Who Read Herodotus' Histories?

Author(s): Stewart Flory

Source: The American Journal of Philology, Vol. 101, No. 1 (Spring, 1980), pp. 12-28
Published by: The Johns Hopkins University Press
Stable URL: [Link]
Accessed: 05-05-2020 05:48 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@[Link].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
[Link]

The Johns Hopkins University Press is collaborating with JSTOR to digitize, preserve and
extend access to The American Journal of Philology

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
 WHO READ HERODOTUS' HISTORIES?

Did Herodotus' Histories achieve wide popularity im

ately after it was published? Modem scholarship routi
sumes that it was an "instant success"' or that its audience
was comparable to the audience for Homer and the trage-
dians.2 Evidence for widespread literacy in the late fifth cen-
tury seems to indicate a large public eager to acquire and read
books. Since reading aloud before an audience was the ancient
norm, Herodotus' work could presumably have reached illiter-
ates as well. Both ancient anecdotes about Herodotus' life and
adaptations and parodies of his work by his contemporaries
appear to show how popular the Histories was. The light tone
of the book itself and its many anecdotes also contribute to the
picture of an author anxious to please a large public and confi-
dent that he could do so. One critic says, for example, of a
certain Herodotean passage: "The object was, as it were, to
get people laughing while you sent the hat around."3
But we know little of how long prose books were used in the
fifth century-compared, for example, with what we know of
how tragedies were performed at Athenian festivals. Schol-
arship, moreover, does not agree on why Herodotus wrote his
book or even on the book's main subject if it has one. Caution,
therefore, dictates that in estimating Herodotus' audience we
discard all analogies with "best-sellers" of any modern era.
The purpose of this paper is to show that Herodotus'
book-if we mean by this term the entire book we now call his
Histories-was too long and therefore too unwieldy to become
truly popular in its author's day. This book requires a taste for

1 T. S. Brown, The Greek Historians (Lexington 1973) 25. Cf. J. E. Powell,

The History of Hdt. (Cambridge 1939) 77 n. 15. But see J. Well, Studies in Hdt.
(Oxford 1923) 170: "It is somewhat surprising that Hdt.'s work should even
have been sufficiently familiar for it to be parodied at all." Even those who
believe death prevented Hdt. from completing his work assume its speedy
publication and success, e.g., F. Jacoby, RE Supp. 2, 376, 378.
2 S. Usher, JHS (1978) 174: "His masters are the tragic dramatists and, most
of all, Homer, his audiences the same as theirs."
3 K. H. Waters, Antichthon 8 (1974) 6.

American Journal of Philology Vol. 101 Pp. 12-28

0002-9475/80/1011-0012 $01.00 ? 1980 by The Johns Hopkins University Press

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
 WHO READ HERODOTUS' HISTORIES? 13

leisurely reading, much more, in other words, than th

facility of literacy.4 As for recitation, the length of t
quired to get through the whole Histories would have
the number of performances and discouraged a large a
If one distinguishes Herodotus himself, as a raconteur-
lecturer, from his book, and if one distinguishes the book's
thousands of anecdotes and snatches of information from the
book itself as a whole, there is no convincing evidence to show
that the work was even widely known in the late fifth century.
To support this thesis we must address three critical issues: the
length of Herodotus' Histories, the level of literacy in his day
and the validity of using allusion in tragedy and comedy to
prove his book's popularity.

The length of the Histories made it clumsy to handle and

awkward to perform. Twice as long as either Homeric poem,
the book would have required a strip of papyrus about one
hundred meters long, divided into at least thirty separate book
rolls.5 Its sheer bulk tells against the number of copies which
could have been quickly made and easily circulated.6 Athe-
nians of the fifth century lacked the copyists, librarians and
"readers" who catered to later bibliophiles.7 Unlike epic po-

4 J. A. Davison, Phoenix 16 (1962) 155, quotes Trollope's distinction be-

tween "the absolute faculty of reading" and "the adequate use of a book."
D.'s whole article (141-56, 219-33) is significant, though Hdt.'s popularity is
not directly discussed.
5 T. Birt, Das antike Buchwesen (Berlin 1882) 444, calculates the length of a
text of Thuc. as 81 meters (though B.'s Grossrollensystem has not won ac-
ceptance). J. Cerny, Paper and Books in Ancient Egypt (London 1952) 8-11,
thinks that a roll about 6 meters long would have been easy to handle. For the
original, pre-Alexandrian, book divisions of Hdt. see: W. M. F. Petrie, "The
Structure of Hdt. Book 2," JHS 28 (1908) 275-76; supported by H. T. Wal-
linga, Mnemosyne 12 (1959) 204-23; objections raised by T. S. Brown, AJP 86
(1965) 72-75. But see now S. Cagnazzi, "Tavola dei 28 Logoi di Erodoto,"
Hermes 103 (1975) 385-423. For a Hdt. papyrus dating from before "our" book
divisions: C. B. Welles, TAPA 70 (1939) 208.
6 T. C. Skeat, PBA 42 (1956) 180-208. Among the difficulties S. cites (183) is
the lack of appropriate furniture.
7 Plin., Ep. 3.5. We do not know how multi-roll works were stored or la-
belled in the fifth century. Alexis (c. 375-275), Fr. 135 (Edmonds), has

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
 14 STEWART FLORY

etry, Herodotus' work depended for its promulgation

cumbrous physical self. Not only professional rhapso
also a traditional education based on memorization of
kept even long poems like the Iliad before the pub
relatively few texts, and the texts themselves were not u
actual performances.8 But because Herodotus' book w
long for memorization, a reading required a text.
A reading of the whole text would have required fi
more hours, a team of readers and a number of days e
the length of any known festival.9 We can dismiss as u
for example, Lucian's story that Herodotus read his b
Olympia, a four day festival, and that each of his bo
there granted the name of a muse.10 Herodotus prob
give readings of excerpts from his book or used mater
public readings in his book, but we can remain skeptica
or others regularly or indeed ever read the whole wo
large, official gathering. A special commission or com
performance before a captive audience might have left
anecdotal tradition, and the extravagance of the conj

Herakles pick a cookbook out of a collection of rolls (on a shelf?) by loo

ETiyQa,uta. Pictorial evidence in T. Birt, Buchrolle in der Kuns
1907). Re-rolling was tedious and damaged books (Mart. 1.66.8; 1
Verginius Rufus (Plin., Ep. 2.1.5) broke his hip and eventually
injuries sustained while reading. Small wonder the ancients quoted fr
ory.
8 Xen., Symp. 3.6. The performing equipment of a rhapsode was
lyre, not a book roll.
9 The estimate for performing time is conservative, the result o
experiments. Cf. J. A. Davison, GRBS 6 (1965) 24 (24 hours for a team to
perform the Iliad).
10 Lucian, Herodotus 1-2. Only this anecdote contains the detail about the 9
books (for others, see How and Wells, Commentary 1, p. 6). For the division
into 9 books see above, note 5 and W. Aly, Rh. Mus. 64 (1909) 591-600.
Paroemiographi Graeci s.v. E rTIv 'HQobo6ov oxtdv (Gaisford, p. 135;
Leutsch-Schniedewin 1, p. 400) preserves an anecdote which explains that
Hdt. did not lecture at Olympia because he could not find any shade. This is
misunderstood by J. L. Myres, Hdt. Father of History (Oxford 1953, repr.
Chicago 1971) 5: . . .lectured overlong at Olympia." Davison (note 4 above)
speculates (155): "[Thuc.] does not mean that Hdt. wrote to win a prize in a
competition, but that he wrote a work to be read (at least in snatches) at an
aycov (Olympia, Delphi, the Panathenea, for examples), and that the people
who were bored with the official program might go to the literary man's booth,
listen for a while and then go away again."

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
 WHO READ HERODOTUS' HISTORIES? 15

necessary to account for Herodotus' book in this w

shows how singularly long the book was."1
The Athenian audience had a limited attention span
see from the relative brevity of individual tragedies a
edies and from Aristophanic banter describing the lon
spectators to slip away in mid-performance.12 How m
Herodotus' book could such an audience have absorbed at a
single sitting? How many would have returned day after d
for further installments? The Homeric poems achieved pop
larity despite their length, but the Iliad and Odyssey prove
exceptions to the rule that in an illiterate culture poets usual
produce single works which require only one session for pe
formance. Bards hired by coffee shops in Novi Pazar to lur
customers back for every night of Ramazan sing thirty diffe
ent songs, not one long one in installments.13 We cannot e
plain why the Homeric poems broke this pattern or how, ov
the centuries, they achieved their unique primacy in Greek
culture except to say that the genius of Homer triumphed ov
every difficulty and attracted a universal audience. Herodotu
surely did not win in only a few months or even years a similar
triumph over equal or greater difficulties.
Herodotus' Histories was very probably the longest single
work written until the fourth century. If only Dionysus o
Halicarnassus commented-and then only obliquely-on ho
Herodotus overreached his predecessors: Trrv rE 7TQay7art
:tQoatQeoatv i rTO r eltov r veyxe xa a T.a QzJord eov (Th. 5
this was because the profusion of long books after Herodot
blinded critics to his work's unusual length.14 If some write

I Dio. Chrys. (Or., 37.7) implies that Hdt. produced a version of his histor
tailored to the biasses of each polis where he performed! Diyllus' 10-tale
bounty for Hdt.'s reading at Athens (Plut., de mal. Hdt. 26) might point to
Athenian commission if the amount were not so implausible and if Hdt
praise of the Athenians were not so tempered with criticism of them (e.g., 1.
and with sympathy for the Persians (e.g., 9.16).
12 Aristoph., Av. 786-89. If the audience's attention span had been longe
trilogies might have been consolidated into a single play.
13 A. B. Lord, Singer of Tales (Harvard 1960) 15.
14 For the myopia of ancient critics, see, in general, G. M. A. Grube
Phoenix 28 (1974) 73-80. Moder scholars too have anachronistic perception
how ancient authors worked: "[Hdt.] kept an interleaved copy of his bo
with him till the last" (J. W. Blakesley, Herodotus [London 1854] p. xliv).

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
16 STEWART FLOR Y

before Herodotus wrote longer books which have not su

vived, these tomes cannot have been numerous nor their
ences large. Even after Herodotus, long prose books ten
be catalogues, geographies or annals, whose length was t
result of accretion, not of organic unity. Thucydides wro
came close to finishing a book which, as we have it, is t
thirds the length of Herodotus. Thucydides, though he
have been trying to surpass his predecessor in length as w
in historical method, surely never expected his work to r
large audience of his contemporaries. Since his subject m
was current history, he could therefore write for later r
an account of each year's events as they happened, and
material lent automatic unity to his long book.s1
Thucydides, however, was an exception to prolific writ
Herodotus' day. They generally wrote many short work
one long one. We possess, for example, more than a sco
titles for Hellanicus, but for Herodotus only one book w
an official title. Where ancient testimonia speak of numb
lines in a poem or of books in a prose work, these numb
not suggest works even approaching Herodotus in s
Hecataeus' longest attested work is in four books and
neti06oc; rF is in only two, "Europe" and "Asia." 7 Down
the fourth century very few single works exceeded in le
even a single five-meter book roll: a tragedy, for exam
collection of Sappho or an :redeiett; by Gorgias. Prose w
tended to be shorter than poetry because prose was mor
ficult to learn by heart. Can we then imagine that the a
mous writer to whom Eryximachus refers in Plato's Sy

15 B. Hemmerdinger ("La Division en Livres de L'Oeuvre de Thuc.,

61 [1948] 104-17) believes that Thuc. wrote, after his long (and I would sa
"Herodotean") Book One, one book roll for every year of the war.
16 I use the unsatisfactory title Histories only because it avoids lab
periphrasis. For Hellanic.: FGrH 1, pp. 104-52 (some of the titles may
blets).
17 There is no way of knowing how long the books mentioned are, but
has no rivals for length even if we imagine that the longer, Alexandr
are intended by the testimonia. For Hec. see FGrH 1, pp. 1-47. Other
chosen at random: a Thebais in 7,000 lines (Homeri Opera 5 [OCT] p
Telegony in 2 books (ibid., p. 109); a nIeQi CPoa5ow; by Heracleitus divi
3 26yoi but in 1 (?) book roll (Diog. Laert. 9.6).

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
 WHO READ HERODOTUS' HISTORIES? 17

posium was able to stretch his subject, the utility of s

yond a single short book roll?18
Finally, is it reasonable to argue that Herodotus went
trouble of writing a long book if he merely intended to ab
shorter readings from it? Such a plan would not ha
rational since the author would have been creating for
the needless difficulty of locating passages in a long s
papyrus rolls. On the other hand, the Histories does h
patchwork quality which tempts one to argue that it
long book at all, but rather a series of poorly-joined s
ones.19 Yet no one claims that Herodotus was not, in th
aiming at unity. If Herodotus' book is not a perfect uni
flaw reflects his difficulty in putting a vast quantity of m
into comprehensible order. Nor has anyone ever se
suggested that a committee wrote or arranged our
Herodotus or that the text contains more than a few minor
interpolations by another hand. Herodotus surely did require a
long period of time to write such a long book, even if he wrote
at the prompting of a single, ongoing creative urge. He must
also have relied at least on the practice gained by earlier writ-
ing or recitation.20 But it would be wrong to think of possible
prior serial composition as an easy explanation for the length
of the book. The Histories was still hard to write and hard to
read.21 If Herodotus published his book piecemeal and then

18 Rhys Carpenter, Folk-tale, Fiction and Saga in the Homeric Epics (Berke-
ley 1946, repr. 1962) 15: "In defiance of what the thoughtless might imagine,
early written literature tends to compactness and brevity." If Eryximachus'
book on salt (Plat., Symp. 177B) became notorious (see Isoc., Hel. 12), this
might have been because few such books were actually in circulation.
'9 Poorly-joined though the parts may be in formal terms, they are often
linked by Hdt.'s casual references forward and back to what he has written or
will write, references which often span hundreds of modem pages and criss-
cross the whole work: 2.38.2 (to 3.28), 2.161.3 (to 4.159), 5.22.1 (to 8.137),
5.36.4 (to 1.92), 6.39.1 (to 6.103), 7.93 (to 1.171).
20 I find convincing the thesis of R. Lattimore, "The Composition of the
History of Hdt.," CP 53 (1958) 9-21, that our text is essentially an unrevised
first draft, though, as K. von Fritz argues (Die griechische Ge-
schichtsschreibung 1 [Berlin 1967] 115-18) some earlier writings could have
been pasted in. If Hdt. used revised versions of earlier writings in our text, the
physical difficulties of copying (notes 5, 6 and 7 above) make it likely that he
wrote them out from memory.
21 Dickens' long, serialized narratives would not have ended up as whole
novels if the author had not sought to satisfy the cravings of an established

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
18 STEWART FLORY

took little interest in collecting the parts into a whole, w

faced with the problem of accounting for our relatively orde
text. The book, moreover, does not fall into clearly-def
episodes nor does any part have a separate textual histor
Though at various times in his career Herodotus may hav
different ideas about what he wanted his book to be, he u
mately created the book we have now.23 If Herodotus' o
creative act was to join together writings which he had e
composed in a different order for different purposes, even t
mechanical act argues his intention to produce a book w
vastly exceeded in size the normal expectation for a prose
book in his day.

II

Evidence for literacy at the end of the fifth century is am-

biguous, and scholars have adopted widely divergent views.24
In general, authorities do agree that this period saw at least a
growth in literacy. We find testimony for books and reading in
art, in inscriptions and in the internal evidence of surviving
texts. Books were, however, enough of a novelty that even the

reading public for "three-decker' books. Hdt. could count on no such audi-
ence.

22 No division of the work into recitations (e.g., the Bude

grand) has been recognized as a true reflection of the author
Cagnazzi's division (note 5 above) or something like it is corr
arose only from the author's practical need to begin a new roll.
91 (1970) 181-91, speculates that Hdt. did write or intended
separate logoi, but the evidence is slender.
23 One way in which Hdt. could easily have made his book s
leaving any promises unfulfilled (see note 19 above) would h
lish Book Two separately, merely inserting a few bridge chapte
and Three.
24 E. G. Turner, Athenian Books in the 5th and 4th Centuries (London 1952),
argues for a very literate Athens. E. A. Havelock, Preface to Plato (Oxford
1963) 36-56, argues convincingly (I think) for considerable illiteracy in the 5th
century. For a complete review of the evidence, see F. D. Harvey, "Literacy
in the Athenian Democracy," REG 79 (1966) 585-635. Though Harvey allies
himself with Turner, the material which he has collected could be used to
support either side. See also: P. A. Cartledge, "Literacy in the Spartan Oligar-
chy," JHS 98 (1978) 25-37.

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
 WHO READ HERODOTUS' HISTORIES? 19

word "book" could be a cue for a laugh in comedy.25 Evidence

about the education of children shows that almost all adult
male citizens must have learned to draw the alphabet and
sound out letters.26 For Athenian democracy to function
smoothly, virtually every male citizen needed this minimal,
name-signing literacy.27 If some voters accepted ostraka which
were already inscribed with the name "Themistocles," this
does not mean that they could not write, any more than today
the acceptance of free transportation to the polls means that a
voter cannot walk.28 And if the average Athenian were as
completely unlettered as the bumpkin who needed help in
writing "Aristeides" on a sherd, the story which we find in
Plutarch would have lacked point.29 Democracy, however, did
not require any more than minimal literacy, and few Athenians

25 For the comic use of "book" see J. D. Denniston, CQ 21 (1927) 118-19.

Also, in Eupolis Fr. 304 (Edmonds): "I scoured the market-incense, frip-
peries, scents, onions, garlic and where the bookstall is" (Xov Tr& fltfi(' `vta),
note that "bookstall" comes last, possibly for comic effect. Most of the shops
here mentioned sell raw products so that a "stationary" store (cf. Mod. Gr.
XaQronTorw( ov) not a book store in our sense may be meant. In Eur., Hipp.
954, Theseus, cursing his son for what he mistakenly assumes are his bizarre
and avant-garde tastes, accuses him of "honoring the smoke of many letters"
(nroRic)v yQaujuJrtov ... xa:rvovgS).
26 See note 24 above and also K. J. Freeman, Schools of Hellas (London
1908); H. Immerwahr (works cited in note 31 below); E. Vanderpool, AJA 63
(1959) 279-80 (5th century schoolboy slates).
27 Harvey (note 24 above) abundantly proves this point. But he also asks the
question (586): "Was this experience [reading Hdt. and Thuc.] denied the
majority of their contemporaries?" In fact, none of the evidence which Harvey
cites specifically shows that Hdt. and Thuc. were read, and Harvey never in
the article directly answers his own question.
28 The "Themistocles" ostraka (Hesperia 7 [1938] 228-43) could have been
provided merely for the convenience of literates by an anti-Themistoclean
club. But I find it significant that of the 14 hands identified on the ostraka, the
clearest and best spelled (i.e., most literate) are most numerous. Thus if we
suppose that 14 men spent a set time writing "Themistocles" as many times as
they could, writers A and B, for example, wrote 32 and 33 sherds respectively,
but writers K and L only managed to do 8 and 4, a strikingly poor effort for
politically active Athenian males. Also, such playful uses of the alphabet
onstage as Eur. Fr. 382 (Nauck) and Callias (in Athenaeus 10.453) would have
intrigued the barely literate but bored the fully literate.
29 If total illiteracy had been the norm, A.'s interlocutor would have been
simply "an Athenian," not: riva rtv ayQapupraTov xai 7ravrei;q dayeoixwv
(Plut., Vit. Arist. 7).

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
20 STEWART FLORY

can have needed to or wanted to read through the long p

inscriptions.30 This minimal literacy, moreover, scarcely
vides the technical facility required for reading Herodotu
The evidence which we have for anything beyond min
literacy concerns the reading and writing of short, singl
books of prose and poetry. These are the only kinds of
to which we find references and which we see represente
vase paintings.31 But scholars do not agree on how comm
even these short books were. Certainly we cannot take at
value the two passages in the Frogs which imply that sailo
the Athenian navy read books when off duty or that the
goers brought along scripts or reference works.32 In add
some references to "books" could be to business documents
or even to blank papyrus. Literate cultures always produce
much more "documentary" than "literary" writing, and a
profusion of documents is no proof of general literacy. Thus,
there might have been substantially fewer literary books in
circulation than even the scanty evidence suggests.33
But no matter how common these short books were, they
were not read as a modern book is read. Silent reading greatly
increases speed of comprehension, but only a few questionable
passages suggest that reading silently was ever even con-
templated. Even when one was alone, reading aloud continued

30 Havelock (note 24 above) believes (39) that these inscriptions were only
used ". . .as a source of reference and as a check on arbitrary interpretation."
The stoichedon style was handsome but not particularly legible. It went out of
fashion in the 4th century partly because it was incompatible with word divi-
sion at the end of the line (i.e., legibility): R. Austin, The Stoichedon Style ...
(London 1938) 111.
31 Some illustrations in Birt, Die Buchrolle in der Kunst (Leipzig 1907), but
the definitive works are: H. Immerwahr, "Book Rolls on Attic Vases,"
Studies in Honor of B. L. Ullman 1 (1964) 18-48; Antike Kunst 16 (1973)
143-47.
32 See L. Woodbury, "Aristophanes' Frogs and Athenian Literacy," TAPA
106 (1976) 349-57 (with bibliography).
33 See note 25 above. Xen., Anab. 7.5.14, speaks of :ro2ai be PiP/ot
yeyeauueyvat among the flotsam from a shipwreck. The words, however, need
not describe literary books. There was ample reason in antiquity to transport
business documents, letters and state papers, writings which could not be
memorized and transported on the lips of men. Whatever Xenophon saw,
entirely too much weight has been placed on this isolated piece of evidence.
The normal cargo of papyrus (which had to be imported in any case) would
have been in the form of blank rolls.

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
 WHO READ HERODOTUS' HISTORIES? 21

to be the dominant pattern for centuries.34 There is also scant

evidence for solitary-and thus reflective-reading of any
kind. When an artist represents a reader alone, the vase or
stele is too small to show an audience.35 Moreover, reading a
book from a text was not generally an end in itself. Athenians
read poetry from a book only to memorize it as a prelude to
performance. Some prose, like the book of Anaxagoras which
Socrates had heard read from a book, was no doubt too techni-
cal to be memorized.36 But other prose, like the speech of
Lysias of which we hear in Plato's Phaedrus, was definitely
read only in order to be able to recite it later from memory in
front of an audience.37 Both Euripides and Aeschylus cele-
brate literacy because it is "a remedy for forgetting," not be-
cause it provides an amusing or instructive pastime.38 Books

34 See W. C. Greene, "The Spoken and the Written Word," HSCP 60(1951)
23-59; E. S. McCartney, "Notes on Reading and Praying Audibly," CP 43
(1948) 184-87. References to silent reading arise in special contexts: An-
tiphanes Fr. 196K (a riddle); Eur., IT 760-63 (a secret letter). Even Thuc.
presumes that even though his work may not give pleasure when it is read
aloud, it will, if it is read at all, be read aloud (dxo6aatv [1.22]).
35 See Immerwahr (note 31 above). Turner (note 24 above) suggests that the
fifth-century Attic grave stele now in Grottaferrata depicts a young man read-
ing a prose book to himself "and pausing perhaps to meditate on what he reads
(15)." G. A. Richter, MDAI(A) 71 (1956) plates 2-5, has excellent photographs.
But the boy's gaze might just as well reflect the artist's perception of the
pathos of an early death. Eur. Fr. 910 (Nauck), however, does sound like a
celebration of private reading, but Eur. was an eccentric known for his library
(Athenaeus 1.3).
36 PI., Phd. 97c. The books of Anaxagoras which were so cheap and avail-
able (PI., Ap. 26D) must have been mere pamphlets, or else the passage is
ironic. Cf. Protagoras reading at the house of Eur. (D.L. 9.54). The Suda (s.v.
'Perikles') has it that P. was the first to deliver a "written" speech (z7rQbro
yQgazTv A6oyov ev 6txaaToruQi) EcE). Surely, however, the contrast here is with
earlier, ex tempore speakers, and P. did not read from a text. For poetry
recitations: e.g., PI., Prt. 325E.
37 Though Plato is a hostile witness on books, the underlying assumption is
that the young man is doing nothing unusual in borrowing a book to memorize;
it is the nature of the book which appalls Socrates. If thoughtful reading,
divorced from memorization, had been more common even in the 4th century,
Plato might have been better-disposed to books. But if even a non-rhapsode
could, with some effort, memorize all of Homer (Xen., Symp. 3.5), one did not
need to own a text. Even for a teacher or a devoted bibliophile to own a
complete Homer was an eccentricity (Plut., Alc. 7; Xen., Mem. 4.2.1, 10).
38 Eur., Fr. 578 (Nauck); Aesch., PV 459-61. Cf. PI., Phd. 275A. But see note
35 above.

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
22 STEWART FLOR Y

were vzro/vriara, "prompt copies." They were precious o

jects, which preserved the ipsissima verba of a work for po
terity and prevented deviation from the author's words by
performers.39 Book roll scenes on Attic vases thus evident
show would-be performers using ltpiAia to master a tex
Where the artist has included a legible text on the book roll
the scene, this writing turns out in every case to be a diffic
piece of poetry and not prose.40 Thus we can compare the
demand for books in Herodotus' day to the demand for musi
scores now, when even halting readers of music are able to
a score to master a new piece.41
In the late fifth century readers wanted short books becau
they could be memorized for later public performance or i
some cases read aloud before an audience at a single sitting
Also, in the absence of long books, readers could not easily
develop a taste for anything else. Therefore, even presumi
that all readers of short books possessed perfect facility in
reading per se, Herodotus could not have written to fill a lar
demand for the kind of book which he did write. Of course,
Herodotus did not write a book which was impossible to read
nor did he write in a total vacuum. Rather he was part of
tradition of logographers and genealogists whose works nece
sarily created a small literary elite and perhaps even a group
salons. But even the members of this elite had formed their
tastes and honed their skills on much shorter books. The His-
tories must have seemed an even longer and more difficult
book to them than it does to us.

39 Thus we hear of a copy of Hesiod on lead plates at Dodona (Paus. 9.31.4).

Cf. the ms. of Heracleitus (D.L. 9.6).
40 Immerwahr (note 31 above, Ullman): "The popular image of the book was
not the prose book nor the great epic or tragic text, but the epic tale, the short
hymn, the gnomic collection and the collection of lyric poems" (48). This
contradicts the view (now generally discounted) of Wilamowitz, Einleitung in
die griechische Tragodie (Berlin 1921) 121-28, that tragedies were the first real
books in the modern sense. The vases also fail to confirm the conjecture of
L. D. Reynolds and N. G. Wilson (Oxford 1968, repr. 1975) 1: "The first
works to reach even a modest public were either the writings of the Ionian
philosophers and historians or those of the sophists."
41 G. L. Hendrickson, "Ancient Reading," CJ 25 (1929) 182-96, esp. 184,
explores the musical analogy.

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
 WHO READ HERODOTUS' HISTORIES? 23

III

References to Herodotus in tragedy and comedy appear,

however, to imply that the book was a popular success. Com-
mentators even allege that this book achieved for a while the
currency of a modern best-seller and that it was widely quoted,
imitated and parodied.42 But the uncertainty of the allusions, if
we examine them cautiously, does not support such a daring
allegation about such an extremely long book. Even if we can
allow in some cases the strong possibility that playwrights had
read the book, we have no way of knowing whether their audi-
ences had too. Doubt remains even in the case of supposed
parodies since the object of the parody, necessarily limited in
scope, could be a lecture by Herodotus, not his book. Only the
appearance on the comic stage of a character called
"Herodotus" carrying bundles of book rolls would convince
us that the average Athenian knew that Herodotus had written
a long book.
Scholars have attempted to use allusions in tragedy and
comedy as a terminus ante quem for Herodotus' publication
date. It is immaterial to our thesis here when Herodotus' book
was published, but studies which attempt to fix this date often
give the impression of proving what in fact they only
assume-namely that Herodotus' book was popular. The pre-
sumption is that only a "best-seller" would be the object of
allusion in a popular medium like the ancient theater. Because
Herodotus' book must have been popular, the argument goes,
a reference in a play to, for example, Egypt must be a refer-
ence to Herodotus, and therefore Herodotus must have pub-
lished his book just prior to the date of the play in which we
find the supposed allusion. Charles Fornara and Justus Cobet,
however, in recent, separate articles re-examining Herodotus'

42 See notes 1 and 2 above. The most complete list of the supposed imita-
tions and parodies is in Schmid-Stahlin 2, p. 318 note 3 (Sophocles); p. 591 note
2 (Aristophanes); p. 663 note 4 (Euripides). For analyses of some individual
passages (none of them, in my opinion, true allusions to Hdt.) see: R. Brown-
ing, "Hdt. 5.4 and Eur., Cresphontes Fr. 449N," CR 11 (1961) 201-2; J. E.
Powell, The History of Herodotus (Cambridge 1939, repr. 1967) footnote, p. 34
(Soph., Phil.-Hdt. 6.75); A. J. Podlecki, "Creon and Hdt.," TAPA 97 (1966)
359-71 (re. Hdt. 3.80-82).

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
24 STEWART FLORY

publication date, are much more cautious and circumspect

sifting the supposed allusions to Herodotus.43 Fornara and
Cobet disagree on the publication date, but, whichever is rig
the net result of their endeavors is to cast doubt on virtually
of the commonly-cited allusions to Herodotus. We can now
doubt in almost every case whether these are actually ref
ences to Herodotus, or whether they in fact presume the au
ence will recognize an allusion. Few of the supposed allusio
remotely recall Herodotus' language, and the most likely
echoes are confined to cliches like: "In a man's long life many
things can happen" or to brief phrases like: "I measured it
myself."44 In other cases it is impossible to tell whether a
playwright and Herodotus may not be using a common
source.45 This latter objection applies particularly to informa-
tion about barbarian lands, information which could easily
have been common knowledge in Athens.
Specifically, Fornara attacks the common view that allu-
sions in the Acharnians show that the Histories had been re-
cently published in 426.46 He shows convincingly how flimsy
these allusions really are. Instead, Fornara attempts to prove
that allusions in the Birds and a cycle of Euripidean plays on
barbarian themes date Herodotus' publication to shortly be-
fore 414.47 Cobet disputes Fornara and shows that Euripides
had also treated similar themes earlier. He concludes that
some of the passages in the Birds do depend upon Herodotu
but, demonstrating that the parody is blunt and general, h
concludes that the historian must have published many yea
earlier.48 Cobet returns to the Acharnians and to the traditional

43 C. Fornara, 'Evidence for the Date of Hdt.'s Publication," JHS 91 (1971)

25-34; J. Cobet, "Wann wurde Hdt.'s Darstellung der Perserkriege pub-
liziert?" Hermes 105 (1977) 2-27. Cf. O. J. Todd, CQ 16 (1922) 35-36. A new
contribution to the debate, by J. A. S. Evans, is forthcoming in Athenaeum.
44 Soph., Phil., 305-6-Hdt. 1.32.2; Ar., Ai. 1130-Hdt. 2.127.1.
45 E.g., Eur. Fr. 646N (a reference to Egyptian mummies). In another case,
Av. 486-87, Ar. gives a detail about the Persian headdress (xvQopaaia) not
mentioned by Hdt. (5.49.3, cf. Xen., An. 2.5.23). Fornara believes (29) that
Hdt. was the first to describe Babylon, but Cobet (13) rejects this.
46 The argument of Wells (note 1 above) 175f., had been based mostly on
Ach. 68-92.
47 Fornara (note 43 above) 25-32.
48 Cobet (note 43 above) 14-18.

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
 WHO READ HERODOTUS' HISTORIES: 25

date, but even then he concedes that many allusions in

are doubtful. Cobet's argument centers on one pass
celebrated Aristophanic conception of the origins o
Peloponnesian War (Ach. 523-28) with its apparent coun
in Herodotus' account of the origins, in mythical time
enmity between East and West (1.4). Since Cobet belie
this is both a genuine allusion and a parody which the
will recognize, we must follow his argument here clos
humor of both Herodotus and Aristophanes, says Cob
pends upon comic reversal or devaluation: heroines be
whores, and affairs of honor become petty crimes. He
parallel between Aristophanes: ravra /Efv bi o
xa&tcX(tLa (523) and Herodotus: rafvra uEv 6rj icaa :r
(1.2). Cobet then sees Aspasia playing the role of the
Helen and Perikles that of the jilted Menelaus. Cobet now
addresses the question of whether Aristophanes could have
expected his audience to understand the parody: "Eine solche
Parodie auf die ersten Kapitel von Herodots Darstellung der
Perserkriege in Zusammenhang mit einer komischen Kritik
des neuen grossen Krieges dem Zuschauer keine intimen
literarischen Kenntnisse abverlangte und gewiss eine grosse
Chance hatte, erkannt zu werden."49
Cobet's argument is powerful, and yet he cannot fully lay to
rest the possibilities raised by Fornara and others that a parody
of Euripides' Telephus or some other play is intended or that
Aristophanes was inspired by an actual event and the echo is
coincidence. And would Aristophanes have created a parody
which had only "a great likelihood" of being understood?
Cobet's ingenious argument that a parody of the opening
chapters of a book had a better chance of being recognized
than a parody from the body of the text rests upon the suspi-
ciously modern analogy that many readers will buy a much-
advertised but difficult book and then read only the first chap-
ter before setting it aside. Also, if Cobet's earlier, detailed

49 Cobet, 10-12. C. also believes that some of the spread of references may
be due to publication in sections or to lectures (20). I cannot agree with C.'s
statement that Hdt.'s book would have remained in the memory longer than a
tragedy (7). K. J. Dover, Aristophanic Comedy (London 1972) 188-89: "A
majority in an audience can be surprisingly tolerant of a parody which only a
few can really appreciate."

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
26 STEWART FLOR Y

analysis is correct, Aristophanes' humorous barb is worki

directions parallel to Herodotus, and Herodotus cannot
therefore be the butt of the joke. Surely Aristophanes intends
his audience to laugh at Perikles and Aspasia, not at
Herodotus, and they can have their laugh without any knowl-
edge of Herodotus at all.
It is unlikely, moreover, for the external reasons we have set
forth above, that Aristophanes would have parodied
Herodotus. The comic poet could not have had the same ex-
pectation of an audience's reaction to an allusion to Herodotus
that he must have had when he parodied Euripides, for exam-
ple. Thirty thousand Athenians saw each of Euripides' plays.
Music, dance and the excitement of a long-awaited festival
impressed such performances upon the expert memories of
their audiences. Herodotus' book, however, belonged to a
completely different genre and was thirty times the length of a
tragedy. All one can say in conclusion is that enough of the
alleged allusions are sufficiently convincing to allow the strong
possibility, approaching certainty, that Sophocles, Euripides
and Aristophanes, at least, had at various times read
Herodotus.50 But these men belong by definition to the Athe-
nian elite. And it was only this small group which, according to
our estimate of the size of Herodotus' book and the literacy of
his day, was equipped to read and appreciate the Histories.

* * *

The size of Her

it that he was t
of poetry or of
smaller circle. T
as a frivolous po
considerably sho
appealed to the
Thucydides disd
(1.22), he may v
But our book by

50 Scholars-perhaps
the two writers-hav
author (see note 42

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
 WHO READ HERODOTUS' HISTORIES? 27

and subjected to affectionate imitation as well as c

could not have been an dycy_bvtoa in a competition
rivals and could not have been heard routinely from b
to end.51
If Herodotus' book was not widely read in his own day, what
external changes in the ancient world allowed books of similar
length to become relatively common so soon afterwards?
Technical innovations and changes in taste may have been
responsible and require no special explanation. But such long
works never became truly popular-and never have become
popular, with the possible exception of the Bible and the
Koran-in the same sense that tragedy was in the fifth century.
Certainly the number of serious readers increased in the cen-
turies after Herodotus' death, but this was more a growth of
erudition in the highest strata of society and a wider dispersion
of such erudition around the Mediterranean than a pronounced
increase in general literacy. The chief effect of the triumph of
books over memory as a vehicle for serious artistic verbal
expression was to create a new genre, "literature," which was
not accessible to mass audiences.
Herodotus may have been inspired, like Homer, by writing
itself. Unlike Homer, Herodotus may also have been inspired
by increasing literacy in his own day to write a work of ex
traordinary length. Herodotus' prose work, like Homer's po-
etic one, seems to partake of both oral and written genres. Bu
when Herodotus came to write his book, literacy had acquire
a culture of its own to encroach upon his oral style.52 Parry and
Lord observed that the Serbian bards who later became literate
produced inferior, pedantic poems because literacy had intro-
duced them to the outside world of newspapers and to the
expression of ideas not suited to their poetic tools.53 If Homer
used writing in the composition of his poetry and yet escaped
this deterioration, it was because in the eighth century writing
was totally new and devoid of any alien cultural associations.

51 The imitative style of Thuc. 1.128-39 confirms, I think, the ancient anec-
dotal tradition that Thuc. was an admirer (and reader) of Hdt. (Marcellinus,
Vita Thulc. 54).
52 G. P. Goold, "Homer and the Alphabet," TAPA 91 (1960) 272-91, esp.
288-91; A. M. Parry, "Have We Homer's Iliad?" YCS 20 (1966) 177-216, esp.
212-14.
53 A. B. Lord, TAPA 84 (1953) 129-33.

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]
 28 STEWART FLOR Y

But Herodotus had to contend with the demands of both writ-

ten and oral modes of thought, and it is this struggle, in addi-
tion to the physical difficulties of writing, which may account
for the many problems we find in his book: the prolix digres-
sions, narrative seams, unfulfilled promises and outright er-
rors. In one important respect, however, Herodotus does not
resemble the literate guslars, for one of the symptoms of their
eroding skills is that their poems become shorter after they
become literate. The great length of Herodotus' work, like
Homer's, suggests the power of the creative will which was
able to break an established pattern.
While the great length of Herodotus' book is evidence of its
author's seriousness and contributes to its grandeur and com-
plexity, this great length, it must be admitted, is also a problem
even for modern readers. Like Pascal who made his letter long
because he had not time to make it short, Herodotus wrote on
and on because he had not the facility to revise and condense.
Opinions will differ as to whether a shorter and tidier book
would have been better, but we ought to give Herodotus credit
for going on when few of his contemporaries, he knew, could
measure his book's worth. Certainly we cannot dismiss any
parts of Herodotus' book as crowd-pleasing ornament, be-
longing to a different and lower level of intent than the whole.
Herodotus attempts to understand the world by writing about
it, and there is no evidence that he is trying to provide casual
entertainment for a general audience. Thus, there is in the very
size of Herodotus' ungainly book a germ, if only a germ, of
Thucydides' bold appeal to posterity.54

STEWART FLORY
GUSTAVUS ADOLPHUS COLLEGE

54 I owe particular thanks to W. Lee Pierson, president of Athens Colle

for hospitality extended to me during research for this paper. I would also li
to thank J. A. S. Evans, Georg Luck and the anonymous reader of this jour
for helpful suggestions and corrections.

This content downloaded from [Link] on Tue, 05 May 2020 05:48:54 UTC
All use subject to [Link]