For most of the past decade I have been engaged in producing a new critical edition of Aristotle's Eudemian Ethics.Footnote 1 This ethical treatise was long regarded as the poor relation of Aristotle's better known Nicomachean Ethics, but in the last half-century has enjoyed something of a change in its fortunes, some regarding it as at least the equal of the Nicomachean. This revival of interest depends at least in part on what is alleged to be the restoration of the Eudemian's three apparently missing middle books—we have what are usually labelled as I–III, VII–VIII—from the Nicomachean. Books IV–VI of the Eudemian are in much of the tradition treated as identical to Books V–VII of the Nicomachean, and indeed as borrowed from there; statistical work done in the 1970s,Footnote 2 however, seemed to show that these books may in fact have a greater stylistic affinity to the Eudemian than to the Nicomachean. Translations have even begun including these books as part of the Eudemian,Footnote 3 and where both works are printed together, omitting them from the Nicomachean Footnote 4—thus reversing a situation that appears to have existed for most of the past two millennia.
This is a striking turnaround for a work that had apparently at some point been separated from the rest of the Aristotelian corpus and for up to a thousand years had depended, so far as we can tell, on the survival of a single copy.Footnote 5 That was the position, so far as we know, at around the beginning of the thirteenth century c.e. Byzantine Sicily and then Renaissance Italy saw a brief surge of interest in the work (as can be seen from the stemma), with that single copy giving birth, eventually, to twenty or so more, but it then again sank into obscurity, accompanied by doubts about its authenticity: the Teubner edition of 1884 was not alone in attributing it to Eudemus of Rhodes.Footnote 6 Because of the vicissitudes in its transmission, the text of the Eudemian Ethics is often in a poor state. For my new edition of 2023,Footnote 7 I made a fresh collation of the four main manuscripts (P, C, B and L in the stemma), one of which (B) was locked away in a private collection until the 1970s;Footnote 8 I have also looked at the descendants of these four, the majority in person. The result is a completely new Oxford edition of the text. It has been generally agreed for some time that existing editions, including not just the 1884 Teubner but an Oxford editionFootnote 9 as recent as 1991,Footnote 10 were inadequate, whether because founded either on an incomplete understanding of the manuscript tradition (in the case of the Teubner), or (in the case of the 1991 Oxford text, postdating HarlfingerFootnote 11) on inaccurate and incomplete collations of the manuscripts,Footnote 12 or for other reasons.
The editions in question also underestimate the singular contribution made to the restoration of the text of the Eudemian Ethics by Petrus Victorius (1499–1585), the subject of the present paper (that is, Pier Vettori, hereafter simply ‘Victorius’), one of the finest Hellenists of the late Italian Renaissance: according to Wilamowitz, the last ‘great figure [from that period] in the domain of Greek studies […] primarily for his conscientious editing of the manuscripts put at his disposal by the Laurentian Library’.Footnote 13 Greek authors edited by Victorius included Aeschylus, Euripides and Plato; he also produced editions of Aristotle's Nicomachean Ethics, Rhetoric, Poetics and Politics—no one, said a contemporary of his in a funerary laudatio, ‘left Aristotle in a cleaner state (purgatior)’.Footnote 14 And this was his intention with the Eudemian Ethics, even though he never published an edition of the text. Instead, he wrote copious marginal annotations in one of his copies of the Aldine Aristotle,Footnote 15 the editio princeps of Aristotle's works that was printed not long after Victorius was born. Aldus Manutius and his colleagues somehow succeeded in sourcing texts of all Aristotle's works (the Rhetoric and Poetics were printed a few years after the rest): an astonishing achievement, involving as it did collaboration with scholars in different parts of Italy and outside Italy. What is more, the Aldine itself made significant improvements to those texts in the process of printing. But at the same time the typesetting introduced multiple new errors, often—in the case of the Eudemian Ethics—corrected by Victorius without comment.Footnote 16
To judge by the variations in his hand, the less neat and tidy marginal notes perhaps being naturally associated with increasing years, he kept coming back to the text of this work over much of his life. He states his ambition late on in his voluminous Variae lectiones:Footnote 17
It is to be regretted that the books about ethics (mores) that Aristotle sent to Eudemus of Cyprus have come down to us in such a maimed and corrupt state. I have devoted myself at certain times (quondam) with the aid of ancient books (ope antiquorum librorum) to remove the stains that have attached themselves to their remains (reliquiis ipsorum). But to tell the truth, I have not been of as much use to them as I was hoping, not just wishing. Yet neither do I regret that labour of mine; it was labour well applied, and to clean up some part, even if not a great part, of so fine a work is no light achievement.Footnote 18
Victorius is not referring here to the mundane correction of simple typographical errors, but to his attempts to ‘clean up’ what is agreed by everyone to be—at least in many parts—a highly corrupt text, ‘with the aid of ancient books’.
There can be little doubt that at least some of the ‘ancient books’ in question are manuscripts that Victorius found in the Medici library, in so far as we know from some of his Latin editions that he had full access to the Medici collection long before the formal opening of what we call the Biblioteca Medicea Laurenziana in 1571. But editors have been slow to accept this. Susemihl divides the alternative readings noted in the margins in Victorius's AldineFootnote 19 into (a) those readings either marked ‘l.’ or ‘m.l.’ or ‘l.m.’ or ‘m.’, together with ‘paene omnia, quibus nullam eiusmodi notam praemisit’, all of which he claimed to be derived from an otherwise unknown manuscript that he labelled codex Victori; (b) conjectures by viri illius temporis docti, marked γρ (written typically as a compendium);Footnote 20 and (c) Victorius's own conjectures, a much smaller number, marked with a ‘fort[asse]’. Confirming Susemihl's own description of the supposed codex Victori, Harlfinger narrowed the codex Victori down to either Laurentianus plut. 81.20 or Laurentianus plut. 81.4, both of which appear to switch from one side of the tradition to the other (that is, from the recensio ‘Messanensis’ Footnote 21 to the recensio Contantinopolitana: see stemma) in a way that is consistent with Susemihl's description of Victorius's annotations;Footnote 22 ‘[d]ie restlichen Randbemerkungen’, Harlfinger says, ‘[…] sind fast ausschliesslich eigene Konjekturen’. Walzer and Mingay, in the 1991 Oxford text, followed Harlfinger in eliminating category (b) and Susemihl's anonymous ‘learned men of the [that is, Victorius's] time’, but apparently rowed back from Harlfinger's proposal that ‘almost all’ the remaining marginal readings were Victorius's by marking items in both category (b) and category (c) as ‘fort. V’.Footnote 23
Several questions arise. (1) Why the insistence on a single codex Victori? If Victorius had access to all the Laurentian codices, there appears no reason why we should in principle rule out any of the four such codices, that is, the four Laurentiani containing the Eudemian Ethics, as Rezenzionsexemplare. (2) What is indicated by ‘l.’, ‘m.l.’, ‘l.m.’ and ‘m.’? (3) What does γρ indicate? I address each of these in turn.
QUESTION 1: WHY ONE CODEX VICTORI?
The restriction to a single codex Victori stems in the first instance from Susemihl, who evidently did not know of Victorius's reference to antiqui libri, plural, and might therefore not have seen any good reason to postulate more than a single missing manuscript. But if Victorius's marginal corrections appear to shift their allegiance, as it were, from one recensio to the other, and Victorius did indeed have access to all the relevant manuscripts, that is consistent with Victorius's using both 81.20 and 81.4 as much as with his using just one of them. It also does not rule out his having at other times used 81.15 itself (L, the hyparchetype for the Constantinopolitani). This is likely to be less visible in so far as the Aldine, which Victorius was emending and annotating, is itself a descendant of L, but in the large number of cases where he notes what are in fact L's readings, for example in the more corrupt EE VII/IV,Footnote 24 the Rezensionsexemplar could as well have been L itself as either 81.20Footnote 25 or 81.4—or indeed 81.12.Footnote 26 Laur. 81.12 is excluded by Harlfinger only because he is looking for a single manuscript, and there is no positive indication in Victorius's annotations of its individual profile (the first four Bekker pages or so from the recensio ‘Messanensis’, the remaining thirty-five from the other recensio). Maybe Victorius missed it: its cover, after all,Footnote 27 announces it as Ἀριστοτέλους ἠθικὰ μεγάλα/Aristotelis: magna moralia, without reference to its second, longer part, the Eudemian Ethics. (Not so the later inscription on the end of the ‘pluteus’, the extended writing desk.)Footnote 28 I wonder, too, how long a philologist of Victorius's calibre would have dallied with Laur. 81.4, which is full of the most elementary mistakes, for example even mistaking instances of psi for phi.Footnote 29 So: just 81.20 and 81.15 (L)? But we cannot definitively rule out any of the four Laurentiani.
Were we still to insist on one of the four (and 81.20 has the most going for it), then a second ‘ancient book’Footnote 30 from which Victorius might have drawn is the Liber de bona fortuna, a translation by William of Moerbeke (1215–86) of Eudemian Ethics VIII/V.2, preceded by his translation of the corresponding section of the Magna Moralia. (Magna Moralia is a work now generally accepted as spurious,Footnote 31 but in the Renaissance and in Byzantium it was regarded as genuine, and indeed preferred to the Eudemian Ethics, being usually copied before it where both works were copied together.)Footnote 32 Since the nineteenth century we have known that this translation was of a Greek text somewhat less corrupt than any of those still extant, and it has been of considerable use in the restoration of this one chapter, chapter 2, of a book (VIII/V) of the Eudemian Ethics that is probably the most corrupted of all in our manuscripts. The Liber was widely distributed: even now we have between five and six times as many mediaeval copies of it than we do of the Eudemian Ethics itself, so that it would not be at all surprising if it had been known to Victorius. What is more, Victorius himself tells us, in the preface to his edition of Aristotle's Rhetoric, that he had used William's translation there, though he did not know the identity of the translator:
But I have also [in working on the Rhetoric] used […] an old translation […]. From reading it I was able to understand what the author of the translation read in his Greek manuscript. The translator was clearly a rude individual devoid of all cultural refinement [!]; yet he did his work conscientiously. So much so that, when I held that barbarous translation in my hands, it seemed to me to be the same as having the Greek manuscript […]; the translator in fact never even changes the word order, he translates literally and frequently even uses Greek words when he does not understand their meaning or does not see how to render them with a single word. But since the translation was put together many centuries ago, Aristotle's books then were less incomplete and in better shape (emendatiores). But I have found that translation for the most part in agreement with [another manuscript spoken of earlier, which Victorius has found superior].Footnote 33
(Victorius also similarly used William of Moerbeke's translation when editing Aristotle's Nicomachean Ethics and Politics.) He is unnecessarily impolite in the passage just quoted about the translator, William, whose aim was precisely to reproduce the Greek in Latin, however ‘barbarous’ the result. But the fact that the Latin reproduces the Greek so well does make it extraordinarily useful for the textual critic attempting to restore a Greek original, as in the case of Eudemian Ethics VIII/V.2; like others, I have made extensive use of the Liber de bona fortuna for that chapter in my new edition. There is, however, insufficient evidence that Victorius used it. HarlfingerFootnote 34 suggests that two outstanding corrections by VictoriusFootnote 35 derive from the Liber, but if he really had this Latin source in front of him we would expect to see rather more evidence of it; those two corrections stand out in a veritable sea of corruption, a much greater proportion of which is addressable—as the last two centuries have shown—with the help of this ‘barbarous’ translation.Footnote 36 I conclude that Victorius's ‘antiqui libri’ were restricted to the Laurentian manuscripts: possibly just one of them, rather more likely (I think) more than one.Footnote 37 So the answer to my Question 1 (‘why one codex Victori?’) is that there is more than enough reason to suppose that there was more than one.
QUESTION 2: WHAT IS INDICATED BY VICTORIUS'S ABBREVIATIONS ‘L.’, ‘M.L.’, ‘L.M.’ AND ‘M.’?
‘On the assumption that they represent Latin expressions of some kind one might hazard that l. = legendum, m.l. = melius legere (or melior lectio), l.m. = lectio melior, and m. perhaps just “melius”’, wrote Christopher Strachan (in correspondence) in response to an appeal to explain abbreviations that neither Susemihl, whose edition contained them, nor anyone else explained, although (I added) they all seemed to be a way of saying ‘fortasse’. I am now certain that Strachan was right.Footnote 38 An alternative, which I now regret preferring in a footnote in my new edition of the Eudemian Ethics, is to take them as ‘l(aurentianus, sc. codex)’, ‘m(anuscriptus) l(aurentianus)’, etc., that is, as referring on each occasion just to a Laurentian manuscript without picking it out specifically.Footnote 39 The chief reason for rejecting this alternative is that we do not know when exactly the Biblioteca Medicea Laurenziana was first called ‘Laurentian’ as opposed to just ‘Medicea’ (certainly rather later).Footnote 40 In any case, ‘to be read’, etc., fits Victorius's actual use of the abbreviations much better, particularly in so far as he notes so many alternative readings undoubtedly from one or another Laurentian that are not so marked, if they are marked at all. The difference between those marked ‘l.’, etc. and those not so marked is that Victorius is expressing a preference for the marked ones, merely recording the others.
QUESTION 3: WHAT DOES VICTORIUS'S ΓΡ INDICATE?
The same applies equally, I propose, to the mark γρ; that is, it too marks a preference for the reading in question, and stands for γραπτέον,Footnote 41 not γράφεται. The abbreviation is of course standardly employed for γράφεται. But the problem with so interpreting it in the present case is that rather few of the marginal readings with this marking in Victorius's Aldine text of the Eudemian Ethics are in fact written/found in the extant manuscripts. Susemihl attributed his category (b) marginal readings (γρ readings) to unknown learned men just because he assumed γρ = γράφεται. But this must be wrong.
Take for example the two signal emendations in Eudemian Ethics VIII/V.2 picked out by Harlfinger as evidence that Victorius was using the Liber de bona fortuna: 1247a18 πλέων Victorius, πλέον PCBLFootnote 42, navigans Liber; a20 ἐν οἷς Victorius, ἐνίοις PCBL, in quibus Liber. Both Victorius's readings are preceded by γρ and are not found anywhere in the tradition (i.e. except in the presumed Greek source of the Liber de bona fortuna, which I have argued Victorius was not using). Given both that no one denies that Victorius is capable of making independent conjectures, and that πλέων and ἐν οἷς are conjectures by somebody, their absence from our manuscripts makes it more economical to attribute the readings to Victorius than to postulate otherwise unknown scholars as their inventors.Footnote 43
If the evidential value of two examples in question is slightly undermined by the very existence of the Liber, there are many other examples of γρ in the margins of the Eudemian Ethics in Victorius's Aldine marking conjectures: 1215a1, 10, 1216b31, 1218a15, 1218b19, and so on—enough to mean either that the viri docti were fairly busy or that there were quite a number of them, and to make it surprising that none of them has left any other trace. In a significant number of cases, as with 1247a20 (for which see above), Victorius—if it is he, as I suppose—is as a matter of fact anticipated, particularly by an anonymous hand in P, but again he would have had no way of knowing that. In short, there is overwhelming evidence that γρ indicates γραπτέον, and where a given reading does not appear anywhere in the tradition, it will stem from Victorius himself. γρ is in fact no different from l. = legendum and the other Latin variants, and functions as an alternative to these, for as well as commending readings that are not conjectures, they too can introduce conjectures:Footnote 44 thus 1216b2 λόγου (also Ravennensis, Biblioteca Classense 210) is marked ‘l.’, as is 1217b32 κοινῶς (not actually an improvement); 1235a18 ἔχει (a proposed, wrong, insertion) is marked ‘m.l.’;Footnote 45 … But sometimes, too, a conjecture appears in the margin with no mark at all: so, for example, at 1241a38 (συμβαίνειν, also found, as it happens, in B, for συμβαίνει).Footnote 46 We should, then, read γρ in Victorius's marginal annotations to the Eudemian Ethics as standing for γραπτέον (not for γράφεται).
***
My overall conclusion, then, is that γρ, l., m.l., l.m. and m. all indicate the same thing: ‘this is to be read’. That is the only outcome that is consistent with the evidence. Another more general argument supports this conclusion: we know, because Victorius tells us (see above), that his aim was to ‘clean up’ the text of the Eudemian Ethics. If γρ is γράφεται, and l., m.l., etc. were (impossibly) to indicate the manuscript source of a reading, then there would be nothing in Victorius's annotations to indicate his preferences at all; so how would he be improving anything, even if it was all (so far) for his own benefit? There are plenty of cases where he records, without any mark, a reading that is plainly wrong, for example, one that we find in the L tradition when the Aldine diverges from and improves on that. If we ask whether its wrongness was necessarily plain to Victorius, he so often corrects mistakes in the Aldine itself, sometimes in the body of the text, that it is beyond doubt that his Greek is very good. Where he needs to record a preference, he does so, using one or other of the marks available to him. The variations in the use of those perhaps merely reflect the fact that he returned to the text many times over, and used whichever mark occurred to him.Footnote 47 Some instances of γρ, I hazard, belong to his later attempts to improve the Aldine, because they are about as far removed as they could be from the neatness of most of the annotations; were his faculties by this point beginning to fail him?Footnote 48
Victorius's creativity and inventiveness as a textual critic—his ability to see just what is needed to heal a corrupt sentence, here and now—have been consistently underestimated, from Susemihl onwards. Susemihl had a partial reason, that is, his reading of Victorius's γρ. But others underestimate the degree of Victorius's interventions for no good reason: Walzer and Mingay in the 1991 OCT constitute a good example. An anonymous reader of my new edition for Oxford University Press roundly claimed that Victorius ‘mostly emends ope codicum and emendationes ope ingenii are very rare’, citing a comment by Martinelli Tempesta relating to Victorius's work on Plato's Lysis.Footnote 49 But there is an important difference between the two cases. The text of Plato, the Lysis included, has come down to us in astonishingly good condition. One of Victorius's own principles, in which he followed his predecessor and spiritual mentor Politian, was to emend only in extreme cases: ‘I think it important (magnum) to safeguard the integrity of ancient authors, and as far as possible to hand them down to posterity pure and uncorrupted’.Footnote 50 But the Eudemian Ethics is already thoroughly corrupted: its text is ‘in a vile state—hideous corruption on every page’, wrote Barnes in 1992.Footnote 51 Given his conservatism as a textual critic,Footnote 52 this may be why Victorius never produced an edition of the Eudemian Ethics, as he did of other Aristotelian works. But thirty or so of his emendations are in my apparatus, and there is roughly the same number of other mentions of him; not infrequently, his interventions are anticipated by correcting hands in manuscripts unavailable to him. In the case of the Eudemian Ethics, his use of his own ingenium is far from rare. That he never published an edition of it is no doubt just because of his awareness that he had not made as much progress with it as he says he ‘not just hoped but wished’, that is, intended.Footnote 53 His judgement is more than equal to that of his nineteenth-century German and British successors, about some of whose rewriting of what the surviving manuscripts do give us he would have been scathing. Here, with the Eudemian Ethics, at the end of the Renaissance, we find another textual critic—I think here especially of Politian—who offered a model of textual criticism that has rarely even now been surpassed.
I suspect, however, that he would not have approved of mine. My question, in dire circumstances, is always: how could Aristotle have got from a to f, say, where f is accompanied by a ‘therefore’, given that b to e, the steps in between the first premise and the conclusion (I here simplify), are muddied and corrupted in the text as handed down to us? This is the kind of leap that Victorius refuses to make, restricting himself almost entirely to replacements of single words or phrases—and unable, because of that, to produce a fully readable text.Footnote 54 This is not the same as the procedure followed by most German and British critics working on the Eudemian Ethics from the nineteenth to the present centuries, which is mainly a matter of—as they see it—restoring Aristotle's Greek, typically without reference to the argument. (That is usually, even now, understood to be a strict division between textual experts and philosophers.) The assumption is that Aristotle would have written proper Greek: after all, as everyone knows, no less an authority than Cicero admired him for his style. But not only is this hardly justifiable from the evidence of his treatises; it also happens that much of the Eudemian Ethics, unlike the Nicomachean, is written in an extraordinarily cramped, staccato style.Footnote 55
Three choices present themselves: (1) we may concentrate on improving Aristotle's Greek (while getting rid of as much obvious corruption as we can); or (2) we may be more attentive to the fact that Aristotle constructs arguments (and usually decent ones); or else (3) like Victorius, we may content ourselves just with getting rid of the obvious textual errors, which sometimes involves real insight and understanding of the microscopic (but not the macroscopic) context. This last option, however, will leave us with a text that is in quite a few places literally unreadable, and not one that an editor and textual critic of the purity of a Victorius could bring himself to bring out in print.
Open access
