Painted Rock Shelters of the Swat-Malakand Area From Bronze Age to Buddhism - Materials for a Tentative Reconstruction of the Religious and Cultural Stratigraphy of Ancient Swat

Preamble In the last ten years 59 painted shelters were documented in Khyber Pakhtunkhwa Province. These rock-art sites were mostly discovered in the upper Kotah valley, in the Swat-Malakand region, at an average height of 1,000 m asl (49 sites). To this group have to be added other 10 analogous sites documented in 2010 in Swabi and Manshera Districts. 1) The painted shelters discovered are similar in regard to their location, techniques utilized, style and subjects. Therefore, they have been considered physically as a relatively homogeneous group. 2) The rock shelters are situated on bare eroded crystalline rock on the edge of steep sloping ground or ravines. The paintings are hosted within natural cavities. The rock shelters rise to approximately the same absolute height, and are sometimes located in the vicinity of major mountain passes, springs, or along ancient paths. 3) In general, it is difficult to ignore the impression that most sites were chosen due to their visual dominancy. The rocks, boulders or pinnacles, which host the paintings rise as landmarks and are visible from a considerable distance. The painted walls are often not easily accessible or visible. For the painter as well as visitors and onlookers, access to most of the shelters was always quite difficult. Arriving to them often necessitated climbing, sometimes painstaking sliding, and then in some cases required one to lie on one’s back. In these cases, the ceiling and other paintings had to be observed from a short distance, only a few centimetres from one’s face, and only one painter or onlooker at a time would have been able to enter the shelter to observe the figures. In general, it is evident that the paintings had been traced, and had been observed, by individuals or very small groups, sitting or crouching. 4) All the paintings were made utilizing mineral pigments made with base metals. Tests performed did not reveal the existence of organic substances such as casein, for example, which should have been present in the composition of the pigment. The ochre tints utilized are listed in order of their frequency: carnelian red, crimson red, orange red, scarlet red, white, and cadmium yellow. The latter color is used at only a single site. 5) The application of color was most likely done with the artist’s fingers. In all cases, the breadth of the application is similar to and differs only slightly from the average width of a finger; in one case, signs of fingerprints were documented. In no case was there evidence of the use of brushes or sticks. 6) All of the figures within each individual painting appear to be part of a complex: apart from few exceptions, one does not have the impression that the figures were added gradually. There is very few evidence of the overlapping of colors or shapes, or of additional layers of colors. Methodology of the research 1) The paintings were contact-traced onto transparent polyethylene sheets using permanent markers, then reduced to the necessary scale. Digital color calibration of the photograhic documentation was carried out when necessary. Thanks to this, in many cases it was possible to integrate the contact copies made on site, where color spectrum was not distinguishable from the rocky surface. 2) To some extent the study of rock art is actually an anomalous sub-discipline of archaeological science. This is true, at least with regard to two aspects: it is the only field where the site is not potentially subject to a destructive analysis (which is after all the final result of digging), and it deals with only the meaningful data itself, which, due to its nature has escaped the analogical-inferential process that characterizes the interpretation of data from an excavation: a dozen beads and fifty pottery fragments do not make up a workshop per se, but only thanks to the associative ability of the three-dimensional mind of their excavator. One important aspect to consider is that while an excavation is a complex site consisting of numerous multi-formed data or objects (ceramics, minor finds, coins, etc.), a rock art site is both a site and an object in itself, no more or less than a vase, painting or a sculpture. From the art history point of view, in contrast to a sculpture or a vase, a cave painting is always contextualized. In this sense a rock art site can provide us with a subject for experimental study, endowed with its testability. However, it is a potential testability. As the site conditions are maintained intact, the conditions exist for reversibility and repeatability of the experimental inquiry. At the same time, however, the nature of experimental investigation lies in the field of interpretation. The broader horizon, which we would like to be able to capture from the data presented, namely the reconstruction of the human landscape, the psychological ratio of the phenomenon of the painted shelters of Swat-Malakand, obviously has its limits. Beyond the rare cases when it has been possible to cross direct/indirect archaeological data with the rock-art evidence, the consistency of this study is based mainly on its logic, which is a parameter which is only partially objective, in and of itself it is, alas, an interpretation. 3) Attempting to interpret rock art not only as a manifestation of self-experience, but also as a mode of communication (especially visual), means trying to recognize semantic patterns which are more or less specialized. As with semantic patterns, where possible, the distinction between pictograms, ideograms and psychograms is useful. This is the case, of course, if by pictogram one intends an objective representation of an object, and by ideogram one intends a more complex, conceptual representation. For example, a central cruciform shield which dominates a scene with warriors armed with shields, is an ideogram, while shields which have been taken up are pictograms; an anthropomorph which is larger than the human figures which surround him, even if not dissimilar to the symbol ‘man,’ could be an ideogram. Instead, a psychogram is a sign that no longer represents an object as much as an action, or a modification of both (as can be said of adjectives and adverbs): an abstract element that is introduced as a visual sign on the rock. Generally, enigmatic symbols such as points, sinuous lines, zigzags, etc.. represent the most common vocabulary present in psychograms. These observations permit us to utilize terms such as ‘lexicon,’ ‘syntax’ and so on, which would otherwise be ambiguous. The interpretation of individual design types (in terms of semantic elements) is briefly reviewed in the following index. 10 Thesis 1) The rock-paintings cultural phenomenon in Swat-Malakand, Swabi and Hazara is comprised within the chronological limits of 2nd Millennium BCE and the first-half of the 1st Millennium CE. The location of the shelters is not associated (inside or outside) with satisfactorily preserved archaeological deposits. One likely consequence is that their dating will depend primarily on the study of the material culture portrayed in the designs themselves, as well as the study of comparative stylistic considerations, and the analysis of topographic contexts. In general any rock art site tends to retain many elements of doubt, which are more numerous than the confirmed facts. The privilege that is accorded to internal chronological stages, or a relative chronological history where detectable, for example in the overlapping of asynchronous paintings on the same medium, often represents the only answer in determining the timing of a site. In Swat, after nearly fifty years of excavation and archaeological research in a geographic basin, which is quite small, it is not absurd to indirectly attempt to cross the archaeological data with the logical deductions suggested by the visual analysis of the painted complexes, even though initially decontextualized. The following summary is based on the visual evidences provided by the paintings. a) Active Buddhist buildings did not exist in Swat either before the 2nd century BCE or after the 10th century CE. Paintings containing representations of Buddhist architecture should probably fall within this period, and more precisely between the 1st and the 4th to 5th centuries CE. b) No representations of horses are present in Swat before second half of the 2nd millennium BCE. Horses are represented with an increased frequency in Swat only after the first half of the 1st millennium; however, representations of horseback riders do not appear before the end of the same Millennium. Vice versa, representations of horses with crenellated manes and covered tails, which are present in some paintings, should not be dated from later than the 4th to 5th centuries CE. Consequently, the representations of horses in the concerned paintings were without doubt produced within a period, which began after the 5th century BCE and terminated within the second half of the 1st Millennium CE. However, in 5 Sites, besides the presence of agricultural ideograms, there are also indicators of horsemanship and breeding, figures of heroes and large anthropomorphs, and the absence of Buddhist structures. All of these Sites are characterized by the physical characteristics of ‘Hermitages’ or ‘Sanctuaries’. The majority of the other Sites, have in common the presence of Buddhist architecture, horsemanship, war scenes, hunting, breeding/livestock/farming and sheep-farming; and the forms are enriched with body modification. Among these, 4 Sites contain representations of triśūla-like ideograms; in two cases they are associated with representations of Buddhist architecture. 2) In Swat-Malakand the rock-paintings area probably corresponded to the living area of the local human communities and/or to their subsistence space. It appears possible that, even if our paintings do not represent not a culture in itself, they none-the-less represent a widespread phenomenon in time. It is found to be parallel to or to fall within a chronological arc characterized by a series of cultural macro-phenomena, all of which have been archeologically documented: (a) The existence of the of small agricultural settlements and centers of exchange, which relate culturally to a period contemporary to the Late Harappan phase (Localization Era; Swat Period IV c. 2nd millennium BCE); (b) the so-called ‘Gandhara Grave Culture’ (henceforth: the ‘graveyards’; end-2nd millennium BCE - mid-1st millennium BCE); (c) the progressive intrusion of Buddhism with its cultic foundations, from 3rd century BCE, that reached its peak, both in quantitative terms and spatial distribution from the 1st to the 4th-5th century CE; (d) urban settlements created and linked to the same local power system described above (end-2nd century BCE - 5th century CE); (e) the foundation of Brahmanic cultic centers under the Shahi dynasties (7th century-11th CE). All these macro-phenomena had an evident and recognizable impact on the social history of the antiquity of the Swat valley and represent phases of acculturation. Acculturation phases might coexisted with phases of cultural marginality (or late persistence of earlier cultural forms) in remote areas of Swat. These macro-phenomena obviously do not represent a totality of ancient Swat and its chronological arc. In fact there is no direct trace of these macro-phenomena in the iconography of our paintings if we exclude Buddhist architecture, and hypothetically ideograms and deity of the Brahmanism of late antiquity. For a moment, let us hypothesize that the community which we credit with the paintings lived in the main valley, far from the area where the painted shelters are, and the shelters were visited by the valley community for ritual purposes. This possibility is not currently supported, in that we have noted that there are irrelevant points of contact between the lexicon of the painting and the graphic and ideological heritage of the cultures that lived in the main valley. On the contrary, we believe that the painted shelters are an expression of communities, which were peripheral or subordinate to the cultures of the main valley; however segregated to the living space of the middle and upper valleys of Kandak, Kotah and Thana. The same context would seem to be applicable to the rock paintings in the areas of Swabi and Mardan as well as for those found in the Tanawal in Hazara. The ecological space available coincides with the economic areas dedicated to hunting and gathering, pastoralism/sheep-farming, and subsistence farming: in fewer words, with the wild. This situation, mutatis mutandis, is similar to that present in Swat. Until today (but the situation is rapidly changing), the most privileged agricultural lands, even in the lateral valleys and the along the main roads of communication, are owned by ethnic Yusufzai landlords (khans). The right of possession is legitimated by the 16th century Yusufzai military conquest, and the permanent distribution of the lands amongst the khans during the Miangul rule in the first half of the last century (Yusufzai State of Swat). Landlords and their kinfolk farmers live mainly in the main valley, as they have means of communication and mobility; they speak Pashto and their principal market town (in the study area) is the urban center of Barikot. In addition to these farmers, the area is occasionally populated by seasonal workers from Sindh, who come to the area during harvest period and live in large camps on the outskirts of villages along the main valley. A third group consists of people who are ethnic Kohistani, and speak Dardic languages such as Torwalak, Khowar and Busharik, and who are related to the possession of pastures, forests and land in Upper Swat. Ajari shepherds from these areas practiced seasonal transhumance. They pass through the territories of the Yusufzai khans in the main valley directly to winter pastures in the plain S of Swat. This use of passes which is paid for in terms of products and especially timber rights (these also pass through the territories of Barikot, both via waterway and inland passage) and the related conflicts, represent the main moment of interaction between the society of the main valley and the population of Kohistan. A fourth group represents actually true stable stakeholders of the khans of the Barikot area: the Gujari communities of the upper valleys of Kotah and Kandak. These depressed areas and the valuable conifer forests in the high mountains are considered valuable common pool resources, but both nevertheless are the patrimony of the khans of Barikot. They are awarded to semi-nomadic communities, Gujars and Ajars, which have been in the area for several centuries, which have come and gone during their various phases of displacement from Kashmir and Punjab. The Gujars are semi-sedentary pastoralist group, who also cultivate products typical of non-irrigated terrains. The Ajars, a semi-nomadic community, are mainly shepherds and breeders. Both are traditionally considered by the Yusufzais as Swat’s aboriginals and slaves. They are tied to the Yusufzais by a relationships characterized by patronage and corvées, from which rises their freedom to utilize the land in exchange for services, among other things the sale of bush products (such as timber, fruit, and berries – wild grapes as well as medicinal herbs), of milk (and butter) and lamb, honey, etc. 3) Two early proto-historic rock-paintings, which might represent various cultic scenes, were documented in the sector of the area closer to the main valley of Swat river. The standard chronological arc of these paintings covers a period from the end of the 1st Millennium BCE (?) to the 2nd half of the 1st Millennium CE. However, some painted complexes remain outside this chronology: paintings where there are no representations of horses, of Buddhist architecture, and pastoral contexts; paintings, which instead contain composite ideograms and pictograms referring to an agricultural context. Late Bronze age facies (Period IV) is undoubtedly characterized by a strong expansion of agriculture and the use of new polished stone utensils such the so-called ‘perforated knives’, scythes typical of the Kashmir and Sino-tibetan area; the dominant role of agriculture is recognized both by those who hold that the Period IV culture, although not uniform was fundamentally sedentary as well as by those who believe it to be a semi-nomadic culture. Agriculture might be the only possible element of contact between the dominant culture in this phase in the valley and two Sites with rock paintings (Sites 03, 04). These shelters display highly symbolic compositions, the syntax of which is based on associations, oppositions and combinations of iconographic themes. The elaborate structure suggests co-literate cultural contexts, that is, typical of primary orality cultures that nevertheless possess a complex mythopoeietic and lexical heritage. The dominant theme in these shelters seems to be agriculture. In the open mouth of the natural gigantic face of Sargah-sar 1, three types of actors, are represented in scenes of plowing, sowing and consecration of the field. Men and women, equipped with farm implements, have the task of plowing; other men, who seem to occupy a hierarchically higher position (they have typical, upside-down U-shaped legs), have the task of sowing the plowed land which is represented through the symbol of a grid-like ideogram. Lastly a large anthropomorphic figure, with the same characteristics as the sowers, is depicted inside a large ploughed and sown field ideogram. At Kakai-kandao the presence of animals in a similar painted ritual is limited to a monkey. At Sargah-sar, a peripheral sector of the painted shelter is dominated by a heroic scene. An individual with U shaped legs armed with a shining shield is represented between two animals facing each other: an ibex surmounted by an anthropomorphic figure and a large feline; the former possibly representing the killing of the ibex. The action of the hero or deity is emphasized by a couple of dots. It is not ruled out that his solar-like shield is a chackra, like those represented in other earlier paintings In this scene we see again, although expressed through a syntactic language, an association between the themes of the feline and the ibex, symbols which are dominant in the earlier phase at Gogdara I (Olivieri 1998) and at Muhammad-patai rock carvings site (Olivieri, Vidale et al. 2006: fig. 19). The ibex icon doubtless corresponds to the hypostasis of one of the many devi of the mountain, known throughout the Karakorum-Himalaya area until recent times. At Sargah-sar 1 and Kakai-kandao 1 the main compositions are surrounded by secondary figures: archers, dancers or figures with body modifications or open hands extending outwards. The open hands theme has often been associated with a state of shamanic trance or ritual death. These observations probably bring us nearer to the meaning of these remotely located paintings. The action of painting and observing could be linked to long periods of isolation and deprivation and associated with forms of initiatic ritual. The composite central ideogram in Site 04 has been interpreted as an icon referring to sacrificial rituals connected plowing and sowing. This ritual finds its direct reference in a Ṛgvedic environment, as well as in related cults in Dardic horizon. 5) From late proto-history onwards, the paintings included in their lexicon icons linked to war, militarism and horsemanship. The icon of the horse was possibly the first foreign borrowing in the paintings’ visual lexicon. The horse icon was the eponymous animal of the Assakenoi, the group living in Swat at the end of the 4th century BCE according to Alexander’s accounts. Representations of horses started being common in Swat in the graveyard’s iconography and culture. One may suggests that the painted shelters and graveyards coexisted. We tentatively imagine a scenario where the main valley and its agricultural land were occupied (for different purposes: grazing, agriculture and funeral rituals) by the community of the graveyards’ culture, while the external areas were inhabited by indigenous communities or otherwise different. The presence of small graveyards in remote areas could be explained by an attempted intrusion, the temporary occupation of grazing areas, such as during the summer season, carried out by small groups associated with the community of the main valley, or as an expression of the process of acculturation to the new burial customs on the part of small groups. In both cases, this could explain the reduced size of the mountain areas graveyards (in Kandak), as well as the simplicity of the cysts, sometimes of a megalithic character, as well as the scarcity of the burial assemblage. The hypothesis of a partial acculturation of the peripheral community could explain the interesting modification of the symbolic heritage of the paintings, which accepts, as revealed in some Sites (such as 05 for example) in addition to the already familiar themes, icons of a culture focused on pastoralism and warmanship (as the cultures of the graveyards probably were). Another aspect converges in the same direction is the engraving of carts, banners, horses and riders, typical of rock production contemporary to the graveyards. When these icons are carved inside a Site with earlier paintings (Site 04), the remaining external space is reserved for them, which does not interfere with the previously painted space. Assuming that agriculture played an important role in the ideology of the author