth
“Come now, I will tell you (and see that you attend to the story you hear) which are the only paths of inquiry that can be thought of” (DK 28 B2.1-2). The goddess’ argument proceeds as follows.
(1) She offers a choice between two paths: “Necessarily (it) is” and “Necessarily (it) is not” (B2.3-5).
(2) She argues against the latter, and hence indirectly in favour of the former.
(3) She warns Parmenides against a third path (B6.4–9), a “backturning” one representing ordinary human acceptance of a variable world – the path of know-nothing “two-headed” mortals, who somehow manage to conflate being and not-being.
If we are to see what this is all about, some preliminaries must be clarified. First, “(It) is” is conveyed by the single Greek verb esti. Greek does not require that the subject always be expressed: hence esti, unlike English “is,” functions as a grammatically complete sentence. As for why no subject is made explicit, the safest answer is that at this stage we are still investigating the logical behaviour of the verb “to be.” Only in the light of that investigation will we be able to answer the question what can stand as the subject of “is.” Thus, identifying the proper subject of the verb “to be” is the final goal of the Way of Truth, not to be prejudged at the outset.
Second, what does “is” mean here? It has become traditional to offer a choice between at least the following: an existential or complete sense, ”… exists”; a copulative or incomplete sense,” … is…”; a veridical sense, ”… is the case” or perhaps ”… really is…”; and a fused sense, combining some or all of these. The main argument that lies ahead may seem to rely on the existential sense, but the third path, that of two-headed mortals who conflate being and not-being, represents acceptance of a variable world, and therefore should include ordinary empirical predications within its scope, for example, that the sky is blue and is not grey, that this animal is alive one day but is not alive the next: and these are incomplete uses of the verb.
The following, however, may be a safer way to proceed. It is widely recognized that the fundamental sense of “be” in Greek is incomplete, to be something. Often this something is made explicit: Fido is a dog, is the dog over there, is hungry, and so on. On other occasions it is left unspecified: Fido is. Modern readers may wish to call this latter a different sense of “is,” equivalent to “exists,” but to a Greek ear it is just a nonspecific use of the fundamental sense. To say, existentially, “Fido is” is merely to say that he is something (unspecified).
To read Parmenides’ poem, we must cling to this fundamental sense of “be.” Ordinary people consider the same things both to be and not to be, because, for example, the sky seems to them to be blue and not to be grey. Why should Parmenides object? Because he is wedded to a principle later expressed as, “The choice about these things lies in the following: (it) is, or (it) is not” (B8.15-16). This amounts to what I shall call Parmenides’ Law 1:
Law 1. There are no half-truths. No proposition is both true and false. No question can be coherently answered “Yes and no.”
Asked whether the sky is, a two-headed mortal is committed to the “Yes and no” answer that it both is (e.g., blue) and is not (e.g., grey). All ordinary human beliefs about change and plurality will on examination turn out to imply the same ambivalence about a thing’s being.
As for Parmenides himself, the reason why his own primary use of “be” in the Way of Truth looks existential is simply that, by Law 1, he can only contemplate total being or total not-being. To specify what a thing is, as mortals do, is implicitly also to specify what it is not, and thus to fall foul of Law 1. It is probably harmless for us to gloss Parmenidean being as existence (and for convenience I shall do so), so long as we do not forget that it arises as a logically sanitized case of ordinary Greek being, namely being something.
It is probably this sanitization that Parmenides means to convey by presenting the first two paths as “Necessarily (it) is” and “Nece
ssarily (it) is not.” Human viewpoints attribute being to things contingently and unstably, so that what-is can also not-be. In view of Law 1, this human outlook does not even start out as a formal possibility, and hence the goddess does not even initially list it among the conceivable paths, which she limits to propositions about necessary being and not-being. She later adds the contingent third path, not because it is even a formal possibility, but because despite its hopeless incoherence it is what ordinary mortals actually believe.
We can now proceed to the goddess’ refutation of the path ”… is not.” Her first argument is: “For you could not know that-which-is-not (it can’t be done), nor speak of it” (B2.7-8). How does this work? We may take it that to reject ”… is not” is tantamount to showing that this negated verb could never be supplied with a subject. And how do you supply a verb with a subject? Either (i) by thinking of that subject, or (ii) by naming it. But (i) to think of something, you must, minimally, know what it is; whereas anything capable of standing as subject of ”… is not” would not be anything at all (given Law 1), in which case, you could hardly know what it is! And (ii) by the same token, since the item in question would be nonexistent, it becomes hard to see how you could succeed in naming it: it simply is not there to be referred to.
Her second argument is even more condensed: “(1) What can be spoken and thought of must be. (2) For it is able to be, (3) whereas a nothing is not able to” (B6.1-2). Typically, Parmenides argues backwards: (1) is the immediate ground for his conclusion, the outlawing of ”… is not” – if you want to supply ”… is not” with a subject, you must either speak of that subject or think of it; however, it is then instantly disqualified as subject of ”… is not,” because anything you can speak or think of must be. The grounds for this last point are then supplied: (2) what can be spoken and thought of at least can be (in that it is conceivable?); but (3) a nonexistent thing (“a nothing”) cannot be (it is inconceivable that there should exist a nonexistent thing); therefore, what can be spoken or thought of cannot be a nonexistent thing, that is, it must exist.
This is a lot of flesh to put on so skeletal an argument. But the goddess now adds “I bid you think that over” (B6.2), acknowledging that her argument needs some fleshing out. She has now established what I shall call Law 2:
Law 2. No proposition is true if it implies that, for any x, “x is not” is, was or will be true.
Laws 1 and 2 will ground all her subsequent arguments.
She proceeds (B6.3-9) to deride the hopelessly confused path of mortals, whose mistake is traced to their reliance on the senses. The alternative approach that she advocates involves abandoning the senses in favour of pure reason (B7).
At this point she launches into her positive account of what-is (B8.1-49). Taken literally, what-is will prove to be an everlasting, undifferentiated, motionless sphere. How is this to be understood? If the sensible world is an illusion, is she describing the reality that actually occupies the place that the sensible world just seems to occupy? Or is she describing a reality as nonspatial and nontemporal as, say, numbers are? To put it another way, how far are we meant to deliteralize the description of what-is? I offer the following reason for retaining an unashamedly spatial reading. This final stretch of the Way of Truth is full of arguments. Most commentators are disappointingly silent on their structure and content. Only if we take them in literally spatial terms, I submit, do they prove to be good arguments.
If I am right, Parmenides’ goal is to reject humans’ woefully perspectival view of the sphere (bounded by the sky) that constitutes their world, and to redescribe as a perfect undifferentiated unity that very same sphere. A familiar objection to so literal a spatial reading has long been that if what-is were a finite sphere it would be surrounded by what-is-not, that is, void, in contravention of Law 2. This objection illegitimately assumes the infinity of space. A century later Archytas still had to argue for the infinity of space,4 and Aristotle, followed in this by a long later tradition, could deny that there is anything, even void, beyond our world. A doctrine of infinite space may have had Pythagorean support by Parmenides’ day, and it certainly acquired considerable currency in the philosophy of the Ionian east, but in the west a philosopher as indebted to Parmenidean thinking as Empedocles could postulate a finite world with (apparently) no void beyond. The very idea of space as an entity that exists altogether independently of the occupying body was slow to emerge in Greek thought,5 and without it the expectation that space should continue even beyond the limits of its own occupant would not present itself as irresistible. Provided that Parmenides’ sphere is imagined from inside, like the sphere of our phenomenal world, and not from outside like a football, the need for empty space beyond need not be forced upon him.
The goddess’ description of what-is starts with a list of its predicates (B8.2-4): it is (a) ungenerated and unperishing, (b) a single whole, (c), unmoving, (d) perfect (teleion) or bounded (teleston) or balanced (atalanton).6 In what follows, these four appear to be proved in sequence. But first a remark about time is added, which it may be easiest to take as parenthetical, since, although supported in what follows, it receives no separate proof: “Nor was it, nor will it be, since it is now all together, one, continuous” (B8.5-6). This is perhaps to justify her exclusive use of the present tense in describing what it “is”: there is nothing to be said about what it was or will be, because once we see that it is a changeless unity we will appreciate that no past or future can be distinguished from its present. Whether this makes being altogether timeless, or simply abolishes the passage of time, is controversial,7 but her retention of “now” may favour the latter.
The proof of the double predicate (a), “ungenerated and unperishing,” starts with the former. The two arguments against the wholesale generation of what-is are: (i) that would mean that “It is not” was previously true, contrary to Law 2 (B8.6-9); and (ii) coming from nothing, there could have been no reason for it to spring into being when it did, rather than earlier or later (9-10) – a celebrated application of the Principle of Sufficient Reason. There follows a separate argument against its piecemeal generation: (iii) “In the same way it must be totally or not at all,8 and the strength of belief will never allow anything to come into being in addition to it, out of what-is-not” (11-13). That is, the generation even of a part would still defy Law 2 as effectively as wholesale generation does.
“Therefore Justice does not loosen it in her fetters and permit it either to come to be or to perish, but holds it firm” (13-15). This is the first mention of perishing in the argument, and “Justice” may represent parity of reasoning: the same arguments that eliminate generation are effective against perishing too. Strictly, however, argument (ii) cannot be reapplied to perishing: in what-is there could well be, for all we know at this stage, ample reason for its eventual destruction, for example, a terminal illness. However, arguments (i) and (iii) are easily adapted to perishing, which, whether wholesale or piecemeal, would entail ”… is not” coming to be true.
The goddess now moves on to predicate (b), “a single whole.” What-is is shown to be “not divided” or perhaps “not divisible” (22-25). It is perfectly continuous, with no distinct parts. Since there are no degrees of being – even limited not-being would contravene both Law 1 and Law 2 – there is nothing true of it at one point that is not equally true elsewhere. That is, it is “all alike,” so that no gaps or distinctions can be found within it.
Predicate (c), “motionless,” now follows (26–33). What-is is motionless in that it is “unstarting and unstopping” (i.e., presumably, it neither starts off nor comes to a halt), “since generation and perishing have been banished” (starting and stopping being, respectively, the generation and the perishing of motion). And it stays exactly where it is because “mighty necessity holds it in the bonds of a limit, which imprisons it on all sides” – that is, filling all available space up to its boundary, it has no room to move. The ground for attributing this boundar
y to it is then added: “For it is not proper for what-is to be unfinished: if it were, it would lack everything.” Absence of a boundary would be a form of incompleteness, and hence a lack; and since, by Law 1, it cannot be both lacking and not lacking, it would be totally lacking, and therefore nonexistent.
“Motionless” here has often been interpreted as “changeless,” and the limit as symbolizing “invariancy.” The danger that such deliteratization faces is that of diluting the argument into the trivial “It does not change because it does not change.” On the spatial reading that his language more naturally invites, Parmenides has a substantial argument. If he does also have an argument against change in general, it is the one against piecemeal generation (11-13), which could well include generation of new properties.
Particularly puzzling are lines 34–41 of B8. They seem to halt the flow, by separating the proof of predicate (c) from that of predicate (d), which follows at 42-49. Some have taken them to be somehow part of that final proof, others to be displaced from their correct position, others to be a summary of the results so far, and yet others a digression against empiricism. My own preference is for viewing this as the place where Parmenides corroborates monism, the thesis which later tradition most strongly associated with him. Before embarking on her final proof, that of the shape of what-is, the goddess must pause to demonstrate its singularity. She has already shown that it is not divided. But there remain three additional claimants to a share of being: (1) thought, (2) time, and (3) the plurality of ordinary empirical objects. Each is addressed in turn.
(1) “Thinking is identical to that with which thought is concerned”: thought is identical to its own object, what-is. “For in what has been said” – that is, in the goddess’ arguments so far – “you will not find thinking separate from being” (34-36). There has been much resistance among English-speaking scholars to attributing to Parmenides any such identification of thinking with being. Yet it is the only natural reading of B3 (of uncertain location), “For it is the same to think and to be.”9 Besides, the price of not identifying thinking with being is to undermine his monism, by separating the thinking subject from the object of thought, that-which-is. Parmenides does not deny that thinking happens, but since being is all that there is, he must deny that thinking is separate from being. So we must take him to hold that what thinks is, and that what is thinks. That may be why in the proem (B1.29) the goddess promised to teach Parmenides the “unshaken mind of well-rounded truth.”10 The conflation is not altogether surprising in a context of early Greek philosophy. Anaximander, Anaximenes, and Heraclitus had all treated their primary existent, the stuff of the universe, as divine. And Parmenides’ follower Melissus, as we will see, likewise speaks about his own One as if it is a living being.