Locus Iste: Modes of Representation and the Vision of Thera

Some of the decisions which the painter has to take relate inevitably to the treatment of spatial relations, involving the range of conventions which the painter may choose to utilise in establishing a coherent visual world upon the two dimensional surface.

While the functional concerns about the role of the paintings within a given building often initiate discussion (in which questions of iconography predominate), more formal conventions relating to modes of representation are central to the nature of the Theran 'style'. One of the salient features of Theran mural art is the careful placement of motifs (such as birds in flight), often with the liberal use of blank space, in a manner which accentuates linear outlines and serves to convey great vivacity. This is a feature seen already in Cycladic representations of earlier date, for instance in the ceramics from the Second City of Phylakopi in Melos. The 'elegance', 'freshness' and 'naturalism' widely recognised as qualities of the paintings of Thera may be seen, when situated within a world context, to derive as much from such formal properties, already characteristic of earlier Cycladic art, as from the careful observation of nature.

The painter with a blank wall in front and brush in the hand is rather like a writer with a blank sheet of paper. What will one write? About what? In what style? In what language, what script? The choice is limitless. Yet in practice the subjects chosen for treatment tend to fall, within a given context of time and space, within quite well defined categories. What are these categories? How are they determined? How large and uniform is the context within which they appear to operate? These are very broad questions, yet they are seldom asked in relation to the figuration of a specific site, where the repertoire of images offered upon the walls is often taken as a given. What is seen by the modern scholar is taken, not unreasonably, as the subject matter for study and interpretation.

The choice taken by the artist in question is generally exercised within a specific context. The context is in the first place social: the artist is working, usually for a client, who has a well-defined range of expectations. The artist is working in a community with a coherent system of beliefs about the world, many of which will be represented in the figurations produced. And the artist is working within a strong tradition of visual representation: whatever the individuality of the single painter the images produced will reflect in part the early formative years, and then the years of experience. The choice taken by the artist is also a visual one: what sort of image will be formed? How will issues of space and time be handled? How is nature conceived and represented? There are subtle preconceptions here, for which each society has formulated conventional answers (often without being consciously aware of the questions) which to some extent set the context within which the individual artist works. These very general issues, which go beyond the specific issues of personal, individual style, deal with what one may term 'modes of representation'.

It is my aim here, in this introductory paper, to situate the art of Thera (and the word 'art' is used loosely in this case to mean no more than the figuration and decoration seen in the Theran paintings) in a wider context in a number of different perspectives. At a later stage it is my hope to introduce some coherent consideration of the mural painting of early Mexico (Berrin 1988; Langley 1986; Miller 1973; Millon 1973) and of Peru (Bonavia 1985; Donnan 1972 and 1984; Schaedel 1951 and 1978).

In the first place there is a purely geographical perspective (Fig. 1) where we see a specific painting in its context within a room in a given building, that building in its place at the site of Akrotiri, Akrotiri and its paintings within the totality of settlements in Thera, Thera in its place among the inhabited islands of the Cyclades (Morgan 1990), the Cyclades within the Aegean region, the Aegean within the wider domain of the East Mediterranean, and so on, each with its own stylistic range and preferred subject matter.

As stated these are essentially designations of location: the perspective is, in spatial terms, geographic. But areas of broadly uniform culture tend to be spatially continuous, and questions remain as to how we recognise and define areas which show some degree of uniformity in cultural terms. Although the focus of our interest is on wall paintings, it is relevant to include other mural figuration from the iconographic standpoint, including mosaics, tile work and reliefs.

According to the Renaissance artist and theorist Leon Battista Alberti the mural decoration within a palace reflects:

(i) the status of the patron,

(ii) the function of the building, and

(iii) the function of the room within the building.

In a sense that is obviously so, but it begins already to conjure up a frame of reference which may beuseful.

The term 'patron' (like that of 'palace') has no doubt a specifically Renaissance connotation here, but in a wider sense it has almost universal validity, for the painter is always working for an audience, and indeed is generally commissioned (or at least invited) to work by a specific group of people, of which he or she may be a prominent member. This would be true as much for Palaeolithic cave art as for a painter in contemporary capitalist society, a Mark Rothko or a Rex Whistler, or indeed for a worker in a socialist inspired collective.

Alberti's assumption that the wall will be situated within a room within a building is reasonable. But in a wider perspective one may profitably go beyond it. Why a building? Why not a cave? A rock-cut chamber? Why not in the open air, or in a rock shelter? What constraints does the location within a building impose? Clearly it bears upon the concept of 'patron', since the building may be in private ownership, or it may be a public building, the work (and the preserve) of some corporate body.

As these ideas develop and unfold it becomes increasingly clear that there is a whole series of constraints operating upon the creation and use of any mural painting. In some cases, careful analysis may reveal what some of those constraints may have been, judging by their consequences. We may learn much about the social context within which they were created, and hence about the society itself. Any painting produced within a given society is inevitably to some extent a mirror to that society, a product of its interests and concerns, even when the society is not deliberately trying to represent itself. But to the extent that they represent a world that is peopled, paintings generally do reflect aspects of the world of those who created them - even when the painters are striving to depict something else. Even when they set out to be exotic (as perhaps in parts of the Theran miniature fresco from the West House) their very definition and treatment of the 'exotic' is revealing, and suggestive and informative about the domestic and familiar as much as of the stange and wild. Even when they set out to depict the kingdom of heaven, the choice of subjects is informative about the present, secular world and its delights and perils.

In my view the juxtaposition of genres and styles from very different contexts can be suggestive and informative for the study of both. To make such comparisons is not to subscribe to some deep-seated underlying structuralist framework in the manner of Levi Strauss and his followers, and still less to posit some generic causation in the manner of Dumézil or of diffusionists such as Perry or Elliot Smith.

The artist with a blank wall in one cultural context is in some ways in an analogous situation in comparison with an artist in a different context: it is interesting, perhaps informative to make the comparisons. For example the juxtaposition or contrast between order and good government on the one hand, and chaos and bad government on the other (very much that outlined so graphically in words by Thomas Hobbes in his antithesis between good government and 'warre') may be contemplated in two very different examples: the two sides of the Royal Standard from Ur, ca. 2500 BC, and the Allegory of Good and Bad Government, the mural paintings by Ambrogio Lorenzetti in the Palazzo Pubblico in Siena, from the fourteenth century AD.

The first case has been discussed by Woolley (1934, 266) and more recently by Irene Winter (1981b). We see on one side a succession of actions of a military nature, and on the other side more peaceful scenes. Although this has been viewed as a continuous sequence of events by Perkins (1957), an instance of the 'episodic method' devised by the Sumerians, it is perhaps permissible, as Winter implies, to see here an allegorical juxtaposition. The second has been re-considered by Quentin Skinner (1986) in a wide-ranging and illuminating discussion.

The comparison is an illuminating one mainly because the analogy in the underlying concept - the antithesis between good and bad government - is a reasonably close one, even though one context is an early Sumerian city state and the other a city state of Renaissance Italy. In the first case the 'patron' was presumably the ruler interred in Grave PG/779 at Ur, in the second the Signoria, the ruling council of the city of Siena.

Such analogies can be illuminating. It is noteworthy that Sakellariou (1980) has suggested a similar polar (or structural) opposition in contrasting the scenes of city life in peacetime seen in the 'Town mosaic' at Knossos and the miniature frescoes from Thera with those depicting war on the Mycenaean silver Siege Rhyton and the stone vases from Epidauros (also Polinger Foster 1988; Morgan 1988). Warren (1979) on the other hand takes the view that all of these works show raiding and/or city siege: he suggests that they are to be regarded as genre works.

In what follows I propose to consider aspects of Theran art within a wider, indeed potentially global, perspective. In doing so I shall consider aspects of space, place and time which offer choice and require decision by any painter, and (related to time) questions of narrative. Then, with reference to the paintings of Thera, I shall suggest that some of the special features of the way space is treated may be particularly Cycladic, since they have earlier antecedents in the Cyclades.

IMAGE, SPACE, ACTION

One of the first questions to be asked is whether the different elements on the wall, or on the several walls of the room, are to be conceived in relation to each other. Are they simply part of some decorative schema? If so they are no doubt chosen for their decorative properties. Or are they part of some larger programme, within which they take a coherently structured place? If the former, their relationships may be governed by a primarily visual logic - for instance by principles of symmetry, by spatial balance, by colour harmony, and those other components of painting which may be considered primarily decorative, as for instance in the case of much Islamic mural decoration. If the latter, other principles may be at work (Marcus 1988).

In some cases, even where there are well-defined figurative subjects, it cannot be assumed that because we see the elements together they are intended to be viewed in that way. In the French and Spanish painted caves of the Palaeolithic period the individual images are believed by some commentators to be conceived and executed individually, in isolation, despite their proximity to other such images, with which in some cases they overlap. Indeed, where engravings are concerned, for instance in Les Combarelles, the position is a complex one, with overlaps forming an almost illegible palimpsest. In such cases, as in some art in rock shelters, it may well be the act of creation which is significant rather than the persistence of the finished product. A comparable observation holds for the sand paintings of the Navajo Indians of the American south-west, whose sandpaintings are produced in the course of a healing ritual, after which they are destroyed. The significance of painting as an activity, sometimes within a ritual context, should not be overlooked. On the other hand Leroi Gourhan has suggested that the animals in the painted caves are placed logically, in positions which are meaningful within a coherent structure.

The production of any image which imitates the natural world is a complex exercise, a mapping. For each element of the image on the wall is a 'mapping' as the topologist might put it, from the corresponding element of the mental map or mappa of the artist, in which a visual memory is stored, itself the product of an earlier process of mapping from the real world to the visual memory store of the artist. In order to create such images there will inevitably be some convention of transformation which involves, among other things, the exercise of elements of choice: the artist both chooses what to depict and decides how to depict it.

It is, however, usually the juxtaposition of elements which forms the picture, and one of the most crucial aspects, as far as the effect upon the viewer is concerned, is the manner in which space is handled, or rather evoked. It should be stressed that in evaluating the spatial qualities of a painting it is indispensable to see the original work itself, if possible in situ. Working from reproductions is not sufficient. For the crux of the matter is the relationship between the viewer and the work. Issues of scale are central, as is the positioning of the viewer within the context in question.

DETACHED OBSERVER SPACE

In the figurations of many rock shelters, for instance among the Australian aborigines (e.g. Lewis 1988), the viewer is often presented with a series of figures at small scale (Fig. 2). They are generally in profile, often proceeding in the same direction and so viewed as a group. In some early cases, for instance in much Spanish Levantine rock shelter art, the figures are interacting, involved in some joint enterprise - a fight or a hunt. Many of the mural paintings at Çatal Hüyük give a similar impression (e.g. Mellaart 1967, pls. 54-57; Fig. 3). The small scale and the schematic nature of the representations make the viewer feel very much an observer, detached rather from the scene, looking down onto it. Much the same may be said for the position of the viewer in relation to the Minoan miniature frescoes of Knossos, and the miniature fresco of Akrotiri (see Laffineur 1990). These are depictions which interest us, but we are not drawn in, involved.

DECORATIVE PLANE SPACE

Sometimes the surface of the wall is treated precisely as a surface, and used as the base for a repeated design of decorative motifs which operate in two dimensions, and do not hint at a third dimension. The picture space is flattened to a single plane. One of the earliest examples is again offered by Çatal Hüyük, where several paintings are composed entirely of repeated patterns (Mellaart 1967, pls. 29 to 38; Fig. 4) which the excavator very plausibly regards as derived from textiles, in particular the woven woollen kilims which are also a feature of modern Anatolia. At Akrotiri such planar treatment is not common, but it appears in the rosette decoration on the second floor of Room 9 in Xeste 3 at Akrotiri (Doumas 1992, fig. 135).

INCLUSIVE SPACE

In some cases, especially when the figuration is at natural size, the viewer is presented with a picture which seems to draw one in to it, to include the viewer within its own pictorial field, so that one feels a participant rather than a remote and distant observer. (As noted above, this can only be judged effectively when the viewer is in the presence of the original work.) Of course the use of perspective in the Renaissance was developed largely for that purpose. Masaccio's Holy Trinity in the Church of Santa Maria Novella in Florence has this effect to an almost mesmeric degree. Such an impression is also conveyed by a number of Roman frescoes such as 'Chiron teaching Achilles to play the lyre' from the Basilica at Herculaneum (Fig.5). Titian and other masters used beckoning or gesturing saints or donors at the foot of their enthroned Madonnas (for instance in the Pearo Altarpiece in the Church of Santa Maria Gloriosa dei Frari in Venice) to reinforce the effect of the perspective and involve the viewer within the scene.

It is interesting that comparable effects have been achieved with very different means in some non-figurative art of the twentieth century, where the large canvasses of Mark Rothko and of Barnett Newman involve the viewer within a space created by the large painted canvas and which, in the case of Rothko sometimes even seems to emanate from it. This same rapport is created in a number of different ways. In the case of some Palaeolithic cave art, for instance in the celebrated ceiling painting of the bulls at Altamira (Bahn and Vertut 1988, fig. 7) and at Le Portel, the contours of the rock are used in such a way as to give the viewer the feeling that one is actually in the presence of the animals (Fig. 6). This, judging from the published illustrations (Clottes 1996, pls. 45 and 46) is the effect conveyed in some cases in the recently discovered Grotte Chauvet, which with an age of 30,000 years is one of the earliest painted caves known. Similar effects occur in the more recent paintings at Lascaux (Fig. 7).

Perspectival arrangement is thus clearly not the only pictorial device which has the effect of involving the viewer in this inclusive space: we find it operating more readily when the viewer is indeed enclosed by painted depictions, as in the Boiana Church near Sofia. Here the directness of gaze (of the Fathers of the Church) is also an involving factor, a feature seen in some Roman painting (e.g. Ling 1991, pl. 170), and indeed on a smaller scale with the mummy portraits of the Fayum. This immediacy of gaze, and a desire to include the viewer within the space created by the principal subject (be it Saint or Virgin and Child) is a long-standing feature of Byzantine art. Emphatic use of perspective can, however, have an exhilarating effect, as Uccello and Mantegna so often demonstrated. Some of the wall paintings of Pompeii, notably those of the Second Style (Schefold 1972, 227-240; Ling 1991, 23-51) sometimes have very little subject matter, other than this sweeping evocation of space (Fig. 8), evidently closely related to the stage sets of the Roman theatre. It is an effect in its way akin to the extensive use of mirrors in interior decoration at the time of Louis XV and Louis XVI, where a profound sense of space is created in rooms already spacious. These effects at Pompeii and Oplontis are effective precisely because there is no specific subject: what is offered is an effect rather than a depiction. The use of space in the frescoes of Thera is further discussed below.

HERE (LOCUS ISTE): THE CONTINGENCY OF PLACE

The relation of wall painting to place can be a very special one, especially if the 'wall' is long antecedent to the mural. This is particularly so in caves and rock shelters when the 'wall' in question is a natural one, so that the artist, or the patron or the community for whom the artist works has chosen this preexisting place as one of special significance for the work. Many rock shelters in the world where rock art is found were already special places, in some cases sacred places, before the paint was applied, while the painting may have enhanced the special sense of place.

This is particularly relevant in the case of the celebrated fifth century paintings on the rock face at Sigiriya in Sri Lanka, the Apsaras (Figs. 9, 10; Bandaranaike 1993, 117) where the presence of these beings from the other world indicate that (in ascending to the palace of the ruler) one is rising to another special and beneficent world.

The position is a different one, when the space is a constructed one. The figured decoration can be of several kinds. It could be purely decorative. Or it could refer to some narrative or myth (see the next section).

But the issue which one needs to explore is the circumstances when what is seen on the wall is in some way particularly apt for the room itself. The choice of decoration or figuration may make direct reference to the patron - as in some of the Assyrian palace reliefs or in the Audience Chamber of the Palace at Mari (Fig. 11). Or it may depict activities particularly relevant to the status of the patron - hunting, or making war for instance. But there are cases when the subject matter is particularly relevant to the activities to which the space is dedicated - for instance reference to the Baptism of Christ in buildings designed as baptistries. Funeral banquets and funeral games seen in Etruscan tombs have a special relevance to the function of the space (Fig. 12).

Of particular interest however are those cases where the wall painting may be construed as having a direct relationship with the actual function of the room in which they are situated - i.e with what went on there. A celebrated case, about which there has been much discussion, is the series of frescoes in the Villa of the Mysteries at Pompeii (Fig. 13). There is little doubt that these scenes depict episodes in the initiation process to some mystery religion. It is possible that they were commissioned to adorn those very rooms where the rites in question were peformed.

We may justifiably ask therefore whether the Thera frescoes in some cases relate to activities which were carried out there. This is of course a central argument for those authors (e.g. Marinatos 1984) who take the view that most of the rooms or buildings at Thera where frescos are found functioned as shrines. In some cases they may be illustrative of the rituals which took place there.

Liminal zones

Much mural decoration deals with the Other World, even when the context is not exclusively a ritual or religious one. In Renaissance iconography the intervention of saints and other figures from the World Above is not restricted to places dedicated to ritual and worship. The great ceiling decorations by Tintoretto in the Scuola Grande of San Rocco in Venice are found in a building dedicated to guild purposes rather than religious observances, and many parallels can be found.

At the same time, however, depictions of deities and divine beings are particularly appropriate to shrines and temples, especially when an epiphany of the deity is to be hoped for.

Tomb paintings have a special place here, for the tomb is itself, in another sense, a liminal zone, a staging post for the deceased to another world.

NOW: THE CONTINGENCY OF TIME

Sometimes the subject matter of wall painting constitutes a story of what actually happened at a particular place and time. In such cases the iconographic programme may be regarded as narrative. Narrative painting has a well defined place in the early east Mediterranean (Breniquet 1992; Frangipane 1992; Kantor 1957; Frankfort 1929; Gaballa 1976; Güterbock 1957; Perkins 1957; Winter 1981a and b) and in the Classical world (Brilliant 1984; Snodgrass 1984; Ahlberg-Cornell 1992). It is now widely accepted to be a major element in Maya art (Miller 1986).

Just as often, however, the scenes in wall paintings depict classes of event rather than a single, identifiable historic happening. This may be the case with many of the banquets and hunting scenes favoured in the mural art of early Mesopotamia (Collon 1992). It may be relevant also in scenes indicative of seasonality. The Thera frescoes certainly take note of the passage of the seasons. The delightful scenes of swallows and plants in Delta have rightly been dubbed the 'Spring fresco'. And the 'Saffron gatherers' in Xeste 3 will have been working when the crocus is in flower, namely in the late autumn. However Nanno Marinatos (1984, 64) on the basis of the presence identifies the scene as a spring one: "I believe that there naturalism has been sacrificed to symbolism" (Marinatos 1984, 64). The scene there involving the presentation to a figure whom several authors have identified as a goddess (e.g. Marinatos 1984, 62) involves activities, perhaps ritual activities, which are seasonally determined.

There are however several dimensions of time in mural art. One is certainly the annual cycle of the world, as seen in the seasons. This rhythm is mirrored in the seasonality of human affairs and especially of ritual. The Christian church offers here a very good example of a ritual cycle which, although dealing with a human time scale, compresses this into a calendar based upon the solar year.

Another very widely followed cycle is that of the life of the individual. This is perhaps at least referred to in the Thera frescoes (Davis 1986) where at least five age grades have been identified on the basis of variations in hair style. It certainly forms one of the central themes of Maya art, where many of the depictions (not least on the carved stelai) refer to the important events in the life of the rulers - the accession to the throne, the jubilee etc... These admittedly refer to the relevant events in the lives of specific rulers and are thus both narrative and cyclical. A good example is offered by the murals of Bonampak where several significant events - the Presentation, the Robing, the Celebration - in the ceremonial life of a single heir are shown, along with a major celebratory battle and a formal display of captives. As Norman Hammond has pointed out to me (pers. comm.) there are three ways in which prominent individuals in these scenes may be recognised: by their individual facial features, by their specific ceremonial headdresses, and by accompanying glyphs by way of caption. In the Palace of Westminster there is a series of paintings of coronations: each similarly depicts a specific event, yet each emphasises salient features in a ritual which is cyclical. The cycle is again not a calendrical one, but with a periodicity based upon the royal life span.

The calendar and indeed the cosmos in Mesoamerica often worked to a very much more extended time scale and in some cases the deities and their activities as depicted in the murals imply this much longer time frame than the annual round or the human life span. For many years scholars (e.g. Thompson 1960) believed that the majority of Maya depictions focussed upon these longer-term events. But it has recently been shown (Schele and Miller 1986) that most of them have a specific historical context within the life cycles of individual rulers.

NARRATIVE AND MYTH

The artist faced with the task of creating a visual programme for a series of walls or a series of rooms must be able to work to some kind of structure. Even in producing a single composition it is important that the viewer should be able to identify the subject matter - in other words, to 'read' it, if one chooses to follow the fashionable metaphor of art as text. Identification implies recognition, and recognition in turn implies the possibility of making an equation or correlation between the image in question and something that is already known. This knowledge may already be in visual form. But in many cases it relates to what is already known rather than to what has already been seen or visualised.

It is very frequently the case that ambitious programmes of figuration choose to follow a pre-existing narrative or story. This story may relate to a specific event, such as a battle, for instance on the Stele of the Vultures (Winter 1981b) or its aftermath, as in the murals in Room 2 at Bonampak (Fig. 14; Miller 1986). Or it may relate to the world of religious legend, as with so many Christian lives of the saints, as in the murals at Assisi depicting the Life of St Francis, or the detailed figuration of the Life of the Virgin at the Church of St Saviour in Chora (the Kariye Djami) in Constantinople (Underwood 1966, 60-83).

In other cases the subject matter is myth rather than legend (if the two can be clearly distinguished) as in the case of so many Greek subjects, preserved now in vases, but reflected also in Etruscan and Roman wall painting. Visual art can narrate myth as readily as legend (Morris 1989 and this volume; Hiller 1990). The latter pertains to human events, occurring at a specific place and time. The former, as the Frankforts described so well (Frankfort and Frankfort 1949) sets out timeless truths in narrative form. But their chronology is usually remote: the narrative of myth belongs to a time before mere human history, and it carries a timeless significance. In this respect it has much in common with the Aboriginal 'dream time'.

THERA: 'HERE' AND 'NOW'

A central problem in the understanding of the Theran frescoes remains the function of the rooms in which they are found. With frescoes recovered from so many buildings within the areas excavated at AkrotiIi we are left with what one might call the 'Çatal Hüyük paradox' - in other words, with so many paintings found in so small a part of the total area of the site, are we really justified in imagining that the entire site was like this, or does it just happen that the limited area so far excavated is particularly rich in mural decoration? Marinatos (1984) inclines to the view that most of the rooms in question served as shrines. So (as at Çatal) either the site as a whole must have been very rich in shrines, or the areas excavated have a preferentially high frequency of rooms with such functions, and we are in a part of the settlement where shrines were particularly numerous.

The case for 'shrine' status is most strongly made on the basis of the murals for the so called 'adyton' of Room 3 and the room immediately above. In the ground floor room or 'adyton' with the lustral basin, the east wall depicts "an architectural structure surmounted by a pair of sacral horns" (Doumas 1992, 129; Marinatos 1984, 75, fig. 53) with a fresco of three girls on the north wall. The 'horns of consecration' is a well-established cult symbol in Minoan Crete. On the floor above in this same room, the north wall shows a seated female figure on a stepped platform, receiving saffron crocuses from a monkey and flanked on the right by a griffin. The presence of this imaginary animal may be taken as a strong identitier of divine status for this figure whom S. Marinatos identified as the 'Mistress of Animals' (Potnia Theron): as Doumas (1992, 131) remarks: "Perhaps in the representation of crocus gathering we have the largest representation of the Nature Goddess in the Aegean world."

We may note that this special figure is seated outdoors, in a scene of activity and action. Nanno Marinatos regards Xeste 3 as a public building and draws an analogy with one of the functions of mural decoration in Egypt (Marinatos 1984, 32): "Ritual scenes depicting actual ceremonies which were carried out in the very room which they decorate". This is an attractive theory, and might be applied also to the naked young men in the ground floor room 3b in the interpretation that "it depicts an initiation rite during which at least one of the actors will achieve manhood, symbolised by his donning the special polychrome textile, the loincloth." (Doumas 1992, 130).

Such a conclusion answers two of Alberti's questions, cited at the outset, the function of the rooms, and of the building: a public building, used for ritual purposes, and in particular for seasonal observances related to a Nature Goddess, and initiation ceremonies or 'rites of passage'. But who was the 'patron'? Or to put the matter more generally, who was the organising authority? Marinatos (1984, 32) pertinently refers to another important function of Egyptian mural art: "Propagation of official authority revolving around the ruler. This is a theme which is either totally absent in Crete and Thera or else thoroughly disguised."

These problems intensify when the other buildings are considered. Eight building complexes have so far been (at least partially) excavated at Akrotiri: Sectors Alpha, Beta, Gamma, Delta, the House of the Ladies, the West House and Xeste 3 and 4. Of these only one - Gamma, itself incompletely excavated - so far lacks any mural decoration. Yet, despite the views of some scholars, there is little in the mural decoration of these rooms or indeed in their contents to argue unequivocally for use as shrines. A bull rhyton was indeed found in one room of Sector A (Marinatos 1984, 20), but that single symbolic artefact does not in itself make a very strong argument.

So how do we answer Alberti's questions? In the case of the West House, the so-called ikria (i.e. palanquins) which decorate Room 4 have been persuasively identified with the officers' cabins seen on the ships on the miniature frescoes in the same building (Shaw 1980 and 1982) and the artists' 'patron' might well have been a prominent seafaring citizen who lived in the West House. But such questions can only be answered when we have some better understanding of the social organisation of Akrotiri. They do, however, bear upon our understanding of the artistic conventions and modes of representation of the time: there are major differences between the use of imagery to enhance ritual activities taking place in the very rooms where these are depicted, on the one hand, and the employment of figures and narrative scenes purely for the purposes of decoration on the other.

NATURALISM AND SPACE

"Unfortunately, what seems most easily accessible in this art will prove most alien, so that our very criteria of judgement become useless. To speak of naturalism, for instance - that is, of a conscious interest in, and respect for, the appearance of the phenomenal world - may be tempting in the face of forms so buoyant with life, yet it means ignoring the fact that several of these forms cannot have been observed at all; blossoms and leaves of different plants have often been high-handedly combined, birds' plumage altered. To speak of landscape may be equally tempting, yet this would make the term quite meaningless. True, all the requisites are there - plants, trees, rocks, watercourses... - but the essence of landscape, the rendering of natural phenomena in space, is lacking completely."  (Groenewegen-Frankfort 1951, 196).

The wonderful opportunity which we have of seeing the various elements of Theran mural painting in their original relationships, an opportunity in the main denied for Minoan fresco painting because of the fragmentary condition of the finds, gives us the occasion to re-evaluate the seemingly rather critical judgement of Groenewegen-Frankfort in her generally very sympathetic analysis of Minoan painting. Certainly I would like to argue here that one of the salient features of the larger Theran paintings is, in fact, their remarkable use of space, although it may be that this is a Cycladic rather than a specifically Minoan quality.

Their apparent naturalism does indeed flow from the choice of plant and animal elements. But if different features are sometimes combined, that is hardly a failure of observation, but as Warren (this volume) has shown, a decision based on a close understanding of the real world. It may prove to be a deliberate pictorial decision rather than carelessness in observation. Walberg (1986, 10) has argued that the source of many plant motifs lies in the pottery painting of the Classical Kamares phase, and very often in abstract elements found there: "indeed it seems that even though the painters still use abstract elements to symbolise features and details in the plants, they appear to be more sensitive to the effect of these elements and to choose elements that have closer relations with natural plants." (Walberg 1989, 12).

In undertaking a general survey of the Theran murals it is in the first place appropriate to re-assert that there is little of the formalism in Theran painting that frequently accompanies art of a more monumental character. We do not see the antithetically placed animals of the Throne Room at Knossos. We do not see (or at any rate not yet) the processions of very uniform figures seen in procession frescoes at Knossos or the Mainland. This may, of course, be a reflection of the functions of the buildings so far excavated (and there are hints of a procession fresco in the staircase of Xeste 4), but then we do not have indications at Akrotiri of a massive structure which could be called a 'palace', even if the ashlar masonry of some of the buildings (including Xeste 4) might reasonably be claimed as 'palatial' by Minoan standards. The paintings are not arranged in superimposed registers like so much Egyptian painting - visual or aesthetic criteria perhaps dictate that one figured register in one area of wall is sufficient. But at the same time the sense of the decorative does not go to the frivolous extremes, seen in Pompeiian painting of the Third and Fourth styles, of placing figurative panels in ornate frames, in such a way that the decorative effect becomes a frivolous one. For all its grace, the Theran style is not a roccoco style.

Secondly it is essential to note that, as in the case of Minoan painting, different conventions of representation are used in the miniature frescoes in contrast with the full-scale murals. To be sure some of the same conventions apply, and the observations of Laffineur (1990) are relevant, but the miniamre frescoes are 'busy': in them it would not be inappropriate to say that 'Nature abhors a vacuum', which would be less just with the larger Theran murals. This may relate to the differing spatial effects. As noted earlier the miniature frescoes operate within 'Observer space': we are observers of a narrative rather than participants in a scene where we feel present on an equal footing.

Precisely the converse is true with the larger frescoes. I clearly remember my sense of wonder when we had the privilege to be present, as a guest of Professor Marinatos in 1970, as the 'Spring fresco' of Delta 2 was being excavated. The spatial relationships in this little room were so well defined, and indeed greatly heightened by the delicacy of treatment of the swallows, that one felt drawn into this charming world. This was 'inclusive space' indeed. I felt a comparable sense of wonder in 1995 when Professor Dournas showed us the reconstructed figures of the youths from Room 3b of Xeste 3. These were real, life-size representations, allowing one to feel (as one does in the presence of the fresco works of the early Renaissance masters such as Masaccio or Mantegna) one's own corporeal presence and scale. Although the drawing and modelling is not, by classical Greek or Renaissance standards, very well defined, the presence is a very powerful one: these figures occupy and thus define a pictorial space in a very effective way.

One secret of their success is that they are not set apart from us by a superfluity of frames, garlands and other decorative devices, as some Roman mural paintings are. Instead they are set on a simple white ground, with a dado of simple bands above and below. In this respect they are presented with the same directness and simplicity which proves so effective in much Etruscan tomb painting, such as the Tomb of the Triclinium and the Tomb of the Lionesses at Tarquinia (Fig. 15; Pallottino 1952, 43-47 and 73-77).

The youths, with their pink bodies, achieve an effective presence using simple blocks of colour. The women, on the other hand (e.g. Doumas 1992, fig. 154), have flesh which is the same colour as the white ground. In one sense one could see them as completely transparent. But even those without coloured clothing achieve an effective presence through the great vitality of the outline and of the accompanying details. There is thus a great difference here in the way these youths and maidens are represented, which is not simply a question of body colour.

PLACEMENT IN CYCLADIC ART

One of the strongest contributory factors towards the general impression of 'naturalism' (which, as noted above, Walberg and Warren have shown is not based on a very close fidelity to natural forms observed at a detailed level) is the use of space in the larger frescoes. We are fortunate indeed that this may be judged reliably at Akrotiri, where the composition is often preserved almost in entirety, a feature rare in most Cretan and Mycenaean finds. The Theran painters were not afraid to use space, blank space, very liberally. This we see clearly with the male youths in Xeste Room 3b: they stand as freely as images on Greek red figure vases (Doumas 1992, 112-115). The same may be said of the saffron-gathering ladies of Room 3a, and this spacious, light and airy quality is enhanced (Doumas 1992, fig. 154) by the avoidance of large blocks of colour, since the female bodies are shown in the same colour (white) as the ground itself. This feature gives extra force to the outline and specific, simple details - eyes, jewellery - giving an impression of delicacy and vitality. The same observations hold for the ladies on the ground floor of Room 3, the 'Lustral basin' area. It is appropriate perhaps to single out the gesture of the lady holding the necklace in her hand (Doumas 1992, fig. 141; compare Fig. 9). Here the strength of the effect comes from three or four qualities: the airy, spacious impression achieved by the minimal use of colour, the delicacy of the outline, the subtle harmony of the few elements of colour which are used, and a special Cycladic character which one might summarise as poise.

Poise in this sense implies balance and gentle movement, achieved by the use of a fine outline on a white ground, and by a rather liberal use of circumscribed emptiness. (I use the word 'emptiness' here to avoid undue repetition of the word 'space' which in the earlier discussion has a rather different sense, referring to the involvement of the observer in a three-dimensional relationship.) The same quality is shown in the way one of the young men in Room 3b is carrying a basin (Doumas 1992, figs. 111 and 115), in the saffron-plucking motion of one of the gatherers (ibid., fig. 156) and in the wing of the griffin accompanying the goddess (ibid., fig. 165). A comparable observation holds for the carefully balanced heads of the two antelopes in Room Beta 1 (ibid., fig. 83), and above all for the swallows in flight of Room Delta 2, most especially the two paired swallows (ibid., figs. 74 and 76) but also one of those in flight (ibid., fig. 75; also Hollinshead 1989; Immerwahr 1990b). Similar comments may be made for some of the painted ceramics of Thera (Davis 1990, 224; Marthari 1987, 376).

It is on these swallows that the case for the particularly Cycladic quality of this art may principally rest (Fig. 16). For while Minoan art of this and the preceding period draws pre-eminently on the plant world, birds are a particularly Cycladic preserve (Vanschoonwinkel 1990). Cycladic inspiration is widely accepted for the well-known examples in the Shaft Graves at Mycenae (Crouwel 1989, 164) and indeed for the full-bodied jug from the Middle Minoan III treasury of the Palace sanctuary at Knossos (Fig. 17; Pendlebury 1939, pl. 37:4).

At Phylakopi in Melos, the Black and Red style is primarily used for the representation of birds (Figs. 18 and 19; Atkinson et al. 1904, 119), mainly of 'round bodied type', but in one case (ibid., fig. 92) a swallow is found very similar to those of the Theran frescoes (Fig. 16). Jugs decorated with birds are not uncommon in the destruction levels at Akrotiri (Marinatos 1972, pl. 58). The Black and Red style is found not only in the Late Bronze Age Third City at Phylakopi, but also in levels of the Middle Bronze Age Second City (Barber 1978, 376).

In this context it is not so much the motif itself but the mode of treatment which interests us, and in particular the way that the bird is represented in space, without being encumbered by numerous other motifs. This is a feature of the Melian Black and Red birds (e.g. Zervos 1957, pls. 306 and 308). But it is even more so a feature of the Cycladic White ceramic style, which is the ancestor and prototype for the Black and Red style. Indeed beaked jugs of the Cycladic White style or its predecessor go back to the very early days of the Second City at Phylakopi or earlier (Fig. 20): their successors are the elegant 'nippled ewers' of the Late Bronze I period, which are common at Akrotiri (e.g. Marinatos 1971, pl. 69).

The point here which I think is particularly interesting is the very free and open spatial disposition of the design on some of the Cycladic White vessels from Phylakopi. Some of these depict curious 'winged demon' figures (Fig. 21; Zervos 1957, pl. 268): one splendid example has a flying griffin seen in magnificent isolation on the white ground of the vase (Fig. 22; Zervos 1957, pl. 271). It is on this and on other Cycladic White vessels that one sees the willingness to show open spaces - to show individual motifs with this circumscribed emptiness - which are so fundamental a feature of the Theran frescoes.

Along with this willingness to use empty space to such good effect, and to evoke it through the depiction of birds poised in flight, or in the balance of a suspended necklace, is a concern to use the arrangement in space of pictorial elements in a deliberately decorative way. This is of course a feature of Minoan art, arising no doubt from the patterns and motifs of Kamares painted ware. But in the Cyclades we do see some especially delicate effects of patterning, not simply in the areas surrounding the field of the fresco (e.g. Doumas 1992, fig. 83), but in some cases within the pictorial space itself. The best example is in the 'Mistress of Animals' fresco of Xeste 3, Room 3b, where the entire white background is embellished with elegantly spaced crocus flowers, now golden but probably mauve when originally painted (Doumas 1992, fig. 122 and rear endpapers) (Pl. 12). The same device is seen on the adjoining fresco on the East wall (ibid., fig. 152).

Several things are interesting about this. In the first place the regularity suggests a deliberate attempt to embellish the work: it shows, then, a deliberate objective not only towards representing the necessary images, but also at decoration. A deliberate aesthetic intention can therefore be inferred here, without contradicting the assertion of Nanno Marinatos (1984, 31; 1985) that the function of the frescoes was not primarily a decorative one. But it is, moreover, a rather daring device in that it might have compromised the 'inclusive space' which the painting so successfully creates. For an 'over all' approach to decorating the surface is a common means of achieving the 'decorative plane space' mentioned earlier, which can so easily subvert the achievement of inclusive space. Tapestries, for instance, generally have a very flat and decorative effect precisely because the level of surface decoration (and the texture of the fabric) results in a two-dimensional planar effect, which perspectival devices cannot entirely overcome. For instance the great series of Raphael cartoons at the Victoria and Albert Museum achieve a more effective 'inclusive space' than do the tapestries actually woven from them, such as those in the Ducal Palace at Urbino.

The use of such a device to give a pleasing decorative harmony to the composition, yet without subverting its spatial properties, is a remarkable testimony to the mastery of the painters of the Thera frescoes.

-------------------------------------------------------

-------------------------------------------------