1) 1134a24–31, 1134b16–19, 1135a3–5

At 1134a24–31 Eth. Nic. has δεῖ δὲ μὴ λανθάνειν ὅτι τὸ ζητούμενόν ἐστιν καὶ τὸ ἁπλῶς δίκαιον καὶ τὸ πολιτικὸν δίκαιον. τοῦτο δ’ ἔστιν ἐπὶ κοινωνῶν βίου πρὸς τὸ εἶναι αὐτάρκειαν, ἐλευθέρων καὶ ἴσων ἢ κατ’ ἀναλογίαν ἢ κατ’ ἀριθμόν· ὥστε ὅσοις μή ἐστι τοῦτο, οὐκ ἔστι τούτοις πρὸς ἀλλήλους τὸ πολιτικὸν δίκαιον, ἀλλά τι δίκαιον καὶ καθ’ ὁμοιότητα. ἔστι γὰρ δίκαιον, οἷς καὶ νόμος πρὸς αὑτούς· νόμος δ’, ἐν οἷς ἀδικία· ‘one must not forget that the thing being sought for is both the simply just and the political just. The latter exists in the case of those who share together a way of life aimed at self-sufficiency, who are free and equal, whether in proportion or number. So all those who do not have this do not have the political just with respect to each other but the just in a certain respect and by way of likeness. For those have justice who also have law in relation to each other, and those have law among whom there is injustice …’. The Eth. Eud. MSS omit the lines italicized above and replace ‘both’ with ‘not’: δεῖ δὲ μὴ λανθάνειν ὅτι τὸ ζητούμενόν ἐστιν οὐ τὸ ἁπλῶς δίκαιον καὶ τὸ πολιτικὸν δίκαιον, ἀλλά τι δίκαιον καὶ καθ’ ὁμοιότητα. ἔστι γὰρ … ‘one must not forget that the thing being sought for is not the simply just and the political just but the just in a certain respect and by way of likeness. For those …’.

How to interpret the Eth. Eud. version, or how to understand what it says about the relation of law and justice, is initially puzzling, and one may therefore be inclined to attribute its readings to a scribe's error of omission, prompted perhaps by the repetition of δίκαιον immediately before and right at the end of the omitted lines. But why then the other difference, the insertion of ‘not’ in place of ‘both’, which can hardly be explained by an error of omission? We should at least consider if the Eth. Eud. readings have a sense in the context. To understand them we must suppose the simply just or the political to be general or legal justice, which was discussed in the first chapter of Book 4/5 and identified with the whole of virtue.Footnote ¹¹ The just in a certain respect we must suppose to be particular justice, namely distributive and corrective justice, which has been the topic from the second chapter on. Justice by way of likeness will then perhaps be the same thing (for particular justice is like general or legal justice in that it exists in relation to another), or perhaps it is the justice found in the household, which is discussed shortly. At all events the omission in Eth. Eud. of several lines from Eth. Nic. clearly has the effect of making Eth. Eud. less focussed on politics and the political community than Eth. Nic., and so less focussed on the general political and legislative interest that marks Eth. Nic. in contrast with Eth. Eud.Footnote ¹²

This result is confirmed by two related passages in the same overall context. The first is at 1134b16–19, where Eth. Nic. has τοῦτο γάρ ἐστι τὸ οἰκονομικὸν δίκαιον· ἕτερον δὲ καὶ τοῦτο τοῦ πολιτικοῦ. τοῦ δὲ πολιτικοῦ δικαίου τὸ μὲν φυσικόν ἐστι τὸ δὲ νομικόν, φυσικὸν μὲν τὸ πανταχοῦ τὴν αὐτὴν ἔχον δύναμιν … ‘for the just in the household is this [sc. the justice in relation to a wife], and it is different from the political just. Of the political just there is the natural and the legal: natural being what has everywhere the same force …’. The Eth. Eud. MSS by contrast (and in agreement with K^b) add a γὰρ and omit some words: τοῦτο γάρ ἐστι τὸ οἰκονομικὸν δίκαιον· ἕτερον δὲ καὶ τοῦτο τοῦ πολιτικοῦ δικαίου· τὸ μὲν γὰρ φυσικόν ἐστι τὸ δὲ νόμιμον,Footnote ¹³ φυσικὸν μὲν τὸ πανταχοῦ τὴν αὐτὴν ἔχον δύναμιν … ‘for the just in the household is this, and it is different from the political just. For the one is natural and the other legal: natural being what has everywhere the same force …’. Again one may attribute the omission to a scribe's error in jumping over the τοῦ δὲ πολιτικοῦ, because he had just written πολιτικοῦ. But then the same scribe would have had to make a second, and wilful, error by inserting, on his own account, the γὰρ in τὸ μὲν γὰρ φυσικόν.

The second passage is at 1135a3–5, where Eth. Nic. has ὁμοίως δὲ καὶ τὰ μὴ φυσικὰ ἀλλ’ ἀνθρώπινα δίκαια οὐ ταὐτὰ πανταχοῦ, ἐπεὶ οὐδ’ αἱ πολιτεῖαι, ἀλλὰ μία μόνον πανταχοῦ κατὰ φύσιν ἡ ἀρίστη, ‘and likewise the non-natural but human just things are not the same everywhere, since not even the regimes are though one alone is everywhere according to nature best.’ The Eth. Eud. MSS have: ὁμοίως δὲ καὶ τὰ μὴ φυσικὰ ἀλλ’ ἀνθρώπινα δίκαια οὐ ταὐτὰ πανταχοῦ, ἐπεὶ οὔδε πολιτεῖαι, ἀλλὰ μία μόνον πανταχοῦ κατὰ φύσιν ἡ ἀρετή, ‘and likewise the non-natural but human just things are not everywhere the same, since not even regimes are, but everywhere according to nature virtue is only one.’ It is easy to ascribe the differences to scribal error. The αἱ in οὐδ’ αἱ is in fact missing in some Eth. Nic. MSS (but not K^b), and an ἀρίστη could easily be misread as ἀρετή. But, of course, the misreading could also go the other way, and besides we do not have to suppose error here. For this passage, like the previous two passages, displays the same feature, namely a de-emphasizing of the political and legislative interest.

The fact is plain in the second passage in the Eth. Eud. version (1134b16–19): ‘for the just in the household is this, and it is different from the political just. For the one is natural and the other legal: natural being what has everywhere the same force …’. These words seem to mean that the natural just is the household just and the legal just the political just. In the Eth. Nic. version, by contrast, the meaning is that the political just is both natural and legal. If we follow the Eth. Eud. version, the point seems to be that marriage and domestic community are natural, while the political community is legal. This idea gets support from the Politica, when Aristotle says that the household of man and wife arises from natural desire and does not need any conscious or deliberate arranging by a legislator, as the political community does (Pol. 1.2.1252a28–30, where the word φυσικόν or ‘natural’ appears). The domestic just does also seem to be the same way ‘here and in Persia’ (1134b26), for marriage naturally arises everywhere. But neither in Eth. Eud. nor in Eth. Nic. is the political or legal just said to be contrary to nature. Both say rather that it is variable but on a natural basis. Still the Eth. Eud. version, despite saying the same in this respect as the Eth. Nic. version, has a different focus and effect, namely the de-emphasizing of the political or legislative interest.

The third passage (1135a3–5) confirms the same fact. The meaning of Eth. Nic. seems clear enough in the context. There are many regimes or constitutions, both good and bad, but only one is by nature everywhere the best (presumably the total kingship of the divinely virtuous man). The meaning in Eth. Eud. also seems clear. Supreme virtue is only one—not in the sense that there is only one kind of virtue (for courage and generosity are two kinds and not one), but that complete virtue is one totality, namely the totality praised at the beginning of this book as ‘fairer than morning and evening star’ (1129b28–9). But why the shift from constitutions to virtue in Eth. Eud. and not in Eth. Nic.?

A suggestion: Aristotle has here finished his excursus on political justice, both in Eth. Eud. and in Eth. Nic., and turns next to distinguish how one does a just or an unjust thing in order to be oneself just or unjust. In Eth. Nic. he ends this excursus with the best regime and in Eth. Eud. he ends it with supreme virtue. Now Eth. Nic. in its last book returns to the question of regimes and ends with a discussion of legislation and of becoming a legislator that leads into the Politica. Eth. Eud. does not return to the question of regimes and ends instead with a discussion of the fully virtuous man, the kaloskagathos, the gentleman, who devotes himself to the worship of the god (8.3). Eth. Nic. talks also indeed of the superiority of the theoretical over the political life and of devotion to god rather than to the city (10.8). But Eth. Nic., unlike Eth. Eud., then descends back into politics (10.9). Eth. Eud. does not. It stays with the god. Its de-emphasizing of politics and the legislative interest in the three passages just discussed fits well with the overall character and focus of Eth. Eud. We need not, then, dismiss the verbal differences between the Eth. Nic. and the Eth. Eud. manuscripts as scribal error in one or other (or both).Footnote ¹⁴ We can instead regard them as reflecting, and confirming, what we already know to be the general difference in character between the two treatises.

2) 1136a15–18, 1136b1–4, 15–17, 26–9

At 1136a15–18 Eth. Nic. has πότερον γὰρ ὡς ἀληθῶς ἔστιν ἑκόντα ἀδικεῖσθαι, ἢ οὒ ἀλλ’ ἀκούσιον ἅπαν, ὥσπερ καὶ τὸ ἀδικεῖν πᾶν ἑκούσιον; καὶ ἆρα πᾶν οὕτως ἢ ἐκείνως, ὥσπερ καὶ τὸ ἀδικεῖν πᾶν ἑκούσιον,Footnote ¹⁵ ἢ τὸ μὲν ἑκούσιον τὸ δ’ ἀκούσιον; ‘for is it truly possible for someone to be wronged voluntarily, or is it not but all of it is involuntary as also wrongdoing is all voluntary? And is it all this way or all that way, as also wrongdoing is all voluntary, or is some voluntary and some involuntary?’ The Eth. Eud. MSS have: πότερον γὰρ ὡς ἀληθῶς ἔστιν ἑκόντα ἀδικεῖσθαι, ἢ οὒ ἀλλ’ ἀκούσιον ἅπαν, ὥσπερ καὶ τὸ ἀδικεῖν πᾶν ἀκούσιον;Footnote ¹⁶ καὶ ἆρα πᾶν οὕτως ἢ ἐκείνως, ὥσπερ καὶ τὸ ἀδικεῖν, ἢ πᾶν ἑκούσιον, ἢ τὸ μὲν ἑκούσιον τὸ δ’ ἀκούσιον, ‘for is it truly possible for someone to be wronged voluntarily, or is it not, but all of it is involuntary as also all wrongdoing is involuntary? And is it all this way or all that way as also wrongdoing is, or all voluntary, or is the one voluntary and the other involuntary?’ The second change, the addition of ‘or’, is found in K^b, but the first change, ‘involuntary’ for ‘voluntary’, is not.

The changes are slight but they produce an interesting difference of division in the two lists:

The Eth. Nic. version may be formalized as follows:

1. 1. Is being wronged all involuntary, as wronging is all voluntary?

2. 2. Is being wronged either all voluntary or all involuntary, as wronging is all voluntary?

3. 3. Is some being wronged voluntary and some involuntary?

The Eth. Eud. version, by contrast, may be formalized, at least initially, as follows:

1. 1. Is being wronged all involuntary, as wronging is all involuntary?

2. 2. Is being wronged either all involuntary or all voluntary, as wronging is?

3. 3. Is being wronged all voluntary?

4. 4. Is some being wronged voluntary and some involuntary?

This Eth. Eud. division is repetitious, since option 3 is already included under option 2.Footnote ¹⁷ Is there another and non-repetitious way of doing the division? Here is a suggestion:

1. 1. Is all being wronged involuntary as all wronging is involuntary?

2. 2. Are both being wronged and wronging voluntary or involuntary (sc. so that whatever the one is the other is also)?

3. 3. Are both voluntary?

4. 4. Are both opposite (sc. so that if one is voluntary the other is involuntary)?

The difference in this interpretation is that the division does not concern being wronged by itself but being wronged and wronging together. The main justification for the interpretation is the way in which the first option is stated (being wronged is involuntary as wronging is involuntary), for it associates being wronged with wronging in the same class of the involuntary. Option 2 also associates being wronged with wronging in the same class of the voluntary. The association in option 1 seems bizarre, because Aristotle has just argued that wronging is always voluntary. If a wrong is done involuntarily, he has said, it is not really a case of wronging save incidentally (1135b2–8). One does a wrong thing but not as wrong, for one did not choose it and do it as wrong. The person who suffers the wrong is perhaps harmed (for he suffers a loss), but he is not strictly speaking wronged. Why then does the Eth. Eud. listing of options speak of wronging as if it could be involuntary? The answer would seem to be precisely that the Eth. Eud. version is not considering the voluntariness or involuntariness of being wronged by itself but rather the relation of its voluntariness or involuntariness to the voluntariness or involuntariness (broadly understood) of wronging. If so, we should, as a result, consider all the options listed in the Eth. Eud. version in the same way.

What are the theoretical options if wronging and being wronged are taken together? Well, there are three about being wronged in the Eth. Nic. version: it is all involuntary; it is all voluntary; it is sometimes one and sometimes the other. So there will, by parity of analysis, be three in the case of wronging: it is all involuntary; it is all voluntary; it is sometimes one and sometimes the other. Combining the two, therefore, will again give three options: being wronged and wronging are both involuntary (option 1); being wronged and wronging are both voluntary (option 3); being wronged and wronging are sometimes one and sometimes the other, and either as being the same way (option 2) or as being the opposite way (option 4). Of course, if they vary, they need not always be the one or the other (always both the same or always both opposite); they might sometimes be the same and sometimes opposite. But we may assume that this possibility is meant to be included: in any particular case of wronging and being wronged, both will either be the same or opposite, regardless of whether they are always the same or always opposite or sometimes the same and sometimes opposite.

The Eth. Eud. division, then, read in this way, produces a plausible logical sense. Accordingly, we are not compelled to regard the textual variants in the Eth. Eud. version as scribal errors. They may indeed be such errors; however, since they need not be, we are permitted, if only for the sake of argument, to pursue the question of what the point or purpose of the Eth. Eud. version is. Now this purpose would, it seems, have to be a strictly philosophical one, for the question of whether being wronged is voluntary or not can be addressed, as the Eth. Nic. version addresses it, by considering only the voluntariness of being wronged. If the Eth. Eud. version says more, the reason must be that the Eth. Eud. version is aiming to satisfy a purely philosophical interest and not to answer something required by the strict needs of the question. Eth. Eud. is, as it were, going off at something of a philosophical tangent. The tangent is, to be sure, interesting in its own right, but rather as a theoretical puzzle than as a practical one, for it seems to add nothing of legislative relevance. So it would interest a philosophical audience but not, or not necessarily, an audience of legislators. Philosophers are precisely theoretically and not just practically inclined, while legislators, the presumed audience of Eth. Nic., would not have this theoretical interest, or not qua legislators.

That the Eth. Eud. version adds nothing of legislative relevance is clear also from the fact that the discussion in the following lines (which are the same in both the Eth. Eud. and the Eth. Nic. versions) answers the Eth. Eud. and Eth. Nic. list of options in one and the same way and without express discussion of the excluded members. The answer given is that being wronged is always involuntary, which is the first option in both lists. Certainly this answer resolves all the questions in the Eth. Nic. list, for that list is about whether being wronged is voluntary or involuntary and whether as always one of them or as sometimes one and sometimes the other. It also resolves all the questions in the Eth. Eud. list, at least to the same extent, namely of showing that being wronged is always involuntary. But it says nothing, or nothing expressly, about the relation of the involuntariness of being wronged to the involuntariness or voluntariness of wronging. This relation is puzzling. For since it is possible to do a wrong incidentally or involuntarily (in a broad sense), and so to do it without wronging or without being unjust, is it possible to suffer a wrong (involuntarily) that someone inflicted incidentally or involuntarily? The case is raised in effect in the later discussion of a judge who makes a wrong award (1136b15–37a4): the one denied the award suffers a wrong but the judge need not himself be doing wrong, if, say, he is following the existing law or acts in ignorance (1136b29–37a4). So Aristotle's response to the questions in the Eth. Eud. version seems to be to adopt option 4, or a variation of it, namely that being wronged is always involuntary but the correlative doing wrong could be either, or could at least be a doing wrong incidentally.

However, the Eth. Eud. version might answer the puzzle more directly, because it also raises it again, if obliquely, while the Eth. Nic. version does not. For at 1136b1–2 the Eth. Nic. version has ἔτι ἑκὼν ἄν τις δι’ ἀκρασίαν ὑπ’ ἄλλου βλάπτοιτο ἑκόντος, ὥστ’ εἴη ἂν ἑκόντ’ ἀδικεῖσθαι, ‘further, someone might, through incontinence, be voluntarily harmed by someone acting voluntarily, so that he would be wronged voluntarily.’ The Eth. Eud. version has ἔτι ἑκὼν ἄν τις δι’ ἀκρασίαν ὑπ’ ἄλλου βλάπτοιτο ἀδικοῦντος,Footnote ¹⁸ ὥστ’ εἴη ἂν ἑκόντ’ ἀδικεῖσθαι, ‘further, someone might, through incontinence, be voluntarily harmed by someone wronging him, so that he would be wronged voluntarily.’

Both the Eth. Nic. and the Eth. Eud. versions state something Aristotelian, and both have the same answer: voluntarily being harmed, whether the one doing the harm is wronging or acting voluntarily or not, is not voluntarily being wronged. But the Eth. Eud. version expressly talks about being harmed by someone doing wrong and not just, as in the Eth. Nic. version, by someone doing a harm voluntarily. It thus recalls the puzzles raised in the original list of options from the Eth. Eud. version (about wronging and being wronged). It thus also gives one to expect that this puzzle will be answered in what follows. And it is answered. That this answer is the same in Eth. Eud. and Eth. Nic. makes no difference. Eth. Nic. has not raised the question, so the answer will not be applied to it. Eth. Eud. has raised the question, so the answer can be applied to it. Indeed, for Eth. Nic.’s audience of legislators this extra question would have no legislative significance, since it requires no extra answer. But for a philosopher it does have significance, because it raises a further philosophical puzzle. That this puzzle requires no different answer does not deprive it of the philosophical interest. Eth. Nic., we may say, sticks to the legislative focus, while Eth. Eud. goes beyond it.

There is, however, a difference in the way in which the answer is given that deserves noting, since it seems to relate back to the different division or listing of questions in Eth. Eud. as opposed to Eth. Nic. In the immediately following lines (1136b3–4), the Eth. Nic. version has ἢ οὐκ ὀρθὸς ὁ διορισμός, ἀλλὰ προσθετέον τῷ βλάπτειν εἰδότα καὶ ὂν καὶ ᾧ καὶ ὣς τὸ παρὰ τὴν ἑκείνου βοὐλησιν; βλάπτεται μὲν οὖν τις ἑκὼν καὶ τἄδικα πάσχει, ἀδικεῖται δ’ οὐδεὶς ἑκών· οὐδεὶς γὰρ βούλεται … ‘or is the definition not right and to “doing harm when one knows the whom and the with what and the how” one must add further “against the other's will”? So one can be voluntarily harmed and voluntarily suffer wrong things, but no one is wronged voluntarily. For no one wants [to be wronged] …’. The Eth. Eud. version has ἢ οὐκ ὀρθὸς ὁ διορισμός, ἀλλὰ προσθετέον τὸ βλάπτειν εἰδότα καὶ ᾧ καὶ ὣς παρὰ τὴν ἐκείνου βούλησιν; βλάπτεται μὲν οὖν τις ἑκὼν καὶ τἄδικα πάσχει, ἀδικεῖται δ’ οὐδὲ εἷςFootnote ¹⁹ ἑκών· οὐδεὶς γὰρ βούλεται … The Greek of this Eth. Eud. version is peculiar and there looks to be scribal error,Footnote ²⁰ but if there is no scribal error what is the sense? One could translate: ‘or is the definition not right and one must add further: “doing harm knowing both the with what and that it is against the other's will”? So one can be voluntarily harmed, etc.’ But if the ‘what’ is going to be mentioned, then it would seem that, as with the Eth. Nic. version, the ‘whom’ and the ‘how’ should be mentioned too.

An alternative is to translate: ‘Or is the definition not right and one must add further: “doing harm knowing both to whom and that it is against the other's will”?Footnote ²¹ So one can be voluntarily harmed, etc. …’. Here the ᾧ is to be read as masculine and as a dative of disadvantage (normally βλάπτειν, ‘doing wrong’, takes an object in the accusative and not the dative case). This construal is admittedly obscure, but it is not impossible (it also fits Eth. Eud.’s general tendency toward obscurity of expression). The construal also introduces just the single factor of the ‘whom’ and not also the ‘what’ and the ‘how’, as the Eth. Nic. version does. But why focus on what the wrongdoer knows rather than on what the wronged wants? For while the wrongdoer must not be ignorant that what he is doing is against the will of the wronged, otherwise his action would be only a harm and not a wrong (or only a wrong incidentally), yet the will of the wronged is what is decisive for something's counting as a wrong. The likely answer is that the shift of focus is to remind or confirm for the reader that, in the list of options given earlier (in the Eth. Eud. version), both wronging and being wronged are at issue and not just being wronged (as in the Eth. Nic. version). So Eth. Eud. expressly draws attention to what the mind of the wrongdoer must be as well as, in the next line, to what the will of the wronged must be, while Eth. Nic. contents itself with drawing attention only to what the will of the wronged must be.

There are additional verbal variants that follow a few lines later. At 1136b15–17 the Eth. Nic. text has ἔτι δ’ ὧν προειλόμεθα δύ’ ἔστιν εἰπεῖν, πότερόν ποτ’ ἀδικεῖ ὁ νείμας παρὰ τὴν ἀξίαν τὸ πλεῖον ἢ ὁ ἔχων, καὶ εἰ ἔστιν αὐτὸν αὑτὸν ἀδικεῖν, ‘further, among the things we chose there are two things to speak of, whether it is the one who distributes the greater share against desert or the one who has it who does the wrong, and if it is possible to wrong oneself.’ The Eth. Eud. version has ὅτι δ’ ὧν προειλόμεθα δύ’ ἔστιν εἰπεῖν, πότερόν ποτ’ ἀδικεῖ ὁ νείμας παρὰ τὴν ἀξίαν τὸ πλέον ἔχωνFootnote ²² εἰ ἔστιν αὐτὸν αὑτὸν ἀδικεῖν; ‘because among the things we chose there are two things to speak of, does the one who distributes against desert do wrong, having the greater share, if it is possible to wrong oneself?’ One may again suspect scribal error here, and in particular over the ἔτι and ὅτι. But if so, which of these is the error (for both can make sense)? Further, the other verbal differences (namely the τὸ πλεῖον ἢ ὁ ἔχων, καὶ and the τὸ πλέον ἔχων) look less like error and more like alternative formulations. The question, then, turns rather on whether the different readings make sense in their contexts, for if so, scribal error will cease to be the only, or the only plausible, explanation.

As for the Eth. Nic. reading, an initial oddity is why one of the two things to discuss is the question about the distributor and the haver, since it was not raised before, whereas the question about whether one can wrong oneself was raised before (1136a31–b1), and solved before (1136b3–9). A first part of the answer would seem to be that the question about distributor and haver is implied by the case of Glaucon and Diomedes (1136b9–13), if one imagined the distribution to be made not by them but by some third person as judge. A second part of the answer would seem to be that the question is related to the other question about wronging oneself, for if the distributor happens also to be one of the parties and he distributes more to the other than to himself, then he would be wronging himself. Such is what the passage immediately goes on to discuss when, in accordance with Eth. Nic.’s way of stating the first of the questions, it says (1136b17–18): ‘For if the first thing said is possible and the distributor of the greater share, but not the one who has it, does wrong …’. It then introduces the example of the equitable man, who gives more to another than to himself. So is the equitable man not a case of someone who wrongs himself? The answer (1136b21–5) is, first, that the equitable man can in a sense be getting more, namely of reputation or the noble, and, second, that, since he acts voluntarily, then, as stated already (1136b3–9), he is not wronged but at most harmed.

As for the Eth. Eud. reading, the two things to discuss are not as obvious. In addition, the ‘because’ (ὅτι and not ἔτι) with which the sentence begins is puzzling. It looks in the context, however, to be referring back to Eth. Eud.’s initial list, which was about the two things of voluntariness on the part of being wronged and on the part of wronging. For these two things certainly appear again in the case of Glaucon and Diomedes, who have engaged in a distribution where Glaucon appears to be wronged in having less and Diomedes to be wronging in having more. We have just been told, of course, that Glaucon was not wronged because he was willing, but if it is possible to wrong oneself (the condition the Eth. Eud. version adds to the question it asks), then perhaps Glaucon has wronged himself, and willingly, in taking the smaller share. So perhaps it was possible too for Diomedes, who takes the larger share, to have done wrong by taking the larger share. For if Glaucon has suffered wrong, though willingly, in taking the smaller share, then could not Diomedes be doing wrong, despite Glaucon's willingness, in taking the larger share?

If such is the question posed by the Eth. Eud. reading, then while it repeats the focus on the philosophical interest that is present in Eth. Eud.’s original list, how does what next follows fit in? For the next words are the same as in the Eth. Nic. version, and these clearly talk of Eth. Nic.’s case (someone engaging in a distribution and taking less) and not of Eth. Eud.’s case (someone engaging in a distribution and taking more). It is possible, however, to understand the words at 1136b17–18, ‘for if the first thing said is possible and the distributor of the greater share, but not the one who has it, does wrong …’, as referring in the Eth. Eud. version just to ‘does the one who distributes against desert do wrong’ and not also to the additional, and here somewhat incidental, remark ‘having the greater share’. Accordingly, there would be two cases to consider: the distributor doing wrong when he does not have the greater share (that is, when he wrongs himself) and when he does have the greater share (that is, when he wrongs the other).

The first case is discussed immediately and the answer given is the same as just explained for Eth. Nic. (1136b21–5). The second case seems to be implicitly included in the immediately following words: ‘It is also evident that it is the distributor who does wrong and not always he who has more’ (b25–6). For these words make it clear that what is decisive is not having more but being a distributor. So the question whether Glaucon was wronged in the exchange with Diomedes is settled not by whether Diomedes got more but by whether Glaucon was a distributor. For if he was, the wrong lies with him. But then, as a distributor, Glaucon was willing and so, by definition, not wronged. Therefore, by definition too Diomedes did not wrong him; and note additionally that Diomedes is only distributor in second place, as it were (he would be a mere taker and not a distributor if Glaucon did not first consent). Or as the text has it (113626–9): ‘For it is not he with whom the unjust thing is who does wrong, but he with whom is the voluntary doing of it; this is where the beginning of the deed starts, and it is in the distributor not in the receiver.’

There is in addition a slight verbal variation to note. The Eth. Nic. version says ‘in the receiver’ (ἐν τῷ λαμβάνοντι), while the Eth. Eud. version says ‘in the receiving’ (ἐν τῷ λαμβάνειν). The Eth. Eud. version would seem thus to be downplaying the person having relative to the person distributing. Perhaps nothing hangs on the difference, but if anything does, the Eth. Eud. version fits better the way it initially posed its question (1136b15–16), which focusses on the distributor, making the correlative haver secondary: ‘does the one who distributes against desert do wrong, having the greater share?’ Similarly, and interestingly, the Eth. Nic. version fits better the way in which it initially posed its question, which sets the correlative haver on the same level as the distributor: ‘whether it is the one who distributes the greater share against desert or the one who has it who does the wrong’. This shift from ‘receiving’ to ‘receiver’, while slight, does cohere with this initial difference. In any event the text, which is otherwise the same, answers the questions posed in the Eth. Nic. and Eth. Eud. versions, and without changing anything in, or by even adding something to, Eth. Nic.’s primarily legislative focus and Eth. Eud.’s primarily philosophical one.

3) 1147b23–31, 1148a22–6, 1148a28–31

1147b23–31: ἐπεὶ δ’ ἐστὶ τὰ μὲν ἀναγκαῖα τῶν ποιούντων ἡδονήν, τὰ δ’ αἱρετὰ μὲν καθ’ αὑτὰ ἔχοντα δ’ ὑπερβολήν, ἀναγκαῖα μὲν τὰ σωματικά (λέγω δὲ τὰ τοιαῦτα, τά τε περὶ τὴν τροφὴν καὶ τὴν τῶν ἀφροδισίων χρείαν, καὶ τὰ τοιαῦτα τῶν σωματικῶν περὶ ἃ τὴν ἀκολασίαν ἔθεμεν καὶ τὴν σωφροσύνην), τὰ δ’ ἀναγκαῖα μὲν οὐχί, αἱρετὰ δὲ καθ’ αὑτά (λέγω δ’ οἷον νίκην τιμὴν πλοῦτον καὶ τὰ τοιαῦτα τῶν ἀγαθῶν καὶ ἡδέων)· ‘but since some things that cause pleasure are necessary and others, while in themselves preferable, involve excess, the bodily ones are necessary (I mean the sort of things to do with food and the use of sex and all the sorts of bodily things we set down license and moderation as being about), but the others, while not necessary, are in themselves preferable (I mean, for instance, victory, honour, wealth and such goods and pleasures as this).’ The Eth. Eud. MSS have the same save for adding two extra ‘nots’, changing τὰ δ’ αἱρετὰ μὲν καθ’ αὑτὰ ἔχοντα δ’ ὑπερβολήν at the beginning to τὰ δ’ αἱρετὰ μὲν οὐ καθ’ αὑτὰ ἔχοντα δ’ ὑπερβολήν, and then changing τὰ δ’ ἀναγκαῖα μὲν οὐχί, αἱρετὰ δὲ καθ’ αὑτά a few lines later to τὰ δ’ οὐκ ἀναγκαῖα μέν, οὐχ αἱρετὰ δὲ καθ’ αὑτά: ‘but since some things that cause pleasure are necessary and others, while not in themselves preferable, involve excess … but the others, while not necessary, are not in themselves preferable (I mean, for instance, victory, honour, wealth, and such goods and pleasures as this).’

A first oddity here is that the additional ‘nots’ seem to destroy the contrast being drawn (emphasized by the μέν … δέ … constructions). An immediate response, then, would be to dismiss these ‘nots’ as scribal error. However, on closer consideration, the ‘nots’ do not so much destroy the contrast as change it. They still draw a distinction between necessary things and non-necessary things that cause pleasure, but, unlike the Eth. Nic. version, they do not identify these non-necessary things (victory, honour, wealth and the like) with things preferable in themselves. Instead, they expressly deny that these non-necessary things are thus preferable. So if a contrast is still preserved, though a different one, what is the reason and point of the difference?

Later, at 1148a22–6, both Eth. Nic. and Eth. Eud. versions talk of desires and pleasures that are noble and virtuous (καλῶν καὶ σπουδαίων) in their kind, and assert that some of the pleasant things are naturally preferable, some the opposite, and some in between. Those in between are then said to be victory, honour and wealth. These remarks seem a better fit with Eth. Eud. and its earlier ‘nots’ than with Eth. Nic. and its absence of them (though there is presumably a difference between saying first that such things are ‘in themselves’ preferable and saying later that they are not ‘naturally’ preferable but in between). However, a couple of lines later, at 1148a28–31, honour (with children and parents) is listed among things naturally noble and good, which seems a better fit with Eth. Nic. But Eth. Eud., in fact, says things ‘naturally noble or good’ and not things ‘naturally noble and good’ (ἤ instead of καί at a30). The ‘or’ of Eth. Eud. allows the thought that honour is not noble but just good, while Eth. Nic. asserts honour to be noble as well as good.

So Eth. Nic. moves from saying that honour and the like are preferable in themselves, to saying that they are not naturally preferable but in between, to saying that they are noble and good. Eth. Eud., by contrast, moves from saying that honour and the like are not preferable in themselves, to saying that they are not naturally preferable but in between, to saying, or allowing, that they are not noble but just good.

We can construe these textual differences in Eth. Eud. as again reflecting the contrasting orientation of the two ethics: Eth. Nic. to legislators and Eth. Eud. to philosophers. There is strictly no conflict in Eth. Nic.’s saying first that something is preferable in itself, second that it is not preferable naturally, and third that it is noble. There is nevertheless a certain going up and down involved. The up-and-down look to reflect the up-and-down nature of the political or practical life (that it is happy and virtuous and noble, but secondarily so, and that it suffers impediments and is more subject to chance, Eth. Nic. 10.8). There is no like up and down in Eth. Eud.’s saying that something is not preferable in itself, not preferable naturally, and not noble but just good. For there is also no like up and down in the philosophical or theoretical life or in a life devoted to worship of the god. This life is wholly focussed on what is always simply best, that very worship of the god (Eth. Eud. 8.3).

Note then a further passage in the same context, at 1148b2, which in both the Eth. Nic. and the Eth. Eud. versions reads: μοχθηρία μὲν οὖν οὐδεμία περὶ ταῦτ’ ἐστὶ διὰ τὸ εἰρημένον, ὅτι φύσει τῶν αἱρετῶν ἕκαστόν ἐστι δι’ αὑτό. These words are naturally translated as ‘now there is in these things [honour and the like], on account of what was said, no wickedness, because each is of itself among things naturally preferable.’ But as so translated the words seem to be in conflict with what has just been said, that these things are not naturally preferable. Another way of taking these words then is to note that the φύσει is outside the τῶν αἱρετῶν and to translate: ‘because each is of itself by nature among things preferable’. Thus taken, these words now cohere with what Eth. Nic. said in the first passage, that these things are in themselves preferable, and they do not conflict with what both Eth. Nic. and Eth. Eud. say in the second passage that they are not naturally preferable. For something can be naturally among things preferable but not itself naturally preferable; it is one thing to be by nature in a certain class and another for that class to be itself natural. A table, for instance, is by nature an artificial thing, but an artificial thing is not a natural thing. So similarly a thing could be by nature preferable without being by nature naturally preferable.

But how must these words be taken in Eth. Eud., since Eth. Eud. has said in its first passage that these goods are not preferable in themselves? The sense must be analogous, that these things are in themselves and by nature among things preferable but not among things preferable in themselves. Why not? Because while preferable for most men most of the time, they are not preferable for the semi-divine philosopher, who prefers the contemplation of the god instead and limits all other preferable goods to that purpose (Eth. Eud. 8.3), even perhaps forgoing them altogether if need be (as religious hermits and gurus do). Eth. Eud.’s sense, therefore, will be that the things political men value and are tempted to be excessive in, victory and honour and wealth, are not preferable in themselves (though excess in them is not disreputable like excess in bodily and necessary pleasures). Nevertheless, legislators can still rightly regard them as preferable in themselves in the context of legislation, and so in the context of Eth. Nic., where the political life is a happy life, though a secondary one (Eth. Nic. 10.8). Within that life honour and wealth and victory are preferable in themselves, at least up to the point the political life needs. Beyond that point, or where even the political life must ultimately be directed to what is beyond it (the contemplation of the god), they take on a different complexion. So the additional ‘nots’ in Eth. Eud. would confirm the difference that Eth. Eud. is directed to philosophers and the contemplative life, while Eth. Nic., which does not have the ‘nots’, is directed to legislators and their concern with the practical life.