Greek, Roman, and Byzantine Studies 47 (2007) 3970

2007 GRBS




Scholiasts and Commentators
Nigel Wilson
R JOHNSONS DEFINITION of the lexicographer as a
harmless drudge is well known. Perhaps the scholiast
was also a drudge, but whether harmless is the right
adjective for him is an intricate question, which I shall attempt
to explore in this paper.
1

To interpret scholia exactly is not always a simple task. So
wrote Eduard Norden, and he gave a striking example. A
scholiasts note on Vergils fourth Eclogue appears to give us the
authority of the excellent scholar Asconius Pedianus for the
view that the child celebrated by the Eclogue was Asinius Gal-
lus.
2
Norden argued that if the scholium is read attentively it
demonstrates the exact opposite: Asconius was in fact refuting
on chronological grounds Gallus claim to be the child in ques-
tion. The matter was incidental to Nordens main purpose, and
so he did not explain how it comes about that scholia are diffi-
cult to interpret. Other scholars have not always recognised the
existence of the difficulty, to their cost, as examples will show.
The nature of the problem and its significance will be one of
my main themes, but in addition I shall try to show that at their
best scholiasts offer something of value to their modern succes-
sors, who disdain the title of scholiast and replace it with that of
commentator.

1
The original version of this paper appeared in Italian translation with
the title Scoliasti e commentatori in Studi classici e orientali 33 (1983) 83
112. It had been my intention to issue a revised and enlarged text within a
few years, but other commitments took precedence. However, thanks to the
kind invitation of Prof. Bernhard Zimmermann to conduct a seminar on
this subject at the University of Freiburg in May 2006, I have had the
opportunity to prepare this new version.
2
Die Geburt des Kindes (Leipzig 1924) 11 n.1, citing Servius Danielis on Ecl.
4.11.
D
40 SCHOLIASTS AND COMMENTATORS

I.
Scholia are generally defined as the commentaries on clas-
sical authors, both Greek and Latin, written in the margins of
the medieval manuscripts that transmit the texts. This defini-
tion makes them a product of the middle ages, for the obvious
reason that in the vast majority of cases we can only read them
in the form they had assumed by that stage in their history.
Some scholia are the work of medieval scholars and school-
masters, whose names in certain cases are known to us, but it is
a mistake to suppose that the content of the scholia as a whole
is medieval in date, or that the form in which they appear in
manuscripts and printed editions can be proved to be a
medieval invention.
The history of scholia goes back to the time when explana-
tions of literary texts first became necessary, in other words the
classical age of Greece, when schoolmasters found that pupils
required some explanation of rare words and other difficulties
in Homer and the lyric poets, the basic texts of a literary educa-
tion. The pupils needs can be inferred from a fragment of Ari-
stophanes first play, Daitaleis, produced in 427 B.C., in which
one speaker asks the other the meaning of some Homeric
words (fr.233). The word scholium itself presumably means a
short lecture.
3
The growth of this form of literary scholarship
was slow. Though few if any of the schoolmasters wrote down
in a form suitable for publication their stock-in-trade of class-
room material, it is a plausible conjecture that the simple ex-
planations of Homer now known as the D-scholia depend
ultimately on their work. The ancients, who liked to be able to
identify inventors, had a tradition that the first person to write
about Homer and by the same token to undertake grammatical
studies was Theagenes of Rhegion, whose floruit they located,
rather strangely, in the reign of the Persian king Cambyses
(529522 B.C.) (frr.1, 1a, 2 in Diels-Kranz).
4
The next identi-
fiable figure is the poet Antimachus, who may be credited not

3
Not so defined in LSJ, but see G. Zuntz, Byzantion 14 (1938) 545614, at
548.
4
Presumably he had no connection with Athens, and so they had no rea-
son to call him a contemporary of Peisistratus.
NIGEL WILSON 41

with a commentary on Homer but some kind of recension of
the text and a life of the poet. That discussion of Homer was
not confined to the classrooms of elementary schools is a rea-
sonable inference from the fragments of Aristotles Homeric
Problems (frr.142179 Rose). The fourth-century Derveni pa-
pyrus with a commentary on an Orphic text is of uncertain
authorship; among the candidates proposed is Diagoras of
Melos, which would make it a work of the late fifth century.
5

One detail of interest for the present purpose is that the auth-
ors vocabulary includes the term hyperbaton.
6

Homer retained his pride of place in the school curriculum
and so won most attention from the Alexandrian critics. Their
efforts to establish the text generated new forms of scholarship.
It was not the commentary that was produced first, but various
sophisticated kinds of pamphlet, corresponding roughly to
articles in modern learned journals. An early example is Apol-
lonius Rhodius pamphlet on the recension of the great epics
prepared by his predecessor Zenodotus, the purpose of which
must have been to discuss disputed readings and inter-
pretations. It is not clear how much Zenodotus wrote,
7
and
Apollonius did not write any commentary on a text. The first
attestation of such a book is the mention of a commentary by
Euphronius on Aristophanes Plutus;
8
the first writer of com-
mentaries in any quantity was Aristarchus, in the second cen-
tury B.C. Aristarchus may also have widened the sphere of the
commentator by writing on a prose author, Herodotus;
9
pre-

5
R. Janko, ZPE 118 (1997) 6194.
6
A. Lamedica, in P. Radici Colace and M. Caccano Caltabiano (eds.),
Atti del I seminario di studi sui lessici tecnici greci e latini (Messina 1991) 8391.
7
A new hypothesis about his work on Homer was put forward by H. van
Thiel, ZPE 90 (1992) 132 and 115 (1997) 1336.
8
Lexicon Messanense, ed. H. Rabe, RhM 47 (1892) 404413, at 411; cf. R.
Pfeiffer, History of Classical Scholarship (Oxford 1968) 161. My debt to
Pfeiffers work, despite certain criticisms expressed below, should be ob-
vious.
9
P.Amherst II 12. Prof. A. C. Cassio drew my attention to schol. Soph.
Phil. 201, which attests a reading of Herodotus (but not commentary) by
Hellanikos. Is this the Hellenistic grammarian? Or could this be an
anecdote about one historian reading the work of another? The view that

42 SCHOLIASTS AND COMMENTATORS

viously Homer and other poets had been the chief beneficiaries
of scholarly activity, as far as we can tell from the scanty
evidence now available. But perhaps we should be wary of
assuming that the fortunes of literary texts are a safe guide to
the history of these developments. The Neoplatonist scholar
Proclus (410485) records that Crantor (ca. 335275 B.C.) was
the first commentator on Plato (In Ti. 20D, p.76.2 Diehl).
Crantors work is clearly earlier than anything achieved in
Alexandria at the Museum, and doubtless he was more con-
cerned with philosophical than literary matters.
10

The distinctions between the types of Hellenistic scholarly
publications are subtle and difficult to define. Between the
short pamphlet at one extreme and the full commentary at the
other there seems to have been an intermediate class of second-
order literature, the nature of which became clearer when the
Berlin papyrus containing a work by Didymus on Demosthenes
was published (P.Berol.inv. 9780). Didymus deals almost ex-
clusively with historical, not linguistic questions; but not all the
points of historical interest are touched on, and so the question
arises whether the papyrus text as we have it is an abridgement
of a full historical commentary. This simple answer was not
accepted by F. Leo in his review of the first edition;
11
he pro-
posed to recognise a separate class of literature, to which
Asconius commentaries on Ciceros speeches also belong. As-
conius goes to work in the same way as Didymus: he selects for
comment historical points that interest him and does not at-
tempt a full treatment. Neither book is a complete commentary
or hypomnema, to give it its technical name, but rather the
discussion of a number of historical questions, zetemata, forming
part of the class of literature known as !"##$%&&'(' )*$+ (,-
.*/0'. In that case the correct title of Didymus book will be
simply )*$+ 12&,!340,"5. The distinction is explicitly made in
regard to the writings of his predecessors by Didymus himself
___
Aristarchus wrote commentaries on Hesiod has been shown to rest on very
uncertain foundations: M. L. West, Hesiod: Works and Days (Oxford 1978) 65.
10
Though his name has not figured in histories of scholarship, he is given
due credit by K.-H. Stanzel, Der neue Pauly 6 (1999) 805.
11
NAkG (1904) 254261 = Kleine Schriften II 387394.
NIGEL WILSON 43

in the scholium on Iliad 2.111. But that is not the end of this
complex discussion. In a valuable paper on Didymus, Steph-
anie West showed that the omissions in the Berlin papyrus
could be accounted for by assuming that these topics had all
been dealt with at an earlier stage of the commentary; the
papyrus preserves part of Book 18 of a series, and opportunity
would have arisen already to explain many historical matters.
12

So perhaps in this case a subtle distinction between types of
commentary is out of place. However, it is important for our
purpose that Didymus made it when speaking of his predeces-
sors.
Commentary and learned discussion were not the only
products of the Alexandrian scholars. They compiled reference
books for the benefit of readers of the classics. Among these
were lexica to various authors, such as Apions list of rare
words in Homer or Harpocrations lexicon to the ten Attic
orators (which in fact is a slightly later work). Other useful
compilations were the lists of persons mentioned in Old Com-
edy, the 67&8.,9&*0,:, and the lists of persons with identical
names, the ;&<0"&,:. Fragments of both are embedded in the
scholia to Aristophanes (e.g. Ach. 214, 703; Clouds 1022; Birds
749, 822; Peace 347), and fragments of the latter in Diogenes
Laertius Lives of the Philosophers.
By the end of the Hellenistic age there was a mass of aca-
demic literature written mainly by members of the Alexandrian
Museum and to a lesser degree by their rivals at Pergamon.
Potentially it was available to educated readers and school-
masters throughout the Greek-speaking world. We do not
know how quickly it spread, but it is found eventually in the
Egyptian country districts such as the Fayum; the most striking
example is perhaps the private letter written in the second
century A.D. (P.Oxy. XVIII 2192) in which the writer asks for
books of 67&8.,9&*0,: and abstracts of the mythology found
in tragedy, and tells his correspondent where some of these
texts may be obtained.
Very little of this literature survives; most of it was lost in late

12
CQ 20 (1970) 288296, esp. 290291. See also G. Arrighetti, SCO 26
(1977) 1367, at 4967.
44 SCHOLIASTS AND COMMENTATORS

antiquity when economic resources declined to a level which
did not permit the copying of the full range of ancient literature
with sufficient frequency to make good the inevitable losses
caused by war, natural disasters, and ordinary wear and tear.
Apart from the texts already mentioned a tiny quantity of
Hellenistic work has come down to us in its original form: the
astronomer Hipparchus discussion of Aratus poem about the
constellations and Dionysius Thraxs tiny but influential
pamphlet on grammar (if it is genuine) are the best-known
examples, to which one should add the partially preserved trea-
tise by Demetrius Lacon on textual difficulties in Epicurus.
13

The efforts of scholars dating from the Roman Empire have
fared only a little better: we have substantial books on gram-
mar and syntax by Apollonius Dyscolus and the lexicon to the
ten Attic orators by Harpocration, and there are in addition
two unusual cases of double transmission, about which I shall
have something to say below. In order to complete the picture
it should be noted that apart from works of literary scholarship
there are many philosophical commentaries, especially on the
Aristotelian corpus, mainly products of late antiquity.
Scholia help us to make good the enormous losses because
they preserve, usually in a much altered form, but sometimes in
verbatim quotations, fragments of ancient commentaries. Two
interesting cases are the metrical commentary on Aristophanes
and Nicanors guide to the punctuation of the Iliad. Numerous
quotations in the scholia seem to preserve to a large extent the
original wording. But more often than not we have to deal with
paraphrase and with scraps of erudition that are not attributed
to their author.
All these texts enjoyed initially an existence independent of
the literary works they dealt with. The ancient form of book,
being a roll with columns of writing often set close to each
other, did not favour the addition of much explanatory mater-
ial in the margin round the text. But good copies of literary
texts had a special series of signs in the margin which told the
reader about certain kinds of note that he might expect to find

13
P.Hercul. 1012, ed. E. Puglia, Demetrio Lacone: Aporie testuali ed esegetiche in
Epicuro (Naples 1988).
NIGEL WILSON 45

in a commentary. The obvious case is Homer, where the
obelus and other signs soon came to have established meanings
and are found in a number of papyri. Some of the signs were
used, but not always with the same meaning, in texts of tragedy
and comedy, Plato, and Demosthenes.
14

But though continuous commentary could not find a place in
the margins of the ancient book, there was room for the school-
master or reader to add short notes or interlinear glosses, as
they are seen in the famous papyrus of Pindars Paeans (P.Oxy.
V 841).
15
A wide interval could be left between the columns, as
is seen in P.Kln IV 185 of Aratus Phaenomena, part of the same
roll as P.Oxy. XV 1807. There was nothing in principle to
prevent the addition of excerpts from learned monographs or
commentaries, especially as cursive script could be written very
small. One curious exception to the general rule should be
mentioned: Lille papyrus 82, a copy of Callimachus dating
from the third century B.C., adopts a quite different format. A
few lines of poetic text are followed by commentary.
16
This
alternation dispenses with the need for broad margins, and it is
surprising that this practice was not adopted for all commen-
taries that followed a text closely. Perhaps it was an experiment
that failed to find favour with the public.
17

II.
Originality and fresh contributions to scholarship seem to
have become rare by the time of Didymus at the end of the first
century B.C. The only other significant figure from this period
may be Theon, whose work on Hellenistic poets is dated to the
reign of Augustus. There is not enough evidence to allow a

14
For signs applied to poetic texts see R. L. Fowler, ZPE 33 (1979) 17
28, at 2428. For philosophical texts see P.Oxy. XLVII 3326 with M. W.
Haslams note, P.Oxy. LII 3656, and Diog. Laert. 3.65. A general survey is
given by K. McNamee, Sigla and Select Marginalia in Greek Literary Papyri (Brus-
sels 1992).
15
For some valuable observations on Pindar scholia see M. R. Lefkowitz,
First-person Fictions: Pindars Poetic I (Oxford 1991), passim.
16
Ed. P. J. Parsons, ZPE 25 (1977) 150. The reader for GRBS kindly
draws my attention to the somewhat similar format in the Derveni papyrus.
17
See A. Carlini, Maia 32 (1980) 225253, at 235236.
46 SCHOLIASTS AND COMMENTATORS

confident judgement about him; one needs to note that work
on these poets had begun earlier, as is clear not merely from
the Lille papyrus of Callimachus but also from such hints as the
mention of the name Asclepiades of Myrlea in the scholia on
Theocritus and the treatise by Hipparchus on Aratus, who tells
us that he was by no means the first person to expound the
Phaenomena (p.4.12 Manitius). The first two centuries of our
era did, however, produce some notable scholarship. Pride of
place must go to the linguistic treatises of Apollonius Dyscolus
and his son Herodian. One should also note the elaborate sys-
tem of punctuation devised by Nicanor; apart from a general
treatise he discussed in detail punctuation in the Iliad.
18
But
there are two further critical stages in the history of scholarship
in antiquity. The first results from a change of interest in the
literary world: whereas the Hellenistic scholars had concerned
themselves to explain rare words, to discuss matters of his-
torical or antiquarian interest, and to criticise the classical texts
as literature, in the Roman imperial age the new fashion of
writing Attic Greek, or as close an imitation of it as possible,
meant a change of emphasis in the classroom, and it was a
change with consequences that lasted right up to the end of the
Byzantine age. Less time was spent on the appreciation of
literature, more on the mastering of Attic words and idioms
that could be exemplified from the classical authors and were
supposed to be the ingredients of good imitation Attic prose.
Hence a high proportion of extant scholia deal with arid details
of rhetoric or inform us, frequently wrongly, that such and such
a usage is a feature of Attic style, as opposed to the koine or
one of the other ancient dialects. These linguistic notes must
often have displaced other information which would have been
of greater value to us.
The second change in this period is the substitution of the
codex for the roll as the normal type of book, a change that for
classical texts was complete by the end of the fourth century.
The new type of book had the advantage of permitting more

18
Many excerpts are found in MS. Venice Marc.gr. 454 = A, where his
name figures in the subscriptions at the end of the books; see also Scholia in
Dionysii Thracis artem grammaticam, ed. A. Hilgard (Leipzig 1901) 26.428.8.
NIGEL WILSON 47

annotation in the margins, so that readers no longer needed to
suffer the inconvenience of referring to another book for most
of the commentary on the text. It was only a matter of time
before not only the jottings made in the classroom or during
private reading but also the full-scale monographs and com-
mentaries were transcribed in whole or part into the margins of
the new copies. The task of compilation and amalgamation
must have been enormous, since there were sometimes several
commentaries or monographs on the same text. The date at
which it was carried out is not certain and has been the subject
of controversy. Extremely few ancient books have extensive
marginal notes; the best example is a famous and very late
papyrus of Callimachus (P.Oxy. XX 2258); it can hardly be
earlier than the sixth century. Probably the process of transfor-
mation began in late antiquity and after an interval continued
in the ninth and tenth centuries during what is sometimes
called the first Byzantine Renaissance. The new arrangement
of the commentaries on classical texts is in some way related to
the standard form of commentary on the Bible. The history of
exegesis of scripture has some analogy to the history of classical
scholarship. The so-called Antiochene school of exegesis in fact
adopted the methods established by their pagan predecessors in
Alexandria.
19
After a time a considerable mass of commen-
taries on the text of the Septuagint and the New Testament
had been accumulated, and early in the sixth century Pro-
copius of Gaza took the step of combining verbatim extracts
from existing commentaries into a large new single commen-
tary, known as a catena, or in Greek !*:$%. Catenae have a
certain similarity to scholia as we know them from medieval
manuscripts, and it is natural to ask which served as the model
for the other. A hint is given by the fact that in catenae it is
very common to cite at the beginning of each excerpt the name
of the author from whose work it is taken, whereas this happens
much less often in scholia; one might infer that the Biblical
scholars made an innovation because they wished to be precise

19
See the important monograph of C. Schublin, Untersuchungen zu Me-
thode und Herkunft der Antiochenischen Exegese (Theophaneia 23 [Cologne/Bonn
1974]).
48 SCHOLIASTS AND COMMENTATORS

in these matters, especially as the orthodoxy of individual
authors might be questioned.
20

Even the ninth and tenth centuries are not necessarily the
end of the story, because we cannot always edit scholia on a
Greek author from manuscripts of such early date. Instead we
are often compelled to rely on copies written late in the Byz-
antine period, and therefore we need to know about the history
of classical studies at that time. Byzantine schoolmasters were
active for many centuries in much the same way as their coun-
terparts of late antiquity. There was no drastic change in the
curriculum; one should note the disappearance of Menander,
but apart from that the most important fact is a reduction in
the quantity of classical literature that the pupils could be ex-
pected to master. Certainly there was no deviation from the
basic aim of turning out pupils well versed in classical Attic. But
each successive generation might wish to alter the inherited
commentary on the classical authors in accordance with cur-
rent taste and needs. Though Byzantine schoolmasters tended
to be conservative, the inevitable result was that ancient learn-
ing in the scholia was gradually reduced in quantity because it
was above the heads of the pupils, or for that matter of the lit-
erary man reading in private. More notes of a purely linguistic
character were put in to help the reader understand the in-
creasingly remote language of the ancient texts.
The process can be seen in the recensions of scholia prepared
by Demetrius Triclinius in the early fourteenth century: he
eliminates a good deal of antiquarian and other information
found in the old scholia, and it must be admitted that he was
not entirely without justification, since the subject matter, al-
though sometimes interesting to us, was often irrelevant to the
text of the author. At the same time he added translations of
words that were no longer current in the spoken language, e.g.
for 0*=>*: he gave the translation ?:,0=@*: (Ar. Ach. 1141), for
ABBC5 B,"6%0:6,0 (Ar. Eq. 143).
21
The same watering-down of

20
For more about this see my articles in CQ 17 (1967) 244256, GRBS 12
(1971) 557558, together with CR 27 (1977) 271.
21
He is found doing the same in his notes on Theocritus: C. Wendel,
berlieferung und Entstehung der Theokriteischen Scholien (AbhGtt 17.3 [1920]) 35.
NIGEL WILSON 49

scholarship took place in the Homer scholia. The rich learning
of the unique tenth-century Venetian manuscript (Marc.gr. 454)
with its quotations from Didymus, Nicanor, Herodian, and
other ancient sources is a striking contrast to the mediocre
notes of the Geneva codex of the Iliad, which probably dates
from the thirteenth century. The compiler of the latter is gen-
erally wrestling with the meanings of the individual words and
no more; he rarely rises to interpretation of the context, and
still less often does one have the impression that in discussing a
difficulty he is satisfying his own need for enlightenment.
22
And
yet neither Triclinius nor the Geneva codex can be entirely
neglected. Both had access to good sources now lost, which
they reproduce in part: in the scholia to Aristophanes Achar-
nians Triclinius is our sole source for two notes stating that lines
of Aristophanes are parodied from Euripides Telephus (schol.
on Ach. 440 and 1188), and the Geneva codex in Book 21 of
the Iliad suddenly produces excellent notes much superior to
the comments on the first twenty books.
23

Since many scholia found in the margins of medieval copies
derive from ancient monographs, it is natural to ask whether
there are cases of double transmission, in which the original
text of the monograph can be compared with excerpts made
from it. Two examples can be cited. A short work by the other-
wise unknown Heraclitus dealing with allegory in Homer is
transmitted as an integral text in a few relatively late Byzantine
copies (Ambr. B 99 sup.; Vat.gr. 305, 871, 951; New College, Ox-
ford 298). Short passages from it are found in the margins of a
few copies of Homer (e.g. Marc.gr. 453; Leipzig, Univ. 1275;
Vienna, Phil.gr. 133). The other case again involves Homer: the
first book of Porphyrys Homeric Questions exists in one medieval
manuscript (Vat.gr. 305, probably copied in 1314), while ex-
tracts from it can be found in the margins of a number of
Homer manuscripts. In this case the text offered by the two
branches of the tradition is just sufficiently different to have

22
I cite the judgement of H. Erbse, RhM 95 (1952) 170191, at 182183.
23
But Moschopoulos recension of the Theocritus scholia is a simple
commentary for ill-prepared schoolboys of the late thirteenth century and as
such valueless for our purposes; see Wendel, berlieferung 2022.
50 SCHOLIASTS AND COMMENTATORS

induced the modern editor to print the text in two columns
rather than try to record all the variant readings in an ap-
paratus criticus. This gives us a hint of the degree to which
other scholia, for which no parallel text can be found, may
have been altered.
24

III.
From this brief historical sketch it will be apparent that
scholia contain material deriving from sources which vary
enormously in character and antiquity. It is important where
possible to distinguish the various strata and to interpret their
language correctly. Even competent scholars have sometimes
been led into error by failure to do so.
First a few words about terminology in scholia. It is wrong to
assume that technical terms or turns of phrase can be con-
fidently assigned to identifiable individuals. An erroneous belief
of this kind would be for instance the idea that scholia intro-
duced by the word &D),(* perhaps, or by rhetorical ques-
tions, come from the pen of Didymus.
25
A moments reflection
will convince us that if we could attribute the invention of a
critical term to a particular scholar, that term might soon be-
come a standard part of the vocabulary for all successive gen-
erations of scholiasts, so that its use does not tell us a great deal.
In the late Byzantine period Demetrius Triclinius took over
much of the commentary written by his mentor Thomas Ma-
gister. Although Th. Hopfner showed
26
how the terminology of
these late Byzantines can be distinguished, it is only because we
possess autographs of some of these scholars or manuscripts
bearing unmistakable evidence of authorship that we are in a
position to identify their individual contributions to the scholia
with some degree of certainty.
The terminology itself is not always easy to understand. The
relationship between Aristophanes and Eupolis was long mis-

24
A. R. Sodano, Porphyrii Quaestionum Homericarum liber I (Naples 1970).
Antonio Carlini kindly draws my attention to the fact that many scholia on
Platos Alcibiades I are excerpts from Olympiodorus commentary.
25
Cf. J. Richter, WS 33 (1911) 3770, at 6364. The idea is mentioned
also by Wendel, berlieferung 147.
26
SBWien 172.3 (1912).
NIGEL WILSON 51

interpreted, even by an authority as eminent as Wilamowitz,
because of a misreading of the scholium on Knights 1291. It
appeared to state that verses 12881315 were the work of
Eupolis. The words in question are E6 (,- F!(:5 ,G0 (,:,-(,0
H0.$' >'!I (:0*5 JK)LB:.,5 *M0': (N0 !'$OP'!:0, *Q #R >2!:0
JS),B:5 T"0*),I2!' (U >'B'6$U (fr.89). It was left to Max
Pohlenz to insist on the correct translation.
27
The words E6 (,-
do not mean from this point onwards. The sense is: because
of line 1288 some people say that Eupolis wrote the parabasis
(i.e. the whole of it), because he elsewhere says I collaborated
with my bald colleague. The original commentator had pre-
sumably hit on another fact, not now mentioned in the scholia,
that verses 128889 in the second parabasis of the play have at
least a superficial resemblance to the corresponding verses of
the parabasis, the antepirrhema, in Eupolis Demoi, which be-
gan F!(:5 ,G0 H$?*:0 (,:,V(,"5 H0.$'5 (fr.99.33). When that
fact is deleted from the scholium the simple phrase E6 (,- be-
comes ambiguous.
Another trap for the unwary is that when a scholiast explains
a word by giving one or more synonyms for it, we should not
mistake any of these synonyms for the word that originally
stood in the text of the author before him. The warning sounds
so elementary as to be unnecessary, but it is easier to fall into
the trap than might be supposed. A case in point occurs at
Apollonius Rhodius 1.219, where the words A6$,(O(,:!:
),.W0 are glossed by the scholiast (,/5 !>"$,/5 X (,/5 A!($'-
#OB,:5. These are intended as synonyms only, but in H.
Fraenkels edition (1961) A!($'#OB,:!: was adopted in the text
because of its occurrence in the scholia; yet there is not the
slightest ground for thinking that the scholiast read this word in
his copy.
28


27
Hermes 47 (1912) 341347 and NAkG 1952.5 96128, at 120121. In
my review of R. R. Schlunk, The Homeric Scholia and the Iliad (Ann Arbor
1974), in Gnomon 48 (1976) 716717, I pointed to some errors of translation
which were blemishes in the authors praiseworthy attempt to treat an im-
portant theme.
28
Notwithstanding Frnkels defence of his choice in Einleitung zur kriti-
schen Ausgabe der Argonautika des Apollonios (AbhGtt 55 [1964]) 109110.
52 SCHOLIASTS AND COMMENTATORS

Apart from terminology the other main difficulty in dealing
with scholia is to distinguish where possible what I would call
the strata within the scholia. Scholiasts very rarely refer to con-
temporary events in such a way that they can be dated, and
when they do it is not usually possible to tell whether the words
in question are part of the original core of a note or a later
accretion. For example a reference to Belisarius (schol. Thuc.
4.83) only gives us a terminus post quem for the sentence in
question. If the Apollonius scholia (on 3.1241) give the text of
an epigram which appears in the Palatine Anthology (9.688), we
may conjecture that the scholia were being revised after the
compilation of the Anthology, in other words in the tenth cen-
tury; but we do not know the extent of the revision, nor can we
entirely exclude the possibility that the epigram was incor-
porated into the scholia at an earlier date from another source.
Nevertheless, study of the manuscripts enables us to discern
facts which are sometimes obscured in modern editions. This
has important results in the Aristophanes scholia, where mod-
ern commentators have occasionally supposed that a scholium
was written by an ancient commentator when in fact it dates
from the middle ages or even the Italian Renaissance. At Frogs
340 there is a crux which has been variously handled by edi-
tors. As transmitted in the manuscripts the line is metrically
faulty and something must be deleted. The scholiast says that
in some copies of the text the word (:0%!!70 was omitted. This
statement was misinterpreted by W. B. Stanford in his edition
(ed. 2, London 1963). He deleted the word, which may well be
the correct policy, but he cited in support of his decision the
scholium just mentioned as follows: sch. notes that (:0%!!70
was absent in some ancient versions. Note the word ancient;
it is pure fabrication; the scholiast only says Y0 (:!:0 (sc. A0(:-
#$%>,:5), and investigation of the manuscripts reveals that the
word was first deleted by Triclinius, and the scholiast is none
other than Marcus Musurus, the editor of the first printed text
of 1498, noting the reading of some Triclinian copies in his
possession. A millennium and a half is a wide margin of error
in dating scholia. A similar case arose at Acharnians 1, where a
NIGEL WILSON 53

metrical note was taken by K. von Holzinger,
29
and apparently
also by K. Zacher and W. G. Rutherford,
30
to mean that the
ancient critic Heliodorus, perhaps writing in the first century,
read in his text stage-directions, which of course had to be
specifically excluded in his analysis of the metres used by the
poet. The words in question are ET':$*=!37!'0 6'+ 'Z )'$*):-
#$'>'=. They do not appear in any manuscript, and J. W.
White omitted them from his edition.
31
Now that we have a re-
liable witness for the Triclinian recension of the text (Holkham
gr. 88) we can see that Triclinius wrote a note mistakenly treat-
ing two short stage-directions as if they were lines of text and
therefore needed to be taken into account in his metrical
analysis. This grave error was corrected by Musurus; no earlier
scholiast had anything to do with the matter, and Heliodorus
name should not have been introduced into the discussion.
Now for a mistake with more serious consequences, leading
to a dubious reconstruction of one aspect of the history of
medicine. The point at issue is the origin of free medical ser-
vices provided by doctors acting as employees of city councils
and other authorities. A modern monograph on the history of
hospitals claims, following the view put forward in an article
some years ago, that such free treatment is a new development
in the fifth century of our era.
32
The evidence cited is a scho-
lium on Acharnians 1030. The article said that the Aristophanes
scholia date from after A.D. 400, citing as authority a valuable if
somewhat polemical monograph by L. Cohn-Haft.
33
In fact

29
Ueber die Parepigraphae zu Aristophanes (diss.Vienna 1883).
30
See respectively Bursian 71 (1892) 102105, and A Chapter in the History of
Annotation (London 1905) 105.
31
The Verse of Greek Comedy (London 1912) 397.
32
T. S. Miller, The Birth of the Hospital in the Byzantine Empire
2
(Baltimore
1997) 47, following D. W. Amundsen, Bulletin of the History of Medicine 45
(1971) 553569, at 556 n.15. What Miller says is: Of greater interest [than
a statement in Chrysostom] regarding the chief physicians in the early
Byzantine period is a scholion to Aristophanes Acharnenses written in fifth-
century Constantinople and reflecting contemporary conditions. The
scholiast defines the public physicians (i.e., archiatroi) as those who offered
their services for free.
33
The Public Physicians of Ancient Greece (Northampton 1956) 34 n.9.
54 SCHOLIASTS AND COMMENTATORS

Cohn-Haft qualified his statement with the word probably
and was wisely sceptical of the value of the scholiasts note for
his purpose, which was to consider free medical treatment in
classical Greece. The result of careless reading of a scholium is
a misleading assertion about the history of medical care.
IV.
After these cautionary tales I come to the merits and
demerits of scholia. First the demerits. One must resist the
temptation to devote too much space to a series of easy jokes at
the expense of scholiasts, but a few examples should be given.
1. Plain ignorance is rampant. (a) At Aristophanes Knights 42
a scholiast in the Ravenna manuscript tells us that the Pnyx
was an Athenian law-court. From scholia in the other manu-
scripts we learn that the Pnyx was a place where the Athenian
assembly was sometimes held. The Suda, citing the same
source, omits, by luck or superior knowledge, the adverb
sometimes. (b) In the same play at 55 the scholiast has to
comment on the Athenian success at Pylos, but he confuses it
with Thermopylae, normally known as Pylae, and talks about
the Spartan king Leonidas. (c) Linguistic ignorance may mas-
querade in another form, by the assertion that something is
superfluous. At Knights 37 the scholiasts are unable to un-
derstand the force of the preposition in the compound verb
)'$':(2!<&*3'. So they say not merely that it is superfluous,
which is a common way of evading the issue, but that Attic
authors add or omit prepositions in this arbitrary fashion.
34

2. Other manias possess the scholiastic mind. (a) One is for
allegorical interpretation, especially of Homer. This is not of
course an aberration of the medieval mind; it goes back to
Theagenes of Rhegion in the late sixth century B.C., was taken
up by the Stoics and then by the Christians, and is most ob-
viously visible in the treatise by Heraclitus mentioned above,
which begins with the observation that allegory is to be found
in Alcaeus and Anacreon. A typical example from chapter 8: in
the first book of the Iliad the plague is said to be caused by
Apollo, but Apollo is synonymous with the sun, and therefore

34
For other cases see e.g. at Acharnians 610 and 835.
NIGEL WILSON 55

what Homer means is not that a god inflicted plague but that
the heat of summer corrupted the air. Allegorical explanation
and rationalism combine to destroy any feeling for saga and
poetry. A closely related phenomenon is the anxiety of schol-
iasts to extract an edifying moral from a text that does not
justify it. This is not a common feature in the Aristophanes
scholia, although there is an example at Acharnians 366; there
Dicaeopolis refers to himself as a very small, insignificant per-
son, the epithet being humorous and nothing more, whereas
the scholiast assures us that the poet is here teaching us to
judge men not by their size but by the merit of what they say.
But such explanations are a substantial element in one class of
scholia on the Iliad, generally but rather misleadingly known as
the exegetic scholia. One example: commenting on Agamem-
nons anger at 1.103104 the scholiast says that we should
avoid anger carefully since it overcame even such a great hero.
But there is nothing in Homer to justify the note, which is a
misunderstanding of heroic qualities and values. (b) There is a
craze for quoting Homer to illustrate Attic usage. Quotations
from Homer can even be inserted without need into otherwise
sound notes. At Acharnians 142 the scholiasts object is to ex-
plain that >:B'3D0':,5 belongs to a class of compound adjec-
tives which take a proparoxytone accent. He lists some of them,
and when he comes to E?4)7B,5, which is found at Iliad 4.458,
he pauses to cite a phrase from Homer, unnecessary though it
is in the middle of a list. This citation seems to me to be a later
addition to a straightforward philological note. The converse
error is also found: Homer is thought of as an Attic author.
This view is expressed by Aelius Aristides in his Panathenaikos
(1.328 Lenz, 13 p.296 D.). The context is the universal adop-
tion of Athenian culture and language as a sign of respect for
the glories of Attic civilisation. Even Homer can be claimed as
one of them, since he came from Smyrna, an Athenian colony;
its language, and therefore his, is derived from Athens. The
scholia on Aristides here adduce the first line of the Iliad in sup-
port, saying that the form [2B2:%.*7 is Attic, being parallel to
the Attic genitive \*04B*7.
35
Aristides is not the originator of

35
Cited by J. F. Kindstrand, Homer in der zweiten Sophistik (Uppsala 1973)

56 SCHOLIASTS AND COMMENTATORS

this erroneous vision of Homer as an Athenian. It occurs in the
D-scholia (on Iliad 2.371) and in the ancient lives of Homer it is
attributed, probably wrongly, to the eminent authorities Aris-
tarchus and Dionysius Thrax.
36
One need hardly add that
muddled thinking continued in Byzantium, where the gram-
marian George Choeroboscus (Grammatici graeci IV.2 86) crit-
icises no less a person than Aristarchus for failing to realise that
Homer wrote an early form of the Attic dialect. Another ec-
centric trait is the desire to assure us that the Athenians were
the autochthonous inhabitants of their country. While this is
not an irrelevant fact at Sophocles OC 947, it is inapposite as
part of the explanation of the words .9!6,B,0 #*$]0(:,0 in
Aristophanes Knights 42. Equally eccentric is the explanation of
the difficult word E)=67),5 meaning up to the hilt at Achar-
nians 230, where the scholiast is led astray by the Athenians
nautical achievements and can only think of a derivation from
the word for an oar.
3. A third failing of scholia from the classical scholars point
of view is that they contain much which only informs us of
practices in the classrooms of late antiquity or the middle ages.
(a) In the first place great attention is paid to rhetorical effects.
Admiration for rhetorical skill in dramatic or other texts is not
in itself out of place, and there are some unobjectionable notes
of this type in the commentary on Sophocles OC 939, 1257,
1760. But as a rule the scholiasts first concern was classifica-
tion into the categories of rhetorical figure, metonymy, synec-
doche, and the rest, that the pupil had to master, rather than
any desire to point out literary merit. These categories were
listed and discussed exhaustively by W. G. Rutherford in A
Chapter in the History of Annotation. (b) Other scholia show us a
procedure of the classroom in operation. The master picks out
a word in the text, asks what it refers to, and has the model
answer in the commentary in front of him. An example from
Soph. OC 25: the scholium on the word 2S.' runs (I 2S.'; F(:
'^('I *_!:0 'Z `3a0':b E&)]$70 .c A0(+ (,- ;.,:)]$70, .:d (d
?. The pupils attention and comprehension are tested in ele-
___
197. He wonders if the assertion is to be taken seriously.
36
See T. W. Allens OCT, V 244.13.
NIGEL WILSON 57

mentary fashion. Other examples are found at 156, 354,
Aeschylus Choephori 661, Euripides Alcestis 999. (c) There is also
a natural preoccupation with mastering Attic usage so as to
earn admiration for the excellence of ones archaising prose
style. A vast number of notes tell us that such and such a word
or usage is Attic. In the commentary on the Acharnians there is a
series of notes which make clear the motive for studying Attic
usage: one will acquire the reputation of being an educated
person (see especially on 207, also 210, 245, 272).
Most of the shortcomings I have enumerated date from late
antiquity or the middle ages. The early and valuable Hellenistic
commentaries have been encrusted with, or more often sup-
planted by, later material. A good example of the shift in
interest is afforded by a comparison of Didymus notes on
Demosthenes and the scholia of the medieval manuscripts;
none of the extant Didymus survived long enough to be incor-
porated into the corpus of scholia that we read.
37
An unusual
exception to this generalisation is provided by a papyrus of
Aristophanes with scholia that correspond almost word for
word with the medieval notes (P.Berol.inv. 13929 + 21105 on
Knights 552 and 580).
38
More often we may expect to find much
the same subject matter and interpretation expressed in some-
what different language, as in the monograph on Iliad 21
(P.Oxy. II 221), perhaps composed by a grammarian with the
common name of Ammonius, which has a reasonably close
relationship with the scholia in the Geneva codex.
39
In this case
we find a papyrus agreeing to some extent with a strand of the
medieval tradition that cannot otherwise be traced earlier than
the thirteenth century.
40


37
The fact was noted by H. Diels, BKT I (1904) lilii.
38
See G. Zuntz, Byzantion 13 (1938) 635690, at 635637; H. Maehler,
Hermes 96 (1968) 287293. Cf. C. A. Gibson, Interpreting a Classic: Demosthenes
and His Ancient Commentators (Berkeley 2002), esp. 6269, and P. Harding,
Didymus on Demosthenes (Oxford 2006), esp. 1320 and 41.
39
H. Erbse, Scholia graeca in Homeri Iliadem I (Berlin 1969) xlii.
40
For similar phenomena in the text of several authors see CR 19 (1969)
234.
58 SCHOLIASTS AND COMMENTATORS

V.
After this survey of the failings of scholiasts let me turn to the
achievements of the scholars who preceded them. First of all,
Hellenistic commentators assembled a vast mass of purely
factual elucidatory material about the ancient texts, just as
modern scholars attempt to do. Unfortunately neither ancient
nor modern commentators have invariably followed Corinnas
precept that one should sow with the hand and not with the
sack. And a good deal of the information that the ancients put
together is of limited interest, for instance the accumulations of
local antiquarian and historical lore that accompany parts of
the text of Apollonius Rhodius. But from time to time an
ancient scholar hit on a fact of importance to us which luckily
none of his successors saw fit to suppress. A famous instance is
the note on Aristophanes Knights 400, which tells us about
Cratinus play Pytine. Another intriguing note, more difficult to
interpret, occurs at Clouds 889: there the entry of the two Logoi
is announced with the statement that they are brought on stage
in wicker cages ready to fight like birds. This detail of pro-
duction is not supported by any hint in the text, but as it is hard
to see why it should have been invented, it may be an inference
from something said in the first edition of the play.
41
It has to
be borne in mind that ancient scholars were not able to see
revivals of Old Comedy, and so their comments on production,
even when excellent, need to be attributed to intuition, not
erudition, if they cannot be derived from the text itself. At Thes-
mophoriazusae 101 the scholium tells us that Agathon sings the
whole of what appears to be a song in two parts. Although Hall
and Geldart (OCT) divided it between Agathon and the
chorus, if one thinks about the humorous possibilities of mak-
ing Agathon sing the whole passage, there can be little doubt
that the scholium is right. In the same play at 295 there is a
note about the use of prose in comedy, to which a note at
Knights 941 provides a supplement; but they are both very brief
and one would have liked to see a fuller original version.
Scholia with valuable information could be listed at length.
An ingenious recent demonstration of their value in matters of

41
See Sir Kenneth Dovers edition (Oxford 1968) xc.
NIGEL WILSON 59

textual criticism exploited a note on the Odyssey to settle a point
in the first line of the prologue of Callimachus Aitia.
42
But the
majority of the best examples come from the commentaries on
the Iliad. A note on 2.665 recognises that what appears to be
the definite article in Homer is a demonstrative pronoun. At
1.277 Aristarchus observes that the trisyllabic form E34B7 is
regular in Homer. From the note at 2.48 one can tell that he
worked out the chronology of the events narrated in the Iliad, a
task repeated in modern times by J. N. L. Myres.
43
Nicanors
book on punctuation showed some awareness of the rules for-
bidding heavy punctuation at certain points in the hexameter
verse (see e.g. scholia on 12.49 and 434, 15.360). Another good
observation is that the first book of the poem has no similes (D
on 1.611, AbT on 2.87). And one may add that scholia are one
of our main sources about difficult points of Greek accentu-
ation, since they contain many fragments of Herodian and
Apollonius Dyscolus.
44

For readers brought up on modern Homeric scholarship the
question that will seem most important is whether the ancients
showed any awareness of the oral tradition that lies behind
Homer. The answer is that the scholia on the Homeric text do
not make the point explicitly, but the scholia on the grammar
of Dionysius Thrax contain the notion that for a time the
written texts were lost, so that the poems were preserved by
memory, until Peisistratus had fresh copies made.
45
F. A. Wolf
was aware of this unduly simplified account of the process of
transmission, and quoted it alongside the better-known passage
of Josephus Contra Apionem 1.2, in his epoch-making study of the
genesis of the Homeric epics.
46
More important, however, are
the scholia which enable us to see how far the ancient scholars
were able to develop the principles of literary criticism and

42
F. Pontani, ZPE 128 (1999) 5759.
43
JHS 53 (1933) 115117.
44
See e.g. the remarks on enclitics by W. S. Barrett, Euripides Hippolytos
(Oxford 1964) 424427.
45
Scholia in Dionysii Thracis Artem grammaticam 179.1113, 481.1720 Hil-
gard.
46
Prolegomena ad Homerum (Halle 1795) lxxviii.
60 SCHOLIASTS AND COMMENTATORS

scholarship; scholia are our best source for the growth of philo-
logical method, and a significant proportion of the most helpful
are found only in the Venetian codex of the Iliad, without
which it would scarcely be possible to trace the history of the
subject. From them one can see how the Alexandrians collated
various copies of Homer and other authors that came into their
hands and noted the variant readings. It is possible that they
occasionally had to deal with copies written in the old Attic
alphabet, removing the ambiguities caused by the failure to
distinguish certain vowels (cf. e.g. Aristarchus on Pindar Nem.
1.24).
47
Since the text of Homer was more subject to variation
than others the scholars are more easily seen at work here. It is
well known how they recorded the readings found in copies
from different Greek cities such as Sinope and Massilia.
48

By the middle of the second century B.C. a standard or
vulgate text had been established, which circulated among the
reading public and was not much affected by subsequent
emendation. But one great question remains unclear: did
Zenodotus and his successors normally make proposals for the
improvement of the text when they found a better reading in
one or more copies? Their procedures remain obscure because
reports of what they did are so brief; one does not get a full
account of how they evaluated the variants they found or learn
whether they proposed readings which had no basis in any
known copy.
49

Exegesis of the text led to the formulation of principles that
are fundamental to modern scholarship. One procedure, the
deletion of unwanted lines in Homer, was admittedly used
much too frequently, and for reasons which may at first pro-
voke laughter. But in defence of the Alexandrians it should be
said that they were wielding a new critical weapon, and they

47
For further details see my remarks in H.-G. Nesselrath (ed.), Einleitung
in die griechische Philologie (Stuttgart/Leipzig 1997) 9293.
48
See e.g. schol. on Il. 1.298. The early papyri with striking textual var-
iants, sometimes known as wild papyri, were studied by S. R. West, The
Ptolemaic Papyri of Homer (Cologne/Opladen 1967).
49
Discussion of Zenodotus activity has been revived by the new hypoth-
esis of van Thiel (n.7 above).
NIGEL WILSON 61

were probably encouraged to declare lines to be spurious by
the discovery that the wild papyri had certain omissions. They
also failed to realise the oral nature of Homeric poetry and its
consequence for the critic: repetition and formulae are needed
both by poet and audience. What is harder to explain is the
Alexandrian zest for objecting to lines thought to contain some
indecent or unworthy idea, especially as the Greeks were never
a prudish race. How can intelligent men have supposed it their
duty to identify as inauthentic lines such as Iliad 1.2931 with
their mild sexual reference? Or 3.423426, where the goddess
Aphrodite carries a chair for the mortal Helen, and Zenodotus
is thought to have rewritten the passage? In these cases the
critics used the term A)$4!*:'. It is not confined to Homeric
commentaries: it occurs in a note on Apollonius Rhodius
1.1207, where Hylas looks for a well and carries with him a
bucket, on which the scholiast observes that it is unseemly for a
young man to be carrying a bucket; he contrasts the correct-
ness of Homer, who made a young woman carry a bucket (Od.
7.20), but does not suggest that any part of the text should be
obelised. Where did the concept of A)$4)*:' as a tool of
literary criticism come from? The answer seems to be from Ari-
stotle, who twice expresses the germ of the idea in the Rhetoric
(1395a5, 1406a13), and is credited with having used the term
elsewhere (fr.100 Rose).
50
It is also found in a scholium on Iliad
7.390 attributed to Democritus (68 B 23 D.-K.), but I should
not care to place much reliance on this ascription. Surprisingly
Pfeiffer in his history of classical philology scarcely mentions
the term at all and does not consider the likelihood of Aristo-
telian influence, which, if correctly postulated, runs counter to
his general interpretation of the history of scholarship. I at-
tempted to deal with other signs of Aristotles influence in my
review of Pfeiffer.
51
Perhaps it is worth adding that the notion
of Aristotle as the inventor of literary scholarship receives some
support from the opinion expressed by an educated man writ-

50
Cf. M. Pohlenz, NAkG (1933) 5292, at 66.
51
CR 19 (1969) 366372. More recently R. Meijering, Literary and Rhe-
torical Theories in Greek Scholia (Groningen 1987), has made a case for seeing
Aristotelian influence in Hellenistic scholarship.
62 SCHOLIASTS AND COMMENTATORS

ing at the turn of the first and second centuries A.D. In his short
essay on Homer Dio Chrysostom says: Aristotle, with whom
they say criticism and the art of grammar began, discusses the
poet in many of his dialogues, for the most part admiring and
paying tribute to him, and so does Heraclides Ponticus (36.1).
It is only fair to add that some critics made deletions in the
texts for other and better reasons. Galen thought a passage in
Hippocrates had been altered in order to make it clearer
(XVII.2 110 K. on Hipp. Epid. 6.3 (a5 )'B':C5 #$'>a5 ,S!25
('V(25 E)+ (d !'>R!(*$,0 'K(N0 &*('(*3*=6'!: ),BB,+ (W0
ET2#2(W0), and there is a similar note on Sophocles Ajax 841,
which to modern scholars seems fully justified (('-(' 0*0,-
3*-!3'I >'!:0 e),PB23R0(' )$d5 !'>f0*:'0 (W0 B*#,&R070).
The allegation that actors or producers altered the text of
tragedy occurs from time to time, most notoriously perhaps at
Euripides Orestes 136668.
52

From athetesis and the grounds commonly adduced to justify
it I return to other principles of scholarship developed at Alex-
andria and implicit in much modern work. The chief of these is
the well-known maxim that each author is the best guide to his
own usage, or in Greek gh&2$,0 ET ih&D$," !'>20=@*:0. Of
course there are many breaches of this rule in the scholia; as I
have said, one of the habits of scholiasts is to cite Homer in
support of usage and vocabulary in other authors. But in the
scholia to the Iliad and occasionally elsewhere the principle is
upheld. Aristarchus used to be credited with its formulation,
but Pfeiffer noted that it is not stated in so many words by any
authority earlier than Porphyry in the third century A.D. and
that the verb !'>20=@*:0 is not used in scholia to mean in-
terpret.
53
However, as Pfeiffer himself did not question, the
maxim describes the character of Aristarchus work accurately
and I found one piece of evidence to support the traditional
view, an anecdote in Aelian (VH 14.13). The poet Agathon

52
Cf. e.g. M. L. Wests edition (Warminster 1987) ad loc. and pp.4041;
also D. J. Mastronardes edition of Euripides Phoenissae (Cambridge 1994)
3949.
53
History 226227. But there is a case in schol. h to Il. 1.279, which prob-
ably does not belong to an early stratum.
NIGEL WILSON 63

responds to a friends criticism that his poetry is too full of
antitheses, which ought to be removed. Agathons reply was
that by deleting the antitheses he would remove the genuine
Agathon from Agathons work, and the Greek text is BRB23'5
!*'"(d0 (d0 jk#O370' E6 (,- jk#O370,5 A>'0I@70. The
linguistic analogy with the famous maxim seems too close to be
disregarded, and its occurrence in Aelian shows that the maxim
had been formulated before Porphyrys day.
54

Rigid application of the maxim entails that any unique word
or expression in the text of an author is open to suspicion be-
cause there is no parallel passage to support it. But an unknown
scholar of great intelligence devised a complementary principle
to protect the classics from misguided emendation on a large
scale: the rule was laid down that there are many unique words
in Homer (schol. on Il. 3.54). This fact is not to be found in
Pfeiffers account, because his history concentrates on the
achievements of individuals, at the expense of neglecting what-
ever cannot be assigned to an identifiable scholar.
As a pendant to this section I draw attention to an interesting
anticipation of an important principle in the textual criticism of
Greek drama. J. C. B. Lowe demonstrated that the assignment
of lines to speakers in Aristophanes must depend on the con-
text, and not on the supposed evidence of the manuscript
tradition. A scholiast on Sophocles Ajax 354, whose date and
identity one would very much like to know, observes in the
same vein: when there is doubt about the identity of the
speaker one must pay attention to the characterisation and
distinguish the person accordingly (E0 ('/5 #l$ A&>:P,BI':5
(W0 )$,!m)70 .*/ (,- n3,"5 !(,?O@*!3': 6'+ .:'!(RBB*:0 (d
)$L!7),0).
55

VI.
From philological method I turn to literary appreciation
among the scholiasts. Dionysius Thrax in his grammar had
defined the task of the grammarian so as to exclude textual

54
On this see CR 21 (1971) 172; A. G. Lee, PCPhS 21 (1975) 6364, and
my reply at PCPhS 22 (1976) 123.
55
BICS 9 (1962) 2742.
64 SCHOLIASTS AND COMMENTATORS

criticism; as Pfeiffer observed,
56
the term .:]$37!:5 is absent
from his little book and it is not to be found in the copious
scholia either. But he defined the ultimate and noblest task of
the grammarian as criticism of poems (6$=!:5 ),:2&%(70).
Dionysius himself does not enlarge on the topic, and so we are
deprived of a Hellenistic statement of the nature and purpose
of criticism; but it is worth pausing to see how the term crit-
icism was interpreted by commentators on his laconic and
difficult pamphlet. The commentators are of course very prob-
ably post-Hellenistic, but I think it is reasonable to suppose that
they or their material are pre-Byzantine. One tells us that Dio-
nysius meant by criticism the appreciation of fine language and
apposite arrangement of material, combined with the ability to
state the reasons for ones judgement (p.15 Hilgard). That is a
good practical statement of an ideal which, if observed, would
have improved many ancient and modern commentaries on
the classics.
But not everyone in antiquity seems to have interpreted Dio-
nysius in this way: another scholiast took him to imply that the
purpose of the grammarian is not to judge the quality of the
poets work but whether it is genuine or spurious (471472
Hilgard, and cf. 303304). By means of analysis of language,
subject-matter and structure we can make a useful comparison
with the undeniably genuine work of the poet. He adds, quite
rightly, that there are a great many texts of doubtful authen-
ticity. This accords well with the famous statement by Galen
(XV 105 K.) about the large quantity of pseudonymous liter-
ature that came into the library at Alexandria. The outlook of
this second scholiast is much less elevated and is not, so far as I
know, taken into account by modern scholarship; but I cannot
suppress the suspicion, which will perhaps be regarded as
foolish or heretical, that our second scholiast has correctly
understood Dionysius obscure text. Questions of authenticity
were never far from the thoughts of the bibliographers who di-
rected the library at Alexandria, and the dry linguistic text of
Dionysius is perhaps more easily reconciled with this interpre-
tation. The same view is clearly expressed by Sextus Empiricus

56
History 269 n.2.
NIGEL WILSON 65

in a discussion of grammar and grammarians (Math. 1.93)
which takes as its starting point the definition given by Dio-
nysius. A letter of Cicero written early in 45 B.C. (Fam. 9.10) is
relevant here. He records an argument between the scholar
Nicias and an otherwise unknown Vidius as to the authenticity
of a document supposedly recording a debt. Ciceros authority
was invoked: ego tamquam criticus antiquus iudicaturus sum utrum sint
(,- ),:2(,- an )'$*&P*PB2&40,:. It may be significant that he
should describe himself as a critic when called upon to
arbitrate; the word antiquus and the fact that Nicias is dubbed
Aristarchus prove that he had the old Hellenistic ideals in
mind.
57

It is worth recording that the view of the second scholiast was
adopted in a Byzantine version of Dionysius treatise, some-
times ascribed to Theodosius of Alexandria and found in
several manuscripts, of which the oldest, Monacensis gr. 310, is of
the ninth or tenth century. Examples of the pseudonymous
literature which the expert is supposed to be able to unmask
are the Acts of Thomas and the Apocalypse of Paul in Christian
literature and Hesiods Aspis and Nicanders Theriaca among
classical texts.
58
Even if incorrect, this interpretation of Dio-
nysius was influential.
After voicing these suspicions of the high ideal normally at-
tributed to Dionysius I turn to consider what traces of it exist in
the scholia. One finds a certain limited range of critical tools.
(a) First the concept of poetic licence. This phrase is not men-
tioned by Pfeiffer; yet three different nouns are used to express
it, all in conjunction with the adjective poetic. First is H.*:',
used for instance in commenting on Sophocles patriotic em-
bellishment of myth in an ode of the Oedipus Coloneus (712). The
noun is not common; used in this sense it occurs only once in
the Iliad scholia, on the very first line, where the commentator
felt obliged to apologise for the poet addressing an imperative
to the Muse, instead of the more urbane optative. Secondly

57
The passage is cited by Suetonius De grammaticis et rhetoribus 14 ; cf. R.
A. Kasters ed. (Oxford 1995) 174175, who notes Ciceros use of Aristar-
chus name at Att. 1.14.3 and Pis. 73.
58
See Hilgard, Scholia 568.1531.
66 SCHOLIASTS AND COMMENTATORS

there is A$4!6*:', again used only once in connection with the
Iliad, at 2.45. This usage appears to be lexicis addendum (it is not
in the DGE). Finally there is ET,"!=', which is fairly common
in the scholia and known to be a Hellenistic term from its oc-
currence in Strabo 1.2.17, which refers to Polybius discussion
of Homer. (b) In assessing similes the critics found it easier to
express their grounds for a negative judgement than for praise.
Two good similes in Apollonius Rhodius (1.1201 and 1243)
were pronounced simply to be entirely sound and powerful,
and the verdict on a third is very similar (1.1265). Negative
comment might take the form of the note on 3.1377, that the
comparison of Jason with a shooting star fails to carry con-
viction (A)=3'0,0), but occasionally led to profitable debate. At
1.879 the first scholiast (schol. L) criticises the simile of the
Lemnian women hovering around the departing Argonauts like
bees hovering over flowers; he maintains that it is unsound and
inappropriate (,K? e#:D5, A0%$&,!(') to compare a flourishing
meadow with a city downcast and depressed at the heroes
departure. The second scholiast (schol. P) takes up the criticism
by explaining that part of the simile is designed to clarify
(!'>D0*:' (,- #*0,&40,"), the rest to embellish the narrative
(6%BB,5 &]0,0 6'+ Y6>$'!:5). By this comment he attempts to
understand why the poet wrote as he did, which is an impor-
tant part of criticism. Curiously enough this second note comes
from the less valuable recension of the scholia, probably put
together by a Byzantine scholar somewhere between 1300 and
1475,
59
and so we must be more than usually aware of the
possibility that what we read is Byzantine and not ancient
criticism; but there is no doubt that our second scholiast was
mainly dependent elsewhere on the old scholia. (c) Occa-
sionally the application of a theory of criticism can be seen. It
was probably the Peripatetics who used the concept of mimesis,
defining the task of the writer of comedy as imitation of life.
60

The principle may also have been applied at times in discussion
of tragedy. A hint of this comes from the third-century Peripa-

59
C. Wendel, Die berlieferung der Scholien zu Apollonios von Rhodes (AbhGtt
III.1 [1932]) 117.
60
Pfeiffer, History 190191.
NIGEL WILSON 67

tetic Hieronymus of Rhodes, who compared a story told about
the inhabitants of the Attic deme Anagyrus with the events of
Euripides Phoenix, and did so in a book entitled On the Tragic
Poets (fr.32 Wehrli). Aristotelian influence was hardly to be
avoided in the realm of literary theory. A famous instance can
be seen in the Homer scholia that describe the parting of
Hector and Andromache (schol. T on 6.467, 474). The poet is
praised for describing a scene from real life with such clarity
that it is conjured up before the readers eye and is perfect in its
imitation; and a note is added that in so doing he does not in
the least detract from the stately tone appropriate to epic. That
is at least the right way to begin an appreciation and perhaps
the original version of this note attempted to analyse in more
detail how the poet achieved his effect. It is interesting to ob-
serve that the Victorian commentator Leaf did not deign to
comment on the merits of the passage, whereas Geoffrey Kirk
in 1990 did far better and acknowledged the merit of the
scholia.
61

Platos misguided attack on Iliad 14 (Resp. 390C) was rebutted
with the aid of critical theory.
62
The story of Zeus and Hera
making love had been explained away allegorically (ps.-Hera-
clitus 39), but another critic invoked the theory usually associ-
ated with Asclepiades of Myrlea that poetic narrative can be
divided into three categories, the true, the false, and the fic-
titious (schol. T on Iliad 14.342).
63
This does not entirely de-
stroy the force of Platos complaint, but it is an attempt to ap-
ply critical theory. Finally a very sound criticism that moderns
and ancients alike should have heeded more: the T scholia on
Iliad 21.269 comment on the strangeness of the battle between
Achilles and the river, and say that it is acceptable in poetry
because the poet here tells his story skilfully so that the reader

61
G. S. Kirk, The Iliad: a Commentary II (Cambridge 1990) 223 on lines
466470.
62
Platos admirers in late antiquity studied his view of poetry with care;
see A. D. R. Sheppard, Studies on the 5th and 6th Essays of Proclus Commentary on
the Republic (Gttingen 1980).
63
Cf. the related sources cited by W. Kroll, Studien zum Verstndnis der
rmischen Literatur (Stuttgart 1924) 60.
68 SCHOLIASTS AND COMMENTATORS

does not have time to stop and think. That is important for the
criticism of oral literature and not entirely irrelevant to the
written word.
64

VII.
The contrast between ancient and modern scholarship was
described by the most eminent of all modern scholars as
follows.
65
Ancient scholarship had no sooner become an aca-
demic pursuit than it sank to the level of textual criticism and
verbal analysis, and the latter deteriorated into establishing the
bare sense of the text. A theory of the nature of poetry, under-
standing of the poet and his work in their historical context,
and even the interpretation of each single work as a whole,
either did not exist or were not attempted after the time of
Aristophanes of Byzantium. No attempt was made to write a
history of Greek tragedy
66
or a critique of tragic drama. The
ideal of modern scholarship on the other hand should be to
assemble for the reader all the information which will enable
him or her to approach the work in question with as nearly as
possible the same knowledge, presuppositions, and attitudes as
the original reader or audience had; so a commentary on a
Greek play should try to put us in the position of the Athenian
as he entered the theatre to watch the performance, and to give
us the same enjoyment as he had.
Wilamowitz made his contrast unduly sharp and not entirely
accurate. He tried to compare a modern ideal that generally
remains unrealised with ancient practice that can only be eval-
uated through the distorting mirror of the scholiastic tradition.

64
See further N. J. Richardson, CQ 30 (1980) 265287, for a survey of
the literary criticism to be found in the Homeric scholia.
65
Wilamowitz, Einleitung in die griechische Tragdie (Berlin 1921) 219220,
256.
66
Photius Bibl. codex 161, 103b1132, refers to a certain Rufus, author
of a 1$'&'(:6N Z!(,$=', of which Book 8 was clearly nothing more than
anecdotes; whether the same was true of Books 13 is less certain (the con-
tent of Books 47 is not relevant for our purpose). Equally uncertain is the
character of the work on the theatre mentioned without an authors name
in schol. Dem. 19.247 (M. R. Dilts, Scholia Demosthenica [Leipzig 1983] 466).
These texts are suggestive but do not amount to a refutation of Wilamo-
witzs judgement.
NIGEL WILSON 69

If we concentrate our attention on the fragments that can be
identified with some plausibility as belonging to the commen-
taries of the best ancient scholars, we shall be able to defend
them against some of Wilamowitzs charges. Certainly scholar-
ship did not come to an end with Aristophanes; the first writer
of commentaries on a large scale was his successor Aristarchus.
It is equally clear that the commentators passed judgements on
poetry which were based on a theory of literary criticism. And
they did do something to help the reader enter the state of
mind of the fifth-century Athenian, when they noted for in-
stance the exploitation of patriotic feeling by the tragedians.
67

Appreciation of a tragedy as a whole was not entirely lacking in
the man who noted the excellent structure (,_6,0,&=') of Soph-
oclean plays.
68
Careful scrutiny of scholia, assisted by modern
editions much superior to earlier ones, enables us to build up at
least a partial picture of what ancient scholars achieved. It is of
course easy to point to defects and shortcomings. But clearly
they also had many merits and were much better than most of
the scholiasts who succeeded them; it is a reflection on us if we
confuse them with the inferior pedants of a less productive
civilisation.
APPENDIX
The authenticity of Dionysius Thrax Ars grammatica has been a
topic for debate for several decades. The scepticism expressed by V.
Di Benedetto, who thought it a forgery of the fourth century or
thereabouts, has elicited a number of responses. A balanced view of
the current status quaestionis is provided by F. Montanari in his article
in Der neue Pauly (3 [1997] 632635). The purpose of the present note
is to suggest that the debate might have been unnecessary, or at the
very least should have been conducted in different terms, if account
had been taken of a remark by Galen.
The last sentence of the pamphlet in which he lists and discusses
his own writings mentions a now-lost work consisting of a single book
with the title Whether someone can be critic and grammarian: *_

67
Schol. on Soph. OC 457 and 712.
68
Schol. on Soph. OC 28, and cf. parallels cited by V. De Marco, Scholia
in Sophoclis Oedipum Coloneum (Rome 1952), from scholia on three other plays.
70 SCHOLIASTS AND COMMENTATORS

.90'('= (:5 *M0': 6$:(:6d5 6'+ #$'&&'(:6]5.
69
It seems not to have
been observed that what induced Galen to discuss this question is
presumably Dionysius famous definition of the grammarians high-
est task as 6$=!:5 ),:2&%(70. For him the grammarian should aspire to
be a critic. Since attacks on grammarians as pedants were common
and often justified, one can imagine that Galen wished to consider
whether denigration had gone too far. If it is right to set Galens
pamphlet in this context, Dionysius work is either genuine or a
much earlier forgery than has been supposed.

August, 2006 Lincoln College
Oxford OX1 3DR, U.K.
nigel.wilson@lincoln.ox.ac.uk


69
Libr.propr. 17, ed. I. Mueller (Leipzig 1891) 124.1617.