• 1 Florence, Museo Archeologico 4209; ABV 76.1, 682, Para 29: Klitias (painter); around 580-570; BA 3 (...)
• 2 The paintings of this villa were excavated in 1903-1905 and roughly published by Della Corte in 19 (...)

1I begin this article with a comparison of two highly incommensurable pictures: the Calydonian boar-hunt on the François Vase (fig. 1),1 an early 6^th century Attic wine-mixing vessel (see most recently Torelli 2007, Shapiro et Lezzi-Hafter 2013), and a landscape decorating a wall of the so-called Villa of Agrippa Postumus at Boscotrecase (fig. 2), a lavish Roman villa of Early Imperial age on the Bay of Naples (Della Corte 1922, Blanckenhagen and Alexander 1962/1990, Peters 1963, p. 69-71, Anderson 1987 and 1987-1988).2 The comparison shall deal with the depiction of landscape, or, in the case of the boar-hunt, with the missing depiction of landscape. Reversing the normal order of enquiry, I will discuss the incommensurability of these pictures only in a second step, in order to reach the main subject of this article: pictorial space as a media phenomenon. In this comparison, ‘landscape’ in Romano-Campanian wall-painting will receive the more extensive treatment, while ‘landscape’ in Attic vase-painting will mainly serve as a contrasting comparandum. The very few examples of Attic vases mentioned here do not, of course, illustrate the depiction of ‘landscape’ in Attic vase-painting in all its complexity, nor can they stand for later developments especially in 4^th century Attic and South Italian vase-painting (for a fuller account of ‘landscape’ in Attic vase-painting, I refer to my monograph Dietrich 2010).

• 3 Panofsky’s article has since been intensely discussed in the field of art-history. See e.g. the in (...)

6The contrast between Archaic and Classical Attic vase-painting and Early Imperial Roman wall-painting concerning landscape is thus very clear. According to the usual methodology of historical analysis, we have to conclude that at some time between the 5^th century BC and the late 1^st century BC, the depiction of landscape in art must have emerged. Some very different monuments standing out by the relative prominence of landscape elements on them would probably let us place the development of landscape in the Hellenistic age. For this, one might cite the so-called tomb of Philip in Vergina from 350-325 BC with its hunting frieze (Andronikos 1984, p. 97-119, Franks 2012), or the Telephos Frieze from Pergamon from 170-160 BC with its rich depictions of landscape (Heilmeyer 1997, p. 99-120, Carroll-Spillecke 1985, p. 18-24, Wegener 1985, p. 56-60). But more fundamentally, this development should then have to be explained and interpreted in terms of cultural history. Though not dealing with landscape specifically, Panofsky has shown us, in his seminal work on perspective in the Renaissance, how an analysis of the changing depiction of space and its significance for cultural history could look like (Panofsky 1927, English translation: Panofsky 1991).3 Classical archaeology has since occasionally tried to adopt such an approach in the study of the development of the depiction of space (closest to Panofsky’s account: Schweitzer 1953).

• 4 Here, Hölscher’s thinking is obviously influenced by French scholarship of the 1960s and 1970s. Se (...)
• 5 Naples, Museo Archeologico Nazionale 10020; for a short review of the vast bibliography on the Ale (...)

7In my sense, the best example of a Classical archaeological attempt to load the changing spatiality of pictures with fundamental historical significance is Hölscher’s work on the development of the iconography of battle from Archaic and Classical times on (Hölscher 1973, p. 25-30, 50-84 and 122-169, Hölscher 1987, p. 20-29 [with English translation: Hölscher 2004, p. 23-37], Hölscher 1995, p. 30-36, Hölscher 2003, p. 4-7). The numerous Archaic pictures of fighting hoplites show war as the confrontation of single warriors, in line with aristocratic values as we find them in Homer, where the bravery of single warriors guarantees the victory of the group. These single combats evoke the agonistic societies of the Greek poleis who fought on the battle-ground in a kind of continuation of the athletic agon by other means.4 The famous painting of the Battle of Marathon in the Stoa Poikile on the agora of democratic Athens shows the victory over the Persians as a collective achievement, though still disrupting the battle into many scenes of individual exploits and single warriors fighting: the democratic polis won the war – but the polis is still not more than the sum of its citizens (on the spatial structure of the Marathon painting see also Dietrich 2010, p. 235-240). The Alexander Mosaic,5 copying a painting of the late 4^th century, for the first time shows a battle as a complex though not less dramatic movement of troops, where the action and suffering of single warriors make sense only in the collective context. Thus, the picture deploys a spatiality that goes beyond the range of the single figure. This corresponds to the new Hellenistic order of the Greek world where large states take the place of the ‘face-to-face’ societies of discrete poleis.

• 6 Munich, Antikensammlung 1549; ABV 383.12, Para 168; CVA Munich IX, pl. 12.3, 15.1-2, 17.5; BA 3024 (...)
• 7 Similar things can be said for landscape elements in Greek architectural sculpture, as exemplified (...)

15That pictures on vases or on walls should not be looked at as isolated works of art but have to be studied within their context has long become a commonplace among archaeologists. But which context are we talking about? There surely are many different relevant contexts for a picture: who is looking at the picture? On which occasions? What other pictures are around? Among the many possible contexts of a picture, the basic one of its material support generally receives relatively little attention. In 18^th and 19^th century Europe, the same kind of landscape may be found on painted porcelain as on canvas. A bone china chocolate cup from 1815-1820 in the Victoria and Albert Museum in London can serve as an example for this (fig. 6). The pastoral scene on the cup, with cows in the foreground on the left and mountains in the background on the right, presents itself just as a framed canvas, opening up a view on a grand landscape going far beyond the tiny dimensions of the fine chocolate cup. Here, the spatiality of the picture does not depend in any way on the dimensions of its material support. In Graeco-Roman visual culture, this is very much not the case, as I would claim. As I tried to show in another context, the landscape elements in Attic images never go beyond the dimensions of the figures. Even the mountain of Sisyphos is depicted as a clearly confined object in space, as seen e.g. on a black-figure amphora in Munich around 510-500.6 Thus, they conform to the spatiality of their material support, the vase, which is too an object in (the viewer’s) space (Dietrich 2010, p. 34-39, Dietrich 2011, p. 296-301).7 Landscapes in Roman wall-painting, on the other hand, go beyond the spatiality of confined objects in space. Indeed, they themselves provide the enclosing space for the objects, architectural structures and figures depicted. As already noted above, the shepherds in the sacro-idyllic landscape painting from Boscotrecase seen in fig. 2 do not stand on the wall like the hunters on the François Vase who stand in the frieze but actually in the landscape. At a first glance, one might interpret this as an eventual emancipation of pictorial space from the spatiality of its material support. However, I would like to suggest another interpretation. By providing the enclosing space for the figures and objects depicted in the picture, the landscape painting still corresponds to the spatiality of its material support, the walls of a room, which provide enclosing space, too – though not for the painted figures in the pictorial space but for the real people in the room. To put it simple: as decorated Attic vases are objects in (the viewer’s) space, the ‘landscapes’ on them are depicted as confined objects too. As decorated walls of a Roman domus enclose (the viewer’s) space, the landscape on them provides enclosing pictorial space too. In both cases, there is thus a correspondence in the spatial structure between the picture and its material support.

• 8 When in later developments of Attic and South Italian vase-painting from the end of 5^th century an (...)
• 9 London, British Museum B 314; ABV 360.2, Para 161: Leagros Group; CVA London, British Museum VI, p (...)

16For Attic vase-painting, the correspondence of the landscape to the spatiality of the vase seems rather clear. Not only does ‘landscape’ always stay within the clearly confined dimensions of an object,8 it even conforms itself to the ‘architecture’ of the vase. This is particularly clear where rocks strictly follow the outline of the picture-field’s frame, turning in a right angle at the frame’s corner, as seen on a hydria of the Leagros-group with Geryoneus reclining ‘in his cave’ (fig. 7).9 This ‘cave’ results simply from fitting the border of the picture field with rocks, without creating any other space than the one provided to the figures by the ‘architecture’ of the vase (Dietrich 2010, p. 156-177). In the case of Romano-Campanian wall-painting, the spatial correspondence between the picture and its material ground might be less obvious. But as I would like to show in the remaining part of this article, Romano-Campanian wall-painting develops much of its aesthetic appeal by gravitating around the wall’s spatial quality of enclosing the viewer’s space, alternately challenging and reaffirming that enclosing quality.

• 10 New York, Metropolitan Museum 20.192.16; for a detailed description and commentary, see Blanckenha (...)
• 11 This is a general trait of third style landscapes: see e.g. Croisille 2005, p. 207-208. Another fa (...)

18What I described here is all well known, and especially the ambiguities intrinsic to Augustan wall painting recently received specific attention (Platt 2007). If third style wall painting is nevertheless generally described as plain surface-decoration in archaeological literature, this is done only in comparison to second style walls, where the element of architectural perspective is much more apparent (see e.g. Zanker 1987, p. 281-283, Ling 1991, p. 52-53, 57, Mielsch 2001, p. 70-73, Croisille 2005, p. 68-71, Barbet 2009, p. 104). But taken for themselves, the painted cubicula of the Boscotrecase Villa do not feature any wall that would not play on the ambiguity between the plain ornamented surface and perspective depth – and the landscapes in the centre of the walls are the focal points of that interplay. As a striking contrast to the surrounding plain surface, the sacro-idyllic landscapes of cubiculum 16 and the mythological landscapes which are found in cubiculum 19 of the same villa function equally well (fig. 10a-b).10 Here, a similarity in the composition of all the sacro-idyllic landscapes of cubiculum 16 and the mythological landscapes of cubiculum 19 is most telling for these pictures’ function in the overall decorative design scheme and the alternate challenging and reaffirming of the wall’s surface. When put side by side as in fig. 11, it immediately becomes clear that all landscape paintings are organised around a central vertical structure, be it a column, an architectural structure, a tree or a rock. Given that these landscapes are all positioned in the centre of their wall, the vertical structure obviously functions as a central symmetry axis of the overall decorative design.11 Thus, the central (mythological or sacro-idyllic) landscapes are not only breaking through the wall into an infinite pictorial space but at the same time, they are still firmly integrated in the wall decoration’s ornamental pattern. The ambiguity between the plain ornamented surface and perspective depth enters even the landscapes themselves.

• 12 New York, Metropolitan Museum 20.192.17. For a detailed description and commentary, see Blanckenha (...)
• 13 The mingling of elements of mythological and sacro-idyllic landscape painting is rather typical. A (...)

19In the case of the Polyphemus-and-Galatea landscape of cubiculum 19 (fig. 12),12 the similarity to the sacro-idyllic landscapes of cubiculum 16 concerns not only the composition but also some iconographic features such as the column supporting a vase or the statue erected on another column (fig. 13). The cyclops’ island is somehow turned into a sacro-idyllic landscape. Insofar as the world of myth and the sacro-idyllic landscape both present a space characterised by their distance from the viewer’s space, the two landscapes can be assimilated in terms of iconography.13 But as much as these landscapes are withdrawn from the viewer’s space by means of their iconography and by their breaking up the closeness of the wall’s surface, their ‘otherness’ is nonetheless firmly kept within the room and fixed on the wall. This appears in another telling detail of the paintings. The (painted) light in the landscapes always comes from the same direction as the (real) light in the room, i.e. from the opening’s side on the south. The same is true for protruding architectural features in the wall decorations, which throw shadows to the same side (this has, of course, already been observed, e.g. by Peters 1963, p. 71), just as if the same (real) sunlight would light the decorated wall and the landscapes painted on it. This reminds us of the fact that every room’s three landscapes were completed by a fourth landscape: the one that could be seen from the room’s opening (on the relationship between real and painted nature, see Bergmann 2002).

• 14 On the landscape paintings of the Villa Farnesina, see Croisille 2010, p. 81-86. For excellent pla (...)
• 15 Again, the Casa di Livia with its famous yellow frieze provides a good example (Croisille 2010, p. (...)

21Other contemporary landscapes in Roman wall-painting show another strategy of maximizing the ambiguity and concurrence of distancing and opening landscape on the one hand and enclosing wall on the other. The sacred and Egyptianising landscape on the black surface of the wall of the so-called black triclinium of the slightly earlier Villa Farnesina in Rome are painted in such thin lines as to keep the monochrome surface of the wall present to the viewer (fig. 14).14 According to how s/he looks at the large black panels in the wall’s middle zone, s/he would see it either as a landscape or as a closed wall, structured by slender candelabrum-columns and adorned with wine-garlands. This strategy of keeping both the displacing landscape and the enclosing wall present explains the painters’ general predilection for monochrome landscapes on wall-zones that might as well count for normal architectural decorative elements such as friezes or wall-panels etc.15

• 16 PPP III 271; PPM VII 1065, fig. 236. On the house and its mythological imagery, see Lorenz 2008, 3 (...)
• 17 On the complex relationships between the mythological figures on the walls and the dominus and mat (...)

23Interestingly, the mythological landscapes of third style walls tend to give way, in subsequent fourth style walls, to mythological pictures much more centred on the figures, leaving little or no place for landscape (Lorenz 2008, p. 38-39). Often, the mythical protagonists appear in front of some object closing the view behind them. The opening up of the space enclosed by the walls that had been achieved on behalf of the mythological imagery in the third style – both in terms of perspective depth and in terms of iconographic ‘otherness’ – is partly abandoned in favour of the mythological protagonists’ more direct reference to the real figures in the decorated rooms: the host couple, dominus and matrona, the actual referents of the Roman house’s decor. The central picture of room 58 in the House of Marcus Fabius Rufus in Pompeii would be one example among countless others (fig. 15a-b):16 Hercules and a female protagonist appear in front of a wall. The picture is very unspecific, neither allowing a definite identification of the woman nor of the precise mythological constellation, nor of the locality. Reducing the ‘dissociating’ elements of mythological imagery, the two figures posing like a couple most directly mirror dominus and matrona – in this way assimilated to the mythical paradigms of manly virtues and feminine attractiveness, Hercules and his partner.17

• 18 For the decoration of the Roman house, the importance of appropriateness and decorum has been brou (...)

26Within such an explanatory model based on the strategies of media for explaining the presence or absence of deep perspective space and landscape, the idea of appropriateness – whether a certain decoration does or does not fit its object, context and aims – is much more crucial than the idea of cultural historical change. This idea of decorum is a highly ideologically loaded subject in most pre-modern societies, and certainly in Graeco-Roman culture.18 Of course, the aspect emphasized here does not exclude cultural historical changes! But it should give us warning of any automatism in assuming fundamental cultural historical change where there is diachronic difference.