Bryn Mawr Classical Review 96.04.31

J. Elsner, Art and the Roman Viewer. The Transformation of Art from the Pagan World to Christianity. Cambridge: Cambridge University Press, 1995. Pp. xix + 375. $60.00. ISBN 0-521-45354-2.

Reviewed by Elizabeth Bartman, New York, NY,

A neighbor, newly arrived to the 1930 apartment building where I live, set about carpeting over the vintage floor tiles by her entrance. "Did you know," she asked, scandalized, "there are swastikas in the design?" Once an innocuous Greek pattern, the swastika cannot now be separated from its later history as a symbol of Nazi Fascism. As Elsner relates, a work of art (or an artistic motif) has different meanings for viewers living in different times and in different cultures. He has written an ambitious book attempting to understand the shift in the meaning of art during a pivotal period in western art, late antiquity, when abstraction replaced naturalism as the dominant artistic mode throughout the Mediterranean. Whereas others have explained the shift in terms of economics, ethnicity, religion, or artistic enfeeblement,1 Elsner points to a change in the prevailing mode of viewing as the impetus. Although recognizing the varied nature of viewing during both the ancient and early Christian periods, nonetheless he discerns an "ironic, deconstructive" mode in Roman art as opposed to an "initiate and exegetic" mode in Christian.

The book, parts of which have been published in a series of articles appearing since 1988, is divided essentially into two parts: Part One, "Ways of Viewing in the Roman World," aims to de-throne naturalism as the goal of ancient art by demonstrating contrary artistic currents. Chapter 1, which includes an extensive discussion of ekphrasis, contrasts the realistic mode of viewing embodied by the Imagines of Philostratus with the allegorical mode offered in a lesser-known text, Cebes' Tabula. Although offering antithetical versions of reality, both aim for aletheia. Chapter 2, devoted to the Roman house, investigates viewing in the domestic context. Elsner expounds his theory of "mystic viewing," with reference largely to the apse mosaics of the Monastery of St. Catherine at Mt. Sinai, in chapter 3; "mystic viewing" abandons the notion of a dichotomy between subject (viewer) and object, a "subjectivity" prevalent among both ancients and western moderns. Pausanias is the subject of chapter 4; arguing against the prevailing view of Pausanias as an antiquarian, Elsner presents him as a religious tourist, a second century precursor to the later Christian pilgrims.

Part Two, "The Transformation of Roman Art from Augustus to Justinian," details how specific monuments exemplify the viewing modes introduced in the first half of the book. Chapter 5 examines the image of the emperor and chapter 6, the representation of sacrifice (I offer lengthier criticisms of these sections below.) In a shorter third part of the book, Elsner considers art of the fourth century, primarily luxury silverplate and catacomb painting, where pagan and Christian imagery co-exist.

In his focus on viewing, Elsner plugs into the current trend in art historical analysis of considering visuality.2 For Elsner, art viewing raises basic cultural issues: how, he asks, "does the act of looking at art condition and create the identity of individuals?" (p. 19). Thus he is not concerned with viewing as it has been previously analyzed, whether from the physiological perspective of biologists, the psychological perspective of scholars such as Rudolf Arnheim or Ernst Gombrich, or the psychoanalytic perspective of Jacques Lacan, formulator of the influential concept of the "gaze."3 Nor is Elsner particularly concerned about who was actually doing the viewing. Audience, of long-standing concern for the Romanist who subscribes to Ranuccio Bianchi-Bandinelli's dichotomy between arte aulica and arte plebea, and more recently for those who raise the tri-color flag of gender, class, and ethnicity, barely figures in Elsner's discussion.

To construct the viewer upon which his notion of ancient viewing is based, he relies (primarily) upon three texts: the Imagines of Philostratus, the Tabula of Cebes, and the Description of Greece of Pausanias. The potentially skewed nature of this evidence -- all three imperial-age writers came from a strongly Greek milieu -- does not deter Elsner from developing a thesis he applies to many centuries. Nor does the reliance upon literary texts to open a window onto the world of seeing; Elsner has a phrase for this, the "rhetorisation of a view" (from Foucault's "nomination of the visible"), but the active presence of literary conventions in shaping these texts would seem to present problems for those who would see them as straightforward commentaries on viewing.4 There are other evidentiary problems as well; in the context of examining such themes as imperial representation or sacrifice, Elsner shapes his argument around the purported contrasts between monuments that he has selected as indicative of either the pagan or Christian mode of viewing. His rationale is not always clear; why these particular monuments and not others? In his discussion of imperial representation, he contrasts the first century BCE Augustus from Prima Porta with a late third century CE painted cycle featuring Diocletian from the Temple of Luxor and finally, with the sixth century mosaic cycle from San Vitale in Ravenna. Why such a wide chronological gap? Would a representation of the emperor from the Antonine period of the mid-second century, a time when abstraction first appeared in public official art,5 alter Elsner's arguments? And why these monuments? Would substituting the famous porphyry ensemble portrait of the four Tetrarchs, embedded in the facade of San Marco in Venice, project the same impression of the emperor as the Luxor murals, a much-damaged painting cycle presented by Elsner as illustrative of the dual human and divine status of the emperor? Virtually all of Elsner's visual arguments for the exegetic mode in early Christian art are mosaic cycles from major churches of the sixth century. Not only are these relatively late in terms of historical development, but they also represent one particular (admittedly a public and prestigious) art form. One wonders if other early Christian artistic media regarded as closely allied to their pagan predecessors -- for example, carved sarcophagi or gold glass6 -- would elicit for Elsner such marked contrasts in viewing. One suspects that Elsner advances his argument on the basis of various "straw men," works located at extreme positions, chronologically, artistically, or ideologically, whose extremism serves to bolster his position.

Elsner understands viewing as a reflexive process. The viewer who sees an object fits it into the framework of what he knows, but that framework is itself transformed by the viewer's experience of the object. In the tradition of Roland Barthes' reader, Elsner's viewer plays an active role in formulating the meaning of the object, although Elsner emphasizes that the artist "forecloses the potentially infinite number of subjective contextualisations that a viewer might choose" by his "choice of subject and form" (p. 39). On a broader level, artistic style depends upon the prevailing mode of viewing -- "... naturalism or abstraction ... is dependent on a great many conceptual, sociological and essentially historical factors rooted in the way art is viewed at particular times" (p. 13). Fundamental for Elsner's argument, such a relationship between object and viewing is nonetheless highly problematic. Does a work of art instruct the viewer in a mode of looking or does it (ultimately) respond to the looking mode? Elsner prioritizes viewing, thus laying the basis for his explanation of the late antique stylistic shift from naturalism to abstraction. But others might construct a different scenario. Consider the modern age, for example, when modes of viewing were themselves transformed by the invention of photography and then of film. Modes of viewing, it can be argued, are responsive rather than directive; thus their elevation in Elsner's scheme is like putting the proverbial cart before the horse.

Even if one were to accept Elsner's contentious notion of viewing, his readings of the particular monuments used to establish the historicity of different viewing modes are so erroneous and misconceived as to undermine his arguments. I will focus on three of Elsner's Roman subjects: the house (Chap. 2), the statue of Augustus from Prima Porta (Chap. 5) and the Ara Pacis Augustae (Chap. 6). In each case Elsner exaggerates certain aspects of the individual monument (or, in the case of the house, the entire architectural type) to build his thesis. His treatment of the house is especially dubious. Attempting to find irony in the early Roman period, he interprets the space and painted decoration of the Roman house as socially and politically subversive, a "licensed transgression" akin to the carnival as conceived by Bahktin (pp. 49, 74). In view of the great expense and labor involved in decorating a house, it is difficult to imagine why its owner would "deconstruct" the very domestic world he has created. While many Roman houses clearly broke down spatial conventions of interior and exterior through views, window, and trompe l'oeil painting, it is hard to see how this illusionism acquires the political implications ascribed by Elsner. In the authoritarian world Elsner envisions, the architectural theorist Vitruvius is Augustus' drill sergeant, writing impassioned invective against decorative forms that threatened the established social order. As the Augustan principate itself overturned the long-established social hierarchy of the Republic, the new regime would have had no vested interest in maintaining an architectural status quo in the domestic realm. Nor does the princeps himself appear to have followed Vitruvius' prescriptions: his Palatine house had at least one room decorated in the condemned third style of painting while its generally modest scale and fixtures belied its owner's social prominence. Houses certainly mattered a great deal to the Romans -- damnatio memoriae included the destruction of a man's house -- and Elsner is correct that "there is nothing natural about how a house is designed or decorated," but as a general rule neither its spatial layout nor its decoration had so ideological a role at Rome as Elsner posits.

Elsner sees instability as well in the concept of the Roman emperor, notably as he is embodied in the statue of Augustus from Prima Porta, now in the Vatican Museum. Because it conflates divine and mortal in its iconography, the statue, in Elsner's view, leaves unresolved the question of whether Augustus was man or god. (Later monuments such as the paintings from the room devoted to the imperial cult at Luxor and Christian mosaics succeed in eliminating this tension; at Luxor paintings depicting Diocletian as a god are physically separate from those showing him in his more human dimension, while Christianity believed Christ to be "God who sanctified man by his descent into humanity" (p. 177). While indeed their respective "followers" may have conceived of Augustus and Christ in different terms, the Prima Porta statue of Augustus is by no means an ideal representative for the Roman imperial cult. Whatever its origins -- the Vatican's marble has been argued by some to copy a bronze statue set up publicly elsewhere -- it was certainly not a cult statue. (Were the statue to have had a cult, Elsner writes, "its ambiguous status as a naturalist invitation to narrative would be in conflict with its ritual function as cult icon" (p. 186). Nor did the Romans conceive of Augustus, even when divinized posthumously, as a god on a par with the others in the Olympic pantheon. Once deified, an emperor became divus, not deus; as historians of religion have argued, these were not the same.7 Elsner makes much of the fact that Augustus was deemed a god simply by virtue of his position as emperor. Of course in large parts of the Empire, particularly the Greek East, many subjects did regard the living Augustus as a god, but few at Rome would have subscribed to this belief before his death and divinization in 14 CE. Nor were all emperors automatically gods; Tiberius, Augustus' successor, for one, was never designated a god by the Romans.

Focusing on the imagery of the statue's breastplate, Elsner contrasts its "narrative" with that of the Byzantine mosaics of San Vitale at Ravenna; the viewer of the statue, he writes, "must construct his own contextualising narrative from the clues of the breastplate" whereas at Ravenna the viewer has no such need because, as an initiate, his own actions mirror those of the figures depicted in the mosaics above. With little knowledge of the circumstances that surrounded the statue's creation -- how the return of the Parthian standards shown on the breastplate was celebrated at Rome and how the story was told through text and other images such as painted banners -- it is unclear today just how "narratologically" adrift the Roman viewer of the Prima Porta statue would have been. For a ritual parallel in which the viewer's action and belief parallels that of the depicted object, Elsner need only have considered the Ara Pacis, a prominent subject of the next chapter.

Elsner discusses the Ara Pacis in the course of tracing the changed depiction (and conception) of religious sacrifice from antiquity to the Christian era. Like the wall painting featured in chapter 2, Roman scenes of sacrifice are deconstructive, casting an "ambivalent reflection on the activity ... which they celebrate" (p. 209). How this is so is never fully explained, but Elsner notes that the actual sacrifice itself is not shown in the reliefs but is deferred ("Thus the main theme and end of the ritual occasion depicted is always an absent goal." [p. 199]). Large sections of the altar's inner frieze, which seem to have depicted further details of the sacrificial ritual, are of course missing today, but it is doubtful that the culminating sacrifice of white heifers was actually portrayed on the altar. Not only was it gruesome in its bloody subject, but the monument itself was not nearly so concerned with sacrifice as Elsner would have us believe. To be sure, the monument is an altar, but it probably hosted sacrifices only a few days a year, perhaps only on the anniversaries of the altar's constitutio and dedication.8 At other times the many viewers who passed it on the Via Flaminia may have focused on other aspects of its polysemic imagery -- Augustan peace, the Golden Age, dynasty, and monarchy are but a few of the themes that Elsner subordinates to sacrifice. Certain decorative motifs such as the carved paterae and bucrania do recall the animal sacrifice made annually at the altar, but their use in a range of non-sacrificial contexts must have diluted any such meaning for contemporary viewers.9 That they are confined to the altar's interior, an area probably off-limits to all but priests and attendants, also underscores the secondary role sacrifice played in the altar's design and usage. Within the Roman tradition of sacrificial depictions, the altar's development of this theme is decidedly limited.10

Flawed in construct and based on questionable interpretations of the individual monuments used to illustrate his thesis, Elsner's attempt to ground a pivotal formal shift in western art in changes in prevailing modes of viewing is ultimately untenable. Elsner does succeed in showing how ancient Romans perceived art and their world in multiple, sometimes contradictory, modes. Compared with art historians whose work centers on the post-antique, those who specialize in ancient art have been slow to embrace new methodologies suggested by related fields of history, linguistics, and anthropology. Elsner's boldness lies in willingness to apply new approaches to the visual evidence from antiquity. Although he has not provided convincing answers to the questions he poses, he has given students of ancient art an arsenal of new tools for the investigations they may undertake regarding the nature and meaning of ancient art and how it was seen by contemporary viewers.


  • [1] Elsner's desire to challenge the long-prevailing view of late antiquity as a period of artistic "decline" (p. 13) would seem to be unnecessary, as no serious observer in recent decades has characterized the period in this way. Elsner cites only the sixteenth century Vasari, important historiographically but hardly an opinion-maker on the subject of Roman art today.
  • [2] See for example, Vision and Visuality, ed. H. Foster (Seattle 1988).
  • [3] J. Lacan, The Four Fundamental Concepts of Psycho-analysis, ed. J. A. Miller, trans. A. Sheridan (New York and London 1978) sec. 6-9.
  • [4] Pausanias' Description would seem to be especially problematic; as Elsner himself notes, Pausanias could not have seen some of the things he discusses (e.g. a destroyed wooden cult image described "from hearsay" [p. 130]). While Pausanias' inclusion of such monuments may indeed reflect a value system or world view, it can hardly be illustrative of his "mode of viewing."
  • [5] G. Rodenwaldt, "Über den Stilwandel in der antoninischen Kunst," Abhandlungen der preussischen Akademie der Wissenschaften. Phil.-hist. Klasse 3 (1935) 1-27.
  • [6] Elsner does devote a section of the epilogue to a discussion of luxury silverplate, a medium where there is strong continuity between pagan and Christian. In this context, he admits, the transformation of Roman art is reflected "iconographically" (p. 261) rather than in terms of altered viewing. While indeed the Communion scene on the Rhia paten reflects a Christian world in contrast to the Boscoreale cup's triumph of Tiberius, in a number of instances the only clue we have to Christian ownership of a piece is an inscription.
  • [7] The entire issue of Roman religious belief, of course, is highly contentious: it has long been suggested that educated Romans put little stock in organized religion as practiced on the state level, but instead put their faith in more personal, family-oriented cults such as those devoted to the household gods (lares and penates) and ancestors.
  • [8] P. Holliday, "Time, History, and Ritual on the Ara Pacis Augustae," Art Bulletin 72 (1990) 544.
  • [9] The motif is found on the exterior of the tomb of Cecilia Metella, for example, a monument that would not have received sacrifices of this type. Regarding the tomb Erika Simon writes, "Friese mit Bukranien ... und Girlanden waren in republikanischer Zeit sehr verbreitet." (Augustus. Kunst und Leben in Rom um die Zeitenwende [Munich 1986] 248 no. 217).
  • [10] No animals, no altar, and no ritual implements specific to the sacrifice are shown on the Altar's exterior. For the tradition of sacrificial representation see I. Scott Ryberg, "Rites of the State Religion in Roman Art", MAAR 22 (1955) esp. 104-19.