2.10.1226b2–5

The main problem here is ἐστι προαίρεσις, ἐστὶν ὡς ἑκάτερον, for not only does there seem to be an unnecessary reduplication of ἐστι but this reduplication also seems to give a false sense. For how can choice be as each and yet not both? Indeed P and C seem to reflect this puzzle because they omit ἐστι προαίρεσις altogether. The OCT, by contrast, follows Allan in deleting the first ἐστι and the comma after προαίρεσις to give the sense: ‘So, since choice is not opinion or will as each of them nor both …’ It also notes in the apparatus Ross's suggestion to read ἐστιν <ἡ> προαίρεσις {ἐστιν} ὡς ἑκάτερον. Susemihl leaves the text as is save for removing the comma, but notes in the apparatus that Bussemaker also deleted ἐστιν after προαίρεσις and that Bonitz (in part anticipating Ross) suggested deleting ἐστι before προαίρεσις and replacing it with ἡ. However, there is no real need to add ἡ (and the OCT does not add it) nor, more importantly, is there need for any other emendation. In fact the emendations give a false sense for, although it looks odd at first sight to say that choice is each but not both, Aristotle's meaning seems to be precisely of this sort when he concludes, in the next line, that choice is from both. For, as he says later, it is from both in the sense that choice is both opinion and appetite (of which will is one species) when these, as a result of deliberation, are brought together in a conclusion (1227a3–5). A better solution, then, is to keep Bekker's text as it is and construe ἐστιν ὡς the way phrases like ἔστιν ὅτε are construed (‘there is a time when …’) and get the meaning ‘there is a way that …’ Hence we should accent differently and print προαίρεσις, ἔστιν ὡς and translate: ‘So, since choice is neither opinion nor will, there is a way that it is each but not both …’ The ‘there is a way that’ will now be qualifying the ‘it is each but not both’ in anticipation of the conclusion about the nature of choice that Aristotle finally reaches.

3.1.1229a11–16

The problem word here is the ὅτι in ἀλλ' ὅτι τὰς βοηθείας τῶν δεινῶν which is not only unnecessary but disrupts the sense. Susemihl prints Bekker's text but notes in the apparatus that Sylburg proposed deleting ὅτι, which is followed by the OCT, and that Bonitz proposed instead adding ἴσασι. A simpler solution is to leave things as they are and regard the construction as an infinitive and accusative clause with pleonastic ὅτι (LSJ s.v. II.2), the ὅτι serving, in this case, to show that the accusatives are not objects of εἰδέναι but an indirect statement dependent on it: ‘… knowing, not terrible things, as Socrates said, but that there are helps for terrible things.’

3.4.1231b38–32a4

The only problem here is the word ὑποδήματι, for the manuscript tradition is unclear. The OCT reports that P has ὑποδη with μτ added as a superscript, that C has ὑποδημ with τ added as a superscript, and that L has ὑποδήματα. The reading ὑποδήματι is proposed by Victorius and is adopted by Susemihl. The OCT, however, thinks that this proposal is not sufficient by itself and follows Richards in adapting from Pol. 1.8.1257a11–12 and writing χρῆται γὰρ <ᾗ> ὑποδήματι <ὁ δεόμενος> (‘For the one who needs it uses it as a shoe’). This more extensive emendation gives a better sense than ὑποδήματι by itself but it has the disadvantage precisely of being more extensive and so harder to explain philologically. The OCT also puts the whole phrase into parentheses, which may well be an improvement but does nothing by itself to favour or disfavour the emendation. An alternative to the emendation, then, would be to follow L and write ὑποδήματα but understand χρῆται as passive (‘are provided’) and not as middle (‘it uses’), and so translate: ‘for shoes [sc. and not weights] are provided [sc. by selling and lending shoes].’ Aristotle will thus be playing on the double meaning of the verb χράω: to provide, and (in the middle) to use; the point is that in selling and lending one is using shoes as a medium of exchange and not for wearing (so that the use is per accidens), but one is nevertheless providing people with shoes and not, say, with weights. The drawback with this suggestion is, of course, that it depends on exploiting a double use of χράω, which, in the absence of any warning of the fact, might be hard to catch, even for native Greek speakers. But perhaps Aristotle intends such readers to catch it when they see that without a double use the phrase will be ungrammatical. Compression of statement, both in grammar and logic, is after all a feature of EE so that to posit another instance of it here is hardly out of keeping with the rest of the text; it speaks more in favour of the suggestion than against it.

3.6.1233b6–15

There are several problems here, as the rather wooden translation shows. Susemihl follows Fritzsche in marking a lacuna after δεῖ δὲ πρέπον εἶναι and in deleting ὃ in ὃ οὐθεὶς τούτων. He also includes in one parenthesis all of καὶ γὰρ τοῦ πρέποντος κατ' ἀξίαν καὶ πρέπον and follows Casaubon in emending ἔστι γάρ τις ὡς ἐλευθέριος, ὅταν ἐλεύθερος to read ἔστι γάρ τις οὔτ' ἐλευθέριος οὔτ' ἀνελεύθερος (‘it is possible for someone to be neither liberal nor illiberal’). The OCT follows Casaubon's emendation, does not delete ὃ in ὃ οὐθεὶς τούτων, accepts Susemihl's parenthesis, and marks as corrupt the whole of πρέπον εἶναι (καὶ γὰρ τοῦ πρέποντος κατ' ἀξίαν καὶ πρέπον). The passage is certainly obscure but there is a way to make sense of Bekker's text as it stands if, in particular, one gives to τοῦ πρέποντος a verbal force and treats it as a compressed genitive absolute: ‘But he who does what is according to worth and as reason says, he is magnificent (for the fitting accords with worth; for nothing is fitting that is contrary to worth). But it must be fitting, for as indeed the fitting is in accord with worth, there must in fact be a fit both as regards what (for example, one thing fits the marriage of a servant and another the marriage of a beloved) and for him, if he is really up to so much or such; for example the thinking was that the embassage Themistocles made to Olympia was not fitting for him (on account of his former low condition), but it was for Cimon. He who is disposed to worth as it chances is the nobody among these. Things are the same way too with liberality, that there is someone who is as a liberal man is – when he is free to be.’ The sense of the last two sentences is that someone might be disposed as the liberal man is but only be able to behave so when the conditions are met, namely when he is free of restraints (financial or physical), just as also in the case of magnificence. For one could be disposed as the magnificent man is but be unable to behave so because of lack of preconditions, like Themistocles who, though he had, like Cimon, the resources to put on a splendid display, did not have the right family status, and so, when he tried to be magnificent, only succeeded in being a swaggerer. He should have refrained because he was out of his league, or a ‘nobody’, in the context.

7.2.1236b1–5

The oddity here is the abrupt change from the neuter ϕίλον in the first clause to the masculine ϕίλος in the second. Susemihl leaves the text as it is but notes in the apparatus Fritzsche's suggested emendation of ϕίλος δὲ τῷ ϕιλουμένῳ to ϕίλος δὲ ὁ ϕιλούμενος (‘the beloved is dear when he loves back) as well as his own to ϕίλος δὲ τῷ ϕιλουμένῳ καὶ ἀντι< ϕιλοῦντι καὶ ὁ> ϕιλῶν (‘dear to the beloved and to the one who loves back is also he who loves’). The OCT adopts Fritzsche's emendation. However, Fritzsche's sense, which seems right, can be found in the text already, though with considerable economy of expression (a particular feature of EE's style): ‘for dear to the lover is what is beloved, but dear to the beloved when he also loves back.’ A remark is first made, then, about the thing loved (in the neuter) being dear to the lover, and then a remark is made about him who loves back (in the masculine) being dear to him who is loved (i.e. in this case, loved back). The abrupt shift in gender is, it seems, deliberate, so as to highlight the point at issue: inanimate things like wine can be loved, but only animate things like other human beings can, when loved, love back and make the lover into the beloved. That is why, as Aristotle says in the opening sentence, friendship is only possible in the second case and not in the first.

7.2.1237b10–13

The problem scholars find with this passage is the οὐ in οὐ ποιεῖ, for it seems the opposite must be true, that things that do not happen quickly and easily do make judgement right and not that they do not make it right. At any rate the following remarks about friendship and trust seem to mean that friendship is only rightly judged after passage of time. Hence both Susemihl and the OCT agree with Bonitz in deleting οὐ. The deletion is unnecessary. The sense is that things that happen quickly and easily make for correct judgement because what they really are is soon seen, while things that do not happen quickly and easily do not make for right judgement because what they really are is not soon seen. Friendship is of the latter kind because it needs trust and trust needs time. So translate more along the lines: ‘For what has passed judgement is firm, but when things do not happen quickly or easily the judgement is not made rightly, and there is no firm friendship without trust and there is no trust without time.’

7.2.1238a4–8

The problems here are clear from the literal translation. Susemihl and the OCT both follow Λ¹ and add δ᾿ to ἁπλῶς <δ᾿> ἀγαθὸς so as to give the new sentence its opening particle. Instead of τῷ τούτου ἄλλῳ Susemihl follows P² and writes τὸ τούτου ἄλλῳ but with what sense is unclear since he marks the whole of ὅ ἐστιν … ὅτι χρήσιμος as corrupt. The OCT follows Richards in writing τὸ αὐτὸ ἄλλῳ and, like Susemihl, follows P and C in writing εἰ καὶ for ἢ καὶ, thus giving the following sort of sense to the whole clause: ‘… so that what is simply good, the same is good to another, even if to one not simply virtuous, but to another he is good, because useful’. Again, however, Bekker's text can be given an intelligible sense as it stands. First, if one removes the stop that Bekker puts after τῷ ἄλλῳ ἀγαθός and makes it one sentence with what follows (so that no δ' needs to be inserted between ἁπλῶς and ἀγαθὸς), one can translate: ‘For a simply good man and friend, when these are both in harmony, is simply good by being good and dear by being good for another.’ Second, if one construes ὅ ἐστιν ἁπλῶς ἀγαθόν rather differently, treats τούτου in τῷ τούτου ἄλλῳ as a genitive of comparison with ἄλλῳ, and retains ἢ καὶ but accents ἢ as ἦ so that it means not ‘or’ but ‘yes’, one can translate: ‘Consequently, what he is simply is a good for the man other than this – yes, even for one not simply virtuous but other [sc. other than virtuous] he is good, because he is useful.’ Alternatively, one could, perhaps, retain the ἢ καὶ or the εἰ καὶ of the manuscripts and, instead of: ‘yes, even for one not simply …’ translate: ‘or even for one not simply …’ or ‘even if for one not simply …’

This sort of construal of the Greek is not obvious but it is not impossible either, and that Aristotle is requiring his readers here to puzzle out his grammar as much as his logic is manifest from the phrase ϕίλος δὲ τῷ ἄλλῳ ἀγαθός (where, to make sense, the τῷ must be construed with an understood εἶναι and not with ἄλλῳ). The sense, at any rate, will be that the simply good man and friend, while he is of course a good for a virtuous friend, will also, because he is simply good, be a good for anyone else, including those who are not virtuous, for he will be useful to them. First, then, he will be useful to other virtuous people with whom he is unable to be virtuous friends (one can only be virtuous friends with a few and not with many, as Aristotle remarks in the immediately next sentence), because he will deal with them justly. Second, he will be useful to the vicious for the same reason (the vicious will be able to rely on his honesty), and also because, by rebuking them as occasion requires, he may help to reform them.

7.4.1239a28–33

The problem words here are ἔνεστι γὰρ ἀνάγκη ἐνεργοῦντα· τὸ μὲν γὰρ ϕιλεῖν συμβεβηκός· ἔστι γὰρ λανθάνειν ϕιλούμενον, ϕιλοῦντα δ' οὔ. The OCT follows Sylburg and Robinson in replacing ἔνεστι γὰρ ἀνάγκη ἐνεργοῦντα with ἔστι γὰρ ἀνάγκη <ϕιλεῖν> ἐνεργοῦντα (‘it is necessary that he love when active’), and follows P² and Λ¹ in replacing τὸ μὲν γὰρ ϕιλεῖν συμβεβηκός with τὸ μὲν γὰρ ϕιλεῖσθαι συμβεβηκός (‘for being loved is accidental’). Susemihl makes the second change but not the first though he remarks in the apparatus that the text is somehow corrupt. If it is somehow corrupt the simplest change would be to follow Richards in replacing ἀνάγκη with ἀνάγκῃ, for then ἐνεργοῦντα (which need not be changed, with Richards, to ἐνεργοῦντι) can be read as an accusative absolute to give the sense: ‘for it [sc. the pleasure in loving] is necessarily present within when he is active’ (pleasure is necessarily present when activity is unimpeded). It is not, however, necessary to change ϕιλεῖν to ϕιλεῖσθαι for, despite the seeming oddity of the remark, we can, if we press the transitive force of ϕιλεῖν, take it to mean that having an object of one's love is accidental (for the object might be unaware that it is the object). The point can be reinforced by noting that the position of μὲν and δ' requires us to contrast ϕιλοῦντα δ' οὔ with τὸ μὲν γὰρ ϕιλεῖν συμβεβηκός and not with ἔστι γὰρ λανθάνειν ϕιλούμενον. The translation should therefore be: ‘for that he has a love is accidental (for it can escape the notice of the object loved) but his being a lover is not’. An act of love has an object that it loves and a subject that does the loving. The object is accidental to the act because the act can exist even if the object knows or cares nothing of the fact, but the subject is not because the act exists in the subject.

7.5.1239b29–32

Both Susemihl and the OCT take exception to the last phrase and instead of … ἀλλήλων· διὸ οὐ τῷ γίνεσθαι write, following Spengel, ἀλλήλων διὰ τὸ οὕτω γίνεσθαι … (‘they are drawn to each other as matching tallies because they thus become a single thing from both’). The emendation gives a false sense. For, as the example of the tallies shows, the point is that the extremes do not, when coming together, cease to be extremes, but just have their defects supplied by each other. Some confirmation is given a few lines later (1240a1–3) when Aristotle speaks of unlike people rejoicing in each other (as the austere in the witty and the hasty in the sluggish) because they bring each other toward the mean. The hasty and sluggish, and the austere and witty, do not hereby become a one; their temperaments are just relieved by each other. We should, therefore, leave Bekker's text (and the translation) as it is.

7.6.1240a33–9

The problem words here are εἰ ἐνεδέχετο, ὅτι μὴ ἐγγύτατα. Susemihl deletes μὴ with Fritzsche but notes in the apparatus Fritzsche's further suggestion of replacing εἰ with ἢ εἰ μὴ. The OCT also deletes μὴ, adopts Fritzsche's further suggestion (P and C have ἢ in place of L's εἰ), but adds a comma, with Robinson, to give ἢ, εἰ μὴ (‘… to thirst with him when thirsty or, if it was not possible, as nearly as could be’). Bekker's text can, however, be given an acceptable sense, and perhaps indeed a better sense, if we take the verb ἐνεδέχετο not in its meaning of ‘it was possible’ but in its primary meaning of ‘to take upon oneself’ and translate: ‘… to thirst with him when thirsty, if he took it upon himself …’ The main problem here is that the verb ἐνδέχεσθαι so often has the modal force of ‘can’ or ‘be able’ that it may seem a stretch not to give it that meaning here too. There is, however, another place in the Aristotelian corpus where the verb seems to require being understood in the sense of ‘take upon oneself’. The passage is Poet. 24.1460a33–35 where Bekker's text reads: ἐξ ἀρχῆς γὰρ οὐ δεῖ συνίστασθαι τοιούτους· ἂν δὲ θῇ καὶ ϕαίνηται εὐλογωτέρως ἀποδέχεσθαι καὶ ἄτοπον … (‘[The poet] should not construct such plots. But if he does posit [such a plot] and it appears fairly plausible, he should accept even an absurdity’). However the manuscripts have ἐνδέχεσθαι in place of ἀποδέχεσθαι and ἀποδέχεσθαι is an emendation proposed in certain late manuscripts.Footnote ³ The emendation is unnecessary. Since ἐνδέχεσθαι in its primary sense means ‘accept’ or ‘take upon oneself’ it is already in this sense equivalent to ἀποδέχεσθαι.Footnote ⁴ Accordingly, one should allow that Aristotle might very well have been using the same verb in the same sense here in EE.

As for the final phrase ὅτι μὴ ἐγγύτατα, it too can be given an acceptable sense, but it is ambiguous. The words ὅτι μὴ can either be taken together with the sense of ‘unless’ or ‘except’, or be taken separately with the sense of ‘because not’. In the first way the phrase will mean: ‘[… took it upon himself,] unless very close by’, and in the second: ‘[… took it upon himself,] because not very close by’. The meaning in either case is more or less the same, that the friend would in solidarity take it upon himself to suffer as his friend is suffering (e.g. to thirst when he is thirsting) unless he is sufficiently close by to come to his aid and relieve him directly. Both construals have, however, grammatical problems. ὅτι μὴ taken as a single phrase normally follows a negative clause (LSJ s.v. ὅτι II), and ὅτι taken separately would have οὐ rather than μὴ as its negative. As regards the problem of the negative particle, since the ὅτι clause is here not only causal but also conditional (it is a subordinate part of the εἰ clause), a μὴ is perhaps not impossible.Footnote ⁵ As regards the problem of the negative clause, one could perhaps write ὅτε instead of ὅτι. For ὅτε μὴ is likewise a phrase but need not follow a negative clause, and it too means ‘unless’, ‘except’, ‘save when’ (LSJ s.v. ὅτε I.2.c).

7.6.1240b28–34

The problem phrase here is ἐπὶ δὲ τῶν ἄλλων ζῴων οἷον ἵππος αὐτὸς αὑτῷ. Susemihl follows Spengel in adding οὔ after ἄλλων ζῴων (‘while in the case of the other animals it is not so …’) and marks, on his own account, a lacuna after αὐτὸς αὑτῷ. The OCT also adds οὔ but follows Rieckher in adding οὐκ ὀρεκτός after αὐτὸς αὑτῷ i(‘… it is not so, for example a horse is not itself desirable to itself’). However, again Bekker's text can be given an acceptable sense as it stands: ‘Hence in the human case each seems to be himself a friend to himself, while in the case of the other animals, as a horse, each seems to be itself to itself, so it is not a friend.’ An animal, unlike a man, is not a ‘two’ in its soul (it is just an ‘itself’) and so cannot be a friend to itself. Like the child mentioned next, animals have desire but no choice and so do not have parts that can either conflict or be brought into friendship with each other. Hence, further, we should not follow either Susemihl or the OCT (themselves following Fritzsche) in replacing παῖς with νοῦς in the next sentence. When a child is old enough to have choice it can be in conflict with itself in relation to desire (it judges and chooses one thing as better but is drawn by desire to something else); hence it is then a ‘two’ and can be, or fail to be, a friend to itself. Animals, by contrast, never get to this point because they never get to have choice.

7.10.1242a20–5

However, the words ὁ γὰρ ἄνθρωπος οὐ μόνον πολιτικὸν are Casaubon's emendation (followed also by Susemihl and the OCT), and what the manuscripts have is ὁ γὰρ ἀνθρώπου μόνον οὐ πολιτικὸς (‘for he of man is only not political’). The emendation seems obvious, but if we supply the word for ‘friend’ from the preceding sentence we can translate: ‘For he who is friend of a man alone does not belongFootnote ⁶ to a city [= is not political], but he is indeed a household animal …' The phrasing is peculiar but it does have a sense. Moreover this sense serves a logical role in the argument in the context. For he who is ‘friend of a man alone’ will be someone who is friends with one or a few and not with many in a city. But such a person, even if he is not part of a city, must at least be part of a family (he is a household animal), for man, unlike other animals, sets up permanent relations with another and does not pass from one chance encounter to the next. Hence, if everyone is either part of a city or not, and if those who are not are still friends and sharers with others, at least in a household, and if, further, being friends and sharers in city and household is a relation of justice, then all relations of friends are relations of justice, which is the statement Aristotle began with (or part of it; the converse, that all relations of justice are relations of friends, is something he argues for in the larger context).

7.10.1243b14–18

The problems here are clear from the literal translation (note, though, that the meaning of ‘not in a straight line’ is clear from the example: those are friends ‘not in a straight line’ who are friends on different bases, as one on that of pleasure and the other on that of utility). Both Susemihl and the OCT (following Bonitz and Fritzsche) transpose γὰρ to follow χαλεπὸν and change the punctuation, writing: … κατ' εὐθυωρίαν, καὶ τὸ δίκαιον ἰδεῖν οὐ ῥᾴδιον. χαλεπὸν γὰρ μετρῆσαι ἑνὶ τῷδε τὸ μὴ κατ' εὐθυωρίαν (‘… in a straight line, and to see the just is not easy. For it is hard to measure by this one thing what is not in a straight line’). The emendations seem obvious. However, Bekker's text can again be given a sense and, though nothing of significance seems to hang on the fact, the sense is worth noting: ‘Many complaints arise in friendships for those not in a straight line even to see the just. For to measure a difficult thing by this one thing is not easy for him who is not in a straight line, as happens in love affairs …’ Those who are not friends on the same basis are ‘not in a straight line even to see the just’ because each is looking in different directions, as one at pleasure and the other at utility. Friendships on different bases are ‘a difficult thing’ and to measure such a difficult thing by ‘this one thing’, namely by justice (each friend getting his due), is ‘hard’ for ‘him who is not in a straight line’, namely for each of the friends, for neither is in a straight line with the other.

7.12.1245b1–7

Again the problems are clear from the literal translation. Note first, however, that μάλιστά γε is from the Aldine editio princeps; the manuscripts read μάλιστά τε. Susemihl and the OCT both write as follows: … ὁτὲ δὲ ἕτερόν τι. εἰ δ' αὐτὸν εὖ ζῆν, καὶ οὕτω καὶ τὸν ϕίλον, ἐν δὲ τῷ συζῆν συνεργεῖν, ἡ κοινωνία τῶν ἐν τέλει μάλιστά γε. διὸ <δεῖ> συνθεωρεῖν καὶ συνευωχεῖσθαι … (‘… sometimes something else. But if it is pleasant [supplied from the previous sentence] to live well oneself and for one's friend also to live thus, and if in living together there is working together, surely their sharing will be mainly in things included in the end. Hence it is necessary to contemplate together and feast together …’). Here the δ' αὐτὸν comes from L, ἡ κοινωνία from D, and δεῖ from Fritzsche. Nevertheless, the readings of all the remaining manuscripts, including μάλιστά τε (which there is no compelling reason to change to μάλιστά γε), can be given an acceptable sense and, though again nothing of significance seems to hang on the fact, this sense is worth noting: ‘And this is sometimes a feeling, sometimes a doing, sometimes something else, provided one can live well, and one's friend in this way too; and it [sc. the feeling or doing] is in working together at living together or at a community of things included in the end, and particularly so. Hence it is to contemplate together and feast together …’

8.1.1246a26–9

First, L and C² omit ϕίλῳ, followed by both Susemihl and the OCT. In addition ἐϕ' ἃ is the reading of Laur. 81.12; the other manuscripts have ἐϕ' ᾧ. Instead of καὶ τοῦτο ἢ αὐτὸ ἡδὺ κατὰ συμβεβηκός Susemihl follows Spengel in writing καὶ τοῦτο ἢ <καθ᾿> αὐτὸ ἢ κατὰ συμβεβηκός (‘… and this either by itself or accidentally’) and the OCT follows Dirlmeier and Jackson in writing καὶ τοῦτο ἢ <ᾓ> αὐτὸ ἢ αὖ κατὰ συμβεβηκός (‘… and this either qua itself or again accidentally’). The emendations are plausible (and Jackson's ἢ αὖ for ἡδὺ ingenious), but if one combines them with the deletion of ϕίλῳ one loses any connection to what has gone before, and so makes the beginning of this book (or chapter; 8.1 is also numbered 7.13) very abrupt. An alternative is to note that the topic that ends the previous chapter is about the bad man using his friend badly by killing him along with himself because he judges it pleasanter, or less painful, to know his friend will die with him than go on living beyond him (7.12.1246a18–25). If, however, ϕίλῳ is kept the question with which this book begins is a natural continuation, for it is about whether one can indeed use each friend, or each dear thing, both for what he is naturally for and otherwise (that is, badly, as the bad man does). The chapter does, indeed, range beyond the limitation of this question to friends and raises it about virtue and prudence, but each of these is just a particular case of the same general puzzle. A suggestion, then, is to keep ἡδὺ as well as ϕίλῳ, prefer ἐϕ' ᾧ to ἐϕ' ἃ, and follow Jackson in writing καὶ τοῦτο ᾓ αὐτὸ instead of καὶ τοῦτο ἢ αὐτὸ: ‘One might raise the puzzle whether it is possible to use in fact each thing that is dear on the basis natural to it and otherwise, namely this [treating καὶ as epexegetic], in so far as it is pleasant incidentally [sc. the way the bad man uses his friend badly, because of the incidental pleasure he gets from knowing his friend will not survive him].’ In this way too the example Aristotle immediately gives of an incidental use of an eye, namely to eat it (b31), falls neatly into place,Footnote ⁷ for an eye might very well be judged pleasant to eat (by those who like eyes).

8.3.1248b39–49a3

Susemihl and the OCT print the same as Bekker save that both follow Victorius and Λ¹ in writing καλοκἀγαθίαν δὲ οὐκ ἔχουσιν (‘but they do not have the quality of gentlemen’), and the OCT follows Solomon in writing τὰ γὰρ ϕύσει μὲν ἀγαθὰ <ἀγαθὰ> αὐτοῖς ἐστίν. The δὲ seems a definite improvement over the γὰρ of the other manuscripts (it nicely answers the preceding μὲν), but the extra ἀγαθὰ seems unnecessary. However, the main issue is that none of the manuscripts have διὸ ἀγαθοὶ μὲν ἄνδρες εἰσί. In place of ἀγαθοὶ they all have ἄγριοι (‘wild’); ἀγαθοὶ is first found in the Aldine editio princeps. The reason for the change is clear. If the people in question, the Spartans notably, have the virtues, though they have them for the sake of natural goods, they must be good men (for to have the virtues is to be good). They do not fail to be good, then; they just fail to be gentlemen (they are ἀγαθοί but not καλοί). Hence it must be wrong to say they are wild.

The reasoning is not compelling. Aristotle does think the Spartans were like wild beasts, and precisely because they pursued virtue for the ultimate sake of gaining the natural goods (Pol. 2.9.1271a41–b17, 7.15.1334a40–b3). Indeed when speaking of the Spartans' training he actually uses of them the word θηριώδεις, compares them to animals and nations that are ἀγριωτάτοις, and says they thus make their sons βαναύσους (8.4.1338b12–13, 18, 29, 33). Recall too that the brutality of the Spartans when in possession of rule over others was notorious. Aristotle's point would seem to be (and it is a point suggested by the whole of the chapter on the gentleman) that virtuous acts, like most acts, can have two ends: the immediate end which is internal to the act, and the remote end which is the sort of overall life for which the act is done.Footnote ⁸ A Spartan soldier, for instance, is brave in battle for he holds bravery to be a noble thing. He is thus really brave (for intending the beauty of the virtuous act is a mark of virtue). But he is brave so that he and his fellow Spartans can live a life of dominance over others in enjoyment of the natural goods, whether the others be helots at home or Greeks abroad. The problem, as Aristotle remarks (Pol. 2.9.1271b3–6), is that Spartans only practised with a view to war and did not know how to be at leisure; hence, while they had the virtues of occupation (courage, endurance, justice, moderation), they did not have the virtue of leisure, philosophy (Pol. 7.15.1334a19–34). In other words, they knew only of the secondary happiness of the political life and not also the primary happiness of the contemplative life (for the sake of which, however, politics exists, EN 10.7.1177b4–6, 12–15). So they were drawn, as simply political men are, to covet the material goods and occasions in which the virtues of the political life are best displayed (EN 10.7.1177b6–12, 8.1178a28–b7). To call such men wild does not seem inappropriate. The gentleman, by contrast, is brave and moderate and just for the sake of a life of philosophic leisure, and not of dominance over others. He intends not only, like the Spartan, the beauty of each act of virtue, but also the beauty of a whole life of virtue, and especially of the virtue of philosophy.Footnote ⁹

7.6.1240b1–3

Both Susemihl and the OCT follow Casaubon in deleting μὴ and replacing ϕιλίαν with ψυχὴν. That there is something wrong with μὴ is evident, for if a negative is necessary it should really be οὐ (an indirect statement regularly retains the negative of the original direct statement, which here would be οὐ) and, further, true friends would seem to be those whose friendship, or whose soul, is most single or united rather than not being so. The deletion of μὴ is philologically easy since it could have arisen by dittography from the μίαν that follows.Footnote ¹⁰ Jackson, however, following Susemihl thinks something more has gone wrong and suggests emending καὶ to κἂν εἰ to give: κἂν εἰ μὴ μίαν ψυχὴν (‘even if we do not allow the exaggeration μία ψυχή’).Footnote ¹¹ But if indeed more has gone wrong here than can be corrected by deletion of μὴ, a simpler solution philologically would be to regard μὴ as a scribal error for μὴν (the same scribal error occurs at 1245a15 where Bekker himself changed the manuscripts' ἀλλὰ μὴ to ἀλλὰ μὴν). The addition of μὴν to καὶ makes a certain sense anyway because Aristotle is giving two common sayings about friendship, but the first he gives in direct speech and in the nominative while the second he gives in indirect speech and in the accusative and infinitive. This grammatical change is no doubt because the first repeats the actual words people use (ἰσότης ϕιλότης is a sort of Ancient Greek jingle), but the second reports people's meaning and not necessarily their precise words. Now, as we learn from Denniston,Footnote ¹² καὶ μὴν can be used to indicate more than another addition to a list but also to express some further sense of difference or newness. Perhaps, then, Aristotle uses καὶ μὴν here to indicate not merely the new item but the newness too of the grammatical construction.

The replacement of ϕιλίαν with ψυχὴν is motivated in part because this very point is made a few lines later at b9 where the manuscripts do have μία ψυχή and not μία ϕιλία. Also, friends being μία ψυχή is a common saying but their being μία ϕιλία seems not to be. Further, Aristotle is here showing how the features of friendship are found in and referred back to the individual, and while an individual is one soul it seems odd to say he is one friendship. Probably the emendation should be accepted, but it is nevertheless worth noting that something can be said in defence of Bekker's text even if that text is finally rejected. First, to say true friends are one friendship covers, and so explains, all the features Aristotle mentions in the next lines while to say true friends are one soul repeats one of them. Second, true friends being one friendship would seem to be a saying that lies behind, and also explains, the phenomena of untrue friends listed later (7.10.1243b15–37). Third, bad men do not have one friendship with themselves but different and conflicting ones, being friends to themselves in pleasure when giving in to temptation but friends in utility when full of blame and regret later (1240b21–4), so that it can make a certain sense, at least as regards good men, to say that they are one friendship, because they are always friends to themselves in the same way at the same time and not friends in this way at this time and in that way at another.

7.12.1244b34–45a1

Both Susemihl and the OCT follow Fritzsche in deleting καὶ in τὸ ζῆν καὶ αἱρετόν. However, the main problem in the passage is καὶ ἐκ τούτων ὅτι τὸ αὐτὸ τοῖς ὑπάρχειν τὴν τοιαύτην ϕύσιν and especially τὸ αὐτὸ τοῖς, which makes no sense in the context (as the literal translation above shows). Susemihl leaves the manuscript readings as they are but reports in the apparatus Brandis's emendation αὐτοῖς and Fritzsche's emendation τὸ αὐτὸ δεῖ ὑπάρχειν τῇ τοιαύτῃ ϕύσει. The OCT prints Mingay's emendation τῷ αὐτῷ [sc. τρόπῳ] αὐτοῖς ὑπάρχειν τὴν τοιαύτην ϕύσιν. Presumably both Brandis and Mingay are supplying αἱρετόν [ἐστι] or the like from the previous lines, for otherwise the infinitive ὑπάρχειν will be ungrammatical. At any rate Rackham and Solomon in their translationsFootnote ¹³ adopt Brandis's emendation and translate: ‘that it is desirable for ourselves to possess that sort of nature [sc. good].’ Fritzsche's emendation will give the sense: ‘that the same thing must be present to such a nature’, which is a bit obscure.

All these emendations suppose that ὑπάρχειν is falling within the scope of the ὅτι clause; hence the need to supply a finite verb to govern it (Fritzsche's δεῖ or Brandis's and Mingay's αἱρετόν [ἐστι]). An alternative grammatical construction is to regard it as falling outside that clause and as being governed instead by the δεῖ at the beginning, thus giving the sense ‘it is necessary … that this sort of nature be present’. The problem now is what to do with the ὅτι clause and the τὸ αὐτὸ τοῖς that follows. Since something has to be done with τὸ αὐτὸ τοῖς in any event, an alternative suggestion is to see if we do not have here another case of EE's common error of misspelling and misdividing words. Now in antiquity and for several centuries thereafter manuscripts were written in capitals without word divisions, hence τὸ αὐτὸ τοῖς would appear as ΤΟΑϒΤΟΤΟΙΣ. But ΤΟΑϒΤΟΤΟΙΣ looks not unlike ΤΟΑϒΠΟΤΟΒ, and might look very like it depending on how carefully formed the letters were in the manuscripts being copied, and ΤΟΑϒΠΟΤΟΒ, when written out with proper word divisions, becomes τὸ Α ὑπὸ τὸ Β or ‘the A is under the B’. The expression is not uncommon in Aristotle's logical works (An. pr. 1.9.30a40, 30b13, 11.31a30, a37, b17, b20) to mean that one term falls under another. Hence the sense will be: ‘it is necessary that, because the A is under the B, this sort of nature be present/available’. How this sense fits the argument will be discussed directly. One immediate advantage of it, however, is that it gives a ready sense to the phrase that next follows: εἰ οὖν ἐστιν ἀεὶ τῆς τοιαύτης συστοιχίας ἡ ἑτέρα ἐν τῇ τοῦ ἑτέρου τάξει. The reference of ‘this sort of ordered series’ (τῆς τοιαύτης συστοιχίας) is usually said to be to some table or column of opposites, probably the Pythagorean, which Aristotle is understood to be pointing to as he makes the remark. But on the alternative suggested here, the ordered series is what he has just said, namely A falling under B or one term falling under another (hence we should follow Bekker in preferring the reading ἐν τῇ τοῦ ἑτέρου τάξει with P and C rather than ἐν τῇ τοῦ αἱρετοῦ τάξει with L).

This advantage will count for little, however, if the overall translation does not make sense in the context. So first, as regards the A and B, these will refer to the two propositions just stated (about the preferable and the good), and the sense will be that the predicate of the first falls under the predicate of the second. The first proposition speaks of living as to be preferred, and living in this sense, as Aristotle has argued in the preceding lines (b23–34), is the living that is our living – the living that we, and not someone else, experience as us being alive (hence we should probably keep καὶ at b35 because, rendered as ‘indeed’ or ‘in fact’, it serves to emphasize this point). However, when Aristotle first introduced this discussion he said that the thing will be plain if we take life ‘in actuality and as end’ (τὸ ζῆν τὸ κατ' ἐνέργειαν καὶ ὡς τέλος, b23–4), and life taken as end is happy life, for happiness is the end and the good. We should probably, therefore, translate the second proposition as ‘living is the good’ and not as ‘the good is to be preferred’, and understand the argument Aristotle is going to give as showing that, because the living that is to be preferred falls under the good, this ‘sort of nature’, that is, life as the good, must be ‘present’ to us, that is, available to us as something we can possess as our own. The argument will thus be that it is necessary to put ‘life as preferable’ together with ‘life as the good’, and then necessary, because life as preferable is under life as the good, that such a life is in fact available to us for our own living.

Interestingly enough, what Aristotle does in the long passage that follows (1245a1–b19) is these two things. First he explains how the life that is preferable is or can become the good. For life as preferable for us exists in perceiving and knowing, and perceiving and knowing are perfected through sharing with friends; but such sharing, when it is a matter of doing the best things with the best friends, is the human equivalent of the divine life; and, of course, the divine life is the good simply. Hence life as preferable falls under life as the good. But, second, this life must therefore be available to us, because doing the best things with the best friends is available to us, at least if we and our friends pursue and achieve virtue together. The argument is lengthy and involved but it is ultimately clear. It also shows, as was Aristotle's intention from the beginning of the chapter, how the comparison with God is leading us astray about the need of friends. For the good life, which is available to God all by himself, is only available to us through sharing with friends. Hence the self-sufficient man (as opposed to the self-sufficient God) will need friends as constitutive of the happy and self-sufficient life and not as an extrinsic or dispensable extra.

8.1.1246b19–22

Susemihl and the OCT follow Λ¹ and change τῆς μὲν ἐν τῷ λογιστικῷ ἀρετῆς to read τὴν μὲν ἐν τῷ λογιστικῷ ἀρετὴν, thus making it object of στρέψει καὶ ποιήσει ἀγνοεῖν (‘will distort virtue in the calculating part and make it ignorant’), and follow Spengel in changing μοχθηρία ποτὲ ἐγγενομένη μὲν τῷ λόγῳ to read μοχθηρία ποτὲ ἐγγενομένη ἐν τῷ ἀλόγῳ (‘wickedness, when once it comes to be present in the unreasoning part, [will distort …]’). The first change, following Λ¹, seems necessary (if one prefers to regard the phrase as a genitive absolute one must provide a verb for it, as Spengel does by adding ἐνούσης from the lines following). The second change is less compelling. Certainly if ἐγγενομένη is left as it is λόγῳ must be changed to ἀλόγῳ. But perhaps we have here another case of the EE manuscripts misrecognizing letters and misdividing words. For while it seems necessary to change μὲν to ἐν, it is not necessary simply to get rid of the μ. One may instead regard it as a misspelling for ν and add it on to the end of ἐγγενομένη to give ἐγγενομένην. One will thus make ἐγγενομένην agree with ἀρετὴν in the line above, and so obviate the need to change λόγῳ to ἀλόγῳ. The sentence will then read: ‘For it is odd if, on the one hand, wickedness will distort virtue in the calculating part (when it [sc. virtue] at any time comes to be present in reason) and make one ignorant, but …’ The philology is perhaps now more intelligible, but in any event the philosophical point remains the same (it is just expressed from the opposite direction). The apparent redundancy (‘virtue in the calculating part … comes to be present in reason’) is only apparent. The phrase ‘virtue in the calculating part’ refers to the kind of virtue in question (namely intellectual virtue and not moral), and the phrase ‘comes to be present in reason’ refers to this virtue actually being present in the soul. Note too that, if this emendation is accepted, it will no longer be necessary, with the OCT, to add ἐν τῷ λογιστικῷ to ἀγνοίας ἐνούσης in the next line (b21) since ἐν τῷ λόγῳ can be supplied from the ἐγγενομένην ἐν τῷ λόγῳ that precedes.

An alternative suggestion is to regard the μ as a misspelling for ς and read the whole passage thus: ἄτοπον γὰρ εἰ τῆς μὲν ἐν τῷ λογιστικῷ ἀρετῆς μοχθηρία ποτὲ ἐγγενομένης ἐν τῷ λόγῳ στρέψει καὶ ποιήσει ἀγνοεῖν ἡ δ' ἀρετὴ ἐν τῷ ἀλόγῳ ἀγνοίας ἐνούσης οὐ στρέψει ταύτην. We will now have a verb for making τῆς μὲν ἐν τῷ λογιστικῷ ἀρετῆς into a genitive absolute (namely ἐγγενομένης), but this genitive absolute will, oddly, split and be split by the main clause μοχθηρία … στρέψει καὶ ποιήσει ἀγνοεῖν. On the other hand we will also now have a balance with the genitive absolute ἀγνοίας ἐνούσης in the next line, which itself will split the main clause ἡ δ' ἀρετὴ ἐν τῷ ἀλόγῳ … οὐ στρέψει ταύτην. This way of emending the text may not be as happy as the former but its peculiar structure makes it, perhaps, worth noting.

8.2.1247b15–18

Susemihl and the OCT both follow Sylburg and Λ² in emending ἂν εἶεν to ἂν εἴη (‘it [sc. throwing high in dice] would be’), and follow Fritzsche in emending μακρὰν βάλλειν to μακαρίαν βάλλειν (‘… but the way that always throwing successfully in dice would be?’). The first of these emendations seems necessary, and perhaps the second too though it is perhaps not of great moment. In any event the problem words are οὐχ ὅτι τοῖς δεῖ which make no sense. Susemihl follows Sylburg in changing τοῖς to οὕτως (‘not because it must be thus’) and the OCT follows Jackson in changing τοῖς δεῖ to τοιοσδί (‘not because he is such’). In the context Aristotle is raising questions about luck and specifically about the puzzle that, on the one hand, some people seem to be fortunate often or always and yet, on the other hand, luck is not the sort of thing that happens often or always. Hence either these people are not really fortunate by luck (but by some other cause) or, if they are, it should be possible, by luck, to go on endlessly scoring high in games of chance (which we do not see happening). Aristotle eventually solves the puzzle by distinguishing two kinds of fortune: the natural or divine kind which is continuous and is a natural tendency to do the right thing because one's desires are naturally well directed, and the ordinary kind which is simply happenstance and is not continuous. Sylburg's and Jackson's emendations both make the same point (the luck involved in throwing high in dice is not a matter of any necessity nor a matter of having a certain sort of nature), but Jackson's is philologically more likely. However, an equally likely or perhaps more likely emendation is to change τοῖς δεῖ to πως δεῖ (‘not because it somehow must’). The sense produced is the same or similar but there is perhaps more likelihood that τοῖς δεῖ is a misreading of πως δεῖ than of οὕτως δεῖ or τοιοσδί. Also, Λ² has hos oportet and it could conceivably be that what the Latin translator read was τοῦς δεῖ and that hos was the best he could make of τοῦς. For a τοῦς could be misread in place of πως as easily as a τοῖς could.

8.2.1247b28–38

In the context Aristotle is again talking about the two kinds of luck and explaining how each seems to work. The first are fortunate by nature because their natural impulse (but not their choice and calculation) is what guides them and it is right, and the second are fortunate by mere happenstance, for they do follow calculation (and their calculation, even if mistaken, is not silly like that of the first) but, in this case, the happenstance stops them following it and suffering something bad as a result.

A first emendation, then, taken from Λ², is to put ἔσωσεν (‘saved them’) in place of the grammatically dubious ἔξωσεν at the end. The problem words, however, are those that precede it, ὅταν μὲν … αὐτὴ δ' ὀρθὴ οὖσα, since, in particular, there is no finite verb for the ὅταν clause. Susemihl changes ὀρθῶς to ὀρθὸς, deletes the δ' in αὐτὴ δ' ὀρθὴ οὖσα, and, in his apparatus, marks εἶναι, τύχη δ' as corrupt. The OCT also changes ὀρθῶς to ὀρθὸς and deletes the δ', but, in addition, deletes the οὖσα in τύχη δ' αὐτοῦ αἰτία οὖσα and, following Allan, marks as corrupt the whole of καὶ τοὺς μὲν ἐνταῦθα … ὀρθὴ οὖσα ἔσωσεν. An attractive emendation, then, would be to follow Mingay (in the apparatus of the OCT) and, in addition to the other changes made in the OCT, change δοκῶν to δοκῇ to give the sense: ‘whenever calculation does not seem to be right but chance is cause here and it, being right, saved them’. Another is to follow Spengel by changing τύχη to τυχῇ, adding ἡ before the following δ', and punctuating thus: εἶναι τύχῇ, ἡ δ' αὐτοῦ αἰτία (with the ἡ δ' referring back, presumably, to the ὁρμὴ or ὄρεξις in b34 just before) to give the sense: ‘whenever calculation happened not to be right, but impulse (or appetite) being cause here and it…’

The problem, however, with both emendations is in the philosophical sense produced. For the sort of fortune in question here seems to be mere happenstance. The people who benefit from it are not those who are incapable of calculating well and who go right because of some natural impulse, but precisely those who are capable of calculating well and who go right because of their correct calculation. Such people are fortunate by happenstance when events turn out contrary even to correct calculation but some lucky intervention saves them from the consequences. An example might be travellers who have calculated correctly that the best way into town is by a first road but are forced by some chance to take a second and so happily escape the thieves who, unbeknownst to them, are lying in wait on the first. In this case, clearly, the calculation was correct, for it was not foolish or silly (unlike the calculation of those fortunate by natural impulse) but soundly based on the known facts (or as soundly based as human calculation can be). That the results of following it in this case would, nevertheless, have been bad is no fault of the calculation, but of things beyond human knowledge and control.

Suspicion, therefore, should fall on the negative μὴ in ὅταν μὲν λογισμὸς μὴ δοκῶν ὀρθῶς εἶναι. So here is a suggestion. Regard this case as another of EE's typical fault of misreading, and replace μὴ with ἦν (ΜΗ misread for ΗΝ); ἦν with ὅταν will then produce the same construction of a past tense of the indicative with ἂν to express a past condition as just before at 1247b32.Footnote ¹⁴ (Or alternatively, if ἦν with ὅταν is thought to be dubious, one could write ᾖ instead of ἦν and get the necessary sense.) If we also retain the δ' in τύχη δ' αὐτοῦ, we will get the sense: ‘whenever there was a calculation seeming to be right, luck however being here cause and it being right saved them’. This sense is now just what we need for fortune as happenstance (like the case of the lucky travellers): the calculation was right, at least as far as all the known facts were concerned (it ‘seemed’ right), but chance intervened (it became ‘cause’ of what they did), and chance turned out to be really right (it was in accord with the unknown facts), and so saved them.

8.2.1248a34–7

Aristotle is still comparing the two kinds of fortune and he is here talking of those who operate by reason (and do not go right because their desires are naturally right), and he seems to be saying that these, like the others, are in a way operating by divination but by some sort of rational and not non-rational divination (not a divination that is ‘apart from reason’). An obvious and first emendation is to follow Ross and Λ² and replace ἀπολαβεῖν with ὑπολαβεῖν to give the sense: ‘Also, one must suppose that the divination of these prudent and wise is swift and is all but (or: almost) the one which is apart from reason …’ The sense is that those who have reason are also able, through their reason, to divine things, including likely events in the future (as, say, coming changes in weather), and their divining can be almost as quick as those whose divination is not by reason. The main problem with the Greek, however, lies not here but with τε ἐν τῷ σκοπεῖν χρῆσθαι, which is without grammatical sense. Susemihl prints the text without change while the OCT deletes τε, but both record in their apparatus Spengel's suggestion to read τοῦ in its place, with the sense: ‘… through familiarity with its use in speculation’. This suggestion gives some meaning to the phrase in itself, but since Spengel thought there must be a lacuna somewhere in the preceding lines he did not endeavour to show how his emendation gave sense to the phrase in the larger context. A better suggestion, then, comes from von FragsteinFootnote ¹⁵ who suggested replacing τε ἐν with δύνανται (from a34) or the like, with the sense: ‘… through familiarity are able to use speculation’. One problem, however, with this suggestion is that its meaning does not fit very happily with what follows next, which talks about how both divinations (that of those who act with reason and that of those who do not) use God, and not of how one group uses speculation. Nevertheless, von Fragstein does seem right to suppose that a verb is to be supplied from the context. An alternative, then, is to adapt both Spengel and von Fragstein and write τὸ rather than τοῦ in place of τε and supply ἔχουσι (from a32 and a33) instead of δύνανται, thus giving the sense: ‘… through familiarity have the use of it [sc. divination] in their speculation’. The meaning will be that those who act by reason have in their acts of speculation some access to divination through their experience or familiarity with things. A case in point might be Thales divining, through his knowledge of things in the heavens, a coming bumper olive harvest (Pol. 1.11.1259a6–19). At all events the whole sentence now reads with fair intelligibility: ‘Also, one must suppose that the divination of these prudent and wise is swift and is almost the one which is apart from reason; but some have the use of it in their speculation through experience and others through familiarity.’Footnote ¹⁶