Escolar Documentos
Profissional Documentos
Cultura Documentos
A STU D Y O F T H E R E L A T IO N S H IP
BETW EEN TH E E U D E M IA N AND
N IC O M A C H EA N E T H IC S O F A R IS T O T L E
ANTHONY KEN N Y
how to use a computer, and to M rs. Susan Hockey who taught me Barnes, Professor G . E . L . Owen, Richard Sorabji, Pamela H uby,
everything that I know about the use o f the computer for literary and M ichael Woods.
studies. I wish also to thank Professor Thedore Brunner and the Finally I must express my gratitude to four scholars whom I have
staff o f the Thesaurus Linguae Graecae at Irvine, and Stephan never met but to whose writings I owe a great deal. Professor J . D .
Gruen who as a graduate student prepared for me a computer- M onan’s M oral Knowledge and its Methodology in Aristotle first
readable text o f the Eudemian Ethics, on which I carried out my made me aware o f the flimsiness o f the arguments on which the
preliminary studies before the tapes prepared by the Thesaurus accepted chronology o f Aristotle’s ethical writings rested. T h e
Linguae Graecae became available. monumental commentaries o f Gauthier and Jo lif and o f Dirlm eier
D rafts o f chapters o f the book were read as papers to the Bs club are classical expressions o f the theories which are the target o f the
in Cam bridge, to the Philological Society in O xford, to the present work : but I am aware that at every step I draw upon their
Am erican Philological Society in W ashington, and to the M oral erudition and industry for the very information on which my
Sciences Club in Cam bridge; and also to societies and seminars at criticism o f the reigning orthodoxy o f Aristotelian scholarship is
K in g ’s College, London, at the U niversity o f Texas at Austin, at based.
Swarthmore College, at the University o f Pennsylvania, at the I am indebted to the Trustees o f the Craven Fund for a travel
U niversity o f M ichigan, and at Bristol University. I am most grant which enabled me to consult manuscripts o f the Ethics and of
grateful to those who took part in the discussions at those places for Aspasius in Florence and in Rome.
many helpful suggestions and criticisms.
T h e Revd. Andrew Q. M orton o f Culross, and D r. Alan Jones o f
the Oriental Institute in Oxford were both extremely helpful and
encouraging to me when first I became interested in the statistical
study o f classical texts. Professor Charles K ahn o f the University o f
Pennsylvania and an anonymous reader for the Archiv für
Geschichte der Philosophie made searching and valuable criticism
o f an early draft o f the central argument o f the book.
D r. C . J . Rowe, whose study o f the relationship between the
Eudemian and Nicomachean Ethics in 19 7 1 awoke my own interest in
the problem, assisted me, in discussion and in correspondence, at
every stage o f my work— bent as I was on the destruction o f the
theory he had so ably defended— with the most admirable
impartiality, generosity, and friendliness. Professor D . J . A llan—
who has for decades stood out, almost alone in the republic o f
classical letters, against the universal acceptance o f the fashions set
by W erner Jaeger, and has given courage to others less learned than
him self to do likewise— has placed me greatly in his debt by his
constant encouragement and willingness to share with me the
unparalleled store o f learning he has acquired in more than twenty
years’ study o f the Eudemian Ethics. T o these and to the other
Aristotelian scholars who have taken a friendly interest in my work
I am most grateful, in particular to Professor J . L . Ackrill, Jonathan
CONTENTS
L ist o f Tables x
Bibliography 240
Index 247
List o f Tables xi
model o f the M M is the E E : the borrowings from the N E are o f the N E is unparalleled elsewhere in the M M . There is little if
entirely secondary. . . . In our section too the E E is the author’s any verbatim quotation o f the E E : but its ideas and structure are
textbook; he treats the N E simply as a kind o f commentary.’ 1 all-pervasive. I f scholars had not been so obstinately convinced that
T h e verdict o f Festugière seems the only possible one for those the N E is later than the E E , they might have been struck by the fact
who do not regard the M M as being the work o f Aristotle himself. that the Nicomachean cento occurs in the M M at a point in the
In my opinion the arguments o f Jaeger, Walzer, and Brink for the argument where the E E itself, in a mysterious phrase which has
inauthenticity o f the treatise have not been overthrown by the become a notorious crux, makes reference to an appendix.1 I f the
counterattacks o f von Arnim , Dirlm eier, Düring, and Cooper. Eudemian treatment were later than the Nicomachean it might well
Whether the Jaegerians were right to assign the work to a post- have carried with it an appendix o f Nicomachean material : in that
Aristotelian date is another matter, which we shall investigate in the case, the M M by incorporating it at the point suggested by the E E
final chapter o f this book. T h e weakness o f the view that the M M is would be following rather than violating its usual practice. But at
an early work o f Aristotle him self (the view o f von Arnim ’s the present stage o f the argument, that is a piece o f premature
followers, though not o f von Arnim himself) is shown above all by speculation.
the need felt by partisans o f this theory to postulate a double What is not speculation is the general closeness o f the M A I to the
redaction o f the work— a manifestly ad hoc device to escape the E E . T h e Nicomachean cento is one of half a dozen places where the
evidence for a comparatively late terminus post quem given by the M M deviates from the E E as a model, as can be seen from T able i . i .
historical allusions. T h is is a question to which we shall return. For
T able 1.1 Parallels to the M agna Moralia in the N E and E E
the moment there is no need to enter into the question o f the date of
the composition o f the M M : what is significant for our present MM book and chapter Bekker pages* Closest Aristo
purpose is that the author, whoever he was and for whatever telian parallel
purpose he compiled his treatise, regarded the Eudemian Ethics, and
not the Nicomachean, as the appropriate Aristotelian work on which A i a (Introduction) 1 18 1 b I 182” (2) None
to base his own.2 A i b (Idea of Good) Il82b- ■i83b (3) EE
A 2 (Divisions of ‘Good') I 184“ (I) None
T h is is significant because the author undoubtedly knew the
A 3-4 (Happiness) I 184b I I 8 sb (2) EE
Nicomachean Ethics and from time to time makes use of parts o f it.
OC
t:
A 5-9 (Virtue) 1 185”- (4) N E (the Allan
D . J . Allan has recently drawn attention to a cento o f quotations cento)f
from the N E occurring in the fifth to the ninth chapters o f the first EE
i i93b (12)
00
A 10 32 (Will, Virtues)
book, beginning with a citation which is introduced by the words A 33 (Justice) • 193°- -1196" (6) AE
‘this can be seen from the Ethics' (i 1 8 sb1 5).3 As Allan says, this use A 34-B 3 (Wisdom) I iç6b 1199“ (5) AE
1. . . . le modèle ordinaire de la G .M . est VÉth. End. L e s emprunts à VÉth. N ie. sont tous
B 3 (5 Aporiai ) 11991 - I200a (2) None
secondaires. E n ce qui regarde notre section, le phénomène est manifeste et bien B 4-7 (Akrasia, Pleasure) I200b--1 2o6b (12) AE
significatif. . . Y É th.E u d. demeure, ici encore, sonlextbook. E iû n e x o it enVÉth. N ie. qu'une B 7-10 (Topics of EE VIII) i2o6b--I2o8b (4) EE
sorte de commentaire destiné à expliciter, ou, parfois, à corriger le livre qui lui sert de base’ B 11 - 12 (Friendship) i2o8b--1 2 12a (9) EE
(Festugière, 1946, lvi).
2. Theiler, 1934, says that it leaps to the eye that M M is closer t o E E ; to save the Jaegerian B 13 -14 (Self-love) I2 I2a--I2 I2b (2) NE
position he postulates a lost M iddle Ethics {M E ) between E E and NE> from which M M is B 1 5—17 (Friendship) I2 I2b--I2I3b (2) EE
derived. Among those who deny the authenticity o f the M M opinions as to its date vary from
the Theophrastean peripatos to the early first century b.c.: there is a neat summary of * T h e number in brackets is the approximate number o f Bekker columns im olved.
positions in Allan, 1957. f E\en within this section, there are passages where the ΛίΛί is dependent rather on the
3. Allan, 19 57, 1966. M ost scholars have refused to see these words as a citation, τ ο υ τ ' E E : e.g. 1 i86a2 0 -b4.
iS tt v €Ot îv €κ τ ω ν ’ Η θ ι κ ώ ν is an unusual way o f quoting, but the text is most naturally
interpreted in A llan’s way, as witness the scholars who have felt obliged to emend the text to
avoid having so to interpret it (e.g. Cooper, 1973). 1. iv τ ο ΐς ά π η λ λ α γ μ έ ν ο ίς , i220bi 1. See D irlm eier, 1962; Allan, 1966.
12 The Aristotelian Ethics The Aristotelian Ethics in Antiquity !3
It will be seen that the M M is closer to the E E for approximately even what is essential, while a third chooses the right time to say just what
he ought. This mean is a mean with regard to us, because it is fixed by us
thirty-two Bekker columns, and closer to the N E for only six.
by reasoning. Thus virtue is a disposition for choice placed in a mean with
Spengel summed up the situation thus:
respect to us, determined by reasoning and the way that the wise man
At the beginning the M M looks back to both, and contributes a certain would determine it. Then, setting out a number of triads in the footsteps of
amount of its own material. But from the tenth chapter of the first book his master, examining each in turn he endeavoured to complete an
until the thirty-fourth there is a manifest attachment to the EE, although induction as follows : for the sake of example the following were chosen :
here too individual expressions can be found which can only have been temperance, intemperance, insensibility; gentleness, irascibility, in
taken from the NE. . . . From A 34 until B 7 the M M follows the N E, and sensitivity; courage, rashness, cowardice; justice; liberality, prodigality,
immediately afterwards goes back to the E E (1841, 515). meanness ; magnanimity, pusillanimity, vanity ; magnificence, shabbiness,
extravagance. Of these dispositions some are bad because they are
B y saying that from A 34 until B 7 the M M follows the N E and then excessive or deficient with regard to the passions ; others are good, clearly
returns to the E E what o f course Spengel really meant was that from because they are means.1
A 34 until B 7 the M M followed the common books o f the A E : he T h e passage then proceeds to a consideration of each o f the triads in
took it for granted that the disputed books belonged to the N E turn. Scholars disagree about the extent and the degree of
because he regarded the E E as spurious. But surely it is on the face
literalness o f the quotation from Theophrastus. Von Arnim , who
of it unlikely that the author of the M M , having followed the
first focused the attention o f the learned world on the importance o f
structure o f t h e £ £ faithfully from 118 7 “ to U 9 3b, should switch for
the passage, regarded it as an extensive, verbatim quotation. He
twenty-five Bekker columns to following the N E , and then switch
pointed out (1924, 12 4 -4 1) that the table o f virtues and vices given
back, by an extraordinary coincidence, just when the disputed
here corresponds in several respects to the one given in the E E
books end, to follow once again the lead o f the E E . I f the disputed
(i2 2 0 b38“ i 2 2 i ai2) rather than to the one mentioned in the N E
books are taken as belonging with the E E , then everything runs
(1107*34) so far as the latter can be reconstructed from the text.
more smoothly : for twenty-four Bekker pages the M M lets itself be
(For example : the two lists are introduced by the same phrase ; they
guided, with minor variations in order and occasional small
both give the defect corresponding to gentleness as insensitivity
interruptions, by the E E . The structure o f the M M is itself, then,
rather than spiritlessness; both mention justice with the major
substantial evidence o f a Eudemian origin for the disputed books.
moral virtues.)2 Som e details o f the subsequent development o f the
And its over-all fondness for the E E , despite its acquaintance with
discussion, likewise, recall the Eudemian discussion o f the table:
parts at least o f the N E , is a striking indication o f the relative
standing o f the two treatises in peripatetic circles at whatever time it
1. T o ovv π ρ ο ς η μ ά ς μ έ σ ο ν α ρ ισ τ ο ν , ο ίο ν , φ η σ ιν ό Θ ε ό φ ρ α σ τ ο ς, ε ν τ α ΐς ε ν τ ν χ ια ις
was written. T h e whole topic will occupy us further in our final όδι μ ε ν π ο λ λ ά δ ιε λ θ ώ ν κ α ι μ α κ ρ ώ ς ά δ ο λ ε σ χ ή σ α ς , ό δ ι δ ’ ο λ ίγ α κ α ι ο ύ δ ε τ ά ν α γ κ α ΐα ,
chapter. ο ΰ τ ο ζ Ôè α υ τ ά α ε δ ε ι μ ό ν α το ν κ α ιρ ό ν ε λ α β ε ν . Λ ύ τ η ( η ) μ ε σ ό τ η ς π ρ ο ς η μ ά ς
( ά ρ ίσ τ η ) , α ν τ η γ α ρ ύφ ' η μ ώ ν ώ ρ ισ τ α ι τ ώ λ ο γ ω . Δ ιο ε σ τ ιν η α ρ ε τ ή ‘ 'έξις
π ρ ο α ιρ ε τ ικ ή , ε ν μ ε σ ό τ η τ ι ο ν σ α τ η π ρ ο ς η μ ά ς , ώ ρ ισ μ ε ν η λ ο γ ω κ α ι ώ ς αν ό φ ρ ό ν ιμ ο ς
ό ρ ίσ ε ιε ν \ ε ΐτ α π α ρ α θ ε μ ε ν ο ς τ ιν α ς σ υ ζ υ γ ία ς , α κο λο υθ ώ ? τ ώ ύ φ η γ η τ η σ κ ο π ώ ν
TH EO PH RA STU S ε π ε ιτ α κ α θ ' ε κ α σ τ α ε π ά γ ε ιν ε π ε ιρ ά θ η τ ο ν τ ρ ό π ο ν τ ο ύ τ ο ν · ε λ ή φ θ η σ α ν 8 ε π α ρ α δ ε ι
γ μ ά τ ω ν χ ά ρ ιν α ΐδ ε · σ ω φ ρ ο σ ύ ν η , α κ ο λ α σ ία , α ν α ισ θ η σ ία ■ π ρ α ο τ η ς , ό ρ γ ιλ ό τ η ς ,
T here have survived two quotations o f the Ethics by α ν α λ γ η σ ία · α ν δ ρ ε ία , θ ρ α σ ύ τ η ς , δ ε ιλ ία - δ ικ α ιο σ ύ ν η * * * ε λ ε υ θ ε ρ ω τ ή ς , ά σ ω τ ία ,
Theophrastus. One occurs in a passage o f Stobaeus’ anthology α ν ε λ ε υ θ ε ρ ία · μ ε γ α λ ο ψ υ χ ία , μ ικ ρ ο φ υ χ ία , χ α υ ν ό τ η ς · μ ε γ α λ ο π ρ ε π ε ια , μ ικ ρ ο π ρ έ π ε ια ,
σ α λ α κ ω ν ία . Τ ο ύ τ ω ν δη τ ώ ν ε ξ ε ω ν α ί μ ε ν τ ώ ύ π ε ρ β ά λ λ ε ιν η ε λ λ ε ίπ ε ιν π ε ρ ι π α θ η
taken by scholars as part o f Arius D idym us’ sum mary of φ α ύ λ α ί ε ίσ ιν , α ί δ« σ π ο ν δ αΓαι, τ ώ μ ε σ ό τ η τ ε ς ε ίν α ι δ η λ ο ν ό τ ι (Wachsmuth, ii. 140 6 ft.).
peripatetic thought. 2. [T h eop h rastus?]: ελ ή φ θ η σ α ν δ ε π α ρ α δ ε ιγ μ ά τ ω ν χ ά ρ ιν α ΐδ ε. Ε Ε : ε ίλ ή φ θ ω δη
π α ρ α δ ε ίγ μ α τ ο ς χ ά ρ ιν . E E and Theophrastus have α ν α λ γ η σ ία where N E has ά ο ρ γ η σ ία .
The mean with regard to us is best; as, for instance, Theophrastus says, Som e o f von A rnim ’s conclusions perhaps need qualification. On the order o f the list, sec
below, p. 23.
one man spends a lot of time in idle chatter, another says very little and not
H The Aristotelian Ethics The Aristotelian Ethics in Antiquity 15
for instance, the comparison o f the insensible man to a stone (Stob. can be traced to the E E itself; commentaries would be superfluous
Ü .I4I.7, E E I2 2 Ia23).' if nothing was contained in a commentary on a text except material
It is difficult to settle with any certainty the boundaries between from the text.
the nested quotations from Arius, Theophrastus, and Aristotle. Aspasius’ commentary on Aristotle’s Ethics ( C A G x ix .15 6 .17 )
Von Arnim thought all the passage set out above was from preserves another fragment in which Theophrastus made a literal
Theophrastus except the words ‘Theophrastus says’ and ‘Then, quotation from Aristotle’s Ethics : this is the sentence ‘ any pleasure,
setting out a number o f triads . . . as follows’ . T h is would include in whether it happens to be the contrary o f the pain involved or not,
the Theophrastean passage a verbatim quotation of the definition o f has the effect, provided it be strong enough, of driving out that
virtue from the N E ( 1 io 7 ai), and the words ‘for the sake o f example pain.’ T h is comes from the section o f A E C devoted to pleasure.
the following were chosen’ , which would thus lead into a verbatim T h e anonymous scholiast on A E A tells us that the quotation from
quotation by Theophrastus from Aristotle. M ore recently M oraux Theognis occurring at A E A, ii2()b30 is used also twice by
has argued convincingly that the Nicomachean quotation, impor Theophrastus, though attributed by him to Phocylides.1
tant features o f which are irrelevant to the context, cannot be part o f
the Theophrastean passage; he is inclined to think that the ( : I ( I RO·’
quotation from Theophrastus is interrupted by a number o f
remarks by Areios (who is certainly the spokesman in the sentence A passage in De Finibus V .12 ( r . 45 b . c .) is often quoted as the
about setting out triads in the footsteps o f his master) and is not earliest evidence for the title o f the Nicomachean Ethics. Cicero
resumed until the triads themselves are set out.2 T h e most natural 1. On Theophrastus' use o f the E ttuts, see also Regenbogen, R J: Suppl. vii, 1354 ft.;
way o f reading the passage is to see the quotation as being resumed W chrli, 1972, 4<)o ff.
2. T h e reader may be surprised to find no section dexoted to Kpicurus' use o f Aristotle.
immediately after ‘as follows’ (ton tropon tout on ) ; it is then difficult T h e reason for the omission is that I am uneon\inced that we have any evidence that
not to agree with von Arnim that the clause ‘for the sake of example Epicurus knew any o f Aristotle’s sur\ i\ ing ethical treatises at first hand : the negativ e, if not
the following were chosen’ is a quotation by Theophrastus o f the the positive, side o f the thesis o f Bignone ( 11)36) seems to me to have survived the criticism o f
later scholars. Certainly, Kpicurus during his Athenian studies had the opportunity to
familiar E E passage beginning ‘ for the sake o f example let there be
become familiar with Aristotle’s doctrines, and two of his major interests— pleasure and
chosen . . .’ and continuing with the list o f triads. Having set forth the relation between freedom and necessity- are discussed at length in the N E and the h h ;
the E E text, Theophrastus then goes on to comment on it : there is but nothing survives to show first-hand knowledge o f the relevant texts. Merlan ( tyfto, 1 37)
has made an interesting comparison between the two treatments o f pleasure in the N E and
no reason to be surprised that not everything in the commentary
Λ -Eand the theory o f K picurus; but he attempts no textual rapprochement. Furley ( 1 <jf>7, l*)l
1. C f. Von der M ühl], 1909, 28. Nor everything in the discussion o f the table can be traced 256) compares the treatment o f voluntary action in Aristotle and Kpicurus. H e begins with
to the E E : von Arnim concludes that it was the M M Theophrastus had before him. B ut the the candid admission: ‘ I have not made a systematic attempt to prove that Kpicurus could
M M neither contains nor mentions a table o f virtues, and von Arnim is obliged to suppose a have read or did read Aristotle’s school treatises’ ; he simply assumes in Kpicurus familiarity
lacuna at the crucial point. M ansion’s judgement is surely sound : ‘ Pour affirmer qu’ il dépend with the N E . But if one is to assume without proof acquaintance with Aristotle's Iùliii s, the
du passage correspondant du M M que nous ne possédons plus, on en est réduit en somme à E E is at least as likely, on the basis o f Fu rle y ’s comparison, to have been Kpicurus'
se baser sur le contenu conjectural d ’un texte perdu. C ’est bien peu de chose’ (19 27, 449). Aristotelian source. T h e closest textual comparison he offers is between N E 3.5, 1 1 1 3b 17 tt.
2. M oraux 19 7 3 ,3 8 3 ff. M oraux concludes that the quotation must be very free: ‘ Kr zitiert and Kpicurus referring to voluntary agents as internally caused agents —ίχ ο ν τ α ς κ α ι ei·
nicht wörtlich, sondern gibt in grossen Zügen das Wesentliche aus Theophrasts έ α υ τ ο ίς τ η ν α ιτ ία ν (On Nature 3 1 .27.3 <), Arringhetti). Kqually good parallels would be E E
Ausführungen wieder. W enn es sich so verhält, so verlieren die gelegentlichen Berührungen 11.6 , 1222’’ 15 2 3a20, and S, t224b2 29. On voluntariness Furley writes: ‘T h e relation
mit E E oder M M beträchtlich an Bedeutung, denn es ist nicht sicher, ob sie bereits in between Aristotle’s two Ethics on this problem is particularly interesting but a thorough
Theophrasts O riginalschrift gestanden haben oder von deren Epitom ator herrühren’ (390). I examination o f it would take too long now. M y own view is that the treatment in the
do not see why A reios' method o f quoting Theophrastus should differ from his method of Eudemian Ethics is consistent with its being intermediate between Plato’s views and that of
quoting Aristotle when he does so at first hand : i.e. not verbatim , throughout, but preserving the Nicomachean E thu s, and that the Nicomachean Etlncs (including the controversial books
many o f the actual turns o f phrase o f the original. I f the quotation o f Theophrastus is o f this K H) rather than the Eudemian Ethics is likely to have influenced Kpicurus' (p. 225). It is
sort, then quite enough turns o f phrase and features o f structure are preserved to make it clear unfortunate that Furley did not have time to give reasons for this judgement: in fact he goes
that what Theophrastus was commenting on was the table o f virtues in the E E . I f von Arnim on to assert that on the principal point -offering a positive rather than a negative criterion ot
exaggerates the amount o f information to be obtained from this text, M oraux undervalues it voluntariness— the E E is more explicit. Rist ( 1970, 101 - 2) argues that the distinctions drawn
to an equal degree. in A E C 11 15 ‘form at least part o f the background’ for Kpicurus’ distinctions between
ι6 The Aristotelian Ethics The Aristotelian Ethics in Antiquity 17
complains that Theophrastus in his treatise on the happy life made o f a Sardanapallus to that o f a bu ll,1 which occurs in E E (1.5,
happiness too dependent on good fortune. ‘L e t us therefore follow I 2 i6 ai- i6 ) and not in N E . But he does so in a way which makes it
Aristotle and his son Nicomachus, whose carefully written books on probable that he is drawing on a different source, which may or may
ethics are said to be the work o f Aristotle, though I do not see why not have been, as Jaeger conjectured, the Protrepticus.2
the son could not have been similar to the father.’ 1 It is commonly In De Fmihus 11.6 Cicero attributes to Aristotle a definition of
taken that Cicero is referring here exclusively to the Nicomachean happiness as the exercise o f virtue accompanied by the prosperity of
Ethics : his attribution o f the work to Nicom achus is taken not very a complete life. T h is definition is perhaps closer to the one given in
seriously, as a mere conjecture based on the title. But the natural the Magna M oralia than to anything in either o f the authentic
way to read the passage is that Cicero is here referring to two Ethics.3
authorities : first, an unnamed but undisputed work o f Aristotle’s, Cicero, then, is a witness that in his time the N E did not yet enjoy
and secondly an ethics with the title ‘Nicomachean’ which is that primacy as the canonical expression o f Aristotle’s ethical
attributed to the father but which Cicero is prepared to attribute to thought which it held for later generations : but it is unclear whether
the son.2 T h e important point is that the ethical work which Cicero his own knowledge o f Aristotle’s ethical teaching, in so far as it
regards as undoubtedly Aristotle’s is not the Nicomachean Ethics.3 depends at all on first-hand acquaintance with his writings, derives
G iven the authority which Theophrastus and the author o f the from the E E , the M M , or the exoteric writings.4
Magna M oralia attach to the Eudemian Ethics, it is tempting to see
here a reference to that treatise; but we cannot be certain that T H E E D IT IO N O F A N D R O N IC U S
model. But it provides no grounds for believing that every reference though not the work o f Nicolaus this central section may well be the
to peripatetic philosophy in Stobaeus is a direct quotation from oldest surviving commentary on the Aristotelian Ethics}
Arius. T h e passage which refers to N E 10 is preceded by a It is regrettable that the date o f the treatise is so uncertain
discussion o f Aristotle’s definition o f happiness (5 0 .11 - 5 1 .1 7 ) and because in the context o f the relationship between the N E and the
o f possible goals ( 5 1.18 - 5 2 .9 ) ; a passage marked by an obsession E E its interest is considerable. A large part o f the discussion is
with triads, which in part duplicates and in part conflicts with the devoted to the ethical virtues, and both in general structure and in
treatment o f Aristotle’s views on happiness presented in the detail the treatment follows the E E and the M M more closely than
epitome itself.1 T h e explicit quotation of, and reference to, book 10 it follows the N E , although from time to time Nicomachean
o f the N E is totally without parallel in the epitome and is quite material is utilized.
unlike Arius’ manner o f utilizing his texts.2 Altogether, it is much T h e order in which the major moral virtues are treated is closer to
more likely that the reference to the Eudoxus passage is by Stobaeus the E E and M M than to the N E . (No great importance can be
himself, and dates therefore from the fifth rather than the first attached to the order o f treatment : but it may be o f interest to set
century a . d .
T able 1 .2 The Ordering o f the Virtues m Peripatetic Ethical Treatises
Unless the treatise has been very heavily interpolated it cannot in Candour 8 9 8 5 I I 12 — 12 9
Wittiness 9 10 — 12 10 — 10 12
fact be the work o f Nicolaus, because it refers to Plotinus and
Friendliness ίο (8)* 9 6 9 II — [ I 7
mentions other proper names from late antiquity. However, the
Dignity — — 10 7 10 8 — 8 8
central section o f the treatise, which breaks the continuity o f the Shame I I I I — —
3 8 9 9 I I
work as a whole, may well be o f earlier date than the surrounding Nemesis 12 10
5 5 J3 7 7 — 7
material: it is a summary o f Aristotelian ethics which follows Justice 12
13 6 4 ‘ 3 >3 4 l3 r3
sources in Aristotle very closely. T h e only indication o f its date is a Hardiness ---- — I I 8 — — — —
critical reference to the Christians. Scholars have suggested that Wisdom --- — H 12 — — — — —
* H ere this virtue is unnamed.
1. Several o f these divergencies are noted by M oraux (19 73, 274) who concludes ‘D aran
geht deutlich hervor, dass Areios aus zwei verschieden Quellen, die eine jeweils anderer i. See M oraux, 1973, 442. Partly because o f the anti-Christian remarks Lyon s suggests
Deutung desselben D efinition bote, geschöpft hat.’ It is surely Stobaeus, not A rius, who is that the Nicolaus o f the treatise may be Nicolaus o f Laodicea, a contemporary o f Julian the
using two incompatible sources. Apostate. B ut there is nothing in the central section which is incompatible with a
2. T h is too is noticed by M oraux. From A rius’ general method o f using his sources, considerably earlier date. It is known that Nicolaus o f Damascus did write an introduction to
M oraux concludes that he may well not have read them at first hand; it is the more the Ethics: it is quoted by the twelfth-century Arabic philosopher Ibn A l-M atran (Peters,
astonishing that he should ascribe this precisely located reference to A rius rather than Aristotle and the Arabs, 270). T he surviving genuine works o f Nicolaus make use o f
Stobaeus. (Compare M oraux, 19 73, 3 13 and 436.) Aristotelian ethical ideas (e.g. that man is an ά ρ χ ή π ρ ά ξ€ ω ς ) but not in a manner which
3. On the Fez manuscript see A rberry, 19 5 5 ; L yon s, Oriens, 1 3 .5 5 ; Paret, 1959/60. enables one to tell which treatise is being drawn upon. See Drossaart Lulofs, 1969.
24 The Aristotelian Ethics The Aristotelian Ethics in Antiquity 25
out the ordering o f the virtues in the treatises we have so far freedom with which he applies it— e.g. to the triad im piety-piety-
considered (Table 1.2).) M ore important, the inclusions and superstition (Immut. xxxiv. 16 3)— leaves it open that he is improvis
exclusions in the list agree with the E E and M M against the N E . ing upon a theme known to him at second hand. Philo divides
T h e principal differences between the treatment o f the virtues in virtues into those that are inborn, those that are acquired by
N E 3 -4 and that in E E I I I is that the N E contains seven major training, and those that come by treaching, in a way which has
virtues and three minor ones (the distinction between the two is reminded some scholars o f N E 10, i i7 9 b2o; but it is equally
drawn at 1 io8a9) whereas the E E contains six virtues and six ‘means reminiscent o f Plato’s Meno (70ai ff.) as is his frequent use of the
o f passion’ (the distinction between the two being drawn at concept o f divine destiny.1 Philo frequently uses the threefold
i2 3 4 a24~34). T h e six virtues o f the E E are the seven virtues listed division o f goods into external goods, goods o f the body and goods
as major in the N E , less the unnamed one which is a proper degree o f the soul : this division is to be found in N E 1, i098bi2 and other
o f ambition ; the six passional means o f the E E are the three minor Aristotelian contexts, but it is also very frequent in Plato.2 Philo
virtues o f the N E , plus shame, nemesis, and dignity. (The first two o f defines happiness as the use o f complete virtue in a complete life:
these are mentioned in the N E in the preliminary discussion in book this is close to the E E definition, but closer to the M M and closer
2 o f means o f passion, but only one, shame, is actually discussed in still to the epitome o f Areios D idym us.3 Like the A E , Philo regards
the full treatment in book 4 ; dignity is not mentioned in the N E at happiness as a combination o f learning and wisdom ; but unlike the
all.) In these respects the Fez manuscript, like the M M and Areios, A E he regards learning as being concerned with the service o f God
follows the E E and not the N E : it does not include the unnamed (Praem. xiv.81). T h e service o f G od, he says, in a phrase which
mean o f ambition and therefore has six and not seven major virtues ; echoes the final section o f the E E , is the beginning and end of
it includes dignity in its list, and it has a discussion o f nemesis or happiness: G od is the standard (horos) o f happiness (De Spec. Leg.,
righteous indignation which is based on the E E text. T o be sure, in ed. Cohn, v.594). Philo’s use o f Aristotelian ethical material is too
these respects the Fez manuscript resembles the M M also; but a free for us to be able to say that he actually quotes the ethical
number o f details indicate direct dependence on the E E : for treatises, or to decide with confidence which of them he is using: we
instance, the naming o f the gourmand who wanted a crane’s gullet can only say he is in general slightly closer to the E E than to the
as the son o f Eryxis (cf. E E III , i 2 3 i ai8). In the discussions o f the N E .A Certainly there is no passage which conclusively proves first
particular virtues the text is sometimes closer to the Eudemian hand knowledge o f any Aristotelian ethical treatise.5
version (as in accounts o f courage1 and temperance) and
sometimes closer to the Nicomachean (e.g. in the distinction δ α π α ν η ρ ία , σ ω φ ρ ο σ ύ ν η --s e e Aitgr. xxvi.147 and Immul. xxiv.16 3 4) φ ρ ό νη σ ές and
between possession and use in the treatment o f liberality, and in the π α ν ο υ ρ γ ία occur only in the Eudemian treatment.
attribution o f wittiness to Socrates as at N E i i 2 7 b5). 1. See e.g. Somn. i.x xv ii.16 4 ; on θ ε ία μ ο ίρ α see Wolfson, 1962, 165 ft.
2. T he relevant texts in Plato are collected in Gauthier and Jo lif, 1959, 6 2 ; for Philo see
Cohn’s edition i.259, v.325, etc.
P H IL O O F A L E X A N D R IA
3. Philo, ed. Cohn, i.2 7 2 ; see p. 17 above. Philo regards happiness as a combination of
Philo (d. A .D . 45) shows acquaintance with a number o f practical and theoretical life, like E E and the Politics, but he uses this division in a way
somewhat different from Aristotle (Praem. 2 .11) .
Aristotelian ethical themes, such as the doctrine of the mean and the 4. Wolfson, 1962, 165 ff., collects together all the texts in which he thinks Philo is
peripatetic definition o f happiness. T h e way in which Philo applies following Aristotle: none o f them amounts to a verbatim quotation.
the doctrine o f the mean to particular virtues makes it more likely, if 5. T h e same observation holds for Seneca, some years later than Philo. Som e scholars have
thought that the clementia o f Seneca is a Romanization o f Aristotle’s ε π ιε ίκ ε ια . G riffin
he is drawing on Aristotle directly, that he was acquainted with the (1976, 159 ff.) has shown this opinion to be without foundation, and M urray, who once
Eudem ian than the Nicomachean version o f the doctrine;2 but the (1965, 1 7 6 ff.) discussed whether the Aristotelian theory o f equity gained widespread
acceptance towards the end o f the Roman republic, has decided that it did not (1967, 355 f.).
1. Especially in the accounts o f the five types o f unreal courage. Even i f Seneca was acquainted with Aristotle’s notion o f equity, he can have derived his
2. W hile most o f the virtues and vices to which Philo applies the doctrine o f the mean knowledge o f it from a reading o f the Rhetoric (1374*26 ff.), a work with which the De Ira
occur in both the E E and N E lists (e.g. α λ α ζ ο ν ε ία , θ ρ α σ ύ τ η ς , δ ε ιλ ία , α ν δ ρ εία , suggests familiarity.
20 The Aristotelian Ethics The Aristotelian Ethics in Antiquity 27
PLU TARCH Favorinus can be traced in both Ethics, in the epigrammatic form
quoted it occurs only in the E E ( V I I .12 , i24 5b2 i ) .1 T h is is made
A moralist as voluminous as Plutarch might be expected to have explicit in the attached comment, which must be a reference to the
had ample occasion to quote from Aristotle’s ethical writings. A Eudemian version since the corresponding passage in the N E
glance at Helmbold and O ’N eill’s index Plutarch's Quotations seems ( 1 1 7 i a1 5—17) occurs in the ninth book.2 T h e comment is evidence
to confirm this expectation, for it lists seventeen reminiscences or also that the version o f the E E referred to must have contained the
quotations to the N E and three to the E E . Inspection o f the three disputed books o f the A E ; otherwise the book on friendship would
E E passages mentioned show that the corresponding passages in be the fourth and not the seventh book.3 It is striking that the E E is
Plutarch are simply common quotations o f third parties (Herodotus cited so firmly as the Ethics o f Aristotle.
and Bion) and so not any evidence of knowledge of the E E . This It would be pleasant to know for certain who was responsible for
fact, plus the prima facie striking fact that not one o f the seventeen the citation. It is possible, though hardly likely, that it was attached
passages o f the N E listed in the index is from the A E , might seem to the epigram by Favorinus. It may be by Diogenes himself, or by
to lend support to the hypothesis that the N E did not originally some earlier writer from whom he has drawn the collection of
contain them: for the hypothesis would explain why Plutarch, Aristotelian sayings. Other scholars have thought that the citation
apparently a keen jV£-quoter and a n on-££-quoter, makes no use was attached to the epigram before Favorinus and originated in
o f them. But, alas, close inspection shows that o f the seventeen some source o f his. Once the edition o f Andronicus had established
‘ reminiscences’ o f the N E all except one are souvenirs so faint as to the Nicomachean Ethics in a position o f pre-eminence, argued von
amount to no more than the common employment o f a com der Miihll (1909, 26), no one could refer to the E E as ‘the Ethics'·,
monplace. T h e single passage in which Aristotle is explicitly quoted therefore the quotation must be by some writer o f the Ptolemaic era
is at 704 E , where Plutarch criticizes him for saying that there is no who included a list o f apophthegms in a life o f Aristotle.4 D üring
incontinence in respect o f the pleasures o f sight and hearing on the went so far as to identify this ancient source as H erm ippus.5
grounds that these are peculiar to humans.1 T h e reference could be In fact, as we have seen, there is no reason to believe that
either to the N E or the E E ; if it has to be to one or other, the E E is Andronicus’ edition even included the Nicomachean Ethics, let
the more likely source because o f details o f language. But more alone that it conferred on it a pre-eminent status. T h e allusion to
likely than either as a source for Plutarch is the passage o f the the E E as Aristotle’s Ethics is consistent with all the ancient
Problems considered earlier.2 Only there is the question o f the allusions we have so far considered. T o whatever date prior to
sharing o f pleasures with animals brought into conjunction with the
1. ο ύ θ ε ίς φ ίλ ο ς ω π ο λ λ ο ί φ ίλ ο ι.
topic o f incontinence rather than intemperance. 2. ο ι Sè π ο λ ν φ ιλ ο ι κ α ι π ά σ ιν ο ικ ε ίω ς έ ν τ ν γ χ ά ν ο ν τ ε ς ούδξν'ι δ ο κ ο ν σ ιν e îv a i φ ίλ ο ι,
π λ ή ν ττο λ ιτικ ώ ς .
F A V O R IN U S O F A R L E S 3· Consequently those who have, implausibly, maintained that the E E is complete in five
books have had to explain away this passage in D iogenes as a late gloss (e.g. T itze : see Rowe,
In the Life o f Aristotle by Diogenes Laertius there is a list o f 19 7 1, 80).
Aristotle’s sayings which concludes thus: ‘Favorinus in the second 4* ‘U t Diogenem vel Favorinum , quem Diogenes excerpsit, Eudem ia perlegisse putem,
vix mihi persuadebis, praesertim cum editione Androniaca pervulgata Nicomachea talia
book o f his Memorabilia mentions as one o f his habitual sayings “ H e dignitati Eudemiis praecellerent, ut Ëthikôn vocabulo semper ad Nicomachea usus
who has many friends has no friend” . T h is can be found also in the scriptorum revocaret. Itaque laudationem illam neque Diogenis neque Favorini sapientiae
seventh book o f the Ethics.’ 3 Though the sentiment quoted by thesauros suppeditasse apparet, sed Ptolemaeorum saeculo a viro docto in vita Aristotelis
cum apophthegmate collocata est, unde Favorinus eam recepit.’
1. 8o k€î 0€ μ ο ι μ η δ ' 'Α ρ ισ τ ο τ έ λ η ς α ιτ ία δ ίκ α ια τ ά ς π €p i θ έ α ν κ α ι α κ ρ ό α σ ιν 5- H is comment on the passage is puzzling, seeming to support two inconsistent
€ ν π α θ € ΐα ς ά π ο λ ν ε ιν ά κ ρ α σ ία ς , ώ ς μ ό ν α ς ά ν θ ρ ω π ίν α ς ο ν σ α ς, τ α ΐς δ ’ άλΛ αι? κ α ι τα interpretations: ‘T h en follows . . . a note from Phavorinus, including an annotation from
θ η ρ ία φ ύ σ ιν ε χ ο ν τ α χ ρ ή σ θ α ι κ α ι κ ο ιν ω ν ε ΐν . D iogenes him self “ this is found in the seventh book o f the Ethics, too” . T h is note proves that
2. ρ. 8. H erm ippus (or the unknown author) had access to or knew an edition o f the Eudemian Ethics
φ η σ ι δ ε Φ α β ω ρ ιν ο ς i v τ ω δ ε υ τ ε ρ ω τ ω ν ''Α π ο μ ν η μ ο ν ε υ μ ά τ ω ν ώ ? έ κ ά σ τ ο τ ε which included the books which it has in common with the Nicomachean E th ics' I f the note
λ έ γ ο ι, ‘ ω φ ίλ ο ι, ο ύ δ ε ίς φ ίλ ο ς \ άλλα κ α ι ev τ ω ε β δ ό μ ω τ ω ν ’Η θ ι κ ώ ν ε σ τ ι. is by Diogenes, it surely proves nothing about H erm ippus’ library (Düring, 1957, 67).
28 The Aristotelian Ethics The Aristotelian Ethics in Antiquity 29
Diogenes we assign the present passage, we have further evidence T w o other points elsewhere in Diogenes are more indicative.
for the ancient preference of the E E over the N E . Among Aristotle’s sayings he gives the definition o f friendship as a
single soul dwelling in two bodies.1 L ik e the saying in Favorinus,
D IO G E N E S L A E R T IU S this too seems to be an allusion to the E E . T h e N E does not figure,
in the same way, in the Life o f Aristotle. Indeed, Diogenes, like
In addition to the apophthegm from Favorinus the L ife o f Cicero, seems to have believed that the Nicomachean Ethics was the
Aristotle contains other material relevant to our problem. Diogenes’ work o f Aristotle’s son : for in his Life o f Eudoxus (viii.88) we read
summary o f the philosopher’s doctrines contains a section ( v .3 1- ‘Nicomachus, the son o f Aristotle, states that he declared pleasure
52) on his ethical teaching. P. M oraux (1949) made a detailed study to be the good’ ; 2 a reference to N E 10.2, i i 7 2 b9 ff.
o f this summary, along with the rest o f the outline in Diogenes, with Diogenes’ most valuable service is his preservation o f the ancient
a view to establishing its immediate sources. After comparing it list o f Aristotle’s writings. This, which raises more problems about
with the Aristotelian texts he came to the conclusion that it must the Ethics than it clearly solves, will be considered later on its own.
have been written by an unknown author between the time of
Andronicus and that o f Diogenes. But there is no reason to believe
A S P A S IU S
that in constructing his ethical summary Diogenes looked anywhere
else than to the epitome o f Arius D idym us. In structure and in From the study we have so far undertaken o f twelve ancient
detail o f vocabulary Diogenes’ summary follows A riu s1 and in its writers, the following conclusions emerge. Every writer, from
whole length there is nothing which demands any different source Aristotle him self until the second century A .D ., who shows a first
except for three insignificant details. These details will not help us hand knowledge o f the Aristotelian ethical writings also shows a
to decide whether Diogenes (or any source he had other than preference for the E E over the N E : either in the sense that he refers
Arius) followed by preference the N E or the E E : for they are to it as The Ethics, sans phrase, or that he quotes it exclusively, or
details which occur in none of the treatises o f the Aristotelian that he prefers its doctrine, or its terminology, or its systematic
corpus.2 structure. N o author (with the exception o f the author of the Magna
M oraha) quotes the Nicomachean Ethics as ‘T h e Ethics’ ; and the
only two authors who show awareness o f an Ethics with that title
i. I will illustrate this from D iogenes’ first sentence:
both regard it as the work o f Aristotle’s son.
Diogenes Arius
τ έ λ ο ς δε εν ε ζ ε θ ε τ ο χ ρ ή σ ιv α ρ ε τ ή ς εν ε υ δ α ιμ ο ν ία ν δ 'ε ν α ι χ ρ ή σ ιν ά ρ ε τ ή ς τ ε λ ε ία ς
T h e comparative silence of antiquity about the N E comes to an
β ία ) τελ εία ), εφ η δ ε κ α ί τ ή ν ε υ δ α εν β ίω τ ε λ ε ιω (130 , 19)· end with Aspasius, the author o f a line-by-line commentary on the
ιμ ο ν ία ν σ υ μ π λ ή ρ ω μ α ε κ τ ρ ιώ ν α γ α It is a whole whose parts are σ υ μ π λ η ρ ω τ ικ ά Ethics which is the earliest o f all the surviving Aristotelian
θ ώ ν εΐνα ι- τ ώ ν π ε ρ ί ψ υ χ ή ν , ά δ ή κ α ί (U0-9)· commentaries.3 In Aspasius’ writing we find the situation with
π ρ ώ τ α τ ή δ υ ν ά μ ε ι κ α λ ε ι· ε κ δεύ ε τ ι τ ώ ν ά γ α θ ώ ν τ ά μ ε ν ε ΐν α ι π ε ρ ί ψ υ χ ή ν , τα
τ ε ρ ω ν δ ε τ ώ ν π ε ρ ί σ ώ μ α , ύ γ ιε ία ς δε π ε ρ ί σ ώ μ α , τά δ’ εκτός, π ε ρ ί ψ υχήν μ ε ν , which we have been familiar for centuries : the Nicomachean Ethics
κ α ί ισ χ ύ ο ς κ α ι κ ά λ λ ο υ ς κ α ι τ ώ ν ο ίο ν ε υ φ υ ΐα ν τ ε κ α ί τ έ χ ν η ν κ α ί α ρ ε τ ή ν κ α ι is the undoubted treatise o f Aristotle, the Eudemian Ethics is the
π α ρ α π λ ή σ ιω ν ε κ τ ρ ίτ ω ν δε τ ώ ν σ ο φ ία ν κ α ίφ ρ ό ν η σ ιν κ α ίή δ ο ν ή ν π ε ρ ί σ ώ μ α
€/CTO?, π λ ο ύ τ ο υ κ α ί ε ύ γ ε ν ε ία ς κ α ί δ ε ύ γ ίε ια ν κ α ί ε υ α ισ θ η σ ία ν κ α ί κ ά λ λ ο ς κ α ί
problematic treatise whose attribution fluctuates, regarded now as
δ ό ξ η ς κ α ι τ ώ ν ό μ ο ιω ν . ίσ χ ύ ν κ α ί ά ρ τ ιό τ η τ α κ α ι π ά ν τ α τ α μ ό ρ ια authentic Aristotle, now as the work o f his disciple Eudemus. B u t if
σ υν τ α ΐς δ υ ν ά μ ε σ ι κ α ί ε ν ε ρ γ ε ία ις · έ κ τ ο ς δε Aspasius departs from the earlier tradition in his ranking o f the two
π λ ο ύ τ ο ν κ α ί δο^αν κ α ί ε ύ γ ε ν ε ια ν κ α ί δ υ ν α σ
τ ε ία ν κ α ί φ ίλ ο υ ς κ α ί σ υ γ γ ε ν ε ίς 1 κ α ι π α τ
1. μ ία ψ υ χ ή δυο σ ώ μ α σ ιν ε ν ο ικ ο ΰ σ α (cf. E E \2<\ο°2\ see D üring, 1 957 ^ 67 )·
ρ ίδ α . (13 6 .9 -16 ). 2. φ η σ ι δ ’ α υτο ί' Ν ικ ό μ α χ ο ς ό Ά ρ ισ τ ο τ έ λ ο υ ς τ ή ν ή δ ο ν ή ν λ ε γ ε ιν το α γ α θ ό ν .
2. T h e three points are ( i ) the statement that the goods o f the soul are π ρ ώ τ α τ ή δ υ ν ά μ ε ι, 3· T h e date o f Aspasius is uncertain: his floruit is generally placed in the first half o f the
(2) the definition o f friendship as equality o f reciprocal goodwill ( ισ ό τ η τ α ε ύ ν ο ια ς second century. It may well be that his commentary is earlier in date than the passages from
ά ν τ ισ τ ρ ό φ ο υ ), (3) the statement that the virtues are not mutually interdependent, which is Diogenes cited above, p. 28. Diogenes himself, according to R . D . H icks (19 25, xvi) ‘ has
in plain contradition to A E C , 114 4*30 ff. been variously dated in every century a . d . from the first to the fourth1.
30 The Aristotelian Ethics The Aristotelian Ethics in Antiquity 31
Ethics, he is at one with it— as we shall see— in regarding the said to be pleasure is that pleasure goes with the greatest good and cannot
disputed books as belonging essentially to the Eudemian version. be separated from it. (C A G xix. 15 1.18 -2 7 )1
Unfortunately, all that survives o f Aspasius’ writing is com T h is important passage has many puzzling features. It is commonly
mentaries on N E i “ 4, part o f A E C , and part o f N E 8. These taken as showing that Aspasius is in doubt whether the passage he is
commentaries are, o f course, in our editions treated as com commenting on— and, by implication, the disputed books in
mentaries on books 1- 4 , 7, and 8 o f the Nicomachean Ethics : but we general— belong to the Eudemian or to the Nicomachean version.
cannot be certain without further ado o f the original context o f his And this is certainly suggested by his reference to the fact that in the
commentary on C . I f the disputed books belonged originally with tenth book o f the N E Aristotle talks o f pleasure as if he were
the Eudemian Ethics, we cannot rule out in advance the possibility discussing it for the first time: this is a premiss which many people
that Aspasius’ commentary covered both ethical treatises, and that since Aspasius have reasonably used as evidence that book seven is
we have four and a half books surving o f his commentary on the N E , Eudemian. But a closer reading o f the text suggests that his doubt is
and half a book surviving from his commentary on the E E . That is, not whether the present book belongs to the Eudemian Ethics, but
indeed, the hypothesis that some o f the evidence from his whether the Eudemian Ethics, to which in his view the present book
commentary would suggest, though it may not be the one which, belongs, is by Aristotle or by Eudem us.2 When he says that
taking the evidence as a whole, is the most plausible over-all Aristotle discussed pleasure ‘in the Nicomachean Ethics'1 it is barely
explanation. L et us examine the evidence in detail. credible that he takes himself to be at that moment commenting on
T h e most important evidence is contained in Aspasius’ com a book o f the Nicomachean Ethics', if he did so, surely he would have
mentary on the following well-known passage in A E C. to say ‘in the tenth book’ or ‘later’ . Aspasius is obviously
embarrassed by what Aristotle here says about pleasure, and is torn
I f certain pleasures are bad, that does not prevent the chief good from
between several ways o f dealing with it and resolving the apparent
being some pleasure, just as the chief good may be some form of
knowledge though certain kinds of knowledge are bad. Perhaps it is even discrepancy between the present passage and the passage that we
necessary, if each disposition has unimpeded activities, that, whether the know as the tenth book o f the N E . One solution he investigates is
activity (if unimpeded) of all of our dispositions or that of some one of that there is no real inconsistency because the present passage is
them is happiness, this should be the thing most worthy of our choice; and 1. δ ία μ € ν o v v τ ο ύ τ ω ν δ ο κ € Ϊ τ α ν τ ό ν ά π ο φ α ίν ζ σ θ α ι τ ά γ α θ ό ν κ α ί τ η ν η δ ο ν ή ν ο ν μ η ν
this activity is pleasure. Thus the chief good would be some pleasure, ό ν τ ω ς ê^€t, άλλα π ρ ο ς τ ο υ ς λ έ γ ο ν τ α ς γ ίν € σ ιν e i v a i ή φ α ύ λ α ς τ ιν ά ς τ ώ ν η δ ο ν ώ ν , ο ΐς
though most pleasures might perhaps be bad without qualification. κ α ι δ ι’ α υ τό τ ό μ ή € ΐν α ι α υ τ ή ν τό α γ α θ ό ν € π ιγ ίν € τ α ι[ κ α ΐ] , € π ιχ € ΐp e f ^ δ ο £ ω ? ώ ? kvov
α ν τ η ν το ά ρ ισ τ ο ν Aeyety, έ π ζ ι 'kv y e τ ο ΐς Ν ικ ο μ α χ € ΐο ις , ένθ α δ ΐ€ ΐλ η π τ α ι κ α ι π € ρ ι
( H 5 3 V 1 3 , trans· R o ss ) η δ ο ν ή ς ’Α ρ ισ τ ο τ έ λ η ς σ α φ ώ ς € ΐρη κ€ν α υ τ ή ν μ ή τ α ύ τ ό ν ζ ιν α ι τ ή € υ δ α ιμ ο ν ία αλλά
7τα ρ α κολο υ0 €Ϊν ‘ ώ σ π € ρ τ ο ΐς ά κ μ α ίο ις τ ή ν ώ ρ α ν \ σ η μ € ΐο ν 6k τ ο ύ μ ή € iv a i t o u t '
Commenting on the second o f these sentences, Aspasius writes : Ά ρ ισ τ σ τ έ λ ο υ ς α λλ ’ Ε ύ δ ή μ ο υ το kv τ ώ ............ λ έ γ € ΐν π € ρ ί ή δ ο ν ή ς ώ ς ο ύ δ έ π ω π ζ ρ ί
α υ τ ή ς δ ΐ€ ΐλ € γ μ έ ν ο υ ' π λ ή ν €ΐτ€ Ε ύ δ ή μ ο υ τ α ν τ ά k a n v eiVe Ά ρ ι σ τ ο τ ί λ ο υ ς , ζ ν δ ό ξ ω ς
ζ ϊρ η τ α ι· δ ιά τ ο ΰ τ ο λ έ γ € τ α ι τό ά ρ ισ τ ο ν η δ ο ν ή , δ τ ι σ νν τ ώ ά ρ ίσ τ ω κ α ί ά χ ώ ρ ισ τ ο ν
These words make it look as if he was proving that the good and pleasure
α ν τ ο ΰ . T h e lacuna o f approximately six letters where one would expect the number o f the
are one and the same; but that is not so: what he is doing is—in a N E book which treats o f pleasure is interesting. T h e number which we would expect, and
dialectical manner against those who say that pleasure is a process or that which a later tradition and H eylbut obviously did expect, is ‘tenth’ . T h e lacuna may mean
some pleasures are bad and draw the conclusion that it is not the good—to that the number which was actually in the manuscripts was not this : which would support the
thesis that the original place o f the disputed books was not in the context o f the N E .
try to show that it is possible to say that it is the supreme good. For in the
2. O f course the ‘t o u t ’ ’ which he is prepared to consider as being not by Aristotle but by
Nicomachean Ethics it is discussed, and there Aristotle says clearly about Eudemus need not be the whole disputed books o f the A E , but a more limited context o f the
pleasure that it is not the same thing as happiness but accompanies it like quotation: e.g. the treatment o f pleasure. B ut (a) within the text o f Aspasius him self there is
the bloom on the cheek of youth. It is a sign that the present passage is not no indication o f a break between the treatment o f pleasure and the rest o f A E book C ; (b)
by Aristotle but by Eudemus that in th e ........... he talks about pleasure as there is no indication, independent o f the present passage, o f a Eudemian origin for the
treatment o f pleasure here other than the fact that the whole o f the A E is by the manuscript
if it had not yet been discussed ; but whether this text is by Eudemus or by tradition regarded as belonging also to a work which some scholars have regarded as written
Aristotle, what is said is dialectical ; the reason that the greatest good is by Eudemus rather than by Aristotle.
The Aristotelian Ethics The Aristotelian Ethics in Antiquity 33
32
only dialectical; another is that the inconsistency is not surprising manner. F ir s t,,it explains why he refers to the N E as if it was a
because one treatment is by Aristotle and the other is by Eudemius. different work from the one he is currently commenting on.
But having offered the fact that the two treatments o f pleasure Second, it explains why, in spite o f this, he comments on the lack o f
ignore each other as independent evidence that the current one is by cross-reference between the two treatments o f pleasure as surpris
Eudemus he then goes on to say that even if it is by Eudem us it is ing. F o r if the only books o f the E E he commented on were the
consistent with the Aristotelian treatment because it is only disputed books, the passage about pleasure, being the only clear
dialectical. doublet with the N E books, would be the most obvious place to
I f we reject as incredible the traditional suggestion that Aspasius, remark on the fact that the N E and E E apparently make no cross-
at the moment when he talks o f ‘ the Nicomachean Ethics’’ as a references to each other. T h is is indeed surprising, given the
different work, was commenting on a text o f the N E which included general frequency in Aristotle of cross-references to treatments o f
the disputed books, there are two solutions which can be offerred of similar topics in other works (e.g. to the Analytics)·,1 it is indeed
the problems set by this difficult passage. One possibility is that something from which a scholar might be tempted to draw
Aspasius commented on both the Nicomachean and Eudemian conclusions about the authenticity or spuriousness o f a work.
Ethics, and that the partial commentary on book seven is a fragment Obiously, this is no more than conjecture, as has been said. I f
o f his commentary on the E E just as his partial commentary on book Aspasius deliberately moved the three disputed books into the
eight is a fragment o f his commentary on the N E . T h e weak point of Nicomachean Ethics from the Eudemian for purposes o f commenting
this hypothesis is the following: if Aspasius regarded the two on a unified and comprehensive treatment o f Aristotle’s ethics, the
treatments o f pleasure as clearly belonging to two different works, place where he would have to announce this would be the beginning
why should he be surprised to find no cross-reference? Other o f his commentary on the first disputed book, N E 5 = E E IV .2 But
parallel treatments o f topics in the two Ethics are not signalized in his commentary on book 5 is unfortunately lost. But there are
either Ethics, so why should the absence o f a reference here be taken several other features o f the surviving parts o f the commentary
as an argument that the E E is by Eudem us? which go some way to confirm our conjecture.
There is another solution which seems to me the more plausible First, a passage in the commentary on N E 8 places it beyond
one. Th at is, that Aspasius was writing a commentary on a text of doubt that Aspasius regarded the N E as containing a lacuna that
the N E which did not contain the disputed books, but which needed filling. Aristotle is critizing those who argue that there is
contained, as our texts o f N E 1 - 4 do, forward references to only one kind o f friendship on the ground that friendship admits o f
treatments o f topics such as justice, wisdom, and continence. There degrees. ‘Things different in kind also admit o f degrees,’ he says,
existed in his time— as is surely most likely from the whole history ‘But these matters have been discussed before’ ( 1 1 35b 13 16).
we have hitherto investigated— texts o f the Eudemian Ethics 1. It has been fashionable to treat the cross-references as untrustworthy and possibly
containing the common books ; but their title gave Aspasius some interpolated. D üring (1968, 19 1) is rightly critical o f this fashion: ‘D ie zahlreichen Vor-,
Querund Riichverweise in seinen Schriften stimmen erstaunlich gut; höchstens ein Zehntel
reason to believe that they were not by Aristotle but by Eudemus,
der Verweise sind problematisch. Ein später Redaktor hätte nie ein so feinmaschiges Netz
just as the title o f the Nicomachean Ethics gave Diogenes and von Verbindungen zwischen den einzelnen Schriften hersteilen können.’
perhaps Cicero reason to believe that the N E was not by Aristotle 2. Earlier scholars who attributed the A E to Eudem us thought they could detect the
suture in Aristotle’s own text. Grant regarded as spurious the last two sentences o f N E 4.
but by Nicomachus. I conjecture that in these circumstances
(‘Continence is not a virtue either, but an intermediate state : T h is will be explained later. We
Aspasius, having commented on books 1 - 4 o f the N E , proceeded to must now discuss justice.’ ) ‘Aristotle’s M S . o f the fourth book having ended abruptly at the
comment on books I V - V I o f the Eudemian Ethics, in order to have a word ε π ιε ικ έ ς Nicomachus, or the editor, whoever he was, in all probability added these
clauses in order to give the book a seeming unity with the three Eudemian books which were
full commentary on Aristotle’s ethical system, in spite o f the lacuna
now to be grafted on’ (1885, ii.94). Aspasius’ commentary on N E 4 ends a column earlier at
in the Nicomachean version. i i 2 8 a3 3 ; the anonymous second-century commentary has a curious comment at this point
T h is can be no more than a conjecture: but it would explain all (C A G xix.126, X X . 204; see below, p. 37). B ut there is no manuscript support for the
excision.
the features o f the puzzling passage o f Aspasius in a very simple
34 The Aristotelian Ethics The Aristotelian Ethics in Antiquity 35
Aspasius comments: ‘ “ T h ey have been discussed” he says expression used to refer across the boundary between the disputed
“ before” . It looks as if they have been discussed in the lost parts o f books and the rest o f the N E , is also used in a case where, from the
the Nicomachean Ethics' (C A G x ix .16 1) .1 N ow o f course we cannot commentary on the N E , Aspasius refers to the E E }
be certain that by ‘ the lost parts o f the Nicomachean Ethics' Aspasius T hirdly, with regard to both the A E and the E E Aspasius is
means the lacuna it displays if it is deprived o f the common books. undecided whether to ascribe authorship to Aristotle or to
But such is certainly the simplest explanation o f his remark : that Eudemus. In the case of the A E we have already looked at the
lacuna is the only lacuna for which there is independent evidence in famous passage in which he assigns at least the section on pleasure
the anomalous nature o f the manuscript tradition.2 I f this is the to Eudemus ; at 2.24 on the other hand Aristotle is given credit for
lacuna that Aspasius means, we cannot tell whether he knew as a the definition o f technëin A E A and at 18.6 the psychology o f book
fact that there were once other Nicomachean books where the A E B (and elsewhere) is attributed to Aristotle, and at 178.22 in the
books now stand, or whether he merely conjectured that there must commentary on N E 8 Aristotle is referred to as the author o f the
have been because o f the forward and backward references in the treatise on justice, which is presumably A E A .2 Sim ilarly, there is
N E itself. vacillation about the authorship o f the Eudemian Ethics. At 11.5
Secondly, an examination o f Aspasius’ method o f citation from Aristotle is cited as the author of the Eudemian view o f riches, and
one part o f the Ethics to another suggests that he regarded the A E as at 132.3 (in a section o f C where Aristotle is treated without
forming a block separate from that o f the undisputed N E . When he question as the author) the Eudemian account o f the insensible man
makes a forward or back reference within a single book, or from one is quoted.3 B u t in the commentary on N E 8 a quotation from the
undisputed N E book to another, he uses an expression such as ‘ as seventh book o f the E E is attributed to Eudem us.4
was said before’ or ‘later’ .3 So too when he refers, within his T o summarize, therefore, the evidence o f Aspasius. It seems
commentary on book C, to the disputed books.4 But when he quotes clear that Aspasius regarded the disputed books as belonging to the
the disputed books from the commentary on the N E , he uses E E for the following reasons : (1) In his commentary on C he refers
different modes o f expression: ‘in the book on justice’ (16 0 .11) to the N E as a different work. (2) H e is certain that the N E is by
‘elsewhere’ (178.22, a reference to book A ; 18.6, a reference to a Aristotle, but is prepared to consider the possibility that C is by
number o f texts including book B as well as N E 1 ).5 Sim ilarly, when Eudemus, just as he is unsure whether the E E is by Aristotle or
referring from the disputed book C to the undisputed Nicomachean Eudemus. (3) With one exception, he quotes the A E when
books, Aspasius says ‘elsewhere’ (138.22, reference to N E 4) ‘ in the commenting on the N E , and the N E when commenting on the A E ,
Nicomachean Ethics' ( 15 1.2 2 ) .6 T h e discrepancy is sufficiently 1. At 1 1.5 Aspasius says that the remark in the N E that riches are desirable άλλου χ ά ρ ιν
systematic to suggest that it is deliberate.7 ‘Elsew here’ , the does not fit his remarks εν α λ λ ο ίς that they are good in themselves. T h is appears to be a
reference to the E E , where several times ( i 2 i 4 b8, 12 17 *3 7 , Ι22 7“ · 4ι Ι2 32ΐ>1°> i 248b28) riches
1. ε ϊρ η τ α ι δ ε , φ η σ ί, π ε ρ ί α υ τ ώ ν έ μ π ρ ο σ θ ε ν , ε ο ικ ε δ ε €ΐρΎ}σθαι εν τ ο ΐς ε κ π ε π τ ω -
are given as an example o f an ultimate end.
κ ό σ ι τ ώ ν Ν ικ ο μ α χ ε ω ν .
2. And o f course throughout the text o f C itself is attributed to Aristotle (13 3 .3 0 ; 136 .7;
2. I agree in this, against Rose ( 18 7 1, 107), with M oraux and Harlfinger (Harlfinger, 1972,
13 8 .19 ; 150-5)·
45 - 7)· 3. ό δ ε ε λ λ ε ίπ α ιν π ε ρ ί τ ά ς ή δ ο ν ά ς κ α ί μ η τ α ΐς ά ν α γ κ α ία ις χ ρ ώ μ ε ν ο ς ό
3. π ρ ό τ ε ρ ο ν is used thus at 19 .2 5 ; 1 2 1 . 1 5 ; 119 .2 7 ; 83.4; 88.2; 1 16 .3 ; 86.9; 1 0 5 .14 ; 53.4 ;
ά ν τ ικ ε ιμ ε ν ο ς τ ω ά κ ο λ ά σ τ ψ , ον εν τ ο ΐς κ α τ ’ ά ρ χ ά ς λ ό γ ο ις ά ν α ίσ θ η τ ο ν ε λ ε γ ε . T his
10 1.8 ; 10 6 .15 ; 168.4; 174.2. ύ σ τ ε ρ ο ν is used in a similar way at 19 .2 ; 23.29.
reference is clearly to E E I I I .2, 12 31*26 ff. : α ν α ίσ θ η τ ο ς μ ί ν οΰν . . . ό ο ύ τ ω ς ε χ ω ν ώ σ τ ε
4. So for references within b o okC , and in the two references back from C to B , 13 6 .3 1 and
κ α ι ε λ λ ε ιπ ε ιν ο σ ω ν α ν α γ κ η κ ο ιν ω ν ε ΐν ώ ς επ'ι τ ο π ο λ ν π ά ν τ α ς κ α ι χ α ίρ ε ιν . T h e N E
140.2 and 30.
does not link the insensible man with the notion o f necessary pleasures, and is indeed
5. ev ά λλ ο ι? ( 17 8 .2 2 ; i8 .6 ); ev τ ο ΐς π ε ρ ι δ ικ α ιο σ ύ ν η ς ( ιό ο .ιι) .
sceptical o f his very existence ( n o 7 b28; 1114 * 10 ). T h is quotation is very interesting, for it is
6. ε ν ά λλ ο ι? (138 .22) ε ν τ ο ΐς Ν ικ ο μ α χ ε ίο ις (15 1.2 2 ).
introduced in C by the expression εν τ ο ΐς κ α τ ' ά ρ χ ά ς λ ό γ ο ις which is used by Aspasius
7· O f thirty-odd cross-references in Aspasius I have found only one that is a real exception
(162.3) and by Aristotle ( n o 4 a2 ; ii4 g b27) to refer to earlier parts o f the same work. It is thus
to the rule that internal citations are introduced differently from citations across the N E - A E
another piece o f evidence that Aspasius regarded the A E as belonging to the E E .
boundary. T h at is 8.30, where, discussing an aporia in book N E 1 Aspasius says that Aristotle
4. λ έ γ ε ι δε κ α ί Ε ύ δ η μ ο ς κ α ί Θ εό φ ρ α σ το ς , ό τ ι κ α ί α ΐ κ α θ ' υ π ε ρ ο χ ή ν φ ιλ ία ι ε ν τ ο ΐς
solves it ττροϊώ ν. T h e reference appears to be to a passage in B ( 1 1 45s 10). T h is, unlike the
α ύ τ ο ΐς γ ίν ο ν τ α ι, ή δ ι' ή δ ο ν η ν ή δ ια το χ ρ ή σ ιμ ο ν ή δ ι' α ρ ε τ ή ν . T h e reference is to E E
other cross-rcfcrences in Aspasius, suggests that he read B as part o f a text continuous with
V II.4 , 1239*1 ff. (C A G x ix .178.3).
the N E .
36 The Aristotelian Ethics The Aristotelian Ethics in Antiquity 37
with the style o f quotation he uses for introducing quotations from M ercken,1 indicates that the writer o f the scholia was a con
other works, rather than that which he uses for forward and temporary o f Atticus and should therefore be assigned to the second
backward references in the same work. (4) From the A E he quotes half o f the second century. T h e commentary on N E 5 appears
the E E with a form o f words he uses for backward reference in the continuous with those on N E 2 -4 and is written in very much the
same work. (5) H e speaks o f the existence o f a lacuna in the N E as a same style : the writer clearly regarded N E 5 as belonging with N E
familiar fact. 1- 4 , and refers in the first line o f his commentary on 5 to the
We shall see that after the time o f Aspasius the common books discussions o f akribeia preceding in the N E . He refers forward to B
are normally taken with the N E : the earlier tradition leaves as its from an earlier book, and cites it by its Nicomachean number ( C A G
only trace— other than the presence o f the A E in the manuscripts of x x .12 5 .3 0 ); on the other hand, at the end of book 4 he appears to
the E E — a continuing confusion about the numbering o f the books promise a discussion o f continence by Aristotle in the book after
o f the N E . I f we are right in our conjecture that the Aspasian next, rather than as we would expect after two intervening books.2
evidence is best explained by postulating that he commented on the In general, this writer witnesses to the Nicomachean tradition as
three E E - books in the Nicomachean context in order to fill a lacuna familiar in later centuries.3
in the N E , then it may very well have been Aspasius’ own activity as
A L E X A N D E R O F A P H R O D IS IA S
a commentator which inaugurated the tradition o f counting the
disputed books with the Nicomachean treatise. I f that is so, the In the writings o f Alexander, who began lecturing at Athens
Nicomachean Ethics as we know them— N E plus A E — might be c. 198, there is no doubt o f the existence o f the ten-book
more aptly termed the ‘Aspasian Ethics’ . Nicomachean Ethics with which we are familiar. Alexander did not
write a line-by-line commentary like Aspasius; but both in his
commentary on the Metaphysics and in his minor ethical writings he
A T T IC U S
refers to the disputed books at belonging to the Nicomachean
Eusebius (Praeparatio Evangelica X V .4) quotes a comment by Ethics,4
the Platonist Atticus on the worthlessness o f Aristotle’s three '· ' 973, 15*·
2. ύ σ τερ ο ν δ ε ε π ί π λ έ ο ν έ ρ ε ΐ π ε ρ ί έ γ κ ρ α τ ε ία ς κ α ί κ α ρ τ ε ρ ία ς κ α ί μ α λ α κ ία ς μ ε τ α το
ethical treatises: the Eudemian, Nicomachean, and Great Ethics.1
έ ξ η ς β ιβ λ ίο ν ( C A G χχ,204)·
Atticus was active in the time of M arcus Aurelius (16 1-8 0 ): he is 3· We can make a guess at the identity o f this scholiast. We know from Athenaeus
thus the earliest testimony to the three traditional titles o f the (X V . 673 e) that Adrastus o f Aphrodisias » rote five books on Theophrastus’ ethics and six (or
a sixth) on Aristotle’s. S ix books must have been the extent o f the second-century scholia on
treatises o f the Aristotelian corpus. H e is also useful in the present
the N E , since the popular composite medieval commentary published in C A G xx appears to
context as determining the date o f the anonymous scholia on books have been built up on the basis o f the scholiast’s work, replacing particular books by works of
2 -5 of the N E , which we shall next consider. more prestigious Christian writers as these became available, as in the case ofE ustratius for
book i. (T he trumping process can actually be observed in the N ew College manuscript of
the commentaries in which both the anonymous scholia on A E A and the preferred
T H E S C H O L IA S T commentary o f M ichael o f Ephesus have been preserved.) T he composite commentary,
however, has to rely on a late anonymous author for book 7, suggesting that the original
T h e scholiast on book 5 of the Nicomachean Ethics ( C A G xx. scholiast had here neither commented nor been trumped. Som e may find it too much of a
248.24-6) is discussing the idea that some things are indifferent coincidence to have both a second-century six-book commentator on the Ethics whose
commentary has been lost and a second-century six-book commentary on the Ethics whose
(adiaphora) : he mentions its proponents ‘among whom previously author has disappeared. In addition to the reference to Atticus the internal evidence o f the
Aristonymus belonged, but now even some who pretend to be scholia suggests a second-century date. T h e latest author quoted is L u cian ; there is no
Platonists, while surreptitiously introducing this doctrine, among evidence o f Christian interest with the possible exception o f the three words ο ύ τ ω ς κ α ι
Ά β ε σ α λ ώ μ (229.22) at the end o f the story in book 5 o f Phoenix’s affair with his father’s
whom also Atticus appears to belong’ . T h is passage, noticed by concubine. M ercken (1973, 15 *) is probably right to see these words as a late gloss, added at
I. α ί γ ο νν Ά ρ ισ τ ο τ έ λ ο υ ς π ε ρ ί ταντα πρα γματείαν Ε ύ δ ή μ ε ιο ίτ ε κα ί Ν ικ ο μ ά χ ειο ι the very end o f a paragraph.
κ α ι Μ ε γ ά λ ω ν Η θ ι κ ώ ν ε τ τ ιγ ρ α φ ο μ ε ν α ι μ ι κ ρ ό ν τ ι κ α ι τ α π ε ι ν ό ν κ α ί δ η μ ώ δ ε ς π ε ρ ί 4. In his commentary on the Prior Analytics ( C A G ii. 1.8 .3 1) he refers 10 hook 1 o f the N E
α ρ ε τ ή ς φ ρ ο ν ο ν σ ι. and then refers to there being ten books in that work. In commenting on the Metaphysics
38 The Aristotelian Ethics The Aristotelian Ethics in Antiquity 39
C L E M E N T O F A L E X A N D R IA between the third and tenth centuries yields rather inconclusive
results.1
In his ‘miscellanies’ , which are commonly dated to the very
A late witness to the uncertainty about the authorship and
earliest years o f the third century, Clement draws freely on both the
content o f the Nicomachean Ethics is contained in the entry in the
Eudemian and the Nicomachean Ethics, quoting the Eudemians
Suda on Nicomachus, which is said to be likely to derive from
without hesitation as Aristotle’s work. His method o f utilizing his
Hesychius (i. 5th cent.). There we are told that Nicomachus wrote
sources makes it difficult to draw any conclusions about the position
six books o f Ethics as well as a commentary on his father’s
o f the disputed books in the texts available to him .1
Physics. Some commentators regard this as a confusion between
Nicomachus and Eudemus, who did write a commentary on the
LATER CO M M ENTATO RS
Physics, and whose name is associated with the E E (some
N o commentary on Aristotle’s Ethics survives from the period manuscripts o f which lump books V II and V I I I together). Possibly
between Alexander and the Byzantine renaissance. In the second so: but equally possibly the six-book Ethics might be the
h alf o f the third century Porphyry wrote a commentary which is Nicomachean Ethics less the disputed books; there is one book short
known only through Arabic sources. Scholars assume that it was a in either case.2
commentary on the Nicomachean Ethics, but there is no decisive
T H E C A T A L O G U E S O F A R IS T O T L E ’ S W R IT IN G S
evidence for this; all we know is that the commentary was on a
twelve-book ethics (or was itself in twelve books).2 Three ancient lists o f Aristotle’s works have come down to us.
A study o f citations o f the Ethics by commentators on Aristotle T h e first, in the Life by Diogenes Laertius (\\22--7) is commonly
held to have been made by Herm ippus, librarian o f the library o f
Alexandria, who wrote a life of Aristotle in about 200 B .C . T h e
(C A G 1.7 .13) he refers to B as book 6 o f the Nicomachean Ethics ; the commentary on the second is in the Life o f Aristotle by Hesychius (the Vita Menagiana)
Sophistici Elenchi in his name ( I I .3.172.20) refers to A as book 5 o f the N E . In his o f the fifth century. T h e third is the list o f Ptolemy which is
commentary on the Topics he quotes the N E a dozen times, but without specific citations. H e
does not appear to have quoted the E E . T h e Ethical Questions attributed to him on doubtful
preserved in two Arabic versions in the Lives o f Aristotle by al-Qifti
authority discuss passages o f the N E at length, and refer twice (Suppl. A r. II. 12 7 .2 9 ; 134.30) (d. 1248) and by Usaibia (d. 1270).3 We have already referred to the
to book C as N E 7. T h e picture is not entirely clear, however, for in A lexander’s second third list in discussing Andronicus : we must now consider the other
treatise De Anima there are two passages which suggest an eleven-book Nicomachean Ethics
(Suppl. A r. II). A t 1 5 1 . 1 1 a reference to N E 8 is described as occurring kv τ ώ iv a r w , nine two.
lines further on a reference to N E 9 is described as èv τ ώ δ € κ α τ ώ , and in 15 3 .14 a quotation T H E L I S T O F H E R M IP P U S
o f E E 10 .3, i i7 5 aio is introduced with the words A eyet Sk kv τ ώ κ τ ώ ν Ν ικ ο μ α χ ε ίω ν .
T h is work may be however considerably later than Alexander (see Gauthier, 1970, 10 1). T h e thirty-eighth item in the Hermippus list is an Ethics in five
1 . T hu s he makes use o f the quotation which introduces the E E and occurs later in the N E y
books ; no other Ethics is listed. M any scholars have seen this five-
at Strom . 7 .1 2 ; he refers to the etymology o f σ ω φ ρ ο σ ύ ν η from the common book B at 7.3,
and the reference to the stubbornness o f m ere belief in book C at 7 .16 . T here is a long 1 . Philoponus in the fifth century quotes both the N E (C A G xv i.3.19 ) and the E E (C A G
development, without explicit acknowledgement o f Aristotle, o f the passage in N E 3 xvi.237.26, 335.22) as sim ply ‘the Ethics'. H e quotes 1139*8, from A E B , as being from the
concerning ignorance o f circumstances. Nicomachean Ethics (xiv.3.84.15). Sim plicus in the sixth century quotes the N E and E E both
A quotation from the E E is introduced with an explicit mention o f Aristotle in the by name (v iii.4 .2 7,6 .14 , 12 .8 ,17 0 .6 ). A to n e point he attributes to the N E a distinction made
following passage: κ α ν ο ’Α ρ ισ τ ο τ έ λ η ς ά ν α λ ο γ η > τ ό μ € ν π ο ιε ΐν κ α ι k n i τ ώ ν α λ ό γ ω ν only in the E E in our manuscripts (x.303.32 ff.).
ζ ώ ω ν τ ά σ σ € σ θ α ι κ α ι €7τι α ψ ύ χ ω ν δ ιδ ά σ κ ω ν , τ ό δ έ ττρά ττ€ ΐν ά ν θ ρ ώ π ω ν e iv a t μ ό ν ω ν 2. έ γ ρ α φ ε ν ’ Η θ ι κ ώ ν β ιβ λ ία ϊ ' κ α ι π ε ρ ί τ η ς φ υ σ ικ ή ς ά κ ρ ο ά σ € ω ς τ ο υ π α τ ρ ό ς α ν το ΰ .
(cf. Ε Ε 2.6, I2 2 2 b20; 2.8, i2 2 4 a28). T h e passage is printed in D üring, 1957, 265, who conjectures that it has been transferred
2. Porphyry’s commentary is mentioned by the Arab bibliographer Ibn al-Nadim ( a .d . from a notice o f Eudemus.
987); see M iiller, 1873, an^ Walzer, Greek into A rabic, 220 if., 240. T h e commentary 3. M oraux attributes the list in Diogenes to Ariston, scholarch o f the Peripatos in the last
certainly included discussions o f N E 1 and N E 10, as we know from the quotations in other quarter o f the second century ( 19 5 1, 2 37 ); but this attribution has not been generally
Arabic writers: see G hourab, in Stern, H ourani, and Brown, 1972, 78, on al’A m iri. From accepted: see D üring, 1956, 1 1 - 2 1 ; K earney, 1972. T h e list o f H esychius is printed in
al’ A m iri we know also that Them istius, the secretary o f Ju lian the Apostate, commented on Düring, 1957, 8 3 - 9 ; the most accessible reproduction o f the Ptolemy list is likewise in
the passages on practical reasoning in book B o f the A E (ibid. 83-7). Düring, 2 2 1 - 3 1 ; see also M oraux, 1951, 2 8 9 -3 2 1.
40 The Aristotelian Ethics
The Aristotelian Ethics in Antiquity 41
book Ethics as being the E E without the disputed books.1 It may be Ethical works appear to be misplaced. One would expect them to be
so, and if it is so, it is further confirmation o f the general thesis that gathered together in the section devoted to practical philosophy,
the E E , during the centuries immediately succeeding Aristotle’s immediately before the political works, say between items 73 and 74
death, was better known than the N E . It would, however, suggest o f the list. In fact they are scattered among the logical and
that the disputed books did not belong with the E E , and this would dialectical works in pairs, so that the first part o f the list presents the
clash with the evidence we have seen and the evidence to be following appearance:
presented later in this book. Other scholars have stressed that the
division o f the Ethics into books may well be later than the catalogue 1- 2 2 Popular and Platonic works
o f Herm ippus, so that the five books may not correspond to any o f 23 Economicus
our present books. T h us Margueritte, reviewing Jaeger in 1930 24 On Friendship, one book
(p. 104) suggested that the title in the catalogue represented the 2 5 -3 6 Logical and dialectical works
complete Eudemian Ethics in five books made up from E E I (1), E E 37 On the passions o f anger, one book
II and I I I (2), A E A (3), A E B and C and E E V I I I (4), E E V II (5) : 38 Ethics, five books
the topics o f the five books being happiness, virtue, justice, wisdom, 39-52 Logical and dialectical works
friendship respectively.2 53 On what is better, one book
A close inspection o f Diogenes’ list permits a conjecture about 54 On the idea (of the good), one book
the five-book Ethics which is simpler than M argueritte’s, and which 55-67 Logical and dialectical works
is more in accord with the evidence o f Aspasius, and the internal 68 On the voluntary, one book
evidence we shall shortly examine, than the common opinion which 69 On the noble, one book
sees in Diogenes’ list a reference to the E E without the disputed 70-3 Dialectical works1
books. From 74 the practical works begin, and the disturbed pattern ends.
Before proposing it, I must draw attention to some features o f the T h e fact that the ethical works occur thus spaced in pairs, at more or
ordering of the titles in the list.3 T h e list begins with some two less regular intervals (from the beginning the intervals are 2 x 1 1
dozen popular and Platonic works. There follow about four dozen titles; 12 titles; 13 titles; 12 titles) suggests that they originally
mainly logical and dialectical works. N ext come sixteen works on belonged together and have been misplaced either owing to an
practical philosophy (politics, rhetoric, poetry). Then, twenty accident in the manuscript tradition or to some feature o f the library
works on physics (including ten on the history of physics) and on whose catalogue is most likely at the basis o f Herm ippus’ list.2
biology, plus half a dozen works on mathematical topics including I f this is so, then inter alia the three titles O n friendship’, O n the
optics and musical theory. T h e remaining thirty titles are passions o f anger’, and ‘Ethics, five books’ may be taken together.
‘hypomnematic’ works, collectanea, letters, and poetry. T h e most
interesting feature o f the ordering, in the present context, is that the 1. T h e Greek titles o f the Ethical works are: Ο ικ ο ν ο μ ικ ο ί α, π ε ρ ι φ ιλ ία ς a, π ε ρ ί
π α θ ώ ν ο ρ γ ή ς â , Η θ ι κ ώ ν α β γ ά ί, π ε ρ ί τ ο ν β ε λ τ ίο ν ο ς α, π ε ρ ί τ ή ς ιδ έ α ς ä, π ε ρ ι
εκ ο υ σ ίο υ <χ, π ε ρ ί κ α λ ο ύ ä. Several points in m y list call for comment, (t) 1 have not counted
1 . T h u s Jaeger, 1948 ; R oss, 19 23 ; M oraux, 19 5 1. D üring, in the apparatus to his edition
34— π ρ ο τ ά σ ε ις π ε ρ ί α ρ ε τ ή ς ä ß — as an ethical title, because π ρ ο τ ά σ ε ις on all topics seem
o f the list (19 57, 43) says : ‘Potest esse editio Ethicae Eudemeae, sine libris D E Z et cum libro
to be regarded by the compiler as dialectical (see 4 5 -7 ,6 6 -7 ). (2) In counting I have followed
H in duos libros ( 1 - 5 , 6 - 1 5 ) distributo, sed mera coniectura est.’ Titze, 1825, thought that
D üring ( 19 5 7 ,4 4 -5 ) in regarding 45-6 and 66-7 as making a single title. (3) I assume that the
the five-book Ethics was the first five books o f the N E ; he sought the remaining books o f that
Idea mentioned in 54 is the idea o f the Good. What other idea deserves the title o f ‘ The Idea’ ?
Ethics under other titles, and located the E E as items 20 (v e p i τ ά γ α θ ο ν α 'β 'γ ) and item 24
(So Titze, 1825, 38.) (4) It is not clear whether the practical section o f the list should be
(7τ ςρ ϊ φ ιλ ία ς·, ά) o f D iogenes’ list.
regarded as commencing at 68 (π ε ρ ί ε κ ο ν σ ίο ν ) or at 74 (π ο λ ιτ ικ ά β I. T h e reason is that
2. In support o f this M argueritte appeals to ‘l’annonce que fait V E E au début du L iv re I
7 0 -3 are all θ έ σ ε ις and therefore dialectical; but are, in at least three o f the cases, about
des deux principaux sujets à traiter, Γ ά ρ € τή et le φ ρ ό ν η ο ις , et aux réferences ou aux
more or less ethical topics. See below, p. 45.
transitions que nous lisons ailleurs’ (1930, 104).
2. M oraux observed that the ethical titles 24, 37, and 38 were out o f place (19 5 1, 190); he
3. On the ordering o f the list, see M oraux 19 5 1, 190, and D uring 1957, 68-9.
has an ingenious reconstruction o f the pinax to explain this and other displacements.
42 The Aristotelian Ethics The Aristotelian Ethics in Antiquity 43
In the light o f this, I conjecture that the five-book Ethics is neither between that work and the eight books o f our Politics, listed at 75 in
the truncated E E minus the A E , nor the hypothetically recon the portion o f the list succeeding the disturbance o f the order.
structed Eudemian Ethics o f M argueritte, but the first five books o f All this o f course is conjecture. I make no claim for it other than
the Eudemians as they are in the manuscript tradition, i.e. E E I —I I I the exceedingly modest claim that it is as plausible as most o f the
followed by A E A and A E B. These books read as a consecutive conjectures which are current about the titles in the H erm ippus list.
whble, each leading naturally into the next and forming as a totality Certainly it accords with all the ancient external evidence as well as
a treatise on the virtues. Book C o f the common books makes a fresh with the internal contents o f the treatises we possess. B u t some o f its
start and applies a new scheme for the classification o f character to more unexpected features can be further confirmed, as we shall now
the material discussed in earlier books. W hile its doctrine is see.
consistent with, and is at many points linked with, the earlier
Eudem ian books, it is capable o f being studied separately as a TH E L IS T O F H E S Y C H IU S
treatise on continence and incontinence and pleasure, and in
Hesychius’ list contains 197 titles. T h e first 139 o f these form a
general the influence o f passion on action. Book V II o f the E E (with
group which resembles so closely the list o f Diogenes that no
or without the chapters which in some manuscripts are treated as a
scholar doubts that in some manner it originates from the same
separate book V III) is also capable o f standing alone as a treatise on
source. T h e remaining titles form an appendix, in four separate
friendship.
parts. Both the main text and the appendix throw light on our
I f we are right to see the five-book Ethics as E E I V , it is natural
concerns.
to look for signs o f E E V I and V I I elsewhere in the list.1 N ext to it in
T h e thirty-ninth title is an Ethics in ten books. M ost scholars
the list is the title O n the passions o f anger’ . T h e plural ‘passions’ is
naturally enough identify this with the Nicomachean Ethics; but
very strange, and all editors emend the text. Rose altered the text to
they do not agree what value to place on the testimony. M oraux
O n the passions or on anger’ ; D üring thought that two titles O n
regards the list as having been transcribed from the same original
the passions’ and ‘on anger’ had been telescoped together.2 A much
catalogue as Diogenes’ list, but as having undergone a process o f
simpler emendation than either o f these is possible. T h e change o f a
correction designed to eliminate apparent duplicates and to bring
single letter in the Greek would make the title O n the influence of
the list into accord with the works o f Aristotle extant in the
the passions’ .3 T h is would be an appropriate title for E E V I (the
Christian era. M oraux agrees with D üring that Diogenes and
common book C) which is mainly devoted to incontinence.4 N ext to Hesychius independently used the same original, and that in some
this title, if the ethical works were originally grouped together small and rather unimportant details Hesychius has preserved the
without the intervening dialectical dozens, we read O n
original better than D iogenes.1 H e thinks, accordingly, that the
friendship’— the obvious title for E E V II. Th us, these three figure ‘ 10 ’ is a comparatively late alteration to the original list. I f so,
adjacent titles would yield our Eudemian Ethics complete. N ext to
it cannot be relied on as giving independent evidence o f a ten-book
them is the Economicus. I f this is the first book o f our Oeconomica, as
Nicomachean Ethics prior to Aspasius.
many scholars believe, it would go naturally with the E E as a link It is possible, however, that the list in Hesychius is not so much a
defective copy o f the Hermippus list, as an independent inventory
1 . T h e several books o f the Topics apparently appear scattered in the list under separate
o f the same library. A comparison o f the two lists yields a number o f
titles; cf. M oraux, 19 5 1, 54 ff.
2. R o se: περι, π α θ ώ ν ( ή π ε ρ ι) ο ρ γ ή ς ; D ü rin g: π ε ρ ι π α θ ώ ν *(α π ε ρ ι) ο ρ γ ή ς à. features which suggest this. For instance, though the titles listed
3· π ε ρ ί π α θ ώ ν ό ρ μ ή ς . largely correspond, there is a great deal o f random variation in the
4· I am not o f course suggesting that the title is Aristotle’s own (though ό ρ μ ή , i i0 2 b2 i ,
h i 6b30, is a word he uses in this sort o f context) nor that he did not him self regard the E E as
order in which they occur. However, no title travels very far from its
a complete and unified whole. I see from Bodéus (1974, 459) that M asellis (19 56 , 348) also original position. It is hard to see what reason a copyist would have
identifies this work with A E C . N ot having had direct access to M asellis’s article I do not
I. M oraux, 19 5 1, 19 5 -2 0 9 ; D üring, 1957, 9 1.
know whether he supports his conjecture with a sim ilar emendation.
44 The Aristotelian Ethics The Aristotelian Ethics in Antiquity 45
had to make so many tiny pointless changes in order : but it is easy 3 1-8 Logical and dialectical works (8)
enough to imagine books being replaced on shelves erratically, and 39 Ethics, ten books
catalogued at different times in different orders.1 40-4 Logical and dialectical books (5)
Once again, though, the ethical titles are scattered among the 45 On the idea (o f the good)
Logical and dialectical books. But this time, instead o f being 46-9 Logical and dialectical books (7)
distributed in pairs, with roughly a dozen books between each pair, 50 On what is better
we find them in approximately the same order,2 but singly, with 5 1-7 Logical and dialectical books (7)
roughly half a dozen books between each one. T h e list goes th u s:3 58 On the voluntary
i-16 Platonic and popular works 59-62 Logical and dialectical books (4)
17 Oeconomicus 63 On the noble
18 -2 3 Platonic and popular works (6) 64 Political works and theseis
24 On friendship, three books Probably it is idle to speculate on the cause o f this striking
25-9 Logical and dialectical works (5) relationship between the two lists. Certainly, it confirms the
30 On the passion o f anger hypothesis o f M oraux that a displacement o f the ethical titles has
taken place: and surely whatever explains the original misplace
i . T h is impression is particularly strong in the case o f the group beginning at no. 6 in each
ment will also explain why the titles occur in one list in pairs
catalogue (nos. 1 - 5 coincide).
separated by a dozen titles, and in the other list in singletons
Diogenes Hesychius
6. Nerinthus, one book 6. Nerinthus, one book separated by half a dozen titles.1
7 · The Sophist, one book 7 · On wealth, one book M ost likely, the relationship between the two lists is to be
8. Menexenus, one book 8. The Sophist, one book
explained by a combination o f displacements o f books in the library
9 - Concerning love, one book 9 - On prayer, one book
10. Symposium, one book 10. Menexenus^ one book whose inventories are their basis, and emendations o f the texts of
I I . On wealth, one book I I . On noble birth, one book the lists between the time of Hermippus and Hesychius. M oraux’s
12. Protrepticus, one book 12. Concerning love, one book
explanation o f the ten-book Ethics on the latter basis has already
13· On the soul, one book 13- On the soul, one book
been mentioned. An explanation on the former basis is also
14· On prayer, one book 14. Protrepticus, one book
15· On noble birth, one book possible. Suppose that our earlier supposition, that the library of
It will be seen that in this section the titles contained are exactly the same, except that the Hermippus contained a five-book E E , plus two separately listed
Symposium has gone m issing in the second list. H owever, the order is varied in such a way
books making up the traditional Eudemian Ethics, is correct.
that the first four titles o f H erm ippus’ list appear as even-numbered titles in H esychius’ list,
and the last three titles o f H erm ippus’ list appear as odd-numbered titles in H esychius. It is i. It is not difficult to imagine circumstances which would explain these features o f the
hard to see what motive a copyist could have for such a sw itch; but if we imagine the books lists. Suppose, for instance, that a section o f the library was, for some reason, arranged as
arranged thus in the library, it is easy to see how the two lists could arise: follows, with the ethical works forming a separate, vertical group:
Nerinthus On Wealth H a lf a dozen Platonic works Oeconomicus
Sophist On prayer H a lf a dozen logical works On Friendship
Menexenus On noble birth H a lf a dozen logical works On the passions
Concerning love On the soul H a lf a dozen logical works Ethics in five books
Symposium Protrepticus H a lf a dozen logical works On the idea
I f one man catalogued in columns, and the other read the titles across, the first man would H a lf a dozen logical works On the better
produce a list very like H erm ippus’, and the second a list very like H esychius’ : the differences H a lf a dozen logical works On the voluntary
would be explained by the loss o f one book, and the slight reordering o f two others, between H a lf a dozen logical works On the noble
one cataloguing and the next. I f H esychius’ cataloguer went from left to right along each shelf in turn, from bin to bin, he
2. T h e order is identical except for the reversal o f the two titles O n the Idea’ and O n what would get the order we find in H esychius’ list; if a cataloguer went zig-zag, reading alternate
is better’ . shelves from left to right and then from right to left, he would get the order in H erm ippus’
3 - T h e numbers in brackets give the total o f intervening titles.
46 The Aristotelian Ethics The Aristotelian Ethics in Antiquity 47
Suppose further that between the time o f Herm ippus and the family Π 1 ), and the Recensio Constantinopolitana, o f Byzantine
inventory which is the basis for Hesychius the remaining origin, of which the archetype is the M S . Laurentianus 8 1, 15 o f the
Nicomachean books had found their way into the library. Unless we 1420s (L ) .1
suppose that at the same time the Eudem ian books had got lost, the Since the time o f Bekker it has been customary to say that the
ten books o f Ethics now listed must be the original five Eudemian manuscripts o f the E E do not include the text o f the common books
ones plus five o f the Nicomachean books, presumably N E 1 - 4 and o f the A E , but merely refer the reader to the text o f the N E .2 In
10 (N E 5 and 6 being already among the five-book Ethics, and N E 7 fact, as Harlfinger has shown, eleven o f our twenty-one manu
being the book on the influence o f the passions1 ). That would leave scripts, all deriving from the Recensio Constantinopolitana and
the two Nicomachean books on friendship. These would pre dependent on L , set out the text o f the A E in full within the context
sumably be placed with the original Eudem ian book on friendship, o f the E E . It is only in manuscripts in which the text o f the N E
making now a total o f three books. And that is exactly what we do precedes that o f the E E that copyists, to save themselves from a
find. In the place where the Herm ippus list has O n friendship : one duplication o f labour, have contented themselves with a backward
book’ the Hesychian list has O n friendship: three books’ . reference to the fifth, sixth, and seventh Nicomachean books. No
In the Appendix Hesychiana we find, under no. 174, the entry O n argument, therefore, to an original Nicomachean home for the
the Nicomachean E thics’ . M oraux thinks this is the N E itself. T o disputed books can be drawn from the failure o f a number of
do so, he has to postulate some very botched editing by a compiler manuscripts to set them out in full in their Eudemian context.
who (a) did not recognize the N E intended, on M oraux’s view, by One might hope, however, that a stemmatological study o f the
the correction in 39 ; and (b) quite misunderstood the work’s title, manuscript traditions o f the books o f the A E might throw light on
inserting an ‘about’ . Surely this is more likely to be a work, not o f the problem. If, for instance, the Eudemian tradition o f these books
Aristotle, but o f some scholar o f the period between Hermippus turned out to be independent of, and superior to, the Nicomachean
and Hesychius, concerning the Ethics. Such a work would be more tradition, that might support the hypothesis that the books belonged
likely to be included in a catalogue if it were itself a fragment of a originally in a Eudem ian context. In a long and patient study
catalogue raisonné, discussing perhaps the troubled history o f the Harlfinger has shown that this hope is delusory. In fact, he inclines
transmission o f the N E to which the present study has borne to think that the E E tradition o f the disputed books, represented by
witness. L and its descendants, takes its origin from a point somewhere in
the genealogy o f N E manuscripts close to the vetustissimus K b.3
T H E M A N U S C R IP T T R A D IT IO N T h is, in turn, does not show that the N E is the original home o f the
common books, as Harlfinger is at pains to point out. For it may be
T h e manuscript tradition o f the E E is very inferior to that o f the
N E . T h e oldest manuscript o f the N E is the tenth-century 1. Harlfinger, 1 9 7 1 , 1 - 4 9 .
2. Such was the case in the unrepresentative manuscripts which Bekker collated. T h u s
Laurentianus 8 1, 1 1 (Bekker’s K b); the E E does not survive in any M arcianus 200 (Q) and 2 13 (M b)atth e end o f E E I I I carried this remark: σ η μ ε ίω σ α ι δ τ ι το
manuscript earlier than the thirteenth century, from the end o f 8 κ α ι € κ α ι s1' τ ω ν Ε ύ Β η μ ίω ν η θ ικ ώ ν ε ν τα ύ θ α π α ρ ε ίν τ α ι δ ια το το μ ε ν 8 τ ώ e τ ώ ν
Ν ικ ο μ α χ ε ίω ν , το 8 ε ε τ ώ ζ, το 8e s1' τ ώ η τ ώ ν Ν ικ ο μ α χ ε ίω ν εν π ά σ ι κ α ι κ α τ α λ ε £ ιν
which date the M S . Vat. gr. 1342 (Pb) and the Cambridge M S .
ο μ ο ια ε ίν α ι. P b contained merely the incipits o f books I V - V I o f the E E .
Cant. Ii 5.44 (C c). Altogether twenty-one manuscripts o f the E E 3. On the basis o f collations o f ten Bekker columns in the A E H arlfinger concludes : ‘Das
survive, the great majority from the fifteenth century. In a masterly in D etailfragen mit gewissem Vorbehalt, in der allgemeinen Tendez aber als ziemlich
gesichert zu betrachtende Ergebnis ist folgendes : D er T e x t underes L a u r. 8 1, 15 (L) ist sehr
study Dieter Harlfinger has shown how they can be grouped into
eng mit dem des L a u r. 8 1, 1 1 ( K b), jenes berühmten vetustissimus der E N und der Magna
two main classes: the Recensio Messanensis, o f Southern Italian M oralia, verwandt. . . . Dam it ist natürlich erwiesen, dass der T ext der kontroversen
origin, whose earliest representatives are Pb and C c ( = Susem ihl’s Bücher, wie er uns in L vorliegt, aus einem K od ex der E N stammt, der im Stem m a dieser
Ethik irgendwo diesseits des Archetypus in der N ähe, aber nicht in der Deszendenz des
i. T h e title being, in our manuscripts o f H esychius’ list, now further corrupted, a well- höchstens eine von mindestens zwei voneinander unabhängigen Fam ilien repräsentierenden
meaning scribe having corrected π ep l π α θ ώ ν ο ρ γ ή ς into π epi π ά θ ο υ ς ο ρ γ ή ς . Lau r. 8 1, i i (K b) angesiedelt werden muss’ ( 19 7 1, 43).
48 The Aristotelian Ethics The Aristotelian Ethics in Antiquity 49
that whereas the earliest E E manuscripts (those available; for view o f the dependent nature o f the Eudem ian tradition. There are
instance, to Aspasius) contained the disputed books, there came a six places in which there are discrepancies between the papyrus and
time when the only surviving manuscripts were ones in which the the manuscripts. In four o f these the papyrus agrees with K b against
E E followed the N E and the common books were therefore L . 1 In two it agrees with L b against L . 2 In no case does it agree with
represented in the E E context only by their mcipits or by a back- L against K b. T h e tradition which— in these passages— the
reference. Then, at a stage later than that represented by the K b readings of L approach most closely to is that represented by M b. 3
manuscript o f the N E , a scribe wishing to have a complete text o f
1. T hree o f these are simple errors in L (α ύ ξ η σ ις for α ν τ η in I i4 2 bi 3 ; où omitted in line
the Eudemian Ethics would need to obtain his text o f the common bi4 ; μ ο ν λ ε υ ό μ ε ν ο ς for β ο υ λ ε υ ό μ ε ν ο ς in bi4). T h e fourth is the reading h iw p ia r a i at
books from an existing Nicomachean version. Harlfinger concludes i i 4 2 bn in which L agrees with M b.
2. T h e readings εά ν τ ε instead o f έά ν τ ε κ α ι at H 4 2 bi5 , and the reading ε υ δ α ιμ ο ν ία ν at
that the manuscript evidence does not permit us to settle the the well-known crux in H 44b6.
question o f the original context o f the disputed books. 3 - In the two readings mentioned in the previous footnote, and the one non-trivial reading
Students o f the text o f the N E differ in their valuations o f its most mentioned in n. i, L is in agreement with M b; it agrees with K b only where the latter agrees
with M b. Since, according to Gauthier (1970, 3 1 1 ) xMb derives from a tradition independent
important m anuscripts; and their relative value appears to vary
o f K b and L b, this tiny sample o f readings does not accord with the results o f H arlfinger’s
from book to book. Bekker based his edition principally on four much more extensive survey.
m anuscripts: in addition to K b he used L b (Parisiensis 1854, 12th
cent.), M b (M arcianus 2 13 , 15th cent.), and Ob (Riccardianus 46,
14th cent.). Bonitz and Susemihl regarded M b, in spite o f its late
date, as being second in importance to K b. Jackson’s edition of book
A o f the A E was based on a wider collation than previous ones: for
that book he regarded L b as the most trustworthy, though he
conceded that K b was the richest in preserving good readings
elsewhere lost. Byw ater, whose edition is the one most familiar to
English readers o f the N E , accorded an enormous primacy to K b,
while at the same time attaching perhaps undue importance to the
readings o f the medieval Latin version, which he regarded as
‘almost as good as a Greek manuscript to us’ .1
T h e oldest witness to the text o f the Ethics is a single papyrus of
the second century.2 It contains only two small fragments, from the
common book B ( i i 4 2 bn - i 7 and ii4 4 z6 - i i ) ; in two places it
supports the readings o f L b against K b, to that extent corroborating
Jackson’s judgement against Byw ater’ s. B y comparing the text o f
this papyrus with that o f the best N E manuscripts on the one hand,
and that o f the Eudemian tradition represented by L on the other,
one might hope to confirm or discredit H arlfinger’s tentative
conclusion that the text o f the A E , as it stands in our Eudemian
manuscripts, is not independent but derives from the Nicomachean
stemma. T h e comparison goes some way to confirm H arlfinger’s
1. On the manuscripts o f the N E see Bywater, 18 9 2 ; Jackson, 1879, ix - x ii; Gauthier,
1 9 7 ° , 30 T-1 4.
2. Oxyrhynchus Papyrus no. 2402; Lobel, Roberts, T urner, and Barns, 1957, 12 4 -6 .
Cross-references in the Ethics 51
T h e early Nicomachean books are not rich in forward-looking
references.1 When the three lives are introduced in 1.5 the
consideration o f the life o f contemplation is postponed until later
CH APTER 2 (1096*7) ; and a little further on we are told that the element o f self-
sufficiency in happiness needs to be looked at again (io g 7 bi4). Both
Cross-references in the o f these promises seem to be kept within the undisputed
Nicomachean books, the latter in chapters 9 and 10 o f book 9, the
Ethics former in the second half o f book 10. B u t after the introduction o f
the Mean in book 2 we read, at the end o f the preliminary list o f
mean states: ‘About these there will be another occasion to speak;
H a v in g examined the external evidence for the relationship but with regard to justice, since the term is used in more than one
between the Eudemian and Nicomachean Ethics I turn henceforth to sense, we will later distinguish between them and show how the two
the internal evidence concerning the original context o f the kinds are severally mean states ; and similarly with regard to the
disputed middle books. For, as I observed earlier, this is the rational virtues’ ( n o 8 b6 ff.). T h e ‘other occasion’ to speak about
question which should be raised and answered before any attempt is the listed virtues is presumably the detailed treatment o f them in
made on internal grounds to settle anything about the systematic or books 3 and 4. T h e reference to justice will fit well the treatment in
chronological relationship between the two Ethics. For if this is not A E A : there he does indeed distinguish between two forms o f
first settled, then that Ethics to which the disputed books rightfully (particular) justice, namely distributive and corrective, and argues
belong— whichever it may be— is being unfairly judged on the that each o f them is concerned with a mean ( i i 3 i ai7 ff. and
basis o f an unnecessarily fragmentary text ; and in any case, as we i I3 2 ai9 ff.). However, the final remark about ‘the rational virtues’
shall see, there is very much more internal evidence available to does not seem to fit the content o f the disputed books : at no point in
indicate the true home of the disputed books than there is to them does Aristotle attempt to show that the intellectual virtues are
indicate which o f the two treatises, as a whole, is chronologically mean states. 2
prior. Failure to observe these obvious methodological points led in A more apt foreshadowing o f the content o f A E B comes earlier
the case o f scholars as gifted in different ways as Schleiermacher in book 2 where virtue is provisionally described as acting in
and Jaeger to fallacious and indeed circular argumentation.1 accordance with right reasoning: ‘Th ere will be a discussion o f this
In such an inquiry the first essential though unexciting task is to later, and o f what right reasoning is, and what is its relation to the
examine the cross-references internal to each o f the ethical treatises other virtues’ ( n o 3 b32-4). Right reasoning on the topics appro
to see whether those in the undisputed Nicomachean books fit the priate to the virtues, we are told at 1 1 44a24 in A E B , is wisdom ; and
content o f the controverted books better or worse than those in the there follows immediately a discussion o f the relation between
undisputed Eudem ian books. In the present chapter I will examine wisdom and the other virtues. (T he points made here about the
the forward-looking references in the first five Nicomachean books, relation between wisdom and the virtues are also made in N E 10,
and the first four Eudemian books— excluding those which find 1 1 78at6, but too briefly to be the fulfilment of the promise o f N E
their fulfilment clearly within those books themselves— and 2·)3
compare them with the disputed books. 1 . Examinations o f the cross-references can be found in Spengel (18 4 1, 489 if.), Fritzsche
i. Jaeger argued that the disputed books m ust belong to the N E because they contained a ( 1 8 5 1 , xxxiv ff.), G ran t (188 5, 47 ff.), Festugière (19 36, xx x-xliv), Fragstein (1974, 398 ff.).
view o f phronësis which suited the N E not the E E ; but A E B is itself the main evidence for 2. F o r this reason, and because the expression λ ο γ ικ α ί aperçu does not occur elsewhere
attributing such a concept to the N E (cf. above, p. 4). Schleiermacher argued that the N E in Aristotle, most commentators regard the clause as an interpolation. However, the reading
could not be by Aristotle because o f the inconsistency between the view o f virtue in the early is attested by Aspasius, and at E E 1 2 2 1 a12 Aristotle lists wisdom as a mean between cu nning
books and in book 6 ; and then went on to attribute the disputed books to the E E and simplicity.
3. Though Aristotle’s words fit A E B well, Aspasius’ commentary at this point does not:
(Schleierm acher, 1835).
52 The Aristotelian Ethics Cross-references in the Ethics 53
Finally, at the end o f book 4 we read: ‘N either is continence a It is clear that all relate happiness to three lives, the life of the politician, of
virtue, but something intermediate; it will be discussed later; now the philospher, and of the hedonist. Now of these the pleasure which is
concerned with the body and its enjoyments1 is no difficult matter to
let us talk about justice.’ G rant thought that these words had been
decide what and what kind of thing it is and how it is produced ; what we
added by an editor, after an abrupt and interrupted ending to book
need to inquire about them is not what they are, but whether they
4, ‘to give the book a seeming union with the three Eudem ian books contribute to happiness or not, and in what way they contribute, and
which were now to be grafted on’ (Grant, 1885, 94). I f there are whether some pleasures should be attached to living a noble life, and if so
independent grounds for regarding the middle books as Eudemian, whether these should; or whether these should be partaken of in some
then there may be reason for regarding the present passage as an other way while there are some other pleasures in accordance with which
interpolation; but at the present stage o f the argument it would be a the happy man may reasonably be said to live a life of pleasure and not
petitio principii to do so.1 merely a life free from pain. But these matters must be investigated later.
Taking the text o f N E 1- 4 as it stands, therefore, we can say that It has been shown by Festugière that this passage admirably fits the
it leads us to expect a later treatment o f justice as a mean, in two discussion o f pleasure at the end of the third disputed book.
parts; o f right reasoning as a virtue; and o f continence. T h is fits Though the end o f that discussion is described as having dealt with
well enough the content o f the disputed books : if one could not the question what pleasure is, the discussion is far less concerned
learn their detailed contents from these slender allusions, equally with the nature o f pleasure than the corresponding one in N E 10
one cannot claim that they contain anything that contradicts them, and is almost wholly devoted to the question o f the relationship
except for the reference to intellectual virtues as means. between pleasure and the supreme good.
T h e references foward in the early books o f the Eudemian Ethics All the questions asked here are answered in A E C : bodily
are very much fuller. pleasures contribute to happiness because the happy man needs
(1) É E 1.5, 12 16 *37 :
bodily goods so that his characteristic activities may not be impeded
( I I 53t>I7); there is a right way o f enjoying food and wine and sex,
and it is only excess in them that is blamable (i i5 4 ai6 ); the happy
S e ife t δ ε ό τ ι 6 ο ρ θ ό ς λ ό γ ο ς ε σ τ ιν ή φ ρ ό ν η σ ις, έ χ ε ι δ ε π ρ ο ς τ α ς α λ λ α ς ά ρ ε τ α ς ή
φ ρ ό ν η σ ις ώ ς η γ ε μ ο ν ικ ή τ ις . δ η λ ω θ ή σ ε τ ο .ι δ ε π ε ρ ί τ ο ύ τ ο ν , ό τ α ν κ α ί α υ τ ό ς λ ε γ η . life is one o f pleasure, not just o f freedom from pain (i i5 3 a28), but
ε κ ε ΐν ο 8 ε ύ π ο μ ν η σ τ ε ο ν , ό τ ι τ ο κ α τ α το ν όρθον λ ό γ ο ν π ρ ά τ τ ε ιν ου μ ό ν ο ν τ ο ΐς ε χ ο ν σ ι the pleasure which constitutes happiness and indeed the greatest
τ α ς τ ε λ ε ία ς α ρ ε τ α ς π ρ ο σ ν ε μ ε ι, άλλα κ α ι τ ο ΐς μ η ε χ ο υ σ ι τ α ς τ ε λ ε ία ς , π ρ ά τ τ ο ν σ ι γ α ρ
good is not bodily pleasure but is the unimpeded exercise o f virtue
ο ν τ ο ι, o>s άν ό ο ρ θ ό ς λ ό γ ο ς ύ π α γ ο ρ ε ύ σ ε ιε ν , ου μ η ν ε χ ο ν τ ε ς τ ο ν λ ό γ ο ν α π ο δ ε ικ τ ικ ό ν
α λλ ά τ ο ν δ ο ξ α σ τ ικ ό ν , ό τ ι τ ο ν τ ο μ ε ν π ρ α κ τ ε ο ν , τ ο ΰ τ ο δ ε οι) π ρ α κ τ ε ο ν . T h ere are several ( n 53bi 3)·
striking differences between A spasius’ remarks here and A E B. First, in Aristotle there is no (2) T he passage o f E E I continues thus : ‘First we must consider
suggestion o f any hegemony o f wisdom over the other virtues : there is rather a partnership.
virtue and wisdom, what the nature o f each is, and whether they, or
Secondly, though Aristotle does distinguish ( ii4 4 b27) between those who act κ α τ α τ ο ν
ο ρ θ ό ν λ ό γ ο ν and those who act μ ε τ ά τ ο υ ό ρ θ ο ν λ ό γ ο υ he does not connect this at all with a the actions which arise from them, are part o f the good life.’ T h is is
distinction between perfect and non-perfect virtue (the terms are used in A E to draw a an accurate prediction o f the structure o f the E E including the
different distinction: ii2Q b2 i ) ; nor would Aristotle agree with the distinction between the
disputed books : the account o f pleasure promised comes after the
apodeictic and doxastic logoi; for him, wisdom is the virtue o f the doxastic faculty ( ii4 4 b14).
T h is passage perhaps suggests that the N E on which Aspasius commented did have a treatment o f virtue (E E II and I I I and IV) and o f wisdom (E E
discussion o f wisdom rather than the complete lacuna conjectured earlier (p. 33), but it shows V = A E B). T h e relation between the moral virtues and the parts o f
also that it is not the one which we have as A E B . U nless we are to say that he took the forward
happiness is stated in A E A at i i29 bi8 , where it is said that legal
reference to be merely to later passages in N E 2 (e.g. n o 7 aI) and that the un-Aristotelian
remarks are his own unprompted additions. justice, which combines the virtues o f courage, temperance,
i. A possible trace o f disorder in the N E text is found in the commentary o f the meekness, etc., is concerned with the things that produce and
anonymous scholiast on this passage: ύ σ τ ε ρ ο ν δ ε ε π ι π λ έ ο ν ε ρ ε ΐ π ε ρ ί ε γ κ ρ α τ ε ία ς κ α ι
κ α ρ τ ε ρ ία ς κ α ι μ α λ α κ ία ς , μ ε τ ά τ ό ε ξ ή ς β ιβ λ ίο ν ; see above, p. 37· T h e singular is odd,
suggesting that A E C succeeded A E A in his text. B u t this in turn is contradicted by the same i . Professor D . J . Allan has pointed out to me that if Stew art’s emendation (18 9 2,220 ) τα ς
scholiast’s commentary on the text discussed above, n o 3 b33, which reads: ρ η θ ή σ ε τ α ι δ ’ σ ω μ α τ ικ ό ς α π ο λ α ύ σ ε ις is read at this point, there is a striking coincidence o f expression
ύ σ τ ε ρ ο ν π ε ρ ι α ν τ ο ν , ε ν τ ω ε κ τ ω ό π ο ν π ε ρ ι φ ρ ο ν ή σ ε ω ς λ ε γ ε ι. with A E C , ii4 8 a5.
54 The Aristotelian Ethics Cross-references in the Ethics 55
preserve happiness and its parts. T h e intellectual virtues o f mark which specifies the mean is merely repeated, but not
practical and philosophical wisdom are described in A E B as answered, in bo okB (i i3 8 b22). For an answer we have to wait until
producing happiness, the latter as ‘a part o f total virtue’ (i i44a5). E E V III. (The question is not raised in the undisputed N E books.)
W hile the promise o f E E I is fulfilled to the letter in the disputed (7) In his discussion o f voluntariness Aristotle commends the
books, it is impossible to see a fulfilment for it within the legal distinction o f events into voluntary, involuntary, and
undisputed Eudem ian books; those therefore, like Schächer, who premeditated ; even if the distinction is not exact, he says, they do
maintain that the E E never contained more than five books, have to grasp the truth in a certain manner. ‘B u t we will speak about these
regard the whole passage as a false start (Schächer, 1940, 66). matters in our inquiry into justice’ (i2 2 7 a2). This promise is kept
(3) A t 1 2 1 8b 1 2 we are told that the good which is the final end o f when in chapter 8 o f A E A a detailed treatment o f voluntariness
things achievable by man is the good which is the object o f the replaces the simple threefold legal division with a many-branched
discipline or disposition which governs all others. ‘T h is ’, Aristotle schema which preserves its main lines while allowing for more
goes on, ‘is politics and economics and wisdom. F o r these complicated cases. T h e exactness o f this forward reference is
dispositions differ from others in being o f this kind ; whether they striking: it is by no means something to be taken for granted that a
differ at all from each other will be explained later.’ 1 treatise on justice will contain a discussion o f the various degrees o f
T h e later explanation is given, as promised, at 1 i 4 i b23 in A E B . mens re a.
W isdom, politics, and economics are all the same disposition, but (8) In the last chapter o f E E II Aristotle asks whether what
they connote different exercises o f that disposition, concerned with virtue does is to safeguard one’s choice o f end or one’s reasoning.
different subject matters. ‘W isdom’ in the broad sense is the general Som e people, he says, think the latter ‘but what does this is
name for the disposition: it may be exercised about an individual’s continence: it is this which keeps the reasoning uncorrupt. Virtue
own affairs (that is wisdom in the narrow sense) or it may be and continence are different. We must speak later about them,
concerned with a wider scope. I f it concerns his household, then it is because this is the reason why those who do so regard virtue as that
economics ; if it concerns the state, then it is politics. Any exercise o f which makes the reasoning right’ ( i2 2 7 bi5 ff) . It is not difficult to
wisdom concerning the state is politics in the broad sense; but the find in the disputed books a treatment o f the difference between
exercise of wisdom in this way may be expressed in legislative, continence and virtue— that is one o f the major themes o f A E C —
executive, or judicial decisions ; only in the two latter cases can it be but it is not at first obvious how what is here said about the relation
called ‘ political’ in the narrower sense.2 between continence and reasoning fits what is said later. On the one
(4) T h e E E , like the N E , after introducing the notion o f right hand, we are told in A E B that it is temperance which preserves the
reasoning into the concept o f virtue, promises a further con wisdom which is identical with right reasoning ( 1 1 40b1 2 ; 1 144b28) ;
sideration o f its nature. ‘What right reasoning is, and what is the yet temperance is the virtue which in A E C is distinguished from
mark to which we must look in order to specify the mean, must be continence ( i i5 2 a i 3). On the other hand, correctness o f reasoning
examined later’ we are told at i2 2 2 b8 .3 A s with the N E reference, seems to extend beyond the bounds o f virtue not merely to the
the promise to examine what right reasoning is is fulfilled in the boundary between continence and incontinence, but to the
examination o f wisdom in A E B ; but the second question about the boundary between incontinence and vice; for we are told in A E C
that the incontinent man has a good prohairesis ( i i 5 2 ai7) which
I. ο ) θ τ ε τ ο ίιτ ' αν ε ιη α ν τ ο τ ο α γ α θ ό ν το τ έ λ ο ς τ ώ ν ά ν θ ρ ώ π ψ π ρ α κ τ ώ ν . τ ο ΰ τ ο δ ’ έ σ τ ί must include right reasoning ( 11 3 9 “24) ; and indeed the incontinent
τ ο νττο τ η ν κ υ ρ ία ν π α σ ώ ν , α υ τ ή δ ’ ε σ τ ι π ο λ ιτ ικ ή κ α ί ο ικ ο ν ο μ ικ ή κ α ι φ ρ ό ν η σ ις.
man is often described precisely as the man who does not stick to his
δ ια φ έ ρ ο υ σ ι γ ά ρ α ί τ α ι a i ε ξ ε ι ς π ρ ο ς τ α ς α λ λ α ς τ ώ τ ο ια ν τ α ι ε ΐν α ι■ π ρ ο ς δ 'α λ λ ή λ α ς ε ΐ
τ ι δ ια φ ε ρ ο ν σ ιν , ύ σ τ ε ρ ο ν λ ε κ τ έ ο ν . reasoning ( i i 5 i a20, a32, b25).
2 · A schematic tree o f the different branches o f wisdom is given by von Fragstein (1974, T h e inconsistency between the E E and the A E is only an
232 f.).^
apparent one : indeed, if it were a genuine inconsistency it would be
3. τ ι ς δ ο ο ρ θ ο ς Aoyoç, κ α ι π ρ ο ς τ ι να δ ε ι ορ ον α π ο β λ έ π ο ν τ α ς λ έ γ ε ιν τ ο μ έ σ ο ν ,
ύ σ τ ε ρ ο ν έ π ισ κ ε π τ έ ο ν . internal to the A E itself. It is true that there are incontinent men
56 The Aristotelian Ethics Cross-references in the Ethics 57
whose reasoning is perfectly correct: but besides these weak (10) At i2 3 4 a28 Aristotle says that ‘as will be said later, every
incontinents there are the headstrong incontinents whose passions virtue is found in one way by nature and in another way with
are so strong that they are lead astray by passion without wisdom’ . T h is fits admirably the account o f natural virtue in A E B
deliberating at all ( i i 5 0 big ff.). Consequently it is only o f the ( ii4 4 bi - i 7 ) which concludes ‘in the part o f the soul concerned with
continent that it can be said without exception that their reasoning character there are two kinds o f virtue, natural virtue and full-
is safe. On the other hand the difference between the continent man blooded virtue, and full-blooded virtue does not come about
and the temperate man is not a matter o f reasoning but o f feeling : without wisdom’ .
the continent man has unruly desires, but unlike the incontinent he ( 1 1) Like N E 4, E E IV ends with an announcement o f a
is not led astray by them ; whereas the temperate man does not feel treatment o f justice (i2 3 4 bi4).
unruly desires at all ( i i 5 2 ai). None the less continence resembles All these eleven forward-looking allusions, then, fit without
temperance, and can be called temperance by analogy ( i i 5 i b32); difficulty the disputed books ; but there are other passages which are
and it is surely that which Aristotle was doing when, in reporting a not so easy to locate as correspondences to previous announce
popular etymology o f the Greek word for ‘temperance’ (sophrosyne ) ments. This was pointed out by Spengel (18 4 1, 489 ff.): Aristotle,
he said that it was what preserves wisdom (sözei ten phronesin). T h e having said that beasts cannot be called happy at 1217*26 , goes on:
imprecision o f this passage is corrected by the later distinction: ‘ N or can anything else whose name does not indicate that it shares
what the temperate man is here contrasted with is not the continent in its nature in something divine; it is by some other mode of
man but the man ‘corrupted by pleasure and pain’— the participation in goods that some o f them have better lives than
intemperate man whose first principles are perverted ( ii4 0 b20). others. But that this is so must be considered later.’ Dirlm eier may
(9) At the conclusion o f the discussion o f intemperance in E E be right to say that Aristotle here allows for the possession of
I I I Aristotle promises to treat with greater exactitude ‘o f the class o f happiness by divine and superhuman entities like the aether (1969,
pleasures’ when he comes to deal with continence and incontinence. 19 1); but it is not this that he promises to consider later. It is more
A E C o f course contains a detailed treatment o f pleasure in naturally taken to be the applicability o f happiness, or sharing
conjunction with a treatment o f continence and incontinence: but goodness, to animals. That is how von Fragstein takes it (1974, 31)
some scholars have complained that despite the full discussion the and refers forward to A E C, 1 14 8 °i5 ff. : ‘ some things are pleasant
class o f pleasures with which intemperance is concerned (those of by nature, and o f these some . . . in accordance with the species o f
touch, o f a certain kind) is no more precisely defined than in E E I I I animals and men. ’ But a passage which is more obviously concerned
(i2 3 0 b2 5 - i 2 3 i a25). But though the class o f pleasures was with the relation between animals and the supreme good is A E C,
precisely defined in E E I I I , A E C shows in chapter 4 how this class H 53b25 ff·
can be viewed as a species o f various genera o f pleasures according
And indeed the fact that all things, both brutes and men, pursue pleasure
to particular classifications, in chapter 5 how the class itself is a
is an indication of its being somehow the chief good. . . . But since no one
genus o f which natural and perverse pleasures are species, and in nature or state either is or is thought the best for all, neither do all pursue
chapter 7 how even in the case o f natural and necessary pleasures the same pleasure; yet all pursue pleasure. And perhaps they actually
one must distinguish between excessive and defective instances.
These passages come not in the section o f A E C devoted exclusively intermediate pleasures o f 1 148*25; (b) the pleasure-making things that are choiceworthy in
to pleasure, but, as the preannouncement stated, in the section themselves (1 i4 7 b25) but admit o f excess = the naturally choiceworthy pleasures ( 1 148a24),
namely honour, victory, riches; and (c) the things naturally unchoiceworthy ( 1 148a24), not
devoted to continence and incontinence.1 naturally pleasant but only pleasant through innate or acquired corruption ( 1 148ai6 fF.). This
seems the best interpretation o f a difficult passage. C f. also D . J . Allan, 19 7 1. T hese divisions
i . T h e objection, first made by Spengel, was forcefully taken up again by K ap p ( 19 12 ,2 3 ) . are necessary to complete the account o f the relationship between temperance and bodily
T h e best treatment o f the passage is by Festugière (1946, xx iv -ix , lxx), who shows that the pleasures : because the vice o f intemperance concerns only excesses o f natural pleasures, the
subdivisions o f objects o f pleasure in A E C 7 give three classes: (a) necessary, bodily disposition corresponding to unnatural pleasures o f touch and taste is not vice but bestiality
pleasures which provide the field o f temperance and intemperance ( i i 4 7 b24 ff.) = the (ii48b25- ii49ai).
58 The Aristotelian Ethics Cross-references in the Ethics 59
pursue not the pleasure they think they pursue nor that which they would the E E was ever concluded by Aristotle according to his plan, and
say they pursue, but the same pleasure; for all things have by nature did not contain the A E as we have them, then it must have
something divine in them. contained something very similar indeed.
o f the E E and A E . M edical analogies o f course occur so frequently ii37ai4-27 T h e doctor’s knowledge compared with knowledge of
justice.
1I3^a3° Comparison between the mean in justice and in medicine.
1 . Burnet, ad loc., recognizes that the reference m ust be to the use o f έ ν δ ο ξ α . Others take AE B: ii38b30-2 ‘What right reasoning prescribes’ no more helpful than
τ ή ν α ύ τ ή ν μ έ θ ο δ ο ν τ ο ΐς π ρ ο ε ιρ η μ έ ν ο ι? to mean n o t‘the method described in the previous ‘what medicine prescribes’ .
discussion’ but ‘the method used in the previous discussion’ . T h e difference is not o f great ii4ia22-33 Ethics, like medicine, is species-oriented.
importance : the introduction o f justice here parallels the introduction o f courage at 1228 a28 1 141 bi 4 ff. Value o f knowledge o f particular in medicine.
in the E E just as it fits the description o f method in the E E . W hen E E I says that we must use H43b2i~32 Apparent uselessness o f medicine and wisdom compared :
the evidence o f the φ α ιν ό μ ε ν α it means o f course the opinions o f others, not ‘the facts’ ; medicine neither sufficient nor necessary for being healthy.
see Owen, 19 6 1. i i4 4a4 Wisdom comparable not with medicine but with health.
2. In the E E ύ π ο κ ε ίσ θ ω and ν π ό κ ε ιτ α ι are used in the technical sense at I 2 i8 b3 7 ; U45a7 Relation o f wisdom to philosophy is like that o f medicine to
I 2 I 9 a IO, a2 9 , b2 8 ; I 2 2 0 a7 , a2 2 ; I 2 2 2 a6 ; I 2 2 3 b3 0 ; I 2 2 7 b2 5 · health.
64 The Aristotelian Ethics The S tyle o f the Treatise on Justice 65
Aristotle’s final methodological remark before proceeding to would be a more accurate translation than ‘absolute goods’
the examination o f justice is that if one o f two opposites has a ‘ unqualified goods’ ‘ intrinsic goods’ ‘things absolutely good’ .1
plurality o f senses, so does the other. T h is is applied in the case o f T h e prima-facie goods are mentioned often in the E E and in the
justice : there is more than one kind o f injustice, therefore there is A E but never in the N E .2 For instance, we are told that ‘ Some
more than one kind o f justice. T h e practice o f distinguishing goods are good in the abstract and others good for individuals
between the senses o f words, and asking in how many different without being good in the abstract; and the same things are good in
senses words are used, is, along with the technical vocabulary which the abstract as are pleasant in the abstract’ (i2 3 5 b32 ff.), and a
goes with the practice, more characteristic o f the E E than o f the comparison is developed between those things which are good and
N E ; the former, it has long been observed, is more interested in the pleasant to a healthy body and those things which are advantageous
technicalities o f logic ( i i2 g a2 ^ -6 ) .1 and pleasant to an invalid. T h e final book o f the E E contains a long
H aving distinguished between ‘ just’ as ‘equal’ and ‘just’ as discussion o f the prima-facie goods; it uses the expression ‘ natural
‘according to the law ’, Aristotle defines the subject matter o f the goods’ (physei agatha) as synonymous with ‘ goods in the abstract’
first sort o f justice and injustice as ‘ goods subject to good and bad (haplös agatha). (T he synonymy is most obvious from the passage
fortune, goods which taken without qualification (haplös) are always I249a8 - i7 , where also ‘exterior goods’ is used as third synonymous
good, but for a given individual are not always good’ (i i29 bi-6 ). H e expression.) There we are told that if a man is good then the things
at once abbreviates this description to the expression ‘ta haplös which are good in the abstract are good for him as an individual also
agatha', literally ‘ the without-qualification goods’ . T h is does not (i2 4 8 b27); and we are given a list o f these prima-facie goods:
mean that they are goods, no matter what may happen, or goods, honour, wealth, the good qualities o f body, good fortune, and
come what may ; just the opposite in fact : it means goods which are pow er.3 These, we are told, are good by nature but can be injurious
good in the absence o f special circumstances to upset their to those o f bad character.4 These goods are contrasted with virtuous
goodness. From the contexts in which Aristotle uses the expression character and virtuous activity (i2 4 8 b37) which are the finest class
it is clear that ‘prima-facie goods’ or ‘things good in the abstract’ o f goods and cannot be misused. A man o f common goodness may
pursue the virtues for the sake o f the natural goods; the man of
A E C: i i 47 a5-8 Exam ple o f medical syllogism. supreme virtue (the kalos kagathos) chooses natural goods for the
i i S ° b32 - 5 Com parison between evil ethical states and types o f disease. sake o f virtue (i2 4 9 a6 - i5 ) .5
E E V II: I 2 35 b32_a7 D istinction between goods simpliciter and others illustrated T h e prima-facie goods, then, play an important part in the
from medicine.
i2 3 0 ai8 -2 4 H ealth taken as prim e analogate for analogy. systematic development o f the theory o f virtue and goodness in the
I2 3 7 b22 W ish not enough for health nor for friendship.
E E V III: I248bi4 Virtue and health wholes consisting o f parts. 1. Rackham (Loeb E E ) and Ross (Oxford N E ) translate ‘things absolutely good’ . G ran t’s
! 248b34 Wealth no more benefits vicious than full diet an invalid. ‘W hat are abstractedly goods’ is more correct, but seems a little clumsy.
I2 38 b2 2 -5 T h e standard by which the doctor judges what is healthy 2. T h e nearest to the concept is at i i 5 7 b27 S o k €Γ γ ά ρ φ ιλ η τ ό ν μ ε ν κ α ί a ip e r o v το
ά·7τλώ5 α γ α θ ό ν η ή δ υ . T h e N E does not talk o f a class o f α π λ ώ ? ά γ α θ ά nor does the notion
I. T h e following lexical facts give an indication o f this. T h e expression π λ ε ο ν α χ ώ ς
λ έ γ ε σ β α ι occurs only once in the N E ( i i 2 5 bi4). In the much briefer A E it occurs twice o f α π λ ώ ς ά γα θ'ον here mentioned play any part in the systematic development o f the N E
(112 9 * 2 4 ; H 4 2 bi7 ) and in t h e £ £ five times (i2 3 0 bg ; I2 34 bi9 and 20; 1236 *7; i2 4 7 b28). Its theory o f goodness.
opposite μ ο ν α χ ώ ί λ ί γ ΐ σ θ α ι occurs only in the E E (i2 3 4 bi9). 3. T h e list which the Scholiast gives in commentary on the passage o f A E A — health,
A synonymous expression, ττολλα χώ ς λ έ γ ε α θ α ι occurs in the A E ( i i 3 6 b29) and in the strength, beauty, good bodily condition, wealth, reputation, and political power—
E E ( i 2 i 7 b27) but not in the N E . corresponds very accurately to the list in the E E , if we take the first four as instances of
A particular instance o f plurality o f senses, 8 ιχ ώ ς λ έ γ ε σ θ α ι, occurs twice in the A E ‘good qualities o f body’ (C A G xx.208).
4. T h is is a different point from that made in the N E that goods can be dam aging: for one
(1 i4 0 b3 1 ; H 5 2 b27) and twice in the E E (12 17 * 3 6 ; I228 bi8) but never in the N E . Iσ α χ ϋ ς
o f the examples given there (i094bi9) that men have died because o f their courage is an
λ έ γ ε σ θ α ι occurs once in each o f the N E (1096*24) and the E E ( ΐ2 ΐ-;°ζ η ) . T h e question
example o f a good which is not just prima facie good but κ α λ ό ν according to the E E
π ο σ α χ ω ς λ έ γ ε τ α ι is raised only in the A E ( i i 2 9 a3 i).
Again, ό μ ο ν ΰ μ ω ς and its cognates, which occur three times ill . 11·. A alone, and from time classification.
5. Other places where the E E uses the terminology are I228 bi8 ; i2 3 0 b34; i2 3 8 a3 ff., b5 ;
to time in the E E (e.g. 12 3 6 * 17 , b25) appear only once in the whole N E (i096b27).
σ υ ν ώ ν υ μ ο ς occurs only in the A E (113 0 *3 3 ). See below, p. 140.
66 The Aristotelian Ethics The Style o f the Treatise on Justice 67
E E . But, as Grant showed long ago, the notion is widespread T o show that the law covers the whole field o f conduct covered
throughout the disputed books also: at i i3 4 aio we hear o f ‘what is by the virtues already considered, Aristotle says :
useful in the abstract’ ; at i i4 3 b3 we are told that the just ruler does
The law commands the doing of the acts of a brave man (e.g. not leaving
not take more than his share o f prima-facie goods for himself, and at
one’s post, not running away, and not throwing away one’s arms) and
i r 37a20 we are told that justice exists among people who share in those of a temperate man (e.g. not committing adultery, not being wanton)
prima-facie goods. T h e distinction between goods in the abstract and those of a meek man (e.g. not committing assault or defamation) and
and goods for individuals is made the basis o f the rebuttal o f similarly with regard to other virtues. ( 122t)biy- 23)
Speusippus’ attack on pleasure in Book A E C (i i5 2 b2 5 ).1 We hear
F o r our present purposes the interest o f this list lies in the selection
also o f things which are bad in the abstract ( i i2 9 b8 ; i i5 3 b2).2 This
o f these three particular virtues as paradigms. It is in the E E that
terminology seems still to be one o f the strongest indications o f a
these three virtues have pride o f place (E E I I I , chapters i, 2, 3 ; also
link between the disputed books and the Eudemian Ethics.
i i 2 i ai5 - 2 3 in book II). In the N E both the preliminary list
Having used the notion o f prima-facie goods to distinguish the
(1 i0 7a2 8 -i io8bio) and in the full-fledged treatment o f the virtues
subject matter o f justice in the sense o f equality or fairness, the
in books 3 and 4 meekness appears well down the list; and so far
justice which contrasts with greediness (pleonexia), the taking o f
from being a paradigm it is twice described as a virtue that is strictly
overlarge shares, Aristotle turns to the wider sense o f justice in
nameless ( n o 8 a5, i i 2 5 b26).1
which it is obedience to law. In this sense, he says, ‘ we call “ just”
O f the justice that is law-abidingness Aristotle goes on to say that
those things that produce and preserve happiness and its com
it is (1) perfect virtue, (2) the use o f perfect virtue, (3) using one’s
ponents for the political society’ (i i2 g bi 8). Once again a tiny point
virtue in regard to others, (4) not a part o f virtue, but total virtue.
o f terminology gives an indication o f the rightful place o f the
E very one o f these expressions is a Eudemian turn o f phrase, as I
treatise on justice. In speaking o f ‘happiness and its components'
shall now proceed to show.
Aristotle is here aligning him self with the Eudem ian view,
At E E 12 19 ^ 8 , when Aristotle has reached his definition o f
according to which happiness is activity in accordance with all the
happiness, we read :
virtues, moral and intellectual, so that each type o f virtuous activity
is a component o f total happiness, and against the Nicomachean Since happiness, we saw, is something perfect, and life can be either
view o f happiness as the activity o f one supreme virtue.3 perfect or imperfect, and virtue likewise (for there is total virtue and
partial virtue) and since the activity of imperfect things is imperfect, it
1. See Festugière, 1946, L X X .
follows that happiness is the activity of perfect life in accordance with
2. G rant notes ‘ It is added that “ men pray for these and seek after them, but they should
not; they should pray that the absolute goods may be goods to them individually, and that perfect virtue.
they should choose what is good for themselves.” T h is is in the same style with Eth. Eud.
Perfect virtue, then, is the totality composed o f the individual
V I I .x ii.1 7 : το ζ η τ ΐ ΐ ν κ α ι eϋ χ ε σ θ α ι π ο λ λ ο ύ ς φ ίλ ο υ ς. But to say what men “ ought to pray
fo r” is not after the manner o f A ristotle.’ Prayer is mentioned in the N E , once, in the not very virtues as parts: it is what is called later in the E E kalokagathia
serious context o f the man who prayed to have a crane’s gullet ( 1118 * 3 2 ) ; but the idea that the (i2 4 9 ai8). Legal justice, he here says, is identical with this perfect
desirable situation is to have the prim a-facie goods be good for oneself individually parallels
E E 12 4 9 *13 even more strikingly than the Eudem ian passage quoted by Grant.
virtue: or, to be more accurate, it is a particular aspect o f perfect
3. O f the twenty uses o f μ έ ρ ο ς , μ ό ρ ιο ν in the N E six are for parts o f physical objects, four o f virtue, its utilization in regard to others ( 1 i29 a26, 31). T h e ability to
parts o f the soul, four o f parts o f motion, two o f parts o f a community, three in expressions make use o f virtue in regard to others is itself added perfection; it
like κ α τ ά μ έ ρ ο ς . T h e N E nowhere speaks o f parts or components o f virtue, vice, or happiness.
T h e nearest it comes to doing so is at 114 0 * 14 (parts o f the common good) and u 8 0 b30
gives another sense in which justice is ‘ perfect’ virtue ( ii2 9 b3 i) .2
(science o f law-giving part o f science o f politics). T h e A E on the other hand constantly speaks
o f parts o f virtues and vices ( 1 13 0 a8, “ 14, *23, *34, bi 1 ; 1 1 3 1 * 1 ; 1144*5) as well as o f sciences 1. See above p. 23 and below p. 232.
( I I 4 I ’ I 3 i I I 43 a3 )· I " the E E we often hear that happiness is a whole o f parts ( i 2 i 4 b5 ; b28; 2. Assuming that the ό'τι in line 1 129b3 i is correct (despite the disagreement o f K b ) the
12 16 *4 0 ; I 2 i9 bi4 )an d parts o f virtue(s) are mentioned at I 2 i6 b6; 1219 *38 , “2 3 ; 12 2 0 * 3 ,“ 14 ; reason there given for the perfection o f justice is not the same as the one presupposed by lines
parts o f vices at 12 3 1* 2 2 ; 12 3 2 *13 . T h e E E and the A E are altogether happier than the N E 2 5 - 3 1 ; cf. the Scholiast, C A G xx.20 9 .31-210 .8 which makes much more sense than the heavy
about applying mereology to abstract objects. See below, p. 126. weather made o f this passage by most commentators, cf. G au th ier-Jo lif (19 59 , 342).
68 The Aristotelian Ethics The Style oj the Treatise on Justice 69
T h e notion o f the use or utilization o f virtue is a Eudem ian one. chapter o f the A E in the same detail would be tedious and
T h e Greek word for ‘use’ is employed in the N E overwhelmingly unrewarding: each chapter, however, can be made to yield the same
for the use o f physical objects (eighteen uses out o f a total o f twenty- minute indications o f the original home o f the disputed books.
five). W e hear o f one science using another (io94b4) and there are Rather than proceed with a line-by-line examination of the style o f
references to our making use o f our senses ( n o 3 a3o ; n i 8 a28) of the A E , I shall present the evidence in a cruder but briefer manner
pleasures ( n o 5 ai2) o f irony ( i i 2 7 b3o). Only once in the whole in the form o f tables exhibiting the results o f vocabulary counting.
work do we hear o f the use of virtue ( i i 7 9 b3). In the Eudemian T h e presentation will be cruder in this way in that it will not permit
Ethics, on the other hand, the notion o f the use o f a hexis is account to be taken o f differences between senses o f the same lexical
ubiquitous ai6 , 18, 2 5 ; I 2 i9 b2, b4 ; i220 a3 3 ; i 225bi 2 ; item ; but it will have the advantage o f providing data sufficiently
1 22 7a23) and there are whole sections devoted to the topic o f the use copious and sufficiently well defined to enable statistical methods of
and abuse o f sciences and virtues (i2 7 ai ff. ; i246a26~b36).1 argumentation to be applied. T h e crudity o f the classification will
It has already been shown that the use o f the part-whole do no harm : it may prevent the detection o f stylistic differences
terminology in connection with virtue is another Eudemian between texts where a finer analysis might enable us to discri
characteristic. T h u s the chapter ends, as it began, with Eudemian m inate; but it cannot have the effect o f producing imaginary
echoes.2 distinctions where there are none in reality. If, then, by vocabulary
T h e close examination, then, o f this first chapter o f the disputed counting and statistical analysis o f a simple kind there appear clear
books, has revealed that almost every sentence contains tell-tale differences between the A E and one of its competing environments,
indications that the original environment o f the books is the it is unlikely that later more refined analysis will overturn any
Eudem ian and not the Nicomachean ethics. T o go through every conclusions thus reached.
1. In the preliminary N E discussion o f happiness we read: ‘ It makes no little difference
whether we place the suprem e good in possession or in use, in state or in activity’ ( ioq8b32 f.).
I f these two pairs o f alternatives are meant to mark a single contrast we have here the
terminologies o f the E E and o f the remainder o f the N E brought together. An act o f a
particular virtue in the E E will be described as an activity o f that virtue ( τ ή ς α ρ ε τ ή ς
ε ν έ ρ γ ε ια ) or as a use o f that virtue (passages listed in text). T h e N E 's preferred expression is
activity in accordance with virtue’ (έ ν ε ρ γ ε ια κ α τ ’ α ρ ε τ ή ν ). T his expression is o f very
frequent occurrence in the N E (e.g. 1098*7, " 1 7 ; i0Q9b26; U 7 4 bi8 ; i i 7 0 b26; H 7 7 a9 ~ i 2 , 18,
24) and does not occur once in the E E . (We do however hear o f activity in accordance with
passion, I2 2 0 bi7 , I2 2 3 a39.) Whereas in t h e A ^ normally a person is said to act in accordance
with his hexis, in the E E his life, or his hexis acts ( 1 2 1 8 * 3 1-3 , 38, b2 ; 1220*8; cf. N E
But altogether the N E is much fonder o f the terminology o f activity ( ε ν ε ρ γ ε ια ε ν ε ρ γ ε ΐν )
than the E E : see below, p. 142.
H ere as usual the A E follows the E E usage. T h e activity words occur only seventeen times,
thirteen o f these in the section on pleasure in A E C. In the m ajority o f cases the subject o f the
ε ν ε ρ γ ε ια is a έ ξ ις , not a person (1144=6; 114 7 * 7 ; I I 52b3 5 ; 115 3 * 14 , bio ; ε ν έ ρ γ ε ια o f life,
1 r 54 a7 )· T h e expression ε ν έ ρ γ ε ια κ α τ ’ α ρ ε τ ή ν is not used. Only in one place is there an
echo o f the Nicomachean term inology: U 3 0 ai 7 : κ α τ ά μ έ ν γ α ρ τ α ς ά λ λ α ς μ ο χ θ η ρ ία ς ό
ε ν ε ρ γ ώ ν α δ ικ ε ί.
2. T h e sentence έ σ τ ι μ έ ν γ α ρ ή α υ τ ή , το δ ’ ε ίν α ι ο ν το α υ τ ό ( i i 3 0 ai2 ) does not give
much help for our investigation. T h e only other sentence o f this form in the ethical writings is
also in the disputed books r I 4 i b24. Though the precise way in which the sentence should be
read is not clear, its general sense is clear from the many other passages in Aristotelian
writings where it occurs (listed by G au th ier-Jolif, 19 5 9 ,34 3). ju stice and virtue are the same
hexis, considered in two different relationships ; justice and virtue are the same thing, though
what it is to be just is not what it is to be virtuous ; in a m ore modern terminology, ‘ justice’ and
‘virtue’ have the same reference but not the same sense.
Particles and Connectives in the Aristotelian Ethics 71
frequency o f most nouns, verbs, and adjectives is; moreover they
occur with sufficient frequency and regularity to enable standard
statistical techniques to be used to describe and draw inferences
CHAPTER 4
from their distribution. Indeed, as we shall see, about one-quarter
o f any Aristotelian text consists o f uses o f these comparatively
Particles and Connectives in the insignificant and topic-neutral words. M oreover, such studies as
have been carried out suggests that the use o f particles and
Aristotelian Ethics connectives in Greek prose may vary in characteristic ways from
author to author.
How far authors are consistent in speech habits such as
T h e problem presented by the books common to the Nicomachean vocabulary choice is a matter o f keen debate : fortunately the scope
Ethics and the Eudemian Ethics is one which particularly invites o f the present inquiry does not necessitate any participation in the
investigation by stylistic methods. It is natural to inquire whether debate. N either here nor elsewhere am I questioning the consensus
the three books which make a double appearance in the manuscript o f scholars that both the Eudemian and the Nicomachean Ethics are
tradition, as books 5, 6, and 7 o f the Nicomachean Ethics and as genuine works o f Aristotle; nor am I putting forward any
books IV , V , and V I o f the Eudemian Ethics, were originally written hypothesis about the fixity or fluidity o f his stylistic habits. I am
for one context rather than another; and one obvious way of merely attempting to discover and compare the regularities to be
undertaking such an inquiry would be to ascertain whether in style observed within the undisputed Nicomachean books, within the
the common books show a greater similarity to one or other o f the undisputed Eudem ian books, and within the books which are
contexts in which they appear. Surprisingly, although scores o f disputed between the two. B y seeing how far the regularities in
Aristotelian scholars have written about the problem o f the the disputed books resemble the regularities in each o f the treatises,
common books and have pronounced confidently on their prov we may hope to obtain an indication o f the context to which they
enance, no systematic study o f their style or comparison with their originally belonged without, for present purposes, making any
Nicomachean and Eudem ian contexts has ever been m ade.1 T he assumptions about the chronological order o f the N E and E E , or the
present chapter contains the beginning of such a study : it confines temporal distance which separates them, or the process by which
itself to a single aspect o f style, vocabulary choice; and within books belonging in one context became included also in another
vocabulary choice it concentrates upon a small fragment of context. Speculation on all these matters should follow, and not
vocabulary, namely the use o f particles and connectives. E ven in precede, an unpartisan study o f the texts o f the disputed books and
such a small area, it will be found, it is possible to detect very their rival contexts.
striking differences between the degrees to which the common Though it is surprising that scholars have so frequently
books resemble the competing environments offered by tradition. attempted to settle the provenance o f the A E without a serious
Particles and connectives offer a number o f advantages as a study o f their style,1 it is not surprising that the particular features
subject o f stylistic study. Students o f Greek style have a o f style studied in the present chapter should not have been
considerable advantage over students o f English style in that Greek investigated. T h e counting o f particles and connectives by hand is a
is rich in brief words which express nuances o f thought-connection tedious and inaccurate business: the availability o f machine-
and mood which in English are more likely to be expressed by the readable texts o f Aristotle, o f computer programs for word-counts
tone o f voice or the tilt o f the eyebrow. T h e frequency o f such i . An honourable exception here should be made for S ir Alexander G ran t’s edition o f the
particles is not affected by variations in subject matter as the Nicomachean Ethics (London, 1885), which contains an excellent and almost comprehensive
index verborum— the work o f John K eb le— and many judicious observations on particular
i . T h e history o f scholarly discussion o f the disputed books can be conveniently studied in features o f the style o f the common books (which he believed to belong to the Eudemian
Schächer, 1940, and in R ow e, 19 7 1, chapter I I I . Ethics, then regarded as Eudem us’ work).
72 The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 73
and concordances, and the development of statistical theory and the final books. In the books in which it rarely appears, Eucken said,
techniques during recent decades make stylometric studies incom it could be said to be replaced by dl\ indeed, in books 8, 9, and 10 de
parably easier to perform, and potentially more fruitful in outcome, is used with a frequency which marks these o ff from every other
than they were in the 1860s when R u d o lf Eucken wrote his work o f Aristotle’s. K a i de and kai de are used as conjunctions in the
dissertation De Aristotelis dicendi ratione (Göttingen, 1866) which, final books with an unparalleled frequency: in book 9 the former
to the best o f my knowledge, is the only previous study o f the expression occurs eleven times and the latter nine times ; whereas in
Aristotelian vocabulary which is the subject matter of the present the three books of the A E they occur only five and two times
chapter. respectively. T h e use o f te also, particularly when used without any
Eucken studied Aristotle’s use o f particles in the Metaphysics, other conjunction, is a characteristic o f the final books; a mark,
Physics, de Anim a, N E , Politics, Rhetoric, and Poetics. H e wrote at a Eucken says, o f their quasi-poetical character. Te . . . te \s to be
time when the almost universal opinion o f scholars held the E E to found in the first and final books, not in the A E ; kaitoi, mentoi, and
be the work o f Aristotle’s disciple Eudem us; but he studied the alia men are more frequent in the A E than elsewhere; whereas ou
usage o f that work, along with Theophrastus’ Characters and a men, occurring once only in the A E , occurs in the third, fourth and
number o f other pseudo-Aristotelian works, in order to investigate tenth more frequently than in any other work o f Aristotle.
how far Aristotle’s style differed from that o f other members of his T h e conclusions which Eucken drew from these striking
school. His work was divided into eight chapters, each devoted to a observations were modest. Unusual though the stylistic features o f
single particle or group o f particles ; his conclusion was that there the later books o f the N E were, there was no reason to regard them
was an astonishing variety o f usages and constructions to be found as alien to the philosopher’s style : it was simply a matter o f particles
in the Aristotelian treatises, which he hoped might be someday o f frequent elsewhere being rare here, and usages elsewhere rare being
use in settling the order o f their composition. O f most interest in here frequent. T h e most that should be said was that the ten books
our present context is an appendix: ‘Excursus de discrimine of the Ethics were written at different times and only later brought
dicendi rationis, quod in libris Ethicis invenitur.’ together. No conclusion could be drawn about the disputed books,
T h e individual books o f the Ethics, he maintained, differed so because the style o f the Eudemian Ethics differed hardly at all from
much from each other in particle use that they could hardly have that o f genuine works o f Aristotle ; there was certainly no reason to
been written in a single stint by Aristotle. T h ey fell into four deny the disputed books to Aristotle once one accepted that the N E
groups: books 8, 9, and 10 resemble each other closely; furthest did not come from Aristotle in the order in which we now have it.1
from them in style are books 5, 6, and 7, which again form a unified It was a pity that Eucken, having noticed striking differences
group. T h e early books stand between, with the first two between the A E and other books in the Ethics, did not compare the
resembling 5, 6, and 7 more than they resemble 9 and 10, and the A E features with the E E to see whether they were not actually
third and fourth most resembling the final books. T h e final books closer to the E E in those points in which they stood out. H e was
are written in a fluent and elegant style, less argumentative than deterred from doing so by the prevailing scholarly dogma o f the
Aristotle’s usual manner : thus epei, which is very common in all time, that to attribute the A E to the Eudemian Ethics would be to
Aristotle’s other works, is rarely found (once in 8, twice in 9, once in deny that they were genuinely Aristotelian. T h e success o f W.
10) ; in book 7 it occurs twenty-five times and nearly as commonly in Jaeger in the present century in convincing scholars that the
books 5 and 6. In the early books, it ranges between six occurrences Eudemian Ethics are genuine— though not based on any close study
in book 2 and one in book 4. Sim ilarly koste : four times in book 8, o f the evidence comparable to Eucken’s— has enabled us to look at
four in book 9, five in book 10, contrasts with twenty-two times in
book 7, twenty-three in book 5, and seventeen in book 6. Books 1 i. ‘Quod si statuimus quam facile fieri potuerit, ut libri V , V I, V II et Ethicis Nicomacheis
et Eudemeis insererentur apparet, utri autem scripto magis apti sunt (de qua re homines docti
and 2, he said, to some extent resembled the middle books (though valde dissentiunt) ex dicendi ratione vix dijudicari potest, quia Ethicorum Eudemeorum
hoste was much less frequent in them) and books 3 and 4 resembled sermo fere non distare videtur ab Aristotelis ipsius scriptis’ (op. cit. 77).
The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 75
Occurrences o f thirty-six Particles and
T a b l e 4 .1. the question with a more open mind. W e shall see that even in the
Connectives in the NE, EE, and AE narrow field o f particle choice there is a great deal o f evidence which
confirms Eucken’s observation that the A E stands out from the rest
o f the Nicomachean Ethics. We shall find also that the Eudemian
NE AE EE
Ethics are not as internally inconsistent as Eucken observed the N E
αλλά (inc. <xAA’) 254 214 31 1 to be.
άν (inc. καν) 363 97 181
T h e first step in our inquiry is to discover the number o f
äpa (inc. dp’) 36 31 31
γάρ ΙΟΙΙ 368 occurrences o f the most frequent particles in the N E , E E , and A E .
577
ye (inc. y ’) 76 20 T able 4 .1 sets out the results o f such a count. T h e counts were made
38
OO
Se (inc. δ’) 723 1034 on the Oxford I C L i960 A computer by the C O C O A concordance
8η 223 46 75 and word-count program, from machine-readable texts o f the
διό 69 44 71 Ethics prepared by the Thesaurus Linguae Graecae o f Irvine,
i àv 31 17 15 California ; the editions used were B y water’s Oxford Classical T ext
>
€i 181 112 168 for the N E and A E , and Susem ihl’s Teubner edition for the E E .
eïirep 25 15 4 Each entry in T able 4 .1 records the number o f tokens o f a given
67T£l 20 42 60 word-type to be found in the text o f the N E (undisputed
€TL 57 48 45 Nicomachean books), A E (disputed books), and E E (undisputed
y 446 182 3X9 Eudemian books) respectively. (T he distinction between word-
καθαττζρ 70 H 10
tokens and word-types is familiar to philosophers: tokens are
κ α ί 2633 994 1633
counted as a typographer would count, and types as a lexicographer
μ έν 591 325 451
μ ,έ ν τ ο ι 2 7 17 would count, so that ‘Break! Break! B reak!’ contains three tokens
Μ 247 144 230 but only one type.) T h e final line o f the table gives the total number
μ ,ή ν 25 II 19 o f tokens occurring in the respective texts. A glance at the table
οθεν 34 I 7 confirms Eucken’s observations, as that de is very much more
ο ΐο ν 136 70 107 frequent in the N E than the A E , and that höste occurs more often in
δταν 34 47 49 the A E than in the N E ; and it shows that in these two respects, and
δτ€ (inc. oV) 34 5 28 others, the A E feature is reproduced in the E E . Som e o f the
o n 163 133 192
expressions which were discussed by Eucken, however, do not
ον (inc. ου/c, ουχ) 534 288 380
figure in the table: a case in point is the conjunction prin. These
ονδέ (inc. οι5δ’) 232 79 113
οΰν 92 absent expressions are ones which are too infrequent to enable a
235 97
OVT€ 82 statistical argument to be based on them. Even if we assume that an
73 74
πότβρον 22 II 21 author is totally regular in his speech habits, so that any difference
πώ ς 47 17 28 in the number o f occurrences o f a given word between two equal
r e (inc. τ ’) 162 56 91 sized samples o f his prose is totally explicable by chance, we still
τ ο ίν υ ν I 4 17 need a sample large enough to contain at least ten occurrences o f the
ώ ς inc. ώ?) 203 67 141 word before we can use the sample to form a reliable estimate o f the
ώσπερ 98 80 123 frequency with which the author uses the w ord.1
αιστβ 38 61 96
Total words in text 39525 17041 26330 i. T h e reason is well explained by Ellegârd : ‘ I f we postulate that the relative frequency o f
a word in an author’s usage is o oooi, or one in ten thousand, an actual sample (using this
76 The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 77
In the present instance, our over-all strategy will be to endeavour in fact a very large proportion of the text is made up o f particles: of
to predict, on the basis o f the undisputed Nicomachean books, how the 26330 tokens in the E E 6861, or 26 05 per cent, consist of
many times a given word would occur in the disputed books if its occurrences o f one or other o f these thirty-six words.
frequency there was the same as that in the Nicom achean; and to Looking at the raw scores recorded in T able 4 .1 is not by itself
predict on the basis o f the Eudem ian books, how many times a word very informative. Apart from the striking discrepancies o f the kind
would occur in the disputed books if its frequency there was the already noticed by Eucken, the over-all impression is one o f great
same as the Eudemian one; and then to compare these expected similarity o f usage: a word like kai which is frequent in one o f the
frequencies with the actual ones and to inquire whether the three treatises is frequent also in the other two, and words like ge
differences between the expected and actual frequencies are such as and ara which are comparatively rare in one are comparatively rare
are explicable by chance alone. F o r this comparison not to be also in the others. T h is is, o f course, what one would expect in three
misleading, we must restrict ourselves to words occurring with works o f the same author— and indeed some o f the features, such
sufficient frequency to enable us to say— either on the basis o f the as the extreme frequency o f kai, would be expected to be com
Nicomachean evidence or on the basis o f the Eudemian evidence— mon to all writers o f Greek. Is there any way in which one
that they would be expected to occur ten times in the A E . Since the can discover whether the over-all pattern o f the A E use o f par
N E contains 39525 words and the A E contains 17 0 4 1 words, a word ticles revealed by this table resembles more closely the N E or
occurring at the same rate in the N E as in the A E would occur the E E ?
(39525/1 7041) X 10 times, or 23· 19 times in the N E if it occurred A very simple way to calculate a measure o f relative resemblance,
ten times in the A E ; and since the E E contains 26330 words, a word which can be carried out by anyone with a pencil and paper, and a
occurring with equal frequency in the A E and E E would occur minimum o f mathematical skill, is to calculate a coefficient o f rank
(26330/17041) X io tim e s,o r 15-45 times in t h e £ £ if it occurred ten correlation. T h is statistic, which is designated by the G reek letter p
times in the A E . Since words can occur only a whole number o f and called ‘ Spearm an’ s rho’ after its inventor, is familiar to many
times, we must restrict our consideration to words occurring either people involved in education since it is frequently used in order to
twenty-four times or more in the N E , or sixteen times or more in measure the degree o f correlation between gradings and assess
the E E . ments o f the same candidates by different examiners. It is a measure
On this basis a word such as prin (occurring twice in the N E and o f how closely two rank orderings resemble each other. T o apply it
thrice in the E E ) must be left out o f consideration. But even with in the present instance the first step is to rank the particles in order
this restriction, we can find thirty-six particles and connectives to o f preference for each treatise— the most frequent particle being
compare. From one point o f view, this may seem an insignificant given the rank 1, the second the rank 2, and the last the rank 36. The
part o f Aristotle’s vocabulary: the E E , for instance, has a results o f doing this are shown in T able 4.2. (Where two particles
vocabulary o f 4750 w ord-types,1 so that the particles might be said occur the same number of times— as poteron and men in the A E —
to make up only three-quarters o f 1 per cent o f its vocabulary. But both are given the same rank— in this case 3 1 , and the next most
frequent is given the next-but-one rank, in this case 33.) In this
word to refer to consecutive pieces o f text also) o f ioooo words by this author may well
table it is already more obvious that certain particles are more
contain no instance o f the word at all, or even as many as threeor four. (T h is can be calculated popular in one treatise than another: an, ge, de, kathaper, hothen,
by means o f the Poisson form ula.) B ut this also means that a sam ple o f ioooo words is far too oun, for instance, are much higher up in the N E 's preference
small to yield a reliable estimate o f the actual frequency o f such a word. If, on the other hand
we take samples o f iooooo words, the observed frequencies will range between at least 4 and
ranking than in that o f the E E , while epei, hotan, hoti, hösper, and
at most 16 in 95 per cent o f the samples . . .. G enerally speaking, to arrive at a realistic haste are higher up in the E E list; and in almost all these cases the
estimate o f the relative frequency o f a word we need a text sample which is about ten times as A E ranking mirrors the E E ranking rather than the N E ranking.
big as the inversion o f the word's relative frequency’ (1962, 14).
i. T h is figure is based on an unlemmatized word-count in which, e.g. δ ιώ κ ω and δ ιώ ξ ω
B u t to get a precise figure for the over-all resemblances, we need to
count as two types. apply the formula for calculating the Spearman rho, which will
The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 79
T a b l e 4 .2 Ranking o f thirty-six Particles and enable us to compare the rankings pairwise. T h e formula is
Connectives in the NE, EE, and AE
P N (N 2 - i )
NE ΑΕ ΕΕ
where D is the difference between the rank numbers o f any given
άλλα 8 6 7
particle and N the total number o f individual particles in the
âv 7 11 10 ranking (in this case 36 ).1 I f we compare the N E ranking and the
apa 26 25 20 A E ranking and go through it line by line, calculating the difference
γάρ 3 3 3 in each line (2 in line 1, for alia, 4 in line 2 for an and so on) and
γε 19 26 25 squaring the difference each time, and then add up all the squares
δε 2 2 2 and multiply the total by 6 we get the numerator o f the fraction,
δη 12 22 20
6 X (D 2 ), which is 6 x 734 = 4404. T h e denominator is 36 x ( 1296 —
διό 22 23 21
1) =4 66 20. R ho is therefore 1 —(4404/46620) = 0 906. I f we
εάν 30 27 29
calculate rho for the A E and the E E in the same manner, we get the
y
€l Η 10 II
result 0 970. Since rho can vary between 1, for perfect correlation,
e ïn e p 3 1 29 36
22
and o, for complete absence o f correlation, the calculation confirms
€TT€L 34 24
u
€TL 20 24
the impression that the three treatises correlate well together (as
23
6 6 would be expected on the assumption they are all by Aristotle) but
V 7
καθαπερ 21 30 34 that the A E resembles the E E more than it resembles the N E . But
καί I I I the difference between the two statistics is not sufficiently large for
μ εν 4 4 4 us to be sure that the greater resemblance between the A E and the
μ εν το ι 35 33 32 E E is not the coincidental result o f chance. T o investigate this we
μή 9 8 8
need to look more closely at the behaviour o f the individual
μην 30 3 ΐ 3 1 particles.
όθεν 28 36 35
Spearm an’s rho takes into account only the differences in ranking
οΐον 17 ι6 15
between the two lists. A more accurate measure o f correlation is
όταν 28 21 23
given by Pearson’s product-moment coefficient (r) which, applied
ΟΤ€ 28 34 27
ότι to the present material, takes into account not only the difference in
15 9 9
ού 5 5 5
the order of preference o f particles, but also the size o f the
ούδε II Η Η preference: it is based not on the simplified data presented in T able
οΰν ΙΟ 12 ι6 4.2, but on the full data presented in T able 4 .1, which gives not just
ούτε 20 Γ5 19 the rank order o f Aristotle’s particles in each treatise, but the actual
ττότ epov 33 3 ΐ 29 number o f times they are used. I f we calculate Pearson’s r for each
πώ ς 24 27 27 pair o f ethical treatises, for the thirty-six particles, we find that the
Τ€ ι6 ι8
!9 correlation is very good in each case,2 though closer between the A E
τοίνυν 36 35 32
and the E E than between the A E and the N E . r for A E and N E is
ώς !3 17 12
1. T h e procedure for calculating Spearm an’s rho can be found in any statistics textbook,
ώ σπερ ι8 ι3 13 e.g. A ry and Jacobs, 1976, 227 ff. An alternative method o f dealing with ties is to give the
ώ σ τε 25 ι8 17 value « + 0 - 5 to two items tying for rank n. With this procedure, and a formula available for
tie-correction, the value o f p for N E and A E is 0-9047.
2. T h e formula for Pearson’s r is not given : it is complicated, and in practice nowadays the
statistic is seldom worked out with pencil and paper. See A ry and Jacobs, 1976, 16 2 - 8 1.
8ο The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 8i
0 9848, r for A E and E E is 0-9954. T h is very high correlation is in T h e relationships between the treatises can be represented
part due to the effect o f the five particles which are commonest in graphically. A scattergram is a graph in which a single dot or cross is
each o f the three treatises (kai, de, gar, men, ou). I f we remove these used to represent the position o f an individual in two dimensions :
particles from our list, and study the remaining thirty-one particles, the pattern formed by the dots illustrates the correlation. Each cross
we find that that the figures for r are now: for A E and N E 0-8015 ; in our scattergrams will represent a single word-type (e.g. alia).
for A E and E E , 0 9753. Each axis o f the graph will be a scale to measure the frequency o f
T a b l e αλ Scattergram o f Occurrences o f Particles and Connectives in the T A BLE 4.4 Scattergram o f Occurrences o f Particles and Connectives in the
NE and AE EE and AE
214 000 214000
X
/
/
/
/
/
/
/
/
/
171-200 /
/ 171-200 /
/
/
/
/
/ /
/
Q.
/ /
X
/
/ /
/ /
X
/ X /
128-400 / 128-400
/ /
/ /
y O /
x / X /
/ Γ3
/
a /
/
/ X
X /
/
/
X
/ x
X /
85-600 85-600 /
/
x / X /
/
X
x x/
X x/
/ X
/
/
X
/
/ *
/ x
/ / *
f
X
X
X
X
X
42-800 42-800 / -x x
* / ?
/ /
/ X /
/ /
/ X /
/ / x
X X /X
/ x x X
' XX *
/ x
/ X
0000 0000 'r χΧ
« e°
N E particles Λ = 0-8015 E E particles Æ= 0*9753
Standard deviation o f Y d ata against regression line is 31-46 Standard deviation of Y data against regression line is 11-60
82 The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 83
that word in a given ethical treatise : if we wish to correlate the N E T a b L e 4 . 5 Frequencies o f Particles and Connectives in the NE, EE, and AE
and the A E , for instance, we can plot the N E frequencies along one
axis, and the A E frequencies along another. I f Aristotle’s use of NE AE ΕΕ
particles in the two treatises was totally consistent, there would be
άλλά frequency (%) 064 1-26 ι·ι8
perfect positive correlation between the frequencies in each treatise,
standard error 0-04 009 0.07
and the crosses in the scattergram could be connected with a single αν f. 0-92 0-57 o-6g
straight line. In fact, o f course, the crosses scatter to a greater or less s.e. 0-05 006 0-05
extent, spreading in either direction from a regression line— i.e. a âpa f. 009 0 18 0* 12
line which can be calculated on the basis o f the data as giving the s.e. 0-02 003 0-02
best possible estimate o f a word’s frequency in one treatise given γάρ f. 2-56 2 16 2·Ι9
information about its frequency in the other. T able 4.3 gives a s.e. o-o8 01 I 0-09
scattergram for thirty-one particles in the N E and the A E : each ye f. 0 19 0 12 ο ·ΐ4
cross represents an individual particle (they are not identified s.e. 0-02 0-03 0-02
Se f. 468 4-24
individually so as not to clutter the graph): the position of each 3’93
s.e. 0 11 0 15 0’ 12
cross is determined by the particle’s frequency in the A E , as given
δη f. 0-56 0-27 0-28
by the vertical axis, and its frequency in the N E , as given by the s.e. 004 004 0-03
horizontal axis. T h e dashes give the regression line around which διο f. 0 17 026 0-27
the actual values cluster: the standard deviation against the s.e. 0-02 0-04 0-03
regression line is a measure of the degree of their spread. Table 4.4 eau f. o-o8 0 10 ο·ο6
gives the same graphical representation o f the data for the A E and s.e. 001 0-02 0·0Ι
>
the E E . It will be seen that just as the correlation between the A E ei f. 0-46 0-66 ο·04
and the E E is higher than that between the A E and the N E , so the s.e. 0-03 0-06 ο·ο5
crosses in the A E graph are much less scattered and more €17T€p f. 006 0-09 002
compactly clustered around the regression line, with a much s.e. 001 0-02 οοι
> /
€7ret f. 0-05 0·25 0·23
smaller standard deviation.1 T h is brings out in a visible manner the
s.e. O-OI 0-04 003
closer relationship between the A E and the E E which the study o f
en f. 0 14 0-28 01 7
particle usage suggests. I f one wished to predict the frequency o f a s.e. 0-02 0Ό4 003
word in the A E , information about its frequency in the E E would be f. 113
V 1-07 1*21
much more useful than information about its frequency in N E . s.e. 0-05 ο·ο8 007
M ore informative than the raw scores for each particle are the καθάττ€ρ f. o-i8 ο·ο8 004
scores expressed as a relative frequency, as in T able 4.5. H ere the s.e. 0-02 0-02 ΟΟΙ
frequency scores are the ratio o f the number o f occurrences o f καί f. 6-66 5·83 6-20
the word-type in question to the total number o f word-tokens in the s.e. 0-13 ο·ι8 ο ·ΐ5
text. T h us, since kai occurs 994 times in the A E , which contains μ έν f. 15 0 i ‘9 1 ι·7 ι
17 0 4 1 words, its relative frequency is 0-0583. T o avoid a superfluity s.e. 0-06 ο·ιο ο·ο8
o f zeros, the frequencies are expressed in the text as percentages, so μ ίν τ ο ι f. 001 0-04 ο·ο6
s.e. o-oo 0-02 002
that the frequency o f kai is given as 5 83 per cent, and the figure o f
μη f. 0-62 0-85 087
ï. T h e graphs, and other sim ilar graphs later in this book, were produced by the B IV A R s.e. 0-04 ο·οη ο·ο6
packagc for bivariate analysis, written by M r. F . Pettit o f the O xford com puter teaching staff μ ην f. o-o6 oo6 00 7
as a statistical demonstration program. s.e. o-oi 0-02 002
84 The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 85
T able 4.5 (contd.) certain word-type in an author’s text we are looking on each word-
token as a binomial event, that is as an event which may have one o f
NE AE EE
two possible outcomes : either the token is an instance o f the type in
O0€V f. 0-09 . 0*01 003 question or it is not; just as each toss o f a coin must be either heads
s.e. O-OI 0*01 001 or tails. T h e frequency given in the table records the proportion o f
otov c1. 0*41 0*41
0-34 such binomial events which had the first o f the two possible
s.e. 003 0*05 004 outcomes. T h e standard error is calculated in accordance with the
όταν f. 009 0*28 0 19
statisticians’ formula for the standard error of a proportion, σ per
s.e. 001 0*04 0-03
on cent = sjpqjn where sigma is the standard error, p is the proportion,
ore f. 0*09 0-03
s.e. O-OI O-OI 0*02 q = i —p, and n is the number o f events (in this case, word-
OTL f. 0-41 078 tokens in the whole text).
o*73
s.e. 0-03 0*07 0-05 We are treating each o f our texts as providing data about the
ού f. I 35 16 9 1*44 vocabulary usage o f its author at the time o f writing : each text is
s.e. 0*06 O'IO 0*07 taken as a sample drawn from the indefinitely large universe or
ουδέ f. 0-59 0*46 o*43 population o f G reek prose which Aristotle wrote, or could have
s.e. 0-04 0-05 0*04 written, consistently with his stylistic habits at the time o f writing.
ουν f. ° ‘59 0-54 o*37 What these habits were is something that we can now only discover
s.e. 004 0*06 0-04 by a study o f the texts themselves ; and we have to call in aid the
ovre f. 0'i8 0-43 0-31
statistical theory o f sampling in order to discover what inferences to
s.e. 0Ό2 0-05 003
features o f the larger population we are entitled to draw from
π orepov f. 006 0*06 0-08
0*02 0*02 features o f the sample text, and what degree o f error our
s.e. 001
πώς f. 0-12 O-IO 0Ί Ι generalizations are subject to. When we compare the A E with the
s.e. 0*02 0Ό2 0*02 N E and the E E in turn, one thing which we wish to discover is
Τ€ f. O4I 0-33 0-35 whether the A E and the N E , or on the other hand the A E and the
s.e. 0-03 004 0*04 E E , can be regarded as two samples from a single population.
τοίννν f. 0-00 002 0-06 Clearly, as T able 4.5 shows, there are differences between the
s.e. 0*00 O-OI 0-02 frequencies in the A E and the frequencies in the N E , and also
œs f. 0-51 0*39 0-54 differences between the frequencies in the A E and the frequencies
s.e. 0*04 0*05 0*04 in the E E : but we wish to know whether these differences, in one or
ώ σ περ f. 0*25 o*47 o*47 other or both cases, can be regarded as due solely to chance or
s.e. 0*03 0*05 0*04
whether they mark a genuine stylistic difference. F o r if any two
αιστε f. 0*10 0-36 0-36
0-04 samples o f any data are drawn from a given population, there will be
s.e. 0-02 0-05
some differences between the statistics o f the samples which call for
o· 12 against does not mean that 0-12 o f the text consists o f ge, but no other explanation than chance variation in selection, as hands
that 0 Ί 2 o f a per cent o f the text consists o f ge. dealt at bridge from the same pack by fair dealers differ from deal
Associated with each frequency given in T a b le 4.5 will be found a to deal.
standard error. T h is is a measure o f how much reliance can be T h e use o f the standard-error formula enables us to separate out
placed on the frequency in question as an indicator o f the author’s chance differences from ones which call for some other explanation.
general usage, on the assumption o f a certain regularity in his It is possible to calculate the probability o f two samples from the
vocabulary habits. When we are counting the occurrences o f a same population having statistics which differ from each other by
86 The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 87
multiples o f the standard error. As a rule o f thumb, where the the A E resembles the E E considerably more than it resembles the
statistics o f two works differ by more than twice the sum o f the N E . O f the thirty-six particles studied, the A E frequency is closer
relevant standard errors, it is improbable that the differences to the E E frequency than to the N E frequency in twenty-seven
between them can be accounted for as chance variations among two cases, and closer to the N E frequency than to the E E frequency only
samples from the same population. T h is rule o f thumb, with the in nine cases. I f we apply our rule o f thumb, and ask whether the
calculations on which it is based, rest on certain assumptions about differences between the respective features are more than twice the
the nature o f the population, which in the case o f literary texts may sum o f the standard errors o f the proportions, the situation becomes
be questioned, and will be questioned later in the present study. But even more striking, as can be seen in T able 4.6. T h is sets out in the
for the moment, let us return to T able 4.5 to see whether a case o f each particle in turn whether the A E frequency ‘agrees
provisional judgement can be made as to how far the A E , the N E , with’ — i.e. differs by less than twice the sum o f the standard errors
and the E E can be regarded as samples from a single population from — the N E frequency and the A E frequency. It will be seen
exhibiting only chance differences, and whether the differences from the table that the differences between the N E and the A E are
between the A E and the N E are more or less significant than those in a majority o f cases— twenty out o f thirty-six— too great to be
between the A E and the E E . attributable to chance in the form o f sampling error. Whereas the
I f we look first at the frequencies alone, we can already see that differences between the A E and the E E in every case but two (eiper
and hote) are such as to require no explanation other than chance.
T a b l e 4.6 Agreements and Disagreements between AE, NE, and EE in T h is result is very striking: can we achieve a more precise
respect o f Particle Use evaluation o f the significance o f the result and o f the probabilities
involved ?
Particle Does Α Ε Does A E Particle Does Α Ε Does A E T o calculate the significance between two sampling proportions
agree with agree with agree with agree with statisticians proceed as follows. First, the value o f a statistic z is
NE ? EE? NE? EE? calculated by dividing the observed difference between the
proportions by the appropriate standard error o f the difference.1
αλλά No Yes μη No Yes T h e theoretical probability o f the difference between two pro
>/
av No Yes μην Yes Yes portions o f samples reaching a particular z value is known in
âpa Yes Yes δθε,ν No Yes advance and can be ascertained from tables to be found in any
γά ρ No Yes oiov Yes Yes
statistical textbook: the probability o f a z value greater than 19 6 ,
ye Yes Yes όταν No Yes
δε Yes Yes ore No for instance, is 0 0 5 , or one in twenty; the probability o f a z value
No
8η No Yes ΟΤΙ No Yes greater than 2 58 is 0 0 1 or one in a hundred. I f the probability that
διο Yes Yes ού No Yes the difference arises from sampling error falls below an acceptable
id V Yes Yes ou8e Yes Yes level, then the null hypothesis, i.e. the hypothesis that the two
î
€L No Yes οΰν Yes Yes samples have been drawn from the same population, is rejected, and
eïnep Yes No ούτε No Yes
€7Τ€ί No Yes TTorepov Yes Yes I. The value of z is calculated from the data by means of the formula
»/
€TL No Yes π ώ ς Yes Yes
V Yes Yes τβ Yes Yes z = ---------- where p =
καθάπερ No Yes τοίνυν Yes Yes , + */«2 JÏ »!+»2
κ α ί No Yes ώ ζ Yes Yes In testing the difference between the N E proportion and the A E proportion p l is the N E
μ ίν No Yes ώ σπ ερ No Yes proportion, p 2 the A E proportion, «, is the number o f words in the N E (39525), n2 the
μ ίν τ ο ι No Yes number in the A E (17 0 4 1). T his formula, and the explanation o f the method o f testing the
ώ σ τε No Yes
significance between sample proportions, is derived from Caulcott, 19 7 3, 85.
88 The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 89
the differences between them are described as statistically signi probability o f two samples coming from the same population is only
ficant. It has to be decided in advance what level o f probability is 0 00005, or one ' n 20000 : the values for an, dë, and hosper are all
regarded as acceptable : the decision is to a certain extent arbitrary between four and five. For z = 5, the probability is only 0-0000006,
and to a certain extent guided by the purpose o f the test (e.g. we less than one in a million and a half ; the £ values for alia, epei, hotan,
must ask whether it would have more serious consequences to reject hoti, oute are all above this. On the other hand, all but three of the
the null hypothesis when it should not be rejected, or to accept it differences between the A E and the E E are statistically insignificant
when it should be rejected). T h e two levels most commonly used at our chosen level. Eiper and hate had already been revealed as
are the 0 0 5 and 0 0 1 level: for our purposes the appropriate mavericks by the application o f our rule o f thumb : oun, where the
procedure is to test for significance at the 0 01 level (it is more difference previously appeared to be just within the limits o f
important not to discover imaginary stylistic differences than it is to sampling error, is now shown to be just outside them.
be sure of discovering all genuine ones). When we say that it is enormously unlikely that the differences
T ab le 4.7 gives the z values o f the differences between the A E between the A E and the N E should have come about by chance,
and N E and o f the differences between the A E and E E for each that does not o f course mean that we can say without further ado
particle. It will be seen that twenty-one o f the differences between that it is enormously unlikely that the A E and the N E were
the A E and N E are statistically significant at the 1 -per-cent level. originally part o f the same work. F o r the differences, though not
Som e o f the values are extremely significant: at 2 = 4-00 the due to chance, may be due to a number o f other things : a conscious
T a b l e 4.7 Significance o f the Difference between Frequencies o f Particles in wish by Aristotle to vary his usage in one part o f his work ; the
the AE, NE, and EE influence o f subject m atter; varied degrees o f revision and
polishing, and the like. What our study has shown so far is that it is a
Particle N E and E E and Particle N E and E E and much more economical hypothesis to regard the A E as belonging
A E value A E value A E value A E value originally to the E E ; for almost all the differences between the A E
of z of z of 2 of z and the E E in respect o f the vocabulary so far studied call for no
other explanation than chance.
αλλά 7'39 0-69 μη 2-90 Q'31 We cannot yet be said, however, to have satisfactorily quantified
άν 4-24 !-5I μην 006 029 the likelihood o f chance accounting for the observed differences
âpa 2-88 1-73 θθ€Ρ 3·52 1 55
γάρ between the N E and A E in respect o f particle usage. F o r the
2-82 0-22 oîov I '2 I 0-07
ye 1-99 όταν statistical procedures which have been adopted are ones which are
0-75 5-48 1-94
Se 2-31 1-63 0Τ€ 2-36 2-84 strictly applicable only in situations where the outcome o f each o f
Sy 4-67 0-29 δτι 060 the binomial events to be counted is independent o f the outcome o f
5-57
διό 2-04 Ο-23 ον 3 09 2-04 each other event. In the case o f the occurrences o f a word in a text
èàv o-8o i*6o ονδβ 1-82 0-53 this assumption is clearly unrealistic : the choice o f a word at one
ei 3-03 0*24 ονν 079 2-65 point in a text affects the likelihood o f its being chosen at other
eiirep 1-02 3’54 OVT€ 5'35 2-09 points in the text. T h e matter has been vividly illustrated by A. Q.
èirei 6-46 0-39 π οτζρον 0-40 0-57 M orton (1966, 70):
£Tl 3-48 2-44 πώ ς 0-62 0-21
Kai occurs at a rate of about 005, so that two successive occurrences of kai
V 0-63 1-37 Τ€ 1-43 0*30
would be found, if the occurrence was a matter of pure chance, at a rate of
καθαπερ 2-69 1-91 TOLVVV 2-43 I9O
005 χ 005 = 00025, i.e. in four hundred words of text you could expect to
καί 3-69 i *57 ώς 19 1 2-10
p.év find two successive kais. In Greek prose two successive kais are found
3-56 1-49 ώσττζρ 4-32 0-03
μ έν το ι about once in a million words of text.
3-12 Γ02 ώστβ 6-83 on
But though the choice o f a word has an effect on the choice o f other
90 The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 91
words in its immediate context, it seems reasonable to expect that study in individual books; and sim ilarly, only those with an
this effect can be discounted if we make our samples sufficiently occurrence in the E E o f fifty or more will repay study in the
large : there seems no reason to expect that the fact that Aristotle individual five books o f the E E . I f we exclude from our list o f
used kai on a particular line o f the first page o f book i o f the N E particles those which do not occur either seventy or more times in
should affect his decision to use it on a particular line o f the last page the N E or fifty or more times in the E E , we reduce our list from
o f book io. One might well hope that samples as large as the A E , thirty-six to twenty-four, with the omission o f ara, ean, eiper, eti,
N E , and E E would be big enough to allow the local in mentoi, men, hot hen, ho tan, hote, poteron, pos, toinun. It will be noted
terdependence o f word-choices to be discounted. But this hope that the new shorter list consists entirely o f particles whose
must be put to the test. What size a sample has to be for the frequencies in the A E and E E differ by only insignificant degrees.
occurrence o f a given word in a text to conform to the conditions of O f the new list, fifteen show significant disagreements between N E
simple sampling is an empirical matter which must be discovered and A E , while nine agree in N E and A E usage.
by counting and calculation. T o discover whether the A E , N E , and Table 4 .10 sets out the occurrences o f the twenty-four particles
E E display the degree o f internal homogeneity which would justify in each o f the seven books o f the N E , and Table 4 .11 in the five
making statistical inferences about their relationship, we must books o f the E E . T h e chi-squared value for each particle is
break up the text o f each o f them into smaller units, and count the calculated in a way which may be illustrated from the case o f ei in
occurrences o f words in each o f these new samples. A significance the E E , whose occurrences are given in T able 4.8. We first set out
test known as the chi-squared test enables one to assess whether the its occurrences in the form of a table known as a contingency table.
occurrence o f the word complies with the conditions o f simple
sampling or whether there are statistically significant differences
between the rates o f occurrence in different samples. T a b l e 4.8 Occurrences o f e i in the EE
T h e obvious smaller units into which to break the treatises are
E E book Occurrences of ei Occurrences of Total words
the individual books that compose them. Considered as samples
other words in book
these books have the disadvantage that they are o f different sizes,
which complicates some o f the calculations; but taking them as I 18 3366 3384
separate wholes has the advantage o f keeping open the possibility, II 27 4672
4645
suggested by many scholars, that they may have existed once as III 52 7043 7095
separate units which were perhaps only later combined into the V II 52 8566 8618
more lengthy treatises that we know. Obtaining statistics for them V III 19 2542 2561
separately will not only help us to assess the degree o f reliance Total 168 26162 26330
which can be placed on our conclusions concerning the treatises as a
whole : it may also enable us to detect any anomalies which may
make individual books stand out from the other books to which they
are traditionally united. We wish to know whether the differences in the figures from
Som e o f the particles we have so far studied occur too book to book can be explained simply as the result o f sampling
infrequently to make it profitable to study their occurrences in error. I f so, we would expect the proportion o f eis in each book to be
particular books. As before, we cannot make reliable estimates o f the same as the proportion in the whole of the E E . Accordingly, we
frequencies from samples unless the expression in question can be work out, for each book, the expected occurrences o f ei and the
expected to occur a minimum o f ten times within them. Since the expected occurences o f other words, on the assumption that the
N E consists o f seven books, this means very roughly that only proportion remains thus constant. T h e results o f the calculation are
particles with an over-all occurrence o f seventy or more will repay inserted in brackets in Table 4.9.
92 The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 93
statistically non-significant and the occurrence o f ei in the E E
T a b l e 4.9 The Calculation o f Chi-squared
conforms to the conditions for simple sam pling.1
E E book Occurrences of el Occurrences of other Total A study o f Tables 4 .10 and 4 .1 1 reveals how far, within the N E
actual (expected) actual (expected) and within the E E , Aristotle was regular in his habits o f use of
particular connectives. T h e tables contain a number o f surprises.
I 18 (21-6) 3366 (3362-4) 3384 K a i, which because o f its extreme commonness and apparent topic-
II 52 (45'3) 7043 (7049-7) 7095
III 27 (29-8) 4645 (4642-2) 4672 T a b l e 4.10 Particles and Connectives in Individual Books o f the NE
V II 52 (55-0) 8566 (85630) 8618
V III 19 (16 3) 2542 (2544-7) 2561 I 2 8 9 10 X 2 f°r
3 4
Total 168 26162 26330 6 degrees
of freedom
άλλα 37 3D 60 38 37 22 30 16-27
>/
αν 60 18 53 50 40 55 87 33-50
F o r each cell we now square the difference between the actual 128 76 170 29-41
γάρ 151 195 135 156
and expected values and divide the result by the expected value: ye 18 20-41
3 17 15 3 5 15
thus for the first cell (eis in book I) we square (2 1-6 -18 ) and divide 8é 267 216 302 295 298 234 239 159 4
the square by 21-6, obtaining the result 0 6 0 . When we have 8η 28 10 42 28 31 38 46 16 37
performed this operation for each cell in the table, we add all the 8 ιό 5 10 Η II 10 10 9 4-51
>
results together, and that gives the value o f chi-squared for the table et 44 l3 37 15 9 21 42 4 12 8
as a whole. In this case chi-squared = 2 46. €7T€L 4 6 4 2 I 2 I N.A.
T h e significance o f a chi-squared has to be discovered by y 63 35 93 67 58 71 59 14-96
consulting a table; but first one must ascertain the number o f καθάπβρ 19 3 3 3 IO 21 II 36-75
και 419 298 398 440 362 320 396 17-97
degrees o f freedom associated with it. T h e number o f degrees o f
μ έν 75 108 104 85 84 67 68 42-85
freedom is the number o f ways in which the data represented in the
μή 28 9 61 59 30 34 26 4 3 11
table are free to vary independently o f each other within the oîov 16 21 20 18 8-50
!3 33 15
constraints set by the characteristics o f the total. In each book, δ τι 21 22 16 10 21 2665
39 34
given the number o f occurrences o f ei, the number o f occurrences o f ον 41 48 91 115 80 65 94 35-24
words other than ei is not free to vary independently but is fixed: ον Se 18 22 43 43 31 27 48 14-50
therefore in each line o f the table we have only one degree o f r 766
ουν 25 32 35 32 30 39 42
freedom. G iven the number o f eis in each o f the first four lines o f the οντ e 16 16 12 15 5 4 5 23-18
table, the number in the fifth is not free to vary but can be deduced τε !7 14 II 22 24 26 48 30-85
by subtraction from the total : therefore the degrees o f freedom (i)S 19 19 42 43 25 23 32 13.89
from line to line are four. T h e number o f degrees o f freedom o f the ώ σπερ 12 14 15 18 6 13 20 8-48
ώ σ τε 11 4 5 4 5 4 5 N.A.
table as a whole is four by one, i.e. four. I f we enter a table o f chi-
squared in a statistical textbook at the entry for four degrees o f Totals 5644 4242 6338 5953 5827 5291 6230
freedom we find that the probability o f a chi-squared o f 2 46 is
F o r 6 degrees o f freedom, a value o f 16 8 1 or more is significant at the i-per-cent level.
above 0 50: a distribution o f ei parallel to that in the E E would be
found by chance in one case out o f two. Clearly, the differences I. T h e procedure for calculating chi-square can be found in any statistics textbook; e.g.
between the rates o f occurrence in one book and the next are Caulcott, 19 73, 1 1 4 ff.
94 The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 95
T a b l e 4 .11 Particles and Connectives in Individual Books o f the EE other particles all occur in at least one o f the treatises in accordance
with the conditions o f simple sampling, and therefore enable
I II III VII V III X 2 for inferences to be drawn from the frequency o f their use in either the
4 degrees N E or in the E E to an expected frequency in the A E .
of freedom Before proceeding to the comparison with the analogous data for
αλλά the A E , we may notice that our tables reveal that the E E , despite the
32 7§ 43 112 46 !4-°7
ay 26 greater variation in the length o f the individual books that compose
34 34 56 31 I 5-38
γάρ 68 104 225 1249 it, displays a considerably greater degree o f uniformity in respect of
Ϊ31 49
ye 5 !3 3 10 7 626 particle usage. Disregarding the four particles irregular in both, we
Se 103 287 198 3Ö2 84 13-01 find that fourteen out o f the other twenty in the E E occur at rates
δή 5 25 8 22 !5 139 3 showing only insignificant differences between books. In the N E
8 to 0 17 ΐ4 35 5 13 9 3 only ten are similarly regular; twelve display significant fluctu
et 18 52 27 52 19 246 ations in rate o f occurrences, and in two cases the chi-squared test
€ΤΓ€1 I ι8 II 22 8 7-16 is not applicable.1
>/
38 ι:5 6ο 67 39 2600 I f the A E was originally part o f the E E , we would expect that
η
καθάττερ 4 3 2 I I N.A. those regularities which are displayed in the particle use in the E E
καί 165 40Q 312 526 ι6 ι 14-26
μ έν would continue to be regular, and regular at approximately the
58 12 1 84 148 40 0-55
μη 26 same rate, in the A E ; similarly, if the A E was originally all o f one
74 40 8ο ίο io-oo
οίον 21 10 piece with the N E one would expect it to reflect the regularities to
15 32 29 1-73
δτι 24 56 20 67 be observed there. W e can therefore use the fourteen observed
25 8 -6 6
ού 45 100 46 !33 56 17-83 regularities from the E E , and the ten observed regularities from the
ovSe 27 28 ι1 29 12 ! 3'25 N E , to make predictions as to the regularities in particle usage to be
ονν 15 19 Π 29 17 8-8o expected in the A E . T able 4 .12 , which sets out the figures for the
ούτε !3 20 25 22 2 13-67 individual books o f the A E , enables us to begin to compare these
Τ€ 17 25 ι8 24 7 4 15 expectations with the observed data to see whether the
ώς Η 25 ι8 67 17 176 5 Nicomachean hypothesis or the Eudemian hypothesis best fits the
ώσττ€ρ 9 23 23 52 ι6 10-94
facts.
ώ στΐ II 35 12 29 9 5-07 T h e most striking feature o f T able 4 .12 is that it shows the A E as
Totals 3384 7095 4072 8 6 ι8 2561 much more uniform than either the E E or the N E . In twenty-two
out o f twenty-four cases the differences between the individual
F o r 4 degrees o f freedom, a value o f 13 28 or more is significant at the i-per-cen t level.
books are statistically insignificant. Only in the cases o f ë and hösper
are the between-books differences significant at the 1 -per-cent
neutrality might have been expected to occur with constant level. T h is fact is not negligible in view o f the fact that some
frequency, is irregular in appearance in both the N E and the E E : scholars have wished to solve the problem o f the disputed books by
there are differences between the rate o f occurrences in individual
books which, in each treatise, are just significant at the i-per-cent i . F o r a chi-squared test to be applicable, at least 80 per cent o f the cells in the frequency
level. Sim ilarly the negation signs ou and ou te occur at significantly table must have an expected value o f no less than five. Because e-nei and w a r e each occur less
different rates in different books within each treatise, and so does than forty times in the N E this is not so. T h e inapplicability o f the chi-squared test in this
case is o f course not because the difference between the N E and E E in this respect is too
the particle an. It is not, then, the commonest or most apparently
small, but because it is too enormous. T h e test is inapplicable for analogous reasons to
topic-neutral particles that occur with the greatest regularity. T he κ α θ ά π ε ρ in the E E .
Φ The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 97
T a b le 4.12 Particles and Connectives in Individual Books o f the AE dë, dio, hoion, oude, oun, hos) where we can compare the regularity
found in the N E with that found in the A E ; and thirteen cases (gar,
χ 2 for 2 degrees ge, de, ei, epei, men, më, hoion, hoti, oude, oun, te, hôste) where we can
A B C of freedom compare regularities in the A E and the E E . We already know, from
T able 4.6, that in six out of the eight N E cases, and in every one of
αλλά 88 43 83 2‘55 the thirteen E E cases, the differences between the frequencies in the
âv 34 32 31 3·78
γάρ 133 102 133 993 T able 4.13 Comparison o f two contexts for the AE in respect
ye 5 5 IO 1-68 o f Particle Use
δε 280 161 282 2-29
δη 12 20 14 8-94 Particle A E and N E A E and E E
διό 15 15 14 2-15 χ 2 for 9 degrees χ 2 for 7 degrees
€1 53 20 39 5·°9 of freedom of freedom
è n ei 10 7 25 8·6ι
V 90 3 0
6 2 1 2 - 2 8 αλλά jr o 8 1 6-86
καθάπερ I 3 10 7'79 âv s ^-44 22-12
καί 362 240 392 1-50 γάρ 40-48 I4'5I
μ ίν 114 71 140 4-27 ye 28-22 8-79
μη 54 3 1 59 i -oo §é 23-84 17-67
οΐον 34 10 26 5-25 δη 46-27 22-95
ότι 45 39 49 1-76 διό 10-90 16 4 1
ού 114 62 I 12 1-54 ei S2 'H 7-68
ούδί 20 28 3 1 7-06 €71-et 66-07 1 6 12
οΰν 21 0-89 >/
3 2 39 27-41 4°'49
η
otire 20 2 4 3 0 4-26 καθάττερ 55'37 N.A.
re 27 IO 19 2-89 και 34 '10 18-50
ώς 3 0 1 4 2 3
1-46 μ έν 57Ί 1 7-30
äianep 2 3 ! 3 4 4 1 0 - 4 5 μη 48-63 11*21
ώ σ τζ 2 3 1 7
21 0-43 ο ΐο ν 15-41 7*01
δη 54-63 10-62
Totals 6430 4198 6413 ον 44-14 22-56
Fo r 2 degrees o f freedom, a value o f 9-21 or more is significant at the irper-cent level.
ov 8e 24-74 20-52
ονν 933 1 5-56
OVT € 52-20 21-33
splitting them up, assigning one or more o f them to the N E and the Τ€ 38-31 7-12
remainder o f the E E . 1 ώς 20-49 25-29
T h e irregular occurrence o f hösper contrasts with the regularities ώσπβρ 40· 8q 21'40
to be found in both N E and E E ; the irregularity o f ë parallels its ώστε 5 r 39 5*55
irregularity in the E E , and conflicts with the prediction o f regularity
from the N E . Otherwise, all predictions o f regularity from either F o r 9 degrees o f freedom, a value o f 21-66 or more is significant at the 1-
per-cent level.
E E or N E are fulfilled by the A E . We now have eight cases (alla, de, F o r 7 degrees o f freedom, a value o f 18-48 or more is significant at the i~
i. T h u s Rowe, op. cit. 9 0 - 1 1 4 sees A E B as being wholly Nicomachean, whereas A and per-cent level.
C, he thinks, have a Eudem ian base. Values o f χ 2 significant at the i-per-cent level are italicized.
98 The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 99
treatises as a whole are not significant ; but we are now in a position T a b l e 4.14b ώ στε tn the traditional EE. Proportions in per cent
to make a comparison which takes account not only o f the frequency 0-6
of the particles in each treatise considered as a whole, but also of the
distribution o f the particles in the particular books o f the treatises. 0-5
We can place the three books of the A E in their rival contexts in
turn, and see which experiment produces the more homogeneous 0-4
resulting whole. We do so by putting the seven books o f the N E
together with the three books of the A E , and applying a chi-squared
0-3
test to the ten-book whole thus produced (the traditional Nico
machean Ethics); and similarly, lumping together the E E and the
0-2 ί-
A E and applying a chi-squared test to the resulting eight-book Eude
mian Ethics. T h e results o f this experiment are shown in T able 4 .13.
0- t
It leaps to the eye that in respect of particle usage the traditional
Eudemian Ethics is a much more homogeneous work than the
traditional Nicomachean Ethics. In the former, there are only nine 0-0
3 4 5 8
cases where the differences between books are significant ; in the Books
T a b l e 4.14 Bar Charts o f the Use o f ε π ε ί and ώ στε in the traditional .... ,j _ l_ . . Ί
00
NE and EE 1 4 5 6 10
Books
T a b l e 4.14a ώ σ τε in the traditional NE. Proportions in per cent
0-6 r
T a b l e 4.140 ε π ε ί in the traditional EE. Proportions in per cent
0-4 r
0-4
C- 0·3
" 0-3 C
α> 1)
u
&
0-2 ê 02
0 1
0-1
00
00 2 3 4 5 6
4 5 6 10 Books
Books
100 The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 10 1
be found in the E E alone; and even those particles (such as alia) T a b l e 4.15 (contd.)
which display irregularity in the narrower context o f the E E no
longer do so if placed in the wider context o f the traditional
Eudemian Ethics. I 2 3 4 8 9 10
We must now take a further look at the irregularities discovered s .e . 0*14 o-i6 013 014 0- 1 4 015 0-13
καθάπερ f. o*34 0-07 0-05 005 0- 1 7 0 40 o-i8
between books within the undisputed N E and the undisputed E E .
s .e . 0-06 006 0-05 005 o-o6 006 0-05
Is it possible to discover whether these irregularities in the use of 6-21
καί f. 742 7'02 6-28 7 ’39 605 636
particular particles are concentrated in particular books o f these 0-38 031 032 0-32
s .e . o* 33 0 -3 3 0 -3 4
treatises which render anomalous an otherwise homogeneous whole ? μ εν f. 164 1 -4 4 1 -27 i 09
1-3 3 2 -5 5 1 -4 3
T h e raw scores given in Tables 4 .10 , 1 1 , and 12 do not render this s .e . o * i6 o-i8 015 016 016 017 0-15
perspicuous : the matter is more easily studied in Tables 4 .15 and 16 μή f. 0-50 0- 21 096 099 0-51 064 0-42
which give the proportions and standard errors o f the particles in s .e . 0-10 011 o-io o-io o-io O II 010
individual books. οΐον f. 0-28 031 0-52 0 -3 5 0 -3 4 0 -3 4 0-24
s .e . o-o8 008 0-07 0-08 0-08 008 0-07
δη f. 0 -3 7 052 0-62 0-27 017 064 0 -3 4
T a b l e 4.15 Frequencies of Particles and Connectives in Individual Books s .e . 0*09 009 0-08 0-08 0-08 009 0-08
o f the NE ον f. o* 73 ΙΊ3 144 1 -9 3 1 -3 7 1-23 151
s .e . 0-15 018 014 0- 1 5 0- 1 5 016 015
I 2 3 4 8 9 10 ονδε f. 0-32 0-52 0-68 072 0 -5 3 051 0-77
s .e . 0-10 O II o-io O 'I O o-io 011 010
αλλά frequency ονν f. o*44 0 -7 5 0-55 054 0-51 0-74 067
(%) 066 071 0-95 064 063 042 0-48 s.e. 0*10 011 o-io 0 10 o-io o-i i 0 10
standard OVT € f. 0-28 038 0 19 0-25 0-09 0-08 0-08
error 0 11 0 12 0 10 o-io 0· 10 o-i i 0 10 s.e. 0*06 006 0-05 005 0-06 006 0-05
αν f. i 06 042 0-84 0-84 069 1-04 I 40 Τ€ f. 0-30 0 -3 3 0 17 0 -3 7 0-41 049 0 -7 7
s.e. 0 13 0 15 0-12 0-12 0 13 0-13 0 12 s.e. 0-09 009 0-08 0-08 0-08 009 0-08
γάρ f. 2-27 17 9 2-38 286 335 2'55 2-50 ώς f. 0*34 0 -4 5 0-66 072 0 -4 3 0 -4 3 0-51
s.e. 0-21 0-24 0*20 0-20 0‘2I 0-22 0-20 s.e. 0*09 010 009 009 009 0- 1 0 0-09
ye f. 0-32 0-07 0-27 0-25 0-05 OO9 026 ώ σπερ f. 0*21 0'33 0-24 0-30 o-io 0-25 0-32
s.e. 006 007 0*05 006 006 0-06 006 s.e. 0-07 007 006 006 0-07 007 0-06
Se' f. 473 509 4*76 496 5 11 4-42 3 84 ώστ€ f. 0*19 009 0-08 0-07 0-09 008 0-08
s.e. 0-28 0 32 0-27 0-27 0-28 0-29 0-27 s.e. ΟΌ4 004 0-04 0-04 0-04 004 004
102 The Aristotelian Ethics Particles and Connectives in the Aristotelian Ethics 103
T a b l e 4.16 Frequencies o f Particles and Connectives in Individual Books T a b l e 4.16 (contd.)
o f the EE
I II III V II V III
I II III V II V III
οΰν f. 0-44 0-27 0-36 0-34 0-66
αλλά frequency s.e. O -IO 0-07 0-09 0-07 0-12
(%) 0 -9 5 IIO 0-92 1 -3 ° i 80 ούτε f. 0-38 0-28 0-54 0-34 0-08
standard s.e. o-io 0-07 0-08 0-07 01 i
error 019 013 016 0-12 0-21 τε f. 0-50 0-35 0-39 0-28 0-27
αν f. 077 0-48 0 -7 3 0*65 1*21 s.e. o-io 0-07 0-09 0-08 0-12
s.e. 014 010 0-12 0*09 Ο 16 (ï)S f. 0-50 0-35 0-39 078 0-66
γάρ f. 201 i -85 2- 23 261 191 s.e. 0-13 0-09 o-i 1 0-08 0-14
s.e. 0-25 0-17 0-21 016 0-29 ώ σπερ f. 0-27 0-32 0-49 o-6o 0-62
γε f. 015 018 006 012 0-27 s.e. 0-12 0-08 o-io 0-07 0-14
s.e. 006 0-04 OO5 004 ΟΌ7 ώ στε f. ο·ι8 049 0-26 0-34 0-35
δε f. 304 4 -0 5 4 -2 4 4-20 3-28 s.e. o-io 0-07 0-09 0-06 ο ι 2
Scores in bold type indicate 2-values (absolute) of more than 2-58 (/><ο·οι).
00 Ό η μ rh ro
binomial, but some other such as the Poisson distribution.1 None
the less, it is regular statistical practice to use the binomial
distribution instead o f the Poisson, and the normal distribution in
turn instead o f the binomial, provided that we are dealing with a
00 N 00 <-> h rsO ■-« m <-< ο -'ί* ·-> Ο O r*-> N N ·-· 'f Ό reasonably large number o f trials ; for in such cases the distributions
O 1· O ' O ir > \C
approximate very closely to each other, and the use o f the normal
distribution greatly facilitates computation. In the case o f the
00 so η o 00 i statistics for the N E , A E , E E as wholes, and for the individual
books as wholes, the length of the books (and therefore the number
ci tj- im *-> t+i o f trials in question, each word-choice counting as a trial) was
i r n nO•Ή *-< Ό
co^foo Lnoc ιη-Φτί-Ν μ μ γον
sufficiently great for this approximation to be reliable. Bu t now that
able
à
(seems), doxeie (would seem), eoike (seems likely), phainetai
vo (appears)— are studied in Table 5.4 which gives the occurrences,
and
Ö
-Θ - between them, add up to over 1 per cent o f the whole Nicomachean
text. B y this crude measure, the E E is only half as tentative as the
NE,
O
OO 00 N m t}· 00 w
>no r-ο o
<P~ ·"< HH VT) N Μ ΙΛ N E (these expressions amount to 0-43 per cent o f the text) and the
00
in In d ivid u a l Books of the
P*
A E is even less tentative (0 36 per cent). Other, less frequent,
* expressions o f tentativeness, not shown in the table, exhibit the
0 0 *-« 0 0 0 0 M Ο N u"> ro N N O OO ro im
<N Tf same pattern, such as tacha (possibly) and adelon (unclear), which
-ö
occur between them twenty two times in the N E , four times in the
C E E , and twice in the A E .
+1 •2 ·-< O' r-** n o 00 n t »Λ h O « o n « ro
4-t ro « Ο Γ>» N N >■* o ro r o r o ro N fO i/ l ΙΛ r o ■φ It will be noted that the tentative expressions occur very much
* Ο ~ Ο ~ ~ ~ « 0 Ö Ö 0 ö ό ö ö ö less frequently in N E 2 than in any other Nicomachean book: the
u ^ 00 00 ro Ο 00 ΓΟ rj- N 1Λ t « r-Ό o m oo
CL, *
>»
o ^ r»so r- o N « N Ο N N Tf
n
hH proportion is only 0-19 per cent. Here, as in particle usage, N E 2
■
<t·
resembles the E E pattern in some respects more than the pattern o f
Certainty
δήλον
δήλονχ χ
/
/
/
/
27-200 . ... y / . 27-200
/
/
/
/
/
/
/
/
20-400 r , 20-400
1
/
/ C
r
\
\
(Λ
T / 0)
\
o
\
\
C/5
/ c
\
8 / U
u
c u
u φχVEpÔVyf 3O φχ νερό V
\
g o
\
\
/ O
\
\
/
\
O 13-600 13-600
/ X X χ ν ίγ κ η
/ 'ανάγκη
/
/
/
/
/
ίσως / Ίσω ς
X εοικε
’έοικεχ x /
6-800 / 6-800
/ φαίνεται χ
f X
/ φαίνεται δόξειε χ
/ δόξειε
/
/
/
/
/
0 000 / 0000
910 14 19 29 41 56 1 5 42 58 70 87 91
R = 0-9322 O ccurrences in E E R = 0-2366 N E occurrences
A E is not significantly different, when tested by the usual test,1 level for six out o f the eight expressions (all except dokei and dëlon).
from the occurrence in the E E ; by the same test, when we compare When we take the tentativeness group as a whole, there is no
the E E and the N E , the null hypothesis is rejected at the i-per-cent significant difference between the figure for the E E and the A E
I . See above, p. 84. (2 = 1*09), but there is between the N E figure and the A E figure
122 The Aristotelian Ethics Prepositions, Adverbs, and Pronouns in the Ethical Treatises 123
( 2 = 8 0 3); similarly with the figures for the certainty group as a P R O N O U N S A N D D E M O N S T R A T IV E S
whole (z for N E and A E is 4 32, for E E and A E it is 2 07). When we turn from indeclinable adverbs to pronouns and
We can present the data graphically in a pair of scattergrams in demonstratives, and later when we come to consider adjectives,
which the values for individual expressions in the A E are plotted nouns, and verbs, we encounter a new difficulty in handling the
against those in the E E and those in the N E respectively. T h e result data. T h e C O C O A program used for word-counting does not
o f this for the A E and E E is shown in T able 5.5, and for the N E and lemmatize : that is to say, each separate form o f a declining noun or
E E in Table 5.6. It will be seen that A E and E E correlate well conjugating verb appears listed in alphabetical order, not grouped
(r = 0-9322), whereas the correlation between A E and N E is so poor under the form under which the noun or verb would appear in an
that the regression line is valueless. ordinary lexicon. Almost all the data considered so far have
There seems to be less information for our purpose to be gained involved only a single entry in the print-out o f the computer word-
from the study o f other adverbs in the ethical treatises. Tw o count (the exceptions being contractions such as αΙΓ which get
temporal adverbs are Eudem ian favourites, hama (at once) and aei counted as forms o f the uncontracted word, and forms containing a
(always) ; in the case o f hama the A E frequency is closer to the N E terminal n, such as eoiken, which are counted with the unterminated
one, and in the case o f aei closer to the E E one, but in neither case form). When counting Aristotle’s usage o f a pronoun, a noun, or
are the differences between the A E frequency and either o f the verb, we have to group together by eye, and total by hand, the
others significant. T h e values are set out below, along with those o f occurrences o f each o f the separate unlemmatized forms which
three other frequent adverbs: together constitute the score for the single dictionary entry. There
is ample scope for human error in omitting a form or incorrectly
NE AE EE adding a total. T h is should be borne in mind henceforth.
T w o o f the commonest pronouns are houtos (this) and autos (self,
Occur Propor Occur Propor Occur Propor he, same). Between them they make up about 3 per cent o f the text
rence tion rence tion rence tion o f each o f the three ethical treatises. Though there are slight
°/
/0 °/
/0 °/
/0 differences between the frequencies with which they occur in the
1 / different treatises, these differences are not significant.
aei 24 ο·ο6 20 0 12 24 0-09 (Approximate frequencies for houtos: N E 1 5 1 per cent, A E 14 6
αμα 18 0-05 9 005 35 0 13 percent, E E 1 7 5 per cent; for autos: N E 1 44 per cent, A E 1-34 per
δλως 35 ο·ο9 15 009 24 0-09 cent, E E 1 5 0 per cent.) Sim ilarly the figures for the two Greek
ο μ ο ίω ς 94 0-24 29 0 17 50 0-19
words for ‘nobody’, oudeis and mêdeis, show no significant
ό ντω ς 51 ο ·ΐ3 20 0 12 34 0 13
differences between the treatises (provided that oudeis and outheis
are taken together, and likewise mêdeis and mêtheis). T h e figures and
proportions are:
In the case o f holös, houtös, and homoios the frequencies in each o f
the three treatises are too close together to be used for any NE AE EE
discriminatory purpose. T h e evidence provided by the use o f these
words cannot be used unless they can be suitably grouped with Occur- Propor- Occur- Propor- Occur- Propor-
other words— e.g. their parent adjectives— to yield statistically rence tion rence tion rence tion
0/ 0/ 0/
significant results. W e shall see later how this can be done, as it can /0 /0 /0
these three expressions account for over 1 per cent o f the N E , but
Οί
o only 0· 57 per cent o f the A E and 0 66 per cent o f the E E . In the case
Q* η
«
N N
μη
r o o
N n
Ο M
00 OO m sO N co O PO o OO 'r> i-<
l··
ro w ci r* H Ht M Ç* o f none o f these words is there significant difference o f frequency
6^
between the E E and the A E , but in the case of the group as a whole,
ro N O ' co ro »-1 to -Tf ΙΛ rt* η o m so co t o ro and of toioutos considered singly, the differences between N E and
& N co (S
A E show up on the test as significant at the i-per-cent level. Table
t/i 5.7 also contains the figures for five words which are relatively more
O
(S fo t m m t> « N ro Γ*· O ' Γ'·» 0 0 O O frequent in the E E than in the N E and which between them
W h-l W Tf
N li*>
* constitute 1 32 per cent o f the E E text but only 0 96 per cent o f the
Co N E text. T h e frequency o f this group— allèlôn (each other), alios
O Tt- <- ►-( 0 0 'Φ r^· ro r ·- r o 0 0 00
Μ N «Ί 11 w N so n r o >-n r>
ro m
N N ri ^
r o r -· i n
N
ro
Tf (other), ekeinos (that), enioi (some), and heteros (the other)— in the
*0
λ A E is much closer to its frequency in the E E than to its frequency in
3 the N E . T h e differences between the A E frequency and the E E
O OO Ο O O' O' Ό (S to <n ro o *-· O ' in o o O ro
ro v© ro in frequency are non-significant in the case o f every word except
-0 allëlôn : it is clear from a glance at the table that the discussion o f the
topic o f friendship (in books 8 and 9 o f the N E and book V II o f the
^ uI _
’ t Ι Λ O '. T t O o c N ro η X Ο Ό Ο O0 T f O
E E ) causes a massive increase in the frequency o f this pronoun.
O ' τί* r o 0 0 t-ι Ο N o Ί- h 'Λ to 00 «η O ' t Ό
ÖÖ Ô Μ Μ Μ M ö O O O O O
When we compare the individual expressions between the N E and
5 Q . <3
ο ?\2 fo ΟΌΟ o ^ ïn v o r>· the A E v /e find highly significant differences not only in the case of
H + c in H 00 U-, o I'' o O i- ι Ό O0 Γ"
O' c o t J- tJ- r o <-*
allëlôn but also in the case of enioi and heteros. It will be seen from
the table that in the N E the figures for the Nicomachean group and
o the Eudemian group commonly approximate to each other ; in the
< !s
o ^ •0 0 O ' ro « (N Tj- O E E and the A E the Eudemian group in every book scores more
ro
b
o heavily than the Nicomachean group, and in the treatises as a whole
h
occurs at least twice as frequently as it.
tn
O Another pronoun which occurs significantly more frequently in
<» t oTt·«(Ti -τ|- vO
M M in Ό hh so the N E than in the other two treatises is the reflexive heautou
o μ cm ro «Ν o
o (oneself) : it is to be found eighty-three times in the N E , fourteen in
h
the A E , and twenty-four in the E E . It has not been included in
T able
(η
O Table 5.7, however, since its between-books occurrence in the N E
h
O
ΜnO Π
t Η
ΜΌΗΌ w t*· O
ro'
b
ö H ff) fN| m m ο n r>
is even more irregular than that o f allëlôn'. forty-eight o f its
»Vy
Nicomachean occurrences occur in the single book 9.
Words concerned with multitude and magnitude, wholes, parts,
o N r o T j-0 0 O' O and shares form a group which repays study in the ethical treatises.
o « S<cuu
ff) §
o 3
o S3
H Though the commonest word for ‘all’, which comes in two forms,
H > > H
120 The Aristotelian Ethics Prepositions, Adverbs , and Pronouns in the Ethical Treatises 127
pas and hapâs, does not tell us anything o f interest— there are no
T H E D E F IN IT E A R T IC L E
significant differences between the frequency o f either o f these
forms, or in the relationship between them, from treatise to One might expect in advance that if any word at all displayed
treatise— many other words in the group are favourittes o f one or regularity within the work o f an author, it would be the definite
other treatise. T h e Eudemian Ethics is more interested in wholes article, so colourless and so topic-neutral. When we compare the
and parts,1 as is shown by the figures for occurrences o f the word for use o f the different cases o f the article in the three ethical treatises,
‘ whole’ (holos) and o f two words for ‘part’ (meros, morion) : this expectation is not fulfilled. Between the E E and the A E we find
statistically significant differences between the frequencies for two
NE AE EE o f the cases o f the article, the masculine dative singular to( i) and the
masculine nominative plural hoi. Between the N E and the A E we
όλος 40 S» 35 find statistically significant differences in seven cases (the frequen
μ (ρος 12 28 14 cies for ho, hë, to, hoi, tous, tois, and tais). But the most striking
μ οριον 8 10 17 feature which differentiates the use o f the article in the N E from its
Total 60 ( = 0-15%) 76 ( = 0-45%) 66 ( = 0 25%) use in the E E is the much greater preference which the latter has for
the singular forms. T h e figures below give the proportion (per cent)
T h is interest, it will be seen, is shared in an exaggerated form by the o f each book constituted by (a) singular forms o f the article, (b)
A E , which in respect o f these words as a group differs significantly plural forms o f the article, (c) the sum-total o f both singular and
from both the other treatises. T h e Nicomachean Ethics has a liking plural forms.
for two words to do with parts and sharing (koinos, ‘common’ , and
loipos, ‘rest’, for which the figures are respectively: N E 48, A E 12 ,
Book NE i NE 2 NE 3 NE 4 NE 8 NE 9 N E 10
E E 2 1 ; N E 39, A E 3, E E 2) ; but has a more general preference for 8-32 7-86 7-96 8-14
Sing. 7-26 9-3I 7-5I
words to do with multitude (such as oligos, ‘ few ’, and polys, ‘m any’) Plur. 5-00 5-82
5-7 I 4'47 5-46 5-05 5-97
and magnitude (such as pleion and pleistos, the comparative and
Total 12-97 14-31 I 3-33 12-79 13-32 I 3-01 14-11
superlative o f polys ; and mikros, ‘sm all’, and its comparative elattön
and superlative elachistos). The figure for occurrences o f these
Book EE I EE II E E III E E VII E E V III
words are as follows. Megas (‘great’ ) with its comparative and
Sing. 9-25 954 8-73 9-98 7'97
Plur. 541 362 4-92 3-78 3-20
NE AE EE 13-76 11-17
Total 14· 66 I 3 J6 13-65
oAiyoç 27 3 8
770Aii? Γ45 26 76
ττλίωνρπλίΐστος 75 33 39 It will be seen that four out o f five o f the Eudemian books have a
μ ικ ρ ό ς 56 3 21 proportion o f singular forms o f the article o f more than 8 5 per cent.
έλά ττ ω νρςλά χισπ ος 27 31 17 Only one o f the Nicomachean books, N E 2, whose similarity to the
Total 330 ( =ο·83% ) 9 6 ( = ο ·56%) ι6 ι ( Eudem ian books has several times been remarked, has a score in
excess o f this. I f we take the plural forms, only the first o f the
superlative is also a Nicomachean favourite, but it will be more Eudem ian books has a score o f more than 5 per cent here; all except
convenient to consider it later in a different context. book 4 o f the N icomachean treatise score over 5 per cent. Here, as so
often, the disputed books display the Eudem ian features in an
I. See above, p. 66. accentuated form. T h e figures for these books are as follows :
128 The Aristotelian Ethics Prepositions, Adverbs , and Pronouns in the Ethical Treatises 129
In the second place, even the most powerful argument to show
Book AE A AE B AE C
Sing. that, considered as a single unit, or even book by book, the
10-90 9-3 I 9’37
Plur. 2-72 4 10 4-04 Aristotelian Ethics belong with the Eudemian would be insufficient
to settle the age-long scholarly dispute about the origin o f the
Total 13-61 I 3'4 I 13-41 common books. F o r, as has been remarked, many scholars favour a
solution which sees the common books as made up o f different
Altogether, the proportion o f all occurrences o f the article strata from different periods, form ing a patchwork. At the end o f
constituted by singular forms is 59-7 per cent in the N E , 69· 5 per chapter 4 we considered the occurrence of particles in each o f
cent in the E E , and 75-7 per cent in the A E . seventeen iooo-word samples o f the A E : the results were
interesting in that they showed no trace o f patchwork composition;
G R O U P IN G IN D IC A T O R W O R D S
but in the case o f many particles the frequencies were too small for
T h « words which we have so far studied amount to some 53 per complete confidence to be placed in the result. It would be desirable
cent o f the entire treatises. In each treatise, roughly 25 per cent to employ a method o f studying the small samples which would
consists o f particles and connectives, 6 per cent of prepositions, 9 generalize the inquiry to include words occurring even less
per cent o f adverbs, pronouns, and similar words, and 13 per cent of frequently than particles, while enabling firmer statistical con
the definite article. In respect o f every feature so far in which there clusions to be drawn.
was a significant difference between Nicomachean and Eudemian We can achieve both these results if we take words together in
usage the common books resembled the Eudem ian pattern more groups and count not the occurrence o f single words, but the
than the Nicomachean. Undeniably the results so far presented occurrence o f a group, the occurrence o f any word in the group
amount to a substantial argument in favour o f assigning these counting as an occurrence o f the group. T o serve our purposes, a
disputed books to an original Eudemian context. group must satisfy two conditions. First, the group must be one
However, there are two ways in which the case so far presented which is characteristic either of the Nicomachean or the Eudemian
has been incomplete. In the first place, many o f the vocabulary Ethics: it must consist in general o f words which are either
habits studied in the present chapter have been too irregular to form Nicomachean favourites or Eudem ian favourites; and the occur
the basis o f any rigorous statistical argument. From time to time a rence o f the group as a whole must be substantially more frequent in
test has been used to determine which frequency differences were one of the treatises than in the other. I f this is not so, the occurrence
significant and which were not. T h is procedure in the present o f the group within the A E will be useless as a discriminator and
context is useful as a rough discriminator: but the result o f the test will yield no information about its likely original context. Secondly,
can only be regarded as giving an accurate measure o f the the group must be one which occurs frequently enough in the
probability o f a difference in frequency being the result o f chance if treatise o f which it is characteristic to enable us to make a reliable
the data in question conform to the conditions o f simple sampling. prediction o f its occurrence in any randomly selected iooo-word
In chapter 4 a chi-squared test was undertaken to ensure that this sample o f material homogeneous with that treatise. I f this is not so,
condition was met, and it was found that in the case o f the majority it will be impossible to tell, when the actual occurrence in the
o f particles studied it was indeed satisfied. In the present chapter no sample differs from the expectation derived from one o f the rival
such test has been undertaken: but it is obvious to the naked eye contexts, whether this is due to a genuine difference o f style or
when we study the book-by-book occurrences o f several o f the merely to a chance sampling fluctuation.
expressions that more than chance fluctuations occur between one We can ensure that our groups satisfy the former condition by
book and another. It would be desirable to find a procedure which calculating, for each word and group a distinctiveness ratio.1 In the
would enable us to tighten up the statistical argument from one
treatise to another. I. T h e notion o f distinctiveness ratio is adapted from Ellegârd, 1962, 20 ff.
130 The Aristotelian Ethics Prepositions, Adverbs, and Pronouns in the Ethical Treatises 131
present context, the D .R . of an expression is the ratio o f its calculate a Eudem ian expectation. B y comparing the expected
frequency in one treatise to its frequency in the other. L e t us number o f occurrences o f each group in each case with the actual
express the D .R . o f an expression occurring in the ethical treatises occurrences, we can determine for each sample in turn whether it
as its Nicomachean frequency divided by its Eudemian frequency. resembles the Nicomachean Ethics more than the Eudemian. Any
T h u s, a word which occurs more frequently in the Nicomachean substantial patching, say, o f an original Eudemian text with
Ethics will have a D .R . greater than unity, a Eudem ian favourite Nicomachean material should display itself in a fluctuation between
will have one between one and zero, and a word which is used with one sample and another in respect o f comparative resemblance.
the same relative frequency in each will have a D .R . o f one. It turns out that it is possible, from the data that we have already
We may expect the D .R . o f many words, in two treatises by the recorded, to assemble six Nicomachean and six Eudem ian groups
same author, to be very close to one: words with such a D .R . will answering to these specifications. T h e first Nicomachean group is
obviously be useless as discriminators. For our purposes we must formed by taking together five particles popular in the N E : de,
look for words, or groups o f words, whose D .R . is some distance eiper, kathaper, hothen, and oun. Taken as a whole the group has a
from unity— say more than 1-4 or less than 0-7; words, in other D .R . of 2 00 and occurs 587 times in the N E , so that words from the
words, which occur about three times or more in one treatise for group could be expected to occur 14 85, i.e. fifteen times, in a iooo-
every twice they occur in the other. W hen we group words, it is not word sample homogeneous with the N E . T h e second group
necessary that every word in the group should have a marked D .R ., consists o f the single preposition en, with a D .R . o f 1-44 and a total
but it is important that the group as a whole should have: that is, o f 482 occurrences in the N E . T h e third group has already been
that the occurrence o f a-word-from-the-group should be a studied as a group : the qualifying adverbs hëkista, hêtton, hikarns,
considerably more frequent event in one treatise than in the other. [fan, malIon, malista, pany occur altogether 4 18 times in the N E , and
T o ensure that our group enables predictions to be made o f its the group as a whole has a D .R . o f 1 9 3 . E ven more distinctive is the
occurrence in samples as small as 1000 words it is necessary to insist group o f expressions o f tentativeness already studied : isos, dokei,
on a minimum absolute frequency as well as a characteristic relative doxeie, eoike, phainetai. T h is group occurs 404 times and has a D .R .
frequency. In practice, we need groups which will enable us to o f 2 3 7 . T h ree Nicomachean favourite pronouns and
predict an absolute occurrence o f not less than ten in each sample of demonstratives— hekastos, toioutos, and tosoutos— form another
1000 words. Since the Nicomachean Ethics is 39525 words long, this group with a D .R . o f 1 -56 and 407 occurrences in the N E . T h e final
means that we can make predictions based on Nicomachean Nicomachean group is made by putting together the two
frequencies o f 400 and above: for if a word or group occurs 400 Nicomachean favourites koinos and loipos with the ‘multitude and
times in a text o f about 40000 words it can be expected to occur ten magnitude’ group of oligos with mikros and polys and their
times in a iooo-word sample homogeneous with it. T h e Eudemian comparatives and superlatives. T h is group has a total occurrence in
Ethics is 26330 words long, and so for predictions based on the N E o f 4 17 and a D .R . o f 1-5 1.
Eudem ian frequencies we need words or groups occurring at least T h e six Eudem ian groups are made up as follows. T h e
260 times in the E E . conjunction alia is popular enough to constitute a group on its own,
W e shall look, therefore, for groups o f words with the following with 3 1 1 occurrences in the E E and a D .R . of 0 54. T h e two
characteristics. A group should be either a Nicomachean favourite connectives dio and holt form a group occurring 263 times with a
group with at least 400 occurrences in the N E and a distinctiveness D .R . of 0 58. A third group can be made up from three more
ratio o f at least 1 -4 ; or it should be a Eudem ian favourite group with Eudemian connectives, hôsper, hôste, and epei : this highly character
at least 260 occurrences in the E E and a D .R . o f less than 0-7. On the istic gr6up has a D .R . of 0 38 and occurs 279 times in the E E . We
basis o f the Nicom achean favourite groups we can calculate a have already seen that anagkê, delon, and phaneron form a
Nicomachean expectation for each o f our seventeen samples o f the characteristically Eudemian set o f expressions, but as they make up
A E , and on the basis o f the Eudem ian favourite groups we can only 0 48 per cent o f the Eudemian text they are not by themselves
T ab le 5.8 Nicomachean Favourites in Seventeen Samples o f the AE: Part One
N I {de, eiper
kathaper, hothen
oun) 587 2-00 1485 88 10 7 7719 9 9 10 12 10 10 12
N II (en) 482 14 4 12-19 10 20 5 6 5 9 ii 6 4 5 7 6 3 6 6 2 5 6-8
N III (hêkista,
he!ton, hikanos,
Itan, mâllon,
malista, pany) 418 1-93 10 58 3 2 2 2 1 5 2 1 5 3 4 3 3 6 3 6 2 3-2
N IV (isös, dokei,
doxeie, eoike,
phainetai) 404 2-37 10-22 9 2 2 2 1 4 2 3 5 4 2 2 1 2 4 3 8 3-2
The Aristotelian Ethics
N V (hekastos,
toioutos, tos-
outos) 407 1-56 10-30 4 3 6 8 2 5 2 7 9 8 7 7 1 1 2 8 5 1 5-5
N VI (oligos,
polys, pleön,
mikros, etc.,
koinos, lotpos) 417 1-51 10-55 20 12 12 5 6 5 2 5 2 2 o 2 8 7 7 3 62
Total N I-V I 27 I 5 6869 42 55 37 21 36 41 28 37 32 34 27 30 36 36 33 31 34-8
<
<N ro
\
Nw
\
60 X α ρ χή
\
*
\
\
y
^ 3 y
5 έ< X N (N ο N tO C m ro 'O
vo m O' t o ofS -< y
■ - i N N CS y
9*·*0
"Ö y
r
y
y '
-j- W<\0 1 * ro n O ' «Λ ι λ χ <N 00 o o 'o' 'o' m ir> m ro 43-800
u->
<
t η v© r o N >Λ h
\
ro t 1 0 Γ» m m t- ι t
£ w o
\
ό 0 Ö ^6 Ö
\
10 ώ ^ P S sO
'—■ e«
\
B.‘C 0 t^ o o m u
u y
O' m t <s 3 ενόεχομαι y
U
υ X X y
O. 3 0 σνμβζti VCtJ y S
S '0 Ul *
y
h j! ϊ ii O ro O OO m o t O' o m o «
y
ro 10
Ί 'δ y
*· t- 29-200
\
\
C/i
\
V )W t η m NH ο Ό Ο
** N
o
S
Vi»
3 ενέργειχ
6.2
^ ^
N Q. OO Ό Γ'·* * * ^ sO O Ο Γ-» O t N X
vVy Q-
Ί» n m r-*
14-600
T a b le
-Ô -è
S
t> 5a. ^τέλειος
Ό N η* f- 00 N O O r o OO O l> N h · Γ" sO N
« η t
" vu
S/ v2
§ N t fO n i r j u*> 10 M 1-H N vO o
c<
X οικείος
The Aristotelian Ethics The Use o f Technical Terms in the Ethical Treatises 151
N O O^toc w oo lO or-i O u"> t three words for instrumental goodbess, or utility, sympheron,
tJ- io r-*
^ i 3, ό ό ö ïïphelimon, chrësimon— are shown in T able 6.6. Once again clear
l>HMir5N Nicomachean and Eudemian preferences em erge; but for once the
O 'C S N NMo i-ifO
iO mO o IH t
Xn O' t w oo
If) N t M Lr-*
T)
A E does not completely share the preferences o f the E E . It
continues to resemble the E E more than the N E , the coefficient o f
correlation between the E E and A E usage being 0-8185 as against
Evaluative Expressions in Individual Books of the NE, AE, and EE
32 ΓΟ CO U"> HH t m n n ro o
/
/
O'' »Λ 0Ο fO N fO O O' /
O -Ι M l- H Q O t < 0 \© /
N 54-400 / ............
/
/
-3 w+ i.S Co o ' θ ' o ' Co' ro t 1> ^ 'το οο O
iO /
O ' θ ' γ-~> O ' ΟΟ GO Μ ο Ο ΟΟ ΙΑ)
ό /
ό Μ Μ μ
Hgb fii 'w ' 3 , 3 - 3 , 3 , ■
— ■'— ' 3 ,3 , /
ο ΟΟ 00 |> Μ t " Μ οο to φ ΟΟ LT1 1-1 ro O ' ι ο /
ι/Ί co -<*■ LD ΙΛ θ ' Γ" t CO Ν σ> ro Ν Ν ro r o /
t
/
/
/
ό
5* 3* Μ CO μ to o ο Μ t Ό u-i Ο IM Μ Μ Γ Μ4 40-800
/
t
rt /
/
/
/
00 * Λ T h -Η Ν ΟΟ Μ ι/ι ιλ N N /
hi /
ε η ιεικ ή ς /
/
27-200 /
/
t O N Ό /
σπουδαίο'; /
X /
/
6.6
j /β ελ τΐω ν
r*><*< *■< t U") θ ' Ο t Ο CO n r- m o
Table
Ö ro ro Π η f#; f ) |s
’ καλός
X
/
13-600 /
N N O t σύμφερόν εν
* /
/
/
<N LO N O' *-i O ro 10 O O' ' *.
κρεά τω ν
χρ ή σ ιμ ο ί
X
ta - « ro tco OO 15 m < ffl U ·« k )"S S S 5 S 0-000
S; o ^ ë ^ « > « S 11 23 38 48 56 73
H ES > H R - 0-8185 E E occurrences
I
152 The Aristotelian Ethics The Use o f Technical Terms in the Ethical Treatises 153
T able 6. 8 Scattergram o f Occurrences o f Evaluative Expressions in the groups o f expressions) while the A E and the E E are very close
NE and AE together (0 58 per cent each)— and this despite the fluctuation in
68 000
χγχθόί the frequency o f epieikès due to the fact that while normally it is
something o f a synonym for agathos it is used in the first common
book, and there alone, in a special technical sense. B u t with the
Eudemian favourites the pattern is unusual : the words are much
more frequent in both the N E and E E than in the A E , and the A E is
54-400 closer to the N E than to the E E .
/
/
/
/ GLO RY AND THE FA M ILY
/
/
*
/ I f we search through word lists o f the N E and E E , looking for
/ words with a high distinctiveness ratio, two rather surprising and
/
40-800 / interesting groups o f Nicomachean favourites emerge. We discover
/ that the Nicomachean Ethics has a consistently higher score first in
/
/ words which concern glory, honour, praise, and their converses
/
f humiliation, shame, and blame, and secondly in words for family
/
/ relationships and for youth and age. Aischros, ‘shameful’, with the
/
/ compounds formed from it, is a Nicomachean favourite, as is
χ επ ιεικ ή ς
/
27-200 / geloion, ‘ridiculous’ , and the verb psegö, ‘blame’, with its adjective
/ psektos. So, on the other hand is the verb epainein, ‘to praise’ , with
/
its adjective epainetos, and the noun time, ‘honour’, with its cognate
> σ π ο υ δ χ ΐο ς /
/ verb timân and adjective timios. T h e N E is also fond o f the adjective
/
β ε λ τ ίω ν ^ / megas, ‘ great’ , with its comparatives meizôn and megistos and of
/
/ χ κχλότ compounds formed from the root mega l such as megalopsychos and
/ εν
13-600 megaloprepës. T h e figures for the treatises are:
συμ φ έρο ν^/
/ à μεινω ν
/ NE AE EE
/
/ X ,
/ κρ α χτώ ν Occur Propor Occur Propor Occur Propor
rences tion % rences tion % rences tion %
0000 004 003
αισχρός etc. 53 0 13 7 9
Λ = 0-8066 S E occurrences γελοίος 17 004 0 ο·οο 5 0Ό2
ènaivéai etc. 64 0 16 7 0-04 31 0-12
ο·8ο66 between the N E and A E , but it does not correlate very
μ ίγ α ς ,
highly with either. T h e result o f plotting the A E uses against each μ ΐγ α λ - 039 6 0-04 80 0·30
156
o f the other treatises is shown in T ables 6.7 and 6.8. τ ιμ ή etc. 92 023 II ο·ο6 19 ΟΌ7
In this group o f expressions it will be seen that the Nicomachean φ ίγ ω etc. 24 006 8 005 12 0-05
favourites follow the pattern with which we are now fam iliar: the
Total 406 10 3 39 0-23 165 0-63
N E stands on its own (with a proportion o f 1 05 per cent for the six
I 54 The Aristotelian Ethics The Use o f Technical Terms in the Ethical Treatises 155
H ere the E E frequency for the group is just under two-thirds that
o ccu r
ren ces
ro Ό u-> NO N rs
for the N E ; the A E is significantly less than either the N E or the in NO O
în rt- U-) N NÔ O' ro
EE. m
O O u*> « tr> O
I f we take a group o f words referring to youth vs. age and family n «ri
relationships— adelphos, ‘brother’ , goneis ‘ parents’, neos ‘youth’, vO O O ΓΟ N
CH
LTi
N O "*·
pais and its cognates, ‘ child’ and ‘ childish’ , and patêr and patrikos, U~) O N iri es
ci
‘father’ and ‘paternal’, uios, ‘ son’, teknon, ‘child’, and mëtêr, 0
3
' - HH O nO O ►H
‘mother’, we again find that the group as a whole is substantially co
more frequent in the N E than in either o f the other two treatises. m N OO OO O' N
OOOI
words with a sufficiently high frequency and a sufficiently marked n N Tt" 9
S e
On Ο fn 0 0 Ö in
distinctiveness ratio. I f we consider first Nicomachean groups, we X 'fi HH
ω
find that the Nicomachean favourite metaphysical terms provide a
group o f 523 words with a D .R . o f 2 74, and the Nicomachean û μ
N
°9 NO ΙΟ
Γ*"·
sO
m N 00
%
NO
words) a group o f 4 12 words with a D .R . o f 2 1 3 . T h e Nicomachean n l-H O
0 0
10 Tj- î vO 00
evaluative words occur 4 18 times in the N E and have a
IC
distinctiveness ratio o f 1 -82 : they will form a third suitable group. A o is H 1? ~ *s? -S -Si •S s o C
a <υ .^ « ^ c a "'s "δ 5
fourth group can be made up o f the ‘glory’ words (406 occurrences, 3 SJ 6 ~ V. . -«ï
a -2 1 1 S, <u ^
with a D .R . o f 1 6 4 ) ; to make a fifth group we can take the ‘fam ily’ « ^ ü ■
S'a .8 „ ® '«■^
«i, » 5δ
c- ■a s u ü 4çj-> -S 1^ 8 β ? -«
fc *I .—
s r
words (145) and the ‘obligation’ words deo and opheilo (295 <U Ü "<Ss *- •<* jr e
u V" ur «,.$<
Si -S
2 <
£ R rf -S ^O ea Sa“ e . -5
. ~ C S
-s
occurrences) to make a group o f 440 with a D .R . o f 1 5 3 . T h e Ο & H '—■-ft <4. o — r Ê i - s <3 «
ÎS u u ^ a !§ w 3 »»Ci a1) 3. «S *
Nicomachean psychological words (2 15 occurrences) are not HN S « G ω «
C s? ίτ" ,s ^ -u 1-11 u V Ss χ __ - I*- ? ^
sufficiently numerous to make a group : we can, however, add them
*». a» g ss
•S J- l* ^ "S -«S o
to some other assorted Nicomachean favourites to make a final, s lu 1 1
H
The Use o f Technical Terms in the Ethical Treatises 157
rag-bag group. Into this bag we can throw the two Nicomachean
verbs o f saying, ero and eipon (139 occurrences), and four other
words which the N E specially favours : (a) diapherô and diaphora,
N E 92, A E 20, E E 30 ; (b) homoiös, N E 94, A E 29, E E 50; (c)
chronos, N E 40, A E 5, E E 12 ; (d) radios, N E 3 1 , A E 5, E E 8. T h is
will give us a total group with 6 1 1 occurrences and a D .R . o f 16 7 .
S ix Eudem ian groups can be constructed to match these six
Nicomachean groups. T h e first is provided by the Eudem ian verbs
o f saying, lego and phemi, plus the Eudemian verbs o f thinking,
oiomai and axioö: these give a group with 27 1 occurrences and a
D .R . of 0-68. F rom the logical and quasi-logical words we can select
the occurrences o f logos, logismos, and its verb, enantion, aporeo and
its cognates, diaireô, and orthos : this will make up a group with a
total o f 279 occurrences in the E E , and a D .R . o f 0 5 1. T h e
Eudem ian metaphysical terms give a group with an occurrence of
267 and a D .R . o f 0 51 ; if we add psyche to the volitional terms we
get a group o f psychological terms occurring 270 times with a D .R .
o f 0 64. T h e Eudem ian ethical terms occur 278 times, and make a
group with a D .R . o f 0-45; the evaluative ones 357 times with a D .R .
o f 0-75.
Arm ed with these twelve groups, we can set out to record their
occurrences in each o f the seventeen samples o f the A E with which
we are familiar from chapters 4 and 5. T h e outcome o f this exercise
can be seen in Tables 6.9 and 6.10. It will be seen that once again the
overwhelming impression is o f a greater similarity with the E E than
with the N E . In none o f the samples does the actual occurrence of
the Nicomachean indicator words reach more than 72 per cent o f
the Nicomachean expectation; in most o f them the actual
occurrence is less than half the expectation, and the mean is only
0 47 o f the expectation. In all but two o f the samples, on the other
hand, the actual occurrence o f the Eudem ian favourites is above 58
per cent o f the Eudemian expectation and in four cases the
Eudemian expectation is exceeded. T h e mean o f the actual
occurrences is 0 8 3 o f the expectation. However, in the case o f these
groups V I I - X I I , unlike the groups studied earlier, the influence o f
subject matter occasionally triumphs over stylistic preference.
W hile in fifteen out o f the seventeen cases the Eudem ian
expectation is approximated by the sample better than the
Nicomachean expectation, in two cases the Nicomachean expec
tation is better approximated. In each case, the reason is not far to
I
158 The Aristotelian Ethics The Use o f Technical Terms in the Ethical Treatises 159
seek: in the case o f sample 3 (which reaches 56 per cent o f the N - them as a group to each o f the seventeen iooo-word samples o f the
expectation, but only 34 per cent o f the Eudemian) the high A E , they indicate in the case o f each individual sample that it is
Nicomachean score is due to the twenty occurrences o f the verb deô closer to the E E than to the N E . N one o f our tests gives any support
in the discussion o f corrective justice and reciprocity. In the case o f to the theory that the A E is a patchwork o f mixed Eudem ian and
sample 6, where the actual occurrences reach 62 per cent o f the N - Nicomachean material. T h e conclusions o f these test-groups are
expectation but only 58 per cent o f the ^-expectation, the confirmed by those o f other vocabulary tests (such as the study o f
explanation is to be found in the technical use o f epieikês in the the singular and plural forms o f the definite article) which are
discussion o f equity, which boosts the score for TV-group-IX to independent o f the vocabulary elements which contributed to the
twenty-two. T h is latter anomaly in fact is the result o f the greater groups constituting the twenty-four tests.
similarity o f the A E to the E E : in the A E and the E E the use o f Since our tests have covered some 60 per cent o f the total
epieikês as a synonym for ‘ good’ is rare, because o f the reservation o f vocabulary o f the ethical treatises, it is inconceivable that their
the term for the technical use in A E A. T h at is why epieikês appears results should be overturned by further study o f the vocabulary.
as an TV-favourite and yet is common in sample 6, where the Since they have been restricted to vocabulary, they leave open the
technical sense is discussed. T h us, the two anomalies reveal the possibility that further information about the relationship o f the
limits o f the comparatively crude vocabulary tests applied rather A E to the E E and the N E may be obtained by the study o f syntax:
than any genuine Nicomachean element in the samples o f the but it is in the highest degree improbable that the over-all picture,
disputed books. Allow ing for discrepancies due to particular o f a close resemblance between the A E and the E E , and a much
subject matters (the high score o f A £ -g ro u p -X II in samples 16 and more distant relationship between the A E and the N E , should be
17 , for instance, is due to pleasure being there the topic o f seriously disturbed by further stylistic investigation, however
discussion) the over-all picture remains as before, o f general sophisticated.1 T h e most plausible and economical explanation o f
similarity between the A E and the E E . There is no support for the I. One stylistic test was carried out which does not depend on vocabulary counting: a test
o f the syntactic category o f last words o f sentences. M ichaelson and M orton have suggested
patchwork theory: the two sections which approximate the N E (1970 , 192 ff.) that a study o f the last words o f sentences (defined as portions o f text
expectation better than the E E expectation are not places which terminating with a period, colon, or question-mark) enables samples from different authors
scholars on other grounds have regarded as particularly to be discriminated. (For a criticism o f this view, see Johnson, 1973, 92 ff.) For this test,
sentences in Aristotle were divided into three classes : those ending in a noun, those ending in
Nicomachean, and in each case if the scores o f the groups I - V I (see a verb, and those ending in some other part o f speech. Each book o f each o f the ethical
above, pp. 13 2 -3 ) are added to the scores o f groups V I I - X I I the treatises was then run o ff against each other to see whether the differences between them, as a
sections emerge as over all closer to the E E expectation. whole and pairwise, were statistically significant, the chi-squared test being used to
determine statistical significance. It turns out that Aristotle differs from other G reek authors
We may conclude our stylometric examination, therefore, with who have been studied in this manner in that an unusually high proportion o f his sentences
the verdict that the overwhelming weight o f the evidence suggests end in parts o f speech other than nouns or verbs. T h e application o f the test to the ethical
that the common books are more at home in the context o f the writings give the following results : D ifferences between individual books o f the E E are non
significant (12 57 for 8 degrees o f freedom). D ifferences between the books o f the A E are
Eudemian Ethics than in the context o f the Nicomachean Ethics. Our non-significant at the i-per-cent level (chi-squared 11-82 for 4 degrees). T h e differences
twelve groups o f Nicomachean favourites and our twelve groups o f between the A E taken as a whole and the E E taken as a whole are non-significant (0 73 for 2
Eudem ian favourites have provided twenty-four independent tests degrees). T h e differences between the A E and the N E however are statistically highly
significant: chi-squared is 30-25 for 2 degrees o f freedom. It is difficult to know how to
to apply to the disputed books. Tw enty-three out o f these twenty- interpret these results, however, for it turns out that in respect o f this feature the N E itself is
four independent tests give an unambiguous answer that the highly inconsistent: taking the seven books o f it together chi-squared for 12 degrees o f
common books, considered as a whole, resemble the E E more than freedom is 72-28. T h is inconsistency is caused by books 4 and 9 : if these two books are
removed the rem ainder o f the N E is internally consistent and differs only in a statistically
the N E } W hen we take the twenty-four tests together and apply non-significant way from the A E and the E E . I know o f no other reason for regarding books 4
and g as standing apart from the rest o f the N E . Both books differ in this respect markedly
I . O nly the twelfth o f the Eudemian groups suggests otherwise. H ere the actual from a test-block o f 6687 sentences drawn from a number o f Aristotelian works in various
occurrence in the A E reaches only 43 per cent o f the Eudem ian expectation. T h is failure can fields. F o r an account o f the application o f the last-word test to the entire Aristotelian corpus,
only partly be explained through considerations o f subject matter. see K en ny, 1977.
The Aristotelian Ethics
the evidence assembled in the last three chapters is that the
common books o f the A E , just as they now stand, belonged
originally to the Eudemian Ethics. H aving established this high
probability on the grounds o f statistical stylistics as well as of CH APTER 7
external historical evidence, we must now turn to the evidence
provided on the topic by the philosophical content o f the treatises. Wisdom in the
Aristotelian Ethics
can be drawn between the N E theory and the E E theory, and which circularity in the argument for the Nicomachean context. F irst, a
o f the two fits better with that o f the E E . We shall not, in fact, contrast between the N E and the E E is built up on the assumption
attempt to draw from this comparison any positive conclusion that B belongs with the N E ; then we are told that the original home
about the relative chronology o f the treatises: for rightly under o f B must have been the N E because it fits the doctrine o f the N E
stood the data do not permit the drawing o f any such conclusions. better than that o f the E E .
Jaeger’s contrast between the sense o f phronësis in the E E and N E
The whole of V I is built up on the distinction between phronësis, the aretë
has been severely criticized by scholars a number o f times— notably
of the deliberative faculty, and sophia, the aretë of the contemplative; and
by M argueritte, Gadam er, and Leonard. These criticisms have since E E actually merges deliberative and contemplative thought, VI
gained widespread acceptance, especially since being canonized in cannot have been intended for it, unless we suppose that Aristotle took up
the commentary o f Gauthier and Jo lif.1 None the less, it is worth two opposing positions in the same work (p. 109).
while to look afresh at the concept o f phronësis in the ethics, since
even those who have discarded Ja eger’s view on this point Supposing we grant, for the moment, that the E E does not
frequently continue to accept his general theory o f the development distinguish between deliberative and contemplative thought with
o f ethical theory in Aristotle, and his assumption that the disputed the clarity that B does : even, so, before concluding that B belongs
book B is Nicomachean. T h us Row e, in his monograph The with the N E , ought not we to inquire whether the N E , outside the
Eudemian and Nicomachean Ethics, while admitting ‘ the kernel o f disputed books, makes the distinction made in B ? Otherwise, may
Jaeger’s discussion o f the ethics, his account o f the history o f the not the situation be that neither the N E nor the E E contains the
term phronësis. . . has been shown conclusively to rest on a fully worked out theory o f phronësis ; and for the simple reason that
misinterpretation o f the texts’ , goes on to argue at length that in whichever o f them B originally belonged, B was the place chosen
for the full working out o f the concept? T h e only sound procedure
there is no sharp distinction between ethics and the theoretical sciences in is to study the theory o f A E in its own right and compare with N E
E E . . . the practical and theoretical tend to merge into one another ; and and E E in turn. T h is we shall now do.
this is reflected in the close relation envisaged between speculative and
T h e doctrine o f the A E , and in particular o f book B , concerning
practical thinking. But in N E, the distinction between the two spheres is
phronësis can be summed up under four heads, (i) Phronësis is an
complete. Ethics and the theoretical sciences no longer have anything in
common, since their subject matters are now established as being totally intellectual virtue concerned with the truth about mutable matters
different in kind. Correspondingly, the rational faculty is now divided into and the whole good o f man. (2) Phronësis is the virtue o f a particular
two; the phronësis of E E becomes the areie of one half, and sophia is part of the rational soul, and is distinguished from other intellectual
appropriated for that of the other. virtues by being deliberative rather than intuitive and practical
rather than theoretical. (3) Phronësis is indissolubly wedded to
He goes on to conclude ‘ in principle, Jaeger’s view o f Aristotle’s
moral virtue, providing the right reasoning necessary for the
development was right’ (op. cit. 1 1 , 70-3).
exercise o f virtue, and dependent on virtue for the correctness o f its
L ik e Jaeger, Row e regards the disputed book B as Nicom achean;
own starting-points. (4) T h e union o f phronësis and moral virtue is
he regards indeed all the disputed books as being Nicomachean as
dependent on the pre-existence o f certain natural qualities,
they stand: books A and C , he thinks, are based on and supersede
intellectual and affective. L et us develop each o f these points in
lost Eudem ian parallels, but B , he thinks, was cut from whole cloth
turn, and compare the A E doctrine with that o f its competing
for the N E . As with Jaeger, so with Row e, there seems a certain
contexts.
i. ‘ Une meilleure intelligence de la “phronësis” de l’Ethique à Nicomaque . . . jointe à une
meilleure compréhension de l’évolution doctrinale d’Aristote et à une connaissance plus exacte
(1) Phronësis as an intellectual virtue. ‘ In classifying the
de la chronologie de son oeuvre . . . achève ainsi de ruiner l’hypothèse de Jaeger sur excellences o f the m ind,’ we are told at the beginning o f book B , ‘ we
l’évolution du concept de phronësis entre PEthique a Eudème et l’Éthique à N icomaque : ici et said that some were o f the moral character, and others o f the
la, c’est la même sagesse, pratique mais normative, que désigne le mot de phronësis’ ( 1959, ii,
469). intellect. Moral excellences we have discussed at length ; let us now
The Aristotelian Ethics Wisdom in the Aristotelian Ethics
discuss the others’ ( i i 3 9 ai - 2 , trans. Greenwood). T h e nature o f these are first introduced (at 1220 *4 -14 ) and indeed it appears at
the aretê o f anything depends on the ergon o f that thing: the ergon of 12 2 1* 12 in what looks like a list o f moral virtues. But in every
every intellectual part is the production o f true and false other respect the doctrine o f the A E is to be found more clearly in
judgements (i 139*29). That, at least, is its ergon in the sense o f its the E E than in the N E . From the last book o f the E E it is clear that
characteristic activity, its ergon whether it is working well or ill ; its phronësis is an intellectual virtue: it is a virtue (i24 0 b35) which is in
activity when it is working well, its good activity, and therefore its the logistikon or ratiocinative part o f the soul ( 1 246b23) ; it is in the
ergon in another sense, is truth alone ( i i 3 9 bi2). T h e intellectual ruling part o f the soul (i240bi2) and it gives commands (epitattei,
virtues are then excellences which make an intellectual part o f the I249bi5). All this accords with the description o f intellectual
soul come out with truth. There are five states o f mind which have virtues as belonging to the rational part which qua rational is in
this effect— technë, epistëmë, phronësis, sophia, nous, which we may command o f the soul (tou logon echontos ho epitaktikon esti tës psychés
translate as skill, knowledge, wisdom, learning, and hë logon echei, 1220*9). As in the A E and the N E , an aretê makes the
understanding— and so five candidates for being intellectual ergon o f its possessor good : as in the A E , the ergon in question is the
virtues. Being truth-productive is a necessary but not a sufficient production o f truth. We are told at i 2 2 i b29 that there are aretai
condition for being an intellectual arete : skill fails to pass some o f corresponding to different parts o f the soul: the aretai o f the
the other tests for being an aretê ( ii 4 0 b2 2 -4 ); the other four rational part are the intellectual virtues ‘ whose ergon is truth, either
qualities remain as virtues, though not all independent o f each other about how things are, or about bringing into existence’ . 1 T he
( i i 4 i a2-8 , b3~ 5 ; 114 2 * 2 6 -3 1). Phronësis is defined as a ratiocina distinction between these two kinds o f truth fits perfectly the
tive, truth-attaining quality concerning what is good for human distinction between the subject matter o f phronësis and sophia in
beings ( ii4 0 b5, b2 i) . T h e fact that it is ratiocinative, and has the book B : sophia, we are there told, unlike phronësis, is not concerned
particular subject matter it has, distinguishes it, as we shall see, with bringing into existence (oudemias esti geneseös, i i 4 3 b2o).
from the other three intellectual virtues. T h us far the A E . Phronësis is ratiocinative in the E E : ‘ It is clear that [lucky people] do
In the N E , too, we find that phronësis is an intellectual virtue : it is not succeed by means o f wisdom, because wisdom is not
listed as a paradigm o f such at 1 i0 3 a6 when the distinction between unreasoning but can give reason why it acts as it does’ (1247*14).
moral and intellectual virtues is introduced. T h e relation between Finally we are told in the first book that the good which is the end o f
having an aretê and doing one’s ergon well is also familiar in the N E human action is the subject matter o f the supreme discipline o f
( i098a8 ff. ; 1 10 6 *17 ff.) ; but we are not told the particular ergon o f politikëkai oikonomikëk a iphronësis. These three hexeis, we are told,
the intellectual part o f the soul. We are, however, told implicitly are alike in this respect: how they differ will be explained later
that phronësis is ratiocinative— is concerned with a logos— since ( i 2 i8 bi5). T h e forward reference well fits a passage in B,
moral virtue is defined as consisting in the observation o f a mean i i 4 i b24 ff. There we are told that politikë and phronësis are the same
‘ which is defined by a logos such as a wise man would use to hexis with a different einai : different applications, we might say, o f
determine it’ (110 7*2). Since we learn from here, and from the final the same mental quality. Wisdom applied to one’s own affairs is
book (117 8 * 10 ), that it is wisdom which sets the standard for the phronësis in the narrow sense ; wisdom applied to the affairs o f the
moral virtues, we might well conclude that the subject matter o f polis or to the affairs o f one’s household is phronësis in a broad sense.
wisdom is the whole good o f man ; but this is not spelt out in the N E Politikë, too, has broader and narrower senses : in the broadest sense
as it is in the A E , though we are o f course told, right from the it includes the wisdom o f the legislator, o f the executive, and o f the
beginning, that there is a discipline called politikë which has human judiciary ; more especially the word applies to the two latter, and par
good as its subject matter (io g4a27). Bu t the relationship between excellence politikë is wisdom in executive decision ( i i 4 i b3o).
this and phronësis is left obscure. (2) Phronësis and other intellectual virtues. In the A E one o f the
L e t us now turn from the N E to the E E . Here we find that
phronësis is not given as an example o f the intellectual virtues when I. ώ ν € ρ γ ο ν ά λ ή θ ίία , η π ε ρ ί τ ο υ π ώ ς « χ€ΐ, 17 xrept y e v e a c ω ς.
The Aristotelian Ethics Wisdom in the Aristotelian Ethics 167
main functions o f phronësis, one o f the characteristic activities o f the accepting V ictorius’ emendation o f the manuscripts). Both treatises
phronimos, is deliberation, bouleusis: this is said not only in the call this an irrational part o f the soul (N E i i0 2 a3 2 : tou alogou to
endoxon which introduces the principal treatment o f phronësis men ; E E i 2 i g b32 : heteros alogon morion), and both o f them dismiss
( ii4 0 a2 5 ff.) but also in the passage contrasting phronësis with it as irrelevant to ethics. T h e irrational part o f the soul that is o f
sophia ( i i 4 i b9 ff.) and in the discussion o f euboulia ( ii4 2 b32). importance is one which, unlike the threptikon, is under the control
Wisdom is concerned with human affairs and with matters that can be o f reason: in the N E it is described as participating in reason but
objects of deliberation. We regard good deliberation as the pre-eminent distinct from reason, fighting reason in the incontinent man,
characteristic of the wise man, and no one deliberates about things that obeying reason in the continent man, and harmonizing with reason
cannot be other than they are, or that are not directed to an end, a good in the virtuous man ( n o 2 bi 4 - n o 3 a2); in the E E too it is said to
attainable by action (i i4 ib9—13). participate in reason, having by nature the power to obey and listen
and to follow the rational part ( i 2 i9 b2 8 -3 o ; i220 ag - i 1). In the
It is because wisdom is concerned with things that can be other than N E this is introduced as part o f the irrational soul, but we are later
they are that it is the virtue o f a different part o f the soul from sophia told that you can if you like call it rational, in which case there will
or learning, which is concerned with unchanging and eternal be two rational elements ( 1 i0 3 a2) ; in the E E , on the other hand, it is
matters. Book B began with a distinction between two parts o f the introduced as a part o f the rational soul ( i2 i9 b28), but is then later
rational soul, based on the principle that like is known by like: the described as an irrational part (12 2 0 “ 10 ); but it is clear that there is
logon echon which was previously distinguished from the alogon is no real difference here between the doctrines o f the two treatises. In
now divided into the logistikon (named from the sense o f logizesthai the N E it is also called the epithymetikon kai ho lös orektikon : the part
in which it is equivalent to bouleuesthai, deliberate) and the o f the soul for desire and appetite ( x i0 2 b30) ; and similarly in the E E
epistëmonikon which is concerned with the eternal truths. the part which is ruled by reasoning is the part concerned with
Repeatedly we are told that each o f these parts has its proper aretë : orexeis kai pathëmata, appetites and passions (i2 2 0 ai) and it is
sophia, apparently, for the epistëmonikon, and phronësis for the called aisthetikon kai orektikon ( i2 i9 b22)— an anticipation o f the
logistikon now also called doxastikon ( 1 i3Qai 6 ; ii4 0 b20; i i 4 3 bi 5 - name the scholastics gave it when codifying Aristotelian psy
17 ; i i4 4 a2~ 3; i i4 5 a4). T h e other intellectual virtues turn out to be chology, the ‘ sensitive appetite’ . It is ‘the part o f the soul that is
parts o f either sophia or phronësis, and to be thus located in one or capable o f following in accordance with prescriptive reasoning’
other o f the soul-parts mentioned : there is no need to go beyond a (i2 2 0 b5). In each o f the treatises the ethical virtues are virtues o f
bipartition o f the rational soul. Th us, sophia consists o f nous (of one this part o f the soul : we gather this in the N E from being told that
kind) plus epistêmë ( i i 4 i b3 -4 ): these two, then, will be, appro the distinction between ethical and intellectual virtue is based on
priately enough, in the epistëmonikon. Similarly, gnome', synesis, and this distinction o f parts o f the soul ; in the E E we are told explicitly
nous (of another kind) share the same subject matter as phronësis that an ëthos is a quality o f this part o f the soul (i2 2 0 b20) and the
(1 i4 3 a25~b6) and therefore inhere in the same part o f the soul as it connection o f this with the notion o f ëthikë aretë via the concept o f
and seem not to be separable from it as a distinct virtue (compare pathos or passion is spelt out at great length (i2 2 0 b8 - i 2 2 i b38). T h e
i i 4 3 a25 with H 4 3bi5 ff.). Aristotelian alogon metechon logou corresponds to the two lower
T h e psychology o f the A E , and especially o f B , is easy to parts o f the tripartite soul o f Plato’s republic: the epithymetikon and
reconcile with that o f either the N E or the E E . Both o f these work the thymoeides. A trace o f this Platonic anatomy o f the soul can be
regularly with a tripartite soul : a vegetable element, an appetitive found in the N E 3, at n i 7 b24, where it is said that bravery and
element, and a rational element. T h e vegetable element, whose temperance ‘seem to be the virtues o f the irrational parts’ ; this
ergon is nourishment and growth, is called in the last chapter o f N E probably also gives us a clue why this part o f the soul is called the
i and in the first chapter of E E II the threptikon ( n o 2 bi 1 ; i 2 i g b22 epithymetikon kai holos orektikon in the N E — the other orexis which
and 38) ; the name phytikon is also used (N E 1 i0 2a32 ; E E i 2 i 9 b38, Aristotle has in mind is thymos. In the E E , on the other hand, orexis
ι68 The Aristotelian Ethics Wisdom in the Aristotelian Ethics
is divided into three species: epithymia, thymos, and boulësis bouleuesthai; the logistikon is called also the dianoëtikon kai
( i2 2 3 a27). T h e first two are no doubt the hormai, or drives, o f the praktikon ( 1 1 39a29) and this term in its turn appears in B to be
irrational part spoken o f at i24 7b9; boulësis, on the other hand, synonymous with dianoia heneka tou kai praktikë ( i i 3 9 a30). T he
which involves a judgement that something is good, is no doubt the logistikon is in B also called the doxastikon ( ii4 0 b26); in an obiter
same as the hormai apo logismou o f i2 4 7 bi9. H ere too is the germ o f a dictum in the E E treatment o f friendship also we learn that there is a
later scholastic faculty: the intellectual appetite or will. special part o f the soul for doxa (i2 3 5 b29).
N either the N E nor the E E makes a distinction between two It is, however, the final books o f the N E and E E that it is most
parts o f the rational faculty with the clarity o f the A E , and indeed instructive to compare with book B in respect o f psychological
both o f them, when distinguishing between the rational and terminology and theory. In N E 10 happiness is identified with
irrational part o f the soul, describe the rational part in terms activity in accordance with sophia (1 i79 a32), and sophia is described
appropriate to the logistikon o f book B rather than to the whole as something which is related to philosophy as knowing is related to
bipartite logon echon o f that book : the E E says that it participates in seeking (1 i7 7 a24-6). T h is fits well the sophia o f book B , which is
reason by prescribing (tö(i) epitattein, i2 i9 b3o), the N E introduces illustrated by famous philosophers and which has as its subject
it as what ‘ exhorts to the best’ (i i0 2bi6). However, it is not difficult matter divine, honourable, and useless things. As in B there is a
to insert the divided rational soul o f B into the slot provided by the contrast between learning and wisdom as between theory and
rational soul in the framework o f the early books o f the N E and the practice, so in N E 10 we are told that the sophos can act in isolation,
E E . In these early books, whichever set o f them is rightly taken as while phronësis appears along with moral virtue as a component o f a
leading on to B , Aristotle would merely be refraining, wisely, from second-rate type o f happiness which is essentially social ( 1 1 77a3 1 ;
introducing a greater number o f technical divisions than he needs at ii7 8 ai6 -22). But there is in addition to this resemblance between
a given stage o f exposition. Whether the s i s is taken as belonging A E B and N E 10 a great difference. Whereas in B sophia is the virtue
with the N E or the E E we have the same didactic procedure as in o f a particular part o f the soul, and there is no suggestion that the
Plato’s Republic : a division o f the soul into parts in a provisional epistëmonikon is transcendent or immortal, in 10 sophia resides not
manner, followed by a further division o f the topmost part into sub in the soul but in the nous, which is divine, or the divinest thing in
elements geared to different subject matters. (Compare Republic us; the life o f nous is superhuman and is contrasted with that o f the
439 d and 478 a.) syntheton or body-soul compound; the moral virtues and phronësis
In the early books o f the N E sophia appears from time to time are virtues o f this compound, but the aretë o f nous is capable o f
alongside phronësis ( i098b24 ; 110 3 * 5 ): whether it is used as a quasi separate existence ( 1 1 7 7 s 14, b2 6 -g ; 78s 14-20). M oreover, it is in
synonym, or to mark a contrast, and if so on what basis, is not clear. this nous that we must see the essential human being ( 1 i78 a2 ff.) : a
In the early books o f the E E there are two hints o f the distinction contrast with B , where a human being is defined as essentially a
between logistikon and epistëmonikon made explicitly in B . T h e first chooser ( i i 3 9 b5)·
is the passage at i 2 2 i b29 already cited where we are told that the With an effort, it is no doubt possible to regard the nous o f N E 10
output o f the intellectual virtues is truth ‘either about how things as being none other than the epistëmonikon o f A E B ; and certainly
are or about bringing into existence’ : in terms o f the psychology o f the tension between the conception o f human nature dominant in
B , truth about how things necessarily are would be the output o f the 10 and that o f B can be found within xo itself (compare i i 7 7 b27
epistëmonikon, and truth about bringing into existence would be with i i78 a7). T h e fact remains that though B uses nous in a variety
that o f the logistikon. T h e logistikon also could very well be referred o f senses, it never uses it in a way resembling that o f 10.
to in E E i220 b25 by the expression to bouleutikon tës psuchës, which Nous in B is often used, like and sometimes with dianoia, as a
we are told is the one which studies the final cause or wherefore (hou general word for the whole human intellectual apparatus, for the
eneka) o f such things as walking; for when the logistikon is cognitive as opposed to the affective half o f the mind. T h u s in the
introduced in B we are told that logizesthai is equivalent to second chapter o f B we are told that the three things controlling
The Aristotelian Ethics Wisdom in the Aristotelian Ethics 171
truth and conduct are perception, nous, and orexis, and man is that there is a nous dealing with ‘ the other premiss in practical
defined as appetitive nous (113 9 * 17, b5). But commonly nous is a reasoning’ (en tais praktika is. . . tes heteras protaseös) I take it that
name for an intellectual hexis : and close inspection o f the text o f B he means a nous dealing with the major premiss o f practical
reveals that there are four different intellectual conditions which reasoning, the premiss which contains reference to the end : this will
bear the name. correspond to the intuitive understanding o f speculative axioms,
First, and most explicitly, nous is the grasp o f the first principles since axioms in mathematical reasoning, play the same role as the
o f theoretical science; the understanding o f unproven necessary wherefore, or hou heneka, in practical reasoning (A E C , i i 5 i ai7).
truths which is the basis o f epistëmë (1 i40b3 i - i i 4 i a9). Together T h e phrase I render as ‘the other premiss’ is commonly rendered
with epistëmë., we are told, this nous constitutes sophia ( i i 4 i ai9). ‘the second premiss’ : most commentators take it that there is just
B u t there is also another theoretical nous, which is concerned with one practical nous, concerned with the changeable particulars which
individuals and is a kind o f perception ( 1 i4 3 b6) : this is the kind o f are the topic o f the minor premiss o f a practical syllogism. T h is
understanding which we must have o f a particular instance if we are makes irrelevant the introductory sentence ‘Nous is concerned with
to grasp it as an instance o f something general; the kind o f vision extremes in both directions' and renders unintelligible the feminine
whereby we see that a triangle is the simplest figure into which more plural hautai in the sentence ‘ these are principles o f the wherefore’ .
complicated figures can be analysed ( 1 1 4 2 “29). What makes it I f we take it that Aristotle is saying that nous is concerned with both
appropriate to use the same name for the grasp o f very general and extremes in both realms, practical and theoretical, then all becomes
abstract principles and for the appreciation o f features o f an clear: we have universal speculative nous (the nous ‘kata tas
instance or diagram before one’s eyes? It is appropriate, Aristotle apodeixeis ton akinëtôn horön kai proton1 o f 1 i4 3 b2 and 1 i40b3 i ff.),
would say, because in each case we are dealing with something particular speculative nous (the nous o f particulars which is aisthësis,
immediate and intuitive: in each case there is a contrast with proof o f i i4 3 b6 and 1 i4 2a29), particular practical nous (the nous which is
and reasoning. ‘en tais praktikais tou eschatou kai endechomenou’ o f i i 4 3 b2), and
Corresponding to these two types o f theoretical nous there are universal practical nous (the nous o f the hetera protasis o f i i 4 3 b3).
two types o f practical nous. Such, at least, I take to be the message o f A s in speculative matters, so in practical matters, nous provides the
one o f the most difficult passages in the whole o f the Ethics, starting-point and the stopping-point for reasoning ( 1 i4 3 bi 1). T h e
1 1 43a35_t>4 : practical reasoning which is the existence o f phroriësis begins with a
conception o f an end, a hypolëpsis about the purpose o f conduct
Nous is concerned with extremes in both directions: it is nous and not ( ii4 0 bi7 ); this right conception o f an end can itself be called
reasoning whose objects are primary terms and ultimate particulars. There phronësis ( ii4 0 bi 3 ; i i4 2 b34). Practical reasoning ends with a
is one nous of the unchangeable and primary terms in the realm of
judgement about what is to be done, a self-addressed command
demonstration, and another of the ultimate and contingent element, and of
(114 3*9) or a piece o f advice to another ( 1 1 4 3 “ ! 5). T h is too can be
the other premiss, in practical reasoning; for these are the principles of the
wherefore; for universale come from particulars, and for these you have to called nous (114 3*27) whether or not it is backed up in a particular
have a perception which is nous. case by a statement o f reasons ( i i 4 3 bi4).
There is thus a parallel between the structure o f phronësis in the
It is common ground among commentators that in this passage logistikon and sophia in the epistëmonikon which can be illustrated as
Aristotle is teaching that there is a practical form o f nous which follows :
corresponds to the particular nous on the speculative side. But in
epistëmonikon logistikon
claim ing that there are here two forms o f practical nous— one
{
corresponding to the general abstract nous o f the speculative nous (of axioms) nous (of end) Λ
faculty, as well as one corresponding to the particular nous— I am logos = epistëmë logos—euboulia yphronësis
departing from the majority o f commentators. When Aristotle says
nous (of particulars) nous (of particulars)J
172 The Aristotelian Ethics Wisdom in the Aristotelian Ethics 173
T h e pattern nous-logos-nous is common to both sides o f the be contrasted with and judged superior to the whole soul-body
division. Ju st as sophia can be described as epistëmë plus nous compound as it is in N E 10 at i i7 7 b28. B u t if we compare the A E
(‘ knowledge with a head on it’ i i 4 i ai9), so phronësis could have use o f nous with that o f the E E we find no such difficult contrast as
been described as euboulia plus nous : instead, euboulia is described we find when we compare the A E and the N E . Nous is uncommon
as phronësis minus nous, as what you are left with if you discount that in the E E until the final book : in the earlier books it appears only in
element o f phronësis which is the right conception o f the end the popular phrase noun echön (having common sense) and in
(1 i4 2 b32~4). T h e particular nous is, in each case, the perception o f conjunction with theos in the refutation o f the Platonic Idea o f the
an instance (of a term or a principle) as an instance : this may be Good ( i 2 i4 b3 i ; i2 3 7 b38 ; i 2 i 7 b3 i). In the final book, however, it
either with a view to acquiring the general term or instance (e.g. appears a number o f times, in senses which match those we have
n 43b5) or to applying it (e.g. U 4 3 bi5). Aristotle’s fullest account seen in the disputed book B.
o f particular nous comes when he is contrasting phronësis with T h us at 1 246“ 1 3 - 14 the incontinent man is described as an
epistëmë : intemperate man with nous (pös akolastos . . . echön noun) : this is no
doubt a popular saying o f some kind, but it suits very well the
That wisdom is not knowledge is clear : for as has been said it is concerned theory o f B and C , according to which the incontinent man differs
with particulars, since action is particular. It is something which perceives from the intemperate man by having a correct appreciation o f the
by nous ; for nous is concerned with terms of which there is no deduction,
end— the nous tës heteras protaseos o f 1 14 3 b3 ■ At i24 7a3o nous
and wisdom with particulars of which there is no science but only sense
appears along with nature and tutelage as a possible cause o f
perception—not the perception of the proper sensibles, but that whereby
we see in a mathematical case that a triangle is an ultimate element, successful action: here it must mean generic intellectual ability,
another stopping-point. This wisdom is perhaps better called perception : including both phronësis and technë. A t i24 6bio epistëmë and nous
it is a different kind from the wisdom we were talking about earlier. are linked together as they are in the account o f sophia in B ; at
( 1 1 4 1 a^4 3 1)1 i24 8 ai 7 f f . nous and bouleusis— the constituents o f phronësis
according to our account o f B — appear together as the conditions o f
Aristotle nowhere gives an example in the A E o f what is perceived
right desire. A t 1 24 8 ^ 9 we are told that (moral) virtue is the tool o f
by this phronësis-nous: he relies on the analogy with the
nous, just as at i240bi 1 we were told that phronësis was the employer
mathematical case o f the analysis o f a figure and with cases o f the
rather than the employee o f virtue. In all these passages in book
exercise o f a skill such as medicine ( 1 1 4 i bi7 ff.). Presumably
V I I I o f the E E nous appears in senses familiar from B : either as the
instances o f the kind o f thing he has in mind would be the
generic cognitive apparatus, or as the intuitive element in learning,
perception that A needs cheering up, that B is being offensive, and
or as the intuitive element in wisdom.
that one is hurting C ’s feelings : an awareness, we might say, o f the
It is the final chapter o f the E E that is the closest parallel to N E
morally relevant features o f one’s action and the situation in which
10, and here the word nous is not used. Aristotle is looking for the
one is acting. Som eone who is blind to this kind o f thing— who does
criterion which the good man should use in his choice o f the good
not, as Aristotle puts it, have an eye for this kind o f thing ( 1 i4 3 bi4,
things o f nature and fortune. H e says :
1 I 44a3 ° ) — will never be able either to acquire or to apply general
principles about the good life or admirable action. Here as elsewhere one should conduct one’s life with reference to one’s
In all these passages nous is an acquired or natural condition o f superior, and more specifically to the quality of one’s superior’s activity. A
the soul or one o f its parts : it is not a substance or faculty which can slave, for instance, should look to his master’s, and everyone to the
superior to whom he is subject. Now a human being is by nature a
i. Reading α ν τ ιλ η π τ ικ ό ν in 1142*25 with Kb, and understanding it in an active sense, as compound of superior and inferior, and everyone accordingly should
suggested by the correction in the manuscript Ο3 α ν τ ιλ η π τ ικ ή ; reading also at 1142*29 30 conduct their lives with reference to the superior part of themselves.
α ΰ τ η μ ά λ λ ο ν α,ΐσθησις ή φ ρ ό ν η σ ις with the majority of the best manuscripts and taking
α ΰ τ η with φ ρ ό νη σ ις. So read, the text agrees with, rather than contradicts, the passage at However, ‘superior’ is ambiguous: there is the sense in which medical
1 1 4 3 * 3 5 ff· science is superior, and the sense in which health is superior; the latter is
174 The Aristotelian Ethics Wisdom in the Aristotelian Ethics 175
the raison d’être o f the former. It is thus, that matters stand in the case of so that it is not surprising that it has not been taken up by other
our intellectual faculty. For God is not a superior who issues commands, commentators. M ?n y distinguished scholars, however, while
but is the raison d'être of the commands that wisdom issues. But ‘raison rejecting the proposal as an emendation have accepted it as an
d'être' is ambiguous, as has been explained elsewhere—this needs saying, interpretation : the word ‘ G od’ here, they claim, does not refer to a
since of course God is not in need of anything. To conclude: whatever transcendent G od o f the universe, but to the nous immanent in each
choice or possession of natural goods—health and strength, wealth, o f us, which is described in N E 10, at i i 7 7 a2 8 - 3 i, as something
friends, and the like—will most conduce to the contemplation o f God is divine. I f this interpretation is correct, then the expression I have
best: this is the finest criterion. But any standard of living which either translated ‘the contemplation o f G o d ’ must be translated as ‘ G o d ’s
through excess or defect hinders the service and contemplation of God is
contemplation’— the genitive must be taken to mark the subject,
bad. (i249b6 -2 i)‘
not the object o f the contemplation; and the phrase rendered ‘ the
T h e word ‘ nous', as I have said, does not appear in Aristotle’s text, service and contemplation o f G o d ’ needs to be taken accordingly as
but von Arnim proposed to insert it in place o f each occurrence o f ‘ the service o f G od, and contemplation’ . T h us there will be found
the word ‘ G o d ’ , whose presence he attributed to a Christian in this passage a twofold division o f the intellectual faculty : one part
interpolator. T h e proposal has no support from any manuscript or in which inheres prescriptive wisdom, the other which is called
version, and is implausible in itself— what Christian would assent ‘ god’ and issues in contemplation. T h e twofold division will
to the proposition ‘G od is not a superior who issues commands’ ?— correspond to the division between the logistikon and the epi
stëmonikon in B , while the divine status accorded to the con
i. M y translation of this passage, like every other translation that has been offered, is
controversial at a number of points. I follow Bekker’s text exactly, even with respect to
templative nous gives the present passage a close resemblance to N E
punctuation; with the exception of the misprint ά ρ χ ο ν τ ε ς for ά ρ χο ν το ς in line 8 . None of io .1
the emendations that have been proposed seems to me to be necessary to make the text Ingenious though it is, this interpretation is not in the end
intelligible, and few of them seem even to improve it. ‘Superior’ is not always the most
credible. T h e previous chapter o f the E E contains several
natural translation of α ρ χ ώ ν but seems the best English word to capture the two senses
present in the Greek of ‘higher in a scale of value’ and ‘higher in a chain of command’ (as in references to G od , which are clear references to the cosmic G od o f
‘superior officer’). There is no reason to emend π ρ ο ς τ η ν ε ξ ιν κατά τ η ν εν ε ρ γ ε ία ν inline7 ; the universe, who is not to be identified with, but is explicitly said to
what is meant is that the slave should help his master to do whatever his master is doing as
be more powerful than, nous and science. T h is God is contrasted
well as it can be done: he should strive to see that his master’s ε ν έ ρ γ ε ια has the ε ξ is of
goodness. The most difficult clause to make sense of as it stands is έκ α σ το ν δή δεο ι π ρος with ‘the divine in us’ , which in that context appears not to be any
τ η ν έ α υ τω ν α ρ χ ή ν ζην. In this clause I think 8η should not be emended: Aristotle is speculative faculty, but human reason in general.2 It is hard to
drawing a conclusion, not adding a second premiss. The genitive plural εα υ τώ ν is puzzling:
accept a quite unprepared switch to a use o f ‘ G od ’ to refer to the
Rackham’s emendation i v α ν τω is attractive, and certainly expresses clearly what I take to be
the sense. The plural is perhaps to be explained, as my translation suggests, as looking back to faculty o f contemplation : this would be so even if there were any
the implicit plural in έ κ α σ το ν ; or it may be due to the self being looked on in this passage as a parallel for such a use, which there is not.3 Bonitz offers no parallel
plurality of parts. The genitive is as it were partitive: ‘that part of themselves which is an
α ρ χ ή ’. I am unable to settle the parallel ambiguities of ο ϋ τω and θ ε ω ρ η τ ικ ό ν in the
either for the subjective genitive with theoria ; and it can hardly be
sentence ο ν τω δ ’ί χ ε ι κ α τ ά το θ εω ρ η τικ ό ν . Does ο ϋ τω mean ‘like the foregoing (viz. said o f the speculative faculty that it ‘ is in need o f nothing’,
“ superior” ) in being twofold’ or ‘like the foregoing (viz. health) in being a raison d ’etre' ? Does
θ εω ρ η τικ ό ν mean the intellectual faculty in general, or the speculative faculty in particular ? 1. See D. Wagner, 1 9 7 0 , passim.
Elsewhere in the E E the use of the verb θ εώ ρέω and its derivatives would support either The text is in part corrupt and is not easy to restore completely with
2 . I 2 4 8 * i 5 - b4 .
reading: the wide sense is to be found at 12 i7b26 ff. ; i226b26; 1245*23, and the narrow sense confidence. Fortunately I do not need to discuss its emendation—other than to say that I
at I 2 i 4 “ i 3 ; I 2 i 5 b i 3 ; I 2 i 6 b i 3 ; I 2 i 9 a i 7 , b2 6 ; I 2 2 i b 5 ; 1227 ' % b2 Q ; other passages are accept Spengel’s reading of “28, τι' οι!ν αν κ ρ ε ΐτ τ ο ν κ α ι ε π ισ τ ή μ η ς εϊη κ α ι νοΰ π λ η ν
themselves ambiguous or use the verb in the literal sense o f‘see’. 1 incline to think that the θ εό ς, which is supported by the old Latin version of the De Bona Fortuna— for the points
faculty meant here is the theoretical intellect, the ε π ισ τ η μ ο ν ικ ό ν of B : this fits the made above remain true on any of the proposed reconstructions including von Arnim’s.
commoner use of the verb in the EE. But provided that the first sense o f‘thus’ is taken with 3 . D. Wagner has devotedly combed classical literature for descriptions of nous, and other
the broader sense of ‘intellectual faculty’ , or the second sense with the narrower sense, there abstractions, as divine (op. cit. 1 2 8 - 3 9 ) . But describing something as divine, or as a god, is
is no difference in the over-all resulting significance of the passage. I owe to Professor D. J. quite different from using the word ‘god’ baldly to refer to it. Politicians and leader writers
Allan the neat translations of ου εν εκ α as raison d’etre and of σ ώ μ α τ ο ς α γ α θ ά as “ health and frequently say that inflation is a cancer ; but if I read that someone has died of cancer, I do not
strength” . take this to mean that his death was due to the depreciating value of money.
176 The Aristotelian Ethics Wisdom in the Aristotelian Ethics 177
particularly in a context in which— on this view— wisdom is said to commands. We thus see the appropriateness o f the ‘for’ : it explains
be issuing commands for its benefit. Finally, it is possible to make why the distinction just made had to be made. It is indeed true, even
perfect sense o f the passage without giving the word ‘ G o d ’ any such according to the final chapter o f the E E , that G od is the supreme
strained sense. archë o f the soul : but not in the sense in which the analogy with
When Aristotle says that a human being is by nature com slave and master would suggest.
pounded of superior and inferior, it is not the composition o f soul Shall we say, then, that it is G od who stands to wisdom in the
and body that he has in mind, but the division between the relation that health stands to medical science? T h e rule about
irrational and the rational part o f the soul. I f we apply to this case the looking to one’s superior would then be applied in this w ay: we
general rule that one should look to one’s superior, we infer that a must obey the commands o f wisdom, and take account o f the needs
human being should govern his life in accordance with what is o f G od. But that in turn would be absurd : since God o f course is not
required by his rational soul. But there are two different kinds o f in need o f anything. T o prevent this absurd conclusion being drawn,
arche : in matters concerning our body we need to take account o f Aristotle has to make a further distinction. Health was related to
the requirements (i.e. prescriptions) o f medical science; we also medical science, we said, as its raison d'etre or hou heneka. But the
need to take account o f the requirements (i.e. needs) o f our health. hou heneka o f an institution or activity may be either that which it
Both medical science and health are archai in different ways : and seeks to attain, or that which it seeks to benefit. Aristotle often
the latter is the raison d'être o f the former. W e must apply this to the invokes the distinction, without offering a clear instance to illustrate
mind : there too there is an arche like medicine, and an arche like it (Phys. I I .2, 19 4 *35-6 ; De An. II.4 , 4 i5 b2 -3 , b2 0 - i ; M et. Λ ,
health. T h e archë corresponding to medicine is wisdom, which like i0 72b2). Everyday examples o f what he has in mind are easy to find :
medicine issues commands. So applying the rule about looking to if I go into a toyshop to buy a toy railway for my young son, the
one’s superior, we must obey the commands o f wisdom, and take railway is the hou heneka o f my shopping trip in the one sense, the
account o f the needs o f that which is related to wisdom as health is boy is the hou heneka in the other sense. M et. Λ , i0 7 2 bi applies the
related to medical science. So far— up to 1 2 ^ 1 4 .— all is reasonably distinction to the F irst M over: as Ross paraphrases the passage, ‘ A
clear. But now we naturally want to ask: well, what is it that stands final cause in the sense o f that whose good is aimed at cannot be
in the same relation to wisdom as health does to medical science? found among unchangeable things, but a final cause in the sense o f
Instead o f giving us an immediate answer, Aristotle goes on ‘For the good aimed at can; it moves by being loved’ (Ross, 1924, 373).
G od is not a superior who issues commands. . . .’ T h is sentence, T h is is a close parallel to the present passage: Aristotle is here
and especially the ‘for’ is at first baffling in the context, and provides saying that G od is not a raison d'être whose good is being aimed at,
the principal reason which has made commentators want to change but a good whose attainment is the raison d'être o f wisdom’s
either the reading or the sense o f ‘G o d ’ so as to make Aristotle refer commands.
here to the speculative intellect. Well, then, what is the hou heneka whose good is being aimed at,
T h e relevance o f the sentence becomes clear if we connect it with and what is it to attain G od? Aristotle gives no explicit reply, but
the previous chapter. T h ere we learnt that the cosmic G od was the the answer is clear enough from the concluding passage : that which
supreme arche in the soul, superior to logos, epistëmë., and nous. If, benefits from the commands o f wisdom is that which serves and
without further ado, we applied the rule that everyone should live contemplates G od ; and this service and contemplation is itself the
according to the requirements o f his superior as a slave lives only kind o f attainment of G od which is possible. And what is it that
according to the requirements o f his master, we would most serves and contemplates G od? Either the soul as a whole, or its
naturally conclude that in matters o f the soul we must live in speculative part : which o f the two is intended must depend on what
obedience to the commands o f G od. Bu t this would be, in is meant by ‘service’ here. And this, notoriously, is a very difficult
Aristotle’s view, to make an erroneous move. We must first problem.
distinguish between senses o f archë: fo r G od is not a ruler who issues T h e only other place in which Aristotle speaks o f serving G od, or
178 The Aristotelian Ethics Wisdom in the A ristotelian Ethics 179
the gods, is in Pol. H , 1329*30 ff., where the expression clearly a difficult one. It has long been recognized that the final chapter o f B
means ‘ worship’ in a literal, liturgical sense. Commentators have is very close to the conclusion o f the E E . H aving established that
found it astonishing that in the last chapter o f E E Aristotle should wisdom and learning are the excellences o f the two parts o f the
offer as a criterion for the goodness or badness o f choices their effect rational soul, at 1 14 3 bi8 ff. Aristotle inquires what use each o f these
on liturgical worship. Accordingly, they have preferred to take virtues is ? Learning does not inquire how to make men happy, and
‘serving G o d ’ in the sense o f ‘cultivating the intellect’ , citing as the acquisition o f wisdom is surely no more necessary in order to
parallels N E 10, 1 i7 9 a23, where we are told that the therapeia o f the become happy than the acquisition o f medical science is necessary
nous endears us to the gods, and Plato, Timaeus 90 c 4 where the in order to become healthy. ‘M oreover, it would seem absurd if
expression therapeuonta to theion will bear this sense. B u t if the bare wisdom, being inferior to learning, should be put in authority over
theos in this context cannot mean nous, as we have argued, this it: as seems to be implied by the fact that whatever produces
interpretation is excluded. something is superior to it and issues commands about it’ (1 i4 3 b3 3 -
I f we are to look for Platonic parallels a much more appropriate 5). These aporiai are solved as follows. Learning does not produce
one is provided by the Euthyphro. T h ere a long section (12 *2 4 - happiness as medical science produces health: it is, rather, a
i4 ai) is devoted to a discussion o f he tön theön therapeia in which constituent o f happiness. Wisdom is not separable from virtuous
Socrates argues against Euthyphro that the notion is unintelligible. activity as the possession o f the science o f medicine is separable
T h ere are a number o f themes paralleling the present text: from healthy activity. Finally
therapeia cannot be for the benefit o f the gods, whose condition we
wisdom is not in authority over learning or the better part, any more than
cannot hope to alter for the better; therapeia is what slaves give to the science of medicine is in authority over health ; it does not make use of
masters ; but what slaves do for their masters depends on what their it, but provides for its coming into being; the orders it issues are not issued
m asters’ ergon is: what then is the splendid achievement, the to it but issued for its sake. To say it was in authority over learning would
pankalon ergon, which is to be achieved by the gods with our be like saying that political science was superior to the gods, because it
service? ‘Polla kai kala’ replies Euthyphro to Socrates, who issues commands about all the affairs of the state. (114 5*6 -11)
promptly turns to some inconclusive teasing.
T h e service o f the gods which Euthryphro has in mind includes Here, as in E E , wisdom is something which issues commands
prayer and sacrifice: but it includes also acts o f justice such as (epitattei: a word popular in the E E , but occurring only once in the
Euthyphro’s attempt to punish a m urderer— the endeavour which N E ) . Here, as in the E E , the prescriptive role o f wisdom is likened
gives the whole dialogue its framework. I f Aristotle does have the to that of medicine. H ere, as in the E E , it is a function o f wisdom to
Euthyphro in mind here, then the service o f G od could well include promote the contemplative activity o f the soul: in the E E its
acts o f moral virtue. Th ese are the kalai praxeis o f the kalos kagathos function is to promote the contemplation o f God, in the A E it is to
which are the subject o f the first part o f the same chapter; they promote the activity o f the better part o f the soul, whose objects are
could well be regarded as the polla kai kala which we, under the unchanging, eternal things, the most honourable things by nature,
arche o f G od ( 1248*25 ff.) find our fulfilment in performing and by the divine constituents o f the cosmos, and in general extraordinary,
which we make our contribution to the splendour o f the universe. marvellous, mysterious, and superhuman things ( i i 3 9 b2 3 ;
T h is interpretation o f ‘ therapeia' is highly speculative: but given i i 4 i a20, bi-~7). In the E E there is stressed the relationship between
the excessively laconic style o f the conclusion o f the E E , no wisdom and the hou heneka in the sense o f hou : namely, G od. In the
interpretation can be anything else. It is certainly not alien to A E there is stressed the relationship between wisdom and the hou
Aristotle’s manner in the E E to defend the moral opinions o f the heneka in the sense o f h o(i) : namely, ‘the better part o f the soul’ .
plain man against the paradoxes o f Socrates. B u t in substance the doctrine is the same.
It is with relief one turns from decoding the cryptogram o f E E T h e nous passage in N E 10 is the closest Nicomachean parallel to
V I I I to looking for a parallel in the A E . T h e search is not, o f course, these two passages o f the A E and E E , but it is much further from
The Aristotelian Ethics Wisdom in the A ristotelian Ethics 181
each o f them than either is from the other. Phronësis in N E does not T h e first question is answered at some length in book B itself: right
appear as the handmaid o f learning, but only as the yokemate o f the reasoning in these matters, we are told at 1 i4 4 b23, is phronësis ; and
moral virtues ; the activity o f contemplation is attributed not to an phronësis is the major topic o f the book. But what of the question:
element of the soul, but to a separable nous contrasted with the what is the horos o f right reasoning? T h ere is no clear answer to this
human soul-body compound. In A E C , as in E E V I I I , 2, there is a question in book B : some commentators take this as evidence that
divine element in us: but as in E E V I I I 2, it is not the theoretical the book is in a fragmentary state, others that Aristotle though the
nous : it is appetitive rather than speculative, the origin o f the right question was misconceived, others that we are to look for the answer
sort o f desire for pleasure (compare i i S 3 b32 with 1248 *17-29 ). to book 10 o f the N E . 1 T he third answer seems to me more nearly
(3) Phronësis and M oral Virtue. In N E 10 we are told: ‘Wisdom correct than the other two, as becomes clear if we consider the
is wedded to moral virtue, and moral virtue to wisdom, since the analogy with medicine. What the Aristotelian doctor possesses and
principles o f wisdom are determined by the moral virtues, and the rest of us lack is not so much a set of rules o f thumb for
rightness in morals is determined by wisdom’ ( ii7 8 ai6 - i8 ) . T h is is prescribing, but a knowledge o f the nature o f health itself : he knows
a neat summary o f the teaching o f all three treatises. In both the N E what the mechanism o f the body is, how its humours are balanced
and the E E a reference to wisdom is incorporated into the definition when it is in good working order, what are the functions and
o f moral virtue. In the preliminary general treatment o f the subject interrelations o f its parts. So in ethics, if we are to know the right
in the N E virtue is defined as a state o f character expressed in choice action to do we must know the nature o f a healthy— i.e. virtuous
lying in a mean determined by reasoning in the way that the wise and happy— mind : we must know what the parts o f the soul are,
man would determine it ( n o 6 b36)— a definition developed by what their functions are, and how they are related to each other.
Aristotle from the commonly accepted account o f virtue as action T h is is why the horos can also be called the skopos (1 i3 8 b23) : it is a
‘ in accordance with right reasoning’ (1 i0 3 b32). In the E E virtue is right understanding of the end o f conduct, namely a good life, that
defined as a state o f character which makes people capable o f is necessary for right reasoning in morals, just as it is a right
performing the best actions and puts them in the best disposition understanding o f health, which is the end o f medical practice, that
with regard to the best and greatest good ‘the best and greatest good is necessary for right reasoning in medicine. In the N E it is in book
being that which is in accordance with right reasoning, that is the 10 that we are given the fullest account o f the nature o f the good
m ean’ (i2 2 2 a9 -io ). A fter the treatment o f magnanimity, Aristotle life— the contemplative activity o f nous— and to this extent those
fills this out: ‘ the virtue concerned with each subject matter judges who say that it is to N E 10 that we must look for the horos are
correctly what is the greater good and what is the less, just as the correct. But o f course in B itself in the concluding chapters the
wise man and his virtue would command, so that all the virtues go relationship o f phronësis to sophia is discussed and the relationship
with that virtue, or that virtue goes with all virtues’ ( i2 3 2 a25-38). o f both to happiness is sketched (114 4 * 3 -6 ; 114 5 * 6 - 11) . T h e
In each Ethics there is a promise to expand, at a later point, the relevance of this to the question of the horos is not however brought
notion o f ‘right reasoning’ (N E 2, n o 3 b3 3 ; E E i2 2 2 b8). T h e out either in N E 10 or in A E B. F o r that we have to turn to E E V I I I .
promise is not fulfilled in the N E , but both A E and E E contain pas In E E V I I I the question of the horos is taken up in language
sages which refer back to it. T o say that the mean is in accordance reminiscent o f the beginning o f B.
with right reasoning, we are told in B (1 i3 8 b5 if.), is true but
uninformative : it is as i f one were to prescribe the medicines ‘that I. T h e first is the view o f Allan (1970, 136) and the second o f Rowe, who writes ‘W ithin the
limits o f A ristotle’s system, no detailed criterion is in fact possible . . . either the question
medical science, and the man who possesses it, would command’
“ what is the standard” means what it says, and the answer is that there is no such thing; or
(1 i3 8 b3 i ) . 1 We must inquire, Aristotle says, what right reasoning else it is sim ply a misleading way o f putting the more general question, how we apply the
is, and what is the standard or criterion which determines it (horos). general rule to the particular case1, in which case the answer is the intuition o f the φ ρ ό ν ιμ ο ς
( 19 7 1, I I 2 - I 3 ) . Both Allan and Rowe are mistaken, I believe, in thinking that Aristotle is
i . The similarity of language between E E III and A E B is striking, a n e p ό φ ρ ό ν ιμ ο ς âv here concerned with the formulation and application o f rules. T h e third view is that o f
κ*λ€νσ€ΐ€ κ α ί ή ά ρ € τη ( ΐ 232β3);οσα ή ια τρ ικ ή κ ζ λ ζ ύ ε ι κα ι ώ ς ό τ α ύ τ η ν ς χ ω ν ( ι ι $ * $ 2 ). D irlm eier (1969, 442) and other commentators.
Wisdom in the Aristotelian Ethics 183
i82 The Aristotelian Ethics
reason for regarding this statement as spurious ; but even if one does
T h e doctor has a standard by reference to which he distinguishes a healthy
from an unhealthy body and the degree to which each activity is advisable so, the discussion in E E V I I I does give the general standard for the
and wholesome, and beyond which in either direction it ceases to be so. exercise o f virtue. T h e ‘things that are good by nature but not
Sim ilarly in regard to actions and choices o f things which by nature are praiseworthy’ are health, strength, honour, birth, wealth, and
good but not praiseworthy, the good man should have a standard o f pow er(i248 b24, b2 8 -3 o ; i249ai o - n ) . T h ese are the subject matter
possession, choice, and avoidance concerning abundance and scarcity o f o f the virtues of magnanimity, magnificence, and liberality (E E III,
wealth and other gifts o f fortune. E arlier we said “ in accordance with I 2 3 i b28 ff.). T h e subject matter o f the virtues o f courage,
reasoning” — but this, like saying in m atters o f diet “ in accordance with temperance, and meekness, on the other hand, are the passions of
m edicine and m edical reasoning” , is true but uninform ative. ( i24Qa2 i - 6 ) 1 the irrational soul. F o r these too a standard is sketched in the last
Aristotle then goes on to give, as an answer to the question about chapter o f the E E : ‘ T h us it is, too, with the soul, and this is the best
the horos, the rule that everything should conduct its life with standard for the soul, to be as little as possible conscious o f the
reference to its superior, which we considered in the previous irrational part o f the soul qua irrational’ (i249 b2 2 - 5 ) .1 T h us the last
section. T h e resemblances to B in the reference to the doctor, and in chapter of the E E does give a standard for the exercise o f the six
the use o f the phrase ‘true but uninformative’ strike all com virtues which are discussed in the official treatment o f the virtues in
mentators. T h e resemblance may seem too striking: if B belonged E E I I I and which together constitute the perfect virtue or
originally in the same context as E E V I I I , would not this passage be kalokagathia o f the first part o f the final chapter. Other praise
a useless reduplication o f the earlier one? Bu t setting aside the fact worthy states, like dignity and candour and wit, are no doubt
that the undoubted E E and the A E itself contain a number o f difficult to bring within the scope o f the criterion offerred by E E
apparent doublets, a closer look at the text shows that it is not mere V I I I : but then according to the E E these states are not virtues
repetition o f the B passage. Aristotle does not say here, as he does in (i2 3 4 a25) and since they do not involve prohairesis (i2 3 4 a26) they
B, that we must inquire what is the orthos logos: the answer he goes do not call for a horos to guide choice.
on to give is an answer to the question what is the horos— the horos As has been said, it is ground common to all three treatises that
of kalokagathia or perfect virtue, as is said in the concluding wisdom and moral virtue are inseparably linked: but how and why
summary (i2 4 9 b24). T h u s E E V III can perfectly well follow A E B this is so is spelt out at much greater length in the E E and the A E
as part o f the same treatise: a twofold question is raised at the than it is in the brief quoted passage which is the N E ’s only
beginning o f B , and the first part o f the question is there answered; contribution to the topic. T h e A E has a long argument to the
the second half o f the question is raised in V I I I , in terms conclusion that ‘ it is impossible to be good in the strict sense
reminiscent o f the twofold question o f B , and the second h alf o f the without wisdom, or wise without moral virtue’ ( ii4 4 b30 -2). It is
question is then answered. indeed possible to do good actions— the actions that a good man
Som e have disputed that the final passage o f E E V I I I is an would do— without possessing either wisdom or moral virtue; but
answer to the question raised in B. Row e writes: ‘E E V I I I , 3 actions only express virtue if they are done out o f prohairesis and for
severely restricts the application o f its suggested horos: it is their own sake (i i4 4 ai3~2o). Even good actions which are not the
introduced as a standard for “ actions in connection with and expression o f virtue do o f course call for some skill in planning: but
choices o f things naturally good but not praiseworthy” . . . there is this does not mean that there can be wisdom without virtue, for the
no justification in the text for supposing that E E is giving a standard planning is the expression not o f wisdom, but merely of
for actions in general’ (op. cit. n o ). L ik e some other scholars, Rowe intelligence— intelligence which could be used for evil ends as well
must be treating as an editorial addition the final statement that it as good. Intelligence is a necessary, but not a sufficient, condition of
has just been said what is the standard o f perfect virtue. There is no wisdom: intelligence only becomes wisdom if accompanied by a
o f the disputed books : bodily pleasures are necessary to prevent the After this prologue ( i 2 i 7 ai8) Aristotle takes his start from the
obstruction o f virtuous activity ( i i 5 3 bi8) and the pursuit of the generally accepted, but as yet unclarified, statement that happiness
right amount o f them is itself an exercise o f virtue ( 1 1 54a1 5 —18 ; cf. is the greatest and best o f human goods ( i 2 i 7 a2 i) . ‘H um an’ is first
i 2 3 i ai7 in I I I .2). clarified : we are concerned with goods what are human both in the
T h e inquiry immediately to be undertaken is that into virtue and sense o f being good fo r men (vs. goods for gods or animals) and
wisdom. (I speak o f ‘the inquiry’ ; they cannot be regarded as two attainable by men. It is of good which are human in both these
separate inquiries entirely, because the division between virtue and senses that happiness is the best ( i 2 i7 a22, a4o).
wisdom presupposed in the traditional trichotomy is something For the rest o f E E I Aristotle devotes him self to elucidating the
which Aristotle is going to correct, substituting for it the distinction word ‘best’ in the popular account o f happiness. People say that the
between ethical and intellectual virtues.) T h e word ‘ wisdom’, used best o f all things is auto to agathon, the absolute good, whose
in the traditional distinction to denote the goal of a philosophical goodness is not dependent on that o f anything else (so that it can be
career, will be reserved by Aristotle for the practical intellectual called ‘the first good’) and which is the cause o f goodness in every
virtue which is the foundation o f the ‘ political life’ but which other good thing. Aristotle is prepared to agree that the best o f all
also subserves the theoretical contemplation characteristic of the things is the absolute good, but there are three rival opinions as to
philosopher. T h e way in which Aristotle concludes his inquiry into what the absolute good itself is. First, there is the Platonic opinion
virtue and wisdom in the E E is something which cannot be that it is the Idea o f Good ( i 2 i7 b6 -i6 ), against which Aristotle
described without settling the question whether the E E contains argues at length ( i 2 i 7 bi 7 - i 2 i 8 a38). Secondly, there is the view
the disputed books o f the A E . I f the disputed books do belong here, that the absolute good is a common, univocal universal goodness :
then we can say that from 1.7 until the end o f disputed book B is this is briskly disposed of with the remark that no practical
devoted to answering the question raised in I2 i6 a3 8 - b2 ; but we discipline— not to mention the highest practical discipline,
cannot say without qualification that books I I to IV o f the inclusive ethics— is concerned to bring about the presence o f so thin and
Eudemian Ethics deal with virtue, and book V deals with wisdom, ubiquitous an attribute in its subject matter ( i2 i8 bi-6 ). Aristotle
since that is to state Aristotle’s definitive treatment in terms o f a concludes by presenting the third possible opinion as his own view.
provisional one which is merely his traditional starting point. It is clear, then, that neither the Idea of Good nor the universal, goodness, is
T h at Aristotle was at this point well aware o f the possibility o f the absolute good that we are looking for: for the one is unchangeable and
structuring the relationship between virtue and wisdom in various unattainable; the other, while it is changeable, is not an object of
ways is immediately apparent : for he goes on to talk o f Socrates, attainment. But the wherefore, in the sense of end (to hou heneka has telos)
who like him self disagreed with the dichotomy between virtue and is the best : it is the cause of what leads up to it and the first of all. It is this,
wisdom, but who unlike him self thought that the division contained then, that is the absolute good : the end of everything achievable by man (to
not too few but too many members, since, on the Socratic view, telos tön anthropö(i) praktôn) (i2 i8 b7~i3)
virtue and wisdom were identical.
Socrates is introduced principally, however, to usher in the ‘B est’, then, in the popular definition o f happiness, is clarified with
methodological considerations which end the prologue o f the E E . reference to the notion of purpose. When we ask for the why and
T h e present inquiry, unlike theoretical sciences, has an essentially wherefore o f any human action, we can be told that it is to be done
practical purpose : our aim is not merely to know what justice is, but for the sake o f something else; we can ask in turn for the why and
to be just. On the other hand, it is a genuinely philosophical inquiry, wherefore o f that something else; when we reach a point where
concerned to discover not just what is to be done but why it is to be there is no further answer to the question, we have reached the
done, operating by means o f argument, but with due respect for ‘ wherefore in the sense o f end’ . It is the worthwhileness o f the end
received opinions which contain a deal o f truth however obscurely o f an action which makes the actions leading to it themselves worth
and confusedly expressed ( i2 i6 b26 -2 7 ai8). while : in this sense an end is the cause of the goodness o f the means
The Aristotelian Ethics Happiness in the Aristotelian Ethics 199
to it. T h e best o f all human goods, then, would be a good which the ergon is an energeia or activity (in the way that the ergon o f sight
appeared at the top end o f every chain o f practical reasoning: that is seeing) rather than a product o f activity (such as the health
would be an absolute good, an independent good on which the produced by healing or the houses that are the output o f
goodness o f every other human good depended, as the goodness o f architectural skill).
health-producing drugs or regimes depends on the goodness of Applying these considerations about ergon to the activities o f the
health itself ( i 2 i8 bi7 ~ 2 i) . T h is supreme good is the subject matter soul, we can draw the interim conclusion that if the activity o f the
o f the supreme science of ethics and the virtue o f wisdom. best state o f the soul (i.e. aretë) is an activity which is also an ergon,
A t this point the first book o f the E E ends. Aristotle has reached rather than the production o f an ergon, then it will be the best thing
this stage by reflection on the dictum that happiness is the best o f in the soul that we are looking for. F o r it will be better than any
human goods. N ext, at the beginning o f book II, he throws into the other activity (by the theorem that an activity takes its rank from the
discussion two further premisses: first, what is common to the state producing it); and it will be better than the best state which
supporters o f the three lives, that happiness is either wisdom or produces it (by the theorem that an ergon is better than the state
virtue or pleasure; and secondly an inductively established whose output it is); and that state was ex hypothesi the best state;
definition o f arete— ‘virtue’ in its broadest sense o f ‘excellence’ — as and everything in the soul is either a state or an activity ( i 2 i 9 a9 - i 8 ) .
‘the best state or condition or power o f anything that has a use or a All that remains, then, to inquire is : what is the activity o f aretë,
function (ergon) ( i 2 i 8 b3 5 ; I 2 i 9 a i ) . W ith the help o f these further and is it its ergon or not? T o discover this, let us ask what is the ergon
assumptions ( i 2 i8 b39) he then proceeds to a close-knit argument o f the soul’s aretë. T h e ergon o f a good X , or o f the goodness o f an A",
designed to elucidate further the nature o f happiness as the best o f is a good ergon-o f-a n -X : thus the ergon o f a shoemaker is a shoe, and
human goods ( i 2 i8 b3 2 - i9 a38). the ergon o f excellence in shoemaking, or o f a good shoemaker, is a
On all three o f the traditional views, the best goods are in the good shoe. Applying this to the case in point, the ergon o f soul is
soul: so that best o f all things that we are looking for will be life,1 so the ergon o f the aretë or excellence o f a soul is a good life
something in the soul. N ow the things in the soul fall into two ( i 2 i 9 a 2 6 - 8 ) . B u t the life that is meant here is active, waking life,
classes: some are states (or powers), others are activities (or and this is identical with the activity o f the soul. So the best thing in
processes). So the best thing in the soul will be either the best state the soul which we are looking for as the essence o f happiness turns
or the best activity. Consider first the best state. I f a thing has an out to be the activity o f aretë or the activity o f a good soul
ergon (as the soul does) then its best state is aretë. So far, then, we ( i 2 i 8 a 3 5 ) . 2 Bu t aretë can be either complete or incomplete,
can conclude that the best thing in the soul is either aretë or the best according to whether we mean aretë as a whole, or a particular
activity. So now consider the best activity (b3 2 - a6, “2 9 -3 1). We partial aretë; and life, whose active stage is the energeia o f the soul,
have to approach this through a consideration o f the value o f an likewise can be complete or incomplete (in the sense o f running its
ergon. T h e best ergon must be the ergon o f the best state: for in
general an ergon takes the rank o f the state whose output it is. But 1 . € σ τ ω ψ υ χ ή ς e p y o v τ ο ζ η ν r r o i e î v . It makes nonsense o f the argument to translate this,
as Solomon, Rackham, and D irlm eier do, ‘the work o f the soul is to produce living’. Ισ τ ω
which o f these two, the ergon or the state producing it, is better? 7Γo € Î v is to be construed on the model o f e a n v i S c i V , with t t o l c i v to be taken in the third o f
Clearly the ergon: for an ergon is an end: and we saw in the final the senses given in Bonitz s.v., namely ‘ to posit1 : ‘L e t us take leave, then, to posit that the
stage o f E E I that an end is the cause o f goodness in the things e r g o n o f soul is life.* T h ere is no need, once this has been understood, to emend the following
manuscript reading τ ο υ B k χ ρ ή σ ι ς κ α ι έ γ ρ η γ ο ρ σ ι ς : it means ‘ . . . is life, and o f life the
whose end it is. (It is a general Aristotelian principle that if X causes active and waking part’ .
Y to be P , X must be P -e r than Y : e.g. only if X is hotter than Y 2. T h e structure o f the argument is simpler and clearer in the recapitulation (a29~34) : the
will X heat Y .) 1 This principle about the relationship o f ergon and best goods are in the soul ; everything in the soul is έ ξ ι ς or c v e p y c i a ; so the best is either the
best £ ξ ι ς or the best è v é p y e t a ; the best I f ι ς as ά ρ € τ ή \ the best i v é p y c t a is the ev ep y eta o f
hexis applies to all kinds o f ergon, and in particular to the case where the best e fts , a n d is b e t t e r t h a n th e b e st e f t? : so happiness, the best good, is the i v é p y c ta o f
i. L ater incapsulated in the scholastic dictum p r o p t e r q u o d u n u m q u o d q u e t a le , et illu d the α ρ ε τ ή o f the soul, i.e. the i v i p y e i a o f a good soul. In the full-fledged deployment the
m a g is . argument is complicated by the introduction o f the notion o f e p y o v .
200 The Aristotelian Ethics Happiness in the Aristotelian Ethics 201
proper course or not). So happiness— which is something perfect makes life worth while and lacking in nothing. Eudaimonia has both
and complete— must finally be defined as ‘activity o f complete life these properties (i0 9 7ai 5 h2 i). What, then is eudaimonia? T o
in accordance with complete virtue’ ( i 2 i9 a39). His definition thus elucidate this, we must consider the ergon o f man. M an must have
completed, Aristotle goes on to confirm it by showing that it is in an ergon, because particular types o f men (e.g. sculptors) do, and
accord with universally received opinions ( i 2 ig a3 9 b26). parts and organs o f human beings do. What is it? N ot life, not at
Though the argument is always compressed, and sometimes least the life o f growth and nourishment, for this is shared by plants,
sketchy, and though the style lapses at one point into mere jotting nor the life o f the senses, for this is shared by animals. It must be a
( i 2 i8 a36), the discussion o f happiness in the E E is clearly very life o f reason concerned with action: the activity o f soul in
carefully thought out and tightly structured. It is also in perfect accordance with reason. So the good o f man will be his good ergon :
accord with the philosophical method sketched in the prologue.1 the activity o f soul in accordance with virtue, and if there are several
When we turn from the first part o f the E E to book i o f the N E virtues, in accordance with the best and most perfect (teleiotatën)
we find a different and far less organized structure. T h e N E starts virtue (i09 7b22-98 ai8).
not from the notion o f eudaimonia but from that of goodness. The Though on a superficial reading the accounts o f happiness in the
subject matter o f ethics (politikë) is the good for man, the end of first books o f the N E and E E appear rather similar, closer
action for the sake o f which all else is desired (an ultimate end whose investigation reveals substantial differences in structure, in argu
existence is presented in a passage o f notoriously suspect validity, ment, and in doctrine. E ven when the same technical terms
i094ai8 -2 2 ).2 M ost people would agree that this supreme good is appear— such as ergon, teleion, auto to agathon— they are put to very
happiness (however much they must disagree about what precisely different use. And the concluding definitions o f happiness
happiness consists in) (i095a 18 -2 6 ); but some people, the enunciate radically different theories o f its nature.
Platonists, argue that the supreme good is an absolute good which is T h e structure o f the N E treatment o f happiness is much less neat
the cause of all other thing’s goodness. Aristotle then goes on to than that o f the E E . T h e flow o f the argument is frequently broken
expand the general view that the supreme good is happiness by by scattered and repetitive methodological remarks (i094bn -
outlining the three traditional lives (lo g s'314-96*10), and to 95ai4 ; i095a3 0 -bi4 ; i098a2 6 -b9). T h e connection between the
criticize the Platonic view by providing a series o f arguments to discussion o f the three lives and the refutation o f the Idea o f the
show the incoherence o f the Idea o f the Good (i096ai2 - 9 7 a4). Good is not brought out. T h e opinions o f others on the topic,
Aristotle begins his own account by saying that the good we are instead o f being worked into the argument as they are in the E E , are
looking for must be teleion by comparison with other ends— that is, presented in an enormous album o f frequently disjointed remarks
it must be something sought always for its own sake and never for in chapters 8 - 12 . T h ere are a number of inconclusive passages
the sake o f anything else; and it must be autarkes or self- where the argument comes back to its starting-point and a fresh
sufficient— that is, it must be something which taken on its own beginning has to be made : sometimes this is even remarked upon in
the text (‘T h e argument has come around to the point where we
1. In recent years Allan (19 6 1, 3 0 8 - 11) and Row e ( 19 7 1, 15 -2 6 ) have done much to were before’, 1097*23).
elucidate the structure o f the E E discussion o f happiness. Though I disagree with Allan on
Rowe, who has detailed the incongruities o f the N E text,1
one or two points and with R ow e on many, my interpretation o f the passage owes much to
theirs. considers that they provide evidence that the E E represents an
2 . In my paper ‘H appiness’ (1966) I attempted to defend the N E from the charge o f earlier stage o f Aristotle’s development. H aving written the E E , he
committing a quantifier-shift fallacy at this point. Subsequent discussions with a number o f
suggests, Aristotle changed his mind on a major point; he wrote the
people (in particular M r. J . O. U rm son and M rs. L . Brown) have made me less confident that
the N E can be defended along the lines I suggested. B u t whatever the case with the N E , it is N E as a revision o f the E E , but the revision was only partially
clear that the E E is guiltless on this point. F or i f Aristotle is guilty o f the fallacy he is charged
with, he must think it is a logical truth that, if there is choice at all, there is a single supreme i . H e even somewhat exaggerates them. T here is no need, for instance, to attribute to the
end. B u t E E I, i 2 i 4 b6 ff. admits it as a possibility— though a mark o f folly— not to organize N E the suggestion that ‘the Idea o f the G ood’ might be an answer to the question ‘ what is
one’s life in view o f an end. happiness?’ (Rowe, 19 7 1, 27). In this section Row e acknowledges his debt to K ap p (19 12 ).
202 The Aristotelian Ethics Happiness in the Aristotelian Ethics 203
complete, leaving ugly traces of the original pattern. It seems a little The question is often raised whether the E E is more Platonic
curious to treat a more careful and polished treatment o f a topic as than the N E . M ost o f those prepared to give an answer give one in
prima-facie evidence o f an earlier stage o f composition: it would be the affirmative : but scholars in recent years have come more and
at least as plausible to suggest that the E E represents a definitive more to realize that the question has too many facets to be given a
ordering of material which had been only partially moulded into simple yes-no rep ly.1 In the present context of the discussions o f
shape at the time o f writing o f the N E . T h e final scrapbook o f happiness we can say that the N E resembles Platonic ethical theory
received opinions, R ow e thinks, provides evidence against those more than the E E does in making the, good, rather than happiness,
who regard the E E as post-Aristotelian. ‘ I f E E is spurious, and was the primary ethical concept; on the other hand the ergon argument
constructed on the model o f E N , then its Peripatetic author shows a in the E E is closer than the N E one is to the parallel section in the
remarkable ingenuity; he has evidently abstracted certain parts o f Republic (352 e-354 a) in appealing to the notion o f the ergon o f the
the list presented by E N , worked them into the structure of I, i -7, soul rather than to the philosophically more suspect notion o f the
and still managed to construct a highly coordinated series o f ergon o f man. T h e criticism o f the Idea o f the Good is more
argum ents.’ T h e rewriting would certainly call for ingenuity. But a thoroughgoing, and conducted from a less sympathetic standpoint,
suggestion that the E E I is later than N E need not be a suggestion in \ h e E E than in the N E : in the words o fD . J . Allan ‘ its manner o f
that E E I is non-Aristotelian ; and it would surely be rash to assume introducing the discussion, placed side by side with the
that the ingenuity involved in turning N E 1 into E E I would be Nicomachean equivalent, can only seem cool and abrupt’ .2 T h e
beyond the powers o f an Aristotle. discussion in the E E , as G . Verbeke has pointed out, is more closely
T h e principal difference in structure between the E E and the N E linked than that in the N E to general Aristotelian theories o f
is that whereas the E E begins with the concept of happiness, and causation, potentiality, and actuality.3
goes on to discuss the nature o f goodness in the course o f T h e most important difference between the accounts o f
commenting on the terms o f a traditional definition o f happiness, happiness in the N E 1 and in the E E I - I I is that the former
the N E begins with the supreme good as the subject matter o f identifies happiness with a single dominant end, the activity o f the
Ethics, and goes on to consider happiness as the most popular highest virtue; whereas the latter views happiness as an inclusive
answer to the question ‘what is the supreme good?’ B u t the end, the activity o f all the virtues o f the rational soul in the broadest
differences between the accounts, as has been said, go deeper than sense. T h e N E says that the good for man is ‘activity o f soul in
mere presentation. In the E E , unlike the N E , Aristotle is prepared accordance with virtue, and if there is more than one virtue, in
to accept the notion o f an absolute good (auto to agathon)} In both accordance with the best and most perfect (teleia/ ( iog8ai6). In
ethics in the construction o f the definition o f happiness use is made the last chapter o f the book we are told that there is indeed more
o f the concept o f ergon·, but the ergon in question in the E E is the than one virtue— there are, for instance, moral and intellectual
ergon o f the soul, in the N E the ergon o f man. T he E E lays much virtues— and hence we conclude, unconditionally, that happiness is
more stress than the N E on the causal role o f the supreme good : the activity in accordance with the best and most perfect o f these many
final cause that is the first cause in the order o f goodness-causation, virtues. (Compare ι ι ο 2 35~7 n o 3 a2 - io .) In the E E we are told
the cause o f the goodness o f other goods. that happiness is activity ‘o f complete life in accordance with
i . D . J . A llan comments well on the significance o f this. ‘[In the N E\ he treats as complete (teleia) virtue’ ; and the word ‘complete’ has just been
equivalent the assertion o f a Platonic Idea and the use o f the expression “ absolute” . T he unambiguously glossed when Aristotle has said that ‘life is either
search for the absolute good is consequently a wild goose chase. T h e author o f the E E joins complete or incomplete, and so also virtue— one being whole
with enthusiasm in the search, and is in full agreement about the description o f the bird that
is being hunted. It would seem to me that the key to this difference in the style o f criticism is virtue, another a part ( i2 ig a35-9). So that when in the E E Aristotle
not that the Eudem ian version shows closer sym pathy with Platonism . . . but that the 1. T his has been so especially since G . E . L . O wen’s classic lecture ‘T h e Platonism o f
passage is integral to the discourse on method which immediately precedes it. T h e Platonist Aristotle’ (1965).
is to be disabused o f his errors by reasoning from the premisses which he admits, and which 2. Op. cit. 309.
are true, but not in the way in which he supposes’ (19 6 1, 309). 3. Verbeke, 19 5 1, 83.
204 The Aristotelian Ethics Happiness in the Aristotelian Ethics 205
goes on to distinguish the parts o f the soul and the virtues that and most teleia virtue’ the word ‘teleia' admittedly has several
correspond to them, and to list examples o f moral and intellectual senses, one o f which is ‘complete’ ; 1 but it is hard to take the word
virtues, we know that their activities are all supposed to be part o f for ‘best’ to mean ‘comprehensive o f m any’ rather than ‘better than
happiness. The virtue o f the soul that figures in the definition o f the the rest’ . ‘A r e të ', like the English word ‘ virtue’ , can be used as a
end o f m an— we learn at 1220*4— is a virtue which is constituted by mass-noun (as in ‘a man o f great virtue’ ) or as a count-noun (as in
the several virtues o f the different parts o f the soul. ‘a man of many virtues’ ); but on A ckrill’s view Aristotle is made to
T h e contrast I have just drawn has long been familiar to switch from the mass-noun to the count-noun use and back again to
scholars.1 It has recently been questioned by Professor J . L . Ackrill, the mass-noun use within a space o f ten words.
who argues that it is a mistake to regard Aristotle’s idea of T h e traditional view sees the clause ‘ if there are more than one
eudaimonia in N E 1 as the idea of a ‘dominant’ end. T h e concluding virtue, in accordance with the best and most complete’, as keeping
clause o f the N E definition ‘if there are more than one virtue, in open a place for the eventual doctrine o f N E 10 that happiness is the
accordance with the best and most complete’ should not be activity o f the supreme virtue o f sophia. Even Ackrill does not try to
understood as referring to a particular supreme virtue, one o f many, deny that in N E 10 a ‘dominant’ view o f happiness is adopted. No
but rather to total and comprehensive virtue, the sum o f all virtues doubt we are wise not to take it for granted that N E i and N E 10
(Ackrill, 1974, 15 - 18 ). were written in a single stint : but there is evidence that when N E i
Ingenious though it is, this interpretation does not seem to me was written Aristotle was thinking o f the topic o f N E 10 (1096*5)
tenable. In the previous chapter of the N E Aristotle has already and that when he wrote N E 10 he had in mind N E 1 ( 117 7 * 11 ) . ‘ I f
made clear that he does not regard happiness as an inclusive end, happiness is activity in accordance with virtue, it is reasonable that
but as a dominant one, in the passage in which he says that when we it should be activity in accordance with the most excellent virtue :
are looking for happiness we are looking for something which, when and this will be the one o f the best part o f us.’ T h is is either nous or
not added to anything else, is most choiceworthy— clearly, if it is so something like it, so ‘ the activity o f this in accordance with its
added it is more choiceworthy with even the least additional good. proper virtue will be perfect (teleia) happiness’ . I f teleia here means
I f happiness were meant as an inclusive end, as the sum total of ‘final’ rather than complete, that suggests that it meant the same in
goods sought for their own sake, it would be absurd to speak o f the N E i passage to which reference is m ade; if it means ‘ complete’
goods additional to happiness.2 When Aristotle speaks o f ‘ the best then again it implies that there is nothing else in happiness other
than the contemplative activity o f nous. Aristotle then goes on to
1. e.g. Verbeke, 19 5 1, 84: ‘L e point de vue adopté dans Y Ethique Eudémienne est
indéniablement plus large que la position prise dans Y Éthique à Nicomaque: dans ce dernier show that theoretic contemplation possesses all the qualities which,
ouvrage Aristote cherche l’activité propre de l’ homme, celle qui le distingue de tous les autres according to Book 1, were, in popular opinion and in truth,
vivants, et pour y arriver, il procède par élimination progressive pour ne retenir que l’activité
intellectuelle ou la contemplation; dans YEthique Eudémienne le Stagirite affirme simplement
properties o f happiness. T h u s: it is the highest activity, most
la supériorité des biens psychiques sur les biens extérieurs et il en conclut qua la perfection de continuous and durable (117 7 * 2 1 ; cf. n o o bi 1 - 1 7 ) ; the pleasantest
l ’homme se trouve dans l’activité de l’âme : mais il ne dit pas que cette activité serait purement (*23 ff.; cf. 1099*7 ff·); most self-sufficient (*27; cf. io97b6 ff.); it is
intellectuelle ou contemplative: on ne voit pas qu’ il exclut d ’autres activités psychiques.’
loved for its own sake ( ii 7 7 bi ff.) and therefore teleion in the sense
2. T h is interpretation o f i097b was adopted in my paper ‘ Happiness’ (1966). It is
criticized by Ackrill, p. 12, who writes ‘ It is indeed only if one is willing, with K en ny, to treat explained at 1097*30 if. I f Aristotle underwent a spectacular change
“ happiness” as a fair translation o f the word eudaimonia that one can feel the slightest in his view o f happiness between Book 1 and Book 10, as Ackrill’s
temptation to take the self-sufficiency passage as he does.’ I agree that ‘happiness’ is an
im perfect translation o f eudaimonia, which means something more like ‘ worthwhile way o f
view seems to imply, he wrote book 10 in such a way as to cover up
life’ ; I use it in preference to such a periphrasis on grounds o f tradition and brevity. B u t my the change entirely. But even if Ackrill is right about N E 1, the
interpretation o f this passage is a traditional one, which was adopted by Eustratius, Albertus Λ7ιό ν τ ο ς α υτο ύ. T h e traditional interpretation criticized by Ackrill is ably vindicated by
M agnus, and Thom as Aquinas, who were comparatively free from the temptations presented
Clark (1975, 15 3-4 ).
by the distracting overtones o f the English word. T hus Eustratius (C A G xx.65) γ ίν ε τ α ι γ α ρ
I . In its immediately succeeding use, the word means ‘ complete’ (i09 8 ai8 ); the use
to TTρ ο σ τ ιθ ε μ ε ν ο ν υ π έ ρ ο χ η τ ο ύ α γ α θ ό ν κ α ί α ύ ζ η σ ις , ώ ς μ ε ίζ ο ν α ε κ τ ή ς π ρ ο σ θ ή κ η ς
previously explained was ‘always chosen for itself and never for the sake o f something else’
τ η ν ε υ δ α ιμ ο ν ία ν γ ιν ο μ ε ν η ν α ίρ ε τ α )τ ε ρ α ν ε ΐν α ι α υ τ ή ν μ e r ’ α υτού, η α ιρ ε τ ή ε σ τ ιν κ α ι
(ί097“3°).
206 The Aristotelian Ethics Happiness in the A ristotelian Ethics 207
contrast between the E E and the N E remains: for at no point does only if A E C , or something like it, preceded in the original
the E E adopt the view o f happiness as identical with theöria which Eudemian Ethics. T h e next points are both made in the book on
Ackrill agrees is presented in N E 10. Friendship: Things which are pleasant simpliciter are fine simpli
T h e complete virtue, whose exercise in a complete life citer because that most rightly deserves t0 be called pleasant which
constitutes eudaimonia according to E E I, is the sum total o f the the wise man calls pleasant; and to him ;t is good and fine things
virtues discussed in the central books o f the E E , and is treated again that are pleasant (i2 3 s b 3 6 -3 6 a7). T h in gs which are good simpliciter
as a unified whole in the first part o f the final chapter o f E E V I I I , are also pleasant, because it is naturai goods which are good
under the description kalokagathia (i2 4 8 b8 -i6 ). Clearly, anyone simpliciter and these are naturally pleasant: this natural pleasure is
who deserves the description kalos kagathos must have all the nature’s road to virtue (i2 3 7 a5-9).
individual virtues, just as a body can only be healthy if all, or at least Thus for the ideally virtuous man the concepts ‘ agathon’, ‘hëdu\
the main, parts o f it are healthy. Aristotle then develops the theme ‘ kalon’ coincide in their application. I f what is pieasant for a man
that natural goods, like health, and wealth are beneficial only to a differs from what is good for him, then he is not yet perfectly good
good man : to a bad man they may be positively harmful. But among but incontinent ( i 2 37a8 -9 ); if what js good for him does not
good men we must distinguish between those o f a utilitarian cast, coincide with what is fine for him, then he is not yet kalos kagathos
who pursue virtue for the sake of non-moral goods, and the kaloi but only agathos: for the kalos kagathos the natural goods o f health
kagathoi who pursue the natural goods only for the sake o f the and wealth and power are not only beneficial but fine, since they
virtuous actions for which they are useful. T h e two marks o f the subserve his virtuous activity (12 4 9 ^ ). So for him ^ goodness,
kalos kagathos— that he has all the virtues, and that he pursues the fineness, and pleasantness coincide. 'J'he bringing about o f this
virtues for their own sake and other things for the sake o f them— are coincidence is the task o f ethics (1247*3). But whereas something
not really distinct : already in the case o f the particular virtues we can be kalon or agathon whether it is a hexis or an energeia ( 1 248b35~
were told that genuine virtue involves the choice o f to kalon 7, b23-4 ) it is only an energeia or praxis that can be pleasant. So it is
(i2 2 9 a2 if. ; i2 3 0 a29~32). T h e new character introduced here is not in the fine activities o f the good man that the highest pleasure is to
the kalos kagathos, but the mere agathos, the ‘Laconian’ character be found and that pleasure, goodness, and fineness meet. But the
who pursues virtue as a means to an end : he— like the continent and fine activities o f the good man are the activities o f perfect virtue
incontinent persons discussed in A E C and from time to time in the with which happiness was identified in book one. So the treatise has
E E — occupies a position intermediate between the virtuous and reached completion: Aristotle has carried out the promise o f his
vicious characters who are presented in an exaggeratedly well- first paragraph to show that happiness combined the three
defined trichotomy by the doctrine o f the mean. T h is last superlatives— finest, best, and pleasantest— o f the Deliac in
distinction between the agathos and the kalos kagathos sets in final scription ( i 2 i4 a7-8).
clarity the ideal o f the happy life as one o f complete virtue T h e quotation just given is not, ho\vever, the last paragraph o f
(i2 4 9 ai8). the E E . It is followed by the passage already discussed concerning
Though we are told that kalokagathia is complete virtue, we are the horos o f kalokagathia. F o r at 1249*2 j Aristotle has yet to finish
not at this point explicitly reminded that happiness is the exercise o f showing the truth contained in the true but unclarified premiss
complete virtue. Bu t Aristotle goes on immediately : that happiness belongs to one of the three traditional lines.1 T h e
discussion o f virtue and o f pleasure and o f kalokagathia has shown
Pleasure has already been discussed: what kind of thing it is, and in what
sense it is a good; and how things which are pleasant simpliciter are fine that what the seekers o f pleasure and the pursuers o f virtue seek is to
simpliciter, and things which are good simpliciter are also pleasant. But be found in Aristotelian ideal. But what 0f the Anaxagorean life o f
there cannot be pleasure except in action: and so the truly happy man will I. I think Jaege r was well inspired to regard A ris to tle ideal as a synthesis o f the three lives
(235-8). Joh n Cooper says ‘ T h e study ot the three lives no mor£ provides the structure o f the
also have the most plesant life. This human quest is not an idle one. Eudemian Ethics than it does that o f the Nicomachean ; Jaeger was at straws here with
(i249ai8 -2 i) which to prop up his theory o f Aristotle’s d e v e lo p m ^ - (Cooperj , 45) T h is seems tQ
T h e first o f the backward references in this passage is intelligible me imperceptive.
2o 8 The Aristotelian Ethics Happiness in the Aristotelian Ethics 209
philosophic contemplation? Are the values o f that life too happiness consisting o f the exercise o f sophia, and an alternative,
incorporated in the inclusive programme o f eudaimonia ? second-class career consisting in the exercise o f phronësis and the
L et us continue to leave aside, for the moment, the A E , and try to moral virtu es.1 N ow which o f these conceptions o f happiness
answer the question from the E E context alone. I f kalokagathia is a matches better the remarks about happiness in the disputed books ?
synthesis o f the virtues o f parts o f the soul in the way that health is a In A E A we are told that ‘we call those acts just that tend to
synthesis o f the health o f various parts o f the body, then it must produce and preserve happiness and its parts for the political
include the virtues o f the intellectual parts o f the soul as well as o f community’ ( i i2 9 bi9). T h is seems to presuppose the E E view of
the passional part. Wisdom is alluded to as a necessary part o f happiness as a whole with parts, rather than the N E view o f
kalokagathia at i2 4 8 b3 i : but is the healthy functioning o f the happiness as a single dominant activity— and one which at its best
speculative part o f the soul part o f the complete virtue which is can be pursued outside a community ( i i 7 7 a33). In B we are told
happiness? T h e final section about the criterion o f kalokagathia is that wisdom studies ‘the things which will make a man happy’ : the
there to tell us not only that is it part o f happiness itself but that it plural perhaps fits the E E conception better than the N E one, but it
sets the standard to which the activities o f the other virtues must is hard to be certain whether the G reek expression (ex hôn estai
conform if they are to remain within the realm o f virtue and euiaimôn anthröpos) refers to means towards happiness or actual
happiness. Not, o f course, that reference to the contemplation and constituents o f happiness (cf. E E I 2 i4 bi 2 f f . and i 2 i 6 b2 i). T h e
service o f G od settles how much o f the other virtues one should crucial passage in B for our concerns is the following answer to the
have ; that would be absurd, since virtue is something o f which objection that learning and wisdom are superfluous because
there cannot be too much. But virtuous action consists in executing unproductive: ‘T h ey are indeed productive: not like medical skill
choices about the right amount o f other things— the passions and in relation to health, but like health itself ; it is thus that learning is
external goods which are the field o f operation o f the moral productive o f happiness : for being a part o f virtue entire by being
virtues— and it is reference to the horos which is necessary to decide possessed and being operative it is productive o f happiness’
what is the right amount o f such things. T h is does not conflict with (114 3 * 3 -6 ).2 T h e passage is difficult enough to deserve five pages of
the doctrine that virtue must be chosen for its own sake, because it is discussion in Gauthier and J o lif s commentary. T h e difficulties are
kalon; the horos provides the measure, not the motive, o f virtuous these. (1) T h e passage begins with a plural verb (‘ they are
action. Brave, temperate, and generous living is an end in itself, productive’ ) and then we are offered a single subject (‘ learning’) :
chosen for its own sake, and as part o f happiness in its own right. how much o f the passage is about both learning and wisdom, and
B u t what particular behaviour in concrete circumstances counts as
1. I agree with C ooper ( 1 9 7 5 ,1 57 - ^°) against G au th ier-Jo lif (19 59 ,8 9 3-6 ) that this is the
virtuous living cannot be settled without consideration o f the only possible interpretation o f i i 7 6 b2 6 -a22, and Aristotle’s doctrine cannot be taken to be
contemplation and service o f God. Aristotle’s favourite analogy that there is a single happy life consisting primarily o f the exercise o f learning and secondarily
o f the exercise o f the other virtues. Cooper shows convincingly that this is not a possible
between spiritual and physical health may help us here. T h e
reading o f the crucial sentences : κ α ι τ ω ά ν θ ρ ώ π ω δ η ό κ α τ ά τ ο ν νο υ ν β ίο ς , ε ϊπ ε ρ τ ο ύ τ ο
healthy man has a good appetite for the right sort and the right μ ά λ ισ τ α ά ν θ ρ ω π ο ς , ο δ τ ο ς αρα κ α ι ε ύ δ α ιμ ο ν ε σ τ α τ ο ς . δ ε υ τ ε ρ ω ς δ ’ ό κ α τ α τ η ν ά λ λ η ν
amount o f food. T h e right sort and the right amount o f food means : α ρ ε τ ή ν . But Cooper’s own view, that the second-rate life o f N E 10 is the same as the mixed
life o f £ £ , seems as untenable as Gauthier’s. T h e only positive argument he offers for it as
the sort and amount o f food that conduces to health. But the healthy
that it would be strange i f the mixed life championed in the E E were not so much as
man, unlike the hypochondriac, does not eat his food because it mentioned in the N E . T h at is so, only if the N E is later than the E E . And for this we need an
conduces to health; he eats it because he has a good appetite and he independent argum ent— which, o f course, C ooper believes he has : the psychology o f N E 10,
enjoys his food. he says, is closer to the de Anim a, which he assumes is late (pp. 175 ff.). D evereux (19 77, 8),
who shares Cooper’s faith in the lateness o f N E 10 , has shown how little support the N E text
T h e contrast, then, between the inclusive, organic view o f gives to Cooper’s interpretation of the two types o f happiness.
happiness in the E E , and the dominant, intellectualist one o f the 2. ε π ε ιτ α κ α ι π ο ιο ΰ σ ι μ ε ν , ο ύ χ ώ ς ή ια τ ρ ικ ή δ έ ύ γ ίε ια ν , άλλ’ ώ ? η ύ γ ίε ια , ο ύ τ ω ς ή
σ ο φ ία ε υ δ α ιμ ο ν ία ν , μ έ ρ ο ς γ ά ρ ο ΰ σ α τ ή ς δ λ η ς α ρ ε τ ή ς τ ω ε χ ε σ θ α ι π ο ι ε ί κ α ι τ ω
N E is clear and profound, Instead o f a single life offering us all the
ε ν ε ρ γ ε ΐν ε υ δ α ιμ ο ν ία ν . L ik e von Fragstein ( ΐ 974 > 2 47 ) 1 prefer to read ε υ δ α ιμ ο ν ία ν with
values sought by the promoters o f the three traditional lives, the Pap. O xyrhync. 2402, L b and Ob and Eustratius; but i f one reads ε ύ δ α ίμ ο ν α with M b
concluding section o f the N E offers us a first-class, perfect (‘ makes a man happy’ ) the sense is much the same.
210 The Aristotelian Ethics Happiness in the A ristotelian Ethics 211
how much about learning alone? (2) W hat is it that health is clearly envisages an inclusive, not a dominant conception of
supposed to be productive of? (3) W hat is meant by saying that happiness.
learning, as part o f virtue entire, is productive o f happiness ? There remains one further passage on happiness in the disputed
I answer the difficulties as follows. (1) It is true o f both wisdom books. It is a long passage in the discussion o f pleasure in A E C,
and learning that they are productive o f happiness in the way that which I shall first quote and divide into numbered sections to
health is productive : but it is true only o f learning, and not also of facilitate further discussion.
wisdom, that it is not productive in the way that medical skill is. For
both learning and wisdom are constituents of happiness; but ( i ) I f certain pleasures are bad, that does not prevent the chief good from
being some pleasure, just as the chief good may be some form of
wisdom contributes to happiness not only in being a constituent of
knowledge though certain kinds o f knowledge are bad. Perhaps it is even
it, but in prescribing for the sake o f another constituent, namely
necessary, if each disposition has unimpeded activities, that, whether the
learn in g(i i4 5 a9). (2) What health produces is neither happiness (as activity (if unimpeded) of all of our dispositions or that of some one of
some commentators have thought in contradiction to E E i 2 i9 b20- them is happiness, this should be the thing most worthy of our choice; and
2) nor health (as others have thought, in spite o f the relation this activity is pleasure. Thus the chief good would be some pleasure,
between health and health being m anifestly one o f tautologous though most pleasures might perhaps be bad without qualification. (2)
equivalence rather than efficient causality) but the activities o f a And for this reason all men think that the happy life is pleasant and weave
healthy body. Health is a hexis which produces, causally, the pleasure into their ideal of happiness—and reasonably too; for no activity
energeiai characteristic o f it. Ross caught the sense exactly in his is perfect when it is impeded, and happiness is a perfect thing; this is why
note on his translation: ‘ I.e., as health, as an inner state, produces the happy man needs the goods of the body and the external goods, i.e.
the activities which we know as constituting health.’ It is in this way those of fortune, viz. in order that he may not be impeded in these ways.
(3) Those who say that the victim on the rack or the man who falls into
that wisdom and learning produce happiness : they are the hexeis
great misfortunes is happy if he is good are, whether they mean to or not,
whose activities constitute happiness, as is spelt out in the final part
talking nonsense. (4) Now because we need fortune as well as other things,
o f the passage. (3) L earnin g is only a part o f virtue entire; it would some people think good fortune the same thing as happiness ; but it is not
be pointless o f Aristotle to remind us o f this here if he intended to that, for even good fortune itself when in excess is an impediment, and
say, as many commentators make him say, that it is the whole of perhaps should then be no longer called good fortune; for its limit is fixed
happiness. It is because it is part o f virtue entire, and because the by reference to happiness. (ii5 3 b7~24, trs. Ross)
activity o f virtue entire (the ‘complete virtue’ o f E E I) is happiness,
that the possession plus exercise o f learning is productive of (1) In the first section o f this passage, it is very striking that
happiness: learning is one o f the hexeis which causally produce Aristotle appears to leave open the question whether virtue is the
energeiai which in their totality are happiness. activity o f all the virtues (for the ‘ dispositions’ he has in mind are
T h u s understood— and this way o f understanding it, I would clearly virtues) or o f some one o f them. T h is is the only place in the
contend, is the natural way o f understanding it even on grounds o f E E or the A E where this is done, and it cannot but remind the
its internal structure alone1 — the passage is in perfect accord with reader o f the definition o f the human good in N E i as ‘ activity o f
the doctrine o f E E I and E E V I I I , but in flagrant contradiction with soul in accordance with virtue, and if there are more than one
that o f N E 10 and more muffled contradiction with that o f N E 1. It virtue, in accordance with the best and most complete’ . I f any
passage o f the A E has the ring o f a Nicomachean interpolation, it is
i . N o one has taken more pains to square this passage with N E 10 than Gauthier and Jo lif, surely this one. Y et the style, in other respects, is not more
and even they adm it that the most obvious way o f taking the passage ‘bouleverse tout l’édifice
de la morale aristotélicienne en faisant entrer la sagesse pratique,— et avec elle la vertu morale
Nicomachean than Eudem ean; and it would be ironic to regard this
qui en est inséparable,— dans la constitution même de la fin suprêm e, qui partout ailleurs[ ! ] passage as evidence o f a Nicomachean origin for the disputed books
consiste dans la seule contemplation’ (1959, 543)· T h eir own interpretation, which on the when it was the very same passage which— because o f its
basis o f comparison with fr. 6 W o f the Protrepticus makes Aristotle say that learning is an
improper part (i.e. the whole) o f happiness, is rightly found unwarranted and fanciful by
identification o f the supreme good with a pleasure— seemed to
Cooper (19 75, 1 1 2 - 1 3 ) . Aspasius (C A G χ ίχ .15 1) to provide evidence that the context was
212 The Aristotelian Ethics Happiness in the Aristotelian Ethics 213
not by Aristotle but by Eudemus. In fact a closer reading o f the happiness and its parts and its relation to virtue and fortune which
passage shows that there is a crucial difference from the is closer in both substance and terminology to that o f the E E than to
Nicomachean parallel. In N E i the disjunctive definition of that o f the N E . T h is impression is confirmed when we consider that
happiness appeared in a conclusion o f Aristotle’s own argument, the Eudemian Ethics itself makes a m uch more coherent whole when
and had as its purpose to keep open an unobvious alternative he was the A E is included in it than if it is omitted. T h e intellectual virtues
later to argue to be the correct one. Here, it appears as an ad are then dealt with at a length proportionate to that devoted to the
hominem argument against those who refuse to identify pleasure moral virtues; the relation of the Aristotelian ideal o f happiness to
with the supreme good— whoever they may be, whether Plato, the Anaxagorean contemplative life is spelt out by the discussion o f
Speusippus, or other Platonists perhaps even including an earlier sophia and its relation to eudaimonia; the promise in E E I to discuss
Aristotle. T h e disjunction can very well be understood as intended the value o f pleasure and its relation to human goodness is kept in
to allow, dato non concesso, a premiss which Aristotle rejects: A E C. I f the A E belongs with the E E then there is consistently
‘W hether happiness is (as argued in the present work) the presented a view which combines an inclusive conception o f
unimpeded activity o f all human virtues or (as our opponents happiness as the exercise of all the virtues with an insistence on the
allege) the unimpeded activity o f one particular human virtue, primacy and supreme value o f philosophical contemplation. On this
either way happiness is unimpeded activity and therefore pleasure.’ view, wisdom and the moral virtues find expression in actions
Certainly this passage allows assumptions to which Aristotle, on his which are both parts o f happiness in themselves and also promote
own account, would never have assented at the time when he wrote the contemplation which is the supreme element in happiness. Acts
either the N E or the E E : as that the majority o f pleasure are bad o f the moral virtues contribute to happiness in the way that good
haplös. breathing contributes to good singing and not in the way in which
(2) T h at the happy life is rightly thought to be pleasant is (say) being early to bed and early to rise may do so. T h ey are things
common ground in the N E ( ιιη η &22) and the E E (i249 a2o). Both which a happy man does qua happy, as the good singer’s good
treatises also allow that happiness m ay be impeded by lack of breathing is an inseparable and essential part o f his good singing.
external goods (N E 1 i7 8 b33 f. ; E E i2 4 g b2o) ; but o f course the way Finally, the summary in E E V I I I .3 o f the whole course o f the
in which lack o f money ‘ impedes’ liberality is more intimate than discussion fits admirably a Eudemian Ethics which contains the A E
that in which it impedes contemplation. T h e present passage would and displays to the careful reader the elegant structure o f the work
fit either type o f prevention. as a whole.1
(3) Th at a virtuous man is happy in the midst o f suffering and T h e -contrast that we have drawn between the N E view of
misfortune is a view which is criticized both in the N E (1096*2) and happiness and the E E view o f happiness, and the comparison we
by implication in the E E ( i 2 i4 bi4) : it was, as Burnet and Gauthier have made between each and the observations on happiness in the
showed, a view dear to Plato and the author of the Protrepticus A E , once more confirms the hypothesis that the disputed books
(G auth ier-Jo lif, 1959, 32). originally belong with the E E . In conclusion, we may note that our
(4) In both ethics the mistaken identification o f happiness with discussion also enables us to throw light on a crux in the E E . In the
good fortune is mentioned ( N E 1, 2099b8; E E 1 2 1 5 * 1 2 ff.); in both discussion (1 i2 0 b3 5 - 2 i ai0) o f the doctrine o f the mean in E E 2
we are warned that excessive good fortune may make virtue and Aristotle presents a list o f triads o f which the final member is
happiness difficult (N E i i 2 4 a30 ff. ; E E i24 9b2o). T h e language of ‘cunning-stupidity-w isdom ’ . As many o f the triads in the list are
the warning in the E E , like the one here, makes use o f the technical
term horos. i. T h is is well brought out by D irlmeier in his excellent note on 12 4 9 *17 : ‘V I I I 3 ist
überhaupt ein K ap . de Rückbeziehungen: am Anfang der Verweis au f E E I I I ; sodann ist die
W e have now considered all the references to happiness in the Existenz von E E Iv ( = N E 5, verteilende Gerechtigkeit) vorausgesetzt, ferner die von E E V
A E . W e can see that, despite the difficulties in interpretation o f two ( = N E 6, Phronësis) une eben auch E E V I ( = Ε Ν η , L u st).’ T his note is the more impressive
o f the most crucial passages, they add up, on balance, to a theory of because Dirlm eier believes that the original E E m iddle books are lost.
214 The Aristotelian Ethics
names o f moral virtues and vices, this passage has seemed to many
to be in such flagrant contradition with A E B that the paradox could
be resolved only by excising the offending triad, or by denying the CH APTER 9
E E to Aristotle, or at the very least by denying B to the E E . N ow in
fact the triad contains many entries which are not moral virtues and
vices (the most obvious being ‘gain -loss-ju st’). Aristotle’s aim is The Dating o f the
not to show that wisdom is a moral virtue, but to show that here too
there can be an application of the mean to the case o f wisdom. Aristotelian Ethical Treatises
Aristotle, at 12 2 1* 3 6 shows how it applies in the case o f the choice of
means. But the contrast between the inclusive and the dominant
notion o f happiness shows us how the doctrine o f the mean applies I t is the almost universal opinion o f Aristotelian scholars that the
also to the choice o f ends. T h e parameter to which the mean is to be E E was composed earlier in Aristotle’s career than the N E . A s to the
applied is given by the question: how many values does a man Magna M oralia, scholars do not agree whether to view it as a
pursue? T h e stupid man pursues too many, he follows one little genuine work o f Aristotle: but those who treat it as authentic are
project after another: he is guilty o f the folly we were warned unanimous in regarding it as the earliest o f the ethical treatises.
against in E E I, i 2 i 4 b9. T h e wise man pursues the right number: Scholars are far less united about the absolute chronology o f
he follows the Aristotelian ideal which combines the values o f the Aristotle’s ethical works than they are about the relative chro
contemplative, political, and hedonic lives. T h e cunning man nology. T h is is because the opinions concerning the order o f the
pursues a single dominant goal and is ruthless about other values. treatises are based rather upon a priori theories o f Aristotle’s
An intemperate man who pursued pleasure, come what may, philosphical development than upon the historical data to be found
would, provided he was intelligent, provide an obvious example o f a in the treatises themselves. T h is is particularly true in the case o f
cunning man (cf. A E i i 4 2 bi8 ; ii4 4 a28). B u t so, if I am right, the Magna M oralia, which is unusually rich in topical allusions
would the man who gave him self to the single hearted and permitting a fairly precise dating: the dating which the allusions
unrelenting pursuit o f philosphy without regard for the moral suggest is quite other than the one which is accepted by the scholars
virtues. A person who organized his life entirely with a view to the who regard it as genuine. But it is also true in the case o f the E E and
promotion o f philosphical speculation would be not wise but the N E , where the rather sparser historical data lend no support to
cunning, not phronimos but panourgos. T h e type o f person whom the view that the E E is earlier in date. T h is is particularly the case if
many regard as the hero o f the Nicomachean Ethics turns out, by the the A E is located in its rightful context in the E E , so that the
standards o f the Eudemian Ethics, to be a vicious and ignoble historical allusions in the A E can be drawn on in order to give a
character.1 terminus post quem for the date o f the E E . T h e present chapter will
examine the evidence for dating the ethical treatises which can be
i. Cooper (19 75 , 164) says ‘Aristotle in the N E conspicuously avoids saying that his
theorizer will be a virtuous person. . . . H e will not possess the social virtues or any other drawn from their texts. It will not attempt to present an overall
virtues, because he will lack the kind o f commitment to this kind o f activity that is an essential theory o f Aristotle’s ethical development, much less o f the progress
characteristic o f the virtuous person.’ D evereux points out on C ooper’s interpretation, there
o f his philosophy as a whole.
is nothing to prevent the contemplative from being quite ruthless in pursuing his goal. F or
example, he may by betraying a friend gain a large sum o f money and thereby assure him self Before examining the textual evidence in detail, let us look briefly
years o f leisure for philosophizing. W hat would hold him back?’ (1978). I am not sure at the chronologies for the ethical treatises which scholars have
that C ooper’s interpretation o f N E 10 is correct, on this point; but i f it is, the contemplative
proposed within the framework o f the known dates o f Aristotle’s
surely deserves the condemnation which the E E , by m y conjecture, would wish to pass on
him. life. We know that he was born in the year 384 ; he came to Athens in
367 and remained there in the Academy until approximately the
time o f Plato’s death in 347. For the next three years or so he was at
2i6 The Aristotelian Ethics The Dating o f the Aristotelian Ethical Treatises 217
Assos under the patronage o f his friend and future in-law, M entor was less vivid : 1 but it tells even more strongly against those
Herm eias the tyrant o f Atarneus. In 343/2 he was summoned to who wish to place the M M in the 360s.
M ieza by Philip o f M acedon to act as tutor to his son Alexander. (2) A t i2 0 3 a23 the writer gives a list o f evildoers on the grand
After Philip’s death and Alexander’s accession Aristotle returned to scale. Along with Phalaris, he names Dionysius and Clearchus. It is
Athens in 335/4, and taught there some thirteen years until just not clear whether the first or the second Dionysius was meant:
before his own death in 3 2 2 .1 either o f them, by their ill treatment o f Plato as well as their other
W. Jaeger, who regarded the M M as inauthentic, assigned the misdeeds, could deserve to figure in such a list. Clearchus is most
E E to the middle forties, during Aristotle’s stay in Asia M inor ; the likely the tyrant o f Heraclea in Pontus, once a pupil o f Plato, whose
N E belonged, he believed, to the final Athens sojourn. Gauthier sanguinary reign lasted from 364/3 to 353/2 .2 T h e reference would
and Jo lif, in both editions o f their massive and authoritative perhaps be more natural after his death ; at the very earliest it cannot
commentary on the N E , accept this chronology, dating the E E be much before 360.
shortly after 348/7 and the N E between 335 and 330.2 Som e more (3) A t i20 5ai9 -2 3 the writer illustrates a point about the
recent scholars, while accepting Jaeger’s order for the E E and the individuation of dispositions by a reference to the knowledge o f
N E , have argued for the authenticity o f the M M as an early work of grammar ‘ in Lam pros and in N eleus’ .3 No likely Lam pros is
Aristotle’ s : thus Dirlm eier proposed that they date from his period known; but Neleus is now universally taken to be Neleus o f
in the Academ y.3 T h e chronology put forward by D üring in his Skepsis, the son o f Coriscus and nephew o f Theophrastus, who
article in R E Supplem ent-Band x i.3 3 1 ff. dates the Magna M oralia inherited the latter’s library at his death c. 285 (Strabo, xiii.608).
before 360, the Eudemian Ethics between 355 and Plato’s death, the N eleus, Strabo tells us, was a pupil o f Aristotle. Unless he inherited
N E and the existing A E to the period between 334 and 332. the library at an improbably advanced age, he can hardly have been
T o date the Magna M oralia as it stands to a period before Plato’s old enough to know grammar during the 360s.4
death is quite impossible, for the following reasons. (4) A t I 2 i2 a4 the writer is discussing the difference between
(1) A t i Kj7b2 1, to illustrate the distinction between demotes and being well-disposed to someone and being their friend. ‘ I f there
phronësis, the author says ‘E ven a bad man may be called intelligent : were people as perhaps there were, well disposed to Darius far away
for instance M entor was generally thought intelligent, but he was in Persia, it does not follow that they were friends o f his.’ 5 T h e
not a wise m an.’ Commentators are unanimous in seeing here a reference is almost certainly to Darius Codomannus, who certainly
reference to M entor the Rhodian, who in 341/0, by ingenious and
1. I agree with D üring (1966) ‘D er Sinn von Nam en wie Alkibiades oder Lysander
treacherous means, entrapped Hermeias and sent him to the court war zu jeder Zeit verständlich. M entor kann nicht zu dieser Kategorie von Persönlichkeiten
o f the Persian K in g where he was crucified. T h e imperfect tense is gerechnet werden. N ur wenn dieser “ heimtückische M ann” noch in aller Gedächtnis war,
hatte es einen Sinn, ihn als Beispiel hinzustellen.’
often, though not always, taken to indicate that the sentence was
2. R E , s.v. ‘K learchos’ . One manuscript adds to the list the name Ε ύ μ ά ν θ η ς . N o such
written after M entor’s death (c. 337/6).4 T h e allusion would be name is known. Ε υ μ α θ ή ς was one o f the T hirty T yrants, but there seems no reason to single
clear to Aristotle’s own circle ; Aristotle him self was deeply moved him out (Xenophon, Hell. 1 1 . 3 2 ; R E vi.1075). Perhaps the text should read Έ ρ ιά ν θ η ς .
Erianthes, the commander who wished on his own authority to destroy Athens after
by H erm eias’ death and wrote a paean in his praise which later led
Aegospotami, might well have lived in Athenian memory as an arch-evildoer. N o other
to his own prosecution for impiety. T h is tells, as supporters o f the credible name ending in α ν θ η ς is given in Hansen.
authenticity o f the M M have insisted, against the theory that the 3. T h e reading Ν η λ ( ΐ o f the manuscript M , though not accepted by Bekker and
Susem ihl, has been universally read since the article by Wilamowitz (1927).
work was written long after Aristotle’s death, when the memory o f
4. Von Arnim , who was anxious to place the date o f N eleus’ birth as early as he plausibly
could, concluded that he was probably born in the fifties, and would be about eighteen in 334,
1. See D üring, 1957, 249 ff. T h e crucial dates are given by D ionysius o f Halicarnassus in the date which at that time von Arnim favoured fo rth e /ΜΛί. H e argues convincingly that the
his letter to Am m aeus: see Chroust, 1973, 16-24. knowledge o f grammar in question is not that o f the professional philologist, but o f a
2. G authier, 1970, 6 1 - 2 . schoolboy (von Arnim , 1928).
3. D irlm eier, 1958, passim. 5. où γ ά ρ et τ ι ς η ν Δ α ρ ε ίω ζ ϋ ν ο υ ς èv π έ ρ σ α ις ό ν τ ι, ώ σ π ε ρ ίσ ω ς η ν, ε υ θ έ ω ς κ α ι
4. So von Arnim , 1924, 10. φ ιλ ία η ν α ύ τ ψ π ρ ο ς Δ α ρ ε ίο ν .
2 i8 The Aristotelian Ethics The Dating of the Aristotelian Ethical Treatises 219
had well-wishers among the anti-Macedonian party in Athens. have impressed other scholars more deeply than they impressed
Whether or not the imperfect tense indicates composition after von Arnim. Consequently, it has been postulated that this updated
D arius’ death in 330, the reference cannot be earlier than his juvenile Aristotelian treatise has reached us through another
accession in 336 and is most likely after his great battles against hand— whether, as D üring supposes, the updating itself was
Alexander at the Granicus in 334 and Issus in 333.
entrusted to someone else, or whether, as Cooper suggests, the
(5) T h e M M , when discussing deliberation, says that no one has updated course given by Aristotle has been preserved only in the
to deliberate how to spell the name ‘Archicles’ . T h e reference may lecture notes o f one o f the members o f its audience.1
o f course be to some member of the immediate circle o f the author Now1 o f course every scholar is at liberty to postulate as many
o f the M M who has, like Lam pros, been forgotten. Certainly it is juvenile treatises by Aristotle as he wishes, and as many revisions as
not easy to find an Archicles whose name would be on everyone’s he thinks necessary to explain the present state o f the manuscripts
lips : but the best known o f those who have come down to us was a o f the corpus. M oreover, if he is prepared to regard as detachable
trierarch in a battle fought in 334/3 (C IA 11.804“ 18). from the postulated treatise all chronologically locatable references
Taken together these references make impossible a date for the and all quantifiable features of style, his hypothesis will be
M M in the sixties, and make most likely a terminus post quem for the conveniently secure against any possibility o f objective falsification.
work of about 335. T h is fact is, o f course, well known to the B u t even safeguarded with this protective covering, a hypothesis
partisans of the view' that the M M is an authentic work o f Aristotle’ s
such as that o f Cooper can hardly escape the charge o f superfluity.
and the earliest o f his ethical treatises. Von Arnim , who was the first
I f we are to postulate, in order to explain the un-Aristotelian style of
and most candid o f the twentieth-century champions o f the
the M M , that w hat we have is the published notes of a student w ho
authenticity o f the M M , accepted that the references proved that
attended his course, we do not need to postulate further a juvenile
the M M could not be earlier than 334, and concluded that it, and version of the course in order to explain the crudity o f the
consequently the two later treatises the E E and the N E , belonged to philosphical discussion. For the crudity to be explained is not a lack
Aristotle’s final years o f teaching in Athens (1924, 1 0 - 1 2 ; 1928, o f overall philosophical vision or comprehensive organizing design :
i ff.).1 it is a constant botching o f the details o f a philosophical argument.
L ater partisans o f an Aristotelian M M have been reluctant to And this kind of incompetence can perfectly adequately be
follow von Arnim here. In the opinion o f all those who have explained by the mediocrity o f the note-taker who has already been
compared the works in the last century and a half,2 the M M is postulated to account for the style. T h e present state o f the M M is
philosophically naïve and crude by comparison with the E E and the overdetermined if we explain it both by the immaturity o f Aristotle
N E . I f we are to attribute it to Aristotle himself, then, it seems that and by the imperfect accuracy o f his recording disciple.
we must assign it to his apprentice years in the Academy rather than In itself, the hypothesis that the M M is a student’s presentation
to his magisterial period in the Lyceum . We shall have to postulate, o f his lecture notes of a course given by Aristotle in his final Athenian
therefore, to explain the historical allusions, a revision of the work, period is an attractive one. It explains, first o f all, the appearance of
or updating o f the course, in the late thirties. M oreover, the stylistic the work in the Aristotelian corpus better than the still most
differences between the M M and the Aristotelian ethical treatises popular hypothesis o f totally post-Aristotelian origin. It explains
1. A t the time o f writing the later monograph, von Arnim believed he had found
why the historical introduction to the work ends with Plato:
confirmatory evidence for this date for the Μ ΑΊ. Rhetoric B must be later than 338, since it something difficult to explain if the work was written by a later
refers to the spccch o f Dcm ades after the battle o f Chaeronea ( i4 0 ib32). But the treatment peripatetic ( n 8 2 a3o). It explains how the author can refer to
o f several passions in chapter 9 o f that work shows that Aristotle had not yet worked out the
doctrine o f the mean, which is clearly to be found in the M M . T h is therefore must be some
Aristotle’s doctrine o f the syllogism as his own ( i 2 0 i b25). T h e
years after 338 (loc. cit). crabbed syllogistic manner of much o f the work, so well brought out
2. T h e qualification is necessary because Schleierm acher, writing in 18 17 , regarded the
M M as the only authentic Aristotelian ethical treatise (18 3 5 , 123).
I. Cooper, I<)73, 327-49·
220 The Aristotelian Ethics The Dating o f the Aristotelian Ethical Treatises 221
in B rin k’s m onograph,1 sorts better with the diligence o f a student really solid evidence there is for assigning an early date o f this kind
polishing up his lecture notes than with Aristotle’s ubiquitous and to the E E . Von Arnim , as has been remarked, thought that the E E
lordly contempt for the technical terms and methodological devices belonged to the final Athenian sojourn. A n examination o f the
o f his own invention. I f not associated with the hypothesis o f an evidence will show that there is nothing to prevent, and m uch to
effectively vanished juvenile treatise, the theory does justice in an commend, such a chronology.
unforced way to the topical allusions which forbid the assignment T h e E E , like the M M , contains a number o f references to
o f the treatise to a period earlier than the middle thirties. Indeed, it contemporary persons and events. T h e mention o f Coriscus at
almost reconciles the parties to the traditional debate. For there is 1220 *19 and i24 0 b25 has been thought to confirm a dating for the
not a very great difference between a scholar who says that the M M E E in the early forties : for Coriscus was probably among the group
is not authentic Aristotle, but is the product o f a Theophrastean around Aristotle in Assos. But Coriscus, whether or not he listened
peripatetic, and a scholar who says that the M M is an authentic to lectures in Assos, occurs as a Jo h n D oe figure in so many o f
work, but has reached us through the hand o f Theophrastus or Aristotle’s writings, assigned by scholars to so many different
another o f Aristotle’s disciples. T h e difference between them, periods o f Aristotle’s life, that it is impossible to base any
rather, is not so much a difference about the historical prob chronological hypothesis on the use o f his name. Other references
abilities, as a difference in the rigour o f their standards o f authen are prima facie more useful for establishing a terminus post quem for
ticity.2 the E E .
I f the M M is a student’s version o f a course given by Aristotle, (1) ‘ Gourmands pray’, we are told at 1 2 3 1 “ 16, ‘ not to have a long
there is no reason why the course it records should not be a course tongue but to have a crane’s gullet: like Philoxenus son o f E ry x is.’
for which we have Aristotle’s own lecture notes : i.e. either the E E or Several gourmands seem to have borne the name : Philoxenus son o f
the N E . In fact, as we have seen, the M M follows the E E slavishly, Eryxis is possibly identical with the dithyrambic poet Philoxenus o f
in the order of topics and the manner o f their treatment, for almost Cythera (c. 435-38 0) who seduced Dionysius I ’s mistress Galatea.
the whole o f its length. B y far the simplest hypothesis to explain the A t all events he must have been old enough to be active during the
present state o f the M M is that it is a student’s published notes o f first part o f the fourth century, for Athenaeus tells us (5.220) that
the course which we have in its authentic form as the E E . With the Aeschines mocked Anaxagoras for having so disgusting a pupil as
exception of a few b rief passages, which will receive special Philoxenus son o f E ryxis. His being mentioned, therefore, is no
consideration shortly, there are no differences between the content help in assigning a date to the E E .
o f the E E and that of the M M which cannot be explained by a (2) ‘A man who is in love endures many dangers,’ says Aristotle
combination o f a pardonable degree o f incompetence in under at i229 a23, ‘like the man at Metapontum who murdered the tyrant,
standing, and a modest talent for editorial revision, on the part o f a and the person in Crete in the story.’ ‘D ie beiden Beispiele sind
disciple o f Aristotle present at the lectures. nicht zu verifizieren’ says Dirlm eier, commenting on the passage
Scholars have failed to see this obvious and attractive explana (1969, 315). T h e second reference is indeed too vague; but it is not
tion o f the evidence largely because o f their conviction that the difficult to find a story to fit the first. Plutarch (M oralia 760 c) refers
E E is an early work. I f the M M is based on notes taken at a course to Antileon o f Metapontum, along with Harmodius and
given during Aristotle’s final Athenian sojourn, then it cannot be a Aristogeiton, as instances o f tyrannicides motivated by love. T h e
record of the E E if that belongs to the period in the Academ y under story o f Antileon is told in detail by Parthenius in a passage drawn
Plato or in Assos with Hermeias. But we must ask ourselves what from the peripatetic historian Phanias o f Eresus (F H G II.298). In
love with the young Hipparinus, he carried out various dangerous
1. Brink, 1933. exploits to gratify his beloved’s whim. He met his death after
2. T o take a modern parallel : students o f W ittgenstein differ in the amount o f authority
they ascribe to G . E . M oore’s notes o f his lectures in the early thirties, or Barrett’s assassinating the tyrant Archelaus o f Heraclea in order to rescue the
reconstruction o f his lectures on aesthetics. young man from his attentions. A monument was erected to him
222 The Aristotelian Ethics The Dating o f the Aristotelian Ethical Treatises 223
when, shortly after, the city o f Heraclea regained its former the education o f his so n .1 But if so, the Pammenes mentioned can
democratic constitution. hardly be the well-known Theban general, though the name is not a
T h is allusion would assist in dating the E E if only we knew the common one. Prodicus o f Selym bria appears in Plin y’s apostolic
date o f the death o f Archelaus. Heraclea was not founded until 432, succession o f Hippocratic doctors between Hippocrates and
so that the story cannot belong to the sixth century, as some Chrysippus, the teacher o f Aristotle’s medical grandson
commentators, misled by Plutarch’s coupling o f Antileon with Erasistratus.2 T h is would place him in roughly the same generation
Harm odius and Aristogeiton, have suggested. T h e time at which as Aristotle, but does not enable us to fix his date with any precision.
the cities o f M agna Grecia, and Sicily, one after the other, were As for the story o f the harpist and the king, M ichael o f Ephesus tells
reverting from tyranny to democracy was during Tim oleon’s march us that some scholars thought it was Alexander while others
down the Italian coast and his expedition to dethrone Dionysius II thought it did not suit that monarch’s character.3 M odern scholars
o f Syracuse in 344 (see Hackforth, 1927, 288 ff.). But I have been on the basis o f a similar story in Plutarch have thought the allusion
unable to find confirmation for the conjecture that Archelaus’ is to Dionysius o f Syracuse.4 B u t would Aristotle have referred to
assassination is to be assigned to this period. either Dionysius as basileus? Despite its rich allusiveness the
(3) In the tenth chapter o f book V I I Aristotle is discussing passage leaves us in total uncertainty about its date.
friendships in which what is given and what is taken differ. Th ese (4) At i 2 3 i an Aristotle quotes an elegant remark o f the
often lead to trouble, he says. musician Stratonicus. No one knows for certain the dates o f this
wit, whose principal stock-in-trade appears to have been the kind o f
This is how it happens in love affairs, since in them one party pursues the
joke that N ew Yorkers like to make about Philadelphia. A survey of
other as a pleasant person to live with, but sometimes the other the one as
useful, and when the lover ceases to love, he having changed the other all the witticisms attributed to him — a depressingly unhilarious
changes, and then they calculate the quid pro quo, and quarrel as Python experience— shows that they cannot all have been made by the
and Pammenes used, and as teacher and pupil do in general (for knowledge same person, unless he lived from about the time o f Alcibiades’
and money have no common measure) and as Prodicus the doctor did with return to Athens in 407 until Ptolemy Soter declared him self king in
the patient who offerred to pay his fee with a discount and as the harpist 304·5
and the king fell out. (i243b17-25, trs. Rackham)1 Altogether, the undisputed books o f the E E offer us tantalizingly
T h e passage is rich in historical allusions. From the Greek it is clear little help to anchor their date. T h e various allusions can be made to
that Python and Pammenes were in the relationship o f teacher to fit with, but in no way enforce, a dating o f the E E during Aristotle’s
second Athenian period.
pupil.* T h ere is no difficulty in finding a famous teacher o f the
name o f Python: the Byzantine rhetorician, who came to Athens in T h e matter is altered if we restore the disputed books to their
343 as an ambassador from Philip ofM acedon, and who may indeed rightful place as part o f the E E . F o r in the disputed book C there is a
have conveyed the monarch’s invitation to Aristotle to undertake reference which is universally taken by commentators to be a
reminiscence o f the court o f Alexander. A t 1 i5 0 bi3 we are told o f a
Xenophantus who was unable to restrain his laughter on some
1. ο ϊο ν σ υ μ β α ίν ε ι έττί τ ώ ν ε ρ ω τ ικ ώ ν , ο μ ε ν γ α ρ δ ιώ κ ε ι ο ις [τον] η δ ν ν e m το α υ ζ ή ν ,
ο δ’ ε κ ε ίν ο ν ε ν ίο τ ε ώ ς χ ρ ή σ ι μ ο ν ό τ α ν δ έ π α ύ α η τ α ι τ ο ΰ έ ρ ά ν , ά λ λ ο ν γ ιν ο μ έ ν ο υ ά λ λ ο ς
γ ί ν ε τ α ι , κ α ι τ ό τ ε λ ο γ ίζ ο ν τ α ι τ ί άντ'ι τ ι ι ’α ς , κ α ι ώ ς Π ν θ ω ν κ α ι Π α μ μ έ ν η ς δ ιε φ έ ρ ο ν τ ο 1 . I f Aristotle was in Athens at the time o f receiving the summons, as seems to be implied
κ α ι ώ ς δ ιδ ά σ κ α λ ο ς κ α ί μ α θ η τ ή ς ( ε π ισ τ ή μ η γ ά ρ κ α ι χ ρ ή μ α τ α ο ύ χ έ ν ι μ ε τ ρ ε ίτ α ι) , κ α ι by his best-known joke (fr. 619, Rose) : έ γ ώ έ κ μ ε ν ’ Α θ η ν ώ ν ε ις Σ τ ά γ ε ιρ α ή λ θ ο ν δ ια τ ο ν
ώ ς ‘ Π ρ ό δ ικ ο ς ό ια τ ρ ό ς π ρ ο ς τ ο ν ά π ο δ ιδ ό ν τ α μ ικ ρ ό ν τ ο ν μ ισ θ ό ν , κ α ι ώ ς 6 κ ιθ α ρ ω δ ό ς β α σ ιλ έ α τ ο ν μ ε γ α ν , έ κ δ ε Σ τ α γ ε ίρ ω ν εις ’ Α θ ή ν α ς δ ιά το ν χ ε ιμ ώ ν α τ ο ν μ έ γ α ν .
κ α ί 6. L ik e Rackham I accept Jack son ’s emendation τ ί ά ντ'ι τ ίν ο ς ; but I reject Ή ρ ό δ ικ ο ς . 2 . Pliny, N .H . 29.2. Editors emend both Pliny and Aristotle to read ‘H erodicus’. B u t this
2. Jackson is surely correct in drawing this conclusion from the oAo k in the manuscripts involves gratuitously attributing to Pliny not only a confusion between master and pupil but
(19 0 0 ,4 2). W hether D irlm eier is correct in concluding that Python and Pammenes were also also a muddle about birthplaces.
lovers seems to me more doubtful : though a stormy affair with a male pupil is o f course not 3 . C A G X X .4 6 4 .
ruled out by the rollicking heterosexual domesticity attributed to Python in Athenaeus, 4. Plutarch, M oralia 33 f.
5-499- 5. R E iv A . 1 . 3 2 6 gives the dates o f his activity as 4 1 0 - 3 6 0 . See G ow , 1968, 7 0 .
224 The Aristotelian Ethics The Dating o f the Aristotelian Ethical Treatises 225
solemn occasion and broke into a guffaw. We know from Seneca i ig 6 b7 it is impossible to tell whether he was alive or dead at the
(De Ira ii.2) that Xenophantus was a musician in the service o f time o f writing, which may therefore be before or after 339. Plato is
Alexander. M oreover, his period at Alexander’s court must have twice quoted in the present tense ( n o 4 bi 2 ; i i 7 2 b28): no
been the very beginning o f his career, since he was still active conclusion can be drawn from this, for so is Hom er ( 1 1 i6 a22 etc.).
enough to play for the funeral o f Dem etrius Poliorcetes in 28 3.1 As Nothing in the N E forces us to a date much later than 3 5 3 ; but
long as book C was thought to belong to the N E , or to a post- equally, nothing disproves the traditional assignment o f it to
Aristotelian Eudemian Ethics, this passage excited no remark : but if Aristotle’s last years. Certainly he cannot have been a young man
it is indeed an allusion to the M acedonian court musician, and the when he introduced his course with the remark that a youth was no
A E belongs with the-Ei;, then we must date the E E well after Plato’s fit student o f ethics (1095*2).1
death, well after the period in Assos, and most probably therefore T h e historical data, then, are perfectly consistent with our
during the final Athenian period : the same period, in fact, to which hypothesis that the M M , while incorporating a small amount of
we were led by the historical allusions to date the Magna M oralia.2 material from an already existing N E , is essentially based on a
I f we turn to the N E , we find a very solid terminus post quem for course given by Aristotle during his last years at Athens which
the work in the allusion, in the discussion o f courage, to a battle ‘At corresponds in substance with our text o f the E E . But we must
the temple o f H erm es’, i i i 6 bi9- Thanks to the well-informed examine more closely the cases in which the text o f the M M
second-century scholiast (C A G xx.16 5) this can be identified with diverges from the structure and content o f the E E .
an event during the Sacred W ar in 353 at Coronea in Boeotia.3 We N o great difficulty is presented by the fact that the M M begins
know therefore that at least this part o f the N E cannot have been with a summary o f the thought o f previous ethical thinkers which
written earlier than the period just before Plato’s death. W e cannot does not appear in the E E . T h e content o f the M M introduction
assume from the casual nature o f the allusion that the battle was still (118 2 * 10 -3 0 ) would not be beyond the ability o f a moderately
news at the time o f writing : a little later, at 1117 * 2 7 , there is an talented student to construct for himself. On the other hand,
equally off-hand reference to a battle at the L o n g Walls o f Corinth Aristotle no doubt gave the course which we possess as the E E on
in 392. T h e reference to the splendid character o f Eudoxus in N E more than one occasion : and most o f us rewrite the beginning o f our
10 .2 ( i i 7 2 bi5) has seemed to many to betray the tones o f the lecture courses more frequently than we rewrite the rest o f them.
obituarist. According to the most usual chronology, Eudoxus died T h ere is evidence in the text o f the E E that the first chapters have
about 35 5.4 From the tensing o f the reference to Speusippus in been more carefully revised than the rest o f the work : they and they
alone are at pains to avoid hiatus.
T h e incorporation o f the Nicomachean material presents more o f
1. See Berve, 1926, s.v., citing Plutarch, Demetrius 53.
2. In the same passage there is a reference to a play o f Theodectes, whom Plutarch a problem. T h e two most substantial Nicomachean portions o f the
(Alexander 17) tells was a pupil o f Aristotle along with Alexander. T h is confirms the M M are the amalgam o f quotations identified by Allan, introduced
identification o f Xenophantus: but the exact dating o f Theodectes’ dramatic career is with the words ‘this can be seen from the Ethics' (1 i8 5 bi5 ) ,2 and the
uncertain. Susem ihl (R h . M us. liv.6 31 ff.) dates his first tragedy to 353, and Pickard-
Cam bridge placed his death (which we know occurred when he was forty-one, Plutarch, X section on self-love in chapters 13 and 14 o f book 2 ( i2 i2 * 3 8 - i3 b2).
Orat. 837 d), in 334. T h is dating fits the ancient literary evidence best; but on the basis o f the
Didaskatiai other scholars (W ebster, 1954, Diehl, R E v A .17 2 2 ) place his career six or more
years earlier. O f Carcinus, the other dramatist mentioned here by Aristotle, all we know is 1. T o illustrate the point that we do not deliberate about what is beyond our control E E
that he attended the court o f D ionysius I I . It would assist us to date the A E if we could be 1226*29 says ‘W e do not deliberate about affairs in In dia’ ; N E i i i 2 a28 says ‘N o Spartan
certain o f the identity o f Σ ά τ υ ρ ό ς ό φ ίλ ο π ά τ ω ρ in 114 ^ '34, the son o f Sostrates according deliberates about what form o f government would be best for the Scythians.’ I f we are to
to C A G X X .4 2 6 . M ore than one Pontic king bore his name in the fourth century (G auth ier- attach any importance to this difference, we m ight say that the Scythians became a poor
Jo lif, 1959, 624). example o f things beyond our ken with Philip’s intervention in Scythia in 339, while the
3. See R E , s.v. Onomarchos. Indians remained a good example until Alexander’s invasion o f India in 327. B y the end o f
4 . So Heath, 1 9 2 1 , 3 2 2 - 3 4 . M erlan, i 9 6 0 (like von Fritz, 1 9 3 0 ) , has argued for a redating Aristotle’s life both examples must have looked a little passé.
o f Eudoxus’ lifespan to 3 9 5 - 3 4 2 , making him survive Plato. 2. See above, p. 10.
220 The Aristotelian Ethics The Dating o f the Aristotelian Ethical Treatises 227
N ow it is a startling coincidence that at each o f the points where the treatise on friendship, Aristotle makes an obscure critical reference
M M draws extensively on the N E , there occurs in the text o f the E E to a book in which the same topic has been treated.
an allusion by Aristotle to an earlier work o f his— allusions which
If therefore one abstracted and treated knowledge as absolute and
scholars, blinkered by the dogma o f the priority o f the E E , have
independent and did n o t. .. (but this is obscure in the way it is written in
universally found baffling. the book, though in fact it need not be obscure) in that case there would be
A t 1 220b 1 1 in the E E , during the discussion o f the way in which no difference between knowing oneself and knowing another instead of
moral virtue concerns pleasure and pain, we read ‘A fter this comes oneself, and likewise no difference between another living instead of
the classification o f passions, powers, and dispositions made en tois oneself: whereas patently to be conscious and aware of oneself is more
apFllagmenois\x T h e three Greek words have often been found desirable. Two of the things in the book have to be put together : life is
puzzling and various emendations have been proposed. D . J . Allan desirable, but so is goodness. (i244b29-45a)1
has suggested two possible translations o f the manuscript reading: H ere a book— presumably one by Aristotle, else the author would
‘ in the cancelled version’ or ‘ in the separate section’ .2 In either case, have to be mentioned2— is being criticized for being obscure, and
Aristotle is referring here to an appendix. Rackham ’s comment on for excessively separating knowledge from goodness and regarding
the passage is interesting: it is, he says, ‘perhaps a reference to N E life, identified with this knowledge, as the only desirable thing. Can
i i0 5 b20, inserted in the belief that the E E is the later work’ (19 35 , we locate an earlier work o f Aristotle which is criticisable on these
247). Certainly, except for very slight variations, the E E text and grounds ?
the N E text are here very similar.3 T h e corresponding passage in M ost authors treat the reference as being to a lost work, and are
the M M (1 i86a9 ff.) comes right in the middle o f the Nicomachean therefore at liberty to invent for it a text suitable to be a peg for these
cento introduced with a reference to the Ethics ,4 If, as the words en criticism .3 But if we turn to the parallel passage in N E 9, we can find
tois apëllagmenois suggest, the E E contained (or the E E lectures an extant argument o f Aristotle’s which passably fits the E E ’s
appealed to) an appendix o f superseded material at this point, and if strictures. First, the passage (1 i7 0 a2 5 - b9) is intolerably obscure: it
the superseded material was in fact the N E or part o f it, then we contains a sprawling sentence o f 1 1 2 words whose construction is
would have a full explanation o f why the M M in this area impermeable even with the aid o f the five braces o f brackets
apparently deserts its usual model, the E E , for verbatim copying o f thoughtfully provided in Byw ater’s edition.4 T h e argument goes
the N E . approximately thus: life consists essentially in perception and
i2 2 0 bi i is not the only place in the E E where Aristotle refers to thought, both o f which are accompanied by self-consciousness : it is
his own earlier work. In the discussion o f self-sufficiency in the
1 . €i ο ΰν τ ι ς ά π ο τ ε μ ο ι κ α ι π ο ιη σ ε ιε το γ ιν ώ σ κ ε ιν α ύ το καθ'1 α ύ το κ α ι μ η . . . (άλλα
τ ο ύ τ ο μ ε ν λ α ν θ ά ν ε ι, ώ σ π ε ρ ε ν τ ώ λ ό γ ω γ ε γ ρ α π τ α ι , τ ώ μ ε ν τ ο ι π ρ ά γ μ α τ ι ε σ τ ι μ η
1. μ ε τ ά τ α ν τ α ή δ ια ίρ ε σ ις ε ν τ ο ΐς ά π η λ λ α γ μ έ ν ο ις τ ώ ν π α θ η μ ά τ ω ν κ α ι τ ω ν λ α ν θ ά ν ε ιν ), ούθ'εν αν δ ια φ ε ρ ο ι η το γ ιν ώ σ κ ε ιν ά λλ ον ά ν θ ’ α ύ τ ο ΰ ■ το δ ’ ό μ ο ιο ν τ ο ν ζη ν
δ υ νά μ εω ν κ α ι τω ν εζεω ν. ά ν θ ’ α υτο ύ ά λλον, ε ύ λ ό γ ω ς δ ε το ε α υ τ ο υ α ια θ ά ν ε σ θ α ι κ α ι γ ν ω ρ ίζ ε ιν a i ρ ε τ ώ τ ε ρ ο ν .
2. Allan, 19 6 1, 3 1 2 ? J 9^6, 148. In support o f the first sense o f α π α λ λ ά τ τ ο μ α ι Allan δ ε ι γ ά ρ ά μ α σ ν ν θ ε ΐν α ι δύο εν τ ώ λ ο γ ω , ό τι τ ε τ ο ζ η ν [κ α ι] α ιρ ετό ν , κ α ι ό τ ι το α γα θόν.
appeals to Plato, R ep. viii.559 b and ix.571 b. T h e text is corrupt in places, and I do not know how to emend it; but the general sense is
3· T h e E E , as is its wont, is a little conciser and more precise, specifying passion by its clear.
accompaniment with sensory pleasure and pain, and explaining that ε ξ ε ις are defined by 2. So D irlm eier (1969, 4 6 1): ‘ wir haben wieder jenes in Corpus Ar. einzigartige, a u f E E
being κ α τ ά λ ό γ ο ν η ε ν α ν τ ίω ς . und Poetik i454bi8 beschränkte, Sich-berufen a u f ein geschriebenes W erk, ein B u ch .’
4- M M 1 1 85bi4 reads: o n δ ε ή ε ν δ ε ια κ α ί ή ύ π ε ρ β ο λ ή φ θ ε ίρ ε ι τ ο υ τ ίδ ε ίν ε σ τ ιν ε κ 3. T h e most promising attempt hitherto to locate a source in the lost works o f Aristotle for
τ ω ν η θ ικ ώ ν · δ^Γδ* ύττερ τ ώ ν α φ α ν ώ ν τ ο ΐς φ α ν ε ρ ο ΐς μ α ρ τ υ ρ ιο ις . Allan comments : ‘T h e the quotation is that o f G aiser (1967) who thinks that the logos referred to is the Protrepticus
attempts o f scholars to emend ε κ τ ώ ν η θ ικ ώ ν (see Susem ihl) show that they failed to (see also W idmann, 1967). T h e two passages which G aiser cites as parallels from modern
recognize the words from δ ε ι to χ ρ ή σ θ α ι as a verbatim quotation o f E N 1 10 4 * 12 .. . . With reconstructions o f the Protrepticus—frs. B 7 1 - 7 , 97-/O J Düring— are far less close to the
this commences a passage o f about five T eubner pages in which phrases from the N E are allusions in the E E than the texts in N R 9 and N E 10 discussed below.
cited with some embroidery and expansion; that we find nothing sim ilar outside these pages 4. According to the figures given by Wake, 19 57, and M orton and W inspear, 19 7 1, this
o f M M ; and that in view o f this it is only reasonable to translate ίδ ε ίν ε σ τ ιν ε κ τ ώ ν η θ ικ ώ ν must be one o f the longest sentences in Aristotle. Wake, indeed, does not record any sentence
‘ it can be seen from the Ethics' (Allan, 1966, 14 2 ; see allan, 19 57, 7). in the N E , E E , or A E longer than 1 1 5 words.
228 The Aristotelian Ethics The Dating o f the Aristotelian Ethical Treatises 229
indeed the consciousness o f one’s own existence that makes life So we must study together and feast together and share the best
pleasant. B u t a good man looks on his friend just as he looks on things in life so far as we can attain them. T h e contrary view 1 which
him self: therefore consciousness o f his friend’s existence will be made the solitary life seem as valuable as life in the society o f friends
just as desirable to him as consciousness o f his own. Hence, a good was based on a false comparison between man and god. I f men were
man must have friends, or he will lack something desirable and will like gods they would not need friends; but if they were like gods
not be truly happy. they would not have any thoughts either, except o f themselves. ‘For
T h ere seems something very strange about this derivation o f the us, our well-being has reference to something other than ourselves ;
desirability o f friends from the value o f self-consciousness. T h e in his case he is him self his own well-being’ . (i2 4 5 bi8 -i9 )
argument in the E E shares some o f the strangeness, but it does Both in the N E and in the E E Aristotle, then, Aristotle argues to
criticize one thing which seems a feature o f the N E argument: the the value of friendship from the value o f self-consciousness and the
exaggerated value placed on the abstract awareness o f one’s own or role o f the friend as an alter ego. But the E E is anxious to criticize a
other’s existence. N o doubt it is nice for the good man to know he is presentation of this argument which divorces knowledge too much
good (i I70b5); but he has to be good, to exercise his virtues, and not from other values, which treats human beings as if they were pure
just to contemplate his goodness with self-satisfaction. It is in the godlike intellects, and which does not sufficiently stress the need for
exercise of the virtues— in the performance of good and worthwhile friends to share in all the elements o f the good life. It is striking that
actions— and not just in the contemplation o f them that we should all these points o f criticism are the very points in which, in the
surely look for the value o f friendship. It seems to be along these previous chapter, we identified the differences between the E E and
lines that the E E develops its own argument. Looking at life simply the N E with regard to the nature and constituents o f happiness.2
in terms o f the possession and communication o f knowledge leads T h ere is no need to look elsewhere than to the N E for the logos
to the conclusion that the society o f others is superfluous (i2 4 5 ai - which is here being subjected to critical commentary.3
19 ); but it is obvious that we all want to share things with our Once again, the conjecture that the E E here refers to the N E is
friends, whether bodily pleasure, or artistic culture, or philosophi confirmed by a comparison with the M M . T h e second large
cal inquiry. Hence there must be something wrong with an Nicomachean borrowing in the M M occurs at I 2 i2 a2 8 -b24, in the
argum ent1 leading to an opposite conclusion. (i2 4 5 a20-8) T h e section on self-love which in the N E immediately precedes the
knowledge of one’s friends who resemble one in various ways is passage corresponding to the E E criticism o f the logos, and which
indeed a way o f knowing oneself; and because a friend is an alter ego first introduces the identification of a human being with his nous
one can look on his good as on one’s own. But the value o f so doing (1 i68b29-69a3) which is developed in the N E 10 account o f
will depend fundamentally on the value o f the good shared. contemplative happiness. T h e M M then goes on, like the E E , to
criticize the excessive assimilation o f man with God which is to be
To share even common pleasures and ordinary life with a friend is
found en tois logois.4 Once again, the presence of the unusual
obviously pleasant (for it is always accompanied by consciousness of him) ;
but it is even more pleasant to share the more divine pleasures. This is ό τ ι a c t ή& ιον Ε αυτόν deœ petv èv τ ω β έ λ τ ίο ν ι ά γ α θ ω . το ύ το δ’ € σ τίν o re μ € ν π ά θ ο ς , ore
because it is always pleasanter to see oneself enjoying a superior good : and δ ί π ρ ά ξις> o re £è erep ov τ ι. et δ ’ α υτόν i ζ η ν κ α ί ο ν τ ω κ α ι τ ο ν φ ίλ ο ν , èv ôè τ ώ σ ν ζ η ν
this may be now an emotion, now an action, and now something different σ υ ν ε ρ γ β ΐν , ή κ ο ιν ω ν ία τ ω ν èv reAet μ ά λ ισ τ α y e.
from either. I f it is pleasant to live well for oneself and a friend, and if 1. λ ό γ ο ς is again ambiguous at I245bi2.
2. N E 1 0 , 1 17 8 ^ 0 - 2 4 , provides an example o f the excessive assimilation o f man with G od
living together involves working together, their sharing will be above all in which exactly fits the complaint o f E E i244b23 and I245bi 3 - i g .
the things which make up the end of life. (i245a38-b4)2 3. Several particular passages in the course o f the argument suggest that E E is
commenting on N E : thus E E 1245*2-^ expands and modifies— admittedly not in a very
1. It is not clear whether logos in line 27 means ‘argum ent’ or is another reference to the perspicuous manner— the brief allusion to what is ώ ρ ισ μ € ν ο ν at i i 7 0 a2 0 ; and E E 12 4 5 * 15 -
book criticized in lines 3 1 and 35 above. 16 corrects H 70bi2.
2. α>στ€ κ α ί τα φ ο ρ τ ι κ ά μ-kv σ υ ν ή δ ς σ θ α ι κ α ί σ ν ζ η ν τ ω φ ί λ ω η δ υ eùAoycoç 4. T h e plural may mean that M M has other texts too in view, perhaps including ones in
( σ υ μ β α ί ν € ΐ γ ά ρ ϊ κ ζ ί ν ο υ ά μ α α ί σ θ η σ ι ς act')· μ ά λ λ ο ν r a s fle io re p a s ή δ ο ν ά ς . α ί τ ι ο ν δ ’ which the comparison with God was made more explicit. M M differs from E E here in that
230 The Aristotelian Ethics The Dating o f the Aristotelian Ethical Treatises 231
Nicomachean material in the M M would be simply explained by further on the way to publishable form than the N E ever w as.1 B u t
the presence in the E E o f a Nicomachean appendix: in this case, scholars are coming more and more to agree that A ristotle’s works
chapters 8 and 9 o f N E 9 .1 cannot be ordered on a scale o f deviation from Plato : his philosophy
M ost scholars will reject these conjectures out o f hand as was a lifelong development o f Platonic themes and a lifelong
presupposing the existence o f at least parts o f the N E at the time confrontation with Platonic problem s.2
when the E E was written or delivered as a course o f lectures. T o be sure, there is evidence that at different times in his career
Certainly it is difficult to find a scholar who will seriously consider Aristotle’s attitude to the Platonic theory o f Ideas varied between
the possibility that the N E antedates the E E : but the time has come different degrees o f sympathy and harshness. But it is unwise to
to look at the evidence for the priority o f the E E . conclude that the sympathetic criticisms are early and the harsh
Probably the commonest ground for assigning the E E to an criticisms are late ;3 it is equally imprudent to draw the opposite
earlier date has been the belief that it is more Platonic than the N E . conclusion, imagining a pupil growing gradually more tolerant o f
In so far as this claim rests on the actual doctrines o f the ethical his teacher’s doctrines under the mellowing influence o f age. I f we
treatises the claim seems unfounded : as we have shown, the E E is compare the criticism o f the theory o f Ideas in N E 1.6 with that in
no m ore Platonic than the N E in its treatment o f wisdom, and E E 1.8 I think it is undeniable that the criticism o f the E E is made
considerably less Platonic than the N E in its refusal to identify from a greater distance. In the N E Aristotle introduces his attack
happiness with the activity o f the separable nous. I f the claim rests with an apology and an admission that the inventors o f the theory
on the constant allusions to Platonic themes and terminology, then are dear and close to him : ‘ amicus Plato sed magis amica veritas’—
it must be admitted that the E E is highly Platonic : it is shot through an apology which tactfully echoes Plato’s own remarks on Hom er
with reminiscences o f Plato’s ethical dialogues from the Euthyphro (i096an - i 7 ) . In the E E he begins: ‘the existence o f a Form not
to the Philebus. But so too is the N E . I do not know that anyone has only o f good but o f anything else is an idle abstraction : this has been
undertaken a count to see which o f the two treatises contains the considered in multifarious ways both in popular and philosophical
more explicit and implicit allusions to Platonic texts. T h e exercise writings’ (i2 4 7 b20 f.). T h e arguments in the N E share more
would no doubt be futile : but it might well be that there are more Platonic premisses than those in the E E , and are less sharply and
clearly identifiable Platonic citations in the E E . T h is would merely crisply developed.4 I f the differences between the two treatments
be an instance o f a general feature o f the style o f the E E remarked on were to be settled by assigning a temporal order between them, it is
long ago by G rant, a more copious and explicit manner o f undoubtedly the E E that an open-minded reader would regard as
quotation. T h is stylistic feature may merely indicate that the E E is the later. But the reason for the differences may well be quite other
than a lapse o f time.
When nineteenth-century scholars sought to show that the E E
instead o f using the theory o f divine self-contemplation o f Metaphysics as part o f a reductio ad was post-Aristotelian they pointed to the fact that there are some
absurdum o f the comparison between man and god, it introduces (whether or not it endorses) virtues for which Aristotle tells us in the N E he has to coin a name,
a comparison between m an and god as part o f a criticism o f the Metaphysics Λ view : i f a man
spends all this time gazing at his own navel we regard him as a dunce; how can we accept a 1. G rant, 1885, 30: ‘T h e Eudem ian writer . . . shows indeed a proclivity to indulge in
self-contem plating G od ? ( i 2 i 2 a3 6 - i 3 a8) abundance o f literary quotations, and he quotes more fully and explicitly than A ristotle.’
i. In one other place £ £ refers to a logos, in the criticism o f P lato’s Ideas. € τ ι κ α ι t o i v τ ω 2. F o r D irlm eier, Aristotle’s development is ‘Reifen, in Verbindung mit Plato, bis zuletzt’
λ ό γ ω γ ε γ ρ α μ μ ί,ν ο ν . η γ α ρ ο ν & ξμ ιά χ ρ ή σ ιμ ο ν a ν τ ο τ ο τ ο ν ά γ α θ ο ν ξ ϊδ ο ς tj π ά σ α ις (1969,568).
ο μ ο ίω ς ■ ζ τ ι ού π ρ α κ τ ό ν . T h is too would fit the N E well: after the arguments which are 3. T h e best treatment o f the E E and N E criticism o f the Ideas is Flashar, in Festschrift
common to both treatises the N E then goes on to an argument which depends on the Schadewalt, 2 23-37 . Flashar notes that the E E contains two arguments not to be found in the
disjunction: either the definition o f good applies only to the idea, in which case μ ά τ α ιο ν N E , and thinks that while all the N E arguments are directed against Plato the E E arguments
€ σ τ α ι το € Γδο? ; or else we have to suppose the same definition in goods o f diverse kinds which are against Xenocrates— though not necessarily after Plato’s death.
are related only analogically (i0 9 0 b8 - 3 i) ; he then goes on to object that the idea ο ν κ α ν ς ϊη 4 - Flashar’s own explanation— that the N E is more exoteric than the E E and therefore
'π ρ α κ τ ό ν ουδβ κ τ η τ ό ν ά ν θ ρ ώ π ω (b3 I -~5)· T h e note again is easily explained as a reference more apology to Plato seemed called for hardly seems plausible in view o f the generally
to an appendix, or hand-out, o f Nicomachean material. more finished and polished form o f the EE.
232 The Aristotelian Ethics The Dating o f the Aristotelian Ethical Treatises 233
but which are alluded to by that name without hesitation in the E E virtue to magnanimity ( i i 2 5 bi- 2 5 ) — it may be called philotimia
and the M M . T h u s, at n o 8 a4 -9 we read: without the pejorative overtones o f that word. T h e E E knows no
such virtue. Instead, in its list of good passional means (i2 3 4 a24) it
In regard to anger also there exists an excess, a deficiency and a mean. has the quality o f semnotës (dignity), a mean between obsequious
Although there are really no names for them we might call the mean ness and self-will. N o doubt it is rash to draw any conclusions from
praotes (gentleness) since we call a man who occupies the middle position
these variations in emphasis. But it is perhaps worth pondering the
gentle. O f the extremes, let the man who exceeds be called orgilos (short-
tempered) and his vice orgtlotês, and the deficient man aorgetos (apathetic) following question. Who is the younger man, and who is the older
and his vice aorgësia. m an: the author o f the N E who prizes highly proper ambition,
magnificence, and magnanimity, or the author o f the E E who is
Again in book 4 Aristotle says that the middle here is nameless, and m ore attracted by dignity and gentleness?1
the extremes almost so (1 i2 5 b26 ff.). B u t in the E E praotes is used W e come finally to an argument which probably most scholars
without apology at i2 2 0 b38, as is orgilotes (the corresponding defect today would regard as the strongest for dating the E E early, or at
is here called analgesia) ; praotes is again used without hesitation in least earlier than the N E : the treatments o f pleasure in A E C and
I I I . 3, i 2 3 i b5 f f ., where various names are offered for the N E 10. M ost scholars are convinced that the treatment in A E C is
corresponding defects, including orgilos. Anaisthësia is listed at Eudemian, even if they are prepared to regard most o f the A E as
i 2 2 i a2 as the contrary vice to intemperance : at I2 3 0 bi4 we are told Nicomachean: for as part o f the N E it makes such a puzzling
that some people call those uninfluenced by pleasures anaisthetoi doublet with N E 10 which makes no reference to it. M ost scholars
and others give them other names. In N E at 1 i0 7 b8 Aristotle says are also convinced that the treatment o f N E 10 is philosophically
that such rare birds are nameless, and offers anaisthetos as a coinage. more sophisticated than that in A E C : consequently, they say, since
Again, at i i 2 6 bn a virtue o f sociability is introduced which is A E C belongs with the E E and N E 10 with the N E , and N E must be
described as having no name (cf. n o 8 ai7 ) : Aristotle settles for later than the E E .2
calling it philia. T h e name is used without embarrassment at 12 2 i a7 Recently, G . E . L . Owen has argued that the two treatments are
and i2 3 3 b30 in the E E . Such examples do not show that the E E is not doublets because they are not about the same topic. T h e A E
post-Aristotelian: but the nineteenth-century scholars may well treatment offers to tell us what our real pleasures are; the N E
have been right to see them as an indication that the E E postdates discussion is designed to tell us the nature o f enjoying by reviewing
the N E .1 the logical characteristics o f pleasure-verbs.3 It is a mistake, he
It is not only differences in nomenclature which distinguish the thinks, to draw a contrast, as many scholars do, between a treatment
treatments o f the virtues in the N E (3-4 ) and in the E E (III). T h e which identifies pleasure with unimpeded activity and a treatment
ampler and more vivid treatment o f the particular traits o f character which regards pleasure as an epiphenomenon supervenient on
is one o f the features which contributes to making the N E longer activity.4
than the E E . In particular the virtues o f magnificence and 1. M y colleague, M r. J . G riffin o f Balliol College, has suggested that if the E E is later than
magnanimity are given longer and more loving treatment in the N E the N E the difference between the two treatments could be explained by disillusionment with
(six and a half Bekker columns) than in the E E (less than three the career o f Alexander.
2. See Festugière, 1946; Lieberg, 1958; G au th eir-Jolif, 19 5 8 ,77 8 ff. Our stylistic criteria
columns).2 Gentleness, on the other hand, while treated like all the support the common view that A E C ’s treatment o f pleasure is Eudem ian and N E io ’s is
virtues at greater length in the N E , is upgraded in the E E to third Nicomachean : T ab le 9.1 shows the result o f comparing the two treatments with expectations
place among the virtues. T h e lists in the two Ethics agree in their drawn from the E E and from the remainder o f the N E . T h e comparison would hardly have
been worth making, for the point is so generally accepted, were it not for theattem pt o f W ebb
content, though not in their ordering, with one exception. T h e N E (Phronësis 1977, 236) to show that it is the C treatment that is the Nicomachean one.
has a place for a nameless virtue o f proper ambition, a satellite 3. Owen, 1972.
4. T hus Anscom besays that pleasure ‘astonishingly, reduced Aristotle to babble, since for
good reasons he both wanted pleasure to be identical with and to be different from the activity
1. See Spengel, 18 4 1, 457.
that it is pleasure in’ (19 57, 76).
2. S e e D . Rees, 19 7 1, for a detailed comparison between the treatments.
234 The Aristotelian Ethics The Dating of the Aristotelian Ethical Treatises 235
T a b le 9.1 Nicomachean and Eudemian Favourite Expressions in the Two them appear both in the A E and in the N E . At the bottom o f the
Treatments o f Pleasure scale there are the pleasures o f the sick (sick either in body or soul)
which are not really pleasures at all : these appear in A E at i i5 3 b33
Group Total in Expectation Actual Expectation Actual and i i 5 4 a32, and in B at i i7 3 b22 and i i 7 6 aig. N ext up the scale
N E less in N E in N E in A E C in A E C come the per accidens pleasures o f food and drink and sex as enjoyed
10 'I -5 10· 1-5 io -1-5 1 1-14 11-14
by the gourmand and the lecher : these appear in A E at 1 15 2 b35 ff.,
H 5 4 ai7 , a35~b3 where they are distinguished from the lower class
A-favourites
o f false pleasures.1 Th ese pleasures appear in N E at i i 7 3 b8 - i5 .
N I 542 42-38 45 24-48 18
N ext up the hierarchy come the aesthetic pleasures o f the senses.
N II 453 35·42 29 20-11 7 T h ese are divided into two classes— the pleasures o f the inferior
N III 38s 30 1 1 33 I 7-44 6
N IV 2 9 17 16-86 12 sense o f touch and taste, and the pleasures o f the superior senses o f
373 31
NY 29-48 30 16-98 6 sight, hearing, and smell. These are mentioned in A E at i i 5 3 a26
377
N VI 403 31-46 H I 7-39 7 and in N E especially at i i7 4 bi4 -7 5 aio and i i7 5 b36~76a2. Finally
there are the pleasures o f mind mentioned in A E at 1 15 3 s! and 20,
2533 19802 182 113-27 56
and in N E at 1 173h 17, i i 7 4 b2 i, and b34 and elsewhere.
Secondly, the nature o f pleasure and its relation to activity. T h e
passage in the N E which is usually thought to distinguish pleasure
from activity in a way which is incompatible with the A E is
Group Total in Expectation Actual Expectation Actual
EE in N E in N E in A E C in A E C H 74b23~32 :
10-1-5 10 1-5 1 1 —14 1 1 —14 Each sense has a corresponding pleasure as also have thought and
speculation, and its activity is pleasantest when it is most perfect, and most
Æ-favourites perfect when the organ is in good condition and when it is directed to the
E I 31 1 33-90 H 19-47 25 most excellent of its objects; and the pleasure perfects the activity. The
E II 263 28-67 16 16-46 24 pleasure does not however perfect the activity in the same way as the object
E III 279 3°"4 I II 17-46 25 perceived and the sensory faculty, if good, perfect it ; just as health and the
E IV 256 27-90 12 16-02 25 physician are not in the same way the cause of being healthy. . . . The
EM 267 29-10 12 16-71 13 pleasure perfects the activity, not as the fixed disposition does, by being
E VI9 380 4 14 2 39 23-79 18 already present in the agent, but as a supervening perfection, like the
190-88 104 109-98 130 bloom of health in the young and vigorous, (trs Rackham)
1756
F r it z sc h e , A. T . H., Eudemi Rhodi Ethica (Aristotelis Ethica Eudemia), J o hn so n , H., ‘ The use of Statistics in the Analysis of the characteristics of
Regensburg, 1851. Pauline Writing’, New Testament Studies, 1973.
F K . von, ‘E u d o x o s vo n K n id o s ’ , Philologus lx x x v (19 3 0 ), 4 7 8 - 8 1 .
r it z ,
K a p p , E., Das Verhältnis der Eud. zur Nik. Ethik, Berlin, 1912.
F u r le y , D., T wo Studies in the Greek Atomists, Princeton, 1968. K ea r n e y , J. J., ‘Two notes on the Tradition of Aristotle’s Writings’,
G adam er , H .-G ., ‘Der aristotelische Protreptikos und die entwick American Journal o f Philology 84 (1963), 52-63.
lungsgeschichtliche Betrachtung der aristotelischen Ethik’, K e n n y , A., Action Emotion and Will, London, 1963.
Hermes 63 (1928). ------, ‘Happiness’, Proceedings of the Aristotelian Society 66 (1965-6).
G a iser , K . ‘Zwei Protreptikos-Zitate in der E. E. des Aristoteles’, ------, ‘A stylometric study of the Aristotelian Corpus’ CIRPHO, autumn
Rheinisches Museum N .F. 110, 4 (1967). 1976.
G a u th ier , R. A., Introduction to Second Edition of l'Ethique à L eonard , J., Le Bonheur chez Aristote, Brussels, 1948.
Nicomaque, Louvain, 1970. L ie b e r g , G., Die Lehre von der Lust, Munich, 1958.
G a u th ier , R. A. and J o l if , J . Y ., L'Éthique à Nicomaque, Louvain, 1958 M an sio n , A., ‘La genèse de l’oeuvre d’Aristote d’apres les travaux
and 1959. récents’, Revue néoscolastique de la philosophie 29 (1927).
G o h lk e , P., ‘Die Entstehung der aristotelischen Ethik, Politik, Rhetorik’, M a n sio n , S., Aristote et les Problèmes de Méthode, Louvain, 1961.
S B Wien 223 (1944). M a r g u er itt e , H., Review of Jaeger, ‘Über Ursprung und Kreislauf des
G o sl in g , J. C. B., ‘More Aristotelian Pleasures’, Proceedings o f the philosophischen Lebensideals’, Revue d'histoire de la philosophie 4
Aristotelian Society (1973) 15 ff. ( i 93°)·
Gow, A. S. F., Machon, Cambridge, 1965. M a s e l l is , V., ‘T raduzione e cataloghi delle opere aristoteliche’, Rivista di
G r a n t , A., The Ethics o f Aristotle, London, 1857. filologia e d'istruzione classica 34 (1956).
G reenw ood , L . H. G ., Aristotle, Nicomachean Ethics Book VI, M auru s , Silvester, Collected Works, ed. F. Ehrle, ii, Regensburg, 1885-6.
Cambridge, 1909. M ein e c k e , A., ‘Zu Stobaeus’, Zeitschrift f.d. Gymnasialwesen 13 (1859).
G r if f in , M . T ., Seneca, a philosopher in politics, Oxford, 1976. M er c k en , H. P. F., The Greek Commentaries of the Nicomachean Ethics o f
H a ckfo rth , R., ‘Sicily, 367 to 330 b .c .’ in The Cambridge Ancient History, Aristotle in the Latin Translation of Robert Grosseteste, i, Leiden, 1973.
VI, Cambridge, 1927. M e r la n , P., Studies in Epicurus and Aristotle, W iesbaden, i960.
H a l l , R., ‘The Special Vocabulary of the Eudemian Ethics’, Classical M ich aelso n , M . and M orton , A. Q., ‘The Authorship and Integrity of
Quarterly n .s . 9 (1959). the Athenaion Politeia’, Proceedings of the Royal Society of Edinburgh,
H a n sen , B., Rückläufiger Wörterbuch der griechischen Eigennamen, Berlin, 1973·
M in g a y , J., ‘Coniunctio inter Homines Hominum’ in Stern, Hourani,
I957·
H a r d y , W. F. R., Aristotle's Ethical Theory, Oxford, 1968. and Brown, 1972.
H a r l f in g e r , D., ‘Die handschriftliche Überlieferung der Eudemischen M on an , J. D., Moral Knowledge and its Methodology in Aristotle, O xford,
Ethik’ in Harlfinger and Moraux, 1971. 1968.
------and M oraux , P., Untersuchungen zur Eudemischen Ethik, Berlin, M oraux , P., ‘L ’exposé de la philosophie d’Aristote chez Diogène Laerce’,
Revue philosophique de Louvain (1949).
1971-
H eath , T ., A History o f Greek Mathematics, Oxford, 1921. ------, ‘Das Fragment Θ, 1 ’ in Harlfinger and Moraux, 1971.
H i c k s , R . D ., Diogenes Laertius (L o e b edn.), H a rv a rd , 19 2 5 . ------, Les Listes anciennes des ouvrages d’Aristote, Louvain, 1951.
J ackso n , H ., The Fifth book o f the Nicomachean Ethics of Aristotle, ------, Der Anstotelismus bei den Griechen, i, Berlin, 1973.
Cambridge, 1879. M orton , A. Q. and W inspea r , A. D., It's Greek to the Computer,
------, ‘ On some passages in the seventh book of the Eudemian Ethics, Montreal, 1971.
Cambridge, 1900. M orton , A. Q. and M c L em an , J. J., Paul, the Man and the Myth,
J aeger , W., Aristotle: Fundamentals of the History o f his Development (trs. London, 1966.
Robinson), Oxford, 1948. M u h l l , P. von der, De Aristotelis Ethica Eudemea, Göttingen, 1909.
------, ‘Aristotle’s use of medicine as a model of method in the Ethics', M u l l e r , A., Die griechischen Philosophen in der arabischen Überlieferung,
Journal o f Hellenic Studies 1957, 58 ff. Halle, 1873.
244 The Aristotelian Ethics Bibliography 245
M unro , H. A. J., ‘Nicomachean Ethics, Book V : Eudemian Ethics, Book S tew a rt , J. A., Notes on the Nicomachean Ethics o f Aristotle, Oxford,
IV ’, Journal o f Classical and Sacred Philology 2 (1855). 1892.
M u r r a y , O., ‘Philodemus and the good King ’, Journal of Roman Studies S to ck , St. G., Introduction in Ross, 19 25.
55 (1965)· S u se m ih l , F., Eudemii Rhodii Ethica ( Aristotelis Ethica Eudemia),
N ew m an , W. L., Aristotle’s Politics, Oxford, 1887-1902. Leipzig, 1884 (Amsterdam, 1967).
N u y en s , F ., L ’Evolution de la psychologie d ’Aristote, Louvain, 1948. ------5 ‘Die Lebenszeit des Theodektes,’ Rhenisches Museum, L IV .
O w en , G. E. L ., ‘Logic and Metaphysics in Some Earlier Works of T h e il e r , W ., ‘Die Grosse Ethik und die Ethiken des Aristoteles’, Hermes,
Aristotle’ in Aristotle and Plato in the Mid-Fourth Century, Goteborg, 69 (1934)·
i960. T it z e , F. N., De Aristotelis Operum serie et distinctione, Leipzig, 1826.
------, Tithenai ta phainomena in Mansion, 1961. V er b e k e , G., ‘L ’idéal de la perfection humaine chez Aristote et l’évolution
------, The Platonism of Aristotle, London, 1965. de sa noétique’ , Fontes Ambrosiani, 25 (19 5 1).
------, ‘Aristotelian Pleasures’, Proceedings o f the Aristotelian Society 72 W a g n er , D ., Das Problem einer theonomen Ethik bei Aristoteles,
(19 7 1-2 ). Heidelberg, 1970.
P aret , R ., ‘Notes bibliographiques sur quelques travaux récents’ , W ak e , E., ‘Sentence-length Distributions of Greek Authors’, Journal of
Byzantion 29/30 (1959/60). the Royal Statistical Society, i960.
Aristotle and the Arabs, N e w Y o r k , 1968.
P e t e r s , F . E ., W e b b , P., ‘The relative Dating of the Accounts of Pleasure in the
R a bin o w it z , W. G., Aristotle’s Protrepticus and the Sources o f its Nicomachean Ethics’, Phronësis, 1977.
Reconstruction, Berkeley, 1957. W ebst er , C., ‘Fourth Century Tragedy and the Poetics’, Hermes 82
R ackh am , H., The Eudemian Ethics (Loeb edn.), London, 1935. (1954)·
R e e s , D ., ‘ M ag n an im ity in the E E and N E ’ , in H arlfin g er and M o ra u x , W alzer , R., Greek into Arabic, London, 1984.
1971. ------, Magna Moralia und Aristotelische Ethik, Berlin, 1929.
R ist , J. N., Epicurus, an Introduction, Cambridge, 1972. W e h r l i , F., Die Schule des Aristoteles, Basel, 1959 ~.
R ose , V., De Aristotelis librorum ordine et auctoritate, Berlin, 1854. W id m a n n , G. Autarkie und Philia in den Aristotelischen Ethiken,
------, ‘Uber die griechische Kommentar zur Ethik des Aristoteles’ , Tübingen, 1967.
Hermes 5 ( 18 7 1) . W ilam o w itz - M o ellen d o r ff , U. von, ‘Neleus von Skepsis’, Hermes 62
Ross, W. D., Aristotle, London, 1923. (1927).
------, The works o f Aristotle (Oxford translation), vol. ix, Oxford, 1925.
------, Aristotle’s Metaphysics, Oxford, 1924.
R o w e , C. J., The Eudemian and Nicomachean Ethics: A study in the
development of Aristotle’s Thought, Cambridge, 1971.
S chacher , E. J., Studien zu den Ethiken des Corpus Aristotelicum,
Paderborn, 1940.
S ch leierm ach er , F ., Über die ethischen Werks des Aristoteles,
= Samtliche Werke, iii.3, Berlin, 1835.
S olom on , J ., ‘Translation of the Eudemian Ethics’, The Oxford Aristotle,
ix, Oxford, 1925.
S p e n g e l , L., ‘Über die unter den Namen des Aristoteles erhaltenen
ethischen Schriften’, Abhandl. der Bayer. Akademie 3 (1841 and 1843).
------, ‘Aristotelische Studien’, Abhandl. d. Bayer. Akademie, Munich, 10
(1864) and (1866).
S t e r n , S . M., H o u ran i , A. and B row n , V., Islamic Philosophy and the
Classical Tradition, London, 1972.
INDEX