Silent Reading in Antiquity Knox, Bernard M W Greek, Roman and Byzantine Studies; Winter 1968; 9, 4; ProQuest pg. 421

Silent Reading in Antiquity 
Bernard M. W. Knox 

SINCE the publication of Josef Balogh's article "Voces Paginarum,"l 
it has become standard doctrine2 that silent reading (and writ­
ing) was, if not completely unknown in the ancient world, at 

least so rare that whenever it was observed, it aroused astonishment, 
even suspicion.3 This conclusion seems at first sight reasonable enough; 
Greek literature, at least up to Thucydides, was intended for public 
delivery or performance, and from the early part of the fourth cen­
tury B.C. on to the end of antiquity, rhetoric was the foundation and 
eloquence the aim of the educational process. The evidence assembled 
by Balogh (which should be supplemented by that of G. L. Hendrick­
son)4 does indeed make it perfectly clear that the normal way to read 
a literary text (non-literary texts are a different matter and will be 
dealt with later) was out loud, whether before an audience, in the 
company of friends or alone. But Balogh's insistence that silent read­
ing was not just unusual but almost unheard of seems to go too far; 
common sense rebels against the idea that scholarly readers, for ex­
ample, did not develop a technique of silent, faster reading. Are we 
really to imagine that Aristarchus read aloud all the manuscripts of 

1 Philologus 82 (1927) 84-109, 202-240. The original version of the article had already been 
published in Hungarian in 1921. The most important items of his evidence were assembled 
in the "Nachtrage" to E. Norden, Die Antike Kunstprosa IS (Leipzig/Berlin 1915). 

2 E. G. Turner, Athenian Books in the Fifth and Fourth Centuries B.C. (London 1952) 14 n.4, 
expresses partial and qualified dissent. Though he finds that "for the Roman period" 
Balogh has "brought forward convincing evidence for reading aloud," he has himself pro­
duced evidence for silent reading in fifth and fourth century Greece, on which see below. 

3 Balogh's aim (p.87) is to prove "dass das Altertum alles, was wir heute stumm fUr uns 
lesen, stets laut las." 

4 "Ancient Reading," C] 25 (1929) 182-96. Though published later than Balogh's article, 
this was read before an audience in 1921 ("rnost of the illustrative material cited was 
gathered before the recent publication of Josef Balogh") and is in any case valuable for its 
more judicial tone and its citations from classical authors (the bulk of Balogh's material is 
from Christian and mediaeval writers). Eugene S. McCartney, "Notes on Reading and Pray­
ing Audibly," CP 43 (1948) 184-87 has little to add; his evidence from Borneo, the New 
Hebrides, Dean Swift, the Babylonian Talmud and the population of Malta in the air-raid 
shelters is interesting but irrelevant. L. Wohleb, "Ein Beitrag zur Geschichte des lauten 
Lesens," Philologus 85 (1929) 111-12, contributes an example of reading out loud from the 
Passion of Saints Firmus and Rusticus. 

421 



422 SILENT READING IN ANTIQUITY 

Homer he used for his edition? That Callimachus read aloud all the 
works from which he compiled his 120 volumes of Pinakes? That 
Didymus wrote his more than 3,000 volumes and read the countless 
books on which he based them, pronouncing every syllable out loud? 
Such doubts, far from being stilled by a second look at Balogh's 
evidence, grow stronger with every fresh example he offers. 

Most of them are Roman, and from late antiquity; he could find 
very little evidence for Greece. Augustine is his main authority (he is 
cited eight times in the article), and the famous passage which de­
scribes the silent reading of Ambrose and Augustine's reaction to it 
(first cited in this context by E. Norden) is Exhibit A. Augustine's 
amazement should however be judged in the light of the facts that he 
was not only a professor of rhetoric but also an African provincial from 
a poor family; Ambrose, the son of the praefectus of Gaul, had been 
educated at Rome, and before becoming Bishop of Milan, had been 
consular prefect of the province of which Milan, then the imperial 
residence, was the capital. The two men came from different worlds, 
and there is no reason to suppose that Ambrose's silent reading would 
have excited such comment in the Italian imperial circles he had 
deserted to become Bishop. 

In any case, Augustine, though he was at a loss to explain Ambrose's 
practice, was quite capable of adopting it, for at the supreme crisis of 
his life, when he heard the child's voice and took up the Epistle to the 
Romans to read, he read it, as he tells us, silently (in silentio, Conf 8.12). 
Balogh explains this as the effect of great emotion, which at this fate­
ful moment, robbed him of his voice. (Augustine does not say so, and 
Balogh has to back up his explanation with a quotation from St 
Gregory Nasianzenus.) Yet, in another passage cited by Balogh, 
Augustine's later reading of the fourth Psalm at Cassiacum, where he 
wishes the Manichees could hear him, he is in a similar ecstatic state of 
religious excitement but has no difficulties with his voice. Moreover 
he is not just reading; he is also talking to himself in the presence of 
God (mecum et mihi coram te, 9.4.). The phrases he uses make it clear 
enough, as Balogh admits, that what he wishes the Manichees could 
hear is not only his reading of the psalm but also his fervid regrets for 
his past errors. Although not one of these passages from Augustine is 
as cogent proof as Balogh thinks, his other example of Bible-reading is 
indeed incontrovertible evidence of solitary reading aloud. But it is 
one which has little bearing on the question of the reading habits of 



BERNARD M. W. KNOX 423 

educated Greeks and Romans; it is the case of the Ethiopian eunuch 
reading the prophet Isaiah in Acts 8.30.5 

With Horace, however, we are in the mainstream of classical culture 
at its most sophisticated. Balogh cites Sat. 1.6.122-3-aut ego lecto aut 
scripta quod me tacitum iuvet-with the comment (p.90), "Horace puts 
special emphasis on the fact that 'he takes pleasure in silent reading'." 
The passage is more difficult and obscure than Balogh seems to real­
ize. His explanation suggests that though he prints iuvet he is really 
paraphrasing the iuvat which is reported in some manuscripts but 
printed by none of the modern editors. This would give the sense, 
rafter reading or writing, which I like to do in silence .. .'; but even so, 
Balogh's case would not be as strong as he claims. Horace would in 
that case be distinguishing his own habit of silent reading and writing 
from that of, for example, Cicero, who in a very similar description 
(Ad Fam. 9.20) of his daily routine says, aut scribo aut lego, veniunt etiam 
qui me audiunt quasi doctum hominem . .. I cannot see that Horace would 
be putting any special emphasis on the fact; the tone of the passage 
shows no consciousness that his procedure might cause astonishment 
or need explanation-in fact it is a parenthetical detail in a description 
of his relaxed and very private daily routine. However, Balogh seems 
to agree with all the modern editors that Horace wrote iuvet, and with 
this reading the lines mean something quite different (and even less 
helpful for his thesis). With iuvat, the quod can refer directly to the 
antecedents implied in lecto aut scripto, as I have translated it and as 
Balogh's paraphrase seems to take it (though it could also refer to an 
understood antecedent in the accusative, which would give the sense: 
rafter reading or writing something which pleases me in silence'). But 
with iuvet it cannot refer directly to the participles, for the force of the 
subjunctive in the relative clause is to characterize the antecedent as a 
member of a class or type. This makes no sense at all if the antecedent 
is scribere aut legere, and the phrase must be understood as Bentley 
paraphrases it, cum iam tunc aut legerim aut scripserim qUid, quod me 
tacitum iuvet-'the sort of thing that pleases me in silence'. With this 
reading Balogh's thesis leaps from the frying pan into the fire, for we 
are now presented with a whole class of writings of such a kind that 
Horace likes to read (and write) them silently. What they can possibly 
be I have not the faintest idea, and the best solution of the difficulties 

G Even less to the point is the citation from the Talmud on pp.l03-04. 



424 SILENT READING IN ANTIQUITY 

raised by Balogh's insistence on understanding tacitum literally is to 
return to the interpretation of the word offered by all commentators 
and translators before and after Balogh: 'in quiet moments' (Fair­
clough), 'von niemand gest6rt' (Mi.ilIer).6 Whatever the lines mean, 
they are in any case dubious evidence for the abnormality of silent 
reading. The situation Horace describes is worlds apart from the terms 
of the rubric under which Balogh places it (p.88): "cases in which the 
reader for special reasons was forced (gezwungen) to depart from the 
general custom and read silently, in which one circumstance or an­
other hindered him (behindert) from proceeding with his reading out 
loud." 

The remaining passages which deal with the reading of literary 
texts come from strange sources; they are, as Balogh says, a 'bunte 
Reihe'. Among their authors are St Gregory Nasianzenus, Paulus 
Diaconus, Johannes Cassianus, St Athanasius, St Benedict, Petrarch 
and Grimmelshausen. I cannot for the life of me believe that they can 
tell us anything about the reading habits of Euripides and Callimachus, 
of Horace and Vergil. Only two passages merit discussion. A carmen 
Priapeum (Buecheler 68) presents us with a Priapus statue which has 
learned Homer from hearing the owner of the orchard read it (domi­
num totiens audire legentem). All this proves is that a certain ancient 
Roman, alone in an orchard, read his Homer out loud-but so, on 
many happy occasions, have I. More important is a passage from 
Lucian. In the diatribe against the uneducated book-collector he de­
scribes his manner of reading his books (Adv. ind. 2): "you read some 
of them moving smartly along (7TCXVV €7TLTPEXWV), your eye keeping 
ahead of your mouth (cpfJaVOVTOS TOU OcpBUAlwu TO UT0f.LU)." Wieland, in 
his note on this passage (which is Balogh's point of departure), drew 
the conclusion that "the ancients, at least the Greeks, used to read all 
books that had any value out loud ... it was a rule that a good book 
must be read out loud." It is hard to see just how this conclusion (in 

8 This is generally held to be the meaning of tacitum at another passage where Balogh 
takes it liter all y: Hor. Sat. 1.3.64ff, ut forte legentem aut tacitum impellat. Balogh admits that 
the phrase could perfectly well mean' er las oder sann schweigsam nach' (and so it is taken 
by its most recent translator, Niall Rudd, The Satires of Horace [Cambridge 1966] 3-'reading 
or in silent thought'), but he prefers of course to take tacitum and legentem as antithetical. 
The support he produces for this (citations from Johannes Cassianus and Petrarch) is liter­
ally far-fetched; closer to hand is the pseudo-Acronian scholium (on 66) which explains: ut 
legentem aut cogitantem. The ultimate provenance of this note cannot of course be deter­
mined, but Keller puts it as far back as his r' which he assigns to the seventh century-and 
this makes it much better evidence than Petrarch. 



BERNARD M. W. KNOX 425 

itself quite sound) is extracted from the words of Lucian. This passage 
is in fact much more difficult to understand than a first reading would 
suggest. Of its three interpreters, Balogh and Hendrickson explicitly, 
and Wieland implicitly, take it for granted that Lucian is finding fault 
with the uneducated man's method of reading. This is indeed what 
one would expect to find in a diatribe which subjects so ITlany aspects 
of its target's life to such blistering invective. Yet in Lucian's presenta­
tion of the indictment the description of his reading sounds more like 
a concession than an accusation. He has just compared the ignorant 
owner of fine volumes to a blind man unable to enjoy the beauty of 
his boy lovers. "You, though (au oE)," he goes on, "do see your books 
with open eyes-you see all too much of them, by Zeus-and you read 
some of them moving smartly along, your eye keeping ahead of your 
mouth." But this cannot be the point of his criticism, for he goes on, 
"But this is not enough to satisfy me (ovo€1rw OE TOVT<) /Lot LKavov):' 
The real indictment follows-ignorance of the merits or faults of what 
is written in the book, failure to understand the meaning of the whole 
and the arrangement of the words, incapacity to judge whether the 
writer has expressed himself accurately in accordance with the rules 
of good diction or has used words that are false, illegitimate or coun­
terfeit. This impressive list of failings cannot be attributed solely to a 
faulty method of reading; they are due to a lack of any education 
whatsoever, and in fact the object of Lucian's satire is portrayed as a 
pretentious ignoramus, who, among other things, cannot even pro­
nounce Greek properly (f3apf3apovs /LEV T~V cpw~v waTTep aV, 4). 

If this analysis is correct, the generalizations of Wieland, Balogh and 
Hendrickson are off to a bad start, but at least they are so far in agree­
ment. But from this point they diverge; they have three different 
explanations of what is wrong with the uneducated man's method of 
reading. Wieland's is only implied, but seems clear enough; an over­
translation is a sure clue to interpretation and he translates TTavv i.TTtTpE­
XWV KTA. with 'aber so schnell dass die Augen usw'. Lucian does not say 
so, but Wieland obviously thought the man was reading too fast; this 
is confirmed by the rest of his remark, which runs: "all poets and 
especially all writers of talent and taste had to be read aloud if half of 
their beauty was not to be lost to the reader." It seems likely too that 
Wieland thought the man was reading silently. Hendrickson certainly 
thought so, for he cites the passage (p.192) as evidence that "reading 
silently" was not only unusual, "it was accounted an imperfect and 



426 SILENT READING IN ANTIQUITY 

defective method of reading." Balogh, who offers no explanation when 
he first quotes the passage, does so much later (p.22S) in his discussion 
of the psychology of ancient reading. Facility in reading, he says (he is 
of course speaking of reading aloud), depended on long practice, and 
"only this practice made it possible for the twofold physical function 
of reading, the work of the eyes and the mouth, to unite in an un­
broken whole, otherwise-to use Lucian's phrase-the eyes would run 
ahead of the mouth." Balogh evidently takes it that our man is read­
ing aloud, but very imperfectly; in the case of a skilful reader "the two 
actions follow the one the other so quickly that the time-difference is 
no longer perceptible" (p.229). Balogh submits no evidence for this 
formulation of the ancient ideal of reading aloud, except an obscure 
phrase-librum ab oculo legit-which comes from the lips of no less an 
authority on matters cultural than Trimalchio (Petron. Sat. 75), the 
same who read about the twelve labors of Hercules in his Homer and 
has a cup with reliefs which show Daedalus shutting Niobe up in the 
wooden horse. 

The passage from Quintilian (Inst. 1.1.32ff) which Balogh cites to 
support his case has the opposite effect. Early training in reading, says 
Quintilian, should for a long time be kept slow, until by practice the 
student attains speed without error (emendata velocitas-a phrase 
which raises doubts about Wieland's interpretation of Lucian). "For 
looking to the right (which is what everybody recommends) and look­
ing ahead is a matter not only of precept but also of practice, since you 
have to look at what follows while you pronounce what precedes ... "7 

But this obviously describes the skill of the fully-trained reader; all 
Quintilian is warning against is teaching it at too early a stage. And his 
words describe exactly what Lucian's uneducated man is doing; pro­
spicere in dextrum ... providere ... ut aliud voce aliud oculis agatur, all 
this is the same process as that defined in 7TCXVV E1TLTPl.xwV, tfo()&'VOVTOS TOU 
'..I..()' ~, , 

O,+, (MJ-LOV TO uToJ-La. 

The conclusion that seems to follow is that Lucian's book-collector 
is reading aloud and reading quite correctly. Be that as it may, one 
thing does emerge clearly from the discussion, and this is the point: 
Lucian's short and apparently simple sentence has given rise to two 

7 Nam prospicere in dextrum. quod omnes praecipiunt. et providere. non rationis modo sed usus 
quoque est. quoniam sequentia intuenti priora dicenda sunt •.. 



BERNARD M. W. KNOX 427 

opposite views of his attitude towards his victim and fours different 
conceptions of just how the man is reading. And this rules it out as a 
reliable basis for generalizations about ancient reading habits-except 
for one, on which all are agreed and which did not need to be proved, 
that ancient reading of literary texts was usually voca1.9 

One more piece of evidence remains to be discussed before the sub­
ject of literary reading is dropped; it was contributed not by Balogh 
but by W. P. Clark, who drew attention to it in 1931 in an article 
which has languished unnoticed ever since.10 It is an extraordinary 
passage. Cicero (Tusc. 5.116) is discussing compensations for the loss of 
hearing, as he had for loss of sight in the preceding section. "If it 
happens that they [i.e. the deaf] take pleasure in songs (cantus),11 they 
must first reflect that before songs were invented, many wise men 
lived happily, and next that much greater pleasure can be experienced 
in reading them than in hearing them (deinde multo maiorem percipi 
posse legendis his quam audiendis voluptatem)." Even though Cicero is still 
a rhetorician even when he is writing philosophy, and even though 
deaf men reading constitute a rather special category, Cicero could not 
possibly have written the concluding phrase if silent reading of poetic 
texts had been impossible or even a nine-day wonder-in fact the 
words imply that he had read them silently himself.12 

None the less, Wieland's statement (even though it may not follow 
from the evidence) is true, and no one would quarrel with his further 
claim that for the ancient world "all poets and especially all writers of 
talent and taste had to be read aloud if the reader was not to lose half 

8 A fifth should perhaps be considered: that Lucian's man is at some indeterminate stage, 
halfway between loud and silent reading. Hendrickson (in another connection) justly re­
marks (p.193): "there are many gradations between vocal and silent reading, descending 
from distinct oral utterance to indistinct murmurs, to whispers, to mere lip motions, and 
so on through unconscious muscular movements of the tongue, throat or larynx, to pure 
eye-reading unattended by any enunciatory effort." The current controversy over the be­
ginnings of literacy in early Greece has had to reckon with the concept of different stages of 
literacy; cf. Sterling Dow, "Minoan Writing," AJA 58 (1954) 109-10. 

9 An important passage which bears on this point seems to have been overlooked: Hor. 
Ep. 1.20--the address to his book. He foresees its eventual fate: it will lose its youth, grow 
dirty as it is thumbed by the hands of the vulgar and either become silent food for unlettered 
bookworms (aut tineas pasces taciturnus inertis), or be shipped off to some provincial town. 
Taciturnus here dearly means ·unread'. 

10 "Ancient Reading," Cj 26 (1931) 698-700. 
11 Clark's impression that cantus "seems to include all harmonious sounds, both prose 

and poetry" cannot be right; the statement is an answer to the objection "at vocem cithar­
oedi non audiunt." 

12 It may be objected that Cicero was thinking of reading aloud as opposed to hearing the 
poems with the music; but this would not be any consolation to the deaf. 



428 SILENT READING IN ANTIQUITY 

their beauty." But Balogh's extension of it, that silent reading of 
literary texts was an 'Abnormitat' and would occur only when the 
reader was Cgezwungen' or cbehindert' can, on his evidence, obtain no 
better verdict than a dour Scottish 'not proven'.13 

But this is by no means the whole of his case. He also claims that 
silent reading of non-literary texts-letters, wills, memoranda, in­
scriptions-was also rare and abnormal; in fact he comes close to 
suggesting that for the overwhelming majority of ancient mankind 
(the only exception he allows is Julius Caesar), silent reading of such 
texts, even of letters in the presence of others, was impossible. Logi­
cally, of course, he needs to make this claim to buttress the other, for 
obviously if people could read letters and wills silently, there is no 
reason why they should not have extended this more efficient, faster 
method to literary texts (especially if they were scholars or volumi­
nous readers). But here too his claim defies common sense and his 
evidence is inadequate. 

The Suetonius passage (Aug. 39) which describes the court of review 
for the equites contributes nothing at all to his argument. "The mildest 
form of reprimand," Suetonius tells us, "was to hand them tablets 
publicly which they were to read silently and at once, on the spot 
(quos taciti et ibidem statim legerent)." The tablets presumably (though 
Suetonius does not say so) contained a record of misdemeanors known 
to the court. Obviously they could have been required to read the 
tablets aloud, but this was lenissimum genus admonitionis-they were to 
be spared public exposure; the court of review contented itself with a 
demonstration that the individual concerned had not escaped its 
watchful eye. The word taciti, far from justifying the conclusion that 
silent reading of such a memorandum was abnormal, specifies the 
lenitas of the reprimand. 

Balogh's next example raises serious doubts about his method. It 
consists of two passages from Horace (Sat. 2.5.51-55 and 66-69). The 
first is the advice of Tiresias to Ulysses about what to do when offered 
a look at his will by the old man you hope to inherit from: refuse, 
remember to push it away from you, but not so that you cannot see, 
with a quick sidelong glance, whether you are sole heir or must share 
the legacy with others. Ulysses is to read the all-important second line 
of the first tablet quickly, furtively and, obviously, silently; the fur-

13 Cf Hendrickson, op.cit. (supra n.4) 193: "silent reading was unusual but in what degree 
exceptional or possible the evidence as yet collected does not permit us to say." 



BERNARD M. W. KNOX 429 

tiveness and the speed are both mentioned-lim is rapias, veloci percurre 
ocuLo-but not the silence. Tiresias then delivers a prophecy so obscure 
that Ulysses has to ask for an explanation; the prophecy, which 
Tiresias now delivers in plain language, refers to some incident well­
known in Horace's day, the details of which are far from clear to us. 
Nasica, evidently expecting to inherit from his son-in-law, will be 
urged to read the will and after many refusals will at last accept it, 
read it silently (tacitus Leget) and find that nothing is left to him and his 
but lamentations. These are two different situations. In the first, the 
legacy-hunter maintains his pretence of indifference but still manages 
to satisfy his curiosity; in the second, Nasica will in the end accept the 
will from his son-in-law and this must mean that he consents to read 
it-there is no question of his pretending not to. We know no more of 
the Nasica-Coranus affair than Horace tells us; interpretation must 
base itself on the text alone. In the dramatic structure of the satire the 
prophecy can have only one conceivable function: it must be a warn­
ing to Ulysses, a salutary example of what will follow from failure to 
observe the advice Tiresias gives him in lines 51ff.l4 If Nasica had 
managed to get an undetected look at the will earlier he would not 
have been so disappointed later. Be that as it may, the two situations 
are clearly different: refusal and undetected reading in the one, 
acceptance and undisguised reading in the other; the mask of indif­
ference maintained in the first and dropped in the second. Balogh 
seems to think that both passages refer to the same actual case,15 and 
uses the first to <explain' the second; <C das <limis oculis [sic] rapere und 
das <veloci oculo percurrere' vereinigen sich schliesslich in der Aktion des 
<tacite legere'." He goes on (p.91) to make confusion worse confounded. 
<First the swindler allows himself to be begged [i.e. to read the will], 
next he takes the tablets in his hands, yet still pretends indifference, 
and only later, at an opportune moment, literally 'seizes' the meaning 

14 Rudd, op.cit. (supra n.6) 304 n.1S, doubts this. "Most editors think that its purpose was 
to instil caution. But if Ulysses was to avoid the fate of Nasica he should have been told 
where Nasica went wrong. This is not at all clear. Certainly in his reluctance to read the 
will (67) Nasica complied with one of Tiresias' own maxims. (51-2)" Where Nasica went 
wrong was his failure, during his repeated refusals to read the will, to catch a glimpse of the 
all-important second line of the first page. The alternative to understanding the second 
passage as a warning is to suspect, with Rudd, that "strictly speaking, there is no relevance, 
and that the tale was included simply because it was too piquant and too topical to be 
omitted." 

15 So at least it would appear from the remarks which follow his citation of 51-55: 
"Offensichtlich spielt Horaz auf einen dem Leser vielleicht bekannten Fall an, denn er 
kommt auf dieses Rezept noch einmal zurlick. Der Erbschaftsjager, Nasica ... " 



430 SILENT READING IN ANTIQUITY 

of the writing (rapit), with a lightening-swift glance for which he has 
so trained himself that he sees only what is essential." 

This is a preposterous conflation of two different situations. The first 
phrase is applicable to both, the second is only fully applicable to 
Nasica, the third is only at home in the first, the words "only later, at 
an opportune moment" have no basis in either text and make sense 
only if understood of Nasica, the next phrase uses the word <rapit' 
which appears only in the first text and the rest of the sentence has no 
warrant in either. Of this contaminated narrative Balogh complacently 
remarks: "this is a not unfamiliar stage-scene; for ancient man its 
strangest, most distinguishing element is precisely the one that does 
not surprise the modem reader at all, namely that the scene is silent." 
On the contrary, there are two different scenes. In the first, where the 
reading of the will must be silent, no attention is drawn to the fact; the 
only proper inference from this is that for Horace and his readers there 
was nothing the least surprising in the silent reading of a tablet so 
quick and furtive as to pass unnoticed. In the second the word tacitus is 
used, but there is no necessity for silence and indeed no reason for it; 
once Nasica accepts the will he cannot pretend indifference any more. 
The only way to make sense of it seems to be, as before, to ignore 
Balogh's insistence on taking it literally and understand it as <by him­
self, in peace'. And I cannot help thinking that Balogh himself had 
some confused inkling of such an interpretation, for it seems the only 
possible explanation of his interpolation of the phrase "nur spater, in 
einem giinstigen Augenblick" into his weird amalgam of the two 
Horatian passages. 

The folk-tale of the apple inscribed with words which, read aloud 
by Cydippe, bound her by oath to marry Acontius is cited by Balogh 
from Aristaenetus (Hercher 140-42). But this text is not as cogent a 
proof as he would like, for, as he is frank to admit, Cydippe, in 
Aristaenetus' version, reads the inscription aloud because her maid­
servant asks her what it says (My€ pm cptATaTT}. Tt T6 7T€ptypafLfUx TOV-TO;). 
What is decisive, however, says Balogh (p.100), is the fact that 
"Acontius has built his whole deception on the natural assumption 
that the girl must unconditionally, according to the law of custom, 
read the oath aloud (miisse nach dem Gesetz der Gewohnheit den 
Schwur unbedingt laut lesen)." If Aristaenetus had been aware of this 
law he would not have bothered to make the handmaid ask Cydippe 
what the inscription said, and Acontius is so far from trusting in the 



BERNARD M. W. KNOX 431 

unconditionality of the law of custom that he rolls his apple not in 
front of Cydippe but at the feet of her servant (AcH)pq. S£€KJA£O'as 7Tp6 

TWV rijr; ()€pa7Ta{vas 7TOSWV). In Ovid, too, (Her. 21) the servant picks up 
the apple and asks Cydippe to read the inscription: 'perlege' dixit (v. 
109). Naturally enough Ovid's Cydippe writes to Acontius that she 
read his letter 'sine murmure'; once bitten, twice shy. Ovid was not the 
man to pass up an occasion for a rhetorical conceit, and in fact later 
in this epistle he pursues it to the limits of absurdity. "If my oath is 
valid," writes Cydippe, "deceive other girls-use letters instead of an 
apple. Deprive the rich of their wealth, make great kings swear they 
will give you their kingdoms ... " (vv.145ff). Balogh does not use this 
passage to prove that the rich and royal of the ancient world were in­
capable of reading letters in silence (though as we shall see this is 
what he believed): perhaps he did not notice it, perhaps even he was 
appalled. These lines are a solemn warning against drawing broad 
historical inferences from a master-rhetorician's exploitation of a folk­
tale motif.16 

Balogh's prize example is a historical incident, which involves a 
letter. Plutarch (Brut. 5) tells the story of the tense Senate meeting at 
the crisis of the Catilinarian conspiracy. Caesar and Cato were engaged 
in debate when a letter was brought in and handed to Caesar; he read 
it in silence (O'£W7Tfj). Cato immediately accused him of receiving com­
munications from the enemy and the Senate burst into an uproar. 
Caesar, who was standing next to Caro, handed him the letter; Cato 
read it and found that it was a love-letter addressed to Caesar by 
Cato's own sister Servilia. 

The conclusions Balogh draws from this story are astonishing. The 
first (pp.92-93) is that "one read even personal letters aloud in assem­
blies; if one did not-as we learn in this case-it caused an enormous 

16 Balogh's interpretation of Ov. Met. 9.568ff (the letter sent by Byblis to Caunus and his 
angry reply) will not stand up against the objections ofE. Norden (printed in n.22), especially 
his reference to pavidum blandita 569, which makes little sense unless the messenger was in­
formed of the contents of the letter. But there is an even more serious objection. Balogh 
(not, as we have seen, for the first time) has constructed his own narrative to bolster his 
thesis. He states that Byblis ordered the messenger to choose the right moment to hand 
Caunus her letter Chat er ja den Befehl usw." p.101) and then explains the reason: Caunus 
would of course unsuspectingly (nichtsahnend) read the letter aloud (p.101), consequently 
the messenger must be sure to choose a moment when Caunus was alone. But of this order 
to the messenger there is not a word in Ovid; all he says is "apta minister tempora nactus adit" 
572-73. Byblis later (vv.61 1-12) wonders whether perhaps her brother's angry reaction was 
the fault of the messenger-Unon adiit apte nee legit idonea, credo, tempora"-but this is no 
firm basis for assuming that she gave him any such instructions. 

6--G.R.B.S. 



432 SILENT READING IN ANTIQUITY 

furore." But Plutarch does not give the slightest hint that the uproar 
was caused by Caesar's silent reading or that this was the basis for 
Cato's accusation. Caesar was already suspected of involvement in the 
conspiracy, and (as we learn from Plutarch elsewhere) he had on this 
very occasion just pleaded for the lives of the arrested conspirators 
against Cato, who had demanded a death sentence. All this is more 
than enough to explain Cato's suspicion and the Senate's reaction. In 
fact, in Plutarch's other account of the incident (Cat. Min. 24)-which 
Balogh does not mention-the detail that Caesar read the letter in 
silence is omitted; evidently it was not necessary for an understanding 
of the story. 

Balogh's second inference from this text is stranger still. He believes 
Cato "might well have read the letter aloud before the Senate (durfte 
... laut gelesen haben)," and in a later passage of his article (pp.101-02) 
he states his belief without qualification: "we have only to think of 
Servilia's love-letter which Cato involuntarily (unwillkurlich) read 
before the whole Senate." In other words, Cato, like Balogh's Cydippe 
(but unlike Caesar) could not read silently at all. 

This is surely too much for anyone to swallow. Even if he could not 
read the letter silently, he could have stopped when he realized what 
it wasI7-Plutarch calls it a 'lascivious letter' (dKbAaa'TOv, Brut. 5). But 
there is no longer any reason why anyone should try to believe it, for 
E. G. Turner has drawn attention to a passage which proves beyond a 
shadow of doubt that in fourth-century Athens silent reading of a 
letter in the presence of others was taken completely for granted. IS 

(And if in fourth-century Athens, why not in Republican Rome?) In 
the Sappho of Antiphanes (Kock 196, Athenaeus 1O.450c) a riddle is 
proposed. "What is it that is female in nature and has children under 
the folds of its garments, and these children, though voiceless, set up 
a ringing shout ... to those mortals they wish to, but others, even 
when present, are not permitted to hear?" A second speaker suggests 
a wrong answer (which has, however, satiric political point and is 
probably the reason the scene was written in the first place), and then 
Sappho gives the correct solution. The answer is €1Tta'TOA1}, a letter; it 
is a feminine noun, and its children are the letters of the alphabet. 
"Though voiceless, they speak to those far away, those they wish to, 

17 As Cydippe does in Aristaenetus: TOV £PWTU(OV MAOV a,7dppu/Jev cxl8ovf'Wr], KCX~ -1Jp.lt/>wvov 
KCXTaA'AO,,"E M,£v ~v £'1T' £C1X&TC~ KE£f"VTJV ••• 

18 Op.cit. (supra n.2) 14 n.4. It was, he says, "recalled by Professor Webster." 



BERNARD M. W. KNOX 433 

but for anyone who happens to be standing near the man who is read­
ing (&:vaYLvcboKOVTOs)19 they are inaudible." This piece of evidence 
could hardly be bettered, for it is the essential characteristic of a riddle 
that the answer to the puzzle it presents must be immediately and 
universally recognized as right-it must be based on common funda­
mental assumptions. 

But there is more evidence still, this time for fifth-century Athens :20 

two passages in which silent reading is not just assumed, it actually 
takes place on stage before the audience in the theater of Dionysus. In 
neither case does the dramatist draw the slightest attention to what is 
happening, and this may account for the fact that these instances 
escaped the notice not only of Balogh but, as far as I can ascertain, of 
everyone else who has concerned himself with this problem. 

The first is Euripides' Hippolytus 856ff. Theseus notices the letter 
which is tied to the hand of his dead wife. After some speculation 
about its contents (all very wide of the mark) he proceeds to open it. 
HCome, let me unwind the wrappings in which it is sealed and see 
what this letter wishes to say to me" (864-65). The chorus now pro­
ceeds to sing five lines of lyric apprehension, followed by three lines of 
apotropaic prayer, and then Theseus bursts out in a cry of grief and 
anger: "Evil upon evil, unbearable, unspeakable." Clearly he has 
read the letter and read it silently-the audience watched him do so. 

The second passage shows silent reading not of a letter but of an 
oracle. In the prologue of Aristophanes' Knights, Nicias comes out of 
the house of Demos at line 115, carrying the oracle which Paphlagon 
guarded most carefully, but which Nicias has managed to steal from 
him as he lay snoring. "Bring it here, let me read it," says Demosthe­
nes (Zv' &:vayvw) and then, like Theseus, Hcome now, let me see what is 
in it." He cries out in astonishment as he reads; indeed, he is so affected 
by the contents of the oracle that he demands more drink. For five 
more lines he continues to express amazement and demand more 
wine while the anxious Nicias presses him with demands for informa­
tion. Finally, at line 127, he begins to explain, and it is clear from what 

19 Here, for once, avay£yvwo"KW clearly means 'read silently'. 
20 Turner (ibidem) offers as evidence for silent letter-reading in fifth-century Athens 

Eur. IT 762, a?J7'~ ~p&.UEL u£ywua 7&YYEypafLfLlva. F. D. Harvey, "Literacy in the Athenian 
Democracy:' REG 79 (1966) 632 n.14, objects that "u£ywua in the EUripides passage ... surely 
refers to the fact that the written letters are silent, not the reader ... " The line could, how­
ever, in view of ar/>wva .•• AaAEi in the Antiphanes passage, be taken as Turner suggests. 
Since the line is ambiguous, it is perhaps better not to cite it as evidence. 



434 SILENT READING IN ANTIQUITY 

he says that he has read right through to the end. At 128 he finally 
begins to tell Nicias what he has read. This is all the more striking 
evidence for quick and skilful silent reading because, while he is read­
ing, Demosthenes is giving orders for more wine, making exclamations 
of amazement and indulging in a rhetorical address to Paphlagon.21 

These three passages, from Antiphanes, Euripides and Aristophanes, 
clearly demonstrate for fifth and fourth century Athens that silent 
reading of letters and oracles (and consequently of any short docu­
ment) was taken completely for granted. But two of them do even 
more. They demonstrate the unreliability of the evidence collected in 
Part II of Balogh's article, which consists of what he calls (p.202) <aku­
stische Belege' in which, "deuten ein Wort, eine Wendung oder 
manchmal nur auch der Sinn auf die als selbstverstandlich voraus­
gesetzte akustische Wirkung eines Textes." These are expressions such 
as the voces paginarum of his title, pagina loqUitur etc., examples of 
which, as he says, are innumerable (though here again his authors are 
of very recent vintage22-they range from Possidius, Optatus of 
Mileve, Augustine, Pontius, Luxorius, Cassiodorus, Salvianus, Phocas, 
Tertullian and Jerome through Bernard of Morlaix and Francis of 
Assisi to Ronsard).23 Balogh takes such expressions literally and con­
structs on this basis a descriptive psychology of ancient reading. The 
possibility that some of them, at least, are purely metaphorical is not 
even discussed.24 It should have been. The Antiphanes riddle speaks 
of letters of the alphabet, which, though voiceless, raise a ringing 
shout (f3o~v taT1]at ')'E"')'WVDV) over the wave of the ocean and all the 
mainland. But, it turns out, they are shouting only to the addressee 
of the letter which contains them, for those standing by him as he 
reads cannot hear a thing. Phaedra's letter to Theseus is just as 
vocal; as Theseus tells us, in terms as <akustisch' as any to be found in 
the Fathers of the Church, mediaeval saints or French Renaissance 

II This passage effectively cancels out Balogh's one piece of evidence from fifth-century 
Greece (one of his 'akustische Belege', p.207): Hdt. 1.48 (Croesus and the oracles), lKacrra 
ava1T'Tvaawv J'/TcfJpa 'l"WV avyypap.p.a'l"wv---6 IlE cfJ'i '1"6 JK tJ£>.rpwv TjKova£ ••• 

22 Hendrickson, who discusses the same phenomenon, has much more respectable 
sources: Plato, Polybius, Longinus, Plutarch, Horace, Varro, Persius, Quintilian et al. 

23 He even includes Ambrose (sonus litterarum) without realizing apparently that this 
example, instead of strengthening his case, raises doubt about his method. 

24 Hendrickson, on the other hand, concludes his discussion of aKOVfW and audire with a 
caution (p.191) against "pressing examples where metaphor or figurative vividness of 
speech may rather explain the usage." He questions the validity of Balogh's use of Augus­
tine's paginarum vocibus and aptly quotes Keats' "when I heard Chapman speak out loud 
and bold." 



BERNARD M. W. KNOX 435 

poets, it sings (pIAos), speaks (CP(}€yyofL€vov) and shouts aloud (f3oif. f3oif.). 
of course it does nothing of the sort; the audience has watched 
Theseus read the letter, and he did so in silence. 

"Can it be proved," Balogh asks at the end of his introduction (p.87), 
"that the ancient world always read out loud everything which we to­
day read silently to ourselves?" He answers his own question a few 
lines later by asking another: "What is the explanation of these strange 
phenomena ... ?" But he was counting chickens which did not hatch. 
The evidence so far assembled (once properly understood) answers 
his first question with a resounding "No!" Ancient books were nor­
mally read aloud, but there is nothing to show that silent reading of 
books was anything extraordinary except the famous passage from 
Augustine'S Confessions, and that is countered by the phrase of Cicero 
which makes sense only if understood as a reference to silent reading 
of lyric poets. As for letters and similar documents, Balogh's evidence 
is inadequate to start with and his case is blown sky-high by evidence 
he did not notice. 

CENTER FOR HELLENIC STUDIES 

July, 1968