JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

Oral Performance and the 
Composition of Herodotus' 

Histories 

William A. Johnson 

O RAL DELIVERY is usually assumed to have something to 
do with the composition of Herodotus' Histories. 

We are not surprised to find that several ancient 
sources report that Herodotus, the great raconteur, gave public 
lectures.1 These reports in turn have led to theories proposing 
either that the Histories provided the material for the lectures, 
or that the lectures formed the basis for the text of the H is­
tories. 2 Oral performance has therefore been championed as the 
solution to various peculiarities of Herodotean style. 3 Latti­
more, in particular, has argued that the Histories as we have it 
represents a first draft, written as if a text for recitation. 4 This 

1 For the best summary and discussion see J. Enoch POWELL. The History of 
Herodotus (Cambridge 1939: hereafter 'Powell') 32ff; see further below. 

2 The Histories as a compilation of lectures: F. Jacoby, "Herodotus," RE 
Supp!. 2 (1913) 726-392, esp. 330, 379f. That Herodotus recited from the 
Histories is the usual interpretation of the ancient biographical tradition and 
therefore an extremely widespread opinion; a convenient summary of early 
views in J. L. MYRES, Herodotus, Father of History (Oxford 1953: hereafter 
'Myres') 20-31. 

J Cf T. E. V. Pearce, "'Epic Regression' in Herodotus," Eranos 79 (1981) 
89f; M. POHLENZ, Herodot, erste Geschichtesschreiber des Abendlandes 
(Leipzig 1937: hereafter 'Pohlenz') 208H; R. V. Munson, "Herodotus' Use of 
Prospective Sentences and the Story of Rhampsinitus and the Thief in the 
Histories," AJP 114 (1993) 27-44. 

• R. Lattimore, "The Composition of the History of Herodotus," C P 53 
(1958) 9. Lattimore's theory has received some acceptance of late: cf J. A. S. 
Evans, Herodotus (Boston 1982) 17; S. Flory, The Archaic Smile of Herodotus 
Detroit 1987) 15. Lattimore is, however, not likely to be right. His assertions 
on the difficulties of writing are overstated: one would think he had never 
heard of composing drafts on wax or wooden tablets, pieces of reused papy­
rus, or leather rolls (convenient as carbon ink was easily erased), or of slaves 
to help in transcribing and copying. The places where, according to Latti­
more, Herodotus corrects himself" in stride" usually find an easy explanation 

229 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

230 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

process has perhaps reached its culmination in Nagy's hy­
pothesis that Herodotus considered himself part of a long 
tradition of logioi, who, like the epic and lyric aoidoi, gave oral 
performances. 5 

Did Herodotus write his history with oral performance in 
mind? Such has been the almost universal assumption. Even 
Flory, who makes a strong case that the Histories could not 
have been performed as a whole, concedes that "Herodotus 
probably did give [public] readings of excerpts from his book 
or used material from public readings in his book," and 
imagines that Herodotus' composition was strongly influenced 
by "oral modes of thought."6 Now, any author may of course 
recite some part of his work to an audience. This is true today 
and would be true a fortiori in a more oral culture. In the sense 
that the sounds of the words are attended, almost all literary 
writers will have 'oral performance' in mind. But this sort of 
commonplace is not what is generally meant by those who 
speak of the performance of Herodotus' work. The assumption 
is rather that much or all of the Histories is either originally or 
prospectively a series of public lectures, and that this mode of 
'publication' had profound effects upon Herodotus' style of 
writing.! 

in a delay of detail intended for some artistic effect. Lattimore's examples are 
more useful as a grouping that defines certain characteristics of archaic style 
than as proof of a lack of revision. See E. Fraenkel, Aeschylus, Agamemnon 
(Oxford 1950) 805; H. R. IMMERWAHR, Form and Thought in Herodotus 
(Cleveland 1966: hereafter 'Immerwahr') 13 n.30. Lattimore's arguments 
regarding the difficulties of book production have more force for revision of 
larger sections of the work. 

S G. Nagy, "Herodotus the Logios," Arethusa 20 (1987) 175-84, esp. 180f. 
See also (in the same volume) criticism of Nagy by M. L. Lang, 203ff, esp. 204 
(on Herodotus' use of logios); W. R. Connor, 255-66; Nagy's reply at 209f. Cj 
also Nagy's restatement: Pindar's Homer: The Lyric Possession of an Epic 
Past (Baltimore 1990) 221-24. 

6 S. FLORY, "Who Read Herodotus' Histories," AJP 101 (1980: hereafter 
'Flory') 14, 28. In recent work (supra n.4: 16) Flory seems more skeptical about 
the influence of public readings on the Histories. 

7 Thus most recently Munson (supra n.3: 27) feels secure in basing her entire 
analysis on the assumption that "the Histories evidently constitutes a 
performance in the ordinary sense of the word, being composed of sections 
which individually or in combination were designed to be delivered orally in 
front of an audience and whose performance has in turn helped to shape the 
text we now have." 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

WILLIAM A. JOHNSON 231 

It may well be that the peculiar habits of composltlon we 
associate with archaic Greek literature-such as ring composi­
tion, epic regression, A£~t~ dp0f.l£vll-are partly or even pri­
marily the result of oral performance, developed originally to fa­
cilitate oral comprehension. But we must be careful not to jump 
too quickly to the assumption that a work that exhibits such 
characteristics is itself the result of actual oral performance. 8 

Reading in classical Athens is still only partly understood. But it 
was, doubtless, a complex social phenomenon. Early readers 
would typically have 'read' the Histories in the sense of hearing 
a lector read the work aloud in the context of a private group. 
Private reading was, in this sense, performance entertainment. 
But the same private intellectual groups would have read 
Thucydides and Xenophon in exactly the same way. What then 
makes Herodotus seem so different from his successors? Many 
scholars attempt to answer to this question by assuming that the 
author himself gave public lectures-with sometimes sweeping 
consequences for our views of the social and literary history of 
the time, as well as for our interpretation of Herodotus' 
methods and aims in the construction of his work. 

Without denying the importance of oral traditions, the 
analysis here will insist upon a view of the changeover from 
orality to literacy that is less cataclysmic, more complex, more 
open to the possibility that varieties of traditions existed side by 
side. Rather than localizing the change from 'oral' to 'literate' on 
a temporal axis, as one that can be dated to the generation 
between Herodotus and Thucydides, we shall maintain here 
that changing use of and attitudes towards the written word 
varied from one type of literary production to the next. The 
work of Herodotus will be seen then primarily as part of an 
ongoing historiographic literary tradition, which had its own (if 
evolving) set of habits, attitudes, and goals. 

Specifically, this study seeks to demonstrate that the assump­
tion of Herodotus' oral performances rests on very flimsy evi­
dence. We shall examine two sets of questions: (1) What exter­
nal sources have been adduced to prove that Herodotus gave 

8 Thus the best studied of these devices, ring composition, is found not only 
in Herodotus but also in the works of Thucydides and Xenophon. For studies 
of ring composition in these and other classical authors, see the references in 
K. Stanley, The Shield of Homer: Narrative Structure in the Iliad (Princeton 
1993) 307f n.21, esp. B. Fenik, who emphasizes its use as a principle of formal 
design in both oral and written literature. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

232 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

oral performances and can they be trusted? (2) What internal 
indications suggest that the work may be originally or prospec­
tively oral? And are there counterindications in the Histories? 

I. Did Herodotus Recite Prize Essays? 

lC'tlllla 't£ ec; aid llaAAOV f1 uywVt<Jlla ec; 'to 1tapaXPlllla UlCOUEtV 
~U"f1C£t 'tat (Thuc. 1.22.4). 

Students only passingly familiar with Thucydides recall his 
claim, apparently a criticism of Herodotus, that he is writing his 
work "not as a prize-essay to be heard for the moment, but as a 
possession for all time."9 The sentence, arguably the most 
famous in Thucydides, also provides, as it seems, the most 
secure evidence for the assumption that Herodotus gave public 
recitations of his work. Thucydides seems clearly to set up a 
contrast between his own written work, composed as a posses­
sion for all time, and the work of Herodotus,10 composed as an 
entertainment to be recited in some sort of competition. But is 
this what Thucydides actually says? It is not easy to overcome 
prejudice implanted at our earliest reading of Thucydides. Com­
mentators and translators almost universally affirm this reading. 
Yet the Greek, I believe, read with care and without prejudice, 
cannot stand this interpretation. 

Let us begin with aywv U:nlU. The investigation of this word 
will lead into an unexpectedly long series of detours, but 
patience will be rewarded, for the definition of aYWVl<JJlU is 
central to understanding the Thucydidean passage, and the Thu­
cydidean passage in turn is central to the question. LSI cites 
aYWVl<JJlU at Thuc. 1.22 under the definition "that with which 
one contends, a declamation"; similarly DGE defines the word 

9 Tr. C. F. Smith (Loeb ed., here as elsewhere in this paper for Thucydides). 
Cf Jowett: "My history is an everlasting possession, not a prize composition 
which is heard and forgotten"; Rex Warner (better): "My work is not a piece 
of writing designed to meet the taste of an immediate public, but was done to 
last for ever." 

10 The argument is the same regardless of whether Thucydides means 
Herodotus in particular or early historians in general. That Herodotus is 
specifically in mind seems likely from the criticisms at Thuc. 1.20; for a 
different view see A. W. GOMME, A Historical Commentary on Thucydides I 
(Oxford 1945: hereafter 'Gomme, HCT) 148. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

WILLIAM A. JOHNSON 233 

in this context as a "competition-piece, declamation."ll But does 
the word by itself signify a declamation in a competitive setting? 
A survey of the lexical range of this word will be instructive. 

Fifth-century examples of ayffivlallu are not abundant, but 
most evidence tends toward a single usage: aywv lallu signifies a 
feat or achievement, the end result (usually glorious) of a com­
petition (be it war, games, or in general); or the popular glory 
and repute stemming from that feat. This meaning is by no 
means unexpected. The word is one of hundreds of Greek 
nouns ending in -IlU commonly formed from verbs (here 
ayo.>vi.~oIlUl); though originally action nouns, they most fre­
quently denote the object or result of an action. 12 As ayo.>vi.~o­
IlUl commonly means "to contend publicly" (whether in games 
or battle), we might well expect the noun to refer to the issue of 
the contest. 

In the six Thucydidean examples besides 1.22.2, the word con­
sistently signifies an "achievement" or the popular "glory" that 
results from such an achievement. At 7.56.2, 59.2, the Syracu­
sans consider that defeating the Athenians would be a glorious 
achievement, KUAOV ayffivlallu, in the eyes of the rest of 
Greece. The same words are used at 7.86.2 of the glorious 
achievement it would be for Gylippus if he brought the 
Athenian generals captured alive to the Spartans. At 8.12.2 
aywvlallu without qualifying adjective denotes the diplomatic 
feat of causing Ionia to revolt, thereby making the Great King 
an ally of the Spartans; Endius plots not to let this become the 
(glorious) achievement of Agis, Ill) "AY180c; 'to ayffivlallu 't01)'tO 

y£vEa8ut. Similarly, at 8.17.2 Alcibiades plots to make sure that 
the glory (of bringing Ionia to revolt) be credited to Endius, 
'Ev8i.cp 'to aywvlallu 1tpoa8£tvut. The word seems then to sig­
nify the accomplishment of a notable act and the glory or fame 
that accrues to its accomplishment. Note that the word seems 
appropriate where the writer's focus is not so much the intrin­
sic value of the achievement as the popular favor that it brings. 
This interpretation is consistent with Thuc. 3.82.7, where the 
man who violates an oath in making a surprise attack is cynically 

11 F. R. Adrados, ed., Diccionario Griego-Espanol (Madrid 1980- ) s.'ll. 11.2: 
"pieza de concurso, declamaci6n"; cf W. Cronert, Passow's Worterbuch der 
griechischen Sprache (Gottingen 1912) S.'ll.: "Prunkrede." 

12 See P. Chantraine, La formation des noms en grec ancien (Paris 1933) 
174-90, esp. 177 (on formation from verbs in -soo) and 182f, 188f (on the 
meaning). 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

234 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

said "not only to win through deceit, but to gain besides 
glorylfame for astuteness," ~UV£o"Ero<; aywvlO"/J.a. 

Other early prose examples support this. At Hdt. 8.76.2 the 
Persians are eager to trap the Greeks at Salamis so as to "punish 
the Greeks for their achievements off Artemisium" (BOlEV 'tlalV 
'trov En' 'Ap'tE/J.lO"icp ayrovlO"/J.a'trov). Hdt. 1.140.2 can also be read 
in this way: "the Magians kill with their own hands every 
creature, save only dogs and men; they kill ants and snakes and 
other creeping and flying things, thinking it a great achieve­
ment" (Kat aywvtO"/J.a /J.£ya 'tou'tO nOtEuV'tal ).13 Similarly the one 
example in Hippocrates (Prorrheticon 2.2) refers to "glory" in 
the sense of popular reputation.14 

References elsewhere are also mostly consistent. In Euripi­
des, the word is found twice. At EL 987, Orestes, having deter­
mined to kill Clytemnestra, says "but for me the achievement is 
bitter, not sweet" (7ttKPOV B' OUX ilBu 't' aywvtO"/J.a /J.Ol). Phoen. 
1355 is more doubtful: the word seems to mean the "issue" (or 
public manifestation) of the curse of Oedipus (apa<; 't' aywvlO"/J.' 
aiBinou). In other early poetic use, Ar. Ran. 284 ('Eyro O£ '1' 
Eu~ai/J.l1v av EV'tUXElV 'tlVt Aa~Elv 't' aywvlO"/J.' a~l6v 'tl 't'n<; oBou), 
the word can be interpreted as either a "glorious accomplish­
ment" or a "contest, fight," depending on whether Aa~Elv is 
understood as "attain" or "undertake." Last, the orators use the 
word in a special way to refer to the central argument or issue 
in a legal contest (Antiphon, Caed. Her. 36; Lys. 13.77). 

That no early instance supports the meaning "declamation" 
may of course be coincidence. But it is interesting that a single 
cluster of meaning accounts so well for usage in Thucydides, 
Herodotus, and most others of the same period. This same 
interpretation works very well indeed in the passage under con­
sideration here: Thucydides writes that this work is composed 
as "a possession for all time, not as a glorious (i.e., popular) 
achievement for the moment." 

But why then does Thucydides choose aywvlO"/J.a to contrast 
with K'tll/J.a? What exactly does he intend, if not a contrast 
between what is written and lasting and what is part of a festival 
(aywv) and thus oral and ephemeral? Before attempting an 

13 Herodotus' subtle humor has misled lexicographers to suppose a special 
meaning here: "an object to strive for" (LSJ) or "an aim" (Powell). 

H The gloss preserved in Hesychius (uY(oV\(Jf.L<lt(OV· 6XAn:rov bn()d~£(Ov) 
probably derives from this passage in Hippocrates (cf Cronert-Passow [supra 
n.ll]), and is itself the probable source of the late scholium (btt()£\S\v) at 
Thuc.l.22. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

WILLIAM A. JOHNSON 235 

answer, it will be helpful to complete this survey with a brief 
look at later usage. Although the meaning "glorious achieve­
ment" continues, in later writers ayWVl<J/.lU is most frequently 
used as a synonym for aywV.15 The commonest meaning, "con­
test," is applied to warfare and games-often for athletic con­
tests, and for dicastic contests and those in the arts as well: 
music, poetry, drama. 16 More to the point is the use of aywv­
l<J/J.CX to signify a literary composition. There are two clear ex­
amples, both late. 1? Both seem to refer to written compositions, 
though at least some of these could plausibly be recited at a pub­
lic assembly (aywv). Thus, [Ps.-JLucian refers to a poetic en­
comium on Homer as an ayWVl<J/.lU (Dem. Ene. 9); and Aris­
tides uses the word generally of prose writings, among which 
he specifies panegyrics, narrations of military campaigns 
(/.lU80l), and legal contests (Ad Sa rap. 148 Jebb). 

Two important passages remain for consideration. In the first, 
Polybius interrupts the narrative at 3.31 to justify his focus on 

15 Thus already in the fourth century: cf Pi. Leg. 830A, 832 E; Dem. Erot. 24; 
Xen. Cyr. 8.2.27, Hiero 11.7; Ages. 9.7; Mag. eq. 3.5; Eq. 11.13. There is a 
possibility that the word is sometimes used as an action noun in the fifth 
century; thus Powell interprets Hdt. 8.76.2 as "fighting" (presumably "suc­
cessful fighting"), and a similar interpretation will work at Ran. 284. Yet if so, 
the word still carries a connotation of glory and is not in any case used in the 
sense of a general contest or competition. The survey here is exhaustive 
through the fourth century B.C.; the summary of later usage is based on a 
review of all instances of the word through the second century A.D. as well as 
in selected later authors (taken from TLG CD ROM "D" [1992J, supple­
mented by all additional holdings of the data bank as of winter 1994). The 
one reference not mentioned is a quotation of Gorgias (D.-K. B8) by Clement 
of Alexandria, where the word is used metaphorically of the conflict between 
two forms of aPE't,,; but it is doubtful whether aYWVtO'llu is to be included 
among Gorgias' ipsissima verba. 

16 Dicastic contests: Arist. In Capito II 321 Jebb; Philostr. VS 1.510; musical: 
Strab. 6.1.9; Photo Bibl. 239.320b; cf Jos. AJ 15.269; poetic: Paus. 10.7.2ff, 6; 
Pluto Mor. 674 D; dramatic: Ael. V H 2.30. In the only use of the word for ora­
torical contests the choice of word depends on the context: Aristides (In Plat. 
II 1) contrasts athletic and musical aywvtcrjlu'tu with "more urbane contests" 
(UO''tEto't£pWV uywvtO'j.Hhwv), by which he means oratory. 

From the general idea of contest develop further semantic variations: prize, 
goal (Ar. Gram. Epit. 2.201; Philo Sam. 2.90; Synesius passim; cf Hesychius ap. 
Suda S.V. &8AOV); labor, ordeal ([LonginusJ Subl. 14.2); difficulty, trouble (Jos. 
AJ 13.111; Aristid. IlEpl 'tau 1tupa<p8. II 393); victim (Jos. BJ 1.226, 565). 

17 The unique meaning found in Himerius the Sophist ("subject, theme," 
applied to tragedy [Or. 4.143] and tales [41.46], but also to statues [48.165]), is 
related rather to the legal usage cited earlier ("central issue" in a case). 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

236 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

the causes of events. He argues vigorously for the instructional 
value of history as follows: 

~honEp oux OU'tffi<; EO"'tt CPPOV'tlO"'tfOV 'tTt<; UU'tOlV 'trov npa~Erov 
E~"'Y11o"Ero<; 015'tE 'tOt<; ypacpouO"tv 015'tE 'tOt<; uvuytvcOmcouO"tv 
'ta<; iO"'topiu<;, w<; 'trov npon:pov !CUt 'trov (il-m !CUt 'trov 
E1ttytvO/lfVrov 'tOt<; £pyOt<;. to"'topiu<; yap faV uCPfAll 'tl<; 'to Ota 
'ti !CUt nro<; !CUt 'tivo<; XaptV Enpax9" 'to npux9Ev !CUt no'tEpov 
E15AOYOV £O"XE 'to 'tfAO<;, 'to !Cu'tuAEtno/lEvov cxu'tTt<; uYcOVtO"/lU 
/lEV /la9,,/lu o· ou yivnut, !CUt nupcxmt!Cu /lEV 'tfpnEl, npo<; oE 
'to /lfAAOV ouoEv OxpEAtt 'to nupanuv. 18 

That Polybius deliberately echoes Thuc. 1.22 hardly needs 
comment. Lexicographers, influenced by their reading of Thu­
cydides, generally translate uYcOvlO"Jla here as "show-piece," but 
impartial consideration will show how little the context de­
mands or even fits this interpretation. In fact the passage seems 
almost to define UYWVlO'lla: an uywvtO'lla is what remains when 
the analysis is removed from history, when the events are 
simply narrated in a way that is pleasing to most people. 19 

Consider, second, Aristotle's sole use of UYcOVlO'lla at Poet. 
1451 b: the worst dramatic plots are those in which episodes 
follow one another with no sense of inevitability. Bad poets 
write such plays as a matter of course, but good poets write 
them to please the actors (or judges?):2o UywvlO'Jla'ta yap 
1tOtoUV'tE~ Kat 1tapa 'tTtV ouvalllv 1tapa'tdvov'tE~ 'tOY llu80v 
1tOAAaKt~ OtaO''tpE<pEtV uvaYKa~ov'tal 'to £<PE~i1~ ("for in writing 
UYWvlO'lla'ta and straining the plot beyond its inherent capacity 

18 "Therefore writers and readers of history ought not pay such attention to 
the narration of events, concentrating on what incidents come before, or at the 
same time, or after others. For if one takes away from history the why and 
how and wherefore each thing was done, and whether the outcome makes 
sense, what is left is an aYcOvto/lO and not an instructive work (/la91l/lo); and 
though it delights for the moment, it has no profit at all for the future." 

19 Further echoes of Thuc. 1.22.4 at Plb. 2.56.11, 38.4.8 are in a similar vein. 
The contrast between a history that is pleasurable and one that is profitable is 
a favorite Polybian theme: K. Ziegler, "Polybios," RE 21.2 (1952) 1503ff. 
Similarly, Lucian's rephrasing of Thuc. 1.22.4 at Hist. conser. 42 clearly under­
stands the contrast not as one of oral against written (the paraphrase omits 
alCoUnv and substitutes ouyypacpnv for SUy1C£t1:m) but rather as the difference 
between what is truthful, lasting and what is fictional, popular: K1:l1/la 1:E yap 
CPl101 /laAAoV r.r; aEl. ouyypacpnv 117tEP r.r; 1:0 7tOpOV aYcOvto/lo, K01. /ll) 1:0 /lueiOoEr; 
ao7taSE09m aAAU 1:l)v aATj9EtaV 1:iOv YEYEVll/lEVWV a7toAd7tnv 1:0tr; U01:EPOV. 

20 Some late manuscripts emend imoKpnar; to Kpnar;. In either case Aris­
totle's point seems to be that dramas of this type are written so as to play to 
the audience, with attendant disregard of higher aesthetic principles. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

WILLIAM A. JOHNSON 237 

they are often forced to distort the internal logic"). The exact 
meaning of aymvlO'jla'ta is disputed. 21 But the 'definition' 
extracted from Polybius fits well here. Good poets are forced 
to compose plays in the popular fashion (to please the actors/ 
judges), that is, to set forth the events in a straightforward, 
pleasurable manner; but the simple exposition of the story line 
takes the poet beyond what the integral plot requires and 
neglects the refashioning (cf. 1451 a36ff) that is necessary to 
produce the desired sense of inevitability. Is this not exactly 
what Aristotle means by the "episodic" plot? 

To return to Thuc. 1.22, Thucydides describes the rigor of his 
investigations and his intention to report the information accur­
ately. The lack of fabulous stories ('to jlll jlu8&8EC;) may make 
the narrative less pleasing to his audience. It will suffice if the 
narrative is profitable (ro<p£Atjla; cf Polybius) to those who wish 
to see clearly the events that have happened, for these events 
will likely happen in much the same way again. In conclusion, 
the narrative is composed as a possession for all time, and not as 
an aywvtO'Jla for an ephemeral audience. The central contrast 
here is between historical narrative that is full of tales (Jl1)8Ot) 
and plays to the popular audience of today, and historical 
narrative that focuses on facts and analysis and hopes to instruct 
generations to come (cf jl&81lJla in Polybius). In our analysis of 
aywvtO'jla we saw that the dominant meaning in early usage is a 
"feat" or "glory," where a focus on popular favor is usually 
present; in later usage, we found some references that showed 
the meaning "(written) literary composition," but passages in 
Polybius and Aristotle seem to imply the more specific 
meaning" a composition in which the events are set forth in a 
straightforward, popularizing fashion." Can it be coincidence 
that here too aywvtO'Jla refers to a literary composition that is 
ephemeral and pleasing to the crowd and (as the previous 
context implies) is full of tales? My suggestion, simply put, is 
that aywvtO'Jla refers to a popular piece of narration; that Thu­
cydides contrasts not something written with something oral 
but a difficult, factual, analytic style of composition intended for 
the ages with the easy, anecdotal, popular style of Herodotus. 

Those who would wish to relate aywvt<JJla too narrowly to 
aymv l~Ojlat in the sense of "contend for a prize in the public 
games" miss the important point that at the basis of aymvl~OJlal 

21 D. W. Lucas, Aristotle, Poetics (Oxford 1972) 125, points out some dif­
ficulties of the usual interpretations without reaching a conclusion. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

238 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

is not only aymv in the sense of "festival" and" contest" but also 
aymv in the more general (and probably original) sense of a 
"public assembly." The verb can mean "to argue the way one 
does in front of the assembly," i.e., in a crowd-pleasing and 
point-scoring manner (PI. Tht. 167E). Derivatives of ayrovi.so~at 
refer to the aymv not always literally, in the sense of the contest 
or assembly itself, but often figuratively, in the sense of that 
which is pleasing to the populace. Thus, the adjective ayrovto­
'ttKOC; can refer to things that are "striking, impressive" to the 
populace, and to people who are "eager for applause" (LSJ s.v.). 
Similarly, ayrov to'tTtC; signifies not only a competitor in the 
games, but more generally a crowd-pleaser: someone impres­
sive before the aymv in speaking or acting (e.g. Isoc. 13.15; PI. 
Ph dr. 269D; Aeschin. 3.43; Arist. [Pr.] 19.15). Thucydides writes 
~tcrJ.ta then referring not so much to a contest as to the 
popular assembly; an aymvtcr~a is a composition written in a 
popularizing, crowd-pleasing fashion, i.e., a composition written 
as though to appeal to the aymv. 

The evidence has led far from any implication of actual public 
performances by Herodotus, and detailed investigation has 
yielded no suggestion whatsoever that aymvtcr~a can by itself 
signify "prize-declamation" or the like. Yet some will argue that 
the word takes on a somewhat different significance in this 
sentence. They will argue that we must construe the word 
closely with E<; 'to 1tapaXPTlJla aKOUEtV, and that Thucydides 
adds the prepositional phrase specifically to make clear what 
aymvtcr~a means. Thucydides contrasts the K'tTl~a, which is his 
written work, with the aymvto~a E<; 'to 1tapaXPTl~a aKOUEtV, 
which is something heard, i. e., a declamation. It is, however, a 
prejudice of our own culture that causes us to take aKOUEtV as a 
reference to a lecture. Readers of the fifth century typically 
read aloud. 22 Thucydides writes aKOUEtV where we would write 
"to read." Thus Thucydides says his work is composed "as a 
possession for all time, not as a popular narrative for ephemeral 
listening (i.e., reading) pleasure." The point of comparison in the 
sentence is the lasting value of the work, not the method of 

22 See]. Balogh, "Voces Paginarum: Beitdige zur Geschichte des lauten 
Lesens und Schreibens," Philologus 82 (1927) 84-109, 202-40; G. L. Hendrick­
son, "Ancient Reading," C] 25 (1929) 182-96; and B. M. W. Knox, "Silent 
Reading in Antiquity," GRBS 9 (1968) 421-35, who accepts that reading aloud 
was general, even while presenting strong evidence for some silent reading. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

WILLIAM A. JOHNSON 239 

compOSItIOn: other works may receive more praise today, but 
the work of Thucydides will be more valued in the end. 23 

The other supposed reference in Thucydides to lectures by 
early historians is similarly misleading. At 1.21.1 the logogra­
phers are said to "compose with a view rather of pleasing the 
ear than of telling the truth" (wc; Aoyoypaq>ot ~uv£8£aav £1tt 'to 
1tpoaaywyo't£POV Tn UKpoaan 11 uA118£a't£pov). But again, we 
forget the ancient habits of reading. "To the ear" translates E<; 
uKpOaatV, which simply means "to the hearing," i.e., "to the 
reader as he listens,» where we would say "to the eye. "24 At 
1.22.4 Thucydides uses the same phrase with reference to his 
own work: Kat £C; Jl£V uKp6aatv taWC; 'to Jlll Jlu8w8£<; a1J'twv 
u't£p1tfa't£pov q>av£t'tat. Yet no one supposes that Thucydides' 
composition is influenced by his public performances. 

Scholars seem to be caught in a circular web of prejudices. As 
we <know' that Herodotus gave public lectures, passages such as 
Thuc. 1.21.4, 22.1 are taken to refer to lectures; when pressed 
for reliable evidence of Herodotus' public lectures, the passages 
of Thucydides are presented as decisive. 25 Other evidence for 
Herodotus' lectures does of course exist. The late biographical 
tradition, widely cited despite Powell's rough treatment, will be 
reviewed below. More important is the only other early source 
for this idea. 

23 It is possible to take <ll-:OUEtV as epexegetic to o.yWVtcr~w and EC; 'to 1to.po.­
XPTlllo. as adverbial. In that case, the parallel syntax predisposes the reader to 
take lC'tllllo. EC; aid as parallel to o.yWVtcrJlo. EC; 'to 1to.po.XpllJlo. in the compari­
son, and o.l(OUEtV applies equally to both. Thus Thucydides' composition is "a 
possession to hear (i.e., read) for all time, not a glorious achievement to hear 
(i.e., read) for the present." The point of comparison remains the longevity of 
the work. Commentators are, however, probably right to insist that 'to governs 
the infinitive; cf EV 'tip n:o.PUUttlCU opav at 2.11.7. 

24 Similarly, Gomme, HCT I 139. A telling parallel is Plb. 38.4.8, where in 
echoing Thucydides (see supra n.19) he speaks of the ears ('to.l~ o.lCoo.l~) of his 
readers ('tour; avuYlVwcrlCOV'tUr;); cf 36.1.7: 'tolC; o.lCouOUcrtV as the audience for 
tOtC; icr'toptOYpUqlOtC;; cf also Hdt. 1.48.1, where the process of reading from 
written tablets is related by the sequence: unfold (avo.1t'tucrcrcuv), look upon 
(EltWpo.), and then "hear" (f\l(oum:). 

25 Thus H. W. Parke, "Citation and Recitation: A Convention in Early 
Greek Historians," Hermathena 67 (1946) 87; E. Badian, "Archons and 
Strategoi, " Antichthon 5 (1971) 23 n.59 (neither of whom gives credence to the 
biographical tradition); cf. W. W. How and J. Wells, A Commentary on 
Herodotus (Oxford 1912) 6£, Evans (supra nA) 7 (who believe parts of the 
biographical tradition, but place special emphasis on the evidence in 
Thucydides). 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

240 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

Plato ([Hp. Ma.], Hp. Mi.) presents Hippias of Elis as an orator 
and teacher who traveled about, often on state business, ex­
acting fees for his lectures. Hippias was a true polymath, 
claiming expertise in mathematics, astronomy, poetry, music, 
grammar, and mnemonics, not to mention metalworking and 
weaving ([Hp. Ma.] 285c-86A, Hp. Mi. 368B-D). He also, how­
ever, claimed knowledge in certain historical topics, such as 
genealogies and foundation stories ([Hp. Ma.] 285D ). This last 
has been taken as evidence for lectures by historians (i.e., nar­
rators of logoi) in this period.26 Yet clearly Hippias was a sophist: 
the depiction here implies it and he is so named by Plato ([Hp. 
Ma.] 282E) and generally (e.g. Pluto Lye. 3); Luc. Hdt. 3; Philostr. 
VS 1.495; Ath. 13.609A; Liban. Decl. 2.1.24). As a sophist, he is 
by definition a traveling educator and lecturer. The historical 
nature of two of Hippias' many lecture topics is irrelevant for 
defining the usual practice of historians or Herodotus' habits. 
Moreover, though the Hippias Maior is set in the last quarter of 
the fifth century (a reference to Gorgias at 282B), the author 
presents the practice of lecturing for a fee as a somewhat recent 
innovation (282c-83D). Protagoras was the first to lecture for a 
fee (282D), thus the practice seems to date to the latter half of 
the fifth century. If the early historians were itinerant lecturers, 
it is remarkable that their contemporaries do not mention this 
practice so frequently cited for the sophists. 27 The case of the 
sophist Hippias, in any event, hardly affords reliable evidence 
for the early historiographical tradition. 

The late accounts of Herodotus' lectures fall into three cate­
gories: (1) the story of Thucydides, who as a boy burst into 
tears after hearing Herodotus speak, thus prompting Herodotus 
to tell Thucydides' father that the boy showed a great passion 
for learning (in one version, this occurs at Olympia); (2) the 
story of Herodotus' recitation of his Histories at the Olympic 
games, resulting in instant fame and the designation of his nine 
books as "the Muses"; (3) various stories that Herodotus was 
paid (or not paid) for telling stories favorable to a city: thus the 
Athenian Council voted him ten talents for his flattery of 
Athens; the Thebans refused to pay him or let him speak with 
their young men, so he wrote malicious stories against them; 
the Corinthians did not think his favorable stories worthy of 

26 L. Pearson, Early Ionian Historians (Oxford 1939) 8; Evans (supra nA) 8. 
27 Athenian writers of the fifth and fourth centuries are almost uniformly 

silent about the early historians: Pearson (supra n.26) sf. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

WILLIAM A. JOHNSON 241 

pay, so he rewrote them to be negative. Powell both con­
veniently collects and effectively disposes of these tales, none of 
which is earlier than the first century A.D. 28 Critics since Powell 
vary in the credence they give to these stories, but remain 
united in their conviction that actual lectures by Herodotus lie 
behind the tradition.29 It is not clear how several fantasies add up 
to a fact, but suspicion mounts that the passages in Thucydides 
remain the decisive evidence. To Powell's critique only a few 
remarks need be added. 

First, detailed studies of the ancient biographies have con­
cluded that there is little evidence of reliable biographical 
traditions for early Greek writers. 30 Factual material, where it 
exists, is derived from the author's own works, or from the 
writings of his contemporaries. Ancient biographers freely in­
vented stories about writers to make history fall into neat pat­
terns, to explain curious details in the author's text, or simply to 
spin a good tale. All the tales surrounding Herodotus have a 
motive: to establish a relationship between the two most 
famous Greek historians; to explain why the nine books are 
named after the Muses; to motivate the favorable or unfavor­
able accounts of various cities. All the tales contain elements 
that are suspicious (the pairing of Thucydides and Herodotus, 
reminiscent of the pairing of Homer and Hesiod)31 or 
demonstrably false (the story of the Muses, which must be false 
because the current book divisions are late),32 or absurd (the 

28 Powell 32ff; further criticism in Badian (supra n.25) 23 n.59; Flory 14. 
29 Myres 3, 5 (unaffected by Powell's arguments); T. Kleberg, Buchhandel 

und Verlagwesen in der Antike (Darmstadt 1969) 4; Evans (supra n.4) 7; K. H. 
Waters, Herodotos the Historian: His Problems, Methods and Originality 
(London 1985) 14. 

30 See J. Fairweather, "Fiction in the Biographies of Ancient Writers," 
AncSoc 5 (1974) 231-75, esp. 275; M. R. Lefkowitz, The Lives of the Greek 
Poets (Baltimore 1981) viii, passim. 

31 Similarly, Lysias is said to have known Demosthenes, although 
Demosthenes was no more than four years old when Lysias died. Inventing 
relationships between literary figures of a given time is a fundamental 
tendency in ancient biographies: see Fairweather (supra n.30) 256--64, esp. 261£. 

32 A reference to an alternative division by books appears in the Lindian 
Chronicle (securely dated to 99 B.C), where an event in Book 3 is cited as Book 
2: 'Hpooo'to:; ... £v 'tat B 'tav icrto(pt]av (29, C38). See C. Blinkenberg, Die 
Lindische Tempelchronik (Bonn 1915) 4 (on the date), 24f (on the text). The 
earliest reference to a division into nine books is Diod. 11.34.6. Similarly, the 
division of Thucydides into eight books seems to have been a Hellenistic 
innovation (attributed to Asclepiades); divisions into nine and thirteen books 
were also known: cf Marcellin. Vito Thuc. 58; Diod. 12.37. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

242 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

story that Herodotus rewrote his work to spite Corinth; the ten­
talent gratuity). Given the imaginative element in ancient 
biographies, there is little reason to trust the whole any more 
than the parts. 

Secondly, in almost all these tales Herodotus' supposed activ­
ities strikingly resemble those of the sophists, including public 
lectures for a fee and exhibition lectures (E1tl8d~Et<;) at the 
Olympic games ([Hp. Mi.] 383c; Luc. Pseud%g. 5). The 
sophists aimed not at "oral publication "33 but at self-promotion 
as wise men able to give profitable instruction. Recitations for 
edification of the public must be carefully distinguished from 
recitations as a marketing tool to enhance the educators' 
reputations and thereby their standard of living. The emphasis 
on paid lectures and exhibitions at Olympia arouses the 
suspicion that the details of the stories are importations from 
traditions surrounding the sophists. 34 

What to do then with the common belief that Herodotus 
traveled about lecturing? The supposed proof in Thucydides 
and Plato does not stand scrutiny; and the late biographical 
traditions, never a secure base of support, are suspect on any 
number of grounds. Is there in fact any credible external 
evidence for this belief? Two final arguments should be 
examined. 

The first is the correlation between Hdt. 3.119 and Soph. Ant. 
905-12. Both passages argue that a brother is dearer to a woman 
than her husband or children: she can remarry or have other 
children, but (with her parents dead) her brother cannot be re­
placed. The phrasing is close enough and the use in Ant. suf­
ficiently awkward to justify the inference that Sophocles drew 
on the passage in Herodotus. 35 Given the traditional dating of 
Ant. to 442, the allusion seems to prove that Sophocles heard or 

J3 Thus E. A. Havelock, Preface to Plato (Cambridge [Mass.] 1963) 54. There 
is no reason to doubt that the sophists distributed written copies of their 
works: see E. G. Turner, Athenian Books in the Fifth and Fourth Centuries 
B.C. (London 1952) 19f. 

H The parallel did not escape Lucian (H dt. 3), who predictably makes 
Herodotus the innovator and his success at Olympia then the source of the 
exhibitions by the sophists. 

35 D. L. Page, Actors' Interpolations in Greek Tragedy (New York 1987) 
86£. To value a brother over a husband or a child was a folktale motif 
widespread in Persia and India (Schmid-Stahlin 3 [Munich 1934] 1.2 318 n.3 
with references). The possibility of a common source cannot be ruled out, but 
the closely parallel sentence structure (as Page demonstrates) seems to point to 
a direct relationship. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

WILLIAM A. JOHNSON 243 

read part of the Histories before it was complete, and some 
have concluded that Sophocles became acquainted with this 
section through a lecture or public reading. 36 But problems 
abound. Aside from the possible interpolation of the passage 
and a recent challenge to the early date of the play,37 even if one 
supposes an acquaintance with Herodotus' work in an un­
finished state, the conclusion of a public performance does not 
follow. In Athens, as in any close-knit intellectual community, 
surely it may have been common practice to circulate drafts or 
to read ongoing work in private groups. Thucydides, Polybius, 
and Tacitus would have done the same. If Sophocles did read or 
hear part of the Histories while it was being written, that need 
have no more import than Pliny the Younger's comments on 
Tacitus' drafts (Epp. 7.20, 8.7). Activity of this kind does not 
differentiate Herodotus, nor does it make his work fundamen­
tally <oral' in some way that the works of Thucydides and 
Tacitus are not. 

Similarly, allusions to Herodotus in Aristophanes' Acharnians 
have been taken as evidence for public lectures: the parodies of 
Herodotus imply widespread knowledge of Herodotus' work; 
as it is unlikely that the general populace could read such a 
work, Herodotus must have given public recitations. It is hard 
to imagine under what conditions Herodotus would have re­
cited to 15,000 Athenians a prose work that would require 
twenty-four hours of recitation,38 but the argument has more 
fundamental problems in any case. There is no reason to sup­
pose that Aristophanes required or expected his entire audience 
to understand every joke. The most secure parody, that of the 
origins of the Peloponnesian War (Ach. 523-28 parodying Hdt. 

36 Thus e.g. J. Cobet, "Wann wurde Herodots Darstellung der Perserkriege 
publiziert?" Hermes 105 (1977) 25. 

37 Interpolation: Page (supra n.35) 86-90; date: Lefkowitz (supra n.30) 81ff. 
38 Flory (13ff) faints out the problems of recitation for a work of this size. 

His calculation a fifty hours of recitation is based on an impossibly slow rate. 
My estimate is based on a count of 185,424 words at a rate of 130 words per 
minute; cf. n.49 infra. For the size of Aristophanes' audience see A. Pickard­
Cambridge, The Dramatic Festivals of Athens2 (Oxford 1968) 263. Suetonius' 
report (Claud. 42.2) of yearly recitations for the twenty books of Claudius' 
Etruscan history and the eight books of his Carthaginian history is surely to 
be taken as an example of imperial excess rather than standard practice. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

244 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

1.4 ),39 makes sense and is humorous without any knowledge of 
Herodotus. This is not necessarily an argument against the pas­
sage as a parody of Herodotus,40 but it does remind us that Aris­
tophanes, like any literary artist, can be understood and appreci­
ated on many different levels. Flory is certainly right in empha­
sizing that the Histories was unlikely to have been a 'best-seller'; 
allusions to the Histories in Aristophanes and elsewhere aimed 
at only the most educated of the audience. Such references in 
any case prove nothing about public recitations of Herodotus' 
work. 

Like any author in any period, Herodotus may well have 
given lectures based on his work, or on which his work is 
based, or which had nothing to do with his work. Yet we must 
be clear about what is fact, what has at least some evidence to 
support it, and what is pure supposition. A public performance 
by Herodotus based on his Histories simply cannot be demon­
strated. In reflecting on this, one must be careful not to 
confound traditions attached to different types of literary 
production. For some literary genres oral performance is ele­
mental: poets continued to perform their works publicly; 
sophists for economic reasons gave public lectures; and local 
memorializers may have continued to work within an oral 
tradition.41 But earlier writers in the historiographical tradition 

39 This parody (like others from the Ach.) has been questioned (c. W. 
Fornara, "Evidence for the Date of Herodotus' Publication," ] HS 91 [1977J 
28) and defended (Cobet [supra n.36J 9-12), but the defense in turn has been 
attacked: see n.40 infra. 

40 Thus Flory (25f), who also goes astray in other arguments against Cobet. 
Aristophanes certainly would "have created a parody which had only 'a great 
likelihood' of being understood"; cf K. J. Dover's shrewd observation 
(Aristophanic Comedy [London 1972] 188f, quoted at Flory 25 n.49): "A 
majority in an audience can be surprisingly tolerant of a parody which only a 
few can really appreciate." Flory also errs in thinking that greater familiarity 
with the beginning of a work rests upon a "suspiciously modern analogy." It 
hardly needs to be documented how much more often passages toward the 
beginning of works or books are quoted in later authors. c. W. Fornara, 
"Herodotus' Knowledge of the Archidamian War," Hermes 109 (1981) 153ff, 
presents a more cogent, if not quite compelling, counterattack against the 
probability of allusion in the Ach. 

41 Cf the cross-cultural evidence collected by J. A. S. EVANS, Herodotus, Ex­
plorer of the Past (Princeton 1991: hereafter 'Evans') 89-146, esp. 96f, 113-22. 
Such evidence reveals little about the habits of Herodotus who, like 
Thucydides, clearly sets himself in the tradition of earlier Ionian writers (cf 
2.15, 17). Evans' survey in fact makes quite plain how different were the 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

WILLIAM A. JOHNSON 245 

explicitly emphasize the written aspect of their work (see Part 
II below). There is no evidence of a tradition of public readings 
for either these writers or Thucydides, writing only a few years 
after Herodotus. The only evidence for Herodotus amounts to 
a few late stories that range from the dubious to the absurd. In 
the common view "the normal mode of publication [in the fifth 
century] was by reading aloud, or reciting, in public."42 Such a 
broad-based assumption is both without evidential support and 
is overly simplistic-a homogeneous model imposed upon a 
heterogeneous and changing literary culture. The possibility 
should be considered that beside the oral culture, still strong for 
certain literary traditions, a culture of readers existed and that 
certain types of literature may have been oriented towards this 
literate culture. Surveys of evidence for literacy in the fifth cen­
tury have uniformly found strong indications of literacy, at least 
among the upper classes in Athens, and a substantial class of 
educated, literate Greeks in this period cannot be doubted. 43 
Whether 'reading public' denotes solitary readers or (more 
typically) participants in small groups where the text was read 
aloud, Herodotus found an adequate reading public among 
intellectuals at Athens and elsewhere to motivate his enterprise. 
In any event, no external evidence supports speculation about 
oral performance and such speculation will therefore have to 
rest on evaluation of the work itself. 

activities and aims of local memorializers from the collecting, analyzing, and 
recording of historiographers like Hecataeus, Herodotus, and Thucydides. 

42 Waters (supra n.29) 14. The assumption is remarkably widespread: e.g. J. 
Welles, Studies in Herodotus (Oxford 1923) 203; Lattimore (supra n.4) 9; 
Havelock (supra n.33) 54; Knox (supra n.22) 421; J. A. S. Evans, "Herodotus' 
Publication Date," Athenaeum 57 (1979) 148; Flory 12, but no evidence is cited 
beyond that reviewed here. 

43 The most exhaustive and cautious assessment of the evidence (with 
references to earlier studies) is W. V. Harris, Ancient Literacy (Cambridge 
[Mass.] 1989) 65-115. Harris' argument against mass literacy in classical 
Athens presents impressive evidence for literacy among the upper classes 
ramong the well-to-do, practically all males must have been literate," 103). It 
perhaps needs to be said that I believe Herodotus was wealthy (how else 
could he have afforded the expense of his travels and writings?) and that his 
intended audience was primarily, if not entirely, the social and intellectual elite 
at Athens and other Greek cities. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

246 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

II. What Does Herodotus Say? 

'HpocS6'rou 'AAtKapV1100EOS iowpLl1S U1tOb£~lS 11b£, ros 1!1)'r£ 'rn 
yc:vol!£va £~ uvElpomwv 'rip xpov,? £~hl1Aa YEVl1'rat, 1!1)'r£ 
epya l!£yaAa 'r£ Kat Elwl!ao'ra, 'rn I!EV "EAA110t, 'rn bE !3ap­
!3UP0tOl u1tOb£xElEv'ra, uKAEa YEVl1'rat, 'ra 'r£ UAAa Kat bt' ilv 
ahtl1v £1tOAEI!110aV UAA1)AOlOl. 

Those who think that Herodotus' logoi reflect a series of lec­
tures 44 would, I suppose, imagine that Herodotus stands before 
the crowd and begins with the words, 'Hpobo'tou 'AAlKap­
VT100fOe; to'topiTle; anobE~le; llbE .... But this will not do. Herod­
otus standing before the crowd would hardly refer to himself in 
the third person; and besides, a lecture would be a poor means 
to insure that "the events not be forgotten.» Obviously the 
opening sentence has been added as a suitable proem for the 
work as a whole. The lecture must have started then at 1.1: ITEP­
Of(I)V !lEV vuv Ot Aoytal <PoivlKae; ai'tioue; <paot YEVf08at Tile; 
bta<p0plk But what bta<popf)? We see at once that this cannot be 
the beginning of the lecture, for it depends upon the opening 
sentence to make clear to what quarrel it refers.45 Perhaps the 
entire prologue (1.1-5) is an addition? The lecture would begin 
at 1.6: Kpotooe; ~v Auboe; !l£v YfVOe;, nate; b£ 'AAua't'tE(I), TupavVOe; 
b£ E8vE(I)V 'trov EV'tOe; "AAUOe; nO'ta!lou .... This is much more 
satisfactory: the sentence easily detaches itself from what comes 
before, and the opening words announce the topic, Croesus and 
Lydia. In fact the Croesus logos (1.6-94) is generally considered 
a well-defined unit in the Histories. 46 Yet defining an end to this 

44 That the Histories is essentially a series of lectures is basic to many studies: 
cf. supra nn.2, 4, 7, 29 and n.51 infra. It may seem only common sense to 

assume a more shifting relationship between oral and written text than is 
allowed in what follows, but these are not the terms in which the discussion 
has been framed. Even a seemingly sophisticated treatment such as Munson's 
detailed analysis (supra n.3) assumes the identity of written and oral text­
typical for studies of 'oral' elements of style and structure in the Histories. 

45 T. Krischer, "Herodots Prooimion," Hermes 92 (1965) 160f, argues that 
the repetition of aittl1 from the opening sentence is a formal aspect of the epic 
proem. If so, the two sentences are not easily separated. 

46 Several analyses delineate the main divisions or logoi of the Histories: 
Jacoby (supra n.2) 283-326; Myres 118-34; Immerwahr 329-62; S. CAGNAZZI, 
"Tavola dei 28 Logoi di Erodoto," Hermes 103 (1975: hereafter 'Cagnazzi') 
421ff, and all agree on the boundaries of the Croesus logos except Cagnazzi, 
who includes 1.1-5 with 1.6-94. Herodotus called the end of this section "the 
first of the logoi" (we; Od;T,AC1l'tat flOt f.v 'tip 7tpOO-tC(l 'twv "A6yrov, 5.36., referring 
to 1.92.2). 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

WILLIAM A. JOHNSON 247 

logos is troublesome. The sentence that clearly intends to close 
the Croesus logos contains /lEV, by which Herodotus looks 
forward to the opening of the next logos: Au801 /lEV 811 uno 
IT€pcrncH £8£()OUA(J)V'tO (1.94.7). Now the point may seem over­
precise, or even tendentious, for in oral performance /lEV could 
be replaced with e.g. ou't(J). More generally, oral presentation 
would by nature have a more fluid relation to the written text. 
But this trifling example yields a basic point: the eHort to avoid 
seams between sections, to present all his evidence as a continu­
ous and connected stream, is a fundamental characteristic of 
Herodotean style. 47 Even if rewriting can be assumed, we must 
ask why, if he had oral presentation in mind, Herodotus so 
consistently does this.48 

What about the length of the logoi? The Croesus logos, as it 
stands, is about 12,400 words and would require about one hour 
and thirty-five minutes to recite-a plausible lecture. 49 Cagnazzi 
(esp. 421H) has attempted to divide the Histories into sections 
manageable for recitation, although her divisions are at times 
debatable. The Egyptian logos, for instance, clearly intended as a 
unit (2.1-3.38, almost four hours), she divides into four discrete 
logoi. Certainly the Egyptian material is organized into subsec­
tions and it is reasonable to posit a series of lectures. But how 
successful are these divisions? The Egyptian logoi range from 
thirty-five minutes (2.1-34) to an hour (2.35-98) to an hour and 

47 Cf H. Frankel, "Eine Stileigenheit der friihgriechischen Literatur," Wege 
und Formen friihgriechischen Denkens2 (Munich 1960) 83ff. 

48 Any rewriting flies squarely in the face of Lattimore's theory that the 
Histories constitutes a first draft (supra nA). The seamless nature of the 
narrative makes uncomfortable bedfellows of Lattimore's two assumptions 
that (1) the text was written for oral recitation, and (2) the text was unrevised. 

49 Word counts are generated by computer from the TLG CO ROM "0" 
(1992). My time estimates follow the" average" rate of 130 words per minute 
calculated by A. Rome, "La vitesse de parole des orateurs attiques," BulAc­
RoyBelg 38 (1952) 596-609. Many of Rome's arguments derive from dubious 
assumptions, most notably that preserved speeches are accurate reflections of 
the speeches as delivered. Rome's calculation of differing rates for specific 
orators in particular is very dubious; the fast rate calculated for Oemosthenes 
(150 words per minute) surely reflects no more than Oemosthenes' habit of 
adding substantially to a speech as he edited it for distribution. Yet Rome's 
basic rate is not likely to be far wrong. He rightfully stresses how awkwardly 
we read ancient Greek and insists on the evidence of rates for modern 
European languages. To be convinced that the Greeks recited as quickly as 130 
words per minute, one need do no more than set up (as I did) a series of tests 
on texts in one's native language. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

248 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

thirty-five minutes (2.99-182), then back down to forty minutes 
(3.1-38). There is a similar diversity in the logoi as a whole: five 
of Cagnazzi's 28 logoi run to twenty-five minutes or less (3.39-
60; 5.1-27, 97-126; 9.90-113, 114-22); four run to an hour and a 
half or more (1.1-94, 2.99-182, 3.61-160, 7.138-239). The remain­
der fall at all points in between. The longest logos is five times 
the length of the shortest (by comparison, the longest books of 
the Iliad and the Odyssey are 2 and 2.5 times the size of the 
shortest). One would expect more consistency if the logoi were 
based on lectures or if they were composed with lectures in 
mind.50 

The Croesus logos exhibits certain model characteristics for a 
lecture: a well-defined subject, complete in itself, and an abun­
dance of colorful stories (Gyges, Arion and the dolphin, Solon 
the wise adviser, the 'tragedy' of Atys, Croesus on the pyre) 
that might be associated with oral performance. Proponents of 
Herodotus' lectures should be satisfied if we choose this as a 
test case rather than, say, the ethnography of Libya. 

Here as everywhere in the Histories many links join the in­
dividual logos to the work as a whole. Some links are formal. 
Thus at 1.56.2-68 Herodotus describes the early history of the 
Athenians and Spartans. This early history is resumed several 
books and many logoi later at 5.39-48, 55-96. The Spartan sec­
tion in the Croesus logos ends with events in the reign of Anax­
andrides (1.67.1). The comment that the king Anaxandrides was 
no longer alive (5.39.1) demonstrates that the later section is 
meant to continue the history left off in the earlier section. 51 

Four other places similarly refer to events in the first logos 
(1.103.2 referring to 1.74; 1.169.2 to 1.92.1,3.47.1 to 1.70,5.39.1 to 
1.67.1). Further, some places in the first logos explicitly refer to 
another logos: 1.75.1 (referring to 1.107ff), "This man Astyages 
was subdued and held prisoner by Cyrus even though he was 

50 Counts do not confirm Cagnazzi's claim (388) that the logoi are generally 
25 to 30 Teubner pages in length. If I seem to attack a point of view that died 
with Jacoby, and to which Cagnazzi is an aberrant adherent, recent books 
warm to Jacoby's idea that the logoi were originally lectures: cf Waters (supra 
n.29: 70): "it seems probable enough that the general plan of Herodotus' work 
was modified, if not totally controlled, by the consideration of the appropriate 
length of a lecture"; and J. Hart's comment (Herodotus and Greek History 
[London 1982] 205 n.17) on the composition that "the most authoritative 
contribution, the starting point of most recent debate on the subject, is that of 
Jacoby." The idea of oral publication (cf supra n.42) also implies units 
manageable for recitation. 

51 The lack of an exact continuation argues against Jacoby's theory that the 
two originally constituted a single logos. An excellent discussion of these 
passages in Immerwahr 35ff. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

WILLIAM A. JOHNSON 249 

his mother's father for the reason which I will declare in a subse­
q uent part of my history" (cSt' ahtT)v TTtV f:ym £V 'totat o1ttaro 
AoYOtat aT)~av£ro). Two later logoi have similar explicit referen­
ces back to the Croesus logos (1.130 to 1. 77ff, 5.36.4 to 1.92.2). 
Readers of Herodotus know that the work is crisscrossed with 
such implicit and explicit references. 52 Parke, arguing for oral 
publication of the Histories, tries to dismiss the cross-refer­
ences: "There is nothing inconsistent in a lecturer saying, 'They 
took Ninus, but how they took it I shall show in another lec­
ture'. "53 We might agree that this seems unexceptional behavior. 
Yet to do this in a series of lectures over fifty times is, I think, 
quite exceptional indeed. The cross-references are both 
ubiquitous, occurring in almost alllogoi, and frequently far apart, 
often spanning several logoi, at times spanning several books. 54 

Implied is a degree of concentration from the audience that is 
hardly to be expected at the Olympic Games. An unprejudiced 
observer might reasonably say that such cross-references imply 
the ability to recall the earlier passage or to reread. Such a 
complex and wide-ranging web of cross-references is in any 
case not at all expected in oral style when no tradition exists of 
large-scale and repeated performance .. 

Besides cross-references, Herodotus links other logoi through 
thematic ties, e.g. in the Croesus logos, the wise adviser, the 
misinterpretation of oracles, the inherited and inescapable down­
fall, the king with unlimited wealth who insists on the conquest 
of poor neighbors, and the violation of the natural boundaries of 
one's domain-themes repeated through the reigns of Cyrus, 
Cambyses, and Darius but most significantly in Xerxes' disas­
trous campaign against Greece. There are more subtle ties as 
well. The story of Gyges recalls the prologue, where dis-

52 For a complete list of explicit and implicit cross-references see Powell 89f, 
to which add 1.148 to 1.139 (which Powell may have considered an inter­
polation) and 1.125.4 to 3.40 (discussed by Powell 49); correct his reference at 
8.92.2 to 6.49 (not 6.50). Some of his implicit references are possible but not 
compelling: 4.165.2 to 3.13.3f, 4.168.1 to 2.77ff, 4.181.1 to 2.32.4 (all three of 
which Powell admits to be ambiguous, 6), 6.104.1 to 6.41, 7.194.2 to 1.137.1. 
Two are in my opinion far-fetched: 1.16.2 to 1.103, 1.91.5 to 1.107. 

53 Parke (supra n.25) 87 contra Powell 32. 
54 To take Cagnazzi's 28 divisions as a sample, cross-references are absent 

only in her logoi 15 and 16 (5.65-96, 97-126), 23 (8.1-39), 27 and 28 (9.90-113, 
114-22). Examples of far-apart references are not hard to find. The following 
lists only explicit references to another part of the work two or more books 
away: 2.161.3f referring to 4.159.4f, 4.1.2 to 1.106.1,4.181.2 to 2.43.4, 5.22.1 to 
8.136,5.36.4 to 1.92.2,7.93 to 1.171.2. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

250 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

honoring a woman motivates a cycle of revenge that culminates 
in the siege and sack of Troy. Gyges' murder of Candaules is of 
course avenged in the fifth generation of his successors, i.e., by 
the siege and sack of Sardis and Croesus' downfall. Or the ties 
can be more subtle yet. As an example of the Persian custom of 
burying people alive, Xerxes' wife Amestris buried alive "twice 
seven" (8ie; £7t'tU) sons of Persian nobles (7.114.2), a number 
recalling the "twice seven" Lydian boys assigned by Cyrus to 
burn on the pyre with Croesus (1.86.2). The repetition of detail 
reinforces the idea that the Persians were both exotic (the 
magical number seven) and cruel (burying and burning people 
alive).55 These are only a few of the most salient, fOints with 
which any careful reader might agree. All students 0 Herodotus 
are aware that the number of links seems limited only by the 
industry of their own investigations. 

The Croesus logos then has many links, both implicit and ex­
plicit, to other logoi. To suggest that Herodotus aimed at 
creating a unity, with many echoes and relations among the 
diverse strands of his material, is hardly a revolutionary position. 
Such in fact is today's communis opinio. This unity has, 
however, important consequences for the suggestion that the 
work was originally or prospectively oral. If the material origin­
ated in lectures, was Herodotus fortunate enough to have 
happened upon lecture topics that happily interrelate in such a 
detailed way? Or is the complex weaving of associations among 
the topics the result of later revision? If later revision, what 
relationship remains between the original lectures and the 
finished written work? On the other hand, if the Histories was 
used as a text for recitations ('oral publication'), did Herodotus 
care so little for the unity he had labored to construct that he 
was eager to present the Histories in part? If he read excerpts, 
does this not imply rather an eagerness to entice the audience to 
read the written whole? Or perhaps the work was recited in 
toto? Yet the work is simply too large for this to be likely. 
Homer is not a good parallel. The Homeric poems became the 
common property of a group of people, the rhapsodes, whose 
function was to present and preserve the poems. Poetic per­
formances had become part of the religious festivals and pro­
vided a traditional context for epic performance. The repetition 

55 To a Greek audience 01.~ btta. may well have recalled the seven youths 
and seven maidens who accompanied Theseus as human sacrifices to the 
Minotaur. Plato (Phd. 58A) refers to them as tOU~ "O1.~ bttCl" hdvo\)~; cf 
Bacchyl. 17.2, Sappho 206, Eur. HF 1327, Isoc. He!. 27. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

WILLIAM A. JOHNSON 251 

of the performances gave adequate opportunity for an audience 
to grasp the unities of the poems piecemeal and over time. 
Herodotus' Histories, on the other hand, was written in prose 
and is nearly twice the length of either the Iliad or the Odyssey. 
In what context would such a work be performed? Even if 
Cagnazzi's ungainly divisions are accepted, are we to assume 
twenty-eight straight days of recitation? Or perhaps two ses­
sions a day for fourteen days? Or four sessions for seven days? 
But where?56 Without a traditional context for the performance, 
what audience would attend a recitation of such length? The 
elaborately constructed attempt at unity implies that the author 
intended the work to be presented as a whole. Yet the size of 
the work seems to preclude oral performance as a suitable 
means for such a presentation. 

In the introduction to the Histories, Herodotus makes it clear 
that he is setting his work in writing. The purpose of the presen­
tation of his inquiry is so that events will not be forgotten and 
the great deeds of men will not go without fame. The opening 
words ('Hpo8o'tou 'AAt1\:apvl1crcrro~ lcr'tOPll1~ cmo8£~l~ ,,8£) 
parallel those of Hecataeus CEKa'tatOC; M lAT]crtOC; Cb8£ /lu8£t-tal' 
'ta8£ 'Ypaq>w, w~ /lOl 80K£t aA118ra dvat, FGrHist 1 F1), Antio­
chus CAV1:l0X0C; 3£voq>av£O~ 'ta8£ O"'\)VEypa'l'£ 1t£pt 'l'taAll1C; EX 
'trov apxalwv AOYWV 'tfx 1tlcr'tO'ta'ta Kat craq>rcr'ta'ta, FGrHist 
555 F2), and Thucydides (80UKU8l811C; 'A811vatoc; ~uvE'Ypa'l'£ 'tOY 
1tOA£/lOV 'trov TI£Ao1tOVVl1crlWV Kat 'A811valWV), all of whom refer 
explicitly to writing. Herodotus several times refers to his own 
activity as "writing." He uses 'Ypaq>w to emphasize what he will 
or will not include in his history. The translation "I record" or 
"I am making a record" has the proper nuance (differently, 
Immerwahr 6f n.14). Thus he says that he is including only one 
variant of a tale: "Various stories are told by people who wish to 
glorify Cyrus, but I will make my record according to those 
people who tell the truth" (1.95.1; cf 2.70.1,6.53.1,7.214.3). Or 
he says that he is including things whose truth he doubts: "I do 
not know whether these things are true; hut 1 make a record of 
what people say" (4.195.2; cf 2.123.1). Or he says what he is not 
including: "Some Greeks have used this doctrine as their own; 1 
know their names hut do not record them" (2.123.3; cf 3.103.1). 
This last example in particular should make clear that Herodotus 
thinks of his work as lasting and with the power to memorialize 

56 None of the major festivals was long enough to accommodate a full 
recitation of the Histories: Flory 14. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

252 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

(pace Evans 101). He is very careful about what he does or does 
not include in his record. He refuses to mention philosophers 
who have stolen their ideas from others, in order to relegate 
them to the oblivion they deserve. When Hecataeus writes 'taOE 
'Ypacpro, 0)(; /lOl OOKEl aATl8£a dvm, the use of 'Ypacpro is emphatic, 
self-confident: he is making a record of what he thinks is the 
truth. He expects the record to circulate in places and times 
beyond those he can personally address. That is why he puts his 
name at the beginning, why he proclaims that he is creating a 
written record. By opening the Histories with his name, 
Herodotus-like Thucydides-proclaims his alliance with the 
tradition of such earlier historians as Hecataeus. 57 This tradition 
was explicitly and self-consciously a written tradition. 58 

A thorough review of supposedly 'oral' elements in Herodo­
tean style is beyond the bounds of the task set here. A brief 
treatment of one well-known study will, however, serve to sug­
gest the fundamental problems that bedevil such analysis. 
Pohlenz (208ff), who thought it probable that Herodotus gave 
recitations, tries to isolate traces of oral performance in his style. 
Now, as noted earlier, the prologue (1.1-5) does not begin in a 
reasonable way for a lecture and there is a clean break between 
the prologue and the following Croesus logos. The prologue can 
be taken then as a brief section deliberately written to introduce 
the work as a whole. This section at least was not written with 
recitation in mind. Yet the prologue contains most of the 'oral' 
elements isolated by Pohlenz: (1) close repetition of nouns and 
avoidance of the pronoun: EcrmtlKV£Ecr8al Kat OTt Kat Et; "Apyo~' 
'to O£ " A pyQ~ 'tou'tOv 'tOY xp6vov 1tPOE'iXE a7tacrl 'trov EV 'tTl vuv 
'EAAaOl KaAEO/l£Vn XWpTl. 'A1tlKO/l£VOUt; O£ 'tOUt; <1>olVl1Cat; Et; OTt 
'to "Apyo, (1.1.1); (2) use of deictic pronouns: a1tO 'tTtt; 'Epu8pTtt; 
KaAEO/l£vTlt; 8aAacrcrTlt; a1tlKO/l£VOUt; E1tt 'tUv OE 'tTtv 8aAacrcrav 
(1.1.1); (3) use of an in trod uctory participle that repeats the main 
verb of the previous sentence (here with a short clause 

57 Herodotus refers several times to Hecataeus' work: 2.143; 5.36, 125f; 6.55, 
137. 

58 Pearson (supra n.26) 8; Turner (supra n.33) 17f; W. C. Greene, "The 
Spoken and Written Word," HSCP 60 (1951) 41; Nagy (supra n.5: 181) puts 
Herodotus within a different, oral tradition of logioi, though it is possible that 
he means to include writers like Hecataeus among his logioi. Cf supra n.41. 

Herodotus' many references to customs or monuments that were still pres­
ent "in my day" (E'n lCU l tr; t"Lt 'Vel sim.) likewise seem to imply that he 
orients his record towards an audience well past his lifetime: thus e.g. 1.66.4, 
93.3,181.2; 2.122.2,130.1,131.3,181.5; 3.97.4. 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

WILLIAM A. JOHNSON 253 

intervening): [<pa<Jt 'toue; <I>oivueae;] EQa1tUevEE<J8at Kat (5it Kat Ee; 
"Ap'Yoe;' ... 'A1ttKOLlEVO'\), (5£ 'toue; <I>OtVtKae; Ee; (5it 'to "Ap'Yoe; 
(1.1.2); (4) loose association of thought between the clauses of a 
sentence: 'tOY (5£ ol(5a au'toe; 1tpiihov 1mupsav'ta a(5tKWV EP'YWV 
Ee; 'toue; "EAA:r\vae;, 't01hov <JT\JlTtvae; 1tpo~TtcroJlat Ee; 'to 1tpocrw 
'tou Aoro'\), OJlOtWC; JltKpa Kat Jlf'YuAa acr'tEa av8p<lmwv E1tEStroV 
(1.5.3; but how exactly is the fate of small and great cities related 
to the story of the man who first perpetrated wrongs against the 
Greeks?). The inclusion of 'oral' elements of style in the 
prologue makes it doubtful that these elements imply oral per­
formance. Pohlenz was careful to note that such elements of 
style are reflected in archaic writing generally. It may be, as 
Pohlenz asserts (210), that the archaic style develops from oral 
performance. Once established, though, the archaic style need 
no longer be restricted to oral performance. Nor is it clear that 
oral performance is the only profitable model for explaining the 
many peculiarities of archaic style. In any case, an argument for 
oral performance on the basis of an author's use of a style 
developed from oral performance is circular. Elements of oral 
style, even if they do exist, are not conclusive evidence for oral 
performance. 

I do not deny the possibility that Herodotus read parts of his 
work orally. Nor should this possibility be denied for Thu­
cydides or Xenophon. The abrupt change in style between 
Herodotus and Thucydides is often attributed to the influence 
of Gorgias. If so, Thucydides' model of style is more 'oral' in an 
immediate sense than Herodotus' models. Oral influences on 
early Greek authors are, I believe, neither so clear nor so simple 
as generally supposed. In trying to define the differences 
between the 'archaic' style of Herodotus and the 'classical' style 
of Thucydides, we may properly speak of the oral traditions in 
which the archaic style is grounded, but to relate Herodotus' 
style and method of composition to actual oral performance 
lacks foundation. 

Epideixis at Olympia is one thing, a monumental history is 
another. The very idea is different. What Herodotus tried to 
accomplish, like Thucydides later, was fundamentally new. 
Prejudicial reaction to Herodotus' style should not induce 
scholars to view the text as a conglomerate or as a work 
proferred piecemeal to an adoring public. Only a few will have 
read Herodotus' monumental history in his own time and a full 
reading with the requisite concentration could only have been 
sustained in a private setting. But that the reading of the 



JOHNSON, WILLIAM A., Oral Performance and the Composition of Herodotus' "Histories" , 
Greek, Roman and Byzantine Studies, 35:3 (1994:Autumn) p.229 

254 ORAL PERFORMANCE AND HERODOTUS' HISTORIES 

Histories caused a stir among the leading intellectuals of the day 
need not be questioned. Moreover, the intelligentsia's accep­
tance of a work of this type and size will have opened the door 
to further innovation; for an intellect like Thucydides, it will 
have demanded it. These innovations of style and method, 
finally, will be what distinguishes Thucydides from Herodotus. 
There is no need to resort to facile and misleading arguments 
based on the determinative influence of oral performance. 59 

THESAURUS LINGUAE GRAECAE, IRVINE 

June, 1995 

59 I am pleased to record my debt to C. J. Herington and R. A. Kaster for 
helpful remarks and criticisms on an earlier draft of this paper.