Walking with Odysseus: The Portico Frame of the Odyssey Landscapes

this article examines the cultural and artistic context of one of the most famous roman frescoes, the Odyssey landscapes. it argues that the painting’s fictive portico frame would have evoked in the roman viewer the experience of the ambulatio, the act of walking for leisure and contemplation that came to be an essential element of a properly hellenized otium. the painted portico thus puts the viewers in the proper frame of mind to appreciate the intellectual associations of the painting as they walk with Odysseus on a parallel journey of philosophical reflection. media vero spatia, quae erunt subdiu inter porticus, adornanda viridibus videntur, quod hypaethroe ambulationes habent magnam salubritatem. et primum oculorum, quod ex viridibus subtilis et extenuatus aer propter motionem corporis influens perlimat speciem et ita auferens ex oculis umorem crassum, aciem tenuem et acutam speciem relinquit; praeterea, cum corpus motionibus in ambulatione calescat, umores ex membris aer exsugendo inminuit plenitates extenuatque dissipando quod plus inest quam corpus potest sustinere. (Vitruvius De architectura 5.9.5) the spaces in the middle, between the porticoes and open to the sky, should be decorated with greenery, since walks in the open air bring great health benefits. First of all, because of the movement of the body, the fine and rarefied air from the greenery flows into the eyes and sharpens the sight; in so doing it removes the heavy moisture from the eyes and leaves behind clear vision and sharp sight. moreover, since the body heats up with movement during a walk, the air reduces saturation by sucking moisture out of the limbs, and thins them out by dissipating whatever is more than the body can sustain.1 american Journal of Philology 128 (2007) 497–532 © 2007 by the Johns hopkins university Press 1 except where noted, all translations are my own. 498 timOthy m. O’sulliVan We do noT knoW hoW many of viTruviuS’ conTemporarieS would have characterized the benefits of walking with such scientific precision, but we can be sure of the popularity of walking as a leisure activity in late republican and early imperial rome. One sign of that popularity is the sheer number of ambulatory spaces that decorated the ancient city, particularly the many public porticoes constructed from the second century b.c.e. onwards.2 these porticoes were multifunctional, but their appeal was in part due to the fact that they gave the roman people a taste of the aristocratic lifestyle; as they strolled in a beautiful setting full of works of art and elaborate landscaping, they enjoyed precisely the experience that wealthy romans of the late republic cultivated in their villas and urban mansions.3 in the roman domestic setting, the leisurely stroll had a decidedly intellectual flavor, exercising both body and mind.4 the intellectual associations of ambulatory spaces in the villa even inspired roman aristocrats to name parts of the property for famous monuments around the empire and decorate them accordingly; 2 Vitruvius’ commendation of walking, for example, appears in his section on theater colonnades, which he surely wrote with the portico of Pompey in mind. On the roman public portico, see richardson 1992, s.v. “Porticus”; senseney 2002. see also Velleius Paterculus, who begins his second book with the familiar topos of roman decline in the second century b.c.e. and uses the importation of public porticoes as a symbol of that decline (2.1.1–2): “in somnum a vigiliis, ab armis ad voluptates, a negotiis in otium conversa civitas. tum scipio nasica in capitolio porticus, tum quas praediximus metellus, tum in circo cn. Octavius multo amoenissimam moliti sunt, publicamque magnificentiam secuta privata luxuria est” (“Vigilance turned to slumber, weapons turned to pleasures, business became leisure: our society was turned upside-down. it was then that scipio nasica built porticoes on the capitoline, and metellus built the porticoes we mentioned above; it was then that gnaeus Octavius built a portico that was the fanciest by far, and private luxury followed public splendor”). 3 in addition to the Velleius quote above, the connection between public porticoes and private leisure is made most explicit by augustus’ decision to replace the vast domus of Vedius Pollio, bequeathed to the emperor, with the Portico of livia; Ovid attests to the didactic value of the portico (Fast. 6.639–40, 647–48): “disce tamen, veniens aetas: ubi livia nunc est / porticus, immensae tecta fuere domus. / . . . sic agitur censura et sic exempla parantur, / cum iudex, alios quod monet, ipse facit” (“But know this, future ages: where livia’s portico now stands were once the buildings of an enormous house . . . . this is how a censor acts, this is how to set an example, when the authority himself carries out the advice he has imparted to others”). similarly, macdonald and Pinto 1995, 54, suggest that “the colonnaded paths leading from one building or vantage point to another” at hadrian’s Villa are “the Villa equivalents of urban porticoes.” 4 cf. Pliny’s comment on spurinna’s rigorous life of leisure (Ep. 3.1.4): “ambulat milia passuum tria nec minus animum quam corpus exercet” (“he walks three thousand paces, and exercises his mind no less than his body”). 499 the POrticO Frame OF the Odyssey landscaPes the educational benefits of travel abroad were imported to the private setting, and a walk through one’s own philosophical playground became, in effect, a kind of metaphorical tourism.5 my aim in this article is to show how an awareness of the particular power of this cultural metaphor—walking as a sign of intellectual leisure—can alter our perception of one of the most studied artifacts from antiquity: the roman fresco known as the Odyssey landscapes (plates 1–2; figs. 1–3).6 unlike contemporary paintings of the mid-first century b.c.e., in which either figural or architectural subjects dominate, the fresco is famous for its emphasis on natural forms; rocks, cliffs, crags, bays, caves, trees, islands, and ocean all compete for the viewer’s attention.7 in the midst of this dramatic natural setting are miniature figures enacting scenes from the homeric Odyssey: specifically, from that part of the Odyssey narrated by Odysseus himself, including his encounter with the laestrygonians and with circe, the nekuia, and the voyage past the sirens. the mural’s modern designation as the “Odyssey landscapes” perfectly captures the interplay of myth and nature that is its most famous feature; in addition to the human figures and the natural setting, however, the wall decoration has a third element, a fact that is sometimes overlooked both in scholarship and in reproductions of the work. Framing the landscape is a series of painted red pillars, evenly spaced, creating the illusion that the viewer must look through a colonnade in order to see 5 cicero’s tusculan academy (e.g., Att. 1.4.3) is the most famous example of this trend; see neudecker 1988, 11–14. On the practice of naming parts of the villa for foreign places and monuments, see görler 1990. On metaphorical tourism in the roman villa, see Bergmann 2001 and O’sullivan 2006. the connection between walking and traveling is also seen in the metaphorical use of perambulare to mean “to travel”; see TLL 10.1.1185.66–72, 1186.23–48 (schmitz). 6 the painting is currently housed in the sala delle nozze aldobrandine at the Vatican, save for a fragmentary panel (fig. 3) in the Palazzo massimo. the bibliography on the painting is vast. Biering 1995 is the most complete treatment; von Blanckenhagen 1963 is still a useful overview, although his central thesis—that the painting is a copy of a hellenistic original—is not universally accepted (further discussion below). coarelli (1998) has more recently provided a thorough reassessment of the painting’s original setting. 7 the dating is newly controversial thanks to Biering 1995, 181–90, who argues on stylistic grounds that the fresco should be dated to the last decade of the first century b.c.e. at least two scholars have already rejected the new date, however: see coarelli 1998, 26–30, and tybout 2001, 35–36. coarelli’s argument is largely based on the fact that fragments of a pre-Julian fasti were discovered at the site, though it is not clear why the landscapes and the fasti were necessarily painted at the same time. 500 timOthy m. O’sulliVan the mythical landscape “beyond” it (fig. 4).8 these pillars serve to divide the frieze into individual units more conducive for viewing. By its very compositional technique, the mural encourages the viewer to look past this architectural frame; yet to do so ignores the fact that these pillars are as much a representation as the landscape “behind” them, and are therefore as worthy of interpretation as the rest of the painting. in what follows, i argue that the portico frame of the Odyssey landscapes serves not only a narrative function but also an interpretive one; the portico would have evoked in the roman viewer the experience of the ambulatio, the act of walking for leisure and contemplation that came to be an essential element of a properly hellenized otium. the aristocratic stroll was an advertisement of intellectual refinement, and it even became an embodied performance of the mental theoria espoused by Plato and his followers as the model for philosophical inquiry.9 as a result, the painted portico puts the viewers in the proper frame of mind to appreciate the intellectual associations of the Odyssey landscapes as they walk with Odysseus on a parallel journey to greater insight. i. the Odyssey landscaPes: an OVerVieW the Odyssey landscapes were hailed as a landmark of greco-roman art immediately upon their discovery in rome in 1848.10 the construction of low-income housing along the Via graziosa on the esquiline came to a standstill when workers discovered a segment of opus reticulatum; subsequent exploration revealed panels 2 and 3 of the Odyssey landscapes (plate 1).11 excavation of the rest of the fresco was then interrupted by 8 throughout this article i will use the term “pillar” for the painted architectural supports, since they are rectangular in cross-section. in so doing, i follow the terminology suggested by ginouvès 1992, 63, who reserves “pillar” for independent supports which are square or rectangular in cross-section, and “pier” for massive or irregular supports; in practice, however, the two terms are often reversed, as he himself notes. 9 On the history of the metaphorical appropriation of theoria by greek philosophers, see nightingale 2004. For the connection between the roman ambulatio and philosophical theoria, see O’sullivan 2006. 10 Braun’s excitement is typical (1849, 27): “Zwischen [den Pilastern] aber entfalten sich landschaftliche scenen von einer großartigkeit und Originalität des Vortrags, wie er kaum sonst wo vorkommt.” 11 although technically inaccurate, “panel” is still a useful way to refer to the individual scenes that appear between the pillars of the painted portico. the first panel was already destroyed when the painting was discovered and left in situ, so by convention the first extant panel (the initial encounter between Odysseus’ men and the laestrygonian 501 the POrticO Frame OF the Odyssey landscaPes disputes both large and small: that year, the city was paralyzed by civil unrest that led to the short-lived repubblica romana (1848–49), while on the Via graziosa, sig. Filippo Bennicelli, who owned the house under which the remaining panels of the Odyssey landscapes continued, managed to hold up excavation by demanding more money from the state.12 emil Braun, secretary of the deutsches archäologisches institut in rome, lamented the fact that a painting of such importance was left to languish in the elements for almost a year, but further damage was inflicted by the archeologists themselves, who ordered restorations of the fresco during and after its excavation.13 a summary of the original circumstances of the excavation is important, since the findspot was buried, along with the Via graziosa itself, by the construction of the Via cavour in the 1880s. moreover, any original excavation reports have been lost, and we are left with the rather unsystematic account of Pietro matranga, a hellenist at the Vatican library who wrote the first monograph on the painting.14 Our knowledge of the original context of the work is therefore somewhat limited. We cannot be sure, for example, what type of room the fresco decorated (atrium, peristyle, and cryptoporticus have all been suggested), nor do we know for certain whether or not the excavated building was a domus at all, though this is usually assumed. matranga actually argued that the structure was the Portico of livia (1852, 103–37), but the location of the portico on the marble plan does not support his thesis. more recently, kuttner (2003, 112) has pointed out that the paintings were discovered in an area that boasted numerous suburban horti, while coarelli 1998 insists that the paintings princess; see plate 1) is referred to as the second (as in von Blanckenhagen 1963 and Biering 1995, among others). 12 see henzen 1850, 167: “der Besitzer, herr Bennicelli, ist mit der alterthumscommission in streit über den ihm zu bewilligenden Preis, und mittlerweile bleiben die gemälde in ihrem dunklen stollen allen Zufälligkeiten ausgesetzt.” For an overview of the available evidence for the discovery and excavation of the Odyssey landscapes, see Biering 1995, 167–80, 195–96. 13 the most damaging restoration was the application of a fixative that altered the original colors; a second round of restorations in 1952–56 removed the fixative but also some of the original painting in panels 4 and 5. see Biering 1995, 14–19. 14 matranga 1852. the bulk of the monograph is only tangentially related to the painting: the first chapter (pp. 1–53) uses the landscape paintings to argue that the laestrygonians inhabited the area around modern terracina, ca. 90 km se of rome, while the second chapter (pp. 55–100) proposes that Od. 7.104 refers to the carding of wool and not to the grinding of grain. the two appendices (pp. 103–48) focus more specifically on the painting and the details of its discovery and excavation. 502 timOthy m. O’sulliVan decorated a grand domus and that the front entryway and atrium of this house are identifiable on the marble plan. the fragment of the marble plan breaks off well before the site of the 1849 excavation, however, and the domus in question is not on the same axis as the Odyssey landscapes wall, so coarelli’s identification is hardly conclusive.15 similarly, there is some confusion in the scholarship about the placement of the frieze on the wall itself. most accounts claim that the frieze was situated high on the wall, though as Biering (1995, 169–70) has convincingly argued, the notion that we have an exact measurement is based on a misreading. matranga (1852, 112, n. 12) gives a measurement from the floor to the beginning of a vault (5.5 meters), which some have used (see, e.g., coarelli 1998, 31) to place the bottom of the frieze about 4 meters high. But, as Biering points out, there is no suggestion in Vespignani’s drawings that the ancient floor was excavated, nor is there any indication of a vaulted ceiling; Biering argues that matranga has misidentified as a vault the beginning of an arch much higher on the wall and that the measurement reflects the depth of the excavation, not the height of the paintings. still, the open vista on the upper part of the wall is a relatively common feature of the second style, and it is reasonable to assume that the frieze was somewhere above eye level.16 the incomplete state of our knowledge has even been exaggerated by scholars: there has been some debate, for example, over whether or not the frieze would have continued on more than the one excavated wall, even though matranga explicitly states that there were panels on the preceding wall that were left unexcavated because they were in such a state of disrepair (fig. 5).17 nonetheless, we still do not know what per15 moreover, as one of the readers for AJP correctly points out, there is still no reason to rule out the possibility that the painting decorated a public building. in this article, i follow the scholarly consensus by assuming that the painting was part of a domestic ensemble, but my argument that the painted portico evokes the idea of walking, and its many intellectual associations in late republican rome, could apply just as easily to public or private decoration. 16 For the open upper part as a feature of the second style, see engemann 1967, 39–57. 17 matranga 1852, 109. One of Vespignani’s drawings (matranga 1852, table 10; see fig. 5) clearly shows the outline of the portico frame on the wall at right angles to the wall with the Odyssey landscapes. andreae has argued in a number of publications (e.g., 1962, 108; 1988, 282) that the first panel must have depicted the Polyphemus episode, on the grounds that it could not have been omitted; since, however, there was another frieze preceding the destroyed panel, the argument does not hold. Biering 1995, 172–74, has tentatively suggested that the fragmentary sirens panel (fig. 3) was originally on a third wall, at right angles to the extant frieze. 503 the POrticO Frame OF the Odyssey landscaPes centage of the original frieze we possess or what the wall looked like beneath the landscapes. yet the central scholarly controversy surrounding the work centers not on any of these unknowns but on the one aspect of the painting that would seem to be beyond dispute: despite the fact that the Odyssey landscapes are one of the relatively few examples of ancient wall painting unearthed in the city of rome itself, scholars have questioned whether the work is really “roman.” Peter von Blanckenhagen argued at great length in an article of 1963 that the fresco was a copy of a now-lost hellenistic original, and the painting has long enjoyed a kind of dual citizenship, serving as a dependable highlight in handbooks on both roman and hellenistic art.18 the temptation to treat the mural as evidence of an elaborate (but almost entirely lost) tradition of hellenistic landscape painting has been too much for many scholars to resist. For one thing, much of roman wall painting, particularly that of the First and second style, has direct or indirect antecedents in greek painting.19 Furthermore, the development of the bucolic genre by hellenistic poets such as theocritus and moschus cries out for the visual analog that a tradition of hellenistic landscape painting would provide.20 the discovery of the frieze of the hunt on the so-called tomb of Philip at Vergina in 1977 has 18 Von Blanckenhagen 1963 was not the first to make this general point, though his is the most complete development of the argument. his article, in addition to being an excellent overview of the painting and its problems, also has a superb summary of the scholarship on the disputed origin of the painting. For a prominent placement of the painting in a work on hellenistic art, see Pollitt 1986, 185–209, esp. 185–86, and 208–9, though ironically his is among the most balanced treatments of its origin and one to which i am quite sympathetic: “to relegate the Odyssey landscapes to the status of copies of some hypothetical greek work of the second century B.c. when all the extant evidence suggests that no such work existed is perverse. On the other hand to deny that the Odyssey paintings represent the natural culmination of a developing tradition of hellenistic landscape painting seems obtuse. they brought hellenistic technique into a world increasingly dominated by roman taste and are a product of the fusion of the two cultures” (208). cf. leach 1988, 41: “[i]n the study of roman painting one must accept the proposition that originality does not preclude a blending of roman and greek.” 19 the First style is essentially a hellenistic style: Bruno 1969; mielsch 2001, 21–25; Baldassarre, et al. 2002, 67–70. tybout 1989, 109–86, argues that what we consider the second style is only the romano-campanian instantiation of a broader hellenistic koine. 20 g. Zanker 2004 argues for parallels between the Odyssey landscapes and moschus’ Europa (50–51) and (pseudo-?) theocritus 25 (89–91), although he does not claim that the paintings are a copy of a hellenistic original. Further complicating the question of the origin of the Odyssey landscapes is that many of the characters in the painting are labeled in greek; rather than proving the existence of a greek original, however, the labels are simply another way in which the painting advertises its “greekness,” as Beard and henderson 2001, 54, aptly note. On the use of greek labels in roman wall painting, see thomas 1995. 504 timOthy m. O’sulliVan done little to settle the debate: the natural setting of the hunt suggests that macedonian painters were interested in landscape as more than a mere backdrop to human action, yet the frieze could hardly prevent us from continuing to call the Odyssey landscapes the first extant example of full-fledged landscape painting in the Western tradition.21 as we shall see, the portico frame has been marshaled as evidence in the debate over the true “nationality” of the Odyssey landscapes, and i shall have more to say on this issue later. suffice it to say for now that, whatever the antecedents of our fresco may have been, its setting requires us to consider the painting in its roman context as it was enjoyed by roman viewers steeped in greek culture.22 ii. the POrticO Frame despite the scholarly controversy surrounding the painting’s place of birth, critics readily agree that the painted frame of red pillars was a roman idea. yet for the most part, and particularly for those scholars who argue that the Odyssey landscapes are a copy of a hellenistic original, the identification of the frame as a roman contribution to the painting has acted as a sort of back-handed compliment, so that the addition of the pillars becomes one of the roman innovations that has done a disservice to the greek “original.”23 numerous reproductions of the mural do away with the pillars altogether, as though by doing so we might have direct access to a lost masterpiece of hellenistic art.24 indeed, the very fact that 21 On the tomb and its frieze (with numerous photographs and line drawings), see andronicos 1984, 97–119. For a similar assessment of the relationship between the landscape of the Vergina frieze and the Odyssey landscapes, see small 1999, 568. 22 as Wallace-hadrill 1998, 91, puts it in a different context, “to be roman, go greek.” For other analyses of aristocratic romans playing greek in the privacy of their own homes, see neudecker 1988, 8–30, and Bergmann 2001. the problem of greek originals and roman recreations has always plagued roman art and is becoming a focal point again as an instance of roman reception of greek culture; see the three art historical essays in the special volume of Harvard Studies in Classical Philology entitled “greece in rome” (Bergmann 1995; gazda 1995; kuttner 1995) and the edited volume by gazda 2002 on emulation. 23 Von Blanckenhagen’s reaction is typical of the critical assessment (1963, 106): “the peculiar and certainly exaggerated fashion of having frames cutting off parts of figures and objects must be explained, all the more because, to my knowledge, this form of ‘illusionism’ is unique in ancient painting.” 24 lydakis 2004, 200–201, goes so far as to reconstruct a hellenistic “original” by putting panels 3 to 5 side by side, without the framing pillars. 505 the POrticO Frame OF the Odyssey landscaPes we commonly refer to the painting with the plural “landscapes” (and the individual scenes as “panels” or “pictures”) leaves the false impression that the portico frame is not part of the original painting, an impression that is further reinforced by the current arrangement of the fresco in four separate frames (with two panels each) at the Vatican (e.g., fig. 2). yet no matter how we understand the origin of the painting, the red pillars deserve consideration in and of themselves, if only because they are a fundamental part of the painting’s organization. a hypothetical reconstruction of the frieze as originally arranged makes it clear that the colonnade would have made an immediate impression on the viewer—if not as striking as the landscapes themselves, nonetheless difficult to ignore (fig. 4). although the pillars at first glance appear to constitute a single row, closer inspection reveals a second row of pillars in the shadows that is partially occluded by the first row, with both rows supporting an architrave and a roof (fig. 6).25 the form, then, is that of a portico (greek stoa, latin porticus): a covered colonnade that was ubiquitous in greek and roman architecture both as a framing device for other spaces (gardens, temple precincts, streets) and as an independent structure.26 in the first four panels (plates 1–2), the back pillar can be seen to the right of the front pillar; starting with the second pillar of panel 6 (fig. 1; cf. fig. 4), however, the back pillar appears to the left of the front pillar. the effect, then, is a trick of perspective: the imagined viewing position of the extant frieze is in front of the sixth panel (that is, at the palace of circe), which is marked as central by the placement of the rear pillars.27 as we 25 it is not clear whether we are supposed to imagine the second row as freestanding pillars or engaged pilasters along a back wall; in the case of the latter, some of the observations below about figures intentionally “disappearing” behind the pillars would not correspond to the portico as “really” painted, but i would argue that we should not press the realism of the architectural reconstruction too far. 26 coulton 1976 offers a history of the greek stoa, although he is primarily interested in free-standing stoai, not the four-sided peristyles that develop in the hellenistic period to frame open spaces; on the porticus, see macdonald 1986, 33–66, and anderson 1997, 247–50. 27 roman wall painting ensembles of the second style usually establish a viewing point (sometimes reflected in the floor decoration) from which the wall is meant to be seen; see scagliarini corlàita 1974–76, 9–10, and clarke 1991, 41–45. the differing perspectives of the portico frame (to be viewed from below) and the landscape itself (to be viewed from above) has vexed interpreters of the painting and fuels von Blanckenhagen’s hypothesis (1963, 111–12) that the landscape was not originally designed for this setting. as Biering 1995, 160, points out, however, the landscape itself shows little internal consistency of perspective (he compares the size of Odysseus’ ships [plate 2] and circe’s palace [fig. 1]), so the expectation of a consistent perspective shared by the landscapes and the portico frame should not be 506 timOthy m. O’sulliVan shall see further on, this centrality is reinforced by the special status of the sixth panel itself, which distinguishes itself from its adjacent panels both in composition and in narrative technique. it seems clear then that the painted frame represents a portico, but what kind of portico is this? Would the ancient viewer recognize the form? the question is not an idle one. Vitruvius famously criticized contemporary trends in roman wall painting for a lack of realism, lamenting the depiction of fantastical architectural features that could not have existed in the real world.28 did a portico such as this, with thin red pillars and golden capitals, exist in the ancient world? it is not immediately apparent what material these red pillars are supposed to represent: are they painted wood, painted stucco, or some sort of fantastically colored marble, like the surreal colors boasted by the panels of the late First style of Pompeian wall painting?29 although it is true that porticoes typically employed cylindrical columns and not pillars for support, there are exceptions. the so-called hall of the doric Pillars at hadrian’s villa has monumental marble pillars for supports.30 On a somewhat humbler scale, the Praedia of Julia Felix at Pompeii boasts a two-sided colonnade with slender pillars that are very similar in impression and scale to the narrow pillars of the Odyssey landscapes (fig. 7).31 similarly, although the greek stoa almost always featured columns, marble pillars were used on occasion; considering the possibility that the Odyssey landscapes portico was displayed at some height, the use of rectangular pillars on the upper level of two-story stoas at delos is particularly noteworthy.32 in addition to these examples, we can only assume that wooden pillars assumed. moreover, there is the possibility that the intended illusion was not that we are looking out at a “real” landscape but that we are looking at a representation of a portico decorated with a landscape painting: a representation of a representation. if this is the case, we would not expect a consistency of perspective between frame and landscape. 28 Vitr. De arch. 7.5.3–8; on the passage, and its relationship to third style ensembles, see ehrhardt 1987, 152–62; clarke 1991, 49–53. 29 the gilded capitals might suggest marble, but the garish red seems a bit too bright for rosso antico, which is how Biering 1995, 21, identifies the material. if this is meant to simulate marble, perhaps we are meant to imagine that the marble is painted? 30 macdonald and Pinto 1995, 80, interpret the space as a basilica, renaming it the “ceremonial Precinct,” “a spacious building for state business and occasions.” 31 On the portico, see Parslow 1989, 267–69, esp. n. 153 for analogs in wall painting (including the Odyssey landscapes) and in italian archaeological remains. 32 On the use of rectangular pillars instead of columns in greek architecture, including their use as “the characteristic feature of the delian upper order,” see coulton 1976,