To: Cornell J.
From: Geoffrey Klempner
Subject: Atoms and voids 2500 years ago
Date: 29th May 2009 11:33

Dear Cornell,

Thank you for your email of 20 May, with your fourth essay for the Ancient Philosophy program, entitled, 'atoms and Voids 2500 Years Ago.'

You start off with a useful survey of all the main theories of the Presocratics leading up to the atomists Leucippus and Democritus. I'll skip this, as there isn't anything that you infer as a consequence apart from the reaction to Parmenides. (Anaxagoras is interesting, because he hypothesized a universe where there are no 'smallest parts', evidently not feeling the grip of Zeno's paradoxes regarding infinite divisibility.)

One point to make is that atomism is not a physical theory, not even in the sense in which we can say e.g. that the Milesians put forward physical theories. (Lacking any sense of experimental method, it might be said that they were more like physical 'speculations'.)

What is a theory? A philosopher of science would say that a theory is an 'inference to the best explanation' (Pierce called this 'abduction' to distinguish it from 'induction'). A theory is an account which we can't prove to be true (cf. the scepticism of Xenophanes) but which we prefer over other accounts because of its elegance, simplicity, and fecundity in accounting for observations. From Dalton's atomic hypothesis, right up to present day particle physics, the aim is not to prove how things must be (as metaphysics aims to do) but rather put forward the best theory which accounts for the things we observe.

By contrast, the atomists did not hypothesize or speculate. Atomism lays down how things must be, in the only logically possible universe. As you observe, the one deviation from Parmenides which the atomists allowed was the existence of void. The reasoning was purely logical. If, as is evident, things do move and change, then Parmenides *must* be wrong, even if we can't find the fallacy in his argument. An analogous situation arises when one gives a counterexample to a putative logical inference. If you can find a valid counterexample, then you know that the inference is invalid, even if you are unable to pinpoint the fallacy.

Admittedly, Parmenides would not agree with this view. This is where something unquantifiable enters the equation, the philosopher's 'judgement'. (This is the point discussed in the units on Parmenides and Melissus, how important you consider 'saving the phenomena'.)

However, it is not immediately apparent why you need void in order to allow for movement. Why couldn't the universe consist of atoms moving about in chicken soup (or 'ether' if you prefer)? It is just not true that things move only if there is 'empty space' to move into.

The answer, of course, is that chicken soup, if it has (or is) Being, cannot alter in any way. Fluidity is the capacity for potentially infinite alterations of shape. So nothing that has Being can be fluid.

But that still leaves the possibility that the universe could consist of perfectly cubical atoms, extending infinitely in every direction, in full contact with one another. I would argue that the atomists were wrong if they thought that the ability of one object to slide against another object requires an interstitial void. Right or wrong, however, this won't do as a model intended to 'save the appearance's for the simple reason that the movement allowed is far too limited.

(I had to stop when I wrote the last sentence, because although it seems obvious enough, I couldn't be absolutely sure. Consider the way that every possible picture is represented on a TV screen or monitor by a simple moving dot scanning across the screen at great speed. Couldn't we somehow reproduce the entire universe with extremely fast movements of cubical atoms sliding past one another? Can you think of a clinching argument, for or against?)

Atoms are tiny, too tiny to see. This seems to go almost without saying but in fact the atomists appear to have held, by the 'ou mallon' principle (='no more reason', cf. Anaximander), that just as there must be atoms of every possible shape, so there must be atoms of every possible size. There just aren't any of the 'big' atoms in our part of the universe. This is just another illustration of the difference between the metaphysical theory of atomism, and the modern atomic hypothesis.

What are the consequences for knowledge? You say, 'Genuine knowledge consists of atomic structure of objects; the shape, arrangement and size of atoms.' Obviously, no-one knows this. Does that mean that there is no genuine knowledge? The atomists would say that what we do know, with absolute certainty, is the truth of the theory of atomism. As for the 'illegitimate knowledge', it can't be that illegitimate if the evidence of the senses (that things move and change) is the necessary premiss that entails the existence of the void. There is a quote relating to this point in the fragments/ testimonia. If you reject altogether the evidence of the senses, then that is indeed the 'downfall' of atomism.

You are right, though, when you talk about Greek atomism anticipating Locke's account of primary and secondary qualities. It is of course also true that modern atomic theory (from Newton and Dalton onwards) owes its original inspiration to the pioneering ideas of Leucippus and Democritus.

All the best,

Geoffrey