A philosophical account of knowledge can be descriptive; or it can be revisionary. A descriptive account will attempt to capture our actual use of the knowledge-vocabulary and the knowledge-concept. (Or -concepts: it will not necessarily aim at a unified account of knowledge.) Describers of knowledge will welcome unity and generality where they find them anyway, and will admit that, unless the knowledge-words had quite general meanings, they could not have quite general usages. Still, for the descriptivist the general tendency of her account of knowledge is not to fix a decreed unity, but to explore a given multiplicity. Extreme cases of the descriptive approach to knowledge will closely approximate, perhaps even instantiate, a dictionary entry for the word ‘knowledge’.

By contrast a revisionary account of knowledge aims not to observe the way we use our knowledge-concept, but to change it. Revisionists insist on the special credentials or the paradigm status of some particular kind of knowledge: maybe first-personal knowledge (Descartes); immediate empirical knowledge (Locke, Berkeley, Hume, the young Russell); immediate logical knowledge (Leibniz); mathematical knowledge (Frege); or scientific knowledge in the modern sense of ‘scientific’ (Quine, Churchland, much contemporary culture or anti-culture). The revisionist insists that, in the interests of epistemic progress, we should judge all other kinds of knowledge or putative knowledge by the canon of this paradigm kind.

Usually just three verdicts are possible outcomes of this judging process: (a) The ‘other kinds of knowledge’ turn out to be not really other kinds – they are actually instances of the paradigm kind. Or (b) they turn out to be not really knowledge – their claims to count as knowledge are unmasked as spurious and illusory. Or (c) these other kinds prove to be derivable from the paradigm kind in one way or another; the way of reduction, for instance, or again the way of emergence. Today's ‘science wars’ usually have precisely this structure.

To us this all seems a terribly modern predicament. Yet suppose a philosopher and dramatist of genius – one rather given to obliquity and indirectness in stating his actual views, at least in his writings – who lived in Athens 24 centuries ago had been struck both by the absolute certainty of the mathematics and geometry that he knew, and also by a certain transcendence or abstraction that he noticed about mathematical and geometrical objects and conclusions. It is a completely certain truth (he might have noted) that the square on the hypotenuse is equal to the sum of the squares on the other two sides. But it is a completely certain truth, just in so far as it is not a truth about this triangle or that hypotenuse, about this particular side here or that individual side there. Geometrical propositions clearly refer to something. But equally clearly they do not refer to the messy and inexact figures, the ‘squares’ and ‘triangles’, etc., that we may happen to draw in whatever sand or clay or papyrus is available to us. None of those figures is the triangle that Pythagoras’ theorem is about. None of them is even, strictly speaking, a triangle, as precise examination of them will always discover. Of any of the individual worldly ‘triangles’ that we actually perceive, it would be generous even to say what Plato says in the Republic: that it is just as much a triangle as it is not a triangle. (Compare e.g. Republic 478d and Theaetetus 182e.)

Suppose then that this philosopher sought, in revisionist mode, to reconstruct as much of knowledge as he could on the geometrical pattern; and suppose that this project led him both to a conviction that ethical claims were transcendent and abstract in something like the same way as geometrical ones, and also to an increased scepticism about the claim of perceptual appearances to count as knowledge at all.

Suppose further that this genius philosopher-dramatist found himself in philosophical confrontation with two other kinds of revisionist. For one of these, perceptual appearances are the only kind of knowledge, to which all other putative knowledge must either be reduced, or shown up as bogus because that reduction proves impracticable. For the other, perceptual appearances are indeed, as our own philosopher suspects, in some sense an illusion; but that is not because real knowledge is about transcendent abstracta, it is because real knowledge can only be about matter in motion – and paradoxically, our only access to this kind of knowledge is via our illusory perceptions.

Given all these suppositions, we might go on to ask what kind of book we might expect our ‘philosopher-dramatist of genius’ to produce, if his aim was to move from the direct exposition of his own revisionism about knowledge to a tentative exploration and critique of these two other forms of revisionism. My suggestion is that he would produce something very like Plato's Theaetetus, and hence, also, that the chief targets of the Theaetetus are the two great Abderite philosophers – Protagoras and Democritus.

This is one reading of the Theaetetus; I defend it in Reading Plato's Theaetetus (2005). (Though that book predates my present conviction that Democritus is the chief, though unnamed, target in Tht. 201–10, where the moral is that semantic structuring of the kind involved in knowledge or even belief cannot be derived from an ontology of atomic simples alone, irrespective of the relation of those simples to perception.) Such a reading presupposes that Plato is at every stage of his career a revisionist about knowledge of some sort. But revisionism about knowledge is a contestable view, like other forms of revisionism; the revisionism/descriptivism contrast plays out in other areas of philosophy besides epistemology, for instance in ethics, in political philosophy and in the area where Strawson originally noted it, metaphysics.

McD.'s classic commentary on the Theaetetus is an important monument to the possibility of reading Plato in as Wittgensteinian, and so descriptivist, a way as possible. That turned out to be a good deal more Wittgensteinian than we might have expected; or indeed than Wittgenstein, or Plato, might have expected. The dialogue is, after all, labelled as peirastikos (‘tentative’) by its ancient editors, and includes a good deal of what is surely intended as exploratory material. Naturally a quasi-Wittgensteinian reading stumbles when confronted with the abundant evidence of Plato's interest in sense-data and what can be inferred from them; at least if one assumes, as McD. rather tended to, that respectably Wittgensteinian philosophers would talk as little as possible about such disreputable entities as sense-data. Above all, such a reading is hard to square with the virtual manifesto for the revisionist ambition with which the Theaetetus opens – Socrates’ famous rejection of Theaetetus’ attempt to define ‘knowledge’ by examples (145e–7c), and his reminder that much more than that is possible in geometrical definitions (147c–8e) – both of which are surely intended to remind us of the Theaetetus’ most significant and most philosophically cognate precursor, the Meno, which is also in its different way a manifesto for revisionism.

McD.'s classic commentary on the Theaetetus was a companion to his equally classic translation of it. In the present edition it is just McD.'s translation, only very slightly revised in light of the new OCT, that is offered The judicious commentary and notes are by L. Brown. Her course through the various controversies wears profound learning with admirable and novice-friendly lightness – and is a good deal more impartial than McD.'s was. Her introduction is not for that reason any less absorbing or entertaining. She reminds us, for instance (p. xxv), of McD.'s claim that ‘by the end of the dialogue Plato has lost interest in the definitional task, given the rich harvest of other philosophical problems that the latter part of the discussion has thrown up’. That would, no doubt, be a rather Wittgensteinian change of tack – though it would also seem a little absent-minded of Plato, no matter how rich the harvest that thus distracted him. The translation itself is a delight to reread, and hardly ever flags either in accuracy or in idiomatic verve. One could confidently use this volume as the primary text in a course on the Theaetetus for undergraduates with no Greek – though every serious Plato scholar will want to have access to it for their own research too.