In the preface to the first edition (1978) K. predicted that ‘when the statistical study of literary texts has grown into a systematically organized discipline, monographs such as the present one will no doubt look very amateurish’. The reissue of this landmark book proves K.’s prediction wrong. Though advances in computer technology have made some of the methods employed in the original study seem quaint, K.’s pioneering methods, painstaking labour and boldness of vision remain a model to this day.

K.’s aims were twofold: (i) to show by stylometric as well as historical and philosophical argument that the common books (K.’s ‘AE’), traditionally printed as Books 5, 6 and 7 of the Nicomachean Ethics (‘NE’) originally belonged to the Eudemian Ethics (‘EE’) and (ii) to argue that the EE is likely a late work, ‘the mature and definitive statement of Aristotle's ethical system’ (p. v), worthy of the same close attention that scholars have devoted to the NE.

By correcting most of the misprints and typographical errors that marred the first version, this reissue sees OUP atoning for past editorial sins. In two chapters, ‘Reconsiderations 1992’ and ‘Reconsiderations 2016’, K. reflects on the reception of the first edition, responding to critics, modifying – and occasionally recanting – his past views. These chapters serve as primers for readers new to the debates that the work spurred and contain a wealth of interesting observations that supplement the original text. K. is incisive, shrewd and wry. Even when his critics appear to have the upper hand, K. bites back forcefully. Chapters 1–9 cover external and internal evidence, style, lexis and philosophical evidence. The final original chapter, ‘Dating of the Aristotelian Ethical Treatises’, demands a longer response than it has so far received. Its treatment of the EE as the template on which the Magna Moralia (‘MM’) is based is ripe for revision.

K.’s first critics objected that we should separate the question of the origin of the common books from the question about their intended home. Even if Aristotle composed the AE as part of the EE, as most scholars now hold, their current form may reflect Aristotle's own efforts to adapt the argument to the analysis in the NE. And so we cannot assume that ‘our’ AE is identical to the original version. This possibility does not receive the attention it deserves in K.’s postscripts. Nor does K.’s stylometric analysis tell either way on this question: we can acknowledge the Eudemian flavour of the vocabulary and style of the common books while still consistently maintaining that the argument they advance shows signs of being revised to fit the specifically Nicomachean account of virtue. The account of full virtue defended in AE B 12–13 seems to contradict the account of kalokagathia and individual virtues of character defended in EE 8.3. Forward references in the special books of the NE match the content of the AE just as well as those in the EE’s special books, even if they are not as specific or numerous.

K.’s omission of backward references is remedied in the 2016 postscript. K. thinks the evidence strengthens rather than weakens his case. However, the one EE passage that K. claims ‘unambiguously’ refers to AE (specifically B 1, 1138b18–34) seems to duplicate rather than refer back to it. In discussing the standard (horos) of possession, choice and avoidance, EE 8.3, 1249b3–6 states: ‘Earlier we said “in accordance with reasoning (to hôs ho logos)”, but this, like saying in matters of diet “in accordance with medicine and medical reasoning”, is true but unenlightening (alêthes men, ou saphes de)’ (p. 277). EE 8.3 now adds the analogy to medicine and medical reasoning. AE B 1 has already given a much fuller and longer version of the analogy to medicine, after observing that ‘saying that the intermediate is in accordance with right reasoning (kata ton orthon logon) is true, but unenlightening’ (alêthes men, outhen de saphes, 1138b25–6, responding to promissory notes at NE 2.2, 1103b31–4 and/or EE 2.5, 1222b7–8). Furthermore, the argument in AE B is Aristotle's attempt to define orthos logos: ‘let us now determine what it says’ (AE B 1, 1138b21, trans. T. Irwin [1999]). K. believes EE 8.3 answers a second question about the horos of right reasoning left hanging by B 1, ‘what is the standard or criterion that defines it’ (pp. 180–2). However, the ‘kai’ at B 1 1138b34 could be epexegetical. If so, there is no second question: alla kai diôrismenoon tis estin ho orthos logos kai toutou tis horos means ‘but we must also determine what right reasoning is, that is to say, what the definition of right reasoning is’ (AE B 1, 1138b33–34). If so, Aristotle's use of the medical analogy in 7.3 would show a remarkable degree of amnesia and lack of authorial control. It would be more likely that EE 8.3 antedates AE B 1 and that Aristotle decided to expand on the medical analogy when he wrote our version of AE B 1. We need a stronger argument than K. offers to show that Aristotle did not intend the AE to form part of the NE.

In the 2016 postscript K. admits that he has never made a systematic study of the Magna Moralia, and yet his argument about dating rests heavily on the assumption that the MM postdates the other treatises. The ‘philosophical crudity’ of many passages in the MM ‘amount[s] to a misunderstanding, rather than a simplification, of the arguments of the other ethical treatises’ (p. 272). Yet he never considers reasons for thinking that the MM is an early treatise by Aristotle in the same sense that the EE and NE, though lightly edited by later scholars, are by Aristotle. K.’s appeal to names and historical allusions in the MM is unconvincing; he does not give due weight to equally plausible explanations of the evidence. To claim that the MM ‘follows the EE slavishly, in the order of topics and the manner of their treatment, for almost the whole of its length’ (p. 220) is to confuse the MM’s greater affinity to the EE than to the NE with an attempt to render an account of the EE – which, contrary to K.’s claim, neither follows the EE slavishly nor for the whole of its length, especially if we accept that the AE belongs in the EE.