SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

Authorial Collaboration? 
Aristophanes' Knights and Eupolis 

Keith Sidwell 

I N A RECENT ARTICLE in this journal, 1 Stephen Halliwell used 
evidence from Old Comedy to suggest that "authorial col­
laboration" occurred in the comic theatre of the fifth cen­

tury. The purpose of this paper is to cast doubt on his findings. 
I shall suggest that "authorial collaboration,» along with 
accusations of plagiarism and other evidence linking the comic 
poets to one another, is part of an elaborate series of jokes 
generated by one underlying characteristic: the tendency of 
poets to attack each other by making comedy out of each 
other's comedy ('paracomedy'). This practice will, I argue, 
explain the well-known relationship between Aristophanes' 
Knights and his contemporary Eupolis, with which we shall 
begin. 

1. The Primary Evidence 

The basic evidence is as follows: 

(1) In the revised parabasis of Clouds, Aristophanes accuses 
Eupolis of EKO'tpbvw; 'taus TtIlE'tEPOUS 'I1t1tEas KaKOS KaKros 
(554) in producing his M arikas of 421. The play was an 
attack on H yperbolus, portrayed as its central character the 
slave Marikas.2 

(2) Eupolis (fr. 89 PCG) in the parabasis of Baptai (417?) 
wrote fK(XKEtVOSt 'taus 'I1t1tEw;/ ~uvETcoil1(Ja 'trot cpaAaKprot (­
u/--) Kaoropl1oalll1v. 

(3) A relationship between Aristophanes and Eupolis is 
evident already in 423 in the parabasis of Cratinus' Pytine. 

1 "Authorial Collaboration in the Athenian Comic Theatre," GRBS 30 
(1989: hereafter 'Halliwell') 515-28. I note that I. C. Storey has recently 
supported the collaboration theory in "Notus est omnibus Eupolis?" in A. H. 
Sommerstein, ed., Tragedy, Comedy and the Polis (Bari 1993) 387f. 

2 This is specifically mentioned in the ancient commentary (frr. 192, 150), 
Quint. 1.10.18 (=fr. 208 PCG), and fro 209. 

365 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

366 ARISTOPHANES' KNIGHTS AND EUPOLIS 

Fr. 213 peG reads: -ruu-ru clK01)O"U<; (the charge of Eq. 531 
that Cratinus is 7tUpUATjPOUV-ru) (, Kpu-rtvoC; EYPU'l'£ -r~v 
llU-rLVTjV, ()ElKVU~ on OUK tATJPTjO"£V· tv ~t KUKro~ AiYEl -rOV 
'AptO"-ro<paVTjV w~ -ra. EU7t6At()O~ A£yoV-rU. 

The date of Clouds II is unknown and the date of Baptai is 
conjectural. 3 So it is not possible to say with any certainty that 
Eupolis is 'replying' to Clouds II, especially since the ancient 
didaskaliai seem not to have recorded a performance of the 
second version at the major festivals. 4 If we postulate that 
Eupolis was replying to a 'reading text' of Clouds II,5 we reduce 
the general intelligibility of the issue as publicly addressed in 
Baptai. It is possible to propose a scale of greater to lesser 
comprehension of detailed reference among the audience of a 
dramatic production. But it does not seem a particularly 
satisfactory way to account for the deliberate artistry of a poet 
who relied on a random cross-section of the Athenian public for 
a vote to win the prize, to assume that he would have wittingly 
risked by-passing any of his potential audience at any point. 6 It 
seems unlikely that we have here two interrelated moments of 
the dispute. I agree with Storey that there is an important role 
for Aristophanes fr. 58 from A nagyros (dated 419-412 by 
Geissler) in Eupolideans from the parabasis: £K 8£ TT)C; £1111<; 

3 See E. C. Kopff, "The Date of Aristophanes, Nubes II," AJP 111 (1990) 
318-29, for the argument that Clouds II should not only be dated after the 
ostracism of l-Iyperbolus (cf S. Bianchetti, "L'ostracismo di Iperbolo e la 
secondo redazione delle Nuvole di Aristofane," StIt 51 [1979] 221--48), but as 
late as 414. 1. C. Storey has certainly damaged the case for a very late date 
("The Dates of Aristophanes Clouds II and Eupolis' Baptai: A Reply to E. C. 
Kopff," AJP 114 [1993] 71-84), but in arguing that the play must have been 
produced before the ostracism of Hyperbolus he avoids the implication of the 
comparison between attacks on Cleon KEql£V(J)l and attacks on Hyperbolus. 
For the date of Baptai see Storey (n.7 infra). Kopff 326 put it as late as 413, 
unnecessarily, to fit in with the later dating of Clouds II. See below. 

4 This is the inference one should draw from I Nub. 522 (=Callim. fr. 454 
Pfeiffer). See K. J. Dover, ed., Aristophanes, Clouds (Oxford 1968) Ixxxff. 

5 The latest proponent of the "reading-text" theory is D. Fowler, "Taplin on 
Cocks," CQ N.S. 39 (1989) 257. On the book-trade see now K. J. Dover, ed., 
Aristophanes, Frogs (Oxford 1993) 34f. 

6 See R. Harriott, "Aristophanes' Audience and the Plays of Euripides," 
BICS 9 (1962) 1-8 for a useful analysis of the level of knowledge required from 
the audience to understand jokes about Euripides. For voting procedures see 
A. Pickard-Cambridge, The Dramatic Festi7Jals of Athens, 3 rev. J. Gould and 
D. M. Lewis (Oxford 1988) 95-99. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

KEITH SIDWELL 367 

XAuvl(5o<; 'tP£l<; imATni8u<; 1tOlWV.7 Since Fritzsche it has been 
usual to see this as another attack upon Eupolis' use of Knights 
(comm. ad PCG Ar. fr. 59). We may be entitled to conclude that 
Eupolidean appropriation of Aristophanic material was a rather 
bigger issue-and more publicly accessible-between 421 and 
(say) 415 than we have been accustomed to suppose. 

But if Clouds II was not answered by Baptai, because it 
belongs to the debate, there is nothing to prevent us from 
conjecturing that it was designed to reply to Baptai. The puzzling 
implication of ~/J.£'tEpOU<; at 554 suggests a complicity between 
the passages. The implied 'we' will on this reading transparently 
be the poet whose voice the parabasis represents (6 <pUAUKp6<;) 
and Eupolis. Part of the line's humour would then rely on the 
apparent acceptance by the poet of Eupolis' claim in Baptai ("so 
what if he did co-write it; he turned it inside out in Marikas").8 
The connection between the image of clothing as comedy in 
Anagyros (XAuvi8o<;, cmAT1yi8u<;) and the metaphor of turning 
cloth to reuse it at Clouds 554 (£KO"'tpt'VU<;: Dover in Nub. 554) 
would fit well with a series of public accusations and replies. 
Anagyros comes first. Baptai replies. Clouds II picks up the 
image from Anagyros and the joke from Baptai as the basis for a 
further complaint. 

In the evidence from Cratinus' Py tin e, it is uncertain whether 
the connection with Knights 531 is that of the scholiast or was 
prompted by something in the parabasis. But it seems certain 
that the charge against Aristophanes for "saying the things of 
Eupolis" was in the text, and it is not easy to account for the 
precision of the scholion if there was not also some allusion to 
the Knights. So it looks reasonable to say that Cratinus was 
alluding here to a noticeably Eupolidean aspect of Knights. 

Halliwell's interpretation of the evidence uses the Baptai 
fragment to unpack the Pytine fragment. Behind the accusation 
of Cratinus in 423 he sees the "collaboration" between 
Aristophanes and Eupolis alluded to in fr. 89. He also claims, 

7 "Dating and Re-dating Eupolis," Phoenix 44 (1990) 22; cf Storey (supra 
n.1) 385f; P. Geissler, Chronologie der altaltischen Komodie (Berlin 1925) 50. 

s Halliwell (n.17) draws attention to the usage, contrasts it with 559f 
(reiterated singular possessive), suggests that it is not the norm for references to 
the poet, and proposes that it might be "an allusion (for those who knew) to 
some Eupolidean involvement in Knights." If my reading is correct, then there 
is no need for the rider "for those who knew," for the matter of the 
relationship would have been public knowledge at least since Baptai. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

368 ARISTOPHANES' KNIGHTS AND EUPOLIS 

apparently basing his view on his (disputed) reading of Wasps 
1018-22, as evidence for an early, secret, collaborative stage in 
Aristophanes' career, that this collaboration with Eupolis was a 
secret that was now revealed to the audience for the first time. 9 

Halliwell writes (524), "it remains unclear how much he 
[CratinusJ gave away.» This is tantamount to an admission that 
the crucial elements of the interpretation come from his 
underlying model and not from Cratinus. One has to say that 
even if that were the correct interpretation of the Wasps 
passage, the Knights collaboration would be a different matter, 
involving a similar exchange of material but between poets who 
were both established (with at least Lenaea victories behind 
them) and hence continual rivals for festival prizes (e.g. at 
Lenaea 425, where Acharnians won against Cratinus' Cheimazo­
menoi and Eupolis' Noumeniai). 

This scenario seems to me to call for a good deal more 
justification than it receives from Halliwell. Is it really 
credible-in the agonistic atmosphere of Athens-that rival 
poets would assist one another in this way, especially when the 
evidence in Old Comedy for relationships between named 
poets reveals very little in the way of positive evaluation (and 
even that is rendered dubious by the weight of antagonistic tes­
timonia)?lO 

9 "Aristophanes' Apprenticeship," CQ N.S. 30 (1980) 33--45. See also G. 
Mastromarco, "L' esordio 'segreto' di Aristofane," QuadStor 10 (1979) 153-96 
for the same view but more detailed reading of the text. D. M. MacDowell 
leads the opposition (Wasps [Oxford 1971] ad 1018, and "Aristophanes and 
Kallistratos," CQ N.S. 32 [1982] 21-26). For a recent assessment of the 
discussion see T. K. HUBBARD, The Mask of Comedy: Aristophanes and the 
Intertextual Parabasis (Ithaca 1991: hereafter 'Hubbard') 227-30 (agreement 
with MacDowell: 228). For a new reading with important implications for the 
relationship between the comic poets and additional support for the 
conclusions of Part III below, see my "Aristophanes' Acharnians and 
Eupolis," CIMed 45 (1994) forthcoming. 

10 Storey (supra n.1: 387f) is content to assume a collaboration that was 
public. I think this needs arguing for, given the nature of the evidence. See 
Halliwell 517ff for the data on "collaboration." Remarks that claim that 
another poet "laboured for others" cannot be interpreted as friendly in a 
culture where the olive-picker is satirized for precisely that (Vesp. 712; cf Oem. 
57.45). See below. The best that can be said for Eupolis fro 89 is that there is no 
open insult. But it is part of a slanging-match between poets and the whole 
notion of "helping a poet to write" is treated elsewhere as adequate grounds 
for attacking the "helped" poet (e.g. Ran. 73-79). See below. On top of these 
testimonia we have in Aristophanes insulting references to Cratinus (Ach. 849, 
1173; Eq. 400, 531-34), Eupolis (Nub. 554), Hermippus (Nub. 557), Lycis, Phry-



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

KEITH SIDWELL 369 

To sum up, the principal objections to this way of reading the 
evidence are (1) that nothing in the scholion suggests or requires 
that Cratinus be accusing Aristophanes of collaboration; (2) that 
without the (chronologically later) Baptai fragment we would be 
likelier to interpret this as an accusation of plagiarism. It is 
necessary for us to accept that the debate that we can see 
opening up in Baptai and Clouds II is couched in terms of 
collaboration. But as things stand, we cannot safely say that the 
way the debate develops later is necessarily indicative of its 
earliest stage. After all, this is comedy and Athenian comic poets 
valued originality (hence the debate). So new twists (such as that 
suggested above between Baptai and Clouds II) are more rather 
than less probable in the treatment of the same material. One 
might add that the impact of an attack on Aristophanes' lack of 
originality (a matter of concern to comic poets; cf Lysippus fro 
4: oub' aVa1(vu\jfa<; Kat 8Etwaa<; 1a<; aAAo1pla<; £1ttVOla<;) would 
have been consequentially greater if one assumed that it was 
made against the background of something the audience already 
knew. 

The inference is possible, then, that Knights at the time of its 
production somehow appeared Eupolidean to its audience and it 
was this open Eupolideanism to which Cratinus was referring in 
Pytine. 

II. Interpretations of the Primary Evidence 

There are four main ways in which these pieces have been 
assembled into interpretative hypotheses. (1) The ancient 
scholars took Eupolis fr. 89 at face value and consequently 
undertook a search for Eupolidean material in Knights. At 1288 
(oan<; oiJv 'tOtO\hov aVbpa Jlil ac:p6bpa ~bEA;UHE'tat) they saw 
what they were looking for and drew the inference that the 

nichus, and Ameipsias (Ran. 14f), Magnes (Eq. 525), Crates (Eq. 538f with 
Hubbard 75), and in Plato (fr. 86) and Eupolis (fr. 62) attacks on Ar. Pax. 
Crates "attacked Cratinus" (having started as an actor in his plays: peG T2a). 
It is true that Cratinus seems to be better treated at Pax 700 and Ran. 357, but 
(1) the evidence for antagonism precedes these passages and may need to be 
understood in their interpretation (this is the wayan audience would operate); 
(2) in Pax the motif of Cratinus as drunken comic poet surely alludes to his 
own Pytine of 423: that play is likely to have been a self-defense (was he cured 
of alcoholism and rejuvenated?), whereas this passage implies (falsely) that this 
love of drink had in a sense killed him. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

370 ARISTOP~IANES' KNIGIITS AND EUPOLIS 

entire second parabasis had been wri tten by Eupolis. ll Pohlenz 
and Colonna suggested that what the scholiast had seen in 
Eupolis was a resemblance to Demes fro 99.33: o<J'tu; o-ov apX£lv 
'tOtau'to'\)<; avbpu<; UiPEt'tUt 1tO'tE. 12 As Sommerstein remarks, 
this was not a very intelligent guess, if it was the basis of their 
conjecture, since Demes must be located in the 410s.13 But 
perhaps Hubbard (85f) is right to suspect that the scholiast was 
not after all so stupid. He points out that it is more likely that 
1288 was an actual line of Eupolis. The scholiasts were quite 
good at spotting direct quotation, and Sommerstein (51£) has 
cogently argued that line 1225 was a line from Eupolis' H eilotes, 
which they did spot (though they did not specifically link it to 
Eupolis). The best explanation of 'LEq. 1291, then, is that the 
discovery of a Eupolidean quotation at 1288 gave the answer to a 
zetema arising from our fro 89. This focus on the second 
parabasis alone, however, puts in doubt Sommerstein's thesis 
about 1225, since the search through Eupolis would surely have 
revealed the H eilotes passage as well. See further under (2) 
below. 

Scholars have rightly seen this ancient attempt to ascribe bits 
of Knights to Eupolis' cooperation with Aristophanes as naive 
and wrongheaded. 14 What we can say, however, is that we are 
able to detect some actual Eupolidean material in Knights. We 
cannot ignore its presence in any attempt at understanding the 
claim of Eupolis in Baptai. 

(2) Sommerstein locates the exchange entirely in Aris­
tophanes' appropriation of Knights 1225 from Eupolis' 

lI:EVE Eq.1291: f:K'tOu' o(J'tt~O~V'toW1ltOVaVOpu' <PUOi'tWE~E\l7t6AlOO~dvat 
'tllV 1tUpo.~UOlV, Et Y£ <PTjOW EU1tOAt~ ';UVE1t01TjOU 'tOOl CjlaAaKpoot'. 

12 M. Pohlenz, "Aristophanes' Ritter," NAkG (1952.5) 12 off; A. Colonna, 
"Aristofane ed Eupoli nella seconda parabasi dei Cavalieri," Dioniso 15 (1952) 
32-37. 

13 "Notes on Aristophanes' Knights," CQ 1'.5. 30 (1980: hereafter 'Sommer­
stein') 52; for dating see Storey (supra n.7) 24-27. 

14 Halliwell (523f) is much more subtle, of course, in his theory of "authorial 
collaboration" (" ... the balance of probabilities is that Eupolis' help had been 
slight or limited to some suggestions, rather than entailing extensive writing"). 
Nonetheless, he is perilously close to following such procedures as production 
of supposed facts from the text for which he elsewhere castigates the scholiasts 
(" Ancient Interpretations of QVOIlUO'tl KWIl'P0EtV in Aristophanes," CQ N.S. 34 
[1984] 83-88, esp. 88: "an assumption that Aristophanic satire accurately repro­
duces the historical truth"). 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

KEITH SIDWELL 371 

H eilotes. 15 Even if we accept the Eupolidean authorship of 
H eilotes, 16 however, we cannot be sure that this quotation 
belonged to a play by Eupolis. We have in fact already given 
grounds for arguing that 'LEq. 1291 implies that 1288 was the 
only detectable citation from Eupolis in the play. Line 1225 
could perfectly well be a citation from another comedy with a 
Spartan theme (e.g. Cratinus' Lakones). This would not have 
emerged in investigations centred on detecting E upolidean 
material in the second parabasis. 

Nonetheless, Sommerstein's argument could still be used in 
relation to the citation at 1288. If there was nothing more, 
however, than this one line of Eupolis in the play, it stretches 
credulity that Eupolis could so confidently use its citation to 
beat Aristophanes in public at least seven years on. Nor, by the 
same criterion, does Cratinus fr. 213 PCG suggest that the 
use/abuse of Eupolis was so limited. We can add Mastromarco's 
point that "strictly speaking ~UVE1tOtll(Ja should mean more than 
this" (ap. Sommerstein 52 n.33). Halliwell backs this up in a 
disavowal of his earlier opinion by adding that "his argument ... 
leaves the specificity of the fragment unexplained" (523 n.16; cf 
supra n.9: 40 n.31). 

If just this one line of Eupolis was embedded in Knights and 
the audience was to see the point of Eupolis' claim of co­
authorship at least seven years later in Baptai (fr. 89), then 
according to Sommerstein's argument they must be presumed 
to have recognized this material as Eupolis' at the time of perfor­
mance. If so, it seems highly unlikely that there was not some 
point to the appropriation of the line. 

The serious question is this, then: is this line borrowed with 
Eupolis' blessing (a version of the collaboration model), 
plagiarized (with the expectation that the audience will not 
recognize it), or employed allusively (with the opposite 
presumption)? It does not seem likely that Aristophanes asked 
the author's permission, since this could hardly be com­
municated to the members of the audience who recalled that it 
was Eupolidean material. The borrowing would thus backfire, 
and be naturally interpreted as plagiarism. It does not seem 
likely that the material was plagiarized either. The poets do seem 
very sensitive to such accusations (otherwise the debate on both 

15 Sommerstein 51ff, and Knights (Warminster 1981) ad loc. 
16 Storey (supra n.7: 7) argues in its favor. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

372 ARISTOPHANES' KNIGHTS AND EUPOLIS 

sides would not be conducted in these terms). And it would 
have been very risky. We know that Aristophanes makes jokes 
with lines of tragedy many years old, where the point some­
times depends on the assumption that the audience recalls their 
original form (e.g. the reuse of Eur. Ale. 177-82 at Eq. 1251£ 
recalls a play produced in 438).17 Much the most likely 
proposition is that the audience is expected to recall that this is 
Eupolidean material. Eupolis seems to have done the same thing 
to Aristophanes in Marikas, where fro 208 (=Quint. 1.10.18) 
looks like a quotation of Knights 188f. The time-gap here is 
certain (L424 to L421) and well within the detailed memory 
requirement of the poets between one comedy and another 
(Knights is recalled in Baptai, perhaps seven years later). The 
reuse of this material may have been highlighted by similarity of 
situation between the plays and even pointed by some 
relationship-invisible to us-beween the character who spoke 
it in one play and the one who uses it in the remake. 

Given Aristophanes' record on tragic quotation (e.g. Ach. 
543=Eur. Telephus fro 709 Nauck-Snell), one might suspect a 
satirical purpose in reusing Knights 1288, and Eupolis would 
presumably be doing the same back to him in Marikas. Cratinus' 
accusation in 423 might be taken to indicate even more 
Eupolidiana. But on the foregoing argument, we can infer from 
the scholiastic method that the Eupolidean material located by 
Cratinus was imitative rather than quoted. By the same token, if 
the material stood in this relationship to Eupolis, the scholiasts 
would not have been looking for it. If the material, however, 
was only imitative, the continuing joke-sequence culminating in 
Clouds 554f outlined above would have made little sense. It is 
equally possible, however, that parody was intended. There is 
overuse of ocrn~ in the second parabasis (1275, 1278f, 1288, 1301, 
1306). This could point towards parody of what we know from 
the Demes fragment to have been perceptible as a Eupolidean 
stylistic trait (Hubbard 85f). 

The upshot of this discussion is that (a) there is Eupolidean 
material, actual and imitated, in Knights; (b) its presence is 
likelier than not to be deliberate and satirical: the audience is 
invited to spot it and laugh. The implication for Eupolis fro 89 

17 See Harriot (supra n.6) 2f; Cf the use of Telephus of the same year in 
Acharnians with many detailed allusions to that text, which it is vital for an 
audience that wishes to follow what is going on to catch. See H. P. Foley, 
"Tragedy and Politics in Aristophanes' Acharnians, " ]HS 108 (1988) 33f. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

KEITH SIDWELL 373 

will be that Eupolis is making a joke about the obvious, 
intended, presence of this material in Knights and turning it to 
his own advantage. 

(3) Halliwell accepts Eupolis' claim, in however attenuated a 
form, and, as we have seen, thinks of a system of authorial 
collaboration as underpinning the references. I have already 
given some reasons for disbelieving his account of the material. 
There are four further objections, however, that touch on the 
whole nature of "authorial collaboration" and the evidence for it. 
First, the notion of co-authorship by poets is highly suspect in a 
tradition where the mythical ideology, a Muse who inspires the 
bard (e.g. Wasps 1022, 1028), supports individuality.18 Second, 
the whole notion of attacking someone for stealing ideas, using 
someone else's work, and pretending it is one's own, or having 
to rely on the help of another poet-the very ideas that 
Halliwell's collection of evidence sets out-fits neatly into this 
assumption of poetic individuality. It follows that, though there 
must be a visible basis for such attacks, it is unlikely in the 
extreme to be an actual collaboration beween poets. A con­
sequence of this is that problematic passages such as Wasps 1018 
and Eupolis fr. 89 need to be interpreted on the basis that the 
revelation of collaboration constitutes an attack upon the poetic 
virility of the assisted party (h£polO"l1tolTj'tal<; at Wasps 1018; 
Aristophancs in Eupolis fr. 89). One should compare Frogs 
73-79 (Sophocles and Iophon), 944 (Cephisophon and 
Euripides), Cratinus fr. 502 (Choerilus and Ekphantides).19 

t8 See e.g. P. Murray, "Poetic Inspiration in Early Greece," JRS 101 (1981) 
96£. 

19 The idea of working for another, embodied in the material under I-Ialli­
well's sections A6-7 and A 9 may well be of a different order. Halliwell 
differentiates the references as follows: (1) in A6 Plato's description of himself 
as having given plays to others because of poverty is taken tentatively to be a 
reference to the same activity as Vesp. 1018-22, viz. "helping other poets" 
(collaboration); (2) the A 9 fragments referring to Aristophanes in Heraclean 
fashion labouring for others are taken as either the same activity as in Vesp. 
1018f as interpreted by Halliwell and Mastromarco (viz. collaboration) or as 
references to the IIeraclean self-portrait in the parabases of Wasps and Peace; 
(3) the statement of Eratosthenes that Plato aAAOt~ i:OtOOU ta~ 1((t)Il(JHOia~ and 
£UOo1(llln with them is taken to mean that more than one of Plato's earlier 
plays was produced through another, that they gained credit for these victories 
(so that the didascalic records gave names of both poet and didaskalos), and 
that these cannot have been first prizes. Here we are dealing with inferences 
made by Eratosthenes. Halliwell says (n.20), "it is hardly conceivable that E. 
was basing himself not on didascalic information ... but on the parabasis of 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

374 ARISTOPHANES' KNIGHTS AND EUPOLIS 

Heath 20 (see next section) is right to see the motive behind this 
usage as satirical, for the very idea that a poet would need or 
indeed accept help from another would in this particularly 
individualistic tradition detract from his nll~. 

Third, all the evidence for such collaboration presented by 
Halliwell comes from comedy. This is in itself suspicious. These 
passages might in fact be jokes, underpinned by some reference­
point that presently escapes us. See further under (4) below. A 
passage from Eur. Andr. 476ff makes the case well by giving the 
exception that proves the rule: '!EKOV'tOlV 8' UIlVOV Epya'!alv 
8uo'iv rplv MODO'at qHAODO'l KpaivElv. It is true that this is 
evidence for poetic collaboration in the writing of UIlVOl. The 
expectation is also that an audience would be familiar with the 
phenomenon. 21 UjlVOl, however, are not comedies and we may 
here be looking at certain quite specific requirements of 
particular cults for certain festivals. Indeed, this hypothesis is 
rendered more likely by the placement of this example in the 
ode. For the chorus is here decrying the idea that the two heads 
(or rather beds) are better than one: OU8E1W'!E 8i8ulla A£K'!P' 
E1tatv£O'U) ~po,!(j)v, etc., 465. The exemplum of poetic collabora­
tion is in this vein. When it happens, the Muses induce strife 
between the collaborators. And why? Because it is normal to 
have one man, one Muse. 

Fourth, one requires an answer to the following question: "If 
Eupolis helped Aristophanes to write Knights, and such collab­
oration was actually quite common, why do we never meet in 
the didaskaliai the formulation 'Aristophanes and Eupolis won 

Peisander; this would leave the use of EUOOKtllEt unexplained." But if Ve sp. 
1018-22 is to be taken as the model (whatever activity it refers to), then 1023 
(on Halliwell's reading) corresponds to a possible claim by Plato in the 
parabasis of Peisander that he did gain fame by this procedure. The treatment 
of these references ignores the possibility (reinforced by the present argument) 
that such statement were part of a cornie poet's laughter-creation kit. See supra 
n.9 for reference to a new interpretation of Ve sp. 1018-22 and further 
discussion in nn.26, 29 below. 

20 M. Heath, "Aristophanes and his Rivals," G{;'R 37 (1990: hereafter 
'Heath') 152. 

21 Both these points were made to me in correspondence by Professor Alan 
Sommerstein. I am grateful to him for bringing the passage to my attention. 
See P. T. Stevens' edition (Oxford 1971) ad 476: u ... it is odd that such 
collaboration should be common enough to provide an obvious instance of 
divided authority and to justify the frequency implied by <jn.A.oooL" 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

KEITH SIDWELL 375 

the prize'?"22 We should now definitively reject Halliwell's 
account of cooperation between the poets Aristophanes and 
Eupolis. The basic explanation of fr. 89 is that material taken over 
from and imitative of Eupolis in Knights was obvious to the 
audience (and to Cratinus), and Eupolis can make a joke on the 
basis of this piece of public knowledge later. We have seen that 
these circumstances also point to an antagonistic and satirical use 
of Eupolis by Aristophanes. We must now turn to current 
views of the broader 'war between the poets'. 

(4) Heath (152) takes the sceptical view that "charges of 
plagiarism are part of a system of ritualized insults ... not meant 
to be believed, but to make the other party lose face." Som­
merstein too (52 n.33) has taken a similar line on Eupolis fr. 89: 
"One must expect ... that the accusations made by comic poets 
(especially against each other) will be wildly exaggerated." So far, 
so good. Heath, however, sees (152) comedy as having a 
"common pool or repertoire of comic material" that was 
contributed to and could be used, with consequential claims to 
originality, by all comic poets. 

The problem with this view is that it ignores the particular 
terms of the claims and counter-claims as we have them. 
Neither Cratinus' nor Eupolis' words, however much designed 
(as I agree) to insult and undermine Aristophanes, make any 
actual sense if the audience could not see something Eupolidean 
in Knigh ts. The same is likely to be true of the way Aris­
tophanes speaks of Eupolis' M arikas at Clouds 554f. In fact, 
Storey has recently tabulated a series of thirteen points of 
appropriation from Knights to Marikas.23 

Heath's view ultimately rests on the assumption that the body 
of comic material is independent of the comic poets who 
produce it. This cuts across the individualistic nature of ancient 
poetic practice and the desire of poets to claim originality that is 
the positive side to these attacks. 24 It is also contrary to what use 
we know was made of tragic poetry.25 

22 I am grateful to Dr David Braund for making this point to me In 

conversation. 
23 Storey (supra n.1) 383f, though he would class them as imitations. 
24 Halliwell (519) recognizes this aspect of the matter in putting Lysippus fr. 

4 (quoted supra 5) in his set of evidence (AD). 
25 It is no use objecting that there was a clear stylistic distinction that made 

tragic diction recognisable. Time and again in Aristophanes the point is that a 
particular poet's tragic style is parodied. In any case, Dionysiades of Mallos 
did undertake stylistic analyses of Old Comedy (Suda t11 169: see J. Hender-



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

376 ARISTOPHANES' KNIGHTS AND EUPOLIS 

As soon as we posit an antagonistic 'paracomedy', we obtain a 
real referent that contains and explains the humour of the 
material gathered together by Halliwell. For if poet A has 
attacked poet B by presenting his own production as though it 
were a composition of poet B, utilizing and satirizing verbal, 
visual, and musical material derived from his rival, it is then open 
to poet B (or others) to attack poet A in return by misrepresen­
ting this satire. The poet has a number of options that all rest on 
the assumption that individual originality and poetic indepen­
dence are central to the comic poet's art and to the audience's 
perception of it. He can accuse poet A of plagiarism (e.g. Ar. 
Nub. 553-56, fr. 55 Anagyros) or imitation (Nub. 559, Cratinus 
fro 213, Pytine), placing emphasis on his own originality 
(Lysippus fr. 4; Ar. Nub. 546£). He can claim co-authorship of 
the piece (Eupolis fro 89, Baptai), on the grounds that the 
parodied material utilized from his own work makes the pirated 
play identical with its original source of inspiration. He can 
humorously suggest that he has allowed others to produce his 
plays (Eupolis fr. 89 Baptai; Plato frr. 106-07 Peisander; Vesp. 
1018ff). A poet may likewise be ridiculed as "working for 
others," because his material, style, and voice have been 
misappropriated by another comic dramatist for satirical 
purposes (Aristonymus fe 3, Ameipsias fro 27, Sannyrion fe 5 
PCG).26 If the poet has produced paracomedies, he may be 
accused of putting on plays by others as though they were his 
own (Hermippus fr. 64, attacking Phrynichus). To this mode of 
attack probably belong references to stealing plays (Eupolis, 
PCG T S=Ael. NA 10,41) or props (Ar. Pax 729-32). 

son, ed., Lysistrata [Oxford 1987] lxii), which differentiated the poets on this 
criterion. In performance, as well as in language, there must have been clear 
idiosyncrasies (cf e.g. the description of Cratinus' )lou<nKl] at Ach. 851). 

26 I suggest in "Aristophanes' Acharnians and Eupolis" (supra n.9) that 
there are two separate aspects to the joke at Eupolis fr. 89: (1) co-authorship 
with 0 <paMxKp6~, (2) the role of Aristophanes as mere producer. I present 
grounds for distinguishing between the two poets involved in the process. 
Halliwell (521) disputes the link usually made between Plato fr. 107 and the 
similar use of 'tE'tpabl YEv£o8at found in Aristonymus, Ameipsias, and San­
nyrion, because there is no reference to Aristophanes in Plato's fragment and 
its surrounding material. But in the current context, it occurs to me to wonder 
whether Aristophanes had not been subject to the same parody as is suggested 
here for Eupolis. He is associated with Heracles (Vesp. 1030, cf 1043; Pax 
752-though I think it is dangerous to assume that the poetic voice referred to 
here is Aristophanes') and could therefore even be the ventriloquial target of 
Plato's Peisander. See also supra n.19 and n.29 infra. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

KEITH SIDWELL 377 

Within this metatheatrical milieu, two related games may be 
being played, which appear in the evidence collected by 
Halliwell. The first involves the notion of collaborative writing, 
either between two poets (Eupolis fr. 89), or between a 'servant' 
(or some other citizen) and a poet (Cratinus fro 502; cf 
Telecleides frr. 41-42 PCG; Callias fr. 15 PCG; Ar. Ran. 944, 
1408, 1452f; cf 73f). There is a scene in Cratinus' Pytine (frr. 
208-09) in which two characters cooperate in writing a comedy. 
Pieters long ago suggesed that the participants were Aristoph­
anes and EupolisY Whether or not his particular solution is cor­
rect, it is certainly obvious that in the context of paracomedy sa­
tirical capital could be made by showing collaboration on stage. 28 

The second involves the ironical effects invited by the meta­
theatre. If a comedy may be presented as though by another 
poet, then part of the humour might lie in the ironical presenta­
tion of that author's paracomic abuse by others. That is, the 
satirized poet may be ridiculed by being made to claim as his 
own plays that were in fact ventriloquial parodies of his 
comedy.29 

27 J. T. M. F. Pieters, Cratinus (Leiden 1946) 151. In Acharnians, the scene in 
which Dicaeopolis borrows costume, props, and speeches (447) from Euripides 
is an example of a sort of collaboration between a tragic and and a comic poet, 
since as early as 416£ the character speaks as though he is an actor, and his role 
as a poet becomes clear at 499, if it is not forced upon the audience by the 
proximity of Dicaeopolis to Eupolis at 406 (so E. L. Bowie, "Who Is Dicae­
opolis?" jHS 108 [1988] 183ff). See my forthcoming article (supra n.9). 

28 It is therefore untrue that the development of the term (!t) jl1tou:'iv to mean 
"co-write" cannot make sense without a consequent referent in the real world 
to support it (Halliwell 520). There is no reason why in comedy a term may 
not be borrowed from the real world, where it has real meaning (e.g. Andoc. 
1.62 of helping someone to do something; INub. 857 of a sculptor's assistant), 
and applied to an invented situation where it is funny precisely because it has 
no application in the real world. The term is not used at The sm. 157 in 
reference to collaborative writing (as Halliwell implies: 520 n.14). Euripides' 
relative is in fact offering Agathon the chance of being buggered, so that he 
will be able to empathize with satyrs, just as in the present scene his 
transvestism helps him to empathize with women. This amounts to material 
assistance but not to co-authorship. 

29 E.g. I do not think we ought to take seriously the claim reported from 
Plato (fr. 106 PCG Peisander) that he handed his plays to others "out of 
poverty," because this is tantamount to admitting he was a wage-earner, 
against which status there was great prejudice, visible in comedy and else­
where (see supra n.10). Plato's remark recorded via Eratosthenes in Ar. fr. 
590.44-51 (on COOC; jlev o.AMnc; tMoo'\) 'tac; KOOjlOOtoiac; fUOOKlJlEt, Ot' a{rtou oe 
1tpOrtov otou~ac; 'toue; 'Pa~oouxo'\)e; Kat Y£VOjlfVOe; 'tE'tap'tOe; u1tfwo8Tj 1tUAtv de; 
'toue; ATjVUtKOUe;) is also suspiciously self-condemnatory and would make sense 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

378 ARISTOPHANES' KNIGHTS AND EUPOLIS 

The upshot of this discussion is that accusations of plagiarism 
and of collaboration are both aimed at ridiculing the party so 
accused. But the audience is invited in these instances to con­
template the misappropriation of specific, not generic, material. 
This means that comic plays could be-and often were-made 
out of other comic plays, not only consciously mimicking the 
techniques of their original authors in order to satirize them, but 
even pretending to be plays by that dramatist. I argue elsewhere 
(supra n.9) that it is this "ventriloquial paracomedy" to which 
Wasps 1018-22 refers, and that Acharnians is a thoroughgoing 
example of the genre, aimed at Eupolis. 

The comic plays of Aristophanes' rivals, then, could be crucial 
to the audience's understanding of his own dramas. Our current 
model suggests that it does not much affect our reading of these 
texts that this material is missing except in fragments: after all, 
the ancient scholiasts had a great deal of it. But to set aside the 
conclusion of the discussion thus would be seriously to misprise 
the priority of performance, which was not available to ancient 
scholars any more than it is to us. It would also involve having 
to explain just why there are so many references in our texts to 
purloined material, but so little is made of this by scholiasts, if 
the material was not being reworked in ways that would qualify 
as original, while still retaining their referential basis (satirical) 
towards their source. It is worth pursuing the trail to see 
whether it can make any contribution to the unravelling of the 
puzzle about Knights and Eupolis, however speculative this is 
bound to be in a context where we have such meagre remnants 
of Eupolis' oeuvre. 

III. A New Hypothesis 

The proposition is, then, that in Knights Aristophanes was 
satirically reworking material from an earlier Eupolis play (or 
plays). Several of Eupolis' plays were in the public domain, 
having been produced at festivals since 429. 30 It now becomes 

as a joke aimed at another poet who had been subjected to paracomic ridicule. 
See also supra nn.19, 26. For an analysis of a satirical ventriloqui'LI paracomic 
parabasis see supra n.9 (forthcoming). 

30 See Storey (supra n.7: 29) for a list of possible plays with dates of pro­
duction. He sets four before Knights, but it is perfectly possible tha there were 
more. See my forthcoming article (supra n.9) for the suggestion that Poleis 
belongs earlier. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

KEITH SIDWELL 379 

possible to see Eupolis' Marikas as revenge for a previous slight, 
viz. the reuse of one of his plays to make Knights (which he 
later comically reclaims for himself in the parabasis of Baptal). 
On this hypothesis, the 'war between the poets' gains con­
siderably in theatrical importance, because it is being conducted 
publicly at an intertheatrical level. For the basic assumption of 
such parody is that its audience will recognize it without being 
told directly "by the way, this is taken from Eupolis."31 

What in Knights is so germane to the playas to constitute an 
unmistakable allusion when briefly expressed? We are fortunate 
to answer this question unequivocally. Commentators on Ach. 
299-302 have no doubt that the linkage of Cleon and the Knights 
there provides a sort of 'trailer' to Knights. Sommerstein in 301: 
"Here the chorus foreshadows the violent attack on Cleon 
delivered in Knights the following year." Hubbard (34) goes 
considerably further in considering this passage as "a hint about 
his next play.» But neither this passage nor 5-8 in the same play, 
which also mentions Cleon and the Knights together, contains a 
speciflc allusion to the plot of Knights. Cleon is not put on trial 
in Knights, as it appears he is at Ach. 5-8. The chorus of Knights 
is not the same as the chorus of Acharnians and in any case, 
Cleon is not cut apart for the knights' benefit, but for that of the 
Sausage-seller and Demos. Besides, how could an audience be 
amused or entertained by such a remote allusion if Knights was 
as yet not even written and certainly was not in the public 
domain ?32 If these are allusions to comedy (and it has often said 

3\ A recent example from modern satirical parody might help. In the first 
programme of the 1993 Autumn series of "Spitting Image" (broadcast Sunday 
7 November), members of the British cabinet were shown reacting with 
boredom to privatisations of the past, but with great pleasure to the 
anticipation of privatising British Rail. The visual and musical framework for 
the sketch was that of a current television advertisement for "Pepsi-Max" 
(itself designed to be amusing). It would have undermined the joke had there 
been any overt reference to the source of the parody. This example incidentally 
produces a good model for the amount and importance of the interpretative 
signals that may be missing from our texts of Aristophanes. 

32 It is, as one of my correspondents points out, possible to envisage an 
author giving "a tantalising (and possibly misleading) indication of the 
author's future intentions; compare e.g. the epilogue to Shakespeare's Henry 
IV Part II, where the author promises (disingenuously) to 'continue the story, 
with Sir John in it'." But the example given from Shakespeare is unequivocal, 
because the above quote is prefaced by "our humble author will." In 
Acharnians on the other hand, the first mention of Cleon and the Knights 
(5-8) is on the lips of the main character (and is often taken to be a reference 
to a comic scene, see below and n.33), and the second (299ff) is spoken by the 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

380 ARISTOPHANES' KNIGIITS AND EUPOUS 

that 5-8 is),33 then they must be referring to plays already in the 
public domain by Lenaea 425. I have suggested elsewhere (supra 
n.9) that this consideration, wedded to a textual crux that sur­
rounds the verb at Ach. 301 and a later self-reference by the 
chorus to itself as the chorus of an earlier play (1153), makes me 
suspect that the text originally contained a past tense very (say 
"taIlOY). Scholiasts related the passage to the Knights and their 
comments eventually crept into the text. So, despite the ever­
popular notion that the old farmer is referring to Aristophanes' 
own Babylonians, the principle of economy of explanation 
suggests that for a central theme of Knights that predates that 
play we ought to be looking at Eupolis, not Aristophanes, for 
the inspiration. 34 

chorus in character to another character, and not (as in the parabasis) directly 
to the audience on the poet's behalf. No indication is given at all that this 
constitutes the author's future plan (contrast the Shakespeare example), as 
would have to be the case where an as yet unwritten, or at least certainly 
unperformed, work is supposedly advertised. Sommerstein on the latter 
passage argues: "there is no particular reason why the Acharnians should be 
hostile to Cleon [he means, no reason is established by this text, which in the 
light of the present argument is probably not the point]; rather, the chorus 
here (in the middle of a sentence) shift to speaking in their capacity as a comic, 
and specifically an Aristophanic, chorus." What theatrical grounds have we for 
believing that an audience could understand such a shift of voice in the middle 
of a sentence? Sommerstein's support is taken from the intrusions of the 
actor's voice in the speeches of the main character at 416 and of the author's 
voice at 377-82, 499, none in mid-sentence. I would in any case be inclined to 
suggest that theatrically, the latter passages could not be understood without a 
simultaneous identification of character and referent that is certainly not 
available from within the text. That (which cannot be what is happening with 
the chorus at 299ff, since they are multiple) is exactly what Dicaeopolis 
suggests to Euripides has to be the case when he is disguised (442ff). I deal 
with the problem of multiple voices in the Acharnians elsewhere (supra n.9). 

3J The first proponent of this view was H. Liibke, Observationes Criticae 
(Berlin 1883) 17f; see most recently E. M. CARAWAN, "The Five Talents Cleon 
Coughed Up (Scho!. Ar. Ach. 6)," CQ N.S. 50 (1990: hereafter 'Carawan') 
137-47. 

H It seems reasonable to trust the evidence of IAch. 378 that Babylonians 
was produced before Acharnians, but the question of whether or not it won 
the prize is still uncertain (Storey [supra n.7] 10f). There is no ancient evidence 
to connect Ach. 5-8 with the play, however. Given the tendency of the major 
school of criticism in antiquity to attend to names as opposed to caricatures, 
the fact the Cleon was mentioned in the play (as the scholion says) is not in 
itself evidence that he was on stage as a character (though reference in Eq. 358, 
976 to two of the play's disguised caricatures, viz. Nicias and Cleon, suggests 
that mention of a name could square with the appearance of that person in 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

KEITH SIDWELL 381 

There may be other Eupolidean connections in Acharnians, 
which would support the conjecture that the Cleon/Knights 
motif goes back to Eupolis. In 1988 Bowie suggested that the self­
presentation of the leading character as a comic poet in 
conjunction with the possible comic relationship between the 
names Dicaeopolis and Eupolis might lead us to conclude that 
Eupolis stands behind the mask, rather than Aristophanes (supra 
n.27; cf n.32). This identification has been strongly contested by 
Parker and Storey, and it has to be said that Bowie's arguments 
in its support are not especially strong. 35 That admitted, Fisher 
recently argues that the concatenation cannot be completely 
dismissed, especially in view of the extraordinary proliferation 
of "voices" in the play.36 Storey, however, at least certainly does 
accept that there is para-Eupolidean material in Acharnians. He 
infers from the relationship between Ach. 3, \jfUJlJlulCocrlOyap­
yapu, and Eupolis fro 308, aptSJletV SeU'tile; 'l'UJlJlulCocrio'Ue;, that 
"Aristophanes has gone Eupolis one better. "37 He is content to 
use this as one of his arguments for setting Chrysoun Genos 

caricature). An anonymous referee points out that it is the enthusiasm of the 
reference that makes it seem more likely to refer to something in Aristophanes. 
I suggest (supra n.9), however, that the character who voices this enthusiasm 
would have been recognized instantly by the audience as a character from the 
play of Eupolis that contained this scene. The humour is not, then, necessarily 
direct but is filtered through a number of ironic structures, including 
paracomedy, which are at present invisible to us. In my view (see Part IV 
below), the humour of this passage also depends very much on who the 
character on stage represents and what the audience knows about his normal 
political attitudes. 

35 L. P. E. Parker, "Eupolis or Dicaeopolis?" JIlS 111 (1991) 203-08; Storey 
(supra n.1) 388-92. See also S. D. Olson, "Dicaeopolis' Motivations in Aris­
tophanes' Acharnians," J liS 111 (1990) 200 n.3. 

36 N. R. E. Fisher, "Multiple Personalities and Dionysiac Festivals: Dicae­
oplis in Aristophanes' Acharnians," G&R 40 (1993) 31--47, especially 37: "this 
possibility is hard to dispel or reject completely." (I am grateful to the author 
for allowing me to see this paper in advance of publication; see also my 
forthcoming article [supra n.9]). But it is worth pointing out that if Eupolis is 
caricatured in Acharnians, then the same basic assumption of Aristophanic 
scholarship under question here will be set at risk there also (namely that plays 
by other poets are not essential to our reading and interpretation of Aris­
tophanes). In that case, to dismiss Bowie's suggestion by adducing arguments 
based on the existing paradigm is irrelevant. His contention can only be tested 
by using it as a positive hypothesis. 

37 Storey (supra n.7) 18, who does not articulate the idea that the relationship 
may be antagonistic and parodic. For the notion of parody see Hubbard 8Sf. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

382 ARISTOPHANES' KNIGHTS AND EUPOLIS 

earlier than traditional, at Lenaea 426. 38 It is interesting to note 
how closely this Eupolideanism is followed by the Cleon/ 
Knights link. This proximity might be indicative of a desire to 
call Eupolidean comedy to the audience's mind and may act as 
further evidence the the Cleon/ Knights scenario may go back 
to Eupolis. It would follow as a matter of course on this 
argument that this material would be amusing because the 
audience had already recognised the Eupolidean parody. 39 

There is a good case, then, for Aristophanic appropriation of 
Eupolidean material before Knights, including, perhaps, allusions 
to a play by Eupolis that had a scene in which the Knights, or 
someone acting on their behalf, forced Cleon to pay back a five­
talent bribe (Eq. hypo II).40 Was it a trial (cl Wasps)? If the 
proximity of the parody of Chrysoun Genos at Ach. 3 is any 
indication, it will now be of interest to examine the remnants of 
this Eupolis play for signs of contact with Knights. 

Chrysoun Genos contains an attack on Cleon in the parabasis, 
which has also two points of detail in common with Knights. Fr. 
316.1£ reads: (1 KUAAtcr'tT) nOAl nucrwv acrue; KA£UlV £<popat, <Os 

38 A similar argument to Storey's on fr. 308 and Ach. 3 might be essayed for 
Poleis, on the basis of fro 248 a<J1to'UooC; 0' UV1lP <J1to'UoapXtoo'U lCalCtO)V and Ach. 
595, 1tOAtTIlC; XPTl<J'tOC;, ou <J1to'UoapxioTlC;, given that the sense is reversed and the 
word <J1to'UoapxioTlC; occurs only in these two passages. See my forthcoming 
article (supra n.9) for the dating of this play to Dionysia 426. 

39 See above and n.31 for this argument and a parallel. 
40 Carawan (140f) puts together an impressive case against Lubke's theory of 

a stage-trial: (1) it requires us to ignore Theopomp. frr. 93-94; (2) the argument 
from context (references to other theatrical events immediately following) is 
unconvincing, because the move from a political to theatrical events would be 
typical comic logic; (3) the fragments of Babylonians have no room for the 
Knights and there is no hint of the five-talent scene. However: (1) There is no 
need at all to see a real political event and a theatrical scene as mutually 
incompatible. Indeed, it is highly likely that we must presuppose some actual 
antagonism between Clean and the Knights for the comedy in Knights to 
operate effectively as satire. Theopompus is testimony to the real events. (2) 
"Comic logic" is too often a pis-aller in the study of Old Comedy. It is our 
own way of dealing with phenomena that may well have had different 
explanations for an audience versed in the political and theatrical life of the 
city and able to see what was going on on stage. Besides, the Eupolidean 
reference of \lfa~~alCo<JtOyapyapa in line 3 already strongly suggests a the­
atrical context, and the play's metatheatre may in fact indicate a thorough­
going allusion to other comic performances. (3) The case against an allusion to 
Babylonians even later on at 634-45 is strengthened by D. Welsh ("The 
Chorus of Aristophanes' Babylonians," GRBS 24 [1983J 137-50), who shows 
that the chorus must have been Babylonians, not allied states. The present 
argument by-passes Carawan's objection by tying the scene to Eupolis. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

KEITH SIDWELL 383 

EuDulllWV 1tPO'tEpOV 't' ~oSu, vuv 8£ llaAAOV EOTll. The two 
points of detail, the first more striking than the second (both 
noted by Kassel-Austin), are wi th Knights 75, £<popal yap o-o'to<; 
1taV't' (where o-o'to<; is Paphlagon/Cleon), and 159, (b 'twv 
'ASTlvrov 'tUYE 'trov EUDUlllOVWV (addressed to the Sausage-seller). 
Fr. 302 is a fragment from a scene in which one character is 
being asked to look at geographical locations by another: (A.) 
opw, (B.) SEW VUV MupwvDUVlUV.41 Fritzsche compared this 
with Knights 162f (PCG loc.eit.). Cf in particular 163: ~Tl. 'tae; 
O'tlXU<; Opal(; 'ta<; 'tWVDE 'trov AUrov; AA. opro, 170f: ~Tl. KUt KanDE 
'ta<; VTtOOU<; (maow; £V KUKAWl. AA. KUSOPW, and 173f: ~Tl. En 
VUV 'tOY O<PSUAIlOV 1tUpa~UAA' de; Kuptuv 'tOY DE~lOV, 'tOY D' 
E'tEPOV d<; KupXTl8ovu. The absurdity with which the basic idea 
is extrapolated in Aristophanes (ef 175 EuDUlIlOVTt0W y' d 
DWO'tPU<PTtOOIlUl) might well indicate the purpose of antagonistic 
appropriation. 42 The coincidence of the mention of do<popa at 
fr. 300 and Knights 924 might in such circumstances be 
significant, but it is hard to tell (the do<popa is connected with 
Cleon by Theopompus FGrHist 115F94). 

That Eupolis was concerned with both the Knights and Cleon 
in his plays might be indicated by the fragments of Poleis 
(whether or not it was earlier than 422 is irrelevant for this 
purpose: supra n.38). Fr. 257 cites the words avuYXl1t1tOUe; and 
avuYXl1t1tElV from Poleis. The first is explained as 'toue; avuy­
KuonKoue; t1t1t£w;. It looks as though the Knights had some 
mention in this play then, and it is interesting to note that 
Theopompus (FGrHist 115 F93) tells us of antagonism against 
Cleon on the part of the Knights, and has Cleon attacking the 
Knights' Ku'tao'tUOle; (with Fornara's emendation)43 and accus­
ing the Knights (in court?) of AEt1tOO'tpu'tOuV'twv. L Eq. 225 and 
Tzetzes ad Nub. 549a tell us that Cleon had been a t1t1tEUe; before 
4281427. Wrapped up in this fog we may perhaps discern the 

41 For the connections of Mariandynia with the Paphlagonians (and hence 
with the caricature of Cleon in Knights), see A. M. Bowie, Aristophanes: 
Myth, Ritual and Comedy (Cambridge 1993) 59ff. 

42 The fact that the same scenario and even the same formulation is used 
again at Av. 175-78 is not an argument against this hypothesis. It is equally 
possible that in the scene in Birds Aristophanes is once again poking fun at his 
rival and misappropriating the same play to do so. His originality would 
consist in (1) the newness of the plot and (2) the characters in whose mouths 
the old material now appears. The sort of attack made by Aristophanes upon 
his rivals at fr. 58 and Nub. 546£ shows that reusing the same piece of comedy 
several times was not unknown. 

43 C. W. Fornara, "Cleon's Attack against the Cavalry," CQ N.S. 23 (1973) 24. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

384 ARISTOPHANES' KNIGHTS AND EUPOLIS 

outlines of an allusion to some aspects of this dispute in the 
locution of Eupolis. Of course, it is always possible that the 
allusion is to an earlier play and/or to something outside the 
play. But here is clear evidence that Eupolis was saying 
something about the Knights in a play that may be earlier than 
Knights. 44 

The evidence that the play concerned Cleon is only cir­
cumstantial, since in forty-one fragments there is no clear 
mention of him. Comparison, however, with Knights, certainly 
a play about Cleon, shows that this need not be an impediment. 
There Cleon is mentioned only once (976).45 The play presented 
the allied cities as women coming to Athens (with their 
dowries? i.e., tribute) and being given (Athenian?) husbands (frr. 
243-47). Now it was Cleon at this period above other Athenian 
politicians who was concerned to keep a hold on the allied states 
by hard, rather than soft, means (Thuc. 3.37).46 This attitude is 
comically transmogrified in Knights 312 and 1070f in particular, 
where it is misrepresented as concern for personal gain from the 
tribute brought to Athens, itself the physical symbol of the 
allies' continued subservience. The arrival of the tribute at 
Athens was mentioned in Poleis (fr. 254). Cleon's watching­
brief over the whole Athenian empire emerges in Chrysoun 
Genos fr. 316, and this is reflected closely, as we have seen, in 
Knights 75. But it is repeated in other passages (237, 839, 1034, 
1319-some of these alluding to transfer of Paphlagon's 
privileges to Sausage-seller, others hinting, like 361, 438, 835, and 
930-40, at Cleon's propensity to get a rake-off). It is not 

44 In fro 94 Theopompus tells us that Cleon was accused of taking bribes to 
lighten the islanders' tribute and then fined because of the hybris towards the 
Knights. Unlike the other references, this looks much more like a comic scene, 
not least because Cleon was in fact extremely hard on the allies (see below). 
Pace Carawan, it does not seem that Theopompus composed his picture en­
tirely independently of comedy, even though he had other sources that evi­
denced a real conflict. An anonymous referee suggests as a possible Eupoli­
dean source for the Knights' defense against charges of effeminacy Astrateutoi 
(or Androgynoi). This is attractive, but (as often) the date is disputed. See 
Storey (supra n.7: 15ff) for dates in the 420s or 414-412. Bowie (supra n.27: 185) 
dates the play to Dionysia 426, though 425 would also be possible. 

45 This comparative evidence might suggest-if the circumstantial evidence 
from Poleis is strong enough to suggest Cleonian concerns-that Cleon was 
caricatured in the play in disguise. 

46 See R. Meiggs, The Athenian Empire (Oxford 1972) 318f on Cleon's 
policy. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

KEITH SIDWELL 385 

unreasonable to see the possibility of Poleis having been in some 
way concerned with Cleon and perhaps with the Knights toO.47 

Hence there are three types of Eupolidean material that we 
may with some justification trace from Knights to Eupolis: (1) 
the major theme of Cleon vs Knights, (2) the quotation at 1288,48 
and (3) the hypothesis that Knights may be a parody of 
Chrysoun Genos and designed as a recognizable attack upon 
Eupolis (not forgetting Cratinus, who appears to have been 
offended by it because he was insulted in the parabasis, and who 
is also attacked by name at 400). It remains to make a broad 
attempt to see what the problems of accepting this line of 
argument might be and what effect this hypothesis would have 
on our view of Knights. 

IV. Puzzles for the Paradigm 

It is worth reiterating first the chief findings so far. When 
poets reuse one another's material, they do so with antagonistic 
intent, as part of a battle at least for the prize.49 The relationship 
cannot be merely imitative for the following series of interlock­
ing reasons: (1) originality is highly prized; (2) accusations of 
plagiarism are the corollary; (3) time relationships between plays 
of known date where such accusations are involved make mere 
imitation in the circumstances of (1) and (2) highly unlikely; (4) 
the fact that the debate goes on inside comedies make it as likely 

47 Another coincidence of language between Poleis and Knights centres in 
Knights precisely upon Cleon's omnipresence and his propensity for theft. Eq. 
78f, <> 1tp<01('to<; tonv U-ll'tOXPl1l!' tv Xao(Jlv, 'too XEtP' tv At't<OAot<;, <> vou<; 0' tv 
KA<01tlOWV has as a parallel in Poleis fro 241 KUt Xuov<ov KUt TIU10V<OV KUt 
Mupoov(J)v. A final, stylistic relationship completes the meagre haul. Kassel­
Austin point to the use of the opening 00<; ... tyoo at fro 228 of Poleis and Eq. 
1100, 1107. As with Hubbard's comments (85f) on the overuse of the 
Eupolidean OO'tl<; in the second parabasis, so here we might be looking at 
stylistic misappropriation with antagonistic intent. 

48 It would make sense to suggest that this belonged to Chrysoun Cmos. 
We have no idea (except for fro 298, if that lists part of the chorus) what sort of 
chorus this play had. But if Ach. 5-8, 299ff (cf 1153) do refer back to it, then it 
may, like Marikas, have had contrasting hemi-choruses, or even two separate 
ones (Knights and old jurors). The wealth and leisure of the Knights would 
well suit the appropriation for them of the Hesiodic motif of the Golden Race 
(Th. 109-19). But equally the term would have a nice irony if applied to old 
jurors. See further on this question my forthcoming article (supra n.9). 

49 It is possible, given the political content of comedy, that there were also 
ideological axes to be ground, though it is difficult at present to differentiate 
between the attitudes of different poets. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

386 ARISTOPHANES' KNIGIITS AND EUPOUS 

(if not more so) that we are looking at jokes as at serious 
discussion; (5) authorial collaboration at any rate cannot be other 
than a joke at the expense of the poet supposedly assisted. Thus 
the speculation involved in section III was a necessary 
consequence of the discovery that the material gathered by 
Halliwell required an intertextual explanation. The coincidences 
that arise between Knights and Chrysoun Genos, then, are 
guaranteed as intended satirical relationships by the direction of 
the argument. Given this framework, it would be wrong to see 
them as trivial, though there is a problem of interpreting the 
satire that we shall face in a moment. Still less should one regard 
them as occurring because the scholia cite them as parallels for 
Aristophanic passages. 50 As a matter of fact, of the particular 
fragments we are dealing with here, none was preserved 
because of its associations with its parallels in Knights. 51 

The main problem with the proposition that Knights is 
"paracomic," rather than imitative, is that we cannot see the 
point of the satire. But if the foregoing argument is sound, we 
may justifiably ask whether we are missing some other struc­
ture vital to the function of such intertextual satire. That could 
only be something in the characters and/or plot of the para­
comedy as contrasted with the original against which its humour 
resounds. It is possible to see what difference such a structure 
would make theoretically. 

Let us suppose, for example, that in fr. 302 of Chrysoun 
Genos the speaker made to look at geographical locations had 
been Cleon. In that case, the humorous point of the reuse of 
this dramatic device would lie in the substitution of Sausage­
seller for Cleon. The humour would have been more pointed if 
(1) the earlier play also charted the rise of a new leader (Cleon), 
(2) Sausage-seller were also recognizable as a caricature of a well­
known Athenian political figure associated with the Knights. 
This hypothesis, of course, shifts the focus of interpretation 
onto what one might call 'disguised caricatures'. It may seem 
especially unsafe to support one hypothesis with another. Yet 
there are two reasons why it is worth taking this suggestion 
seriously. 

50 This point was made to me by one of my correspondents. 
51 Fr. 300 was preserved for its information on barbers' equipment (Poll. 

Onom. 10.140, 2.32; Photo Lex. p.2S0.18), not for its mention of the da<popa; fr. 
302 was preserved for the geographical datum (Steph. B)'z. p.433.S), not for the 
thematic and dramatic parallel (Fritzsche compared E q. 163f: see supra n.SO); 
fro 316 was preserved by Priscian (De metro Ter. 26) as a metrical example. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

KEITH SIDWELL 387 

First of all, the disguised caricature is already found in Slave 1, 
Slave 2, and Paphlagon in this play. It is not especially radical to 
add another character to the list. 52 Second, the hypothesis pro­
duces a structure within which one can attempt to understand 
the dynamics of both plays, if one once accepts the possibility 
that the references to Cleon and Knights in Acharnians go back 
to Eupolis, and specifically to Chrysoun Genos, rather than 
pointing back to Babylonians or forward to Knights. Eupolis' 
play will have dealt with Cleon within his favoured legal milieu 
(if Ach. 5-8, 299-302 really do imply a trial) and to the political 
advantage of the Knights. Knights in contrast sidelines Cleon 
and makes comedy out of the idea of an even viler figure (e.g. 
180f, 183ff, 213-219, 276£, etc. )-supported by the Knights­
defeating Cleon on his own ground. The play will be fundamen­
tally ironic if (1) the Sausage-seller represents a political figure 
actually supported by the Knights (and Eupolis?); (2) the play is 
presented as though by Eupolis (cf Cratinus fro 213).53 The 
paracomedy, then, on this hypothesis, is not merely a means of 
artistic criticism. It is serving the end of political satire. 

52 One might also mention: (1) that the recognition of a disguised caricature 
is given dramatic prominence in this play at 232f; (2) that the opening scenario 
contains an anti-Cleonian (Slave2/Nicias) in cahoots with a person who had 
certainly helped Cleon at Pylas (Slave lIDemosthenes), a paradoxical partner­
ship that surely produced humour because the caricatures were meant to be 
recognized before a word was spoken (as Platonios suggests [in W. J. W. 
Koster, Prolegomena de comoedia (Groningen 1975) 5] against e.g. E. W. 
Handley, "Aristophanes and his Theatre," in J. M. Bremer and E. W. Hand­
ley, edd., Aristophane [=Entretiens Hardt 38 (Vandoeuvres-Geneva 1993)] 
101£); (3) that the scholiastic tradition marked a change from "open" to 
"enigmatic" attack as beginning with the career of Eupolis (see S. Halliwell, 
"Comic Satire and Freedom of Speech in Classical Athens," JHS 111 [1991] 56 
n.9) and that this could have been a result of legislation fll) KWflWlOEtV ES 
6v6j.1Ctt~. Halliwell dismisses this idea casually (58): "we should not believe 
that the Athenians engaged in the casuistry of banning names but not 
personalites." But it is hard to see why ancient scholars should have invented 
legislation formulated thus when they believed that naming individuals was 
the primary mode of attack in Old Comedy; moreover, the disjunction 
between disguised caricatures and named attacks on the same individuals 
within Knights (Nicias 358, Clean 976) might very well be the result of 
precisely the casuistry that Halliwell scouts. There is much more to be said on 
this matter, which I intend to deal with separately elsewhere. 

53 See supra n.9 on Acharnians, where I argue that this is what is happening 
in that play. 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

388 ARISTOPHANES' KNIGHTS AND EUPOLIS 

V. Conclusion 

To go any further at present would involve an even higher 
level of speculation, unacceptable until the merits of these 
arguments have been assessed. But it is important to recognize 
that speculation will ultimately be a necessary consequence of 
testing the two structures that have been outlined here, both 
because they clearly cut across ways we currently manage our 
negotiations with Old Comedy and because they hypothesize 
the importance of the plays of Aristophanes' rivals, of which we 
know so very little. 54 I shall look in conclusion at the general 
consequences for methodology of accepting the reality of dis­
guised caricature and paracomedy. 

If, say, Sausage-seller is a caricature of a real individual, then 
we shall have to find a better method of penetrating his disguise 
than the present idea that" Aristophanes was never coy about 
his caricatures."55 It is simply not obvious who he is. By the 
same token, however, the argument that "because it is not ob­
vious, he is not a caricature," is unsafe. If it docs seem likely on 
the foregoing arguments that he is a caricature, we shall be 
forced to admit that the text was not constructed to give clues, 
but to produce laughter consequent upon prior recognition, by 
portrait-mask and other visual and aural signals. In that case, we 
shall have to accept that our philological tools must be subor­
dinated to a much wider enquiry about the humour of these 
plays, which may edge us towards accepting identifications that 
would not be arrived at by current methods. The way that the 
figure of Socrates in Clouds is constructed so as to contain both 
reflections of the individual and attributes culled from others 

54 I am not encouraged, however, by the way in which scholars have gen­
erally responded to Bowie's radical proposal that Dicaeopolis=Eupolis to sup­
pose that such testing will be allowed. I point out (supra n.9) that a sugges­
tion such as Bowie's cannot be dealt with adequately by the technique of 
measuring it against what we already know, since it inherently challenges the 
framework by which that knowledge has been determined. Hence, it can only 
be assessed by being used as a positive hypothesis. So with the suggestion 
offered here. 

55 R. F. Moonon, U Aristophanes on Alcibiades," G R BS 29 (1988) 346; cf 
Storey (supra n.3: 82): U Aristophanes does not hide his targets behind subtle 
disguises." In that case, why is there a debate about the identity of Slave 2 in 
Knights (see Sommerstein 4M)? 



SIDWELL, KEITH, Authorial Collaboration? Aristophanes' "Knights" and Eupolis , Greek, 
Roman and Byzantine Studies, 34:4 (1993:Winter) p.365 

KEITH SIDWELL 389 

should be a warning about the omnipresence of irony and 
calumny in the caricatures of Aristophanes. 56 

The consequences of accepting paracomedy are just as far­
reaching. Simple correlations between text and author cannot be 
trusted if we suspect that a whole play may be misappropriating 
the voice of another poet to produce satire. Even the referent of 
the parabasis may not be Aristophanes, but the object of his 
satire. This again destabilizes current readings. Paracomedy 
makes a global difference to the way we read and·-like disguised 
caricature-can only be dealt with by a method that accepts its 
~xister:ce a?d is p.repared to make the considerable hermeneutic 
InverSIOns It reqUIres. 

In conclusion, it must be said that neither of these tools is of 
itself either unevidenced or practically impossible. Caricatures­
both disguised and named-are a well-recognized part of Old 
Comedy's dramatic repertoire. The audiences who watched 
Aristophanes also saw the plays by other comic dramatists at the 
same festivals. They were expected to recall many details of past 
tragic plays. The parabases of Acharnians and Clouds (among 
others) make references to past comic productions by their 
own poet and others. What rational grounds have we for 
rejecting the possibility-and further for rejecting the investiga­
tion of its ramifications-that material from other comedies was 
an Important factor in the audience's understanding of these 
plays ?57 

KING'S COLLEGE, LONDON 

5T PATRICK'S COLLEGE, MAYNOOTH 

September, 1994 

56 Cf the grounds of the complaint put into Socrates' mouth by Plato (Ap. 
19c). See C. W. Willink, "Prodikos, 'Meteorosophists' and the 'Tantalos' Para­
digm," CQ N.5. 33 (1983) 25-33, for the attractive suggestion that several attri­
butes were borrowed from the popular conception of the sophist Prodicus. 

57 This piece has benefited greatly from discussion with colleagues at King's 
College, London, the Institute of Classical Studies, and the fondation Hardt 
in Geneva. I am especially grateful to Eric Handley, Bernard Gredley, Jane 
Rowlandson, and Alan Sommerstein for reading an earlier draft. It has also 
been considerably sharpened as a result of the comments of anonymous 
referees. None of these should in any way be held necessarily to agree with 
anything in the argument I have presented.