Objective Evidence for Actors' Interpolations in Greek Tragedy Hamilton, Richard Greek, Roman and Byzantine Studies; Winter 1974; 15, 4; ProQuest pg. 387

FOR R. A. PACK 

Objective Evidence for Actors' 
Interpolations in Greek Tragedy 

Richard Hamilton 

CONSIDERING ITS IMPORTANCE for textual criticism the question 
of the type and quantity of actors' interpolations in Greek 
tragedy has received little attention.! The few who have 

studied the evidence in any detail disagree substantially about its 
value. The disagreement continues today, but no attempt has yet been 
made to summarize, evaluate and harmonize the views.2 The evi­
dence is of two kinds-a very few <documents', scattered in all sorts of 
unlikely places, and fourteen or so scholia to the plays of Euripides, 
primarily Medea and Orestes. The mere distribution of the scholia 
might make one suspicious, and careful examination reveals that 

1 By 'actor' I mean anyone connected with the stage production; by 'interpolation' any 
interference with the text, not just an insertion. The most important studies of the question 
are found in: U. von Wilamowitz-Moellendorff, Einleitung in die griechische Tragodie [= Hera­
kles I] (repr. Darmstadt 1959, from the 3rd ed. 1895); W. Rutherford. Scholia Aristophanica ill: 
A Chapter in the History of Annotation (London 1905); W. Malzan. De scholiis euripideis quae ad 
res scaenicas et ad histriones spectant (Darmstadt 1908); J. Viirtheim. Aischylos' Schutzjlehende 
(Amsterdam 1928); R. Cantarella. "n testo di Sofocle" (RivIGI 9.1-2 [1925] 3-15) and 
"L'influsso degli attori sulla tradizione dei testi tragici" (RivIGI 14.3-4 [1930] 39-73). both 
conveniently found in his Seritti minori suI greco (Brescia 1970). from which I cite them; 
D. L. Page. Actors' Interpolations in Greek Tragedy (Oxford 1934); G. Zuntz, An Inquiry into 
the Transmission of the Plays of Euripides (Cambridge 1965); A. Tuilier, Recherches critiques sur 
La tradition du texte d'Euripide (Paris 1968); and J. Baumert, ENIOI A9ETOYEIN (diss. 
TUbingen 1968). Hereafter these works will be cited by author alone. Euripidean scholia 
will be cited from the edition of E. Schwartz (Berlin 1887). 

I Doubts about actors' interpolations can be found in W. Biehl, "Die Interpolationen in 
EUripides' Hekabe," Philologus 101 (1957) 66 and 68 n.2; A. M. Dale, Euripides. Helen (Oxford 
1967) xxxii; and C. W. Willink, "A Problem in Aeschylus's Septem," CQ 18 (1968) 8 nn.3 and 
5. Generally, however, endorsement of the idea is wholehearted: W. Ritchie, The Authen­
ticity of the Rhesus of Euripides (Cambridge 1964) 39; OCD2 (1970) S.V. SCHOLARSHIP, GREEK, 
IN ANTIQUITY; J. M. Bremer, "Euripides' Hecuba 59-215: A Reconsideration," Mnemosyne 24 
(1971) 243; M. L. West, Textual Criticism and Editorial Technique (Stuttgart 1973) 16; M. D. 
Reeve, "Interpolation in Greek Tragedy, ill," GRBS 14 (1973) 171; and, most positively, 
W. S. Barrett, Euripides, Hippolytos (Oxford 1964) 46. In all of these Page's is the only study 
mentioned. 

387 



388 OBJECTIVE EVIDENCE FOR ACTORS' INTERPOLATIONS 

quite a few of them are speculative, guesses by scholars rather than 
verifiable fact. Those that are factual, as we shall see, have no effect on 
the text but rather record actors' gestures and phrasing. In any case 
both speculative and factual scholia attest at most only minor changes. 
The documents, on the other hand, attest wholesale revision of plays. 
The fact that we have no clear examples of such suggests that the 
acting texts, in which they must have been recorded, had little or no 
effect on the manuscript tradition as compared with the booksellers' 
texts. This interpretation is implicit in the work of earler critics and 
can, I think, be clearly demonstrated from the material itself as well. 

I 

Wilamowitz's Einleitung (1889) contains only a few paragraphs 
devoted to the question of actors' interpolations, but they are crucial. 
He balances trust in the literary tradition with the realization that 
actors' interpolations did exist. Book form protected the tragedies 
from corruption (p.130). Actors, however, inevitably made changes 
and the Lycurgan law restricting them was ineffectual, so that when 
our texts depend on actors' copies their reliability is small. Nonethe­
less, the condition of our extant plays is quite good (pp.132-33). 

In regard to the actual scholia, Wilamowitz (p.162) is critical, in 
particular, of Didymus, the great compiler of scholia in the first 
century B.C., who is mentioned in several scholia concerning actors' 
interpolations (schol. ad Med. 169, 356): Her hat von ihrer Tatigkeit 
weder eine klare Vorstellung, noch gibt er sich die Mlihe die Vorwlirfe, 
die er gegen sie richtet, zu beweisen. Er braucht die Schauspieler 
vielmehr .. , als Deus ex machina urn kritische Knoten zu durch­
hauen." Wilamowitz is critical, in general, when the wording of the 
scholium is vague or a scholarly question can be seen to lurk in the 
background.3 Yet this scepticism does not extend to the hypothesis of 
Rhesus (p.13l) or to the scholium at Orestes 1366 (p.154 n.63) even 
though, as we shall see, the language in both places is circumspect 
about the possibility of actors' interpolation. Nor does he question the 
statement by Aristophanes of Byzantium about the use of an ekky-

• That is "wo der Grammatiker durch ein tcwc selbst eingesteht. dass fUr ihn das Drama 
Dieht mehr auf der Biihne existiert ... Auch wenn iiber das Umkostiimieren geredet wird. 
ist die Verkehrtheit der Bemerkung Beweis genug. dass das am Schreibtisch ausgedacht 
ist" (p.154 n.64). 



RICHARD HAMILTON 389 

klema at Hippolytus 171, even though the play itself seemed to con­
tradict this (p.154 n.64). 

Wilamowitz's trust in Aristophanes stems from his belief that 
Aristophanes, living in the third century B.C., unlike Didymus in the 
first, saw plays performed and that all the factual (as opposed to 
speculative) scholia concerning actors go back to him. The evidence 
for this belief is the scholium to Orestes 1287, where Aristophanes 
defends a reading by quoting Stesichorus, and so, Wilamowitz argues, 
when Stesichorus is quoted at Orestes 269 it is likely that Aristophanes 
is responsible for this scholium also and, since this scholium is con­
cerned with actors, Aristophanes must have been concerned with 
actors. The reasoning is ingenious but not totally satisfying. None of 
the five other citations of Stesichorus mentions either Aristophanes or 
actors,4 nor is it certain that Aristophanes and Stesichorus are to be 
joined in the scholium to Orestes 1287: in the first place a quotation 
from Homer intervenes and, secondly, Aristophanes was concerned 
with the grammatical subject (and consequently the number) of the 
verb, whereas Stesichorus is cited for the idea that Helen's beauty 
overpowered her attackers, an idea which is in the play regardless of 
the verb's number and subject.5 Finally, Aristophanes never mentions 
actors and those who do are Didymus and his contemporaries, 
Apollodorus of Tarsus (schol. ad Med. 148= 169) and Philoxenus 
(scho!. ad Phoen. 264).6 

Rutherford in his four-page treatment (1905) is more thoroughly 
sceptical; the language of the scholia gives them away as guesses (p.59) : 
<CAt Med 169, the note reveals that nobody had thought of solving a 
notorious ~~'T7Jf.La by supposing that it was the creation of the actors 
until a scholar in the first century B.C. suggested that the difficulty 
might be got over by this assumption." Yet, in the end, Rutherford 
concedes that the stage tradition may have been responsible after all 
for the obvious corruptions in our texts. 

, Elsewhere Stesichorus is one of a series of authors attesting a point of mythology. A 
fact that Wilamowitz might have used to advantage is that of the seven citations of Stesich­
orus in the scholia, four are to this play. 

S A stronger argument for connecting Aristophanes and actors is the high density in 
Orestes scholia of references to him (schol. ad 488,713, 1038, 1287) or his pupil Kallistratos 
(ad 314, 434, 1038) and to stage business (ad 57, 176,643, 1366, cf. also the hypothesis). 

6 Philoxenus and Didymus were contemporaries (R. Pfeiffer, History of Classical Scholar­
ship [Oxford 1968] 274-75), Apollodorus presumably a bit earlier (Rutherford p.59; F. 
Susemihl, Geschichte der griechischen Literatur in der Alexandrinerzeit II [Leipzig 1891-92] 178). 



390 OBJECTIVE EVIDENCE FOR ACTORS' INTERPOLATIONS 

Malzan in his dissertation on scenic scholia (1908) is only tangentially 
concerned with actors' interpolations as such. His main purpose is to 
carry on the work of Trendelenburg and Weissmann by showing that 
not only the scholia criticizing Euripidean dramatic technique but 
also those dealing with stage business depend on Peripatetic rhetori­
cians (p.32).7 He finds a close connection between criticism of actors, 
precepts for actors and, ultimately, the use of actors to solve textual 
problems (p.34). The scholia involve all three. 

Viirtheim's purpose in his appendix (1928) is simply to justify 
speaking of the possibility of actors' interpolations in Supplices. 
He first catalogues the evidence for actors' interpolations and then 
argues for their existence in Aeschylus. Although he is uncritical of all 
but one of the scholia (Med. 169), he carefully points out when the text 
actually remains unaffected or only slightly altered. 

Cantarella at first (1925) thought that the manuscript tradition was 
based on stage texts rather than book texts (p.306), but he became 
more doubtful when, in 1930, he examined the actual evidence (p.149): 
uStando cOSI Ie cose, una critica troppo sottile potrebbe, se non 
negare del tutto, ridurre abbastanza l'attendibilita di tali testimoni­
anze." His discussion of the scholia themselves is generally perfunctory 
but he does point out two problems that others have ignored: the 
usual explanation of why Euripides alone attracts such scholia, that he 
alone was frequently revived (see e.g. Rutherford p.57, Tuilier p.26 
n.2), is insufficient.8 The second problem is that the scholia invariably 
speak of'the actors' rather than <the actor' (or even a proper name), 

7 K. Weissmann. Die szenischen Anweisungen in den Scholien zu Aischylos. Sophokles. Euripides 
und Aristophanes und ihre BedeutungfUr die Buhn.enkunde (Bamberg 1899). collects quite a few 
passages "im Anschluss an spiitere Auffiihrungen" where "Vortrag oder Spiel getadelt 
wird" and suggests they go back to Apollodorus of Tarsus. but he discusses only five. very 
briefly (pp.33-34). His general intent. following A. Trendelenburg. Grammaticorum grae­
corum de am tragica iudiciorum reliquiae (Bonn 1867). is to show the theoretical. philosophical 
nature of the scholia. in particular their dependence on Aristotle's Poetics. 

8 Cantarella argues that Euripides had lost the dramatic sense the other two tragedians 
retained by being actors themselves (p.152) and that Euripides was much more popular in 
the schools and therefore attracted more scholia (p.l48). The relative popularity of Euripi­
des is a commonplace (see e.g. Schmid-Stahlin, GGL 1.2.60 n.3, 1.2.507,1.3.824 n.3). Sophocles. 
however, was quite popular as well (F. G. Welcker. Die griechische Tragodie [RhM suppl. 2, 
1841] 912; A. Pickard-Cambridge, The Dramatic Festivals of Athens- [Oxford 1968] 100; 
B. E. Donovan, Euripides Papyri [Toronto 1969] 6). P. Orsini (REA 38 [1936] 110) asks perti­
nently why the popular Ion and Alcestis were not also victimized (of course we have no 
scholia to Ion). 



RICHARD HAMILTON 391 

and this might lead one to suspect the scholia were not first hand 
(p.149). 

Page's book (1934), by far the longest treatment of the subject, is 
disappointingly uncritical in its approach. His general thesis is that 
the manuscripts depend on actors' copies and are therefore terribly 
corrupt. The bulk of his book is taken up divining where the corrup­
tion lies. Defense of his position is confined to a short and rather late 
chapter (pp.I06-12) and is preceded by only the most dogmatic 
statements of his view. His acceptance of the scholia is total. 

After a brief reply to Rutherford,9 Page outlines his argument: (1) 
actors' interpolations exist as we know from explicit (schol. ad Phoen. 
264) and implicit evidence (interpolations in Heraclidae, Iphigenia at 
Aulis, Orestes and Phoenissae, p.107). Page does not raise the possibility 
that some of these presumed interpolations could be caused by men 
other than actors. (2) Our manuscripts derive directly from Aristoph­
anes and Aristophanes visited the theatre, and so our manuscripts 
are from the theatrical tradition (p.107). Page here depends on 
Wilamowitz for his evidence (which we have shown cause to doubt) 
but ignores his conclusion, that book texts rather than stage texts 
were the main carrier of our manuscript tradition. (3) "The publisher 
who made a book wrote what the actors said at the time, and so 
perpetuated the actors' alterations in the written texts which were 
soon to lie open before Aristophanes" (pp.108-09). This compromise 
between his second argument and Wilamowitz's position ignores 
several problems: (a) there is no reason to assume that it was the 
actors' copies rather than the autograph that provided the exemplar 
for the booksellers ; (b) even if the first book was dependent on 
actors' copies, both book and performance were presumably con­
trolled by the poet (Wilamowitz p.128, Viirtheim p.231, Cantarella 
p.138);10 (c) subsequent texts, then, could come from two sources and 
it is likely that they came from the book tradition rather than the 

I Page makes one general criticism: "we need not take the Scholiasts seriously, in the 
sense that we may write them off as murmuring pedants; but we are bound to take their 
charge seriously, in the sense that we must earnestly scrutinize its relationship to our 
manuscript text" (p.106). Yet this is precisely what Rutherford (in a rather abbreviated 
fashion) has done and what Page leaves undone. His particular criticism is also peculiar: 
the scholium to Med. 169 "only proves that its compiler derived his information from a 
note by Apollodoros; it does not say whether Apollodoros merely gave his own view or 
based his comment on older material" (p.107). 

10 Strictly speaking, interpolation requires a preexistent text, which would be the text 
of the first production. 



392 OBJECTIVE EVIDENCE FOR ACTORS' INTERPOLATIONS 

acting tradition for (d) the book trade would reflect a more contin­
uous and probably greater demand for the texts than would stage 
production, which in terms of the whole tragic corpus was inter­
mittent and highly variable;l1 and (e) the actual form of actors' texts 
may well have been unsuited to reading. 

A reconstruction such as Page's requires that actors' interpolations 
not post-date Aristophanes.12 But there are two problems with this 
thesis. First, as noted above, the only names we find in the scholia are 
much later: Didymus, Philoxenus, Apollodorus of Tarsus. To meet 
this difficulty Page is content with Wilamowitz's pronouncement on 
Didymus (p.lIO), not recognizing that this course requires elimination 
of much of his evidence. There is also the problem of the phrasing in 
the scholia, Hthe actors now." Page argues that <now' does not show 
that the scholium was late but simply that it was a verbatim quotation 
from an earlier source (p.ll0). Still, this argument does not eliminate 
the problem, raised by Cantarella, of the generalizing <actors'. 

Page has some positive arguments as well: (1) Hnotes about actors' 
interpolations are not strictly relevant to any commentary save that 
which explains a textual recension; and the only great textual 
recension was that of Aristophanes" (p.109). Yet (a) most of the notes 
are not relevant to any commentary at all, as we shall see; (b) it is 
unlikely that Aristophanes wrote a commentary to accompany his 
text;13 (c) it is not certain that Aristophanes' was Hthe only great 

l1 Page's view requires, as Orsini has pointed out (op.dt. [supra n.8] 111), "que la trans­
mission des textes tragiques au moins pour les pieces a succes, fut, jusqu'aux Alexandrins. 
surtout orale." This is unlikely in view of the common use of books as early as the fifth 
century (see Pfeiffer, op.dt. [supra n.6] 27ff; F. D. Harvey, "Literacy in the Athenian Demo­
cracy," REG 79 [1966] 585~35). So G. Jachmann ("Binneninterpolation n," GiittNach 1 [1936] 
193 n.1) : " alles in allem iiberwiegt in unserer 1Tap&&CLC doch bei weitem die rein literarische 
Interpolation, etwa im gleichen Verhaltnis wie die Tragodie, auf die Gesamtzeit gesehen, 
starker, Hinger und kontinuierlicher als Lesebuch denn als Biihnenwerk fortgelebt hat." 

11 That they should date from 400-200 B.C. is accepted by all: Cantarella 142ff, 172f; 
A. Lesky, DLZ 57 (1936) 2205; F. Solmsen, CR 49 (1935) 131; Tuilier 26; Barrett 46; Ritchie 
34; West 16 (the last three cited supra n.2). 

13 pfeiffer, op.dt. (supra n.6), is firm on this point (pp.173, 190,212), but Page's assumption 
finds some support in scho!. ad Tr. 47: "the line is marked [since it sounds] as if Troy were 
no longer inhabited, for Aristophanes suspected [this line] on the basis of II. 20.307," which 
suggests that the mark was geared to a commentary. Likewise a scholium to Pindar, 01. 
2.48c Drachmann: "Aristophanes athetizes this colon for he says it is metrically superfluous 
in regard to the antistrophe" (see also the scholia to Od. 23.296 and Hipp. 171). E. G. Turner, 
Greek Papyri (Oxford 1968) 116-17, notes that the mere presence of signs suggests a com­
mentary, but unfortunately the examples he offers are all much later than Aristophanes. 



RICHARD HAMIL TON 393 

textual recension"14 and, even if were, this does not preclude other 
editions or other commentaries to his edition. (2) "The existence of 
the substitute [for the unpronounceable EKCPPWC' at Phoen. 264] and its 
universal acceptance can only be explained by the assumption that it 
is very old indeed" (p.ll0). This scholium will be discussed in detail 
shortly, but one should note here that the opposite argument is just 
as easy: the scholiast maintains that the actors said OU JLEOWC' because 
they could not pronounce OVK EKCPPWCLV; since it is unlikely that 
Euripides would write what his actors could not pronounce, the 
actors in question must be performing some time later, when pro­
nunciation has changed. (3) The actors' interpolations are too good to 
be late (p.ll0). This might suggest to the sceptical that they are not 
interpolations at all. (4) Histrionic accidents exist in the texts, and the 
only way to explain this is to assume that a prompter's copy was 
used as a source for our manuscripts (p.ll!). Yet the existence of 
prompters at this date is unlikely.I5 Moreover, most of the examples 
of histrionic accidents that are clearly such, i.e. mispronunciation, do 
not affect our manuscripts. The exception, Phoenissae 264, could hardly 
be called <accidental'.16 

U The evidence of Aristophanes' impact is the uniformity of the colometry all the way 
back to its introduction (by Aristophanes) and the uniformity of the texts after Aristophanes 
as opposed to their wild fluctuation before. The evidence for the former, however, is 
exiguous (see Pfeiffer, op.cit. [supra n.6] 185ff), and the evidence for the latter has recently 
been questioned (see Barrett, op.dt. [supra n.2] 56 n.l, and Donovan, op.dt. [supra n.8] 24, 
versus Turner, op.dt. [supra n.l3] 107 and n.14). Page's statement that "the wide publication 
of annotated texts ensured a high degree of stability" (p.7) ignores the fact that hypom­
nemata were published separately from the texts they discussed until the codex form 
allowed space for their inclusion (Pfeiffer 218; for some exceptions see N. G. Wilson, "A 
Chapter in the History of Scholia," CQ 17 [1967] 244-56). On the whole question of actors' 
interpolations, Page is much more restrained in "Some Emendations in Aristophanes' 
Acharnians," WS 69 (1956) 127. 

15 The possibility of a prompter is discussed in some detail by P. D. Arnott, "The Dis­
sociated Actor," in Greek Drama, ed. G. L. Beede (Vermillion, S.D. 1967) 40-51. The actual 
mechanics to get us from improvisation to textual interpolation is Page's greatest weakness 
(so Lesky, op.dt. [supra n.12] 2206). 

11 One would expect Page to start with attested actors' interpolations (i.e. mispronuncia­
tion), but a few pages earlier (lOOff) we find that the "accidental interpolations" he 
discusses are not only unattested but also quite different from the attested accidents of 
mispronunciation: substitution of a word or line for another from a similar context; 
repetition of a line from a different context. The scholium to Med. 84 should have been a 
sobering reminder how hard it is for an actor's version to infiltrate the MS tradition: the 
reading preserved in this scholium shows no transferral of either TOIYrO or ca.pwc from actor 
to text, and the actor's version of the disputed form i'tyv<fJclm( c) is found only in one MS and 
in corrections of two others. The actors' reading preserved in the scholium to Med. 910 had 



394 OBJECTIVE EVIDENCE FOR ACTORS' INTERPOLATIONS 

Zuntz (1965) clearly believes in actors' interpolations but only as 
one of many sources of textual corruption (p.252). He finds many of 
the scholia speculative and actors used as "mere whipping boys" in 
scholarly arguments (p.254 note). 

Tuilier (1968) exhibits the same paradoxical attitude found in 
Wilamowitz, Rutherford, Cantarella and Zuntz. The idea of actors' 
interpolations is attractive (p.26); the evidence, however, is weak 
(p.222). 

Baumert's aim in his extended discussion of the scholia concerning 
actors' interpolations (1968) is to show that they, like all other external 
evidence beside manuscripts, are of no value. His method is to argue 
either that the text in question is perfectly satisfactory and therefore 
the scholium must be wrong or that, since there is no distinction 
between reading and acting texts, actors are invoked only to explain 
variations in the manuscript tradition. Such circularity, unfortunately, 
mars his thorough, well-documented and often penetrating analysisP 

II 
Everyone believes that actors changed their texts in performance 

(it would be impossible not to), but there is little agreement about 
how many of these changes can be discerned in OUT texts. A chart will 
make this immediately apparent (x indicates that the scholar heading 
the column was sceptical of the scholium's factual basis, 0 that he was 
not):18 

no effect whatsoever on the text. One can argue the same for the scholium to Med.228 :all 
the MSS read 'Y'YV~Cl(t"W, but the scholium suggests that this was a misreading by the actors 
("the actors not understanding the turn of phrase say 'Ytyvc.:,CI(££V I(MWC"). The result has 
been, as J. Kaiser notes CBemerkungen zu Euripides' Medea," Gymnasium 66 [1959] 503), 
that the MS reading has been emended "mit jeder der drei Person en von 'Y,yvc.:,CI(f!£v." Yet, 
Kaiser argues, 'Y,yvc.:,Cl(w' makes perfectly good sense, so why change it: "warum soU der 
Fehler nicht vielmehr im Scholion stecken und hier zu andern sein 7" See also O. Regen­
bogen, "Randbemerkungen zum Medea des Euripides," Branos 48 (1950) 26: "seine Mei­
nung durchaus unklar ist;" and Baumert, who concludes that interpolation by actors is 
possible but not compelling (p.95). 

17 For Baumert's faults of logic see M. D. Reeve's lengthy critique, "Interpolation in 
Greek Tragedy, I," GRBS 13 (1972) 247-65, and infra n.30. 

18 The question marks and parentheses should be explained: Wilamowitz's condemna­
tion of Didymus would presumably include Med. 169 and 356. Malzan says on Med. 
148/169, "quod sane factum esse putandum est non nisi post saeculum quintum chori 
partibus neglectis" (p.26); on Med. 228, "quid de histrionum culpa iudicandum sit equidem 
ignoro" (p.27); on Or. 643, "si vere rettulit ille [criticus]" (p.17); and on Phoen. 264, "si re 
vera ab histrionibus orta est" (p.Z4). Viirtheim thinks that at Med. 910 "fanden die Schau-



RICHARD HAMILTON 395 

REFERENCB WILA- RUTH- MAL- VURT- CANTA- PAGB ZUNTZ TUI- BA.U-
MOwrrZ BRFORD ZAN HBIM RELLA. LIER MERT 

Andr.7 X X 0 0 0 X X 
Hipp.l71 0 X 0 (0) 0 0 

Med.84 X 0 0 0 0 X 0 X 
Med. 148/169 (x) X x? X 0 X X x? 
Med.228 X ? 0 0 0 X X 
Med. 356/380 (x) x X 0 0 0 X X ? 
Med.910 0 0 07 0 0 x X 

Or. 57 0 0 0 0 0 0 0 
Or. 268 0 0 0 0 0 0 0 0 
Or. 279 0 0 
Or. 643 0 0 o? 0 0 0 0 0 
Or. 1336 0 X x 0 0 0 x x x 
Rhes. hypoth. 0 X 0 o? 0 0 0 
Phoen.264 0 07 07 0 0 X X ? 

No scholium is doubted by all and only three are accepted by all. 
If we separate the scholia into two categories, however, those dealing 
with stage business or mispronunciation and those dealing with a text 
problem, a clear pattern is discernible: 

STAGE BUSINESS OR MISPRONUNCIATION TEXT PROBLEM 
REFERENCE X 0 REFERENCE X 0 

Hipp.l71 1 5 Andr.7 4 3 
Med.84 3 5 Med. 148/169 7 1 
Med.910 2 5 Med.228 3 3 
Or. 57 7 Med. 356/380 5 3 
Or. 268 8 Or. 1336 4 4 
Or. 279 2 Rhes. hypoth. 1 6 
Or. 643 8 Phoen.264 2 5 -Total 6 40 Total 26 25 

spieler einen korrupten Satz" (p.234), that concerning Phoen. 264 "die Aenderung war 
jedenfalls keine Besserung" (p.239), and that the second prologue given in the hypothesis 
to Rhesus is prosaic "aber Euripides hat mehr solche geschrieben" (p.233), although the 
existence of another variant prologue suggests that this is "Schauspielermachwerk." Page 
does not discuss Hipp. 171, but his citation of Wilamowitz implies his acceptance of it. 
Tuilier (59 n.4) follows the older reading (B) in Hipp. 171 EYKAT//La: "n s'agit ici d'un rep roche 
precis qu' Aristophane de Byzance adressait a la mise en scene, et non de la machine 
scenique." Aristophanes, he says, is correct: Phaedra should be inside. Baumert thinks that 
Apollodorus had a paragraphos in his text which led him to his false conclusion about Med. 
148/169 (p.75), that Med. 356/380 records a transposition, not an interpolation, but actors' 
interference cannot be ruled out (p.88), and that Phoen. 264 is not a certain example of 
actors' interference (pp. 93, 269). Here as elsewhere I am not considering speculative stage 
business, i.e. the frequent stage directions deduced from the text (schol. ad Or. 225, 1567, 
1573, Med. 96,97, 1317. Hipp. 215, Tro. 98). It has long been recognized that these are the 
work of'armchair' critics (see supra n.7). To the list in my table might be added the arrivals 
via mechane described in the scholia to PV 397 and Med. 1320 and the placement of Electra at 
the foot, not the head of Orestes' bed in the Orestes hypothesis (see now V. Longo, "L'Hypo­
thesis di Aristofane di Bisanzio e Ie posizioni di Elettra nella parados dell' Oreste euripideo," 
Dioniso 41 [1967] 39()-97). Scholia such as these sometimes clearly refer to revivals (Or. 57, 
268) and perhaps they always do. 



396 OBJECTIVE EVIDENCE FOR ACTORS' INTERPOLATIONS 

Scholia concerned with stage business or mispronunciation are 
questioned only 15 % of the time, while those concerned with text 
problems are questioned 50 % of the time. The logic of this distinction 
is simple and was expressed in some form by Wilamowitz, Ruther­
ford, Malzan, Viirtheim, Zuntz and Baumert: stage business and 
mispronunciation do not reside in the text alone and so a scholar 
looking only at the text would not be apt to invent them, whereas 
text problems might lead him to invent a solution involving actors.19 

The criteria for deciding when we are dealing with a scholar's inven­
tion are (1) presence in the scholium of language which makes it clear 
that it is an hypothesis, or (2) discussion of a problem (&.7Topla, '7jT7}p.a) . 
to which an appeal to the actors is one solution.20 

By these criteria, all but two of the scholia dealing with text prob­
lems can be seen to be inventions, and even these two are arguably 
SO.21 

Four notes guess that actors were responsible for a problem in the 
text. This is clearest in the hypothesis to Rhesus, where the problem 
is that there is evidence of two versions of the prologue not found in 
the manuscripts; the speculative solution is advanced that «perhaps 
some of the actors might have reworked it."22 Similarly in the scholium 
to Orestes 1366, although the problem is not explicitly stated, it is clear 
that a contradiction was felt between the Phrygian's described leap 
from the roof and his (described and visible) exit from the door. The 

11 One must then discount the few critics of such scholia: Zuntz gives no reasons for 
doubting Med. 84 and 910 nor does Rutherford for doubting the former. Baumert thinks 
that the rejection of a variant reading in both cases led to the false assumption of actors' 
interference, not noticing that the actors' versions in either case could not be derived from 
our MS tradition (see supra n.16). Malzan's distrust of Hipp. 171 is certainly understandable, 
but it is hard to imagine what would induce Aristophanes to read an ekkyklema into those 
lines other than something like a stage direction that later dropped out of the text (so 
Weissmann, op.cit. [supra n.7] 26). "It is unwise to reject Aristophanes of Byzantium's 
testimony," said T. B. L. Webster, CR 13 (1963) 33. 

10 For a similar phenomenon in Pindaric scholia see H. Fraenkel, "Schrullen in den 
Scholien zu Pindars Nemeen 7 und Olympien 3," Hermes 89 (1961) 385-97. 

II For Med. 228 see supra n.16. 
II The scholiast may be correct: faced with a defective text an actor would have to 

invent while a scholar would not. The ancient evidence is inconclusive: we know from the 
hypothesis that the play, without any prologue, was read by Aristophanes and that the 
first prologue was read by Dicaearchus, a pupil of Aristotle; but there is no assurance that 
Dicaearchus did not know of the other prologue or that Aristophanes was ignorant of 
both. To use this as an example of an actor's interpolation means agreeing (1) that the 
play had a prologue to begin with; (2) that this original prologue was removed by an actor 
and not simply lost; and (3) that it was replaced by a prologue written by an actor. 



RICHARD HAMILTON 397 

solution is again speculative: "one would readily agree that these lines 
[about the door exit] were not from Euripides but rather from actors 
who come out of the door so that they will not be hurt leaping from 
the palace."23 The language is less hesitant in the scholium to Medea 
169 (= 148), but the situation is dearly the same: "this too is one of the 
celebrated ~7Jn}l-'aTa: how Medea calls on Themis and Artemis, but 
the nurse says that she called on Zeus, not Artemis, as witness. Now 
Apollodorus of Tarsus says that the actors are the cause of the am­
biguity by running together the choral part and Medea's words." 
Didymus disagreed, rightly as the meter assures us. In the notes on 
Medea 356 and 380 we find that Didymus thought that v.380 should be 
placed after v.356, not v.379. It is not clear what bothered him, but 
once again the solution is criticism of the actors: "he blames the actors 
for their poor arrangement."24 

To these four clearly speculative references to actors' interference 
may be added the scholium to Andromache 7: ot (J7TOKptTa~ TOV tal-'/3ov 

'f) • I "",..~ 5W. I "'''' tf •• ~ 5W. TTpOCE 7]Kav VTTovo7JcavTEc ELvat T7Jv ypv"'f"lv VII Ttc (.V TI OVTWC: vvv VII 

I """ " ,~ ~ '''~ '" [5W. I Ki h ff TtC al\l\7] Kat aVT£ TOV CVYKp£nKOV TO OVCTVXECTaT7J VII T£C rc 0 , 

i}T£C 0 A om. M, 1'V1' o~nc A, vV1' 0' i}nc M, vVV 0' Etnc 0]. The implica­
tion is that the actors were wrong, that the text should not read 0T] 
TtC (ori}T£c); indeed most of the manuscripts read 0' E'nc. Yet two read 
ounc, and so there is clearly a textual problem distinct from anything 
the actors may have suggested. At least as far as v.6 is concerned, then, 
it looks as if the actors were simply reacting to a corrupt text. One 
should note also that the actors' &] Tic (or ijnc) is not reflected in the 
manuscripts. The last part of the scholium only adds to the confusion: 
it could mean that the actors substituted the superlative for the 
comparative or that they took the superlative as a comparative. If 
the former, there would be no reason to add v.7;25 if the latter, then 
we must assume the actors both added a line and changed the existing 

23 It is now generally recognized that there is no contradiction; see Baumert 76fffor the 
literature, to which may be added Dale, op.cit. (supra n.2) xxxii, and J. Roux, "A propos du 
decor dans les tragedies d'Euripide," REG 74 (1961) 28ff. 

2' The usual assumption that 356a is a conjecture by Didymus (see Baumert 86-88) is 
attacked by Baumert on the grounds that when the scholiast says "it is well placed here. 
Didymus marks it because the actors arrange it poorly," the scholiast is not opposing 
Didymus since there is no S.[ separating them. Moreover, Didymus' mark would not 
reflect a textual variant but "eine Besonderheit"; but see Turner, op.cit. (supra n.l3) 116f. 

25 And so the latter interpretation is always accepted, even though the elimination of a 
textual basis for the scholiast's inference about the actor's interpretation considerably 
weakens his objectivity, 



398 OBJECTIVE EVIDENCE FOR ACTORS' INTERPOLATIONS 

context to fit it in.26 Further doubt of actors' interference is raised, as 
Baumert notes (pp.81-83), by the scholium to v.6, in which v.7 is 
attacked without reference to actors. It looks, then, as if v.7 was in the 
text to begin with and that textual corruption led to questioning of it. 
To accept Andromache 7 as an example of actors' interpolation means 
agreeing (1) that the actors inserted a line into "un contexte en soi 
tres clair" ;27 (2) that to accomplish this they altered the surrounding 
context with not one but two grammatical peculiarities (&J, super­
lative for comparative); (3) that only one of the three changes, the 
insertion of v.7, had any effect on the manuscript tradition even 
though this renders the context incomprehensible; and (4) that the 
actors' role in this was ignored by all but the scholiast to v.7. 

We are left, then, with one note, Phoenissae 264, that clearly attests 
an actual change in the text by actors, and it is on this note that Page 
most depends. But here too there is cause for doubt. First there is the 
variation in the scholium itself. The accepted version is, "the text is 
OUK EK4>pWctV; the actors changed the phrase on account of difficulty 
in pronunciation. Philoxenus in his <About Monosyllables' in talking 
about 4>pw cites this use." One of the two manuscripts (A), however, 
reads ol /LEV "J.p, "some say OUK EK4>pWCtV ••• ," and if this is closer to 
the original form of the scholium (no other explanation for its pres­
ence forthcoming) we might reconstruct the earlier form in this way: 
"some say OUK EK4>pWCtV but others say OU IU(}wc,v; the solution is that 
the actors introduced the latter to replace the former which they 
could not pronounce."28 A more substantial cause for doubt concerns 
what Philoxenus would have been likely to say; if he explicitly 
quoted this line as from Phoenissae and actually said that the actors had 
changed it, one would be bold to doubt his word. But Philoxenus 
may merely have cited among his examples of 4>pw a line identical to 
Phoenissae 264 except for having EK4>pWC' instead of /L£(}wc'; someone 
noted the resemblance and resolved the discrepancy by reference to 

II Baumert notes (p.82) that 8-rj does not fit with v.5. Reeve, op.cit. (supra n.17) 261 nAl, 
eases the difficulty here considerably by positing the insertion and later disappearance of 
the comparative. 

17 So A. Garzya, "Quelques notes sur l'Andromaque d'Euripide," RBPhil29 (1951) 1143, 
who argues that therefore v.7 is clearly interpolated; but this is precisely what should 
make us suspicious of such a solution. See P. T. Stevens, Andromache (Oxford 1971) ad loc., 
and Reeve, op.cit. (supra n.17) 261, "there is no clearer case in Euripides of an interpolation 
designed to obviate an elliptical construction." 

18 The ,,&'p that introduces the variant reading suggests that even more must be supplied. 



RICHARD HAMILTON 399 

actors.29 If so, we have no text changed by actors but yet another 
example of speculative actors' interpolation.30 The arguments in 
favor of the latter are: (1) This is consonant with all other actors' in­
terpolation scholia involving textual changes. (2) It is not likely that, 
in fact, the word was difficult to pronounce (cf €Kcppiw, €KCPPWV LS]).31 
(3) Philoxenus was interested in usage, not pronunciation; it is there­
fore doubtful that he did more than quote the line without attribu­
tion.32 The scholium itself suggests no more. Thus we find in Photius' 
lexicon OilK €K~PWC'JJ· OVK €ga~wcL· EO~OKAfjc: that is, OVK €K~pWCLJJ, 
found in Sophocles, means OVK €g~wn. This may well have come 
from the same spot in Philoxenus but with a different attribution 
eventually attached. There are parallels for the insertion of material 
from Philoxenus' very popular book into scholia via lexica (see C. 
Wendel, RE 20 [1941] 197-98 S.v. PHILOXENOS). (4) Philoxenus was 
contemporary with Didymus; both lived during the height of the 
mania for speculative actors' interpolations. This increases the likeli­
hood that a false ascription on the basis of his reference would be 
explained by reference to actors' interpolations.3s 

20 Page collects (and tries to eliminate) a number of such 'echoes', twenty-five in Eurip­
ides alone (pp.104-05; he adds four more in a note on p.220), most of which involve the 
change of only a word. 

30 Baumert's reconstruction (pp.91-93) is similar, but it rests, as usual, on ill-defined and 
circular arguments: iKCPPWC' must be a rejected variant since "it is hardly possible that an 
expressly rejected variant could replace the vulgate reading." Therefore there is no 
division into reading texts and acting texts. Neither of the two possible sources for the 
variant, a Kollationsexemplar or Philoxenus, suggests actors, and so there is no sure proof of 
actors' influence. 

31 "Le temoignage de Philoxene d'Alexandrie repose certainement sur une conjecture 
de philologue, puisque Ie verbe iKcpplw est un mot tardif qui est:inusite dans Ie vocabulaire 
d'Euripide ... Il est donc evident que l'auteur de l'archetype conserve une grande inde­
pendance a regard des notes de philologues qu'il joint au texte d'Euripide" (Tuilier p.223). 
Baumert, on the other hand, says (p.92) "sprachlich sind beide Worter in dem hier von 
Euripides angewandten Jagbild moglich." For a thorough study of the verb see Barrett, 
op.dt. (supra n.2) ad Hipp. 866-{)7. 

32 In a scholium to Dionysius Thrax (Gramm. Graee. 1.3 p.247. 21-22 Hilgard) Philoxenus 
is cited as saying that cPpw is the 'vivid' word for 7TpoLlvaL. 

83 The limited influence of the theory of actors' interpolations is suggested by the 
paucity of places in which such types of errors are so explained. Thus at Med. 169 it is a 
question of misattributed lines, but the same problem arises without mention of actors at 
Med. 520f, Hipp. 58,776, and Or. 140. In Andr. 7 and the Rhesus prologue actors are accused 
Qf haVing added their own lines, but there are many places where lines are added (or 
removed) and the actors are not accused: Ale. 820, Andr. 1254, Hipp. 871, 1050, Med. 87, Or. 
957, 1394, Phaen. 973, 1075. See Baumert 269ff for a variety of explanations of interpolations 
not attributed to actors. 



400 OBJECTNE EVIDENCE FOR ACTORS' INTERPOLATIONS 

The general conclusion is clearly negative. Careful examination of 
the scholia inevitably leads one to doubt almost all of them. An 
appeal to actors' interpolations seems grounded even less than before 
on valid external evidence. 

III 
What then of the evidence other than the scholia? The non­

scholiastic evidence is small in quantity and of uneven quality. In a 
sense Cantarella (p.143) was right in saying that Plutarch's statement 
(841F) about Lycurgus' law was the only surviving document. Ruther­
ford refers also to a gloss by Phrynichus {"restoring the old: a meta­
phor from putting other soles and heels on old sandals; they are 
speaking of those who rework, restitch the old dramas," Anecd.Bekk. 
39.19), but this probably refers to comic practice (the gloss was of a 
comic line) of unknown date and may simply reflect the common 
Roman practice of contaminatio, the blending of two Greek originals, 
to make one new play.34 Wilamowitz (p.132 n.18) refers to a passage 
in Dio Chrysostom (Or. [XIX] 69.487 R) to witness the disappearance 
of the chorus: "the strong parts, that is the iambic (Ta la/Lf3Eia). of 
the (old) tragedies remain; parts of these (TOVTWV /Llp1]) they go 
through completely (SLEgtaCtv) in the theatre while the softer parts, 
the choral (Ta 7TEP;' Ta /Ll;\'1]). have wasted away." This remark might 
be taken to refer to alterations ofthe tragedies for reviva1.35 Viirtheim 
refers to Quintilian's confusing statement (10.1.66) that "the Athen­
ians allowed later poets to enter corrected versions of Aeschylus' 
plays (correctas eius fabulas) in contests, and many won the crown that 
way." As Cantarella notes (p.154): "Le parole stesse di Quintiliano d 

3' One should note that Phrynichus speaks of restiching shoes and that it is other soles, 
not new soles, that are used. 

35 A. Pickard-Cambridge, The Theatre of Dionysus in Athens (Oxford 1946) 196, counters 
that "this only shows that 'parts of the iambic portions: i.e. select scenes, were acted, and 
does not refer to reproductions of the whole tragedies at all." He was probably thinking of 
the analogical treatment of the choral parts, which were commonly performed separately 
and in isolation; see SIG3 648, and E. G. Turner, "Dramatic Representations in Graeco­
Roman Egypt," AntCl32 (1963) 128 and n.16, Yet this is not what Dio Chrysostom means: 
the iambic parts are clearly being contrasted to the lyric in toto, dialogue to chorus (so E. 
Bethe, "Die griechische TragOdie und die Musik," NJbb 19 [1907] 84 n.1), and the verb 
8t"~la(,,v suggests a thorough rendition. For further evidence of a decline in the choral 
part see the passages assembled by Pickard-Cambridge pp.l60f and 240fI, the scholium to 

Ajax 693, and P.Oxy. 2458, which Turner thinks is "an excerpt made for acting purposes" 
(p.l26). 



RICHARD HAMILTON 401 

dimonstrano oramai che gia al tempo suo non si sapeva pili esatta­
mente come Ie cose fossero andate." 

Page adds a number of documents, only one of which is really 
relevant. The opening of the third book of Aristotle's Rhetoric simply 
states that actors are now more important than poets. The anecdote 
in Aelian (VH 14.13) merely indicates that some of Agathon's friends 
thought he should clean up his language. The reference to Theodorus 
in Aristotle's Politics (1336b28ff), however, is more to the point: 
Theodorus "allowed no one ever to appear on the stage before he did, 
not even the bad actors, on the grounds that the audience is influenced 
by what they have heard first." Aristotle's phrase "what they have 
heard first" suggests that Theodorus' alterations went beyond scenic 
effects such as the alteration recorded in the scholium to Orestes 57, 

where Helen and her baggage were paraded in the prologue-"in­
correctly, since he [the poet] says explicitly that she slipped in during 
the night and the drama takes place during the day"; if the protag­
onist is to speak the first words in every play, this would often mean 
considerable relocation and adaptation of the original. 

We come finally to the law ofLycurgus: "that the tragedies of them 
[Aeschylus. Sophocles and Euripides] were to be copied and preserved 
in the public archives and that they were to be read to the actors [who 
were planning a revival] by the city recorder." Scholars debate the 
law's effectiveness,36 but in any case it assumes the existence of a 
< clean' copy of the plays and gives no reason to suppose that such a 
copy would be obtained only with difficulty.37 Furthermore. in light 
of the two relevant documents (Dio Chrysostom and Politics), which 
both refer to major alterations, one could argue that it was changes 
such as these that were the object of the law and that it was the 
performance rather than the text that was being protected. 

If we are to give any weight to the documents, then, it is major 
alterations that we should look for as the result of actors' interference. 
Whether we have any in our extant tragedies is an issue that is and 
always will be debated (e.g., the end of Septem). The prologue to 
Rhesus seems the likeliest candidate, and Page has added the prologues 

36 Wilamowitz (p.132) is negative, followed by Page (p.2) versus Cantarella (p.143), 
Tuilier (pp.28ff) and A. C. Pearson, The Fragments of Sophocles I (Cambridge 1917) xxxiv: 
"it may be presumed that the net was cast as wide as possible, and that the most authori­
tative sources were consulted." 

87 Whether it was an acting copy (page, Cantarella p.305, oeDS supra n.2) or not (Wila­
mowitz, Cantarella p.143). 



402 OBJECTIVE EVIDENCE FOR ACTORS' INTERPOLATIONS 

of three lost plays (ArcheZaos, Meleager and MeZanippe Sophe) with more 
or less validity.Ss 

In any case, we can say that although there was clearly reworking of 
plays for dramatic production, there is no objective external evidence 
that the dramatic texts had any influence on our texts. The apparent 
evidence turns out to be a few notes on two plays, concerned with 
stage business and actors' slips. Probably in the first century B.C. the 
theory of actors' interpolations was applied to some textual problems, 
and it is this that has been wrongly backdated by several centuries and 
elevated to the status of fact.s9 

BRYN MAWR COLLEGE 

June, 1974 

38 The only source of the alternate beginning of Melanippe is Pluto 756c, who tells us 
explicitly that the poet changed the first line for a second production. With Me/eager Page 
again contradicts the only explicit testimony: the scholiast to Frogs 1238 tells us that the 
line is from Meleager but that it is not the first line. With Archelaos the external evidence 
partially supports Page: the scholiast to Frogs 1206 says "this is the beginning of Archelaos 
some say, wrongly, for no such ;\oyoc is now in Euripides; for it is not, Aristarchus says, 
from Archelaos unless the poet himself changed it later, but Aristophanes said this was the 
beginning." We know from other sources that this is not the beginning of Archelaos (see 
Nauck, TGP fr.228) but, whether we then accept Aristarchus' tentative explanation (Page) 
or not (F. Stoessl, RE 23 [1957] 2340 S.V. PROLOGOS II), the possibility of revision by actors 
has not even been raised. 

38 I wish to thank Professors G. W. Dickerson, J. M. Hunt, M. L. Lang and Miss Lisa 
Montagno for reading and criticizing this paper, which was measurably improved by the 
journal's referees as well.