Orality, I-,Iteracy, Memory in the Ancient Greek and Roman World

Home , Apelles, Marsyas, Parrhasius (painter)

Orality, I-,iteracy,Memory in the Ancient Greekand Roman World

(Oraliq. and Literacy in Ancient Greece,vol. 7)

Editd b,l E. Anne N,'Iackav

?.t G ln j -a -$

:#2f/$/{ -) ftAA\ { /6g-j

BRILL

LE,II)[,N. BOSTON 2008 Mnemosyne

Supplements

Nfonographson Greek and Roman Language and Literature

Editortal ]Joard GJ. Boter A. Chaniotis K. Coleman IJ.F. deJong P.H. Schrijr.ers

\ T)LU\IE 298 CHAPTERELEVEN

MEMORY VISUALCOPIES AND '

JOCELYNPENNY SMALL

We live in a rvorld of copies not just of books and art, but of virtually el'erything we use from computers to cars to the furnishings of our home and the games we play. We are so surrounded by facsimiles and reproductionsthat it is dilficult for us to imagine a world with limited meansof making copies.It is jolting to rememberthat the assemblyline rvas an invention of the Industrial Age and did not become a major eco- nomic force until Henry Ford producedhis Model Ts in the early 1900s. It is not that copies did not exist in classicalantiquity, but rather that their nafure differs in some cases dramatically fiom modern ones. We expectour copiesto look so like their originals that not even an expert can distinguisha digital reproductionfiom its original. In antiquity,except lbr certain restrictedcategories of die- and mould-made objects, like coins, seals,and lamps,each copy could generallybe distinguished from every other. While classicistshave long been accustomedto the idea of variationsbetrveen stories and manuscripts,classical art historians approach the problems of copies with an ingrained bias toward Greek art that makes them treat Roman copies, if they judge them aes- thetically fine, as exact replicas of lost Greek originals. Although that bias has begunto shift in recentyears in the study of sculpfure,painting

I It rvas a grcat honour to have been invited to give the keynote addressat the SeventhIntemational Orality,lliteracy Conference.In particular I would like to ex- pressmy deepgratitude to Anne Mackay for her exemplaryorganization of the con- ferenceand lor her gracioushospitality. The reactionand commentsfrom the atten- deesn'ere most helpful and are reflectedin the notes.I would especiallylike to single out Ed Cararvanlbr our refreshingdiscussion. It is a pleasure,as always,to ac- knorvledgethe help of A. A. Donohueand SusanWoodford, both of who made the supremescholarly sacrifice of readinga draft of this paperwithout the notes.I also thank Brunilde S. Ridgrvayand Miranda Man'in for their observations.Please note that referencesare kept to a minimum both for objectsmentioned and the extensive literatureon copies. All translationsare from the Loeb ClassicalLibrary editions unlessothenvise noted. All rvebsites were accessed in March2008. tt8 JOC'EL\'NPF,NNY S1\lALL

hasreceived little attention.rNor haveclassical art historiansconsidered the implicationsof theresults fiom studiesof oralityand literacy.In this paperI shalltry to redressthat lack ofbalance. I begin rvith a considerationof r.vhatGreeks and Romansthought about copies. The English r.vord"copy" comes from the Lattn copia, rvhich,however. does not mean"copy" but "abundance"or "plenty"- meaningswhich explainits laterextension to our senseof "copy.": ps1- litt's extremelyuseful compendium of technicalGreek and Latin words for art history, TheAncient View of GreekArt, containsno entry in the indicesfor "copy." With a knowledgeof Greekand Latin,however, one can find nopoberypo and exentpluir together with exemplar. Pollitt notesfor the Greekterm that its "basicmeaning ... is 'model' or'pat- tern'."'1 Similarly, he saysthat "the |erms exemplumand exentplarcan meanboth 'model' and 'copy.' When the u'ord means'copy,' however, it alwayshas the senseof a 'representativecopy' and henceis still very close in meaningto 'model'."s ln other words, the Greek and Latin rvords focus on the sourcefbr copies rather than on the copies them- selves,ironically like scholarstoday.0

2 Lippold(1951) remains the basic study fbr theidea that Greek paintings stand behindalmost elery Romanpainting. Bergmann (1995) is oneof the fervto consider painting.Hallptt (2005: .133-35) has a brief sectionon paintingin his reviervof Gazda(2002)and Perry'(2005). Even the recent lascicle of lrt Hrstor_r-(Trimble and Elsner[2006]), devoted to theproblem ofclassical copies, has no articleon painting. J Accordingto theOED Online(s.v. copy A 11.3),the meaningof "copy" as "a pictureor otheru'ork of art,reproducing the featuresof another"dates to I 584.The earliermeaning, more literally after the Latin, as "abundant"or "copious" is dag- geredas obsolete(A I.l). The earliestcitation is 1596for "copy" as "something madeor formed,or regardedas madeor formed,in imitationof somethingelse; a reproduction,image, or imitation"(A II.4a).It is probablynot coincidentalthat the modem meaningof "copy" as artistic reproductionfolloll's the inventionof the printingpress with its multiplecopies that are portableand hencecan be compared to eachother. Compare Muller (1989),who similarlydates the beginningof the desirefor "authenticity"to the sixteenthcentury. 1 Pollin (1974 2ll). tvno5 is anotherproblematic word, when usedin sculptural contexts.It probablydoes not mean"model" but rather"mould" or "reliel" both of which termsremove it lrom my currentconcem about "copies." See Pollitr (1974:272-93) for a summaryof the scholarshipand especially291 for the''best" usage.I thankA. A. Donohuefbr bringingthis term to my attentionin this context. 5 Pollirt,l914:367\. 6 This usageparallels the classicalinterest in firsts.Pliny the Elderrecords who inr.entedrvhat artistic technique. That sometimesthe stories,such as for the invention of portraitsin clay (HN 35.151)and paint(HN 35.15)are the samedid not botherhim. if he notrcedat all. It n'ould appear,then, that the classicalinterest in firstsparallels the moderninterest in originalsexcept that Greek and Latin seemjust VISU,\L COPTF-S,ANDNII1NIORY 229

It is thereforeno surprise.as isagernotes, that "the extensiveprivate marketin modificationsor adaptationsof Greekart constitutesan area rvhichPliny [the Elder] fails and probablydid not wish to include."In fact, Pliny refersonly once to a copy of a painting.rThe referenceis instructive:

In hisIouth Pausias [the painter] loved ... Glykera,thc inventor of flou'er n'reaths.imitating her in rivalrv lcertandoqueimitotionel he extendedhis methodof encausticpainting to representa very numerousvariety of 1'lorvers.... A copyof [his] panellhuius tabulae eremplarf [of Glykera], an onoypoqov as they say,by Dionysiosin Athensr.vas bought by LuciusLucr,rllus tbr tu'o talents.8

I find it interestingthat Pliny falls backon a Greekword, becauseLatin lacks the appropriatervord.e Norv the absenceof a particularword does not mean that a particularphenomenon does not exist, but ratherthat no need $'as f-eltfor such a word. For example,Latin u'as quite contentto use the same rvord, pollex, for both big toe and thumb.10 Sometimes contextis all. In this case.horvever, I do not think that contextfully accountsfor the absenceofour senseof"copy." In the first partoithe passage,Pliny refersto "imitating . in rivalry"-two tenns we are accustomedto seeing in classicaltexts on copying."Rivalry" obviouslymeans competi- tion and a numberof anecdotesdescribe both formal and intbrmal artistic competitions.llFor the most part,I am not concernedwith that as- pect here."lmitation," however,is a more complexterm that may in- cludecopying but doesnot haveto.tz I could,for example,be inspired by Seuratto paint a picfureusing only dots of paint.My paintingneed

as parsrmonious rvith rvords for "original" as they are for "copy," since Pollitt ( 197'l) similarly doesnot have a listing fbr "original." 7 Isager ( l99l: I 74) for both the quotation and the information. Lucian (Zeuxis 3-5) refersto an ''extremelyaccurate copy" (3.10) ola painting by Zeuxis in terms remarkably' similar to the rvay scholars today refer to copies. Yet, as will be seen, there is no x'ay for Lucian to have knorvn horv accurate the copy is. since the original. according to him. rvas lost at sea. Lucian, like Pliny the Elder, uses a similar uord to ref'erto copy. ouriypoqo5. 8 Plin. l1N 3-5.125(my translation). e A similar situation exists rvith "svmmetrv." Comnare Plin. l1N 34.65: non ha- hct Lulinum numen .\)mmelrio. t0 OLD I i9". s.r . 1',,//,:r ll The most f'amous"contest" for artists that rve knorv of may be the one among live sculptors to make the best Amazon, on rvhich see Plin. HN 34.53. For another example in painting, consider that betrveenZeuxis and Parrhasios(Plin. HN 35.65). I2 In general.on artisticimitatio seePerry' (2005: I I I -22). 230 JOCEL\N PENNYS]\{ALL

not sharethe samecolours rnuch less the same subject as any of Seurat's paintings.It rvouldonly looselybe an imitationof his sfyle.In a sense this is the kind of imitationPseudo-Longinus (Or the Sublime,13.2_4) describesrvhen he sa1'sthat Platoimitated Homer. when art historians, horvever,ref-cr to "copies," they generallyare not talking about inspira- tion as imitation.They mean somethingthat has the samesubject and elementsas the original and is portrayedin the samemanner. The three requirementsof subject,elements, and style must all be met. rnthe Acodemica(2.85-86) cicero talksabout such exact replication: Tellme. could not Lysrppus, by meansof thesame bronze, the same blend of metals.the same graver and all theother requisites, make a hundredAl- exandersof thesame shape fmodil? then horv fqua ... notione]would you tellthem apan? Well, rf I imprinta hundredseals with thisring on lumps of waxof thesame sort, rvill there possibly be any means of distinctionio aidrn recognizing them? Or will youha'e to seekout some ring_maker?13 It is significantthat cicero chosefrvo types of reproductionthat really can produceidentical copies. Because the case for identicalsealings from a signetring is obvious,I discussonly bronzeshere. classical bronzestatues are a rariry today, becausebronze was presumablyivortn more as money as material than as art. Moreover, what has survived seemsto be r,'ariantsrather than exact replicas.A stock type received modifications from minor adjustmentsin pose to the treatmentof details. Mattuschpresents the somervhatsurprising example of the Riace bronzes.r+At first. and even second,glance the differencesin their headsmask the samenessin their bodies,in part becausewe are ..hard wired" to notice heads,and not just headsbut faces-a fact rvhich ex- plains, in part, ivhy the Romansconcentrated their efforts on the heads for their portraitsand oftenused stock bodies.r: Gazda presents the example of Vespasianand ritus from the Shrine of the Augustalesat Misenum,made afler both had died.roLike the Riacebronzes, only the headsvary. The skill needed,however, to replicatestone images may be greaterthan that for bronzes,which can repeatedlyuse the samemoulds.

l3 Translationadapted lrom theZCZ. Compare platt (2006). rq . Ivlattusch(1996: 64 and66-6i, fig. 2.18). For example, Botbein (1996:72) re_ f'ersto "the sf.v'listicallyearlier of the bronzervarriors from Riace,,un.l h.n". io.s notsee them as trvins in bodv. ri Massironi2002. 14-4i . r6 Norvin Baiae.Castello, after AD 96. Gaztla(lgg5. l4l-42 and 155.fig 7). Steri'art(2003: .17-59) discusses the practiceof the individualizedportrait"head ioinedto a qenericbodv. VISUAL COPIESAND N,IEN{ORI zJl

The mostobvious extant example of Romancopies of a Greekorigi- nal is that of the Erechtheioncaryatids with repiicas in the Forum of Augusfusin Rome and at Hadrian'svilla at Tivoli.tr Becausemoulds can be taken from eristing statues,as Lucian mentionsfor a Hermesin the Agora at Athens,there is no logisticalreason why the threesets of caryatidsshould not match.18 Moreover, for us today it is a relatively simple matter to compare the three sets through photographs,which demonstratethat the copiesmeet the criterionof "close enough."The Romans,however. rvould not have beenable to seeeven two of the sets of caryatidstogether. Statues.no matter the material, can be shipped from site to site. Bartman suggeststhat copies of official Roman portraits in lighter weight materialslike plasteror wax would have beensent to various parts of the Roman Empire for copying locally.rr Yet that does not mean that they are identicalportraits, such as for QueenE,lizabeth II in British embassiesthroughout the rvorld. Bartman, like Maftusch, comments on the fact that "variants are frequentin Livia's portrait corpus, perhapsmore the norm than closecopies ... [becauseof] the rudimen- tary natureof the systemby u'hich it [the poftraits] rvasproduced."2o In particular, she notes that "the Roman sculptor ... seemsoften to have reproducedassiduously those aspects of the image that were unfamiliar while executingmore freely those he alreadyknew."2r To put it in Thu- cydideanterms, even w,herewe might expect precision,generally only

l7 Schmidt(1973) is the basicstudy fbr all copiesof the Erechtheioncaryatids, includingthe threementioned in the text. Shealso provides full photographicdocu- mentation:for the carl'atidsfrom the Forum Augustum,Rome: pls. 1-5; for the caryatidstiom Hadrian'sVilla: pls. 6-32. For the Erechtheioncaryatids see, among nranyothers, Stervart (1990: pls. 431-32).On "exactcopies," see Perry (2005:90- 96)u'ith trvo caryatids from Tivoli illustratedon92-93,ligs. l9-20. l8 The Hermeswas so lrequentlycopied that it becameblack from the pitch used.Lucian, Iupp. trag. 33. CompareMattusch ( 1996:l9l ). Ie Bartman(1999: l8-24) discusses the logistics ofcopying. 20 Bartman(1999 20 and24). CompareAlbertson (2004: 300) who, in a study !'|as of portrartsof MarcusAurelius, says that we progressfrom the lst throughthe 2nd centuriesthe actualcopying of an official model becomesmore accurate,the dependenceon modelsgreater and greater."Baftman obviously illustrates a number of Livia portraits,but one of thoseon u'hich she focusesin this sectionis the head norvin Baltimore,The Walters Art Gallery23.21l: Bartman (1999: 19, figs. 13-14). I also know of one instancewhere "copies" were madebased on a verbaldescrip- tion, but obviouslynot of a portrait.The deviceof the Marsyasin the Forumshows a distinctdilference on GreekImperial bronzes compared to the originalin the Forum Romanum:his right handis no longerraised above his head,but in an adloartio.See Small(2003: 1l,l-16 with figs.58-59). 2l Bartman(1999: l9). 232. JOCELYNPEN.NY SN'{ALL

glst is captured.22If sculptorshave trouble accuratelyreproducing heads,consider rvhat may liappenwhen copying statuary groups. To understandthe problem, first try a thought experiment.Imagine the Laocoon,a three-figuredstatuary group.23 We now know that Lao- coon's right arm no longer extendsmore or less straightup in the air, but is bent back at the elborvtoward his head,which falls to our right in intenseagony. Now think of his two sons.Which is the older boy and rvheredoes the snakervrap aroundhim? Where is the headof the snake that bitesLaocoon'? Where is the secondsnake's head? Which is Lao- coon'srveight-bearing leg? Are therethe usualbits and piecesof drap- ery and, if so, rvhere are they? The more questionsI ask, the more I hopeyou will realizethat, like me, you reallydo not havea clearpicture in your rnind of this u'eil-knorvngroup. If you look, fbr example,at an illumination from the Vatican Vergil, Laocoonlooks quitereasonable, even if his two sonsare awfully small and his red cloak in contrastrather voluminous.2a His left leg is the weight-bearingleg, becausehe kneels on the altar with the other one. The snakesare a bit hard to find, but they encircleLaocoon and the boys aroundtheir torsosand aroundthe arms of Laocoon.I passover that the image of Laocoon, on the lett, clearly labelied, has no beard and no cloak, but is dressedllke a victimarius. Let us try anotherversion, the marvellouscartoon by CharlesAddams (Figure 5).zsTypical of twentieth-century artists, he has placed the group in a specific, three- dimensionalsetting and has basedhis rendition on the earlierrestoration of the Laocoon with the right arm extendedupward. What about his sons?They, too, are raisingtheir right hands.ls that correct?Let us look

22 Thuc.1.22. See my discussionin Small(1997: 191-93). 23 The Laocoonremains the subjectof long debateas to whetherit is a Roman copy of a Greekoriginal or a Romanoriginal, and even whetherthe Laocconwe haveis the onePliny the Eldermentions. Deciding the answerto thesequestions has no beanngon my useof it as an iconicexample thar everyone'.knorni." See Bril- liant (2000:2-3 figs. 2-3 rvith the two differentrestorations, and 98 fie. 20 for the backview); also Ddcultot et al.(2003), and Varner (2006: OTql with biblrography. 21 Verg.Aen.2.l91-98: Folio 18v. See Wright (1993:22-23), Brillianti2d00: l3 fig. 8),and Small (2003: 149-150 rvith fig. 69). 25 Addams(1991: 215). First published inThe New yorkerfor April 17. 1975. To date,The Nett'Yorkerhaspublished four cartoonsspooting the Laoioon:anorher one by Addamsfor November22, 1982,one by William O'Brien for January25, 1958,and one by VahanShirvanian fbr January12, Iggi. I treasurethem all. but thisone the most. For all of thesccanoons, see the rwo CDs'thar came uith Mankolt (2004);search under the date, the artist, or "Laocoon." \iISUAL COPIESAND NIElVtORY LJ) at the "real" Laocoon(Figure 6).:t' 4ntt,'. licenseis perhapstoo kind a rvord. The illuminator of Ihe VctticanL'ergil not only gave Laocoonfwo little bo-vsinstead of an older boy and a youth, but also has changedthe poseofLaocoon, rvho now helplesslyraises his handsrather than vainly trying to removethe snakes.In addrtionthe statuaryLaocoon is more or lessseated on the altar u,ith his left foot touchingthe ground to the side of it. The older son on our right is trying to stepout of his snake,so to speak.u,hile the younger one is moresecurely ensnared. The headofthe secondsnake is difficult to discern,because the left handof the lefi son coversits head.as he triesto pushit alvay. The phenomenonyou havejust experiencedis known as recognition memorv. When l,ou seethe Laocoon,you knorv it. To understandwhat I mearr,consider the infamouspenny test from 1919.21Diabolical cognitive psy'chologiststhey areahvays diabolical-showed fifteen possible obversesfor the Americanone-cent coin and then askedAmerican col- lege studentsto identily the real one.Most could not pick out the right one.I havererun the testrvith my studentsand had the sameresults. As Nonnanpoints out, "the students,of course,have no difficultyusing the money:in nomal life. ti'e have to distinguishbetween the penny and otherU.S. coins.not betweenseveral t'ersions of one denomination."28 SinceI am talkingabout a basichuman skill, the Romansand their art- istswould havethe samekind of recogniticlnmemory that we haveto- day. Artists making copies.hou'ever, should be better at this task than \\'e are. Some experimentalevidence exists to suppotl that position. When chessmasters are askedto memorizethe arrangementof men in the middle of a game,they recall it with remarkableaccuracy. On the other hand. rvhen the same chessmasters are askedto memorizea chessboardrvith randomly sprinkledmen, they do no betterthan anyone else.reHence it is likely that Romanartists might not havebeen able to

26 As Brunilde Ridg$a.v (pers.comm.) points out, the illumination inthe vati- con L,ergil does not neiessariiy depend on lfte Vatican Laocoon sculptural group Yet. $hin rve think of a representltion of the death of Laocoon, rve think of the sculptural group. Moreover, because $e knorv the sculptural group, we can recog- nize the same subject in the lratican vergil. At the same time we consciously or unconsciously "n]easure" all Laocoon '"'ersionsagainst that group. On the icono- graphicalhisio' of Laocoon. seeLIMC 6, pp. 196-201rvith pls. 94-95: "Laocoon" (Erika Simon). lr Nickersonand Adarns( 1979: 297. fig. 6). lE Norman (1988: 57). :9 Cognitive psychologists refer to this phenomenon as the differences between erperts arid no.iiei. the rdsultsofrvhich uould also apply to the abilities ofexpert 234 JOCEL\N PENNY SN{ALL

haverecalled today's modern art accurately.but only piecesrvithin their eXPertlSe.l0 If artistsare able to copy an object right in front of them, as in a sec_ ond cartoonof the Laocoonby William O'Brien,lr they shouldhave no problem with accuracy.Yet Bartman says that, at least for porlraits of Livia, accuracyis a real issue.32I do not know ofany sneakycognitive tests of artists' memories comparedto ordinary folks. I do, however, have the cover from a TV Guide that appearedshortly after I had ac- ceptedthe invitation to deliver the keynoteaddress represented by this paper.33TV Guide re-creatednine famous covers,one of rvhich shows RebaMcEntire taking the pose of Lucille Ball in the iconictrampling of the grapes.:+Because the originalcover was availableto there-creators.

artists and of the lay public to remember precisely hou,a statue looked. Here I have chosen a particular variation that tested visual memory. Also note that the chess masters recall onll' the layout and not rvhat the pieces look Iike. I thank Barbara Tversky for this observation- The bibliography on the topic is noiv quite large: for a summar)' see Didierlean et al. (2004). l0 ''Scurptors compare Dion. Hal..Derr.50: and paintersrvihout long experience in training the eye bv studf ing the works of the oltl masters *,ould nJt b.l able tu identily them readily, and would n_otbe ableto say rvith confidence that this piece of phidias, sculptureis by Polyclitus,this by this by Alcamenes;and this painting is by Polvgnotus.this by Timanthesand this b1,parrhasius." 3r William O'Brien: The Nev.yorkei, January25, l95g; Mankoff (2004: CD). Cohen (2005: 997) "attempts to determine rvhat those rvho . ^12 draw o..r.ui.ty do differently than those rvho do not" and concludes that "high gaze lrequencies iray facilitate drari'ing accuracy by (l) allorving the artist to hold less information in $'orking memorl', (2) reducing memory distortion, and (3) facilitating the reduction -a of context elfects through inattentional blindness." In other words, copyisr must look frequently at the original to get it right. presumably the sculptors gu;r'ru* ..oriqinal." di. cussesdid not comparetheir Livias to the )\ TI Gttide(October q-lb) 2005. il l\lichael uagarin suggesteriar the Lonlerence thar there is a subsrantivedil- ference "re-creation" behveen a and a "copy" and that therefbre I *,as .*p""ting o greater yet ..nine accuracy than TV Guide intended an examination of the tribirte covers" shor.vsthat TV Guide went to great trouble to choose look-alike stars and then to pose and dress them as closely as possible to the original .ou.rr. I u"ti.u. that g'ord they used the "re-create' rather than "copy," becausJthey viewetl the cor,- ers more.as."copies" of live performances and hence "re-create" mo.e precisely captured iheir intention. one does not "copy" a performance. rn fact, TV t;uiai 1p 30 of the same issue as the cover) said in the iaption to a photograph of Reba-"s blouse being adjusted beforettre shoor, "every sritc-hhad to mitch ttie originat.'; itre onginal cover can be seenat Ihttp://www.tvguide.com/ce]ebrities/lucille-b"alllphotosl 16302514):the re-creation is at [http://wwrv.tvguide.conr,/celebrities/reba-mcentirel photos/171072/341. In any case,.my basic point remains valid: except ror aigital reproductionsit is nigh impossibleto get all the detailsright, especiallyin cc,mllex scenesrvith perry'(2005: numerous figures and objects. 91) commerus...it w;s r. basic tenet of stoic thinking ... that there rvas no'such thing as ar exact duplicate of any object in the phenomenal world." The TV Guide re-creations also should be VISUAL COPIESAND N{EMORY 235

thereshould be no inaccuraciesexcept tbr the fact thatthe newerrendi- tion is in colourand the originalr',as shot in blackand white. Both Lucy and Reba are similarly dressed,but their headscarfshave different pat- terns.Reba manages to keepher blousefirmly on both shoulders,while Lr-rcy'shas slippedoff her right shoulder.Their handsdo not match. Rebaholds her right thumbout and extendsher lefl fingers,while Lucy has fornied loose tlsts r,i,itheach. Lucy is looking more downward witn her eyeshalf-closed compared to the open-eyedReba. Like a good art historian, I could go on, but I think I have mentionedenough cliff-er- ences.Making an accuratecopy, even in the best of circumstances,is not easy.For the purposesof TV Guide the trvo imagesare closeenough ,.close evenin a directcomparison, but for art historians enough,,is of_ ten not enough.we rvantto knou' exactlyw'hat the originallooked like. Unfortunatelyeven today rve live in a Thucydideanworld. I wonder if your mind dritted off during my comparisons,because basically many of us do not care about that much precision.Good enoughis good enough.I think Romansmust have been similar. Some cared for accuracy; most rverehappy $'ith gist.ii Thereis additionally,of course,the f-actthat most could not easilycompare original and copy in the absence of photographs. with this backgroundlet us switch our focusto classicalcopies of classicalpaintings. The situationimmediately becomes more compli- cated. The only securely identified copies I know fbr painting are on The Museumof Fine Arts in Boston 'ases. owns two Attic recl-hgure kylikes by Aristophanesrvith identicalsubjects and scenes.i6The mu_

distinguishedlrom parodies.rvhich reproduce the poses,dress, and setting,but *,ith t\\'rsrs.onthe originals to amusethe vierver.For eiample,smithsonsian lloos: t to.1 compiledten takeson the classicGrant wood paintingAmerican Gothic that range tiom cartooncharacters (Beavis and Butthead)to vizs'ladogs to paul Newmanaid hisdaughter Nell. on theproblems rvirh our terminology, sei Ragghianti(1964: 141. comparethe title of Bergmann'sarticle (1995): "crJet uaste-Siecesand Roman RecreativeFictions." l5 (irmpare ._ Perry-(2005: 60): "This [the'ariety in the postureand proportions ol'the olympias-Aphrodite.sculpt'res] implies thit it migLt havebecn u'g.n.rol risual tamiiiarity.and not,!....I.:.t replicationof a particilar model,to rvhichpa- tronsand responded."Fullerton (2006: 4g3) .,Alternatively, 'ie*ers suggesrs, per- hapsthe imageswe haveare signs. the form of *hich der"iielrom a mentaliri"gi the-things being signified . .. but rvhichemploy a fe*. salilentfeatures ... thatrv"ould "r :ul-fiecto indicrtethe suhjert. ... in an1case. no onedisputes that these represenra- tronsare nor pictorialll' accur.rte, but we might questionmore carefully rvheiher they wereeven intended to beso " 16 Bosron,Museum of FineArrs 00.3214 and 00.315(ARI,' t319.2and 3 respec- tj'ely. Para. 478;Add.:363). They dare ro c.425-400BC. For onlinephotograplis of PENNY S]VIALL 236 JOC'EL\N

Seumevendisplaysthevasesnexttoeachother.Likethephotographsol separatesthe two' For example'in Rebaand Lucy, only a closeanalysis who still holdsDeianeira' In addition ih. tonao, Herakleshghts Nessos the two vasestoday' Boston to the issuesof preservationthat distinguish and Boston00'345 doesnot' 00.34,+has an inscriptionin the exergue at the ankle differs slightly in its Th. lo'u., ecigeof Deianeira'sdrapery fbrmer' Similar discrepanctes treatmentand she has larger feet on the fighting centaurs'For ex- can be found in the exteriorscenes of Lapiths reverseshas slightly difl'erent ample,the letlmost centauron eachof the extendmy analysisof gestureslbr his armsand hands'37 Again' I could suchdetails,butnoneisasstrikingasthefactthattheyaresoclosely copyingdoes not concemme rnatched.More importantly,this kind of producedat the sametime in the here,because the trvo uur., u.a nl'ins, ln a senseboth are originals' same rvorkshopby the same painter'18 rvaspainted first or even if since there ls no way tbr us to knorv which Nor is there any way to tell more of theservere made at the sametime lvhetheranothervasewasthemodelforthesetwo.Similartwinsexistin Immediateduplication of objects sculpture,such as Kleobis and Biton'le iseconomicalnomatterrvhatthematerial,becauseoncetheartisthas the next one will be that figured out how to make a particular object' to be carved or painted fiom much easier to prodttce even if it has scratch. of copiesseparated tn I am, horvever,concerned lvith the production sole example I know of a time and spacetiom their "originals'" The in both model and copy occurson an Etmscan fainting that has sun'ived from an Attic red-figure red-figurekylix that adaptsthe exterior scenes y76tt- differ-satyrs for the Etrus- tyfi*ifigu.. 7)''ro the interior scenes

Boston00.344:BAD220531and220535respectively;alsosearchbymuseumano **Uer athttp:/irvwu'Perseus tufts'edu' tt""i".v , armby hishead and his 37 On Boston00.3'+4 the leftmost centaur has his right hands'areraised on eitherside of leftarm extended; uu, on"ioiton ilo.:+r uothhis his head. jsConnor(1981)drscussesthesamephenomenonofcontemporary..replicas.'' the Painterof Louvre F 6 (ABV but in this iase tbr A;;;j";[ lt!u"'puinttt' i',". 50-53;"tAdd2 31'35)' \)ijb, inv 467 and 1524' Stewart le Marble, early 6ttr cent B'C: Delphi Mu,se^1m sso' t -s)' Boardman(2006: l7 fig 9 and ( 1990:pl. 56) on "trvins,';;; M;"tti t 29n.35rvithbibliograpl,yl..n'u.r.'thattiretrvofiguresare..intermittentlyregarded asbeing rather the Dioskouroi" 980;c 425-'100BC' Martelli l0 Etruscan,.a-f,gu"ityfo:Paris' Rodin Museum kvlix oedipusPainter' No. tdo with pls 210-2ll) Attic red-fisure trqgl:J)o-il Para 37"6;ttli.:z+t; BAD20s312). Varican16541: c.150 tst'i;I}.rii.i. ris+, vISU.\L C,OPIF,SAND NIEN{ORY 231

kylix the exterior can kylix and Oedipusand the Sphinxfor the Attic Scenes0|bothshorr,satyrsrevelling.TheEtnrscanvase-paintersimpli- side with the result iled the r\ttic sceneb1' removing one f-tgureon each becausethe that one of the stdesmakes less sensethan the origrnal. On the other youth about to be beatenwith a sandalhas been omitted' a pot (Figure8)' On ,ia., tt.rcEtruscan artist ornitted the satyrpissing in awklvard,but the tuhole the Etruscansatyrs Seem iess elegant and more is actuallymore the leltmost Satyf,Seen tiom a three-quartersrear r'ierv, one Henceit is accornplishedon the Etruscankylix than on the Greek is clearlya very importantto keepin mind thatjust becausesomething than the original' .opy do.. not mean that the copy cannot be better u,hetheroverall or only in sections,as in this case' presentgreat Large-sca1epaintings' either on panelsor on the n'all' for making difficultiestbr the cop-vist,because the methodsavailable for mould-made copiesof paintingsdo not permitthe accuracypossible transportable obj..ts. Panelpaintings do hal'e the advantageof being a painter'sr'vork- uni ur. thereforecapable of being directlycopied in paintersettlng up shop.Alternatively the reversecould happen, with the "Shall we follow his easelin front of the original'Quintilian remarks' is to be able to the erarnpleof thosepainters lpictoresf rvhosesole aim measurementsand lines copy pi.iu.. s ltlescribere tabulasf by using to examine vocabu- f.rrri,ri, oc lineisf'l"ttlt is imporlant Quintilian's for he usesthe iary. First, he may be referring only to panel paintings' yvordtabtt|a.second,nlenstracaneasilybetranslatedas,.measure- mean"lines" or' ments." but lineaels more cornplicated'lt can simply accordingtoPollitt.n-rayrefertothedrarvingofoutlines'+2HenceQuin- practtce tilian is not talking about imposinga grid on the original-a meansmaking that uould destroythe original-but insteadhe probably surethatthebasicsketch'thatisthelinesoftlrescene,isaccurateby precisionof such checkinglts measurements.alHe doesnot assessthe

ltisimportanttonotethattheAtticred-lrgurekylixrvasfoundinVulcr,theprob- Ior the Ftruscankylir' -"^i,.rble pr.rrenience atlapted from LCL (lst ed.). Butler..translates [uint ]nst. 10.2.6.ri"rir"tii," "b1- and the measuring rod"'based on th. ptlrir" ,t"r, uris ac, lineil as using the ntler a grid l have provided a more the assumption that the orrginal ivas dii icletl into literalreadtng. rr Pollitt-(197'l: 392-195 s.v. liteanrenta' linea) they could have used string for a ll A. A. Donohue 1pers.comm.) suggeststhat temporarily to the tcmporar)' grid. Such a gnd could be aitached Picture !I-bl-t:jl such a grid on a wooden lrame ruax. It mie'ht also be possible. ibr example, to mount we have no evidence for or against rhese sugges- ;;'pl;,"..';i il"inr,,i-l. original. 238 JOCIfTLYNPENNY Slv{ALL copies.Nor doeshe mentionthe problemof colournor how thosecol- ours are laid on, suchas with broad,visible strokesor small,nigh in- visibleones. In contrast,both Pliny the Elder and Piny the Youngerdiscuss colour. Pliny the Elder deploresthe inaccuracyof illustratedtexts on bot- any: Crateuas,Dionysius and l\4etrodorusadopted a most attractivemethod, thoughone *'hich makes clear little else except the difficulfy of erriploy- ing it. For theypainted likenesses leffgies) of the plantsand then wrote underthem their properties. But not only is a picturefpictura] misleading [/a11arJshen lhe colours are so rnany.parlicularly as the aim is to copl foemulationenrlNature, but besidesthis, muchimperfection arises from themanifbld hazards of thecopyists.aa Pliny doesnot rate the ability of the copyistshighly, becausethey are virfually incapableof achievingaccurate colours. Even todaywe simi-

tions. The best evidcnce rle have ibr grids comes liom Egyptian art, but as lar I knou'. the grids uere used for the paintings being made and not on ones being copied. The grrds u,ere used not as a mechanism for making accurate copies but fbr rnaintaining the appropriate proportions for figures: the Egyptian "canon." Robins (1997: 109) sa1.s,"Grids, uhich were usually laid out in red paint, u'ere often un- even, and it is clear that artists $ere not aiming at mathematical accuracy. The lines were merely'an aid to drauing acceptablyproportioned human figures, and artists drd not have to lbllo*'them slavishly." Compare Davis (1981:64-65).I thank An- thorry Spalingertbr discussingthis matter rvith me. Cennini (1954:1.23-26, pp l3- 1.1)suggests using tracing paper, "fastening it nicely at the four corners with a little red or green rvax." He also giles instructions on horv to make tracing paper by scraping parchment and then soaking it in linseed oil. While parchment was avail- able from the second century BC and later, this particular use is not mentioncd, to my knoii'ledge. in ancient sources. It should also be noted that it is one thing to make a grid for an "original" painting and quite another to use one for copying. For example, the scene rvith the rape of Persephonefrom the "Tomb of Persephone" at Vergina may shorv faint traces of a grid with very large squaresthat are more useful in roughing out the figures than in making precise copies of existing figures: see Gallazzi and Settis (2006: 40-,11,figs. 15-16). Scheller (1995) offers a rvonderful compendium on model-books that also considers the Egyptian and Classical evidence. Note especially that (p.72) "the second half of the lzlth century saw an up- surge in the application of labour-savingdevices." He also remarks (p.383), "ln 1839 the French archaeologistAdolphe Didron called at Mount Athos on his tour of Greece. He was amazed to see how the Greek fiesco painters designed their large compositions directly on the wall, \r'ithout any preparation. In the West this method had been supersededin the late Middle Ages by a step-by-step,complex and time- consumingdesign process." aa Plin. HN 25.1-5 (8-9). Translation adapted fiom the LCL. On copyists "im- proving the original." recall how the Etruscan artist improved one of the figures he rvas copf ing from an Attic red-figure ky-lix. Compare my discussion on the "Repro- duction ofPictures" in Small (2003: 134-38). VISUr\L COPIESAND N4F.llloRY 239 larly lamentthe lack of accuracyin colourphotographs' some of which, even in expensivecoffee table books, are often wildly off from the originals.Pliny the Elder's nephew,Pliny the Younger,addresses the difficultiesof accuratelycopying portraits painted on panels,a process that involvesmany of the sameproblems as copyingmanuscript illumi- nations.He writesto Vibius Severus: Thelrell-kno*n scholar Herennius Severus is Iery anxiousto placein his library portrairclimaginesl of your fellorv-townsmen,Comelius Nepos and Titus Catius,and asksme to havethem copied ferscribendasf anc colouredlpingendasl if, as seemslikely, theyare ln your possesslon" ' All I askis thatyou find asaccurate fdiligentissimrn] a painterfpiclor] as youcan, fbr it is hardenough to makea likenessfrom life lex verof,butan imitationof an imitationfimitationis imitatiol is by far the mostdiftlcult ofall. Pleasedo not let theartist you choosedepart from the original even to improveon it.45 Pliny the Youngerrecognizes the variabiliryin individualreproductions of riorks of art.16He describesthe processof reproducingpainted portraits as requiring rwo steps:the drawing or outlining of the figure, like thelineae of Quintilian,and then the additionof colour.It makessense that the sameorder would be followed no matter rvhat the subjectand thus appliesto all painting.What is interestingfor us is that Pliny considers"an imitationof an imitationby far the most difficult of all." In other words, Pliny the Younger, rvho should be acquaintedwith both originalsand copiesof paintings,implies that most copiesof paintings fall far shortof theoriginals.+r To get someunderstanding of the problemspainting presents, con- sidertr','o Roman uall paintingsof Perseusand Andromeda,one from the House of the Priest Amandus at Pompeii and the other from

a5 Plin.Ep 1.28.Translation adapted from Radtce(1963). 46 CompareDion. Hal. Dln. 8, rvho expressesstrikingly similar thoughts about copiesand originals:"a certainsponaneous charm and fieshnessemanates from all theoriginal *Jd.lr, *hereasin theartificial copies, even ifthey attainth€ height of imrtatiie skill, there is presentnevertheless a certainelement of contrivanceand unnaturalnessalso. [t is by this rule that not only oratorsdistinguish other orators, but paintersthe rvorksof Apellesand his imitators,modelers the rvorksof Polycli- Translationadapted from the I'Cl tus,and sculptors the rvorks ofPhidias." ^ a7 l\{irandaMan,in (pers.comm.) suggests that Pliny the Youngeris refeningto Plato'sdenigration ofpainting and sculpture(Resp. 10.598b) rather than aclual copies of copiei. Perry(2005:95), on the other hand.has the samereading of the pas- sageas I do. 210 JOC'EL\NPENNY S\,1AI-L

Boscotrecase(Figure 9).+s 4 glanceis enoughto tell thatthey represent the samesubject lvith the sameelements: Perseus flying in liom the lefl; Andromedamanacled to the clilf jutting up in the centrewilh the ketos on the left and her motherKassiopeia belorv on the right; and finally Perseusrepeated and being receivedby her father Kepheuson the right. The iconographicaldifferences are minor. Kassiopeia,for instance,sits on a separateoutcropping in the one liorn Pompeii,but at the bottomof the samecliff in the Boscotrecasepainting. Despite their iconographica. similarities,their renderingsare strikinglydifferent. The Boscotrecase painting is seenfrom farther away and is rather atmospheric.The ketos standsout in the Boscotrecaseexample, rvhereas the palaceof Kepheus is much clearerin the Pompeianpanel. Finally, of course,as both Pliny the Elder and the Younger rvould have noted, the colours differ. For instance,the Pompeianpanel dependsmore on a bluish-greenfor its background,while the Boscotrecasepainting uses a deepergreen over- laid w'ithmore greys and browns. Which is the original?Should it be the one from Boscotrecasefrom the decidedlyupper class villa of Agrippa Postumus'lRemember. horvever, that accomplishedexecution is not ahvaysthe most reliableguide or all the satyrson the Etruscankylix rvould hal'e been more poorly drarvn than their Attic models. Then. again.another painting. no longerextant, may be the original. What about the other Pompeianfype of Perseusfreeing Anclromeda? A slightly later momentis chosenin a versionfrom the Houseof the Dioscuri (Figure 10).'lop..t.us has releasedAndromeda's right arm frorn its manacleand is helping her step do"vn.I passover the fact thal her 1eftarm, r.vhichPerseus awkwardly supports,remains pinned to the cliff. Neither is looking at the other. This painting focusesless than the othertrvo on piacingthe protagonistsin an overallsetting, and moreon the pair alone. In fact, the only subsidiaryfigure is the dying ketos,as usualon the lower left. Perhapsthis versiongoes back to a Greekorigi- nal, sincethe Greeksnever lost their focuson the main figures.Recall

48 For the'Boscotrecaseversion: Nerv York, MetropolitanMuseum of Art 20.192.16:Anderson (1987188: 53, colour),and LINIC AndromedaI 32 rvith pl. 629.For the Pompeianversion: Kraus and von Matt (1975:186 fig. 250,colour). LIMC AndromedaI 31.Note thatRichardson (2000: 36) anributesboth paintings to the BoscotrecasePainter, but he hasassumed that becausethey showthe sameele- mentsthey must be b;-'thesame hand. Compare my discussronof theiruse of space to portraycontinuous nanative: Small (1999: 568. with 569fig. 8). r'r Pompeii6 9.6-7.Naples 8998: Hl 186.LIMC AndromedaI 69,p.781 rvith pl. 63.1.where it is assignedto theFourth Style. Bergmann (1995: 95-96 and I l3 fig.6, bottomleft). \ ISL \I ( I )PIfS \\D \IENIOR\ 241 tlratPliny the Elder (NttturalHistotl:35.116) said that it was a Roman, Studius. r.r'lro invented landscape painting. Pliny (Natural History 35.132)even offers us a Greekcandidate for the representationofsuch an Andromeda:Nikias of Athens.It rvas among his "large pictures" (grandespicltrro.s) not to speakof the fact that Nikias rvas knorvn for "paint[ing]wolnen most caretullY."50 This Andromedaand Perseusdil'- f-ertoo much from the other two to be their model, however,and so let us eliminateNikias. There is a secondpossibiliry: a Romanpainter took tlie model and betteredit. Quintilian,in the samepassage I quotedabove, directly con- tinues."It is a positivedisgrace to be contentto owe all our achievemenf to imitationlimiterisl. For rvhat,I askagain, would havebeen the result ifno onehad done more than his predecessors?"5iIn otherrvords, Quin- tilian not only believesthat Romanscan do betterthan their predecessors.but also that they shoulddo so. If that is true of rhetoric,rvhy u'ould it not be tnte of painting?Would Roman artistsbe bettering Greek paintings,horver'"er, or rvould they merely be betteringthe works of their peersand their orvn ancestors?I can only partiallyanswer this question.I think that it is highly unlikely that absolutelyno changesor only changesfbr the worse occurredin painting sincethe lburth century BC u'henNikias lived. I think that the Romans'nvo greatestcontribu- tionsto painting*'ere full illusionism,and landscapeswith figuresinte- gratedu'ithin thoselandscapes and not drvarfingthe setting.The Greeks never lost their belief in the idea that "man is the measureof all things.":: The fourth centuryBC painting that lve do have from Vergina emphasizestl.re figures, as with the secondpair of Perseusand Andro- meda,rather than having the landscapedominate the ligures.sl At this point let us expandour discussionto considerthe most fa- mousand complexexample from Pompeii:the AlexanderMosaic from c. 100 BC in the Houseof the Faun (Figure 1l). Without its borderit measuresa little over five metresby nearlythree rnetres.sl With its bor- dersit expandsto nearlysix metresby just over threemetres the size

50 Plin. HN 35.131, my translation. Compare Pliny the Younger's remarks quoted above.Lippold (195l) providesa useful compendiumof"traditional" schol- arlv attributionsof Pompeian paintings to Greek artists.He discussesNikias (93- l0l) and the Perseusand Andromeda(9'1 n'ith 1tg.76). -i1 Quint. Inst. 10.2.1. translationliom the lCZ. 1sted. i: ProtagorasupudPI. Tht. 160d. 5l Andronicos( I 991: 97-I l9 rvrthcolour picturespa.r.sln). 5l The measurementsand statisticsare taken fiorn (N{oreno2001: 1l and l6). ) t) JOCEL\'NPENNY SM.\LL of an averageManhattan living room. It contains over fifty men anc approximatelyrwenty horses.The vaguenessof thesenumbers is due to the incompletepreservation. No doubt exists that the mosaic represents Alexanderand Darius.Which of their fwo major encounters,Issos in 333 BC or Gaugemelain 331 BC, remainsdebated. For my purposes here it doesnot matter.Similarly. at leastseven different Greekpainters have been proposedas the artist.5sRecent attributions have centredon Phiioxenosand Apelles.Pliny the Elder (NantralHistory 35.110)says that Philoxenusnot only was the court painterof Kassandrosin Mace- donia after the death of Alexander, but also the painter of a battie fttroeliumf befiveenAlexander and Darius. Apelles is a possiblecandi- date,first becausePhny (Natural Histo: 35.85) recordsthat Alexander the Great ofien visited his u'orkshop.Second, according to Pliny, Apelles "surpassedall the paintersthat precededand all who were to come after him."soIn addition,scholars tend to assignextant works of high quality to the artistrated most prestigiousin the ancienttexts.s? Pliny describesApelles'style as "unrivalledlbr gracefulcharm. ... he knerv rvhen to take his hand au'ay from a picture . ... he used to ac- knowledgehis inferiorityto Melanthiusin grouping,and to Asclepio- dorus in nicely of measurement."Comments like these are uselessin making attributions,especially for a Roman copy made in anotherme- dium three hundredyears later. Even more important,not only have no paintings by Apelles surr,'ived,but also rve do not have even a single scrap painted b.v any classicalGreek painter mentionedin the extant literarysources.58

s5 Cohen (1997: 138-142). 56 Plin. HN 35.79.Also seeertended discussion in Moreno (2001: 29-18). 57 Compare Ridgrvay (2004: 733). There are a number of resonancesbetween Ridgway's essay on the Laocoon and my treatment of copies, although I read the article after I gave the keynote address but before I prepared it for publication. The origrnal article. "Le Laocoon dans la sculpture helldnistique," appeared in Ddcultot et al. (2003: l3-31). 5E Steuart (1993: 150-157). among others, relates the Alexander l\losaic to Apulian vasesby the Darius Painter (c.330 BC according to Stewart, [50]) that portray encounters of a Greek wanior on a horse pursuing a Persian. In the end, horveler, Stervartconcludes (152-53), "Clearly, the Darius Painter cannot have seen the original of the lvlosaic, a sketch of it, or any other painting of Alexander pro- duced for the Macedonian court. He simply did not know what Alexander looked like. Had he done so, he rvould not have shou'n him bearded.... but in the absence of any precise indication of Alexander's real appearance had to resort to guess rvork." \tlSU.\L C--OPIES.\ND lUEivlORY :+)

Justbecause we do not hayethe original and cannotknow its painter, it may still be possibleto posit ail original Greekpainting that stands behindthe AlexanderMosaic. First, Alexander the Greatby dint of his date obviously datesthe subjectto no earlier than the last third of the fourth century BC. Furthermore,as I indicatedabove, we know of at leastone painting from the tburth centuryBC depictinga battlebetween Alexanderand Darius.Hence the appearanceof Alexanderand Darius in a Romanmosaic fiom c. 100BC is not in andof itselfsurprising. The secondmajor supporttbr a fourth century BC original lies in the palette of the mosaic rvhich follorvs earlier practiceand is limited to four colours-"red," black, yelloiv, and rvhite-though Cohen notes "some greeneiements."se Third, the focusof the sceneis on thefigures, not the setting,w-hich has the proverbialGreek lone, barren tree. The depiction of the spearsextending above the fray reflects the rvay they actually looked in a battle,as I discoveredin an otherwiseforgettable movie about the battle at Marathon.Other elementsare variously interpreted. Some scholarsmaintain that the afrnour is authenticfourth century BC, othersthat it is a mixture of elementsfrom the Hellenisticperiod.e0 Somemistakes are apparenton closeexamination. For example,there are tracesof "a rvhite horse that anatomicallycannot be put together" amongthe tbur blackones on the right quadriga.otThese errors are con- sideredto prove that the mosaic must be a copy, becausesureiy the originalpainting got it right.Again. this argumentdoes not matter,since originals can have mistakes.Even the scholarswho know about and noticethese errors still considerthe mosaican "excellentcopy," as Hcil- scherputs it, of a Greekoriginal painting from the fourth centuryBC'02 At this point rve must considerthe logisticsof making a copy of a paintinglarger than a Persianrug in an averageNew York living room. We should not worry about whether the original was in Pella or even

5e c_-ohen(1gg7. 167-69. esp. 16{l). It is interesting to note that the restriction to fbur colours i; not apparent unless pointed out, in part, I believe, because battles sceneson banen plains are naturally often limited in their palette 60 while most scholars accept the realiu as fourth century BCl, Michael Pfrom- rner (1998) has devoted a monograph to a study ofthe individual elements,espe- "Die.Re- cially of the armour and the dress olthe ligures. He concludes (215) that alien des Mosaikgemiildes entsprechen keinesfalls alexander- oder diadochen- zeitlichen vorgabeir." (Nly emphasis.) He believes (216) that rhe realia indicate that the mosaic datEsmost likely inthe late third or early second century BC' 6r Cohen (1997:19) u'ith othermistakes discussed. 62 Hcilscher (2004: 2l); "an excellent copy ofan important painting dating from " the late fourth century BC soon after the death ofAlexander the Great 211 JOC]ELYNPENNY SNIALL

Alexandria,as otherssuggest, since either locationpresents tlie same problemsfor copying.o:I beginrvith the obviouslimitations: no photographs,no casts,no imposinga grid overthe precious original. A sketch couldbe made.but not to scale,since the availablematerials-papyrus, wax or u'ood tablets,clay slab-were not manufacturedon such grand scales.Even photographingthe rvholemosaic in one shot is not easy, and detaiisblur and sometimesdisappear. We still needdetails. In the absenceof a picfureof the rvhole,joining thesedetails together is also not easy.Consider that the copy being made in Ravennatoday has a lif-e-sizephotograph displayed in the rvorkshopand that thar photograph showsfaint verticallines indicating that it was piecedtogether.6a Next theRavenna mosaicists made: ... a tracingofthe photorvith a darkmarker and covered it rvitha thin layerof tissueto makea negativeimpression. Norv they had their design. ... Insteadof r.rsinga single large u'ooden [frame] ... coveredin limeas the ancientmosaicists might have done. the ltaliansdecided to use44 separateclav frames and u,ork on the mosaic section by section.6-s Eventoday ll'ith all of our technicalequipment it is no simplematter to makea copy of somethingthat size.What is the likelihoodthat sorne 4zl sketches,each totallir accurate, rvere precisely pieced together in antiq- uitv?66

6l For Alexandria: Fehr (1988). Cohen (1997:59) suggeststhar the so-called original paintingu,as "perhaps ... carriedoff to Rome fiom Macedoniaas part of the booty fiorn the battleof P1,dna(168 B.C.)." br N{erola(2006: 36-37). 65 N'lerola(2006: 38). One scholar(Donderer [19901) belier,es that rhe mosaic is an "original" Greek mosaic that tvas removed in sections lrom a Flellenistic palace' Against this idea Dunbabin (1999: 43) argues,"ln rny opinion this rheory remains unlikely in view ofthe size and fragility ofthe rvork and the difficulty that would be involr.ed in dividing it into sections."She then adds, "the mosaic was laid on the spot by a team of craftsmen, who may safely be taken to have been Greek. Every- thing else is uncertain...'One cannot assumeprima.faciethat all good artistsmust be Greek. Even Greek names are no guaranteethat they are not Romanized Greeks, as the variety of names in the United States so eloquently testily. On the mechanics of lay'inga mosaic,see Dunbabin (1999:279-90). 66 There is vtrnrally no extant evidence of actual copybooks. While most scholars uould agree that artists must have shown some kind of designs to prospective clients and that artists must have had accessto designs fbr their own use, nonethe- less, it remains highly speculative what fbrms these designs took. The closest example mav be the sketcheson the reverse of the Artemidorus papyrus, but these com- prise separatedrawings of animals and monsters, as well as details like hands, feet. and heads.Canfora (2007), among others, questions the papyrus' antiquity. Nothing comparable to what would be needed to replicate the Alexander. Mosaic is pre- sened. The best colour pictures appear rn Gallazi and Settis (2006: esp. l,l2-55) VISUALC.OPIES AND N'tElvlOI{\ 245

Next tlie AlexancierMosaic, to statethe obvious' is a mosaic'It is not a pairrting.The closcst altalogy to rrlratit is like to copya paintinginto a mosaic-likeformat is that twentieth-centuryinvention of paintby num- bers.Consider thc cor.erof Esquirewith the portraitof LyndonJohnson by RichardHess liorn June 1967.0;Johnson's head is divided into a numberof irregglarsections that are then numberedwith the colourof the paint to be use

Note that Settis(31) refersto the Alexander l!{osaic as a "riproduzione intenzional- mente f'edele" a judgement similar to that of [{cjlsher. 6l In addition to the actual cover, an online reproduction may be fbund at lbj html] Ihttp:riamericanhistory.si.edurpaintr'lmages'rlarge-lmagesiIMAGE-HTN'lLi ttii image was part of an exhibition (at the Smithsonian Natronal Museum ot AmericariHistoryl on paint by number pictures, for u'hich see Bird (2001) with the Johnsonportrart appearing on p. I 12 5s \{lrola (2006: 38) for both thc size of the tesseraeand their number. 6e Ragghianti (1964: 2a-36) is especiallysalutary to read on the idea of exact cclpics oil-ost Greek paintings. FIe heads one chapter (p. 30) "The Impossibility o1' fvtaking 'Perf-ect'Reconstructlons of the models of 'Classical' Artists " l0 (ome scholars' desire for a Greek original is so great that they make s.me remarkablearguments. For exanlple,Cohen (1997: 52) says,"to arguethat^the sur- ViVing image is solell, a Roman Creation rvould be to fbrestall discussion of its rich fuurtf,-ccnt'u' Greek imagerl. and the histclrical associatio's of this imagery, and contlne oneselfto issuesof reception." 216 JOCELYN PENNY SI\IALL originalfor comparison.At this point,then, I think it is necessaryto pull togetherthe strandsof my discussionto understandwhat the Alexander Mosaicis andhow it works.From the outset,making a copy of a painting presentsmore problems than that of a singlestatue. In thecase of the AlexanderMosaic fwo issuesare paramount:its sizeand the colours.I hope I have estabiishedthat neitherwas likely to be copiedwith any degreeof exactness.Even today rvhena model is right before us, as in the caseof Lucy andReba, we areunable to makean exactreplica without digital assistance.In classicalantiquity, however, one thing was on the sideofthe copyistthat is not the casetoday: very few people.ifany, couldor would checkto seehow rvell thecopy matchedthe original.At the same time, in part becauseof these limitations, their standardsof precisionrvere dif-ferent tiom ours. Considerthe preface to Cicero'sTopics:

on reachingVelia I sa\\'your family[Gaius Trebatius Testa] and your home,I rvasreminded of ... [my]debt [to u'rite a translationof Aristotle's Topicsl.... Therefbre.since I hadno bookswith me,from memory re- calledI rvrotedown these things on the vo)'age itself.Tl Sincer.','e have Aristotle'sTopics, we arein a positionto judge the quality of Cicero'stranslation. Most scholarsfind little in commonbetween the trl'o *'orks exceptthat they both do discussthe use of "topics" "for inventingarguments ... [using]a rationalsystem."72 As the arlistof the AlexanderMosaic changed paint into tesserae,so Cicerochanged Aris- totle'sexamples into legalones u'hich *'ould be more appealingto Tre- batius,a larvyer.Similarll' Trebatius is not likely to compareAristotle's text to Cicero's.but, like many an art historiantoday using Ihe Loeb translationsand not checkingthe Greekor Latin original,not only ivill Trebatiusbe relievedthat he doesnot haveto slog throughAristotle's text, which he found difficult and obscure,but alsohe will believethat he does,indeed, have Aristotle's Topics. In fact, while I havejust used the Loeb translationmyseli so much easierthan making translations from scratch I have actually adaptedthat translationto more literally captureu'hat Cicero says. The translatorsays "I wrote up what I could remember," rvhen Cicero never uses "could"; instead he says corz- scripsi-"1 wrote down" for Cicero believeshe has rememberedeve-

tt Cic. Top. 1.5;translation adapted fiom the lCZ. Seefuller discussionin Small(1997:217-t9). 12 Cic.Top. 1.2. 1/1-7 VISUAL COPIESAND IVIEN{OR\

Like mostRomans, especially rything.He needsno text in front of him. thoser,,,ho ,uvere larvyers, he prided himself on his recall.Because he lived in a *,orld still dominatedby oralityaud not literacy,he felt freeto "copy" s$itch eranrplesto lcgal oncs becausehis standardlor is .'equivalency"not "identity."ln Thucydideanterms the gist is suificient' Hencerve can recognizethe threeLaocoons discussed above as all being in.ritationsof the "real" Laocoon.In the caseof the AlexanderMosaic precisionis neitherpossible nor desired.what we haveis a Romancrea- tion in the spirit of cicero's Topics aftera Greekoriginal-something re-creeteAris- thatQuintilian lould haie approredof. As rvecould not totle's loplcs from Cicero's Topics,So we cannotreconstruct any Greek original fiom the AlexanderMosaic.?3 Nor in truth can we reconstruct any Greekpainting tiom an.vRoman painting or mosaic.TaIn classical antiquitygist ahvaystrumped precision, because even in the rare cases rvhereprecision lvas possibleno one could really check.Orality governs not iustthe u'orld of textsbut alsoof art.

Bibliography

Abbreviations

.Ldtl.: = Carpenter, T. H (1989). Beazlel'Addenda,2nd ed Oxfbrd: Oxford Univ. Press(for The BritishAcademy). 2nd ed Oxford: lRr': - Beazley.J. D. (1963).Attic Red-FigureVase-Painters. Clarendon. B.4D'- BeazleyArchir e Database: ault htm]' Ihttp:iiu u,rv.beazle-v.ox.ac uk/BeazleyAdmirL/Script2/def Harvard Univ. Press LC:L- TheLoeb ClassicalLibrary'. Cambrrdge" lv{A:

1/- il Such a reconstruction \\ ould be analogous to reconstructillg portions of the from the iad, Ocll,ssg1,and the llirtuperris from Vergil's Aeneid The Aeneid srems "an excel- same tradition, but it is an utterly' Roman creation that no one suggestsis lent cop)"' of anY Greek u ork 'a bompare Ragghianti (1961:2i): "lfu:e look_atcopies actually made by artists we (and those irhich Rubens and Cezanne made of Carvaggios' Deposition exrst) quite nex' i,nd ,r'e are laced s'ith another phenomenon. that is that they are really *irf r. :prre of the fuct that they dependon f,n alli5tic P1ec5d':n1:.thcl lre.uork: 'n cll' IS rihere the link rtith a fomr thet rias alreadypersonal and dcfinite' if tt extstsat r,er1:slight,ovenvhelmed and obliterated by other vafrlel, by lnew.l.angua.e^1.S9m9 (82)lhat it tim'esth"e moclel is completelyburied." Bergmann (1995: 8l-83) believes Polykleitos] is a "common illusion t-hat*e can anchor them [also the Doryphoros of She later in nvo distinct cultures (ireece and Rome-and make sense of them." ..1 to repli- (g7) says. would suggestthat the primary aim of most muralists was not the Alex- cot. e*uctlv a Greek oifulnat." She does not. ho*'ever. appearto consider rndcr Mosaic a Rt'tnln crcation. 248 JOCELYNPENNY SNIALL

LL\.[C - Lexicott lconogruphicunt ltlv thologioe C)las s icae ( 198 1 - 1997). IV{unich, Ziirich, and Dtisseldori Artemrs. OED Online: The OxJbrdEnglish Dictiontnt Online. [lvwr.v.oed.com]. Ot.D - TheOrJbrd Latin Dictionary,(1968-1982), P. G. W. Glare,ed. Oxfbrd: Clarendon. Para - Beazley.J. D. (1971). Paralipotnena:Additions to Atic Black-Figure Vase-Paintersand to Attic Red-F'igureVase-Painters,2nd ed. Oxford: Clar- endon.

Addams, Charles(1991). The lI'orld of ('harles Addant.y.Nerv York: Alfied A IQiopf. Albertson,Fred C. (200.1)."The Creationand Dissemination of RomanImperial PortraitTypes: The Caseof N,larcusAurelius Type IV." JdaI l19:259-306. Anderson, Maxrvell L. (1987-1988).Pompeian Frescoes in The Metropolitan IILtseLrnto.f Art. Bulletin of The ,\lettopolitort llusettm ty';lrl 45.3. Nel" York: N'letropolitanMuseum of Art. Andronicos. l\lanolis (.1993).Vergina. The Rotal Tornbsancl the Ancient C'itt,. Athens:Athenon. Bartman.Elizabeth (1999). Portroits o/ Lit,ia: Imctgingthe Imperial Ll'ontanin .4ugustanRone. Cambridge:Cambridge Univ. Press. Bergmann.Bettina (1995). "Greek Masterpiecesand Roman RecreativeFic- tions."HSCP 97. 1() -1 20. Bird. William L. (2001).Puint hl Nunher. New York: PrincetonArchitectural Press. Boardman,John (2006). "Sources and models."ln Gret'k Scu/pture.Function, ,\,faterials,and Techniquesin the Arc.haicancl Classical Periods. Olga pala- gia,ed.: l-3L Cambridge:Cambridge Unir'. Press. Borbein.Adolf H. ( 1996)."Polykleitos." In PersonctlStyles in GreekSculpture. Olga Palagiaand J. J. Pollitt. eds.:66-90. Cambridge:Cambridge Univ. Press. Brilliant, Richard (2000).,{1r Luococjn; Alternative Claims in the Interpretation ofArnrorks.Berkeley and London: Univ. ofCaliforniaPress. Canfora.Lucrano (2007). The True Histon' ol'the So-calledArtentidonts papv-- nrs.Bari: Edizionidi Pagina. Cennini,Cannino d'Andrea (1954). The Craft.sntan'sHandboctk. The ltalian "It Libro dell' Arte," Iransl.Daniel V. Thompson,Jr. New York: Dover. Cohen. Ada (1991). The Alerander Mosctic.Stories of Victory and Defeol Cambridge:Cambridge Univ. Press. Cohen.Da[e (2005)."Look Little, Look Ofien: The Influenceof Gaze Fre- quency on Drarving Accuracy." Perception und Psychophl,sics6l: 997- I 009. Connor.Peter J. (1981)."Replicas in Greek Vase-Painting:The Work of the Painterof Louire F6." BABesch56'.37-44. Davis. Whitney lvl. (1981)."Egypt. Samos.and the Archaic Stvle in Greek Sculpture."JEA 61: 6l-81. VISL].\LC'OPIES AND NIE\10RY 249

Ddcultot,E.. J. Le Rider.and F. Queyrel.eds. (2003). Le Laocoon:histoire et re(:eption.Revue Genttatritlue Inte rnotionole 19 (SpecialIssue). Didierjean. Andre, Vincent Ferrari, and Evelyne Marmeche (2004) "L'expertise cognitile au jeu d'dchecs:quoi de neuf depuis de Groot (19.16)1"L'annte pstchologique 101 77l-93. Donderer, Nlichael (1990). "Das pompejanischesAlexandermosaik Ein ostlichesImportstuck')" ln Das cntike Rom tutd den Osten; Festschrift.ftr Klaus Porlasca zutn 65. Gebtolstag, Christoph Bdrker and Michael Don- derer.eds.: l9-31. Erlangen: Universitiitsbibliothek Erlangen. Dunbabin, Katherine lvl. D. (1999). ivlosaicsof rhe Greek ontl Roman Ll/orld. Cambridge:Cambridge Unir'. Press. Fehr, Burkhard(1988). "Zwei Lesungendes Alexandermosaiks."In Bathron.' Ileitrtige zur .lrchitektur untl venrortdlen K[insten .fir Heinrich Drerup ztt seinem 80. Geburt.\togtctn seinenSchiilern und Frettnden.H. Biising and F Hiller,eds.. l2l-3-1. Saarbriicken: Saarbriicker Druckerei und Verlag. Fullerton.N1ark (2006). "(Re)presenting the Roman 'Clultureof Statues'."JR4 l9:479-83. Gallazzi.Claudio and Sah'atoreSettis (2006). Le tre vite del Papiro di Artemi- doro. L'ocie sguardi dttll'Egitto greco-romono.Milan: Electa. Gazda,Elaine (1995). "Roman Sculptureand the Ethosof Emulation:Recon- sideringRepetition." HSCP 97: 121-56. Gazda.Elaine. ed. (2002). The .lncient Art o.f'Emulatictn.Sntdies in Artistic Originalitl ond Traditionfi'om the Pre,sentto ClassicalAntiquilj'. Ann Ar- bor: Unir'.of NlichiganPress. Hallett,Christopher H. (2005)."Emulation versus Replication: Redefining Ro- manCopying." JRA 18:419-35. Holscher.Tonio (2004). TheLangttage of Imagesin RomanArt. rransl.Anthony Snodgrassand AnnemarieKiinzl-Snodgrass. Cambridge: Cambridge Univ. Press. Isager,Jacob (1991). Pliu,on Art and Society: the Elder Plin;"s Chapterson tlte Historl' o/lrr. Odense: OdenseUnir'. Press. Kraus. Theodor and Leonardvon Matt (1975').I'ompeii and Herculaneutn:the Liting Cities o/ the Dead. transl.Robert Erich Wolf. New York: Abrams. Lippold.Georg (195 1). Antike Gemcildekopien.Munich: Bayerischen Akademie der Wissenschaften. N{ankoff.Robert. ed (2004). The CompleteCortoon,\ of the Nev',Yorker.Nev' York: Black Dog & Leventhal. Nlartelli. lvlarrna.ed. (1987). La ceramica degli etrttschi: la pittura va,scolare. Novara:Istrtuto Geografico de Agostini. lr,lassrroni. Nlanliedo (2002). The Psycholog.t' o.f Graphic Images. Seeing, Drtntittg, (-omntuticLttins,transl. Nicola Bruno. MahrvahNJ: Lawrence Erlbaum. Nlattusch,Carol C. (1996). Clluslcal Bronzes;the Art and Craft o.fGreek and RotrrtnStatuon. Ithaca and London:Cornell Univ. Press. N,lerola,Ivlarco (2006). "Alexander Piece by Piece.''Archaeolog 59.1:36-u10. 250 JOCELYNPENNY SMALL

Nloreno, Paolo (2001). Apelles: the Alexander lt'Iosaic.transl. David Stanton New York: Skira. Muller, JeflreyM. (1989)."N{easures of Authenticity:The Detectionof copies in the Early Literatureon Connoisseurship."In Retaining the Originol. Mul- tiple Originals, Copies,and Reproduclions,Kathleen Preciado,ed. Studies in the Historyof Art, Vol. 20: l4l-49. Washington:National Gallery of Art, Hanover,NH. Nickerson,Ra-vmond S. andMarilyn JagerAdams (1979). "Long-term Memory for a Common Object." CognitivePsychologt ll:287-307. Norman, Donald A. (1988). The Psychologyof Every:daylftlngs. New York BasicBooks. Perry, Ellen (2005). The AestheticsoJ Emulation in the VisualArts of Ancienl Rome.Cambridge: Cambridge Univ. Press. Pfrommer,Michael (1998). Untersuchungenzur Chronologie und Komposition des .4lerundermosaiksauJ antiquari,scher Grundlage. Marnz am Rhein: von Zabem. Platt,Veri5" (2006)."Making an Impression:Replication and the Ontologyof the Graeco-RomanSeal Stone." In Trimbleand Elsner (2006): 233-257. Pollitt, J. J. (1974). The Ancient Viett' of Greek Art: Criticism, Histor,r, and Terntinolog,r-.Nerv Haven and London: Yale Univ. Press. Radice,Betty. transl. (1963'1The Letters of the YoungerPliny. Bahimore: Pen- guin. Ragghianti,Carlo Ludoi'ico (1964). The PaintersoJ-Pompeii, transl. Shirley Bridges.Milan: Edizionidel MiLone. Richardson.Lawrence Jr. (2000).A Catalog of ldentifable Figure Painters oJ Ancient Pompeii, Herculaneum,and Stabiae.Baltimore and London: Johns HopkinsUniv. Press. Ridgway, Brunilde Sismondo(2004). Second Chance: Greek Sculptural Stuclies Revisited. London: Pindar. Robins, Gay ( 1997).The Art of Ancient Egypt. Cambridge,MA: Harvard Univ. Press. Scheller,Robert W. (1995).Exemplum. Model-Book Dra*-ings and the Practice of Artistic Transmission in the Middles Ages (ca. 900 ca. 1170), transl. Michael Hoyle. Amsterdam:Amsterdam Univ. Press. Schmidt,Erika E. (1973').Die Kopien der Erechtheionkoren.Antike Plastik 13. Berlin:Gebr. Mann. Small, Jocelyn Penny (1997). Iltax Tabletsof the Mind: Cognitive StudiesoJ trIemory and Literacy in Classical Antiquity. London and New York Routledgd. Small, JocelynPenny (1999). "Time in Space:Narrative in ClassicalArt." ABull 81:562-515. Small, Jocelyn Penny (2003). The Parallel Worlds of Classical Art and Text. Cambridge:Cambridge Univ. Press. Sntithsonian(October 2005). "This SideofParodies": I 16. Stewart,Andrerv (1990). Greek Sculpture.An Exploration. 2 vols. Nerv Iiaven andLondon: Yale Univ. Press. VISTJALC'OPIES AND NIE\lOR\ 251

Stenart. Anclreu(1993 1. l'ace.r o.f Pott'er: Alerontler'.s Intage ond Hellenistic I)olitic,s.Berkeley and Los Angeles:Univ. of CalitbrniaPrcss. Ste\\'art.Peter (2003). Stotue.s irt Ronon Societl'.Representation and Response. Oxford:Orfbrd Unir'.Press. Trimble. Jennrferand Jas Elsner,eds. (2006)."Art and Replication:Greece, Romeand Beyond." .7rt LIi.story 29.2: 201-312. Varner,Eric R. (2006).Rer iew of Perrl'Q005). A.IA 1l0: 678-79. Wright. David H. (19931.The l'otican L'ergil. a trIusterpieceof Late Antique .lrt. Berkelevand Oxford: Unir'. of CalifomiaPress. ..,"=1W,

Figure5. "Laocoon."Charles Addams' lVetr Yorker (April 11, 1915)' O Teeand CharlesAddams Foundation.

Photogranh: Archive Timothv Figure 7. Satyrs:Attic red-figurekylix, OedipusPainter, c.450 BC. Vatican 16541.Photograph: Alinari I Art Resource,NY ART319401.

Figure 8. Satyrs:Etruscan red-figure kylix, c.425-400BC. Paris, Rodin Museum Co. 1387(980). Photograph: Luc et Iala Joubert/ Collection d'Antiques du Musde Rodin, Paris. Figure 9. Perseusand Andromeda:wall painting from Boscotre- case.Photograph: New York, Metropolitan Museum of Art 20.192.17.Rosers Fund. 1920. Figgre 10. Perseusand Androntecla:u'all painting liont Pontpeii. Holtsc ol-the I)ioscLrri.Naples llc)98;H I l8(r. Photograph'Erich Lcssing / Art ILcsortrcc'NY .\ Itt-6l.llt8. .:.;*' ',siT ,.,9G Wrz""t-,\i\o I ;i..:a: .,f : i'".''l ;;.1' .2 ! ---:h- € 7/./ s ;:i 'a '.:j:at .,s ::"-? = .i! ,11ri;r, ,.1 t.& a\ 1'g,., .. .:) ' .Y{.;i '41 . 6*\. -.Ai$. llt -::-l -.n 'rr.-\,iii '' ffi ci-s'; . s4'.5 I : 1:-:\ #; .t!, 1t,, t*+ t;t ;4 - H: ,,:wl-. € Wr'.t5.:.*' ;.' rin"t3 'rN..tr+,,.:.. !E i'' - 1 .W ,, o& ,:.r; 5'*-, o -'1, ' .iF- ,--. &:- i si; ,, .Ni$* F, , ,,",. -r;-J-.{,- fu'' , i- t :i';:i( .{ 1!) -.:'- ffi t t{ & ;: F