[In the following excerpt, Barnes discusses Democritus's skepticism regarding humankind's ability to know anything with certainty.]

DEMOCRITEAN SCEPTICISM

Metrodorus of Chios, a pupil of Democritus (e.g., Clement, 70 A 1) who held solidly to the main tenets of atomism (e.g., Theophrastus, A 3), purveys an extreme scepticism which foreshadows, in its ingenious comprehensiveness, the most extravagant claims of Pyrrho: at the beginning of his book Concerning Nature Metrodorus said:

None of us knows anything, not even that very fact whether we know or do not know; nor do we know what not to know and to know are, nor, in general, whether anything is or is not.

(505: B 1)¹

Of Metrodorus' book little else survives and nothing tells us what his scepticism rested upon, or why he wrote Concerning Nature at all. His scepticism, however, like his atomism, was inherited. For according to Democritus,

In reality (eteêi) we know nothing; for truth is in a pit.

(506: 68 B 117)

Our main source for Democritus' scepticism is Sextus; and I quote the chief Democritean fragments in their Sextan setting:

Democritus sometimes does away with what appears to the senses. … In the Cratunteria, though he had promised to ascribe the power of conviction to the senses, he is none the less found condemning them; for he says:

We in actuality grasp nothing firm, but what changes (metapipton) in accordance with the contact (diathigên*)² between our body and the things which enter into it and the things which strike against it [= B 9].

And again he says:

Now that in reality (eteêi) we do not grasp of what sort each thing is or is not, has been made clear in many ways [= B 10].

And in Concerning Forms he says:

A man must know by this rule that he is separated from reality (eteê) [= B 6].

And again:

This argument too makes it clear that in reality (eteêi) we know nothing about anything; but belief (doxis) for each group of men is a reshaping (epirhusmiê) [= B 7].

And again:

Yet it will be clear that to know what sort each thing is in reality (eteêi) is inaccessible [= B 8].

In those passages he pretty well destroys apprehension in its entirety, even if he explicitly attacks only the senses. But in the Canons he says that there are two kinds of knowing (gnôseis), one via the senses, one via the intellect (dianoia); he calls the one via the intellect ‘legitimate (gnêsiê)’, ascribing to it reliability for the judgment of truth, and he names that via the senses ‘bastard (skotiê)’, denying it inerrancy in the discrimination of what is true. These are his words:

Of knowledge (gnômê) there are two forms, the one legitimate, the other bastard; and to the bastard belong all these: sight, hearing, smell, taste, touch. And the other is legitimate, and separated from that.

Then, preferring the legitimate to the bastard, he continues:

When the bastard can no longer see anything smaller, or hear, or smell, or taste, or perceive by touch, † but more fine † [= B 11].

Thus according to him too, reason, which he calls legitimate knowledge, is a criterion.

(507: adv Math VII. 135-9)

Fragments B 7 and B 10 show that Democritus' scepticism was not merely a glum asseveration of intellectual impotence, but the melancholy conclusion of a set of arguments. Two of Democritus' arguments can, I think, be reconstructed.

First, there is doxis epirhusmiê of B 7. I suppose that ‘doxis epirhusmiê’ means ‘belief is a rearrangement of our constituent atoms’, i.e. ‘coming to believe that P is having certain parts (e.g., cerebral parts) of one's atomic substructure rearranged’ (cf. Theophrastus, Sens §58 = A 135).³ Belief, then, cannot ever amount to knowledge, because it is never anything more than an atomic rearrangement. I guess that Democritus is supposing, if only tacitly, that knowledge is essentially reasoned belief: opinion not arrived at by rational considerations cannot qualify as knowledge. But if every belief is simply a cerebral alteration (caused, no doubt, by our changing relation with other atomic conglomerates), then no belief can be rational. To put it crudely, causally determined cerebral mutations cannot be identical with rationally accepted beliefs.

The argument has connexions with Xenophanes (vol. 1, p. 142); but it is less subtle and less persuasive than Xenophanes' argument. According to Xenophanes, certain types of causal chain prevent a caused belief from counting as knowledge; according to Democritus, any belief, being the physical result of a causal chain, is disqualified from knowledge. Democritus, I think, is simply wrong: my belief that P may constitute knowledge even if it is itself a physical state (a state of my nervous system) and even if it stands at the end of a causal chain (as surely it does). Roughly speaking, the belief is knowledge if the physical state which embodies it was caused, mediately or immediately by the fact that P (i.e., if it is true that because P I believe that P); and the belief is rational if the physical state which embodies it was caused by certain other beliefs (i.e., if because I believe that Q I believe that P, where Q in fact gives good grounds for P). If a causal theory of knowledge can be worked out in detail, then Democritus' argument for scepticism in B 7 must be rejected.

Second, there is B 9. Sextus evidently thinks that Democritus means ‘perceive’ by ‘grasp (sunienai)’; and he may be right. But Democritus is not simply ‘condemning’ the senses: he is offering an argument. The point, I think, is this: cognitive processes are interactions between observers and objects of observation; the processes, atomically construed, consist in the impingement of atoms from the object on the body of the observer. Now any such process involves a change in the object; for it loses at least those atoms which impinge upon the observer. Consequently, we can never know the state of any object; for any attempt to discover it thereby changes it. We grasp nothing ‘firm’; for our very grip disturbs. Knowledge alters the known; and therefore knowledge is impossible.

According to modern physical theory, we discover the position and characteristics of an object by way of some physical interaction with it: in the simplest case, I see where the cat is by shining a torch on it and receiving the reflected rays. What goes for cats goes for sub-atomic particles; to tell where a particle is I must fire a ray at it and receive it on the rebound. But sub-atomic particles are delicate things, and when a ray hits them they are shaken; thus the reflected ray will not give me the information I want. It cannot tell me where the particle is and how it is travelling; for the impact, without which I can know nothing of the particle, will change the particle's trajectory. (That is meant as a kindergarten version of the reasoning behind Heisenberg's Indeterminacy Principle; science for the infant is usually bad science, but I hope that the point of my parallel is not wholly blunted by my puerile exposition.)

Atomic structures cannot be known; for the process of acquiring knowledge necessarily distorts those structures. The quest for knowledge is like the search for the end of the rainbow: we can never discover the pot of gold; for our journey towards the rainbow's end in itself moves the rainbow to a different and ever distant location.

The argument that I have dredged from B 9 is not a priori: it depends on Democritean physics and psychology. I guess that it may present a plausible deduction from those Atomist theories, though I doubt if there is enough evidence for us to test its validity. In any case, there is no philosophical way of attacking it: it fails if the physics and psychology are false (and I assume that they are).

Metrodorus of Chios said that no one knows anything: the things we believe we know we do not strictly (akribôs) know; nor should we attend to our senses. For everything is by belief.

(508: Epiphanius, 70 A 23)

Leucippus insists that we have belief, but no more (Epiphanius, 67 A 33); and in many of the fragments I have quoted, Democritus denies that we have genuine knowledge. Many sceptical philosophers seem to be making what is little more than a verbal point: we do not, strictly speaking, know anything, but we can, of course, have reasonable beliefs. Such thinkers set the canons of knowledge artificially high: knowledge must be certain, or infallible, or necessary, or indubitable, or whatever. If the canons are set high, then knowledge is indeed beyond us; but ordinary men are quite happy with relaxed canons, and those sceptics who allow reasonable belief in fact allow precisely the thing that ordinary men call knowledge.

The Atomists, however, do not even allow reasonable belief: their arguments against knowledge, in so far as we know them, are equally arguments against reasonable belief. We have beliefs: that is an incontestable empirical fact. Our beliefs do not amount to knowledge: that is the argument of the Abderites. Yet our beliefs are not even reasonable: being atomically caused, they are not founded on reason; and the physics of the cognitive processes assures us that no impressions of external reality are accurate. If there is no room for knowledge, by the same token there is no room for reasoned belief: ‘everything is by belief’—but that, far from being a consolation, is only a cause for despair. The urbane scepticism of Locke allows a decent wattage to the human candle: our light extends as far as we need, but not as far as we like to boast. Abderite scepticism is Pyrrhonian: the light of the mind is an ignis fatuus.

That conclusion did not please Democritus; indeed, as Sextus observes, his fragments do not exhibit consistency. Fragment B 11 tails off into corruption; but the general sense of Democritus' remarks is clear enough: ‘the bastard way of knowing (skotiê gnôsis)’ will not carry us to the finest or ultimate constituents of stuff; for that, ‘the legitimate way of knowing (gnêsiê gnôsis)’ is needed. That coheres with Democritus' approval of the Anaxagorean slogan: opsis tôn adêlôn ta phainomena—what the senses cannot apprehend must be grasped by the intellect. There seems, then, to be an empiricist Democritus rising in revolt against the sceptic.

And perhaps the sceptical fragments have been misread: the Heisenbergian argument, after all, at most shows that we cannot directly apprehend the atomic elements of things; it does not show that we may make no inferences from perceptible things to their elemental structure. B 9 and B 10 consistently say that we cannot ‘grasp’ things in their reality; but that only means that atoms are not open to perceptual knowledge.⁴ Thus we may find a positive epistemology for Democritus: ‘All knowledge rests on perception: and perception will not, directly, yield knowledge of what exists eteêi. But by perception we may come to know about what is nomôi, and intellectual attention to those sensual pronouncements will enable us to procure an inferential knowledge of genuine reality.’

Alas, that happy picture is mistaken. The doxis epirhusmiê argument is resolutely sceptical; and B 6, B 7, B 8, and B 117 leave no room for any knowledge at all. Moreover, Democritus recognized that the empiricist intimations of B 11 were misleading:

Having slandered the phenomena … he makes the senses address the intellect thus: ‘Wretched mind! Do you take your evidence from us and then overthrow us? Our overthrow is your downfall.’

(509: B 125)

In a puckish mood, Russell once observed that naive realism leads us to accept the assertions of modern science; and that modern science then proves realism false. Realism is false if it is true; hence it is false. And if science rests on realism, then it is built upon sand. The parallel with Democritus is plain: the observations of the senses give us a set of facts upon which an atomistic science is reared; the science then proves the irrationality of all belief and the unreliability of the senses. If the senses are to be trusted, they are not to be trusted; hence they are not to be trusted. And if atomism rests upon the senses, then atomism is ill founded.

Did the mind answer the senses? Had Democritus any solution to the problem which 509 candidly poses? There is no evidence that he had; and I am inclined to think that he had not. It is, I suppose, a tribute to Democritus' honesty that he acknowledged his plight; but it derogates somewhat from his philosophical reputation that he made no move to escape from the impasse he found himself in.

Notes

1. Cf. Sextus, 70 A 25; Philodemus, A 25; Epiphanius, A 23; Diogenes Laertius, IX.58 = 72 A 1. The text of 505 is reconstructed in part from Cicero's translation, and the details are far from certain.

2. The MSS. read diathêkên (‘disposition’). …

3. The word ‘epirhusmiê’ itself has foxed the scholars; de Ley plausibly suggests ameipsirhusmiê (cf. B 139), which would have the sense I give to epirhusmiê.

4. And some translate idmen in 506 as ‘know by experience’ (see especially Cleve, 428-31). Compare Fränkel's version of Xenophanes' sceptical fragment (vol. 1, p. 138). But idmen means no more than ‘know’; and B 7 and B 8 are more than enough to impose a scepticism on Democritus.

Appendix A: Sources

Our knowledge of the Presocratic philosophers is almost entirely indirect; for even where we possess their actual words, those words are preserved, fragmentarily, as quotations in the works of later authors. The sources we rely upon for testimonia and fragments span two millennia: they differ widely, one from another, in their literary aims, their historical competence, and their philosophical interests.

This appendix lists in chronological order the ancient authors I have quoted from or alluded to in the text and the notes. Some of the authors are (from a Presocratic point of view) of minor or minimal importance. A single asterisk is prefixed to the names of the more freely flowing sources; and those few gushing streams are marked by a pair of stars. Each name is followed by a date, often roundly given, and the briefest of biographical sentences. When a ‘principal work’ is named, that is not necessarily the author's major opus, but rather the book which holds most interest for students of the Presocratics.

Where no edition of the ancient text is mentioned, the reader may assume that I have used only the excerpts printed in Diels-Kranz. In citing editions I use these abbreviations:

CIAG: Commentaria in Aristotelem Graeca (Berlin, 1881-1909)

OCT: Oxford Classical Texts

SdA :Die Schule des Aristoteles, ed. F. Wehrli (Basel, 1967-9²)

theophrastus: 371-287; Aristotle's greatest pupil and his successor. Only fragments survive. Abbreviation: Sens de Sensu Edition: Diels [4].

philodemus: c.80-c.35 bc; Epicurean philosopher, fragments of whose works were discovered in the lava of Vesuvius.

diogenes laertius:? third century; scissors and paste historian of philosophy. Work: Lives of the Philosophers. Edition: OCT, Long.

sextus empiricus: fl. 180-200, massive compiler of sceptical topoi and our main source for ancient scepticism. Abbreviations and Editions:

adv Math: Against the Mathematicians (Teubner, Mau)

Pyrr Hyp: Outlines of Pyrrhonism (Teubner, Mau).

epiphanius: c.315-403, bishop of Salamis. Edition: Diels [4].

F. M. Cleve: The Giants of Pre-Sophistic Greek Philosophy (The Hague, 1965)

H. Fränkel: Early Greek Poetry and Philosophy, trans. M. Hadas and J. Willis (Oxford, 1975)

H. de Ley: ‘… critical note on Democritus Fr. 7’, H 97, 1969, 497-8