Like Plato, B. writes good openings. For instance: ‘God, as we know, did not leave it to Aristotle to make men rational. But he did leave it to Aristotle to make philosophers take account of the fact that there is more to men's rationality than the ability to construct syllogisms’. So begins ‘The Origins of Non-Deductive Inference’, one of 29 articles gathered in this new two-volume collection (quoted from vol. 1, p. 112). All but one previously unpublished piece appeared between 1971 and 1998.Footnote ¹ This material forms a substantial part of one of the most widely admired and influential bodies of work on ancient philosophy in the English language. Most scholars in the field will be revisiting a good number of the pieces, rather than reading them for the first time. This threatens to make the task of reviewing the volume superfluous. It may well be that you want to know only whether the volume is well-produced, to which the answer is yes: particularly helpful is the indication of original pagination. I also found almost no typographical errors, apart from a few Greek letters that do not appear properly and one more significant mistake (on p. 243 in vol. 2, ‘All soul is mortal’ should read ‘All soul is immortal’).

It is particularly welcome to have B.'s studies collected like this, since his work has been distinctive in appearing largely in article form. There are exceptions, of course: his book A Map of Metaphysics Zeta (2001), and his invaluable commentary on Plato's Theaetetus, accompanying the Levett translation (1990). In fact many of the pieces here form the background to that commentary – the Theaetetus is by far the most frequently discussed text. One of the pleasures of this new collection is the opportunity to see how B., over the course of about three decades, explored not only the Theaetetus but also a number of issues in ancient philosophy related to that dialogue. (For this sake I would have preferred that the pieces be presented in chronological order, rather than in thematic groupings, which are in any case fairly loose.) Most obviously there is B.'s pioneering work on ancient Scepticism. This goes hand-in-hand with his work on the self-refutation accusation thrown against Protagoras in the Theaetetus (see the two pieces on ‘Protagoras and Self-Refutation’, both from 1976). From there, B. went on to explore fundamental issues concerning this tradition, including one of the most important studies of what ancient Scepticism actually amounts to (‘Can the Sceptic Live His Scepticism?’ [1980]). B. is right, in my view, to insist on a radical interpretation according to which the Sceptic is actually trying to live without beliefs.

His work on Scepticism and sceptical arguments also touches on more specific issues. For instance he uncovers the meaning of an obscure metaphor in Lucretius, where he accuses the Sceptic of self-contradiction by saying that he stands on his head in his own tracks (B. proposes that this reflects an Epicurean phrase, περικάτω τρέπεσθαι). He contributes a subtle discussion on the paradox of the heap, which again rightly points to the importance of the dialectical context in which the sorites question was posed and answered – or as Chrysippus advised, first answered, then not answered, and then answered again. That emphasis on dialectical contexts is another leitmotif. Repeatedly, B. points out how objections that may strike us as unconvincing or incomplete could have been devastating in a dialectical exchange. This helps us to see why the Sceptic cannot get away with denying the possibility of proof (vol. 1, p. 13), and why Socrates might be on firm ground in rejecting lists of examples in place of definitions (vol. 2, p. 13).

Another striking philosophical theme that runs throughout these studies is B.'s interest in what one might call ‘second-best’ epistemic states. Most central here is a tremendously insightful article on Aristotle's Posterior Analytics, which points to the high standards Aristotle thinks need to be met by ἐπιστήμη in the strict sense. This need not prevent Aristotle from recognising other cognitive states – for instance in which one has true beliefs about particulars instead of universals – as a kind of ‘knowledge’. It is just that Aristotle is out to describe a systematic kind of knowledge that we might better classify as ‘understanding’. This is a crucial insight which helps us to understand other ancient thinkers too, for instance Plato (e.g. vol. 2, pp. 9, 67, 112) and Augustine (vol. 2, p. 178). It can be extended further still, to other periods of the history of Aristotelianism.Footnote ² For ancient authors, there are at least two broader implications. First, that not all desirable cognitive states need qualify as knowledge in the strict sense. For instance, B. shows that Aristotle was quite open to the use of ‘merely’ persuasive arguments (‘Enthymeme: Aristotle on the Logic of Persuasion’ [1994]). Second, and on the other hand, certain modes of attaining knowledge that we would now typically find unproblematic were rejected as defective in the ancient world. A good example here is Plato's view that knowledge cannot be passed along ‘second-hand’ (‘Socrates and the Jury’ [1980]).

This last point is typical of B. His style of writing and philosophising brings ancient texts into contact with contemporary debates. He describes himself at one point as an ‘analytic philosopher’ and few have done more to make ancient philosophy interesting to the analytically-minded reader. Yet he often emphasises how different the ancients were from us – to the point that philosophical issues and problems that now seem obvious were not even raised in antiquity. The most notorious example is idealism. In ‘Idealism and Greek Philosophy’, B. argued that ‘idealism … is one of the very few major philosophical positions which did not receive its first formulation in antiquity’, and that ‘Greek philosophy does not know the problem of proving in a general way the existence of an external world’ (vol. 1, pp. 245–6, 258). Predictably this provoked attempts at counter-example, both in the form of general studies of the (pre-?) history of idealism,Footnote ³ and in one of the few prominent philosophical debates over a Greek church father, namely Gregory of Nyssa.Footnote ⁴ It seems to me that B. was right to emphasise the absence of external world scepticism in antiquity. He also tacitly invited us to fill in a more detailed story between antiquity and Descartes, offering Augustine as the first to emphasise our infallible grasp of our own ‘subjective states’ (vol. 1, p. 269). B. is wrong to say, though, that it is only with Descartes that such states take centre stage in epistemology. Subsequent research has shown that the same is true of Avicenna.Footnote ⁵ (Medieval philosophy, incidentally, is something of a blind spot in these volumes. But this is compensated by B.'s command of a wide range of ancient authors – he will casually refer, for instance, to Philo of Alexandria or Theon of Smyrna, vol. 2, p. 82 n. 28 and p. 157 n. 33.)

Another nice example of B.'s ability to get inside ancient ways of thinking is his work on what he calls the ‘window model’ of perception, which is explored in ‘Plato on the Grammar of Perceiving’ (1976) and ‘Conflicting Appearances’ (1979). On this model, which appears implicitly in Plato and more or less explicitly in Hellenistic philosophy (see references at vol. 1, p. 289 n. 21), the mind is a unified centre of awareness that simply looks ‘through’ the sense-organs. In this case, B. not only shows the centrality of this model for Plato and other ancient authors, but also its surprising vitality. Even twentieth-century authors, as he points out, have falsely believed that people with jaundice see everything as yellow, because their eyes are yellow – like tinted windows (vol. 1, p. 288).

B. also calls the window model a ‘picture’, one of numerous echoes of Wittgenstein in these volumes. He frequently quotes Wittgenstein, and one piece – which is among the most perceptive studies ever written on Augustine's De magistro – takes its starting point from a passage in Wittgenstein. Occasionally this leads him astray, I think. For instance the compresence of opposites of which Heraclitus teaches is not, or not only, a contradiction within language (vol. 1, p. 277). This is one reminder that a number of these pieces were written in a very different time, when Wittgenstein's ghost still haunted British philosophy and cultural norms were less enlightened (in vol. 2, p. 35 an otherwise terrific piece ends with an unfortunate paragraph about Plato's possible homosexuality, written in 1977). B. was also writing before Hellenistic philosophy came to be accepted as a standard object of interest for analytically-trained historians of ancient philosophy, a change that these articles helped bring about (see his plea for this very development at vol. 1, p. 4). B. himself acknowledges the old-fashioned approach to Plato's chronological development in the one previously unpublished piece included here, which is on the Phaedrus.

Even when B. is responding to then-contemporary issues and authors, his contributions remain well worth reading. The best example is a devastating discussion of Straussianism first published in the New York Review of Books (‘Sphinx Without a Secret’ [1985]). Among other things it hilariously applies Straussian methods of interpretation to a posthumous collection of Strauss' own writings (‘the two-page “Note on Maimonides’ Treatise on the Art of Logic” is pregnant in its silence about, for instance, logic', vol. 2, p. 199).

These articles have stood the test of time in every sense. For all their seminal influence, they have not been superseded by the research they inspired – perhaps because, as B. aphoristically remarks in a different context, ‘time moves slower in philosophy than in life’ (vol. 1, p. 347). His writings remain required reading for anyone with a serious interest in ancient philosophy.