Document name: Alexandria: Library of Dreams

Document date: 2002

Copyright information:
OpenLearn Study Unit:
OpenLearn url:
Activities 2 and 4: extracts from: Roger Bagnall ‘Alexandria:
Library of Dreams (2002) in Proceedings of the American
Philosophical Society, published by American Philosophical
Society. Courtesy of Dr Bagnall and American Philosophical
Society.
The Library of Alexandria
http://www.open.edu/openlearn/history-the-arts/library-
alexandria/content-section-0

Alexandria: Library of Dreams

Roger S. Bagnall

Source: Proceedings of the American Philosophical Society, Vol. 146, No. 4 (Dec. 2002), pp. 348–362

Published by: American Philosophical Society

www.open.edu/openlearn Page 1 of 1
Alexandria: Library of Dreams
Author(s): Roger S. Bagnall
Source: Proceedings of the American Philosophical Society, Vol. 146, No. 4 (Dec., 2002), pp.
348-362
Published by: American Philosophical Society
Stable URL: http://www.jstor.org/stable/1558311 .
Accessed: 09/09/2014 11:18

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
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@jstor.org.

American Philosophical Society is collaborating with JSTOR to digitize, preserve and extend access to
Proceedings of the American Philosophical Society.

http://www.jstor.org

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
 Alexandria:
Library
ofDreams1
ROGER S. BAGNALL
Professorof Classics and History
Columbia University

Y TITLE does not intendto suggestthat the Alexandrian

Librarydid not exist,but it does pointto what I regardas
1JvJ. z the unreal characterof much that has been said about it.
The disparitybetween,on the one hand, the grandeurand importance
of this library,both in its realityin antiquityand in its image both
ancientand modern,and, on theother,our nearlytotalignoranceabout
it, has been unbearable.No one, least of all modernscholars,has been
able to acceptour lack ofknowledgeabout a phenomenonthatembodies
so manyhumanaspirations.In consequence,a whole literature of wish-
ful thinkinghas grown up, in which scholars-even, I fear,the most
rigorous-have cast aside the time-tested methodsthatnormallycon-
straincredulity, in order to be able to avoid confessingdefeat.After
sketchingbrieflythe main lines of our ignoranceof the Library'shis-
tory,I shall talk about threetypesof dreamsthathave beguiledcom-
mentatorsancientand modern:dreamsabout thesize of theBibliotheca
Alexandrina;dreamsabout placingthe blame forits destruction;and
dreamsabout theconsequencesof its loss.2But thereare some positive
lessonsas well, as I hope to show.
There is no ancientaccount of the foundationof the Library.3We

I Read 10 November2000.
2The bibliography on the BibliothecaAlexandrinais enormous;I referto it very
inwhatfollows.The following
selectively worksarecitedbelowbyauthor'sname:Mostafa
El-Abbadi,The Lifeand Fate of theAncientLibraryof Alexandria(Paris,1990); Rudolf
Blum,Kallimachos,the AlexandrianLibraryand the Originsof Bibliography, tr.H. H.
Wellisch(Madison, 1991); Lionel Casson, Librariesin the AncientWorld(New Haven,
2001); Diana Delia, "FromRomanceto Rhetoric: The Alexandrian Libraryin Classicaland
IslamicTraditions,"AHR 97 (1992): 1449-67; P. M. Fraser,PtolemaicAlexandria,3 vols.
(Oxford,1972); K. S. Staikos,The GreatLibraries:FromAntiquityto the Renaissance
(Londonand New Castle,Del., 2000). More extensivereferences to theancientsourcesthan
are possibleheremaybe foundparticularly in El-Abbadi,Delia, and Fraser.
3 Blum,100, suggests thatCallixeinosmayhavegivensuchan account,and thatsomeof
theinformation in laterwriters
mayderivefromhim.I can see no evidenceforthisview.

PROCEEDINGS OF THE AMERICAN PHILOSOPHICAL SOCIETY VOL. 146, NO. 4, DECEMBER 2002

[348]

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
ALEXANDRIA: LIBRARY OF DREAMS 349

have onlybriefand glancingreferences. The nearestthingto evena

briefhistoryappearsin theprefaceto a commentary on Aristophanes
writtenbytheByzantine polymath JohnTzetzesin thetwelfth century.
KindercriticssaythatTzetzes"preserves muchvaluable,thoughto be
surenot alwayscorrectly reported,information on ancientliterature
and culturalhistory."4
The lesscharitablecall him"copious,careless,
quarrelsome" and "extremely inaccurate.... His uncorroboratedevi-
denceis accordinglyviewedwithmuchsuspicion"5 or "quiteunjustifi-
ablyconceitedabouthis own attainments."6 Tzetzes,liketheancient
tradition treatedPtolemyII Philadelphos
generally,7 as thekingwho
createdtheLibrary. He describeshowthreemen,Alexandros ofAeto-
lia, Lykophron of Chalkis,and Zenodotosof Ephesos,workedwith
Ptolemy to acquirebooks.8
One mightthenthinkthatthe foundation by Philadelphoswas
secure.But no. Tzetzes,like othersources,also mentionsthatPtolemy
collectedthe books "through"Demetriosof Phaleron.Now thisDem-
etrios,a pupilofTheophrastosand earlierofAristotle, had ruledAthens
fortheMacedonian kingCassanderfora decade (317-307); afterCas-
sander'sdeath,he fledto Egypt,joiningthe court of PtolemyI Soter,
the fatherof Philadelphos,wherehe certainlycontributedmuchto the
royalprojectof makingAlexandriaa worthyrivalto Athens.He made,
however,the strategicmiscalculationof supportingas Soter'ssuccessor
theolderhalf-brother of Philadelphos,and whenthelattercame to the
throneinstead,the sexagenarianDemetriospaid forhis mistakewith
internalexile,dyingsoon thereafter.9
He is, in short,not a good candi-
date forcollaboratorwithPtolemyII.
Demetriosis alreadypresent,however,in theearliestsurvivingtext
to talk about the Library,namelythe curious Letterto Philocrates,a
workof thesecondcenturyB.C. thatclaimsto be theworkof a courtier
of PtolemyII named Aristeas.10 As faras we know,therewas no such
personas thisAristeas.11
Althoughsomecompetent modernscholarshave
been at pains to praise Pseudo-Aristeas's
knowledgeof the Ptolemaic

4W. 0. Schmitt, KleinePauly5 (Munich,1975), 1033.

5Oxford Classical Dictionary,2d ed. (Oxford,1970), 1102 (P.B.R. Forbes,Robert
Browning).
6 L. D. Reynoldsand N. G. Wilson,Scribesand Scholars,2d ed. (Oxford,1974), 62.
7 See Fraser1:321.
8Prolegomena de comoediaAristophanis 2.
9The major source is Diogenes Laertius 5.75-85 (F. Jacoby,Die Fragmenteder
Griechischen Historiker IIB [Leiden,1962], 642-43, no. 228 Ti).
10Andr6 Pelletier,
s.j., La lettred'Aristeea Philocrate(Paris,1962).
11ProsopographiaPtolemaica6 (Leuven,1968), no. 14588, considershim probably
fictitious.

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
350 ROGER S. BAGNALL

milieu,12to theextentthathe reflectsany realityit is thatof thesecond

century, not the third,and the work is fullof incrediblethings.13 The
court detail is, indeed, "merelycorroborativedetail,intendedto give
artisticverisimilitude to an otherwisebald and unconvincingnarra-
tive," as Pooh-Bah would put it (Mikado, act 2). It was Demetrios,
accordingto Pseudo-Aristeas,who persuaded PtolemyII to commis-
sion thetranslationof theJewishscriptures thatwe call the Septuagint
in order to help complete the royal library'sholdings; this story,
indeed,is thecenterpieceof thispiece ofJewishpropaganda.
Now mostphilologists,facedwithtextsfullof misinformation and
a flatcontradictionsuch as the juxtapositionbetweenDemetriosand
Philadelphosoffers,would normallybe extremelyskeptical,or dismiss
Demetrios' role as fiction.14Not here, however.Everything reported
mustbe keptin some fashion.So, almostunanimously, thereactionhas
been to suppose that PtolemyI was the real founderof the Library,
assistedby Demetrios,while Zenodotos was eithera subordinate" or
came to theforeafterPhiladelphoscame to thethrone.16 The onlyreal
basis for such a view, otherthan a desperatedesire not to abandon
the sources,is a statementof StrabothatAristotletaughtPtolemythe

12 Fraser1:696-704 gives a detaileddiscussion,dwelling(699-700) on the author's

knowledgeofthePtolemaiccourt(hispictureis clearlythatofthesecondcentury, afterthe
introduction of courtranks).Fraserdatestheworkto thereignof PtolemyVI Philometor
(180-145 B.C.). Otherviewsdiffer, buta mid-second-century date is plausible.For recent
generaldiscussionsof "Aristeas,"seeJ.M.G.Barclay, Jewsin theMediterranean Diaspora:
FromAlexanderto Trajan(323 BCE-i 17 CE) (Berkeley, 1999 [Edinburgh, 1996]), 138-50
and E. S. Gruen,Heritageand Hellenism:The Reinvention ofJewishTradition(Berkeley,
1998), 207-22, withthediscussioncomparing thembyD. R. Schwartz,ClassicalPhilology
95 (2000): 352-54.
13 For example,Ps.-Aristeas believesthatthereare stilltwelvetribesin Judaea,and he
claims that Ptolemyliberateda hundredthousandslaves in Ptolemaicpossessionby
purchasing themfromtheirowners.How Fraser(1:700)canthinkthisis a "genuine document"
is mystifying, althoughhe is not alone. The textcitedas a parallel,C. Ord. Ptol. 22, is,
despitesomeverbalsimilarities (acceptedevenbytheusuallyskepticalGruen[above,n. 12],
211), radicallydifferent. Captivestakenby PtolemyI to Egyptwere,if slavesat all, in
preciselythe class (slaves sold by the crown)thatthe ordinanceof 260 B.C. leftin the
undisturbed possessionoftheirowners.
14E. A. Parsons,The AlexandrianLibrary:Gloryof the HellenicWorld(Amsterdam,
1952), 83-105, in discussingthe foundationand buildingof the Library,recognizesthe
weaknessof Ps.-Aristeas's evidenceand thedifficultieswithTzetzes(whomhe discussesin
greatdetail),butrefusesto giveup theinformation theyprovide.Staikos,60-61, also notes
theinsecurity oftheevidenceforDemetrios, butby71, n. 22, he has succumbed to thinking
that"theeventsdescribedbyAristeascannotbe fictitious." Gruen(above,n. 12), 209, is
moreconsistently criticalin regardingDemetrios'involvement as fiction.
15E.g.,Blum,102.
16 So Delia, 1460. The latestversionof thisis Casson's formulation (34): "It was the
brainchildof PtolemyI, eventhoughit maynot have come intobeinguntilthereignof
his son."

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
ALEXANDRIA: LIBRARY OF DREAMS 351

formation ofa library.17

Thisremark, whichcanhardlybe literally true
(Aristotle diedin 322), is takento meanthattheidea ofsucha library,
broadand scientific in character, was Peripatetic and cameto Ptolemy
through Demetrios. Thatis notunreasonable, butithardlyshowsthat
Ptolemy I tookanyspecific action.And,tobesure,Alexander, Lycophron,
andZenodotos,thetriomentioned byTzetzes,wereactiveduringPhil-
adelphos'reign.So muchforourlackofpreciseinformation aboutthe
foundation and earlygrowthoftheLibrary. It mustbe addedthatwe
are hardlyin bettershapeconcerning thefamousMouseion,therela-
tionship ofwhichto theLibrary is also a matterofspeculation.18
It is to Pseudo-Aristeas also thatwe owe theearliestsurviving fig-
uresforthesizeoftheLibrary. He has Demetrios tellPtolemy thatthe
Library nowhasmorethan200,000books,buthehopesto bringitup
to 500,000 beforelong.19 Tzetzestellsus thatthePalaceLibrarycon-
tained400,000 "mixed"(symmigeis) books and 90,000 "unmixed"
(amigeis).He also reportsthattherewas an "externallibrary"with
42,800 books.Although therehas beenmuchcontroversy, it is likely
that"mixed"refers torollscontaining morethanonework(andperhaps
morethanone author),"unmixed"to worksoccupyingbook-rolls
(oftenmultiple rolls)bythemselves.20 Laterwriters giveotherfigures:
AulusGellius(Noct.Att.7.17.3)says700,000rolls(butsome"inferior"
manuscripts give70,000).Seneca(De tranq.animi9.5) reproaches Livy
forshowingregret at thedestruction of40,000 volumes(an excessive
luxury,in Seneca'sview) in the Alexandrine War; modernscholars,
witha bentforgigantism, have suspectedthisof beingan errorfor
400,000,21on thebasisof a figure in thelatehistorian Orosius(Hist.

17Strabo608c. See Fraser1:320,on theproblemsofthispassage.As carefula philologist

as HughLloyd-Jones, in a reviewcastigatingL. Canfora(below,n. 49) foruncriticaluse of
evidence,takestheroleof Demetriosas a given(Greekin a Cold Climate[London,1991],
115-22; fromNew York Review of Books, 14 June 1990). Similarly, RobertBarnes,
"CloisteredBookwormsin the Chicken-Coopof the Muses: The AncientLibraryof
Alexandria,"in Roy MacLeod, ed., The Libraryof Alexandria:Centreof Learningin the
AncientWorld(London,2000), 61-77, withoutengagement oftheliterature on thesubject.
It is truethatIrenaeusand ClementofAlexandria(moreconfusedly) indicatePtolemySoter
as thefounder(see El-Abbadi,79-80), butit is notclearthatthisrestson anyindependent
tradition;theymay simplyhave recognizedthe problemof connecting Philadelphosand
Demetrios.
18See Fraser1:312-19 and El-Abbadi,84-90, foraccountsoftheMouseion.As withthe
Library, ouraccountsofitcomemainlyfromtheRomanperiod.Lloyd-Jones (above,n. 17)
correctly remindsus ofhow muchwe do notknow.
19Ps.-Aristeas'saccount,withfigures, is repeatedinJosephus,JewishAntiquities 12.13.
20 See Fraser1:329 on thispoint.WhyFraser(328) saysthatTzetzesis
"ouronlyevidence
as to thenumber"I do notknow.See Fraser2:474 n. 108,demolishing theviewthatTzetzes'
figures comefromCallimachus.
21 See,e.g.,Delia, 1458 n. 38.

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
352 ROGER S. BAGNALL

adv.pagan.6.15.31-32), whereonce again some manuscripts

give
40,000 insteadof the majority400,000. AmmianusMarcellinus,writ-
ing of the Serapeum,tells us that it containeda libraryand that "the
unanimous testimonyof ancient records declares that 700,000 vol-
umes,broughttogether bytheunremitting energyofthePtolemaickings,
were burned in the Alexandrinewar" (22.16.13). He has of course
been reproachedby modernsfor confusingthe Palace and Serapeum
libraries-moreon thislater.22 It is reasonablyobvious thattheancient
sourcesthoughtthe librarieswere enormousbut had no good figures
to work with.23In any case, figuresin ancienttextswere easily cor-
ruptedin transmission and oftensurvivein multiplereadings.
We have alreadyseen thatPseudo-Aristeashas thatleast attractive
quality in a source: to be trustedonly where corroboratedby better
evidence,and thereunneeded.The qualityof therestof thelatertradi-
tion about the size of the Libraryis not muchbetter.But let us turnto
askingabout the inherentplausibilityof the numbers.The basic ques-
tionswe shouldask are,how manybooks probablyexistedin theearly
third century,how likely it is that large-scalecollectingcontinued
underthe laterPtolemiesand the Romans, and whetherthesefigures
are at all in linewithwhat we know of otherancientlibraries.24
The computerdatabank of ancientGreekliterature, the Thesaurus
Linguae Graecae, containsabout 450 authorsof whom at least a few
words survivein quotationand whose livesare thoughtto have begun
by the late fourthcentury.No doubt therewere authorsextantin the
earlyHellenisticperiodof whom not a line survivestoday,butwe can-
not estimatetheirnumbers.Of most of these450, we have literallya
fewsentences.Thereare another175 knownwhose livesare placed, or

22SeeJ. C. Rolfe'snotein theLoeb Ammianus(2:302 n. 1), confidently and precisely

informing us that"at thetimeofthebattleofPharsaliathetotalnumberwas 532,800 [i.e.,
490,000 in themainlibraryand 42,800 in theSerapeum]and itmayhavereached700,000
bythetimeoftheAlexandrine war."
23 See Delia, 1458-59. Onlybycollapsingantiquity intoa singlechronological
horizon
could one saythat"contemporary accountssuggestthattheyamassedas manyas 500,000
texts"(The Economist,8 April2000, p. 92). Staikos,70, claimsthat"thereis no doubtthat
theLibrarydid have a stockof severalhundredthousandrolls,and whenall thereliable
contemporary evidenceis evaluateditis reasonableto suggestthatthehighestfigure
ofall-
700,000 rolls-does not sound excessiveand may even be an exaggeration." What "the
reliablecontemporary evidence"consistsofis hardto see.
24 Barnes(above,n. 17), 65, oddlycitesthelibrary at Pergamon, forwhichPlutarchgives
the(probablyunreliable) figureof200,000 volumes,as evidenceinfavorofthehighnumbers
of volumesat Alexandria.On the otherhand,he also says "it has been suggested"that
Alexandriahad only70,000 different titlesin thethirdcentury(he does not footnotethis
statement, butLloyd-Jones [above,n. 17], 117, citesE. G. Turner,GreekPapyri[Oxford,
1968] forthisassertion, withoutpage number;it does notappearin Turner'sdiscussionof
theAlexandrian Libraryon pp. 102-03).

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
ALEXANDRIA: LIBRARY OF DREAMS 353

whosebirths areplaced,inthethirdcentury B.C. Mostoftheseauthors

probablywrotewhatbymodernstandards was a modestamount-a
fewbook-rolls full,perhaps.Eventhemostvoluminous authorsofthe
group,liketheAtheniandramatists, probablyfilledno morethana
hundredrollsor so. If theaveragewriterfilled50 rolls,our known
authorsto theend of thethirdcentury wouldhaveproduced31,250
rolls.Wemustthenassume,to savetheancientfigures forthecontents
oftheLibrary, eitherthatmorethan90 percent ofclassicalauthorsare
notevenquotedorcitedinwhatsurvives, orthatthePtolemies acquired
a dozencopiesof everything, or somecombination of theseunlikely
hypotheses. Ifwe were(moreplausibly)to use a loweraverageoutput
figure per author,thehypotheses neededto save thenumbers would
becomeproportionately moreoutlandish.25
To lookat matters anotherway,just2,871,000wordsofGreekare
preserved forall authorsknownto have livedat leastin partin the
fourth century Addingthethirdand secondcenturies
or earlier. brings
thetotalto 3,773,000words(or about 12,600 pages of 300 words
each).26At an averageof 15,000 wordsper roll,thiscorpuswould
requirea mere251 rolls.Evenat an averageof 10,000wordsperroll,
thefigure would be only377 rolls.It was estimated by one eminent
ancienthistorian thattheoriginalbulkofhistoricalwritings in ancient
Greeceamountedto something likefortytimeswhathas survived.27 If
so, our estimate wouldrunto an originalbodyof 10,000to 15,000
rolls.Thismaybe too low,butis itlikelythatitis too low bya factor
of thirtyor forty,and thatonlyone wordin 1,500 or 2,000 has sur-
vived?Again,we would be requiredto believethatwe do not even
have the namesof thevastmajority of ancientauthors,or thatthe
Library possessedthirty orfortycopiesnotonlyofHomerbutofevery
singleauthor.
Wecannotsavethefigures bysupposing thatgrowth afterthethird
century, or evenafterthesecondcentury, accountsforthedifference.
For one thing,noneofourevidenceforbook acquisition is laterthan
thethirdcentury, and mostofit concernsPtolemy II and Ptolemy III,
thelatterbeingthesubjectofthefamous,butprobablyunbelievable,
anecdotes inGalenaboutseizingbooksfrompassingshipsandhijacking

25H. Strasburger, Studienzur AltenGeschichte3 (New York,1990), 178-79, lists32

historicalwritersforwhomwe know exact or approximate numbersof books originally
producedbutnowlost.Theaverageis 28.2; itwouldfallto24 ifwe excludedAristotle
(thecity
constitutions),an altogether Andhistorians
exceptionalfigure. wererelatively
long-winded.
26 Thesefigures arecomputedfromthefilesoftheinvaluableThesaurusLinguaeGraecae
(Irvine);I am gratefulto Maria Panteliaforsupplying
them.
27Strasburger (above,n. 25), 180-81. He worksmainlywithTeubnerpagesratherthan
rollsin hiscomputations, buttheresultscometo muchthesamething.

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
354 ROGER S. BAGNALL

theoriginalsofthetragediansfromAthens.28 It is mostunlikely, at all

policywas pursuedin thewake of
events,thatan activeacquisitions
theexpulsionofmostoftheMouseion'sintellectuals in 145 B.C. More-
over,ifwe areto giveanycredenceto thesenumbers, whyshouldwe
notbe consistentin ourcredulity
and believethatDemetrios ofPhale-
ron alreadyhad amassed200,000 volumesin thefirstdecadeof the
third B.C., as Pseudo-Aristeas
century says?
An amusingsidelight to suchreflections
is providedbya blockof
granite,inthetopofwhichis a hollowed-out spacemeasuring 19.5 by
23 centimeters and 8 centimeters deep. Foundin 1847 and now in
Vienna,ithas thelegend"Dioskourides, 3 rolls"inscribed on itsface.
It hasgenerallybeenseenas a storagecontainer forthreepapyrus rolls;
becauseit was foundnearwheretheLibraryis thought to havebeen
located,it was quicklyidentifiedas partoftheLibrary's equipment.29
Although othersrejectedthisidentification,
almosteveryone has agreed
thatit was indeeda book-storage device.A libraryof a half-million
rollswouldhaverequired166,667ofthesecontainers. It is noteasyto
imaginea structure and shelvingsystem in whichsuchgranite contain-
erswouldhave stood;thereis no lid,either. No wonderone scholar
hastilyassuresus, althoughwithoutany evidence,that "onlyrare
manuscripts would have requiredsuch custom-made stonebins for
theirpreservation."30Actually,thereis no reasonto thinkthatit held
papyrusrollsat all. Its traditionaldepictionin drawings(Fig. 1) no
doubthelpedencouragesuchideas,buta soberlook at therealthing
(Fig.2) showsthatonlya smallfraction oftheblockconsisted ofthis
hollowspace.31It was in factsurelya basefora statueor bust.
In sum,theancientfigures forthesizeoftheLibrary orthenumber
ofvolumeslostin theAlexandrine Wardo notdeserveanycredence.32

28Theseare quotedin everytreatment ofthesubject;cf.,e.g.,El-Abbadi,73-102; more

briefly,Barnes(above,n. 17), 65-66. Hardlyanyonehas eversuggested thattheymightnot
be factual.
29 The actuallocationin thepalacequarter is unknown, butithas beenarguedthatitwas
intheareanearthemodernNabi DanielStreetandnorthofHorreyaAvenue.Forthehistory
of the question,see MieczyslawRodziewicz,"A Reviewof theArchaeologicalEvidence
Concerningthe CulturalInstitutions in AncientAlexandria,"Graeco-Arabica6 (1995):
317-32.
30Delia,1455.
31 Fora comprehensive bibliography,
seenowE. Bernand, Inscriptions
grecques d'Alexandrie
ptoMmaique(Cairo,2001), 167-69, no. 65, butevenhe merelyreprints a drawingof the
nineteenth century.I am grateful to Dr. AlfredBernhard-Walcher of theKunsthistorisches
Museum,Antikensammlung, Vienna,forthe photographprintedhereand access to the
originalin April2002. I discussthisobjectin detailin an articleforthcoming in theBulletin
de la Sociteoarcheologiqued'Alexandrie.
32Blum,107, is one ofthefewscholarsto doubttheancientfigures.

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
ALEXANDRIA: LIBRARY OF DREAMS 355

17?"~1

r TOM0I1/
LAlOYLKOYPIAHY
He 1 71
FIGURE ofgranite
1. Drawing block(from
Delia)

Theydo notappearto reston anygood ancientauthority,theywere

repeatedfromauthorto author,and when theirconsequencesare
examined,theylead to impossibilities
and absurdities.The actual
numberswereprobablylower,perhapsby as muchas one orderof

-al

2. Photograph
FIGURE of granite
block(photograph Vienna,Antik-
courtesy
Kunsthistorisches
ensammlung, Museum)

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
356 ROGER S. BAGNALL

magnitude.33 The Libraryof Alexandria,howevercomprehensive for

itstime,was noton a scalecomparable withthegreatresearchlibraries
ofthetwentieth century.
Indeed,howcouldithavebeen?One has onlyto imaginethediffi-
cultiesinvolvedin catalogingsucha collection.Book-form catalogs,
evenwithall theadvantagesof thelargecodex,ceasedto be useful
whenmodernlibraries startedto reachthekindsof middlesix-figure
sizesimagined forAlexandria,andhadto be replacedbythecardcata-
log, unknownin antiquity. My own university'slibrarygrewfrom
20,000 volumesin 1856 to 100,000 in 1889 and 362,000 in 1903.34
Eventhegiantsdid notreachthemiddlesix digitsuntilthemiddleof
thenineteenthcentury, thepointat whichthecardcatalog
precisely
startedto comeintouse.The British
Museumhad onlysome200,000
volumesin 1830,reachinga milliona thirdof a centurylater.35
Calli-
machus'sfamousPinakes,a systematiclistingof genres,authors,and
worksin 120 books, could not have held the information necessaryto
catalog hundredsof thousandsof rolls.36
Nothingin theLibrary'shistoryhas quite inflamedtheimagination
so muchas its destruction.But how was it destroyed?This is a murder
mystery witha numberof suspects,each at leastwithopportunity and
means. The most popular candidate has been Julius Caesar, whose
operationsin 48 B.C. in the harborof Alexandriaare oftenblamedfor
settingfireto the librarynear the shore.The turbulentpoliticalhistory
of the thirdcenturyof our era also offerssome possibilities,including
the emperorsCaracalla, Aurelian,and Diocletian,all of whom did sig-
nificantdamage in Alexandria.The anti-Christian partyinsiststhat it

33Cf. AndrewJ. Carriker, The Libraryof Eusebiusof Caesarea (Ph.D. diss.,Columbia

Univ.,1999), 32, remarking that"a librarythatwas a tenthofthissize [sc. the500,000 in
wouldstillhavebeenverylargein antiquity,"
Ps.-Aristeas] and collectingfigures forancient
libraries.
34JamesH. Canfield,in A Historyof Columbia University, 1754-1904 (New York,
1904), 437-41.
35See AllenKentand Harold Lancour,eds., Encyclopediaof Libraryand Information
Science4 (New York,1970),295, fortheBritish Museum'sgrowth;4:277, on theriseofthe
cardcatalog,whichwas dominantin theU.S. by 1893.
36The Pinakeswere not themselves the library'scatalog (see Fraser1:453), but were
certainlybasedon it.Ifitsbookswerestandardrollsof20 sheets,and iftheyusedrelatively
narrowcolumns(yielding 27 columnsto a roll)and werewritten in smallletters(44 linesto
a column),theywillstillhavecontainedno morethan142,560 lines.As muchofthework
was biographical,onlypartofthattotalis availableforlistingworks.Of coursesomeworks
had multiplebooks,butCallimachusseemsto havelistedsomeworks(likePindar)poemby
poem and arguedpointsabout them;a numberof entriesper book-rollwillhaveresulted,
balancingthemulti-roll works.Iftwo-thirds ofthespacewas usedfortitlesand on average
each line representeda title(bothassumptionsprobablytoo favorableto the numberof
books),thetotalwouldstillnothavereached100,000rolls.

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
ALEXANDRIA: LIBRARY OF DREAMS 357

wasthemobofmonksresponsible forthedestruction oftheSerapeum in

391, who wipedoutclassicallearning. The pro-Christian,anti-Muslim
sentiment can believethestoriesthatblameinstructions givenbythe
caliphto Amr,theArabconquerorofEgypt,to feedthebooksto the
firesin 642, buttheseoriginate centuriesafterthefactand are surely
fiction.37
Passionsstillrunhighon thismatter. WhenGlenBowersockfirst
invitedme to presentthispaper,I hesitatedbecauseof a traumatic
earlyexperience. I wrotean articleon the Alexandrian Libraryon
for
commission a short-lived magazinecalledThe Dial, publishedfor
Channel13. The editordidnotlikemycautionabouttheaccountsof
thedestruction oftheLibrary and,without tellingme,rewrote thearti-
cleto blameeverything squarelyon theChristians.38 Whether he hated
Christianity or justlikeda simplestoryline,I do notknow.
Thematter is,truth to tell,notso clear.39
Thesubjecthas beenend-
lesslydebatedbymodernscholars, butwithlittleresult.Therewas cer-
tainlystillsomesubstantial libraryinRomanAlexandria. Thisis evident
fromSuetonius's lifeofDomitian(20), wherewe learnthathereplaced
bookslostto firein Romanlibraries inpartbysendingscribesto Alex-
andriato copymanuscripts there.40 And someof thescholarly work
thatwenton in theRomanperiodin Alexandria is difficult
to imagine
withouta substantial library.As theMuseumwas certainly stillopera-
in
tive the Roman period,beliefin a Caesariandestruction of the
Libraryrequiresthe uneconomical assumption thatthe Librarywas
destroyed in the firebut theMuseumwas not. Recentlytherehave
been signsof a consensusin formation thatthemostlikelydate of

37 See El-Abbadi, 167-72. A. J.Butler,

TheArabConquestofEgypt,2d ed. byP.M. Fraser
(Oxford1902, 19782),401-26, alreadypronounced thestorya fable,althoughnotall ofhis
arguments are persuasive.See also Fraser'saddendato Butler,pp. lxxv-lxxvi,and Delia,
1465-67.
38 "LessonsoftheAlexandrian Library,"The Dial 1.2 (Oct. 1980): 96-100.
39Matters are madeworsebythefailureof our bestsource,Strabo,to speakclearlyon
thematter. In his Geography17.1.8,he saysthat"theMouseionis also partofthepalaces,
possessinga peripatosand exedra and large oikos, in whichthe commontable of the
philologoi,menwho are members oftheMouselon,is located.Thissynodoshas property in
commonand a priestin chargeoftheMouseion,formerly appointedbythekings,butnow
byCaesar."(I havekepttechnical termsintransliteration.)WhydoesStrabonotmention the
Library?His odd allusiveness in 2.1.5 has also arousedsuspicion:"For Eratosthenes takes
all thesemattersas actuallyestablishedbythetestimony of themenwho had beenon the
spot,havingencountered manyhypomnemata, withwhichhe was wellfurnished, havinga
library suchas Hipparchoshimself saysitwas." Was itno longersuchin Strabo'stime?And
yet,itlooksas ifthepalacequarterhad beenunscathedbythefires,tojudgefromtheoverall
tenorofStrabo'sreport.
40On what basis Staikos(83) thinksthisepisodemightindicatetheexistenceof Latin
worksin theLibrary, I cannotsee.

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
358 ROGER S. BAGNALL

majordestruction forthePalaceLibrary is 273, withAurelian's recap-

tureofAlexandria fromthePalmyrenes theoccasion.41 Butthereis no
directevidenceconcerning theLibraryin theancientsourcesforthis;
theargument, rather,is thatthepalaceareawas devastated at thistime
andthegreatLibrary was probablya victimofthislargerdestruction.
The argument is evenmorecomplicated, however,becauseit is
generally thought thatthereweremultiple inAlexandria.
libraries John
Tzetzes,youwillrecall,speaksofan outerlibrary. He doesnottellus
whereit was. Modernscholarship has uniformly filledin thegap with
thestatement of a Christian writer,
Epiphanius, thebishopof Salamis
bestknownforhiscompendium on heresies,to theeffect that"later
anotherlibrarywas builtin theSerapeum,. . . whichwas calledthe
daughter ofthefirst one."42OtherChristian writers echothisinforma-
tion.Archaeological workat theSerapeumhas shownthattherewere
spacesthatcouldhavehousedbooks,butthatis as muchas excavation
has revealed.43 The age ofthislibrary is unknown, althoughit is usu-
allythought, on notmuchevidence, to go backto thetimeofPtolemy
III's constructionat theSerapeum.NeitherCaesar'sfirenorAurelian's
destruction wouldnecessarily haveaffected theSerapeum; thusa library
couldhave survivedin Alexandriauntilthedestruction of the Sera-
peumitself.
Whatis lesscommonly recognized44is theexistence ofwhata film
about brittlebookssomeyearsago called "slow fires."Papyrusis a
goodmaterial, acidfreeand highly durable.It can lastforhundreds of
yearsundergood conditions. ButAlexandriahardlyrepresented ideal
conditions. It has a Mediterranean climate,not a Saharanone,with
humidity enoughto be detrimental to books.No papyrihavesurvived
therefromantiquity to thepresentday,unlikein drierdesertareasin
Egypt.Booksdeteriorate also withuse,and who is to say thatthere
were no mice or insectsin the greatlibrary?These certainly were

41 Most recently,
Casson,47, adoptsthisview.
42Epiphanius, De mens.et pond. 11, quotedalongwithTertullian, Apol. 18, in Fraser
2:478 n. 132. See Fraser'sdiscussion,1:322-24,citingin footnotes
theotherevidence.
43And eventhere, we findlittlecomfort.M. Rodziewicz(above,n. 29), 321, pointsout
thatthecolonnadedspacesusuallythoughtof as a possiblelocationin theSerapeumwere
"destroyed in theearlyRomanperiod,"so thatthelaterRomanlibrary's"locationin the
temenosremainsunknownuntilnow."
44An exception is mycolleagueAlan Cameron,quotedin theNew Yorker, 8 May 2000,
p. 97. The notionputforwardbytheauthorofthatarticle,thatparchment is more"stable"
than papyrus,is, however,fiction.JamesO'Donnell, Avatarsof the Word(Cambridge,
Mass., 1998), 52, also pointsout thatrecopying intocodiceswouldhave beenessentialto
survival.(His statementsthatMenanderwas notcopiedintocodexformandthatparchment
predominated overpapyrusin codices,however,are erroneous.)Staikos,89, also concludes
finally thatdeterioration
was theculprit.

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
ALEXANDRIA: LIBRARY OF DREAMS 359

present in archives evenin drierpartsofEgypt.Wehaveplentyofevi-

denceforpapyrusrollsremaining in use fora century,and somefor
survival as longas twoor eventhreehundred years.45Butthatis about
thelimit,as faras we can see. The likelihoodis thatbythereignof
Tiberiusrelatively littleofwhathad beencollectedunderthefirst three
Ptolemies was stillusable.
Even withouthostileaction,then,the Library,or Libraries,of
Alexandriawould not have survivedantiquity. Indeed,any library
almostcertainly would have been a sorryremnantwell beforelate
antiquity, unlessits books wereconstantly replacedby new copies,
withtherollsbeingsupplanted in
bycodices thefourth century. The
ancientsalreadywereawareof thisnecessity: Jerome reportsthatthe
library at Caesareafoundedbythetheologian Origenwas restored in
themid-fourth century bythecopyingofthebooksontoparchment.46
Butthereis no evidencethatanysuchreplacement wenton inAlexan-
dria,noranyindication thattheimperial Romangovernment provided
anybookacquisitionbudgetto theLibrary. Thatdoesnotmeanthere
was none,butitis notlikelyto havebeenon thescaleneededto main-
taina trulygreatlibrary.
It is hardto giveup villains, butitlooksas ifwe mustabandonthe
searchforsomeindividual or smallgroupto blame.Thedisappearance
of theLibraryis theinevitable resultof theend of theimpetusand
interest thatbroughtit intobeingand of thelack of thekindof sus-
tainedmanagement and maintenance thatwouldhaveseenitthrough
successive transitionsinthephysical mediabymeansofwhichthetexts
couldhavebeentransmitted. It is idle,giventhisreality,to indulgein
such Gibbon-like reflections as the following claimof Hugh Lloyd-
Jones:"If thislibraryhad survived, thedarkages,despitethedomi-
nanceof Christianity, mighthave beena good deal lighter; itsloss is
one ofthegreatest ofthemanydisasters thataccompanied theruinof
theancientworld."47 Thisis to getthingsbackward.It is notthatthe
disappearance ofa library ledto a darkage,northatitssurvival would
haveimproved thoseages.Rather, thedarkages-if thatis whatthey
were,and in theEasternRomanEmpirewe maydoubttheutility of
sucha concept-showtheirdarknessby thefactthattheauthorities
botheast and westlackedthe will and meansto maintaina great
library. An unburned buildingfullof decayingbookswouldnothave
madea particle's worthofdifference.

45Forreferences,seeNaphtaliLewis,Papyrusin ClassicalAntiquity (Oxford,1974), 60-

61, withaddendain Papyrusin ClassicalAntiquity: A Supplement (Pap. Brux.23, Brussels,
1989), 32-33.
46See Carriker(above,n. 33), 22-23, citingJerome,Ep. 34.1 and De virisill. 113.
47Lloyd-Jones(above,n. 17), 117.

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
360 ROGER S. BAGNALL

Indeed,no morebookswouldhavesurvived antiquityiftheLibrary

had notbeendestroyed or accidentally)
(deliberately thandid so any-
way.The destruction simplyis not important. This mayseemlikea
bleakassessment, butitneednotbe so. It suggests thatwe shouldturn
our attention away fromthe dramaticsingleeventand towardthe
forcesandpersonalities thatcreateandsustaincultural for
institutions,
it is theirabsencein the Roman period,not the presenceof some
destructive force,thatdecidedthe fateof the books of Alexandria.
Whyshouldanyonebe disillusioned by the realizationthatcreative
achievements survive onlyifwe foster a cultural
milieuthatvaluesthem?
Most booksexistedin multiple copies,and it is thefailureofmostto
survive thatis mostimportant. The rarities
oftheAlexandrian Library
too owe theirdisappearance as muchto omissionas to commission.
I havedevotedquitea bitoftimeto showingthatthosewhohave
written abouttheLibrary ofAlexandria haveuseddubiousmethods to
arriveat improbable conclusions, pursuing I
what believeto be false
dreams.ButAlexandriais also a libraryof validdreams,and I shall
close by evokinga fewof them.First,and mostdirectly, theLibrary
and Mouseionsustainedforthefirst timea philological in
enterprise,
whichscholarstriedto establishcorrecttextsand to thinkaboutthe
artofdoingso. Theirearliestefforts werenotterribly sophisticated
by
ourstandards, buttheylaid thefoundations forall thathas followed.
Thisworkhad tangibleresults:In theliterary papyrifromEgypt,we
can see the point-startingabout 150 B.C.-at whichthe messy,
unstandardized tradition ofHomer'stextwas replacedbythestandard
textthatwe owe to Aristarchos ofSamos,whichliesat therootofthe
entireHomerictextualtraditionsince the second centuryB.C.48
Althoughour copiesof classicalliterature todayare notthoseof the
LibraryofAlexandria, mostofthemundoubtedly owe theirquality,if
nottheirsurvival, to thescholarsoftheMouseion.49
Second,theLibraryservedas thebase fora widerangeof other
scholarly activities,
scarcelypossiblewithoutitsricharrayof texts.I
cannotevokehereanything likethefullrangeof intellectual pursuits

48 See mostrecentlyJohannesKramer,"Die Geschichteder Editionstechniken und die

literarischen
Papyri,"ArchivfarPapyrusforschung 46 (2000): 19-40 at 22-23, puttingthe
Alexandrian workin thecontextoftheentirehistory ofcriticalediting.
49 LucianoCanfora,The Vanished Library(Berkeley,
1990), 197,concludesbyminimizing
theimportance of theLibraryof Alexandriain thisregard,claimingthat"whathas come
down to us is derivednot fromthe greatcentresbut from'marginal'locations,such as
convents, and fromscatteredprivatecopies."Actually,
muchofwhathas survived comesto
us throughConstantinople, whereverit ultimately wound up; Canfora'sclaim is thus
spurious.Butevenifitwerenot,itignorestheimpactofAlexandria(andothergreatcenters)
on thetransmission ofthetextsthatwoundup inmoreremotelocations.Cf.thehalf-hearted
rejoinderofR. Barnes(above,n. 17), 75.

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
ALEXANDRIA: LIBRARY OF DREAMS 361

supported bytheLibrary's collections,buttheyincluded manyattempts

to compilesystematic information about different subjects.50One
exampleis geography, whereEratosthenes was able to makedecisive
progressin creating themathematical foundations ofthatsubjectand
in enablingthedevelopment ofcartography.51
Third,and probablymostimportant ofall,theLibraryofAlexan-
driabequeathed theimageofitself, theidea ofa large,comprehensive
libraryembracing all of knowledge.As JamesO'Donnellhas put it,
"thelibrary at Alexandria has longloomedas a chimeraofpowerand
mystery on thehorizonof ourculture."52 The sourcestellus thatthis
reachextendedbeyondGreekcultureto theliterature ofitsneighbors,
rangingfromtheJewsto India.Theyprobably exaggerate,butitis still
significantthatalreadywithina centuryor so of its foundingthe
Libraryhad becomea symbolofuniversality ofintellectual
inquiry and
of thecollectionof written texts.53EvenifPseudo-Aristeas's storyof
thecreationoftheSeptuagintis fictitious, itshowsus thatinclusion in
theLibrarywas a kindof universally recognized validationto which
people would aspire.The Librarywas so farbeyondanything else
antiquityhad knownup to thatpointthatitembodied theseaspirations
and appealedto theimagination of all who wroteaboutit.Itsgripon
themindsofall whocontemplated itwas alreadyin antiquity as great
as itwas later,and ithardlymattered whatfanciful numbers theyused
to expressitsgreatness. Although the'authorswhoseworkssurvived
antiquitytold posterity littleof any concretesubstanceabout the
Library,theytransmitted itsindelibleimpression on theirimaginations.
This imagewas passed on to the Renaissanceand the modern
world,and everyone of ourgreatcontemporary owes some-
libraries
thingto it. By way of example,thepaperof mycolleagueCarmela
Franklin(below,p. 372) describeshow a Vaticanlibrarianof thefif-
teenthcentury wrotea LatinversionofTzetzes'pottedhistory of the
Libraryin themarginof a manuscript of Plautus.The contemporary
attempt to createa new universallibraryin Alexandriaitself,which
has receivedenormouspresscoverage,is onlythelatestrepresentative

soFraser1:447-79 givesa surveyof "Alexandrian scholarship,"butmanyothersections

ofhisbook are also relevant.
5iSee on this point Mostafa El Abbadi, "The AncientLibraryand its World-wide
Connections:The Makingof a WorldMap," Proceedingsof the 1st AnnualBibliotheca
AlexandrinaSymposium, 17-19 October,1998 (Alexandria,n.d.), 22-26. For a general
discussionofgeography in Alexandria,
see Fraser1:520-52.
520'Donnell,Avatars(above,n. 44), 33.
53The widestclaims,however, comein late Christiansourcesand maybe no morethan
embellishmentson Ps.-Aristeas;cf. Barnes(above,n. 17), 67. Theyare, however,quoted
withoutchallengebymostauthors;cf.,e.g.,Lloyd-Jones (above,n. 17), 116-17.

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions
362 ROGER S. BAGNALL

ofthistradition.Manyaspectsofthisprojecthavebeencriticized, per-
hapswithreason,54 butwe willhavetherightto denigrate theaspira-
tionsit embodiesonlywhenwe becomewillingto giveup our own
pursuitsof theAlexandrian dream.Thankfully, I see no signsof such
renunciation.Although it is too lateto recovermuchoftherealityof
thePtolemaic itsdreamis verymuchstillwithus.55
library,

54The mostseriousproblemat presentbeingthelackofa coherent collection

development
policyand fundsto carryit out. The beautifulworkingspace in thelibrary,however,is a
worthysuccessorto theMuses' bird-cage.
55Thanksto GlenBowersockfortheinvitation to deliverthispaper;to Alan Cameron,
CarmelaFranklin,G. N. Knauer,and Maria Panteliaforvariouscomments and references;
and to Mostafa El-Abbadifor offprints of rare publicationsand a visit to the new
AlexandrinainJanuary 2001.

This content downloaded from 137.108.145.45 on Tue, 9 Sep 2014 11:18:09 AM

All use subject to JSTOR Terms and Conditions