epigraphic and archaeological evidence for Christian reidentification of statuary

This article examines literary, epigraphic and archaeological evidence for the Christian reidentification of statuary and reliefs as biblical scenes and protagonists, saints and angels. It argues that Christian identifications were promulgated, amongst others by local bishops, to make sense of imagery of which the original identity had been lost and/or was no longer meaningful. Three conditions for a new identification are discussed: the absence of an epigraphic label, geographical and/or chronological distance separating the statue from its original context of display, and the presence of a specific attribute or characteristic that could become the prompt for reidentification. In their manipulation and modernization of older statuary, Christians showed a much greater appreciation of the statuary medium than generally assumed. Adresse: Dr. Ine Jacobs, The Ioannou Centre for Classical and Byzantine Studies, Oxford University, 66 St. Giles’, Oxford OX1 3LU, United Kingdom; ine.jacobs@classics.ox.ac.uk The statues on display in the Constantinopolitan cityscape after Antiquity came to be surrounded by stories, urban legends and folklore. Moreover, the Parastaseis syntomoi chronikai and the ensuing Patria Konstantinupoleos make it clear that some of these statues were identified in interesting manners by the eighth and late tenth century respectively.1 In addition to or instead of the gods, heroes, Versions of this paper were presented at the University of Warsaw, the University of Birmingham, the University of Oxford and the Universität Bern. I am grateful to Hugh Jeffery, Panayiotis Panayides, Efthymios Rizos, Grace Stafford and Mary Whitby for their feedback on an earlier draft. I would like to thank in particular Ben Anderson for his insightful comments, as well as the anonymous reviewers. All errors are my own.  The reception of statuary in both collections has been studied in manifold publications, including but not limited to: C. Mango, Antique statuary and the byzantine beholder. DOP  (), –; G. Dagron, Constantinople imaginaire. Études sur le recueil des “Patria”. Paris ; S. Bassett, The urban image of late antique Constantinople. Cambridge ; B. Anderson, Classified knowledge: the epistemology of statuary in the Parastaseis syntomoi DOI 10.1515/bz-2020-0035 BZ 2020; 113(3): 789–836

or mythological personae that they initiallyw erei ntended to depict,contemporaries also sawthem as Byzantine emperors and Christian protagonists. Modern research has often been successful in finding out the originallyintended identity of these and other statues, despite the occasionally 'misguided opinions' of younger eras.The originallyi ntended viewing of as tatue remains the focus of attention especiallyf or archaeologists, despite the growingb odyo fr esearch on alterations, reinterpretations and repurposing.² Attempts to trace, let alone engagewith, alternative identificationsorradicallynew takes arising in late antique and Byzantine centuries, have been few and are largely limited to Constantinople.³ This article focusses on statues thatweregiven identificationsentirely different to the ones they originallyh ad, and on the ways in which such new identifications came about.Although the complicated phenomenonofstatuary reception throughout time cannot be captured in monocausal explanations, Ia im to demonstrate that reidentification was common and,e venw hen explicit literary or epigraphic sources are not available, cannot be neglected if we want to understand the importance of statues in the centuries after theirp roduction. Iw ill focus in particularo ns tatues and reliefs that werer eidentified as Christian scenes and protagonists, saints, biblical personae and angels, since the misconception that such three-dimensional imagery did not exist is ap ersistent one.⁴ Usingacombination of literary,epigraphic and material sources, Iwill contend that allocations of alternative Christian identitiesw erearegular occurrence alreadyi nL ate Antiquity for which therei se vidence also outside of Constantinople. After brief overviews on newlyc arved Christian statues and reliefsa nd the overall presenceofstatuary in Late Antiquity,Iwill focus on the earliest phase of Christian reidentification in the ageofConstantine. Subsequent sections will explore various elements thatl ed to an ew identity beingassigned and made permanent.Iwill arguethat reidentifications, Christian or otherwise, should not be considered as cases of mistakeni dentity,b ut as positive and often conscious choices to make sense of imagery of which the original intention had been lost and/or was no longer meaningful.⁵ Iw ill suggest thatn ew identitiesw ere purposelyassigned by specific individuals in society.Although physical remains will occur occasionallyinthese earliest sections, the focus will eventuallyshift to material evidence and archaeology.S tarting with ac ase-studyofaradicallydifferent Christian identity being assigned to images of Ares and Athena at Sagalassos, Iwill examine the characteristics of statuary for which new Christian identifications could help explain whyt hey are preserved up until today.

Christiansa nd statuary
Christians apparentlyw eren ot taken by the medium of statuary.P rotests first pertained to idols, but later extended to other pagan-mythological and also honorific statues.⁶ Christian suspicion towards and the resulting reluctancet oc on- Such reidentification has in the past been indicated by the term interpretatio christiana. Due to the problematic natureofthe term and the manywaysithas been applied by past scholars (D. Kinney,Interpretatio christiana, in B. Harvey,Jr. /C.Conybeare (eds.), Maxima debetur magistro reverentia: Essays on Rome and the Roman tradition in honor of Russell T. Scott. Biblioteca di Athenaeum, .Como ,  -), Ip refer not to use it.  Reidentifications,whereby an entirelyn ew identity was assigned, therefore differ from reinterpretations, wherebyt he impact and interpretation of ac ertain motif shiftedb ut was not entirely changed.  Fora no verview of the development of negative Christian attitudes in Late Antiquity,s ee I. Jacobs,F rom production to destruction?P agan statuary in latea ntique Asia Minor. AJA  (),  -;B .C aseau,R eligious intolerance and pagan statuary,i nL .L avan /M. Mulryan (eds.), The archaeology of latea ntique 'paganism'. Late Antique Archaeology, .L eiden/Boston ,  -.R area ttempts to continue the statuary habiti naChristian guise arediscussed in B. Ward-Perkins,F our bases fromStratonikeia: a(failed) attempt to Christian-I. Jacobs, Old statues,n ew meanings tinue statuary have been debated on several occasions.⁷ Stand-alone statues supposedlyi nvited veneration and weret hereforet oo reminiscent of pagan idolatry.⁸ Michael Peppardh as explained the averseness to statues of Christ by stressingt hat Christians would have considered them too similar to statues of the emperor,t he worship of which they had so strongly protested, whilst also being aware of the vulnerability of statues,⁹ whereas Paolo Liverani has argued that two-dimensionalp ortraits allowed for the development of ac ode-language that conveyedp articulard etails about the honorand.¹⁰ Regardlesso ft he exact reason why, the established list of purpose-made freestandings tone and metal statues depicting Christian subjects is short,w hereas also relief decoration had clearlybecome far less popularthan in preceding centuries. It is worthwhile brieflyenumerating the known examples and categories of evidence in order to demonstrate how scarcet raditionalt hree-dimensional imagery seemingly became especiallyo nce the fourth century was over.¹¹ In addition to general references to heretics worshipping statues of Christ,¹² literarysources mention specific examples of Christian statuary as well. Thus the ize the statue habit,inA.Busine (ed.), Religious practices and Christianization of the late antique city (th-th cent.). Leiden ,  - and A. Avdokhin,C hristianizings tatues unawares?I mperial imagery and New Testamentp hrasingi nalate antique honorific inscription (IEph .). ZPE  ()  -. Historia Augusta claims thatthe emperor Severus Alexander included astatue of Christ in his lararium.¹³ Farmore famous werethe donations by Constantine, includingt he 5f th ighs ilvers tatuary groups of Christ and the Apostles and of Christ and four angels thatd ecorated the Fastigium of the Lateranb asilica.¹⁴ Constantine also bestowed ag oldl amb,s ilvers tatues of Christ and John the Baptist as well as statues of deer to the Lateran baptistery.¹⁵ Accordingt oE usebius the emperor moreover decorated fountains in Constantinople with Good Shepherds, as well as ag old-plated brass statueo fD aniel and lions.¹⁶ Material evidence consists of some thirty under-life-sized third-a nd fourth-century statues of shepherds that mayo rm ay not have been identifieda st he Good Shepherd.¹⁷ Almost all of them originated as table supports, but some were found in funerary contexts, whereas others show signs of having been reused as fountain figures.¹⁸ Other examples include the third-centuryG ood Shepherd icons (as footnote  above),  and eadem,P anel paintings and earlyC hristian icons,i nJ ensen/Ellison, Routledge Handbook (as footnote  above),  -,o n.M arsengill interprets the images mentioned as panel paintings,though the additionof"formed of other materials" could indicatet hat three-dimensional portraits existed as well. Both Irenaeus' and Epiphanius' passagesa re discussed brieflyi nP eppard,P resence (as footnote  above), .  Historia Augusta, Severus Alexander . -.See Peppard,Presence (as footnote  above),  -.Marsengill,Portraits and icons (as footnote  above),  againprefers to see these as two-dimensional depictions.  Liber Pontificalis . -;B.Ward-Perkins,The end of the statue habit,AD -,in R. R. R. Smith /B.Ward-Perkins (eds.), The last statues of Antiquity.Oxford ,  -,LSA- and ;Peppard,Presence(as footnote  above),  - and Liverani,Sunset (as footnote  above),  - both give further references and comments on the reception of such Christian statuary in ecclesiastical contexts.  Marsengill,Christian reception (as footnote  above) ,lists two possible statuettes of St. Peter,one possiblyofSt. Paul, and another possible apostle, all from the fourth or fifth century, I. Jacobs, Old statues,n ew meanings two sites had aproduction line of ceramic statuettes continuinginthe sixth century:t he pottery kilns of Sagalassos in SW Turkey produced bothz oomorphic and anthropomorphic small-scalef igures.S urviving examples mainly consist of armed riders, some of them marked with crosses, but there is also amusician and ab ishop performing an act of consecration.³⁴ At the pilgrimages iteo fA bu Mina, rider figurines weremanufactured, though the main product line consisted of women shownp regnant or holding children.³⁵ In view of the apparent scarcity of material evidence,itissomewhat surprising to find the matter of Christian statuarye xplicitlya ddressed in letters by highlyp laced men of the church who take au nexpectedlyp ositive stance towards the medium.H ypatius,a rchbishop of Ephesus between 531 and 538 as well as aclose advisor to the emperor Justinian, condones the usageofChristian statues for pedagogical reasons in his response to objections against statuary veneration raised by his suffragan, the bishop of Adramyttium.³⁶ Hypatius did not see anyh arm in using religious sculpture,a sh ec onsidered the danger of idolatry to be at hing of the past.³⁷ Theses entimentsa re repeated in letters sent by Pope Gregory the Great in July 599 and October 600 to Serenus,b ishop of Marseille, who destroyed images of saints present in the churches of his see.³⁸ Although it is impossible to examine what exactlyt hese statues depicted, how they werebeing used, or even how many therewere, these letters again confirm that three-dimensionalC hristian statues still existed as late as c. 600.
It is entirelyunclear if the statues mentioned by Hypatius and Gregory were newlycarved for Christian usageorifthey were older items that had been repuras wellasaseventh-century statuette of apilgrim or saint.Two of these arediscussed in Hornik,F reestandings culpture (as footnote  above),  - as well.  P. Talloen,From pagan to Christian:religious iconographyinmaterial culture from Sagalassos,i nL avan/Mulryan, Archaeology (as footnote  above),  -,  -;I .J acobs /M. Waelkens, 'Christians do not differ from other people'.The down-to-earth religious stanceo f latea ntique Sagalassos (Pisidia), in W. Amelung (ed.), Die ChristianisierungK leinasiens in der Spätantike. Asia minor Studien, .B onn ,  -,s ee  -. posed. Both scenarios are imaginable. There certainlystill was plentyofstatuary around that could have served this new purpose.³⁹ Especiallyinthe cities of the Eastern Mediterranean, the advanceofChristianityhad not led to the eradication of older statues. Honorific as well as pagan-mythological statues are regularly uncovered during archaeological excavations of urban centres that wereo nly (largely)a bandoned at the end of antiquity or even later.W et hereforec annot denyt hat,despite the manyv oices raised in protest against the medium, statuary remained acomponent of the late antiqueand sometimes alsothe Byzantine cityscape. We also must acknowledge that contemporaries chose to preserveand use at least certain images. In some cases, therei sc lear evidence for conscious selection. The fact thats pecific reasons for such selections are not explicitly mentioned in literary sources should not stop us from examining how these decisions werem otivated.⁴⁰ Older interpretationso fs urviving material evidence wereo ften negative,⁴¹ with scholars maybe affectedb yE usebius' assertion thatC onstantine collected statues in his capital to be derided.⁴² In the meantime, interest and also evidence for morepositive takesonstatuary in Late Antiquity has been growing, with stat-  Vita Constantini ..T his and similar claims ared iscussed in H. Saradi-Mendelovici, Christian attitudes toward pagan monuments in lateantiquity and their legacy in later Byzantine centuries. DOP  (),  -,s ee .I ti sc ontested in Mango,A ntique statuary (as footnote  above),  -;C.Lepelley,Lemusée des statues divines.Lavolontédesauvegarder le patrimoine artistique païen àl ' époque théodosienne. Cahiers Archéologiques  (),  -, here ;Bassett, Urban image( as footnote  above), ;J acobs,F rom production to destruction?( as footnote  above), .
I. Jacobs, Old statues,n ew meanings ues being moved around for what have been called 'beautificationr easons',⁴³ and 'antiquarianism',⁴⁴ used as political talismans and meanso fp ropaganda,⁴⁵ or as signifiers of specific honourable qualities and pleasant aspects of late antique life.⁴⁶ Both Christians and pagans appear to have enthusiasticallyp articipated in such reinterpretations,especiallyinthe fourth and earlyfifth century.⁴⁷ Over the course of the sixth centurya tt he latest,s uch usagesa re assumedt o have lessened, and contemporaries supposedlyl ost interest in the medium of statuary and at best passively let them be.⁴⁸ An ever greater number of statues disappeared as they werer eused as buildingm aterial, melted down, or burned to lime for economic reasons.N evertheless,r esults of recent research in cities From use to refuse, in Kristensen/Stirling,A fterlife (as footnote  above),  -,s ee  -.  Ibid., : "an unquantifiableamount of statuary remained passively or neutrallyinplacein cities and landscapes." such as Miletus, Aphrodisias, and Sagalassos would suggest that the role of statuary was not playedo ut after all. Statuary relocations at these sites have been (re)dated with certainty to the sixth century,due to better and more careful excavations and reinvestigations of older research, with meticulous recording of find contexts.⁴⁹ Of course, the further forward in time we push the active role of statues, the more pressingitbecomes to explain whysome surviveduntil the end of antiquity and even beyond.

An earlys tart in the Constantinian period
The earliest literary attestations of Christian refashioningofidentitiesdate to the reign of Constantine. The most informativea uthor is Eusebius.⁵⁰ In al etter to Constantia, which is best known for discussing EarlyC hristian reception of images of the divine,⁵¹ Eusebius mentions what he esteems to be rather confounding identifications: Once -Idonot know how -awoman broughtmeinher hands apictureoftwo men in the guise of philosophers and let fall the statement that they werePauland the Saviour -Ihave no means of sayingw here she had had this fromo rl earned such at hing.⁵² Although theirp ortable depictions are rarelyp reserved today, philosophersa nd portraits of individuals in the guise of philosophersare known to have been popular in second-and third-centuryfunerary art.Onsarcophagi, philosophers and other men of learning are usually depicted in profile, though sometimes frontally, dressed in tunic and mantle either with partiallyn ude torsoa nd barefoot  R.R. R. Smith,Statue life in the Hadrianicbaths at Aphrodisias,AD -.Local context and historical meaning, in Bauer/Witschel, Statuen (as footnote  above),  -;I . J acobs /L.S tirling,R e-usingt he gods. A th-c. statuary displayatS agalassos and ar e-evaluation of pagan-mythological statuary in EarlyByzantine civic space. JRA  (),  -;I. Jacobs,P agan-mythological statuary in sixth-century AsiaM inor,i nI .J acobs/H. Elton (eds.) AsiaM inor in the long sixth century.C urrent research and futurep erspectives. Oxford ,  -.  See also the discussion in Marsengill,P anel paintings and earlyC hristian icons (as footnote  above),  -.  Translated in C. Mango,T he art of the Byzantine empire  -:s ources and documents. Sources and Documents in the Historyo fA rt. Prentice Hall ,  -;S .G ero,The true imageo fC hrist: Eusebius' Letter to Constantia reconsidered. TheJ ournal of Theological Studies  (),  -;J ensen,F ace to face (as footnote  above),  - for ac omprehensive discussion. The authenticity of the letter is not undisputed.  Eusebius, Ep.C onst. (transl. Mango,A rt,a sf ootnote  above, ).
I. Jacobs, Old statues, new meanings or entirelyc overed and wearings andals, but invariablyr eadingf rom as croll.⁵³ The deceased as philosopher could be surrounded by other philosophers. They signified areflective and peaceful life to non-Christians and Christian alike.⁵⁴ Individuals could have their own features added to the general type,thus personalizing the message. At the same time, one particularC hristian viewing of the seated philosopher identified him as Christ the Teacher.A round the year 300 afrontal-facing philosopher type appeared, still in the same garb but now holding ascroll and making agesture of speech, which was almost certainlyintended to depict Christ (Fig. 1).⁵⁵  Manyofthe earliest Christian motifs are reinterpretations and adaptations of such Graeco-Roman prototypes, with the Good Shepherd being the best-known example.⁵⁶ They are probablyi ntentionallya mbiguous or polysemic. Their perceivedi dentity dependedo nt heir association with other figures and scenes and theirp hysical context,a sw ell as the religious conviction of the viewer. Against the background of astill crystallizingChristian iconographyand the omnipresenceofpolysemic imagery,⁵⁷ it is hardlysurprising that aChristian woman in the earlyfourth century sawChrist and Paul when Eusebius still sawmore traditional philosopher types.
Eusebius' discomfort about depicting the divine expressed in his letter to Constantia was obviouslyn ot shared by his contemporaries, nor is it repeated in his other works.A sa lreadym entioned, Eusebius describes Christian statues dedicated by Constantine, including Good Shepherds, in his Life of Constantine. As just indicated, their more general iconography mayhavebeen givenaspecif-icallyC hristian interpretation by Eusebius. In his Ecclesiastical History (7.18), paintingso fP eter and Paul are mentioned in am uch more matter of fact manner,a nd Eusebius seems more prepared to accept and endorse his contemporaries' view on imagery in general, and statuary in particular. Eusebius here tells about astatue on displayinCaesarea Philippi (Paneas).⁵⁸ Icite the text in full, as it illuminatess ome of the processes involved in identifying as tatue.
But since Ih avec omet om ention this city,Ido not think it right to omit as tory that is worthyt ob er ecorded also for those that come after us.F or they sayt hat she whoh ad an issue of blood, and who, as we learn from the sacredG ospels, found at the hands of our  Fordifficulties in distinguishingthe Good Shepherdf romamoretraditional kriophoros,see Studer-Karlen,Verstorbenendarstellungen( as footnote  above),  -;F euser,M onopodia (as footnote  above),  -;L idova,R ise (as footnote  above),  -.The influenceo fc lassical art on Christian iconographyi sd iscussed in K. Weitzmann,T he survival of mythological representations in earlyChristian and Byzantine art and their impact on Christian iconography. DOP  (),  +  -;G .M. A. Hanfmann,The continuity of classical art: culture, myth and faith, in K. Weitzmann (ed.), Ageofspirituality.Asymposium.New York ,  -;A .Grabar,Christian iconography. As tudyo fi ts origins. Princeton ,  -.S ee also J. Huskinson,Some pagan mythological figures and their significanceinearlyChristian art. Paperso ft he BritishS choola tR ome  (),  - for associationsb etween Orpheus,S ol and Ulysses with Christ mainlyi nf unerary art up to the mid-fourth century.  Fora no verview,s ee Grabar,Christian iconography( as footnote  above).

I. Jacobs, Old statues, new meanings
Saviour relief fromher affliction, came from this place, and that her house was pointed out in the city,a nd that marvellous memorials of the good deed, which the Saviour wrought upon her,still remained. For[they said] that therestood on alofty stone at the gates of her house ab razen figure in relief of aw oman, bendingo nh er knee and stretchingf orth her hands like as uppliant,while oppositet ot his there was another of the same material, an upright figure of aman, clothed in comelyfashion in adouble cloak and stretchingout his hand to the woman; at his feet on the monument itself as tranges pecies of herb was growing,w hich climbed up to the border of the double cloak of brass,a nd acted as an antidotet oa ll kinds of diseases.This statue, they said, boret he likeness of Jesus.A nd it was in existenceeventoour day, so that we sawitwith our own eyes when we stayedinthe city.And thereisnothingwonderful in the fact that those heathen, wholong agohad good deeds done to them by our Saviour,s hould have made these objects, sincew es aw the likenesses of His apostles also, of Paul and Peter,and indeed of Christ Himself, preserved in pictures painted in colours.And this is what we should expect,for the ancients werewont, according to their pagan habit,t oh onour them as saviours,w ithout reservation, in this fashion.⁵⁹ Eusebiushimself does not appear entirely convinced of the identity of the statue, as suggested by the addition of the narrative voice. However,h es uspects that such identifications are there to staya nd thus feels it is his duty to spread the story.I nt hat regard,i ti sq uite interesting that he concludes his account with ar eassurance thats uch statues of Christ had been made in the past.C reating them mayn ot have been en vogue among Christians, it wast he common thing to do for pagans. Thus, even though Eusebiusm ay have had doubts himself, at the same time he participates in the promulgation of the Christian identification and lends it additional weightb yoffering an explanatory foundation story.
Philostorgius, in his own account written roughlyacentury later,a dds that the statue was rediscovered by the Christians of CaesareaPhilippi, after aperiod of neglect during which the identity of the statue, the reasons for its dedication and its significance had all been forgotten.⁶⁰ The success of the statue'sn ew identity is confirmed by ag rowinga mount of pilgrims but especiallyb yt he actions of the local pagan community who, under the reign of Julian, attack the statue, tear it from its base and drag it by its feet through the streets of the city -much likep agan statues are draggedb yC hristians -leading to its destruction.⁶¹ The headwas salvagedand givenaplace in one of the local basilicas, wherei ta gain became an object of veneration.⁶² This identification of an older statuary group as Christ and the Haimorrhousa has been categorized in modern scholarlyl iterature as am isunderstanding and is considered to be misguided.⁶³ Yet, such qualifications do not do justice to the Christians of Caesarea Philippi nor to otherg roups of Christian believers who sawtheir heroes and champions eternalized in the statues surviving in their cityscape. It would help our perception greatlyi fw ereferred to this and similar phenomena in amoreneutral manner,stating that (some members of)the Christian population of Caesarea Philippii nterpreted as tatuei nawayt hat was aligned with theirw orldview.⁶⁴ By the time Eusebius reported on the Christ statue at Caesarea Philippi, the allocation of new identities was widespread, includinginthe new imperial capital. John Ma has recentlya ssembled examples of statues that wereg iven new identitiesi nt he Constantinian ageo nce they had been brought togetherf rom all over the empire and givenanew home at Constantinople.⁶⁵ As part of this more general repurposing and reimagining of older statuary, some items were indeed givenaChristian identity.C onsidering that Constantine ordered the creation of Christian statues in Rome, it is possible that the emperor also partook in the naming practice in Constantinople. He is certainlycredited with assigning aC hristian identityt oa no lder statue by John Malalas in the sixth century.M alalas recounts how the emperor,when visiting the temple of an unnamed deity near Byzantium, "recognized" the monumental statue inside and "said that it represented an angel in the clothing of am onk of the Christian faith." The pre- ForChristians draggingstatues through the streets, see P. S tewart,Statues in Roman society.R epresentationa nd response. Oxford/New York ,  -.  Wilson,Caesarea Philippi (as footnote  above),  -.The topographer Theodosius, writingb etween  and ,s till mentions that statue but it remains unclear if he himself went to Paneas.S ee Y. Tsafrir,T he maps used by Theodosius:o nt he pilgrim maps of the Holy Land and Jerusalem in the sixth century C.E. DOP  (),  -;Wilson,Caesarea Philippi (as footnote  above), .  Fori nstance, P. S tewart,C ontinuity and tradition in late antique perceptions of portrait statuary,i nB auer/Witschel, Statuen (as footnote  above),  -,h ere ,a rgues that "We need not be surprised if Eusebius, or his contemporaries, have misunderstood these old statues." Liverani,S unset (as footnote  above), ,n ote  likewise uses "misunderstanding" for Byzantine allocations of Christian identities.  Rous,Reset (as footnote  above),  - suggests the term "upcycling" for intentional mean- cise identity of the imagea st he archangelM ichael was apparentlyr evealedt o him in ad ream.⁶⁶ Likewise, the case alreadym entioned above, in which Eusebius accredited Constantine with using images of the Good Shepherd to decorate fountains, in all likelihood was aC hristian refashioningo fo lder ram-bearing shepherds rather than new statues specificallyrepresenting the Good Shepherd.

The importanceo fl abelling
Establishing the 'correct'-in the sense of 'originallyi ntended'-identity of a statue in the round or afigure depicted on arelief, in apainting,orinamosaic is not always straightforward, not even in today'sh yperliterate academic community.Without clear attributes or other iconographic clues,p agan-mythological subjects are difficult to identity.⁶⁷ Distinguishing between as tatueo fa divinity and areal person is not always possibleeither,⁶⁸ and honorific portraits especiallyp ose manyd ifficulties of exact identification. Specific characteristics usually make it possiblet oc ategorise them broadlya sa ni mperial ruler,d ignitary,learned individual, and so on,⁶⁹ whereas, basedoncomparisons with other statues, paintingso rc oin portraits, we can sometimes find out theirn ame.B ut even today, with all the resources and comparative material at our disposal, we oftenf ail to do so.
Problems of identification were infinitelylargerinthe past and could onlybe avoided by adding name labels.⁷⁰ Aristotle said of ancientart "one could not recognize what each thing was unless someone had inscribed it."⁷¹ Similarly,P au- sanias (10.25) tells us that he was unable to identify some of the characters in an enormouspainting thathesaw in Delphi, because not all of them werelabelled. Even though labels became much more frequent in Late Antiquity,especiallyfor biblical figures as well as saints, manyn on-Christian and non-Jewish figures in narrativescenes in paintings, mosaics and reliefsstill remained unnamed.⁷² And even though statuary in the round usually wasidentified in the inscriptiononits base, at least when displayedinpublic space, alreadyfrom the Late Republican period onward,h onorific statues werer epurposed and reidentified by recutting or replacing the inscription.⁷³ As statues werei ncreasingly moved around and used for entirelyn ew purposes in Late Antiquity,e xamples separated from their original identifiers onlybecame morenumerous.InI talyand North Africa, statues werer e-erected on new basesf eaturing new inscriptions, which almost never mention the exact subject of the statue.⁷⁴ The far-awaylocations of origin of the statues re-erected in Constantinople meant that they as well wered etachedf rom their heavy bases and placed on topo fn ew ones at their final destination.⁷⁵ They would have had to be reidentified, and this new identity had to be eternalized, again, in writing.A ss uch, at least some of the statues in the baths of Zeuxippos apparentlywerepositioned on top of bases with rather careless inscriptions simplyc larifyingw ho they wereo ra tl east who they were thoughtorpropagated to be.⁷⁶ Labels could thus have been added at anypoint in time and do not necessarilys ignpost a 'correct' identity.They onlyi ndicate that  R. Leader-Newby,Personifications and paideia in late antique mosaics from the Greek East, in E. Stafford/J. Herrin (eds.), Personifications in the Greek world: from antiquity to Byzantium. Aldershot ,  -;R.Leader-Newby,Inscribed mosaics in the lateRoman Empire: perspectivesf rome ast and west, in Z. Newby/R.L eader-Newby (eds.), Art and Inscriptions in the Ancient World. Cambridge ,  - givesa no verview of the different conventions per region; S.V. Leatherbury,I nscribed within the image:t he visual character of earlyC hristian mosaic inscriptions. Ph.D.d iss., University of Oxford , .C onversely, Christian figures wereo ften accompanied by labels. See Leatherbury,i bid.,  -;M arsengill,P ortraits and icons (as footnote  above), passim.  B. Longfellow,The reuse and redisplayo fh onorific statues in Pompeii, in Ng/Swetnam-Burland, Reuse (as footnote  above),  -,o n -.S ee also Rous (as footnote  above),  -.  See footnote  above.  Ma,T ravellings tatues (as footnote  above).  Three statue bases were recovered fromt hese baths,o ne identifying 'Aischines',o ne referring to 'Hekabe',w hereas at hird was uninscribed. See S. Guberti Bassett, Historiae custos: Sculpturea nd Tradition in the Baths of Zeuxippos. AJA  ()  -.M a ,T ravelling statues (as footnote  above)  consideredthese explanatory labels to be "of an almost 'museographical' type". the owners or caretakers of as tatue viewed it in ac ertain waya taspecific moment in time and wished to conveyt his vision to others as well.
The most intriguing attestation of labellingi nt he context of this article is a poem writtenb yt he empress Eudocia after avisit to the healing baths of Hammat Gader (Gadara) either in 438 -439o rbetween 443and 445. The text was integrated in front of apassageconnectingtwo major halls of the bath house as an epigraphic panel,probablyin4 55.⁷⁷ The second part of the poem takes the form of an ekphrasis referring to separate areas of the bath: "Indian and Matrona, Repentinus, Elijah the Holy, Antoninus the Good, dewyGalatea and Hygeia herself, the large lukewarm pool and the small lukewarm pool, the Pearl, the old clibanus,I ndian, and alsoa nother Matrona, Briarat he Nun, and the [spring of the] Patriarch."⁷⁸ Someo ft he areas are therefore identified by specific pools or water features (the large lukewarm pool and the small lukewarm pool, the Pearl, the old clibanus,t he [spring of the] Patriarch⁷⁹), others by statues. Some of these characters we would expectinaRoman bath complex, includingstatues of femaleb enefactors (the two matronae,p ossiblya lso the Indians, if they represented honorands in exotic garb, Repentinusprobablyw as alocal donor), imperial statues, and personificationsofwater creatures (Galatea the sea nymph).⁸⁰ Conversely, Elijah the Holya nd Briarat he Nuna re quite untypicalf or aR oman bath house.⁸¹ In these cases, Eudocia'sekphrasis no doubt reflects contemporary The text was found in AreaD ,oppositethe passageleadingt oAreaA .F or an extensive discussion on the text,see S. Busch, Versus Balnearum. Die antikeDichtungüber Bäder und Baden im römischen Reich. Stuttgart/Leipzig ,  -.  DiSegni,Greek inscriptions(as footnote  above), ;Busch, Versus Balnearum (as footnote  above), ,with adjustment to "Briarathe Nun" suggestedbyIda Toth. J. Green /Y.Tsafrir,Greek inscriptions from Ḥammat Gader:apoem by the empress Eudocia and twobuilding inscriptions. Israel Exploration Journal  (),  -,on as wellasDiSegni,ibid.,  opt for the translation "Briara and the nun",a lthough the first also suggest considering μονάστρια as an adjective, "the mighty nun".B usch,i bid., ,o pted for "Briara, nun."  Green/Tsafrir,Greek inscriptions( as footnote  above),  -;E .R. Habas,Apoem by the empress Eudocia: an ote on the patriarch. Israel Exploration Journal  (),  - argues for ac ontemporary Judaic patriarch; Busch, Versus Balnearum (as footnote  above), ,l eavest he question open.  Green/Tsafrir,Greek inscriptions( as footnote  above),  -.  Overviews of statue subjects present in baths can be found in H. Manderscheid,Die Skulpturenausstattungd er kaiserzeitlichenT hermenanlagen. Berlin ;M arvin,F reestanding sculptures (as footnote  above); R. Bol,M armorskulpturen der römischen Kaiserzeita us Milet.A ufstellungskontextund programmatische Aussage.F unde ausMilet . Milet, ..Berlin, ;Jacobs,From production to destruction? (as footnote  above) Appendix; L. M. Stirling, identification.⁸² The Elijah most likelyrepresented the prophet Elijah, who in late antiquel iterary sources is named as ar olem odel for monastic life and depictions of whom are known from wall paintingsa nd sarcophagi as well as from the wooden doors of SantaS abina mentioned above.⁸³ Who Briarat he Nun was and whys he mattered to contemporaries remains entirelyu nknown.
In this ekphrasis, Eudocia mayh avei ncorporated information conveyedt o her orally, but,a st he bases from the Bathso fZ euxippos indicate,s he may also have repeated what was written on the bases of the statues. Aconfirmation of the existenceo fs uch Christian epigraphic labels comes from Echinos in Thessaly. Excavations of as mall church outside the walls of the city uncovered a cubic stone base of ab ronze statue, bearing an inscriptiont hat identified the honorand as Saint Athanasios of Alexandria. The production of ab ronze statue of an Alexandrian bishop for such amodest building -the church is not even 10 meters long -in as mall town in Thessalyseems highlyunlikely.⁸⁴ It is far more plausible thatanolder item deemed well-suited for the representation of ah igh ecclesiastical official had been taken from elsewhere. Adding an ew inscription was enough both to fashion and to stabilize an ew identity for the statue.
In contrasttothe baths of Gadara, whereElijah and Briarathe Nunwerepossiblyalreadyondisplaybefore theiridentity change, the example at Echinos was very much an intentional act.The people responsible for the new identity recovered the statue from elsewhere, had it transported to this small church and added the inscription to the base.Presumablythis process mirrored that at Caesarea Philippi described above. Philostorgius mentions an inscription clarifying the proper identification of the statues there as well. Wilson argued thatthe original text on the base mayhavecontained general titles that could easilybereinterpreted in aC hristian sense.⁸⁵ However,s ince Eusebius makes no mention of Patrons, viewers, and statues in latea ntique baths,i nS .B irk /B.P oulsen (eds.), Patrons and viewers in latea ntiquity. Aarhus Studies in Mediterranean Antiquity, .A arhus ,  -.  The phrasingused in Roueché,Seeing statues (as footnote  above), ,that the "subjects do not all sound 'modern'",isquiteunlucky,asthis was the fifth-century,modern interpretation of them. I. Jacobs, Old statues, new meanings the inscription at all, it is more likelyt hat al abel was applied in the century in between the twoa ccounts, to make the new identity permanent.

Incentivesf or and instigatorso fC hristian identifications
Although it can thus be established that imaginative manipulations of identity weretaking place and wereonoccasion also 'eternalized' in the form of inscriptions, we rarely find information on how intentional such reidentifications were or who was responsible. Eusebius remains very vagueo nt he latter point,using the narrative voice presumably to indicate the local population of CaesareaPhilippi. In his letter to Constantia, he is explicit about his own ignorance concerning the origin of the Christian identification. Eudocia mayhavebeen expressing her personal vision, but it is more likelythat she wrotedown what she was told and/or read on the baseswhen visitingthe baths. In line with what is known of informantsi np ast centuries,⁸⁶ local Christian populations weren od oubt responsibleb oth for the conveyance and alsot he conception of new identities. I would moreover like to point to the involvement of local bishops in these processes. ⁸⁷ We geta ni dea of the role of bishops in conveying information to travellers and no doubt alsot ot heir community from the travel account of Egeria. When Egeria visits Edessa, it is the bishop who shows her "huge" marble portraits of King Abgara nd his son.⁸⁸ When she is travelling through the desolate city of Rameses, she comes face to face with "Theban stone" (porphyry) statues presented to her as Moses and Aaron by the bishop of the nearby city of Arabia.⁸⁹ It is safe to assume thatc ertainlyt he latter weren ot originallyc arved to depict two Old Testament prophets.Yet information provided by the bishop can be considered trustworthyb ecause he was, as Egeria adds, "am an of some age, of a godlyl ife since the time he became am onk, and an approachable man, who is very good at welcomingp ilgrims and alsov ery knowledgeable about God's Scriptures".⁹⁰ This stressingoft rustworthiness founded on superior knowledge is ar ecurrent theme; when Eusebius mentions the Good Shepherds at the fountains of Constantinople, he adds that they were "evident signs to thosew ho start from the divine oracles."⁹¹ Egeria'sexperience in Rameses moreover is very much reminiscent of the late antique ekphrastictradition, with the bishop taking the place of the late antiquepoet.Abrief comparison between, for instance, Christodorus of Coptus' ekphrasis on the baths of Zeuxippos and the explanation receivedby Egeria -and manyC hristians like her,both at Rameses and elsewhere -shows they are two sides of the samec oin.⁹² In bothc ases, an unknowing audience is present: the viewers of the baths in the first and travelling Christians in the second. Though they themselvesare not capable of interpreting what they see, they can relyonexternal sources of information: Christodorus and the bishop of Arabia respectively.B oth informants have the credentials to provet hat their explanation is the correct one: Christodorus has areputation as apoet in the Homeric tradition and can thus identify characters from the story of Troy,whereas the bishop is very knowledgeable of the Bible. Both are thereforeq ualified to guide viewers towardst he correct identification:⁹³ specific characters from the Iliad or Aeneid with Christodorus, specific biblical figures with the bishop of Arabia.
Mary Whitby'se stimate of Christodorus' achievement goes as follows: "Facedw ith ac ollection of statues thati ncluded an umber of beardless naked youths, figures wreathed in laurel, and draped women, Christodorus gives them life and meaning by relatingt hem to the familiar material of myth, often the Trojan story […]."⁹⁴ It is not hard to imagine aChristian variant of this,where- I. Jacobs, Old statues,n ew meanings by aC hristian bishop, faced with an umber of unidentifieds tatues untethered from their intended identity,g avet hem new life and meaningb yr elatingt hem to the familiar stories of the Christian past,oftenthe Bible. Although other mechanisms behindthe allocation of new identities can and should not be excluded, it seems very plausible that local imaginative bishops wereinvolvedinthe Christian rebranding of old images. In some cases, reidentification mayh avet aken place in order to explain the presenceo fn oteworthys tatues in the landscape. Their new identity could even be accompanied by ac ompletef oundation story referringt oe vents from the Bible or to the deeds of remarkable Christians from the ageofpersecution. Thus Malalas tells the lengthytale of five bronze female statues "standing at thatpublic bath [of Antioch] to the present," who supposedlydepicted five pious Christian woman martyredb yt he emperor Trajan.⁹⁵ Yeti ti sa lso possiblet hat older statues were deliberately soughto ut to depict local Christian celebrities. Parallel to what had happened in the pre-Christian past,s uch celebrities as well as tangible traces of al ocality'si mportance in Christian history were potential sources of patriotic pride.⁹⁶ Theses tatues certainlyd id not do their communities anyh arm. In the case of Caesarea Philippi for instance, the statue of Christ ensured astream of pilgrims who kept on coming even after the destruction of the group under the reign of Julian.

Formal prompts forr eidentification
We can ask ourselvesw hat or whom the statues eventuallyr eimagined as various Christian figures originallyw erem eant to depict.A nalysingt he original identity of reidentified statues is not the purpose of this article though, and, as alreadys uggested, the generous attention givent ot he originallyi ntended viewing of statuary has often led to the neglectingofl ater identities. My reason for turning to this aspect here is to point out that alternative identities werei nstigated by specific visual cues -postures,a ttributes,a nd dress styles -that weremore meaningful to late antique and later viewers thanthey are to academ- Malalas, Chronographia . (Jeffreys/Jeffreys/Scott,C hronicle, as footnote  above, ).  Pausanius (..)for instancepromotes areliefofwhat probablyoriginallywas abearded mountain godcarved in the thirteenth century BC near his native Magnesia as the oldest depiction of Cybele, see F. Rojas /V.Sergueenkova,TracesofT arhuntas:Greek,Roman,and Byzantine interaction with Hittitem onuments. Journal of Mediterranean Archaeology  ()  -,on.Rous,Reset (as footnote  above) researchesthe relation between purposeful reuse and community buildingi nA thens throughout the centuries. ic researchers. Dale Kinney could confirm this for the rider statue of Marcus Aurelius in Rome,a ne xceptional case wheret he statue can be examined together with multiple sources pertaining to later, ever-developing viewingsf rom the tenth century onwards.⁹⁷ By contrast,asn one of the statues featuring in Byzantine literarysources have been preserved, and vice versa, testimonies pertaining to the material examples Iw ill discuss later on in this article do not exist,a ll suggestions made here must remain hypothetical.
Ag ood place to start is the statue group at CaesareaP hilippi. Modern researchers,a fter analysis of Eusebius' testimonya nd other, related passages from later historians,h aves uggested that the figures on displayw ereo riginally intended to depict the emperorH adrian accepting homagef rom the provinceo f Judaea personified as ak neeling femaleo r, more likely, Asclepius or another healing deity,w ith one hand lifted up in the air and as uppliant in front of him.⁹⁸ Christians would have had little difficulty seeing Christ in the bearded protagonist,a st herew erem anys imilarities between the iconographyo fA sclepius and some late third-century depictions of Christ,i ncludingt he bare chest and full beard (Fig. 1).⁹⁹ In this case, the bearded protagonist,t he suppliant woman, and the granting of aw ish expressed in an outstretched hand were ample cause for reidentification.
The abovementioned gold-plated brass statue of Daniel and the lions that, accordingt oE usebius, was dedicated by Constantine at Constantinople may have been ar eidentification as well.¹⁰⁰ There are not manyRoman male statues with accompanying lions -astatueofHercules fighting the Nemean lion would not have been appropriate for this particularr einterpretation -but perhaps a new combination was made between am ale statuea nd freestandingl ions, of which therewereplenty from Roman and pre-Roman times. Also at Constantinople, ab ronze statue of ab ishop holding as taff originallym ust have been an imageo fA sclepius,c onsidering Byzantine sources mention thatt he staff was  Kinney,H orse (as footnote  above), especially .B .A nderson,Constantinople'sm edieval antiquarianso ft he future, in: Cities as palimpsests?U rbane volutions in the Eastern Mediterranean. Cambridge (forthcoming), makes avery similar argument about the interpretation of statuary by the authors of the Patria beingd rivenb yc lose and "direct observation."  See Weber,Statuengruppe (as footnote  above), ;Wilson,Caesarea Philippi (as footnote  above),  -.B eatrice,P ilgerreise (as footnote  above), ,c ontests that the group depictedHadrian. Avdokhin,Christianizing(as footnote  above) provides auseful overview of the iconography.  Mathews,Clash of Gods (as footnote  above)  -.  This was alreadys uggested by Cameron/Hall,E usebius (as footnote  above), . snake-entwined.¹⁰¹ Agroup of one mana nd three women standingo nc olumns in the hippodrome was identified in the Parastaseis as Adam and Evew ith the personificationso fP lenty (Εὐθηνία)a nd Famine (Λιμός), but originallym ay have depicted Herakles and the Hesperides sisters.¹⁰² The prompt for the alternative viewing in this casemust have been provided by the tree, apple, and serpent with which Herakles was commonlyd epicted.¹⁰³ In the archaeological record, ag ood candidate for such reidentification through 'recognition' of ac ertain prompt can be found at Aphrodisias. One of  the double half-columns in the peristyle courtyard of the Triconch House at Aphrodisias features reliefsofwinged females set in small panels (Fig. 2).¹⁰⁴ The top panel,with one winged figure, is badlyp reserved. The lower shows two winged females in long garments standing underneath an arch. The reliefs,which predate the late antique house, werer eused also in the Mid-Byzantine phases, from the late eighth or earlyninth centuryuntil the twelfth century,when it functioned as the residence of the bishop of Aphrodisias.¹⁰⁵ In this phase at the latest,t he figures could easilyh aveb een read as depictions of angels.¹⁰⁶ The winged angel, with flowing robes and asexual appearance,h ad indeedb ecome aw idespread motif from the late fourth century onwards.¹⁰⁷ The fast-growing popularity of angels is obvious in their importance in scenes such as the Annunciation, the Nativity,a nd Resurrection but even more so in their appearance as stand-alone guardians,m essengers, and psychopomps in paintingsa nd mosaics, on sarcophagi, ceramic lamps, and so on. Givent hat such winged angels wered erived from and thus very close in appearance to depictions of winged genii and Victoria/Nike,¹⁰⁸ it is no surprise that they would have been 'recognized' in older material culturea sw ell. TheM edieval Mediterranean, .L eiden ,  -,s uggests that,consideringt he popularity of the archangels Michael and Gabriel, the reliefs werereinterpreted in this specific manner.  Beforet hat period, angels werec ommonlyd epicted as wingless,h uman-likeb eings.A fter  as well, wingless angels appear,a st he presenceo fw ingsd epended on the natureo ft heir activity and the contexti nwhich angels were depicted, see G. Peers,S ubtle bodies.R epresenting angels in Byzantium. TheTransformation of the Classical Heritage, .Berkeley/Los Angeles/ London ,  -.For the late antique iconographyofangels,see G. Stuhlfauth,Die Engel in der altchristlichen Kunst.F reiburg ;T h. Klauser,s .v.E ngel X( in der Kunst). RAC  (),  -;P eers,ibid.,  -;C.Proverbio,Lafiguradell'angelo nella civiltà paleocristiana. Pian di Porto ,  -.E .J astrzębowska,N ew Testament angels in early Christian art: origin and sources. Światowit. Annual of the Institute of Archaeology of the University of Warsaw  ( -),  - focusses on the fourth to sixth-century iconographic development of New Testament angels.  Peers,S ubtle bodies (as footnote  above),  - for asummary of resemblances and differences in gender and dress, with further bibliography;  - discusses further iconographic proto-types.

I. Jacobs, Old statues, new meanings
In other cases, prompts mayh aveb een less specific. The identificationo f statues in Rameses as Moses and Aaron mays imply have been basedo nt he fact that the Hebrew Bible states that Aaron livedi nt his eastern border-land of Egypt.Likewise, acertain pose or dress mayhavebeenenough reason for reidentification. Thus, the posture of as eated statue at Constantinople, the chin propped on ah and,w as enough for Byzantines to perceive the wisdomr epresented in the statue and to identify it with Solomon. Originally, it must have depicted ap hilosopher,poet,o rs ome otherm an of letters.¹⁰⁹ Fort he statue identified as the bishop Athanasios in Echinos, we maysurmise aseated philosopher as well. The statue identified as Briarathe NuninEudocia'spoem mayhavebeen an older honorific portrait statue, the veil of which covered the hair,i fn ot the lower part of the face as well.¹¹⁰ Such ah ypothesis is basedo nt he assumption that nuns alreadyi nthe fifth century distinguishedthemselvesf rom layw omen and that the statue reidentified as such must have had at least some resemblance to ac ontemporaryn un. As far as we can tell, like layw omen, nuns woreatunic and cloak.¹¹¹ In addition, they invariablyc overed their head with av eil or maphorion.¹¹² Yet, once observation had established aprompt,itmade contemporary viewers overlook otherfeatures that, at least to us, contradict the new identification. Returning to the Christ statue in Caesarea Philippi, the prompts that caused the association with the story of the Haimorrhousaa re fairlye asy to identify:E arly Christian renderings of the scene in the catacombs,ons arcophagi, textiles,a nd so on depict the woman, ofteno nh er knees, sometimes with her left hand stretching out in supplication.¹¹³ In front of her stands Christ,who is turned to- LeoG rammaticus  - and Patria .,d iscussed in Mango,A ntique statuary (as footnote  above),  -;B assett,Urban image( as footnote  above),  - cat.n o. .  The veil originallyi ndicated the woman was of marriageable ageo rm arried (Dillon,F emale portrait statue, as footnote  above, , , , ). L. Llewellyn-Jones,Aphrodite's tortoise. The veiled woman of Ancient Greece.Swansea ,discusses the practice of veiling in the Greek world. Foru se of the veil with latea ntique statues,s ee K. Schade,F rauen in der Spätantike -Status und Repräsentation. Eine Untersuchungz ur römischen und frühbyzantinischen Bildniskunst.M ainz , .  J. Ball,D ecodingt he habito ft he Byzantine nun. Journal of Modern Hellenism  - ( -),  - discusses literary and iconographic evidencef or Byzantine nuns; Krawiec, "Garments" (as footnote  above).  Ball,D ecoding (as footnote  above), . wards her and points in her direction, indicating healing.¹¹⁴ However,a tt he same time, it is clear that some aspects of the statuarygroup at Caesarea Philippi did not match the details of the biblical story or other earlydepictions. There is no biblical explanation for the plant growinga tt he statue'sf eet nor does it appear in other renderingso ft he scene. Elsewhere, the woman is always depicted touchingthe hem of Christ'srobe with her right hand, obviouslythe key element of the miracle but apparentlyabsent in the iconographyofthe Caesarea Philippi statue.
Cues for reidentification are of course not limited to Christian examples. Laurel wreaths wereenough to make Christodorus see Homeric prophets in the Zeuxippos baths.¹¹⁵ Interestingly,the Parastaseis ( § 43)describes astatue thatpresumablystood at the Milion as a "porphyry statue […]ofthreestones with threestone heads, which some said wasConstantine the Great in the middle, Constantius on the left,and Constans on the right with twofeet but six hands -astrangespectacle for those who sawi t, each one looking in ad ifferent direction -and one head."¹¹⁶ The three upper bodies indicate that the statue originallye mbodied the goddess Hekate, of whom statues with one head but three profiles looking in diverse directions survive.¹¹⁷ The most obvious cuefor the imperial reidentification is the material that the statue was made from, porphyry.¹¹⁸ In addition, the connectedness of the three porphyry torsos would have been reminiscent of the porphyry groups of embracing tetrarchs.They themselvesh ad been reidentified as Constantine'ss ons and gave their name to the square wheret hey werer eerected, the Philadelphion.¹¹⁹ With this in mind, the reidentificationo ft he por- Even though according to the original story the woman with the issue of blood secretively approaches Christ from behind to touch the hem of his robe,E arlyC hristian healingm iracle iconographyh as Christ turned towards the recipient of the miracle (Jensen,U nderstanding, as footnote  above,  -).  Whitby,Christodorus (as footnote  above), .  Transl. Cameron/Herrin,E usebius (as footnote  above),  -.  The statue is discussed in Bassett,U rban image( as footnote  above), ,c at.n o. . phyry Hekate as asecond monument to Constantine'ssons and successors Constantius and Constans is less surprising,event hough the femaleh eads and the joined bodies to the modern mind are far from apt for imperial portraits.
Some of the Byzantine viewingso fa ne questrian statue on the Forum Tauri at Constantinople likewise are extreme cases of reidentification to modern eyes.¹²⁰ The imperial statue, acolossal bronze rider statue, is described by Niketas Choniates as arider with aglobe in his left hand. He had his right hand outstretched, as is common for depictions of Romanrulersassign of adlocutio, adventus,o rs imilar.U nder the uplifted foot of the horse, as mall figure of a barbarian was present,c onfirmingt hat this was an imperial rider statue.¹²¹ Though several Byzantine sources identifiedt he statue as one of Theodosius, some inhabitants of Constantinople apparentlys aw it as Bellerophon whereas others claimed that is was Joshua, two rather divergent and not obvious alternative interpretations.¹²² The identification of Bellerophon mayhavebeen based on the motif of arider conqueringanopponent,eventhough,atthe same time, contemporaries must have recognizedt hatt he opponent depicted in this statue group bore no resemblancet ot he Chimaeraa nd that the horsew as wingless.
Choniates recounts how the identification as Joshua was basedo nt he outstretched hand of the rider,agesture related to the events at the Battle of Gabaon (Gibeon) as can be read in the book of Joshua 10:12 -13.I nE arlyC hristian and Byzantine depictions of the scene, Joshua is shown, hand outstretched, but on foot,n ot on horseback.¹²³ There weret hereforep lenty of visual aspects that wereh ighlyc ontradictory to Joshua as well as to Bellerophon. In particular, lung in Konstantinopel. Istanbuler Mitteilungen  (),  -.Inthis publication, aTheodosian dateo ft ransfer and reidentification is preferred.  The origins of the later identifications made arediscussed in R. H.W. Stichel,Bellerophon oder Joshua -zur vermeintlichen Missdeutungspätantiker Kaiserstatuen Konstantinopels in byzantinischen Quellen, in R. Harreither (ed.) Frühes Christentum zwischen Rom und Konstantinopel. Vienna ,  -.A nderson,A ntiquarians (as footnote  above), comments on both identifications as well, but moreover offers aconvincingexplanationwhy the rider statue was no longer identified as al ate antique emperor based on close readingo fr elevant passagesi nt he Patria and Constantine of Rhodes. The statue and literary sourcesr eferring to it arem entioned in Bassett,Urban image( as footnote  above), cat.n o. .  (Historia , )  Similar identifications are alreadyp resent in the Patria Konstantinupoleos .;D agron, (as footnote  above), , , ;A.Berger,Untersuchungen zu den Patria Konstantinupoleos. Poikila byzantina, .Bonn , .A sthe Patria is acompilation of older sources, it is likelyt he identificationi ss till older than the tenth century.S tichel,B ellerophon (as footnote  above), ,s uggests the reidentifications of both Bellerophon and Joshua go back to associationsb etween them and late antique emperors,m ade in latea ntique imperial panegyrics.  Stichel,i bid.,  -;s ee now also Anderson,A ntiquarians (as footnote  above). the left hand holding ag lobe is somewhat puzzling in either case.¹²⁴ Both identifications are based on as ingle aspect that hardlys eems meaningful today.
Interestingly,the conclusion that one 'recognizable' elementinastatue was enough to make anew identification alluringand convincing,ismatched by conclusions reached by Richard Krautheimerc oncerning medieval architecture in his 1942a rticle "Introduction to an 'Iconographyo fM ediaeval Architecture'". Based on ac omparison of tenth-and eleventh-century 'copies' of the HolyS epulchre,h ef ound that,a gain to moderne yes, there was both an "inexactness in reproducing the particulars hape of ad efinite architectural form" and a "peculiar lack of precision in medieavald escriptions not onlyo fa rchitectural patterns but of all geometric forms".¹²⁵ With aw ealth of architectural remains but also associatedl iterary and related iconographic sources at his disposal, Krautheimerw as able to infer that the minds of medieval viewers were indeed tuned to specific elements rather than the whole architectural prototype.E ven though they transferred measurements,numbers and forms in splendid isolation and combined them with entirelynew featuresintheir new context,these single elements always remained representative of the original building.¹²⁶ In some cases, the content and meaningo ft he original weres op aramount thatt he mere dedication of ab uilding( its new 'label' so to speak)w as sufficientt o turn it into a 'copy',e veni ft he two physicallyh ad nothing in common.¹²⁷ These,t ou sp eculiar, manners in which 'recognition' of the HolyS epulchre and in particularo ft he Anastatis Rotundaw as achieved in medieval churches as well as in late antique and medieval iconographic representations, runs parallel to what we have seen for the ways in which statues werer eidentified: aselection of one or more essential elements, which could differ from the ones modern viewers value most,e nables identification.
Finally, before we turn to material evidence for Christian statuary,itisworthwhile repeatingthat statues could be simultaneouslyidentified in multiple ways. As such, the rider statue on the Forum Tauri could be seen as both Joshua and  Stichel,B ellerophon (as footnote  above),  -.  R. Krautheimer,Introduction to an 'iconographyofmediaeval architecture'. Journal of the Warburga nd Courtauld Institutes  (),  -,h ere .  In the case of the Holy Sepulchre, therew as for instanceafocus on the numbers of the eight piers supposedly present in the original Anastasis Rotunda or the twelve columns described by Eusebius as surroundingt he apse of the Constantinian basilica. The imitation of either one or the other in medievalb uildings was apparentlys ufficient to produce a 'copy' of the original sanctuary,r egardless of the architectural form of the buildingt hey appeared in; see Krautheimer,I ntroduction( as footnote  above),  -.  Krautheimer,i bid.,  -.
I. Jacobs, Old statues,n ew meanings Bellerophon by some and was still an imperial rider statue to others.¹²⁸ Likewise, the introduction of anew identification could be highlycontentious. Although I do not know of anyliterary reference pertaining to acontested reidentified Christian statue, contemporary passages do speak of gods and goddesses 'misidentified' as personifications or secular individuals. Zosimus thus comments on the rebranding of as tatue of Rhea as the Tyche of Constantinople in orans pose, achieved in this case with minor physical alterations.¹²⁹ The Parastaseis explicitlyr efers to the existenceo fc ontemporary competingo pinions regardingafemale seated statue in the Hippodrome.¹³⁰ Disparate identifications, as in the past,m ay have reflected ad ifferent personal background or differing levels of knowledge,b ut also mayh avef avoured personal observation and visual prompts over tradition or vice versa.¹³¹ Am ore in-depth studyofthis phenomenon, though obviouslystronglyrelated to the current topic, lies outside the boundaries of this article. When reviewingt he material evidence in the following sections of this article, it is, however,important to be aware that assorted stories circulatedamong citydwellers and visitors,most of which weren ever written down, and thatt he Christian identification could be just one of many.¹³²  Contentious and competingidentifications occurred in previous centuries as well. Herodotus,for instance, mentions identifications of aHittiterelieffigure in Karabel as the pharaoh Sesotris (the identity Herodotus himself prefers) and the Asian hero Memnon (Rojas/Sergueenkova,T races, as footnote  above, ).  Zosimus .. -.  Parastaseis § ,i dentified as Verina, wife of LeoI(  -)o rA thena. The story, with competing interpretations,i sr epeated in the Patria ..S ee also LSA-.  On the authorshipo ft he Parastaseis,s ee Anderson,Classified knowledge (as footnote  above). Anderson,A ntiquarians (as footnote  above), contrasts the viewings of Theodosius' rider statue with that of Constantine of Rhodes.Kinney, Horse (as footnote  above),  is able to distinguish between opinions expressed by pilgrims,l ocals and the clergy.  Although outside of the scopeo ft his article, the fact that the opinions of ordinary citydwellers weren ot always receivedf avourablyb yi ntellectuals is exemplified in ac omment by al-Muqaddasi( d. )i nh is Ahsan al-Taqāsīmf īMa'rifat al-Aqālīm,amajor geographical work written around .A mongt he sights of the city of Hims (modern-dayH oms), he lists a "figure of am an wrought in brass,s tandingo nafish which the four winds cause to rotate." Al-Muqaddasi continues: "About it manystories aret old, but not to be believed. This town has suffered violent disorders,and is falling into ruin." Afterwhich he concludes: "The people there arestupid." See: Al-Muqaddasi (d. ), The best divisions for Knowledge of the Regions, trans. B. Collins.R eading , .Ithank EdwardZ ychowicz-Coghill for pointingm et othis passage.

Archaeological case-study: archangels at Sagalassos
In view of the above, it is highlyprobablythat some of the statues and reliefsstill preserved todayw erea ts ome point in their history viewed as Christian protagonists or scenes.None of these remains werementioned as such in literary sources though,and none of them preserveepigraphic labels corroboratingr eidentification. Consequently, we have to relyonthe statues themselves, acareful study of their find context and,when possible, an equallya ttentive reconstruction of their displaycontext,aswell as an investigation of the pertinent Christian iconographic repertoire. Due to wordconstraints, Iwillanalyse one case-studyinfull as am atter of example, whereas other material remains will be discussed more succinctlyi nt he following sections.
In late antiquetimes, Michael, archangeland archistrategos of the heavenly army, was ah ighlyp opulari ntercessor between God and humankind.¹³³ It was alreadymentioned how,accordingtoMalalas, Constantine identified acult statue as the archangelMichael. In keepingwith the broader cult of angels, the cult of the archangeldeveloped earlyinAsia Minor, notablyinGermia and Colossae/ Chonae.¹³⁴ Michael'sp opularity in Asia Minor is furthermore confirmedb yt he highnumber of inscriptions mentioning his name, especiallyinPisidia, Phrygia, Asia and Galatia.¹³⁵ At Sagalassos, ap rovincial town in Pisidia, the ruins of a bouleuterion weret ransformed into ac hurch dedicated to St.M ichael possibly in the second or third quarterofthe sixth century.¹³⁶ Aconcentration of inscriptions and graffitimentioningt he followers of Michael, the Michaelitai,wered is-covered in its immediate surroundings,onboth ceramic patens and elements of architecture.¹³⁷ At the time of the conversion, manyoft he buildingblocks and much of the original decoration of the bouleuterion werer eused in the church'sw alls. Judging by the discovery of frieze fragments depicting helmets, shields and cuirasses in the debris of the church courtyard, some frieze blocks of the council house werer eincorporatedi nto the basilica'so uter walls.¹³⁸ They would indeed have made afitting decoration for achurch dedicatedt othe angel-warrior.Moreover, two piers from the original gallery werea lso recovered duringexcavations.One pillar bore ar elief of af ull-sized Ares,the other one of afull-sizedA thena, with respectively af emalea nd male prisonera tt heir feet ( Fig. 3).¹³⁹ Their find position in the debris indicated thatthey had been placed at the top of amonumen-  tal staircase connectingthe church complex with the neighbouring Upper Agora, with the reliefs framing the entrance into the atrium and greetingvisitors to the precinct.
The heads of both gods are missing.Had this decapitation been intentional, the reuseo ft he reliefs in this conspicuous location could cautiouslyh aveb een considered as tatement of the victory of Christianityo verp aganism. However, there are no clear marks on the stone to confirm deliberate decapitation and, in view of their overall poor state of preservation, the damaget ot he reliefs mayj usta sw ell be the resulto fm oren atural processes.¹⁴⁰ Narrative scenes such as these, with figures in full action who bear very few distinguishing attributes,generallywerenot targeted in Late Antiquity.¹⁴¹ Considering that they were probablystill (largely)intact in the sixth century,itisfar more likelythat churchgoers looked favourablyu pon these reliefs and had givent hem an ew identity. Since the Bouleuterion Church wasdedicated to St.Michael, anew identification as archangels is very attractive.The billowing robes of Ares and Athena could have been taken for angels' wings, whereas the cowering figures,w ith their raised knees and arched bodies, are reminiscent of the serpents or dragons trampled in Christian triumphal iconography.¹⁴² Because of theirperceivedrole as defenders, the archangels Michael and Gabriel weremore frequentlydepicted as watchers and guardians at thresholds and gateways,for instance at the doorway of the late fifth-centurybasilica at Alahan (Fig. 4).¹⁴³ Both of them have vanquished pagan enemies underneath theirf eet: Gabriel stands on "the back of abull, below which is amale bust with head uncovered," whereas underneath Michael "busts of two femalefigures,each with a Phrygian capo nh er head" wered epicted.¹⁴⁴ Thei conographyu sed here, with the angels Michael and Gabriel depicted static and frontally, was common for late antiqued epictions of archangels, but it was not the onlyw ay angels were  R. Turban,D es enfances de Bacchus àl aN ativité, in V. Gaggadis-Robin /N.D eL arquier (eds.), La sculpture et ses remplois.Actesdes II es rencontres autour de la sculpture romaine. Bordeaux ,  - discusses ac omparable example whereareliefd epictingt he birth of Dionysiosw as integrated as lintel into the Church of Corcolle. Up until the modern dayi tw as viewed as the birth and infancyo fC hrist.  Jacobs,F rom production to destruction? (as footnote  above), . imagined and,asusual,there was plentyofroom for alternatives.¹⁴⁵ In later medieval iconographic sources, the archangelM ichael is often depicted in full armour,h overing over av anquished enemy, ofteni nt he form of as erpent.T his fullyd eveloped iconographyo fS t. Michael slaying ad emon or dragon shares al ot of resemblancet ot he Sagalassos reliefs,yet the earliest securelydated exampleso ft his particulari conographyo fM ichaeld ateb ack onlyt oa round 700.¹⁴⁶ It has been suggested that the more dynamic Michael iconographyw as based on the battle between Michael and Satan as described in the Book of Revelation 12:7-9.¹⁴⁷ Other evidence suggests that these verses had an impact alreadyi nL ate Antiquity.E usebiusf or instance recounts how ap ainting of Constantine and his sons standing on top of as erpent pierced by aw eapon was put on displayi nf ront of the imperial palace of Constantinople. Considering that Constantine had the imagemade in the aftermath of his victory over Licinius near the Michaelion of Chalcedon, the sanctuary dedicated to St.M ichael con-  Glen Peers has argued that the metaphysical natureofangels made it very difficulttoeither describe or depict them. See Peers,S ubtle bodies (as footnote  above), esp. ,  -,  -,G .P eers,A pprehendingt he archangelM ichael: hagiographic methods. BMGS  (),  -.  Schaller,E rzengelM ichael (as footnote  above),  -.  Ibid.,  -. structed by the emperor several years earlier,¹⁴⁸ it is likelythe emperorwanted to referencet he war in heaveni na na ttempt to associate himself with St. Michael.¹⁴⁹ The iconographyo faC hristian victor trampling as erpent became widespread in the following centuries,w ith examples in the palace mosaic in Sant'Apollinare NuovoatRavenna, and mid-to late-fifth-centurygold emissions in both Ravenna and Constantinople,¹⁵⁰ as well as on some late antique amulets and African Red slip ware,¹⁵¹ whereitispresumablyChrist who is trampling the serpent.I nL aodicea in Phrygia,apilgrimagea mpulla was found with ad epiction of either Christ or as aint treadingo nt he snake/basilisk.¹⁵² Consequently, even though thereisnoevidence for newlycreated renditions of Michael in combination with the serpent in the sixth century and the Sagalassos reliefss howed much more movement than known contemporaryi mages of the archangel, they fit into the existingi conographyo fC hristian triumph. When the builders of the church came across these two reliefs in the debris of the bouleuterion, the presenceo fh uman-likef igures and vanquished enemies at his/her feetm ay have been incentive enough to connect it to ab iblical story that was most apt for ac hurch dedicatedt oM ichael. To moderne yes, not all details of the bouleuterion reliefsatSagalassos are fitting for representations of archangels as demon slayers, even if the iconographyw as already known or developing around the middle of the sixth century.H owever,a sd iscussed above, the aptness of an imagew as evaluatedi nadifferent wayb y late antique and Byzantine viewers.Not all details had to match, far from it.One prompt sufficed to make the association and overlook nonconforming aspects.

The importanceo fp hysical context
Arguing for aChristian reidentification of the reliefs at the Bouleuterion Church at Sagalassos would not have been possiblei ft heir find context had not been carefullyr ecorded and hence their displayc ontext at the entrance to ac hurch atrium not reconstructed.F ind and displays contexts of statuary are invariably instrumental when interpreting how it was viewed in its laterl ife. When tracing Christian reidentifications in the material record, church contexts are an obvious starting point.Eventhough newlycarved statuary has onlyinvery rare occasions been found here, older items increase the total numbers. Next to the Sagalassos example, the statuary base and hence astatue reidentified as bishop Athanasios of Alexandria at ac hurch in Echinos has alreadyb een discussed above.¹⁵³ The best-known and probablya lso the most contested example of Christian reidentification, the famous preserved metope from the northwest corner of the Parthenoni nA thens, likewisec omes from ac hurch context (Fig. 5). At an unknown moment,atthe earliest when the temple was turned into ac hurch dedicated to the Mother of God in the late fifth or sixth century,¹⁵⁴ the metope sculptures of the east,w est and norths ides wered efaced. They had carried mythological narrative scenes of Amazons( west), Giants (east),a nd the Sack of Troy (north), which seemingly lacked aspects that wereoffensive from aChristian point of view.B yc ontrast,t he Centauromachym etopes on the south side, the frieze on the inside of the peristyle and the temple'spediments were not targeted in the same manner,despite the latter being much more eye-catchingand depicting loaded scenes,the birth of Athenaonthe east and the contest between Athena and Poseidon on the west.The preservation of much of the architecture and figurative decoration of the monument has been interpreted as ac onscious decision fuelled by local patriotism to preserveamonument thatwas at the core  See also Turban,E nfances (as footnote  above). of Athenian identity.¹⁵⁵ This makes the concerted effort to erase three-quarters of the metopes all the more puzzling.¹⁵⁶ I. Jacobs, Old statues, new meanings Onlyt he metope at the western corner at the north sider emains in ar elatively good state of preservation. It is visible to av isitor of the Acropolis when facing the formerP arthenon. The metope shows af igure of aw oman, who approaches aseated and dignified femalefigure with agesture of salutation. Modern scholars agree that this scene was originallyintended to depict Athena with possiblythe seated Hera. Yet, by the time the metopes wereselectively mutilated, this relief mayhavebeen spared thanks to a 'recognition' of the scene as aChristian Annunciation,¹⁵⁷ which would be avery fitting decoration for achurch dedicated to the Mother of God.
Although acomparison of iconographic details is made difficult by later mutilations to the metope -the heads and some of the limbs werek nocked offsome observations can be made.¹⁵⁸ If this relief indeedw as reidentified, the prompt no doubt was the combination of seated femaleand approaching figure.
In contemporary Annunciation scenes,which werea lreadyw idespread both in monumental art and on small finds, items of jewellery,a nd in dress by the time the Parthenon underwent its transformation,¹⁵⁹ Mary is generallys eated on the right,while the angel approaches from the left,h is announcement indicated by the raising of his right hand.¹⁶⁰ Sixth-century Annunciation scenes are furthermore characterized by Mary holding up her right hand, parallel to the gesture made by the goddess originallyd epicted in the relief.¹⁶¹ The clothes of the standing goddess weres imilar enough to the typical angel'sd ress of tunic and cloak. The waving folds maye venh avef unctioned as stand-in wings.F inally, as at Sagalassos, alsoa tA thenst herew as even ac onnection between the dedication of the church -to the Mother of God -and the imagery.
Even for more generic statuary,its find and supposed displaycontext within achurch can make aChristian identification avery alluringoption.For instance, ahead of Zeus or Asclepius found in afifth-century context in the Church of St. Theodore at Gerasa is difficult to explain without turning to this hypothesis.¹⁶² Or what to think of across-marked black porphyry femalehead dated to the second or third centuryf ound in the pilgrimagec entre of AbuM ina outside Alexandria?¹⁶³ As remarked by Troels Kristensen, AbuM ina started to develop into ap ilgrimagec entre onlyi nt he fourth century,t he pilgrimagei ndustry being its sole reason for existence. Considering the Christian character of the site and the fact that the head must have been brought from elsewhere, aChristian reidentification and new purpose become very possible.¹⁶⁴ Furthermore, aC hristian identification of statuary is alluringa lso outside churches proper,i na ssociatede cclesiastical buildingsa nd structures. The presence of wingedfemales-made-into-angel-iconographyinthe mid-Byzantine bishop'spalace at Aphrodisias has alreadybeen discussed. Another example deserving of mentioning are the reliefs applied abovethe so-called Gate of Persecution on the Ayasoluk Hill at Ephesus, which because of their particularlocation may even have been intentionallyp olysemic. When the Basilica of St.J ohn was surrounded with af ortification wall, probablya round the time thatt he Justinianic church was built,its main gatewas adorned with reused reliefs of third-century date, recovered from sarcophagi from one of the city'sn ecropolises.¹⁶⁵ The dec- Lidova, ΧΑΙΡΕ ΜΑΡΙΑ (as footnote  above),  -.  Amman, Jordan Archaeological Museum inv. J .A .B. Cook,Z eus.Astudyi na ncient religion ..A ppendixes and Index. Cambridge ,  - suggested it was reinterpreted as Christ.S ee also Th. M. Weber,G adara -Umm Qes. Untersuchungenz ur Topographie,G eschichte, Architektur und der bildenden Kunst einer "Polis Hellenis" im Ostjordanland. Wiesbaden , ,c at.n o. C ,Taf. b.  T. M. Kristensen,M iraculous bodies:C hristian viewers and the transformation of 'pagan' sculptures in lateantiquity,inPoulsen/Birk, Patrons (as footnote  above),  -,see  cat. A.  Kristensen,M iraculous bodies (as footnote  above), .  The construction date of the walls and the gatei sd isputed, see J. Crow,F ortifications, in Ph. Niewöhner (ed.), The archaeology of Byzantine Anatolia. From the end of lateantiquity until the coming of the Turks.N ew York ,  -.The current excavators connect it with the school system long until after these walls wereb uilt.¹⁶⁸ It is thus likelyt hat the reliefs werestillrecognised as depicting the fall of Troy when they weremoved to their new position abovethe gate. Yet, their location abovethe main entrance of one of the most important pilgrimagecentres of the Byzantine world means that they also lent themselvesparticularlywell to aChristian viewing.¹⁶⁹ Pilgrims visiting the site, as well as the majority of local Ephesians, had not had the same education. To them, the much more familiar world of martyrs and saintsw ould have come to mind on this saintlysite. Even though the earliest references to the presenceofascene of martyrdom on the gate and the ensuing toponym 'Gate of Persecution' dateb ack onlyt ot he later seventeenth century,t he origins of this particularv iewing therefore probablyg ob ack much further in time.¹⁷⁰ Af inal example confirms once more whyagood recording of find circumstances and reconstruction of displaycontext are absolutelyessential in arriving at an interpretation of reidentification. Ajootian in an article of 2000 describes a partiallyp reserved Roman table leg in the form of Hermes Dionysophoros. The find location of the support in the debriso faFrankish structure at Corinth, together with its badlyw eathered right side, fuelled the hypothesis that it had been immured in an external wall of aroom associated with aFrankish funerary I. Jacobs, Old statues,n ew meanings chapel and thus exposed to the elements for quite some time.¹⁷¹ The religious nature of the buildingc omplex in which the statuary was reused, like at Sagalassos, prompted Ajootian to suggest aC hristian identity.M ores pecifically, the iconographyo fa na dultc arrying an infant,h eads close together,i nt he twelfth and thirteenthc entury was reminiscent of St.Christopher carrying Christ in the guise of ab aby.¹⁷² Athree-dimensional imageo fS t. Christopher,patron saint of travellers, would have been av ery aptd ecoration for ac hapel associated with what mayhaveb een ac omplex intended to offer accommodation to pilgrims.¹⁷³

Applying Christian identifiers?
So far,cross-marked statues have barelyfeatured in this article.¹⁷⁴ In conjunction with the application of an ew name label, one would expect thatar eidentification could also have been made permanent by am ore general marker of Christian identity.H owever,r eidentifications and cross-additions cannot simplyb e equated. The action of applying ac ross or multiple crosses to as tatue could have served quite different purposes, includingexorcism and general updating, bringings tatues into the Christian worlda nd ensuring their continued survival and display.¹⁷⁵ To establish that cross marking sealed reidentification, we need information on the displayand use contexts of cross-marked statues and reliefs, which for most examples we do not have.Onlyoccasionallywas it recorded that cross-marked statuary had been found near achurchb uilding, as in the case of the porphyry femaleheadatAbu Mina.¹⁷⁶ In afew other cases, reidentification is certainlyconceivable, for instance with cross-marked philosopher types. They include aportrait of Marcus Aurelius with the headcovered as velatus from Italica and as tele from Andros.¹⁷⁷ In the first case, ac ross has been carved rather crudelyo nt he left chest,i nt he second, an elaborate and carefullyc arved cross as well as the christogram ΙС ΧС and ΝΙΚΑ have been added behind a scene with as eated man teaching ay oung girl, not on the figures themselves, but almostasacertificate in the background. As described earlier in this article, the philosopher-teacher type could have served well as astand-in bodyfor Christ or as aint.Both the Marcus Aurelius statue and the teacher on the stele are visuallyq uite similar to the bearded Christs appearingo nf ourth-century sarcophagi. Yet, without more information on their find and displayc ontexts, all such suggestions remain highlyc onjectural.
In addition to crosses and sending am uch clearer message, halos could be added to reliefs to enable future viewers to identifyt he subject as Christian saints more easily.¹⁷⁸ Mangopointed to the halos added to some of the fragments of the Barbii monument framing the entrance door to the eleventh-centuryc athedral of S. Giusto at Trieste.¹⁷⁹ In this case, the church context makes aC hristian viewing undeniable. Late antique examples of added haloes must have existed, though once again we lack information on their displaycontexts. Astele in the Sparta ArchaeologicalM useumh ad ah alo carved around the head of the central figure.¹⁸⁰ Astanding male in military garb on afragment of asecond-century sarcophagus from Constantinople and now in the Louvrewas provided with ah alo at an unknown moment in time,t hus transforming him into am ilitary  Thus ap ortrait of ab earded man with al arge and elaborate cross carved on its forehead was found in  close to the fifth-century Basilica Ba tP hilippi (Kristensen,M iraculous bodies,a sf ootnote  above, cat.A ). The head of ay oung man with as mall and crude cross was found close to the Estauromenos church in Aigaleo in western Athens in the s (ibid., cat.A ).  Delivorrias,I nterpretatio Christiana (as footnote  above), pl. ,K ristensen,M iraculous bodies (as footnote  above), ,  -,c at.n o. B fig. .  Jensen,Understanding(as footnote  above), .The halo has avery longhistory that is not uniquelyconnected to Christianity, see M. Collinet-Guerin,Histoire du nimbe: des origines auxt empsm odernes.P aris ,  - for the appearance of halos beforeC hristianity,i n the Graeco-Roman and eastern traditions and  - for their appearance and usage in Christian visual culture.  Mango,A ntique statuary (as footnote  above), .  Delivorrias,I nterpretatio Christiana( as footnote  above), ,p l. . saint.¹⁸¹ An Apollo relief from Radheim in Germanyhad its/his genitalia removed and ah alo added,a gain at an unknownm oment.¹⁸² Nonetheless,ashas alreadybecome clear,such physical alterations werenot required for the allocation of an ew identity.I nt he literaryp assages mentioned earlier,there is no mentioning of the paintingsorstatues having been marked by crosses or altered in anyw ay.

Rediscovery of statues
Besides as pecific prompt and the absence of ap revious label identifying the statue, both of which have been discussed at length above, at hird important condition for facilitating the possibilityofreidentification is achangeincultural circumstances creatingdistance from the previous identity of astatue or relief. A new identity could thus be fashionedb yt aking the statue elsewhere, and creating geographic distance, or it could come into being because time elapsed between the previous identity of the statue and its new one. The statues taken by Constantine to Constantinople are an example of the first scenario. They mayh avef ulfilled ac ertain role in their original locationo fd isplayb ut were cut off from this by theirp hysical relocation to Constantinople. Chronological distance in particularw as apparentlyavital factor,a st he previous identity of the statue, the reasons for its dedication,and its significancecould all be forgotten duringt he time that it was not actively in use.¹⁸³ Chronological distance is sometimess ignposted in literary sources by the fact that the statuew as said to be 'rediscovered' by Christians. Fori nstance,P hilostorgius describes how the Christ statueatCaesareaPhilippi wasrediscovered after aperiod of neglect. The siteo fR ameses, whereE geria sees the statues of Moses and Aaron, had all but disappeared when she came across it.The bouleuterion at Sagalassos had been abandoned and was in ruins before the late fourth century when some  See also Rojas /Sergueenkova,T races( as footnote  above) for shifts in identifications throughout time, fromt he Bronze and Iron Aget oO ttoman travellers. of its reliefswerer eused in the new fortification walls.¹⁸⁴ At least 150 years had passed before the reliefs wereturned around to greet visitors to the church dedicated to the archangelMichael. The Triconch House at Aphrodisias is thought to have been abandoned around the middle of the seventh century and reoccupied onlyi nt he late eighth or earlyn inth century,a tw hich point the 'angels' in the peristyle courtyard werer ediscovered.¹⁸⁵ The builders of the fortifications on Ayasoluk Hill at Ephesus turned to funerary reliefs that weres ome 300 years old and possiblyc ame from ac emetery that had been abandoned.
With the later examples, such as the Frankish reuseo ft he table support in the form of Hermes Dionysophoros, but alsot he much better studied reused reliefs at Merbaka or the Little Metropolis Church at Athens, the effects of the passing of time are even more obvious.¹⁸⁶ Another illustration of providential 'rediscovery' comesf rom Rome: in 1551,at hird-century headless statuew as discovered by Pirro Ligorio at ac hurch site near the ViaT iburtina. It began a new life as ar epresentation of BishopH ippolytus of Rome, at hird-century theologian and martyr.¹⁸⁷ Originally, the statue depicted af emale -probablya seated muse. Presumably in the third century,the sides and backo ft he throne on which it was seatedwas used to inscribe Greek inscriptions: acalendar,acalculation of east,and, at the back, alist of books, some of them ascribed to Hippolytus. Consequently, when the statue was discovered, it was identified as Hippolytus, after which it was creatively altered and restored. Today, it is labelled 'Hippolytus' and is on displayi nt he entrance hall of the Vatican Apostolic Library.M odern scholars rarelyq uestion the history of the statue itself and use it uncriticallyasaconvenient illustration of the theologian.¹⁸⁸ Such rediscoveries continue into the late modern period. Forexample, in 1817apharmacist living in Athens dugupafunerary stele. It was again recognized as a 'holyimage' by the locals, who boughtt he item and installed it in the narthex of the Church of the Megali Panagia.¹⁸⁹ These instanceso fl ater rediscovery contrast stronglyw ith the situation encountered at the Parthenon however.Ifweassume that the metope on the northwest corner of the buildingw as indeed viewed as the Annunciation at least by some and thatt his reidentification took place alreadyi nL ate Antiquity, then there was little or no time to forgett he original identity of the protagonists, especially since Athensr emained acentrefor pagan intellectuals until the closing of its philosophical school in 529.¹⁹⁰ As alreadymentioned, competingidentifications werecommon and the Christian one maysimplyhavebeen pushed through in this case. It maye venb et hatt hese particular circumstances -the lack of chronological or geographical distance to ease an actualisation of identity and contemporaryr ival identities -led to the destruction of the other metopes. Whereas the reidentificationo ft he remainingr elief as an Annunciation has found relatively wide acceptance, we maygravely underestimate its importance in the history of the buildingand its potential to influencelate antiquedecision making.I ts preservation is generallyc onsidered an impromptu side-effect of a mutilation campaign targeting the other metopes.¹⁹¹ Iw ould like to suggest that the reidentification of the 'Annunciation' metope, which as by divine providence was placed at the most visiblecorner of the monument,was made apriori and prompted the destruction of the other reliefso nt he north, east and west sides, whilst leaving the much less visible souths ide. When it wasd ecided to convert the Parthenon into achurchdedicated to the Mother of God, local patriotism safeguarded the pediment sculptures,b ut the changei no wnership was propagated through the highlightingofone relief with ascene central to the worship of the Mother of God,which onlyw orked if it was not lost amidst other reliefs. In other words, the preservation of this one relief maynot have been coincidental, but ac learlyi ntended objective,w hereas the destruction of the Parthenon'sm etopes with theirf airlyi noffensive mythological subject matter was not the goal, but an ecessary corollary.  G. Despines, Ενα επιτύμβιο ανάγλυφο από τη Μακεδονίασ την Αθήνα. Egnatia  ( -),  -,s ee  -;A nderson,D efacement (as footnote  above), ,n ote .  E. J. Watts,City and school in latea ntique Athens and Alexandria. TheT ransformation of the Classical Heritage, .B erkeley/Los Angeles ,  -.  Either because it was decided there and then that it would be au seful Annunciation or, according to Anderson,D efacement (as footnote  above),  -,because it was too difficultt od eface it in the same thoroughm anner.

Conclusion
In manyc ities in the Eastern Mediterranean, classical statuarya nd reliefss urvivedu ntil the end of Antiquity and sometimes beyond. Rather thans imply remaining puzzled over continuity of classicaliconography, from time to time even integratedinto church contexts, and rather thanconstructingcomplicated narrativesofreligious triumphalism, Ihaveargued thatweneed to consider Christian reidentification as an additional explanation for the continued preservation of statuary.E vent hough examples discussed in this article will always remain hypothetical, the totality of literary,e pigraphic and archaeological evidence testifies to the fact thatC hristians at times used statues to give shape to their celebrities and heroes.
Assigning new identifies to statues was not purelyaChristian prerogative, but part of am uch wider and much older phenomenon whereby statues and other kindsofdepictions werecontinuously updated to changingcircumstances. The practice acquired an additional Christian facet during the reign of Constantine at the latest,with the emperorhimself apparentlyaparticipant in assigning both Christian and alternative secular identities. Christodorus' ekphrasis of the baths of Zeuxippos givesu sah int of how widespread the phenomenon was in Constantinople.I nt he newlye stablished capital of the Empire, there was ample opportunity for updated identifications, as statues wereb rought from elsewherei nt he empire often without their bases. Separated from theiroriginal culturalcircumstances and accompanying inscriptions, it became possibletoassign them new personalities and meaning, not limited to but certainlyincluding Christian ones. As opposed to newlyc arvedC hristian statues and reliefs, the phenomenon of Christian reidentification was not limited to an earlya ge of experimentation but continued throughout late antique and Byzantine times and far beyond.
Several conditions for reidentification have been identified in this article: the absenceofanidentifying label, geographical and/or chronological distance separating astatue from its original context of display, and the presenceofaspecific attribute or characteristic that became the prompt for reidentification. In some cases, aChristian identification was easy since the iconographyofcertain motifs, such as the Good Shepherd or Christ as Teacher,was so similar to its classical predecessors that identification depended mainlyo nt he religious conviction of the viewer and aC hristian physical context,i ncludinga ssociation with other figures and scenes.I no ther cases, aspects unbefitting the Christian identification wereo verlooked. Consequently, manyo ft he reidentified statues and reliefs discussed in this article weren ot textbook reflections of alreadye xisting and fixed iconographies,b ut creative reimaginations instigated by the 'recognition' of sometimesonlyone specific detail. In order to solidify the new identity of astatue and probablytoensure that it won out over other competingidentifications, an ew epigraphic label could be applied. Although the archaeological exampleso fr eidentified statues discussed in this article are not attested to in literary sources, they further testify to the three conditions for reidentification.
Apart from Constantine 'recognizing' as tatue of the archangelM ichael, explicit testimonies for who was responsible for Christian reidentifications do not exist.Iargued that local bishops playedas ignificant role in the process. Faced with (unidentified) statues, they either repurposed them by relatingt hem to familiar stories of the Christian past themselves, or,a tt he very least,t hey supported and helped spread such new identities, which became all the more convincing once they werec orroborated by church authorities.
These new identities, Christian and others, cannot simply be categorized as mistakes but need to be appreciated for what they were: creative manipulations and modernizations of the cityscape, executed by populations who continued to value the mediumo fs tatuary.