Monumental Entrance to Gandharan Buddhist Architecture Stairs and Gates from Swat

The article presents a series of pieces excavated by the ISMEO Italian Archaeological Mission in two Buddhist sacred areas in Swat (Pakistan). The pieces are chosen for their connection to the theme of monumental entrances of cultic buildings. In the first case (Gumbat), the building is a shrine. In the second, (Amluk-dara) it is a Main Stupa. The pieces belong to three different entrance parts: lower sides of the stairs, decorated steps or stair-riser friezes, and decorated frames of doors. Pieces like these, which belong to specific architecture, can be hypothetically positioned in their places, allowing thus a more vivid reconstruction of the original appearance of the monuments. The decorative apparatus of the entrances to Buddhist monuments, although apparently extraneous to the religious language, is not less rich than the Buddhist iconographic programme illustrated on the stupas or inside the shrines. The second part of the article deals with the interpretation of the language of the entrance as ‘symbolic capital’ of the political élites, who were the donors of the great Buddhist architecture in Swat.

0.17 m, and always within the rise/run ratio 17/29 (r/R; expressed in cm). In Gandhara the rise/run ratio is always between 20/30 (Gumbat), 24/28 (Amluk-dara) and 25/30 (Saidu Sharif I) (data from Olivieri et al. 2014c). The ratio is confirmed by the h/l (= r/R) ratio of the Gandharan strings, which is always in the same range. Consequentely, the bodies of the stairways in Gandhara are shorter, the number of steps is lower, and the inclination is steeper. The inclination of stairways in Gandharan architecture is always ~ 45° (see Faccenna, Spagnesi 2014), while the ideal architectural angle of inclination is below 37°. In ideal stairs though, decoration will not be appreciable. Higher risers (i.e. steeper inclination) offer more visual space. Such visual space is appreciable as a whole, in perspective from the bottom level, and as single stair-riser during the climb. 9 Interestingly, at the site of Amluk-dara, which will be analysed in details in a following paragraph, the inclination and length of the lower flight of stairs was modified at a certain stage. The main stupa had two flights. The lower, the bigger one, led to a landing at the podium. From here, a second, smaller flight, rose up to an upper landing just in front of a frontal niche, of which only the lower pediment survives. In period III, which is dated to the end of the 3rd century CE, after a destructive event, probably one of the earthquakes that hit the nearby ancient city of Barikot, the entire decorative material was made anew with imported limestone heavily coated with lime plaster (see Olivieri, Filigenzi 2018). The original blue-schist flamboyant materials were removed and reused as construction material, only few fragments were found (see Olivieri et al. 2014c;Olivieri, Filigenzi 2018;Olivieri 2018).
It was in this (late Gandharan) phase of the monument that the lower flight was massively reshaped, and the entire body of the stairway was elongated. The original 'Gandharan' lower flight (the one which will be analysed below) had 21 steps (r/R 24/28), was 6 m long, and had an inclination of approx. 45°. The new flight was longer (c. 11 m), less inclined (c. 38°). Its 32 steps with a r/R ratio of 17/28 make it closer to the architectural ideal. That interfered with the symmetrical prospect of the original Main Stupa, and the final appearance was weird, with two successive flights of steps at strikingly different inclinations. The new stairway, though, was plain with no 'visual fringes' attached.
9 If the standard measurement unit was the so-called 'Gandharan foot' or Gft, which is equivalent to 0.324 m (see Ioppolo in Faccenna 1995, 168), the run was always ~ 0.9 Gft, and the rise ~ 0.65 Gft.

Finds Associated to Architecture
New information on the features of the stair-riser friezes is provided by two sites excavated in 2011 and 2012 by the Italian Archaeological Mission in Swat, namely Gumbat/Balo Kale (GBK) (Olivieri et al. 2014b) and Amluk-dara (AKD) (Olivieri et al. 2014c), two sites located respectively 5 km south-west and south-east of Barikot/ Birkot-ghwandhai, the major archaeological urban site of Swat. The two Buddhist sacred areas were both founded towards the end of the 1st century/ beginning of the 2nd century CE. Both are marked by a major architectural landmark. Gumbat's signature is represented by a still standing double-domed shrine, flanked by two same-size stupas, unfortunately almost razed down. The shrine (or the Great Shrine) is square in plan and is surrounded by a square ambulacrum, with the entrance facing east [figs 1-2]. The shrine stands on a square podium 3.3. m high, reachable through a frontal stairway 5.1 m long, provided with 16 steps (r: 0.21 m; t: 0.30 m; w: 2.5 m).
The Buddhist sacred area of Amluk-dara is marked by a colossal stupa (the Main Stupa: total height including the chattrāvali: 32.8 m [figs 3-4]) is built upon a massive podium marked by pilasters surmounted by modillions (height: 4.7 m), reachable from the north side through a massive stairway 7.20 m long, with 25 steps (r: 0.25 m; t: 0.30 m; w: 6 m).
The chronology of both monuments, which is based on the analysis of C14 data, is consistent with the earliest phases of the respective sacred areas (end-1st century CE). 10

2.2.1
The Series from the Main Shrine of Gumbat A series of five pieces were excavated next to the stair of the Great Shrine of Gumbat [figs 5-9]. 11 They belong to the earliest sculptural production of the site, and are certainly associated with the main monument. 12 All the five pieces were found in the same layer 4 in the space between the southern (left) side of the stairs and the side stupas 3. We cannot consider the location of discovery as a primary deposition, for example after a collapse, since it was a nothing but the refilling of a modern pit (pit <109>) dug most probably in 1938 by E. Barger and Ph. Wright. The two (or rather Wright) 13 largely plundered the site and after their dig left several dozens of pieces buried in pits, after earmarking the best ones for the VAM and University of Bristol who sponsored their trip to Swat. 14 In fact many of the pieces belonging to the figured friezes of the minor stupas found as leftovers of the 1938 dig match those which are preserved in the VAM (Ackermann 1975). 15 In the collection brought to the United Kingdom by the Barger and Wright there are no other pieces such as the five we recovered in pit 109. Probably the reason for their being discarded was the apparent crudeness of the carving compared to the visually rich and crowded vivacity of the genre scenes of the minor stupas, with their meddling ladies at the balconies, quarrelling actors (we will soon come back to them), flying amorini etc., which attracted the interest of the two Britons. 16 By the way, the isolated metopes of Gumbat somehow recall the upper frames of the merlons of Surkh Kotal (Tissot 2006, 59-62). The associations between the artistic school and ateliers active in this important site and those active in Swat, especially in the Barikot area, should be carefully studied in future. 17 12 Some of them were briefly re-examined in the framework of a contribution presented at the second international workshop of the Gandhāra Connection Project at the University of Oxford in 2018 (Brancaccio, Olivieri 2019, 130, figs 16-17). The reader can find there the previous reference to the site, architectural studies, excavation reports.
13 E. Barger left for Kabul in mid-Summer as soon the permission arrived, leaving to Ph. Wright the task of conducting the excavations in Swat (Barger, Wright 1941, 12).
14 The 1938 expedition to Swat and Afghanistan of the two was sponsored also by the Royal Geographic Society, the Royal Society of Arts and the University of Cincinnati (Barger, Wright 1941, iv). The two researchers were not exactly archaeologists. In fact, they dug random pits rather than conducting regular excavations at the sites where they halted during their journey. Nevertheless, their mind-set was innovative. For example, they were amongst the first to stress the need of paleoclimatic studies and settlement excavations, and they had quite an eye for the then so-called 'minor finds', e.g. the terracotta figurines, as potential markers for a more reliable reconstruction of the chronological sequence.
15 On this topic see Brancaccio, Olivieri 2019. 16 This is how they describe these genre scenes: "Several small friezes are carved with figures grouped in pairs between Indo-Corinthian pilasters [...]. With their togalike dress and declamatory attitudes, these figures smack strongly [of] the Roman forum. Such friezes bear more than a superficial likeness to the carvings on early Christian sarcophagi, with their rows of saints grouped in pairs beneath the arches of a colonnade" (Barger, Wright 1941, 17).    The five pieces recovered in pit 109, despite the lack of a proper context, can be attributed to the stairs of the Great Shrine for two reasons: proximity and measurements. To that one should add also the absence of a real alternative. Of both the stupas flanking the Shrine only the layout is preserved; they were almost razed in ancient times, and the stones re-employed elsewhere; the Shrine was amazingly preserved. 18 The measurements are consistent with the reconstructed height of the steps (c. 0.20 m), but of course they could have also matched the risers of the two side stupas if their stairs had been preserved.
Interestingly, both at Gumbat and at Amluk-dara, the stair-risers have the tenons on the lower side, not on the upper side, as in most of the published cases. Evidently, the assemblage system in use in these two sites was different from the one commonly followed in Gandhara.
At this stage, it is possible that the decorated patterns may reveal something more interesting. As a matter of fact, the stone and the decoration of four of the pieces from Gumbat/Balo Kale (namely GBK 54, 52, 61 and reg. 409, see fn. 9) is identical: same grey/greenish schist, same lower moulding formed by two flat rebated fillets, same decoration of isolated metopes, and same treatment of the flat surfaces left uncarved. GBK 54 are both part of the left side of a stair-riser: the framed metope in both sides presents a half, closed tuft of acanthus, and, to the left, a flat partly dressed zone, which was meant to be inserted either behind the string, or under the body of the railing.
GBK 50 [ fig. 5] differs slightly in the quality and tone of the stone, but it is still a grey schist. It has the same height of the others and the same flat-moulded base with two rebated fillets. The decoration differs: a framed left-handed ivy scroll. Normally one would have determined it as a scroll of pipal branches (Ficus religiosa, L.). As rightly pointed out by Tanabe (2017-18), in some cases the 'cordiform' vegetal pattern of these scrolls and decorative patterns with cordiform leaves should be interpreted as ivy scrolls or parts of them (see also Brancaccio 2018, 165). 19 So far GBK 50 is probably the best evidence recovered to support Tanabe's interpretation: here the presence of the corymbi definitely points to the representation of ivy. The Dionysiac 18 It was found used as a shelter and shed (Stein 1930, 13;Barger, Wright 1941, 16).
Later it was occasionally used as a mosque, especially in periods of drought (local informant: personal communication).
19 Ivy grows forming scrolls, with spiral-shaped sprouts (from which the Latin term Hedera helix, L.), while pipal branches hang laterally. Ivy (Hedera nepalensis, L.) is common in the mountains of Chitral and Swat as reported also by Curtius Rufus, VIII, 10, 13: "Multa hedera vitisque toto gignitur monte [...]". value of ivy in general and of corymbi in particular is too well-known to be worth of further elaboration in this context. 20 The importance of this motif either with pipal or ivy is also stressed by its presence of the decorated band on the kaftan of Kanishka (?) in the celebrated statue II from Surkh Kotal (Schlumberger, Le Berre, Fussman 1983, pl. 60).
Moreover, almost all the Gandharan pieces with ivy scrolls illustrated by Tanabe -on the basis of their shapes, features and measurements -are actually part of the decorative apparatus of the stairs: e.g. the stair-risers from Butkara I (Swat), and other sites of Gandhara (Tanabe 2017-18, figs 1, 9-11), a stair-side element ( fig. 8, from Gandhara, Peshawar). A further interesting piece of comparison comes from the Main Stupa of Saidu Sharif I. This monument is extremely important since it is possibly the earliest amongst the 'Gandharan' stupas with podium and frontal stairs (Faccenna, Callieri, Filigenzi 2003). 21 Certainly, it was a stupa celebrated from Gandhara "to Miran" (Filigenzi 2006; see also Provenzali 2016). We do not have much preserved from the monumental stairway, if we exclude the railing and part of the first steps of the first flight, leading up to the podium (see Faccenna 2001). However, one single fragment [ fig. 10] in green schist is preserved, pertaining probably to a stair-riser of the second flight leading up to the pradakṣiṇāpatha. This piece (SSI 21) shows a beautiful left-handed ivy scroll, with spiral-shaped sprouts. In this case also the tenon is on lower side as demonstrated by a mason-mark, the Kharoṣṭhī akṣara 'ga' carved on the lower fillet (personal communication by Stefan Baums).
To conclude this section, we should briefly consider the hypothesis that four GBK pieces were part of the same stair, which possibly was the one of the Great Shrine; if the fifth piece with the ivy scroll (GBK 50) was part of the same structure, it should be positioned ei-20 Bacchus corymbifer. Again, see Tanabe 2017-18, 101-2. 21 This and other themes will be further elaborated in a forthcoming study recently financed at Ca' Foscari University of Venice (Research Project: Tecnologie della pietra e dei cantieri nel Gandhara: Saidu Sharif I). ther at the bottom or at the top of the flight; most probably at the top, on the last step. The first step as usual, and because of its higher moulding, is a bit higher than the others.

2.2.2
Side-Elements and the Stair-Riser Series from the Main Stupa of Amluk-dara 22 We have already mentioned how rich and frequent is the repertoire of Classic patterns in the step-rise friezes, and especially in the strings (or stringers). 23 We should here introduce another element, which is part of the architectural structure of the monumental stairway: the stair side-element (AKD 97 and AKD 98 [figs 11-13]). Pia Brancaccio in a masterly study (2014; 2018) has clearly demonstrated that many pieces mistakenly interpreted as 'bases' or 'stools' or 'pedestals', were actually the typical bases for the newel of the railing of the monumental stairs in Buddhist monuments: the so-called side-elements (Faccenna, Filigenzi 2007, 93, pl. 58.2). It is possible that these Gandharan stairs' side-elements derive from specific antecedents in Mathura. 24 However, these bases, became a typical Gandharan production. It looks as the latter was an elaboration, a kind of projections at ground level (as schematized in [ fig. 14]), of the toraṇa arches, the monumental gateways of the earlier Andhra and North Indian Buddhist architecture. The metamorphosis can be explained by the fact that the typical Gandharan stupa, being on a podium, requires a stairway. These side-elements, carved in rare blue schist (see below), have often the same di- the length can be shorter. They are typically heavy massive parallelepipeds with rounded ends, one of which is carved, the other being left plain with the inner part undressed. The long sides of these pieces are treated differently: one is beautifully carved (external), the other is left undressed (internal) except for the frontal (or visible) portion, which is carved. On the top, which is treated differently, we often find sockets or hollow depressions dug to house vertical parts. The scheme is always the same; the only difference is that we have pieces carved on the right or left sides. From the analysis of these elements one can immediately conclude that these are pieces that were meant to be placed at the bottom, directly on the floor, either to the left or to the right of a structure, and that they were meant to support a vertical piece, like a newel or a front pilaster of the railing. One more element is presented by a beautiful side-element, confiscated and preserved in the Kabul Museum (Nettl 1991, fig. 5;Tissot 2006, 529, figs 1a-b). 26 'Persian dancers' and musicians (along with beautiful men and women drinking and making merry) are often represented on stair-risers. Other exquisite examples with scenes associable to sea thiasoi come from Shotorak (Tissot 2006, 327, 858.64;328, 863.67, 863.69). 27 If the toraṇa gates are the marker and the signature of the Indian architectural schools, the new stairs and the podia are the signature of the Gandharan religious architecture (Kuwayama 2002). The architects of Gandhara, with the revolutionary introduction of monumental stairs and podia, offered a privileged space to express the visual imagery the 'Graeco-Iranic' culture of the local courts (the Oḍi, the Apraca etc.), who was patronising the construction of the great stupas and shrines, as it happened certainly at Amluk-dara. The relationship between the flamboyant imagery of the stairs and the 'Graeco-Iranic' local culture will be elaborated in the second part of this article. 28 Let me detail a bit more on the problematics presented by the stair of the Main Stupa of Amluk-dara (AKD). This Stupa had two stairs: the 26 On this see the seminal contribution by Lo Muzio (2019, 80). Lo Muzio, though, still follows the identification of these pieces as pedestals. Tissot describes it as if the piece were sculpted on both sides: "Dionysiac scene on the pedestal of a Buddhist image. On either side of the pedestal, between lion-paws, three men and a woman playing musical instruments" (2006,529). A closer analysis of the scene and of the fractured and chipped parts shows that the two images reproduce the same side, which means that the back was most probably not sculpted. Therefore, the piece can be legitimately included in the list of the side-elements of monumental stairs.
27 At the Asian Art Week 2020 in New York stair side-element was presented as "a gray schist relief with lotiform motif" (no. 604: 0.23 h.; 0.438 l.). It is decorated with a flamboyant palmette enriched by ivy scrolls, and corymbi or pomegranates. Unfortunately, the front face, where one guesses the presence of a standing figure, is not reproduced. That is typical, since this class of materials is generally misunderstood, and labelled as 'stools', 'footstools, 'thrones', or generic 'reliefs'.
28 We use the term 'Graeco-Iranic' in the sense established by D. Faccenna in his article on Kuh-e Khwaja (Faccenna 1981, 94-5 lower, or first, and the upper. The lower one -as we have seen -was reconstructed and elongated in period III of the structural phasing of the monument and stupa terrace. Traces of the older stair (both architectural and sculptural) were documented in situ. In fact, amongst the few original schist pieces found almost in situ, we recovered in 2012 the two massive stair side-elements we briefly introduced above [figs 11-13]. The decoration of these pieces shows on the external sides a richly elaborated open flame palmette with tuft (which is partly replicated on the short visible part of the inner side). The front sides differ: on the right element there is a flexuous standing Aphrodite, while on the left element there is a standing Herakles leaning on his club. We recall these pieces because they are part of the elaboration on the evidence provided by the stair-riser friezes from this monument (and here we are absolutely sure that the pieces belong to the Main Stupa).
Three stair-riser pieces were recovered (notwithstanding the site having been heavily plundered for ages): AKD 61 They were carved in the same kind of stone of the two side-elements, a pretty rare variety of compact blue schist, a kind of stone which was evidently the 'signature' chosen by the architects of Amluk-dara, exactly as green schist was the 'signature' chosen by the Maestro of Saidu Sharif for the Main Stupa at the latter site. 29 All the three pieces show the same height (0.24 m). Their stratigraphic provenance is absolutely unreliable since they were found in the refilling of a modern pit dug by treasure hunters. The position of this pit is interesting though, as it was located to the left of the stair, and it possibly included leftovers thrown out by the diggers while they were working on the stair's body.
The scheme is again the same as observed in Gumbat, marked by deeply carved framed metopes; the only difference is that here the space between the metopes was filled with figurative, iconic representations. AKD 61 [ fig. 15] is the right side of a stair-riser: to the right a framed metope with a schematic full-blown open lotus with one corolla, stamen in evidence, and framed round rosetta-like pistil. The scene on the left shows traces of vegetal elements, a tree, maybe a standing figure. AKD 60 [fig. 16] is again the right side of a stair-riser: to the right a framed metope with a heraldic animal, a prey-bird or a gryphon represented in profile. The scene on the left shows again traces of vegetal elements, a tree, followed by a round element, maybe related to a figure. AKD 64 [figs 17, 18] is the most interesting of the three pieces. It is part of a stair-riser, but not the final one. In fact, the study of the back side of right end of the piece reveals behind the metope the presence of a vertical rebate where the following portion In the scene we clearly see a male figure facing left, with head and shoulder enveloped by an ample velificatio, sitting or riding a triton or an anguiform dragon. To her left is barely visible a sitting female seen from behind. The figure represented on the back holds an elongated object in her/his right arm (a helm of the drapery? A fish?). The analysis of this and other associated images (e.g. a male god velificatio with 'Persian dancer' from our excavations at Gumbat, GBK 34) will be elaborated in the second part of this article.

Building Doors or Gates
Before concluding my part, I would like also to present a serendipitous discovery, which is somehow connected to these architectonic parts associated to monumental entrance and gates of Buddhist buildings in Gandhara. In May 2013, during a survey in the valley of Kandak, I took a halt in the mansion of an old acquaintance of mine, Mr Muqam, an affluent khan, one of the few Pashtun landlords of that valley, a gentleman and exquisite host. It was after a pretty heavy 'chicken party' that Mr Muqam led me to a backyard of his house where he showed me an extremely beautiful Gandharan frieze. Honestly, my reaction was slow. I did not understand immediately its importance. To partly excuse myself, it may be worth saying that when I find myself in similar situations, I am always in doubt whether to show or not my interest. In circumstances like these, I am always afraid of giving the impression that I am interested in buying the piece (which for a professional archaeologist is not exactly the best impression to give). So, I took some photographs, the measurements, and then I left the courtyard with a dismissive smile. 30 Only after some years did I realise that in that hot afternoon I had seen a very beautiful example of Gandharan architrave of a building gate, or an architraved door of 'Western type' (Faccenna, Filigenzi 2007, pl. 57 The object is c. 0.90 max. long, c. 0.22 high, and c. 0.10 thick. It is broken to the left, while to the right it presents an inclined side, where it is meant to join to a jamb. The stone is a pretty refined, a very compact variety of chloritoschist, grey to deep grey in colour. The markedly concave upper cornice is decorated with a ABAB 1 scheme of vegetal motifs: row of acanthus tuft (A), open palmette (B), and open palmette with trumpet-shaped central shoot (B 1 ) (Kökdemir 2004). Below, on a rebated band, is a row of reversed border petals; this is followed by an empty rebated high band marked below by a sunken half-round, a horizontal flute, and a projecting fillet. Below is a band with a right-handed ivy or grapevine scroll marked by a projecting fillet, a half-round. The lower band is decorated with a eight-petal rosette (type) within filletted lozenges and half-rosettes in the resulting triangles.
The object is clearly the architrave of a tapered door, the typical conventional door reproduced in architectural representations in Gandhāra. We do not know where the object was found. The area surrounding the village of Kandak is full of majestic ruins, the best preserved being those in Dur-bandai on the eastern mountain uphill the village. Here were documented the ruins of "a building with a false domed roof, probably a vihāra" (Olivieri et al. 2006, 111-12, fig. 65). This type of buildings (vihāra or shrines) are accessible build-30 However, I forgot to take a picture of the back side and to note down the presence of sockets or tenons.

Postcript 1
When this article was in its final stages I saw an interesting contribution by Tanabe to Rienjang and Stewart 2020, where it was presented as a 'vertical relief' (Tanabe 2020, 88) whose caption read "Gandhāran Dionysiac relief with peopled vine-scroll. H. 124 cm, c. second to third century AD. Boston, Museum of Fine Arts, Charles Amos Cummings Fund 39.36" (Tanabe 2020, fig. 5). The article is online: I invite the reader to see that beautiful piece, which is nothing but a jamb of a door, exactly how the piece described above is a door's architrave. 32 The presence of an angular rebate shows clearly that it is a left jamb.
LMO 31 From this, we may assume that door architraves were typically composed of two pieces fixed to each other and to the wall, and supported by the jambs.
These pieces are frequently misunderstood and described as "relief[s] with foliate scroll". At the Asian Art Week 2020 (New York) was shown a wonderful piece, the lower segment of a door's left jamb, features a lower metope decorated with ivy or grape scroll, and a continuous vertical scroll with palmettes and pomegranates. The vertical decoration is framed by fillets, and on the left (closer to the wall) by a thicker set formed by two projecting fillets. Staging Power in Threshold Space: The Ambiguity of Access

A Preamble to the World of the Buddha: Gandharan Stairway
The enquiry into the correlation between Gandharan Buddhist monuments and their figurative programme have yielded -with the exception of Saidu Sharif main stupa (see Faccenna 2001) and the stupa of Sikri (Foucher 1903) -quite inconclusive results, and this situation is unlikely to change. Indeed rearrangements, substitutions and removals of the figurative apparatus from the bodies of stupas and shrines, both in antiquity and in modern times, have prevented comprehensive understanding of the overall visual communication of such structures beyond their monumentality. Moreover, even though general guidelines for figurative programme in sacred places might have existed, the selection, disposition and combination of specific elements, subjects and narratives on the body of religious monuments had to cope with several interrelated factors such as local repertoires, personal preferences of the patrons commissioning the work, religious features of the place, and craftsmen's skills, besides more practical matters. 33 In this sense, every complex had a micro-programme, so to speak, and the mobility and networks of craftsmen must have played a key role in the diffusion of motifs and development of interand intra-valley schools (see Faccenna 2001, 197-8; see also Brancaccio, Olivieri 2019). Despite this, the almost coherent figurative repertoire appearing on stair-risers invites reflection on the internal logic of theme selections at specific sectors of Buddhist monuments and to look for possible social or religious dynamics associated to this phenomenon.
Significantly, from the general repertoire of stair-risers, it seems that almost all the subjects on display depict narratives that lie outside the temporal domain of Buddha Sākyamuni's historical life: mainly jātakas, unidentified narratives and scenes displaying music, dance and drinking iconography or marine figures. Although jātakas are not a common subject in Gandharan art, their location on stair-risers seems to be quite popular, as suggested by the friezes found in situ in Jamal Garhi in the Mardan district and in Chakhili Ghundi near Hadda (Zwalf 1996, 56). As highlighted by Zwalf, most of the jātakas "are traditionally situated by Chinese pilgrims in or near Gandhāra" where several stupas are said to have been erected to mark and commemorate a sacrifice of the Bodhisattva (Zwalf 33 Although the type of sequential narration for the life of the Buddha on the body of main stupas was probably the rule (Taddei 1993 A larger number of stair-risers, however, follows another pattern, depicting 'non-Buddhist' motifs. This is so in the case of the widely debated 'Dionysian scenes' variously read as bacchanalia, religious events or urban rituals (e.g. see Carter 1968Carter , 1992Brancaccio, Liu 2009;Galli 2011;Filigenzi 2019). While processions of nāgas and nāgis and figures in Indian attire are not unusual (see for instance Jamalgarhi, BM 1880.36, 40), a large number of stair-risers yield a distinctive Hellenistic-Central Asian/Iranian (hereafter 'Graeco-Iranic', see above) blueprint in both iconography and composition of the scene (Soper 1951;Rowland 1956;Goldman 1978). 35 Drinking scenes are animated by satyrs and richly-dressed women, dancers and musicians, often performing an Iranian dance and mostly playing instruments of western origin while wearing Hellenistic, Scythian and Central Asian dresses (Lo Muzio 2019). Interestingly, most such stair-risers seem to come from northern Gandhara, in particular from Swat, Buner and Dir, which is also the area from where reliefs with dancers performing the 'Persian snap' apparently come (Lo Muzio 2019, 77).
Another, less celebrated but recurring subject, which recalls a Hellenistic prototype, is that of marine deities and sea-monsters. The most famous example is from the series of bearded and moustached men holding paddles, probably coming from Swat or Buner (Marshall 1960, 37, fig. 47; MMA access no 13.96.21), 36 but also images of triton-like figures are frequent (e.g. BM access no. 1880, 57). Finally, an unmistakable recall of western repertoire is offered by the myth of the wooden horse attested, for which a possible 'Buddhist translation' has been suggested (Foucher 1950;Stewart 2016).
This Hellenistic imagery is not restricted to the stair-risers but encompasses the whole stairway's sector including side-elements and triangular strings. The formers are mostly decorated with vegetal scrolls, drinking scenes, marine figures and classical deities (Brancaccio 2018, 165-7), while the latter can feature a variety of marine beings, such as triton-like figures or ichthyocentaurs and hippocampi whose anguiform bodies easily fits into the triangular field of the architectural piece (e.g. MMA access no. 13.96.19). 37 34 On the re-localization of jātakas in Gandhara see Neelis 2019. 36 One piece is at the BM (access no. 1889,1016.1), the other at the MMA (access no 13.96.21). In the former, one of the men holds what is usually defined as a dolphin seen frontally, while in the latter, the animal held by the man seems to be a snake. Indeed, on the Gandharan stairways, the so-called 'Graeco-Iranic' lexicon and syntax is striking when compared to the other sectors of the monuments. The direct relation between marginality with regard to Buddhist narrative and marginal space may seem evident. 38 Indeed, as noted by several scholars, foreign and 'non-Buddhist' motifs, though not uncommon, are "set aside from the main and most sacred narrative" (Lo Muzio 2019, 72) and are often accommodated in small continuous friezes or on pedestals of statues with an ancillary role. In the case of stair-riser, instead, such repertoire steps into the foreground.
Though properly outside (or better, on the approach to) the religious focus, from the point of view of the visitor or devotee approaching the main monument from the court of the stupa, the monumental Gandharan stairway actually represented a sort of visual axis converging attention on the symbolic vanishing point of the sanctuary, the dome (see also Brancaccio, Liu 2009, 229). Yet, the particularly steep inclination of Gandharan stairways (~ 45°) guaranteed visibility to the stair-riser friezes from a lower perspective and the eyes were inevitably focused on them while approaching the monument. I argue that the classification as 'marginal' here is almost certainly misplaced. Rather it seems that monumentality and architectural centrality make the stairway a real stage. Therefore, motifs of foreign and 'non-Buddhist' origin -not necessarily classified as such by contemporary users -illustrated on it would have actually introduced, not only the main frieze on the body of the dome, but also any ritual activity and event taking place at the monument.
These observations raise the questions: why were 'non-Buddhist' themes with a rather Hellenistic prototype put on stage? Was the figurative apparatus on the stairway and that on the rest of the stupa -displaying more interest in Buddha Sākyamuni's historical life -always serving the same message and the same master? Have urban phenomena influenced the Buddhist figurative programme and architecture?
In the following pages the few motifs known from the main stupa of Amluk-dara will be contextualised within the socio-economic scenario of the middle Swat valley. By arguing that in northern Gandhara, Buddhist spaces offered to urban élites both logistic support for the utilitarian exploitation of rural territories and an arena to permanently materialise local authority, I suggest that: (a) Buddhist monumental stairway were often used by the urban actors for staging their power, marking and solidifying their role as patrons and their 38 But not 'polluted' in the sense of Behrendt (2007, 27): "That this sculpture was placed in a polluted context -where devotees, presumably barefoot, would have walked -perhaps explains why the panels were carved with non-devotional subject matter." e-ISSN 2385-3042 57, 2021, 197-240 224 social position by using Hellenistic imagery as 'symbolic capital'; (b) the introduction of a monumental steep stairway in Gandharan Buddhist monuments was an architectural response to new urban aspirations and necessities.
The site of Amluk-dara will be used as case study to test the hypothesis that interplays between Buddhism and the urban in the first centuries of the Common Era influenced the spread of specific artistic models in Buddhist monuments as well as their architectural layout.

The Hellenistic Motifs on the Stairs of Amluk-dara
The starting point for our reflection is offered by a rather badly preserved piece of stair-riser from Amluk-dara, AKD 64 [figs 17, 18]. Despite its fragmentary state, the piece is quite interesting when considered together with the stair side-elements and the subsequent rearrangement of the stairway.
The piece shows a Nereid and a male figure, seen from threequarter back (left) and three-quarter face (right), riding an anguiform monster while looking at each other. The figure on the right is actually hard to identify, but seems to be a male wearing a dhotī and partially framed by an ample velificatio, which usually characterises gods or Nereids (see Postscript 3). The details of the Nereid are also hard to define with any degree of certainty (see above). The scene evidently derived from the Hellenistic iconographic model of the marine thiasos so often represented in Western art. A reference to the marriage of Poseidon and Amphitrite would be rather a wild guess but the more general motif of Nereids on a ketos is unmistakable. Although, as mentioned above, references to figures connected to water are frequent on stair-risers, the representation of the marine thiasos is quite exceptional in Buddhist complexes.
While comparisons for such iconography are unknown on Buddhist monuments, the same iconographic model -though in different formula and more clumsy style -seems to have been quite popular in the urban context. Nereids or male riders on sea-monsters (sometimes represented alone) with anguiform tail and the front part in the shape of a dragon, horse, lion or ram were depicted on the so-called toilettrays found in a variety of sites in Gandhara and beyond (for an updated bibliography see Falk 2010). These are small round stone dishes with a decorated field on the inner side, apparently only found in cities. Although the exact function of toilet-trays remains a controversial matter, there is a general consensus on the ritual use of this artefacts as liturgical or libation vessel in urban contexts. Most of the toilet-trays found in archaeological excavations come from the city of Sirkap at Taxila (Marshall 1951), where toilet-trays, as highlighted by Coningham and Edwards (1997-98, 58) and Michon (2015, e-ISSN 2385-3042 57, 2021 Luca Maria Olivieri, Elisa Iori Monumental Entrance to Gandharan Buddhist Architecture. Stairs and Gates from Swat 225 171-83), were found in connection with 'domestic' sacralised spaces along with other artefacts (stele of local goddesses, 'votive tanks' and oblong dices) hinting at a rituality related to worldly matters, such as fertility, prosperity and protection of children.
In general, the motif of marine monster derived from western iconography (ketos) seems to have met a certain success in Central Asian regions where water, and especially rivers, played an important role in both economic and religious matters, 39 however, the motif of the marine thiasos does not appear to be attested there. In the north-west of South Asia, where life largely depends on the good management of water (Olivieri et al. 2006, 131-5), the iconographic repertoire for water semi-divine beings (nāgas) and water monsters (makaras) has always been quite rich. 40 It is therefore peculiar that in some specific context Gandharan craftsmen gleaned the marine motif from the Hellenistic models -probably diffused in Central Asia through small luxury artefacts (e.g. metal objects, textiles etc.) and through travelling craftsmen -rather than from the well-known Indian iconography. The intentionality of this selection certainly had some specific reasons, suggesting that the meanings and values these figures were conveying went beyond meanings that could have been expressed by the figures of nāgas or makaras. 41 To complete the picture, it is worth remembering that the two stair side-elements of the same stairway also present a Hellenistic decorative and figurative pattern. Brancaccio's reconstruction demonstrates that at the base of the first pillars of the railing, in fron- 40 Representations of nāgas are diffused both in Indian and Gandharan art. Makaras, the mythical composite creatures with a propitious and apotropaic function, are represented in the volutes of toraṇa lintels (like in Bharhut), within the medallions on vedikā (like in Bharhut, Amaravati and Nagarjunakonda), and later in the form of 'makaraarch' above the entrance of monuments and niches in several Buddhist sites of South Asia (5th to 7th century), and Ajanta, Ellora etc. For further bibliography see Bautze-Picron 2010. The Indian makara was probably at the origin of the volutes with 'monster heads' that characterise the arches and semi-arches of the so-called 'false-niches' (Iori 2018, 115). A makara also decorates the terminal part of Kanishka's club in the statue from Mathura (Rosenfield 1967, pl. 2b).
41 On the toilet-tray, the image of Nereid or male rider on ketos often lacks a coherent narrative feature that enables any reference to specific (or known) mythical narratives. At the same time, we cannot ignore the possibility that these isolated figures on ketos could have evoked, to the ancient users, well-known tales and narratives. Anyway, the connection to the feminine and erotic component is likely, especially if we consider both the assemblage in which toilet-trays were found and the general figurative repertoire displayed on these set of objects, often recalling mythological subjects, related to the feminine sphere and drinking couples. tal position, are two figures sculpted almost in the round following "the classical type of Herakles and Aphrodite known in Gandhara through coins and precious objects" accompanied by the use of decorative scrolls that "evokes the ivy branches often found in contexts associated with images of the god Dionysos" (Brancaccio 2018, 165; [figs 11-13]).
The pair Herakles-Aphrodite and the marine thiasos from the stairway at Amluk-dara confirm the Hellenistic pattern attested in several other pieces known from museum collections, yet adding new interesting elements. Although data on the figurative apparatus are sparse, the stupa of Amluk-dara, located in the well-explored area of the middle Swat valley, offer us the opportunity to refocus the orientation of studies on step-risers by broadening their perspective beyond the 'art historian' approach.

The Socio-Economic Context of Amluk-dara: The Rural Landscape as Source of Empowerment
The main stupa of Amluk-dara [figs 3-4] was constructed at the end of the 1st century CE (Olivieri 2018, 60-7) when Swat was already under Kushan control. The site, located along the main track leading up to the sacred Mt. Ilam, is the largest stupa ever excavated in Swat and its construction, together with those of the coeval complexes of Tokar-dara (Faccenna, Spagnesi 2014, 331-7) and Gumbat ([figs 1-2]; Olivieri et al. 2014b, 255-319) was most probably patronised by families ruling the area. From several epigraphic sources (see Baums, Glass 2002;Baums 2012) we know that between the mid-1st century BCE and the 1st century CE, the territories of Swat included within the kingdom of the Sakas and then the Kushans, were locally controlled by clients known as Oḍi-raja. 42 Their role as clients of foreign kings is made clear in the inscription of the Oḍi king Senavarma (ca. mid-1st century CE) who mentions his political brotherhood with the scion of the Kushan king Kujula Kadphises, thus referring to a political pact with the new ruling entity. Although Senavarma is the last Oḍi-raja known from inscriptions, we can assume that this successful strategy of 'feudatory' alliance continued until the political upheaval of the Kushans (mid-3rd century CE) since no particular social changes seem to occur in the area by that time.
The alliance matrix interwoven by Oḍi-raja in the first centuries of the common era ran not only on a vertical level (clients-foreign kings), it also included horizontal partnerships with local agents. Archaeological evidence, dedicatory inscriptions and textual sources attest that the progressive religious dominance assumed by the Buddhist communities (or saṃgha) in Gandhara was strongly associated with the patronage activity pursued by the local urban élites. In particular, between the 1st-3rd centuries CE, the rural landscape of Swat -before used as graveyards and by marginalised communities -radically changed under the increasing pressure of the Buddhist communities that progressively intruded into the ecological space of the mountain people by acquiring both their vital and ritual spaces (on the 'Dardic' communities, see Olivieri 2015). However, this is only one part of the story.
The fertile territories of the countryside in fact, while becoming Buddhist land, were at the same time transformed into the economic pool of the valley. The foundation of hundreds of Buddhist establishments in the proximity of strategic locations (mountain passes, surveillance points, springs) and fertile lands along with the construction of hydraulic infrastructure (dams, aqueducts, pit-wells, tanks) was an intensive phenomenon in Swat and the hypothesis that Buddhist communities managed the agricultural lands and controlled trade routes through mountain passes on the behalf of local élites in exchange for direct commitment seems to be substantiated by archaeological evidence (Olivieri et al. 2006). Indeed, one might reasonably claim that, in Swat, 'domestication' of the saṃgha and 'domestication' and exploitation of the landscapes by urban community went hand in hand. 43 Over this process the Buddhist community assumed the role of an essential 'human infrastructure'.
As we move through the interaction between the urban and religion, 44 we find ourselves in a tangle of triggered phenomena that cannot be exhaustively discussed in this paper. However, what directly concerns us here is that Buddhist religion became, from the 1st century onward, the main sources of empowerment for urban élites. Buddhist space and rural landscape (exploited under the supervision of the Buddhist community) were transformed into a competitive ground where both economic power and social prestige were built. In other words, religion created for the urban agents a market where both economic and symbolic capitals could be invested.

Technologies of Power
As it has been pointed out by DeMarrais, Castillo and Earle (1996, 19) in their reappraisal of power strategies "Monument and ordered landscapes domesticate unused territories and symbolize the appropriation of space (Kus 1982), organizing and materializing social relationship and boundaries".
Materialisation of power, however, is not only about monumental architecture. In fact, monumental buildings as "public material embodiment of the power" (Trigger 1990, 126) were embedded into a constellation of practices and material culture instrumental to social and political strategies. Among others, dedicatory inscriptions, artistic and craft productions, written and oral narratives (sometimes regardless of their adherence to fact) 45 were used as tools to materialise and permanently fix discourses on power, vertical social relations and boundaries. Buddhist art and architecture, as urban phenomena, were probably not excluded from this process.
My basic argument here is that the recall of a Hellenistic imagery 46 on step-risers and the new architectural vision of the stairway in Gandharan monuments were an integral part of a wider process of political and social strategies which took the form of both spatial practices and targeted building interventions.
As for the first argument, one should say that the use of the Hellenistic models in Gandharan art as objectified mark of social distinction by the bourgeois is a sociological reading which, launched by Taddei (1969, 156), 47 did not receive due follow-up in Gandharan studies, much more concerned (of course with some exceptions) with stylistic and iconographic analysis sometimes trickling into obsessive mannerism. This is not to say that an art historian approach is not fundamental to this field of research, which it is indeed. Howev-45 In this regard, we may also mention the attempt by the Oḍi kings to artificially draw through inscriptions and art a direct connection with the Buddha himself. Senavarma's family claims a genealogical connection between the Oḍi-raja and the Ismaho royal family, the lineage of the Buddha Śākyamuni (Salomon, Baums 2007). It is probably not by chance that king Utarasena, mentioned in the Senavarma inscription as ancestor of the Oḍi kings, was included in the panels of the frieze of the main stupa of Saidu Sharif while bringing back his share of the Buddha's relics on the elephant (S 241;Faccenna 2001, 73). The construction of religious memories and identities was instrumental in solidifying the relation between élites and Buddhist communities by also fixing the privileged role of the former in the symmetric alliance.
46 On the complexity of discussion on Hellenism in Gandhāra see Filigenzi 2012. 47 "We cannot of course content ourselves with a bare description of convergency or derivation phenomena in iconography and style. Nor can we accept facts as mere links of a chain of stylistic evolution under the impact of a foreign artistic culture [...] Gandhāran sculpture is to be considered as the art of few social groups that needed a foreign model enabling them to distinguish themselves from the majority of the population" (Taddei 1969, 156). er, if we want to gain any insights into the social and religious world of ancient Gandhara we should shift our approach to a different level of magnitudes and scales.
In an article adopting new approaches, Galli (2011, 281-4) sharpens Taddei's argument on the social function of the Hellenistic imagery in Gandharan art by introducing the concept of 'court imagery' and 'symbolic capital' borrowed from sociology and widely applied in studies of Antiquity in the Mediterranean area. 48 His approach has had more or less similar success as Taddei's. The main challenge to widening the perspective of Gandharan studies is due to an objective (and frustrating) lack of archaeological contexts for artefacts, and a general apathy to cities where people conceiving, transforming, using and visiting religious buildings temporarily or permanently lived. 49 As briefly sketched above, in the Swat valley studies of Buddhist art and architecture, landscape archaeology and urban archaeology have been conducted side by side. 50 This offers us solid ground on which enquiries into the complexity of historical phenomena can be conducted without stumbling into free-floating arguments.
Without pretending to take any step forward in this direction, I would simply like to raise again the perspective indicated by Taddei and Galli in order to attempt, on the basis of the evidence and considerations reported above, answering the following question: to what extent and how did Buddhist monuments give urban élites the space to manifest their authority into physical reality? In particular, were figurative programmes and the architecture of Buddhist monuments shaped by urban aspirations?
Practices here had certainly played a crucial role in temporary appropriation or materialisation of power, as Buddhist monuments must have also been the setting of processions, festivals, civic and religious events (e.g. see Schopen 2014).
Nonetheless, one may also suggest that the process of selecting iconographies on Buddhist monuments was also partially shaped by the urban aspirations (Goh, van der Veer 2016) of patrons who wanted to fix and reinforce their distinctive position within the vertical 48 The first concept is drawn from Norbert Elias' work Die höfische Gesellschaft/The Court society  and the second from Pierre Bourdieu's approach to social classes' distinctions (see references in Galli 2011).
49 Lack of context, however, does not seem to be perceived as a big issue by most scholars, since new inscriptions, manuscripts, statues and friezes from the black market are often enthusiastically welcomed, announced and published. In his article on the Hellenistic court imagery Galli (2011, 296-302) suggested that the use of aulic references to Greek mythology and Hellenistic iconography on toilet-trays (which also includes the marine thiasos) has to be interpreted as the results of the paideia of the owners, the urban élites, who wanted to demonstrate their high social status in a domestic context. 51 If we follow this line of interpretation, we might speculate that displaying an image from the classical marine thiasos on the threshold of a stupa could have aimed at publicly flaunting the 'learning' and the distinctive social position of the patrons commissioning the work, though maintaining the reference to elements of fertility, prosperity and auspiciousness, as appeared at the entrances of the Indian stupas. Notwithstanding the architectural innovations, the Gandharan repertoire at its threshold is in fact still deeply rooted in the Indian tradition (see also Iori 2018). In fact, Gandharan craftsmen continued to refer to the same values but translated them into a Hellenistic imaginary: both Dionysian scenes of drinking and dancing (Brancaccio, Liu 2009, 231) and marine deities and monsters could indeed fit in the same set of values. 52 What is of interest here is the reason behind this iconographical translation of Indian values.
What was put on stage, through this process, was indeed the 'symbolic capital' of the patrons, namely those practices, narratives and symbols that, by referring to the court imagery, marked their social position.
As first highlighted by the French sociologist Pierre Bourdieu, power was culturally and symbolically created and constantly re-negotiated and re-legitimised through practices and interplays between social agents and social structures. In particular, Bourdieu, who distinguished different types of capital in the social world, says that: Social groups, and especially social classes, exist twice, so to speak, […]: they exist in the objectivity of the first order, that which is recorded by distributions of material properties; and they exist in the objectivity of the second order, that of the contrasted classifications and representations produced by agents on the basis of a practical knowledge of these distributions such as they are expressed in lifestyles. These two modes of existence are not independent, even as representations enjoy a definite autonomy with respect to distributions: the representation that agents form 51 He considers the 'Dionysian scenes' on the stair-risers a representation of the aristocratic social praxis (habitus) of a court society (Galli 2011, 302-21).
52 Drinking and dancing scenes of nāgas represented on stair-risers from Jamal-garhi (south of Swat) also would fit in the same set of values. of their position in social space (as well as the representation of it that they perform -in the theatrical sense, as with Goffman) is the product of a system of schemata of perception and appreciation (habitus) which is itself the embodied product of a condition defined by a definite position in distributions of material properties (objectivity I) and of symbolic capital (objectivity II), and which takes into account, not only the representations (which obey the same laws) that others have of this position and whose aggregation defines symbolic capital (commonly designated as prestige, authority, and so on), but also the position in distributions symbolically retranslated as lifestyle. (2013, He goes on to claim that: objective differences, inscribed in material properties and in the differential profits these provide, are converted into recognized distinctions in and through the representations that agents form and perform of them. Any difference that is recognized, accepted as legitimate, functions by that very fact as a symbolic capital providing a profit of distinction. (2013,297) Urban élites patronising the construction of the Buddhist foundation, as was certainly the case at Amluk-dara, probably felt the necessity to self-represent their social status and their closeness to knowledge and 'habitus' of the ruling court (Galli 2011, 324) through an artistic production that in itself was already a manifesto of their position in the 'distribution of material properties' through its exclusivity. One should reflect on the fact that craftsmen producing the Hellenistic court imagery were, if not a rare commodity, an exclusive class of specialists (in terms of skills, knowledge, technical devices and tools) 53 probably travelling from valley to valley through Gandhara. The very fact of having the resource for employing such (certainly expensive) craftsmen might in itself have represented some exclusivity. Moreover, the advantage to appeal to artistic production as communicative strategy is the fact that this medium strategically cuts across differences of language and religious affiliation, survives the temporality of public events or rituals and could have a wider audience than foundation inscriptions often hidden in inaccessible reliquaries. That held the possibility to amplify the receptivity of the message, making it comprehensible to the whole multi-ethnic society (from lowest to highest classes) of the north-west territories accustomed, since almost three centuries, to the Hellenistic imagery and the use of Hellenistic material culture as status symbol. 233 hypothesis is accepted, then one may actually speculate that the introduction itself of the steep stairway in the Gandharan stupa was the best architectural response to manifest the social and economic power of the emerging urban élites.
To conclude, I argue that the aspirations and social strategies of the urban élites actually played a role in shaping the figurative programme as well as the architecture of access to the Gandharan stupas.
Indeed, the Gandharan stairway seems to be the physical threshold where Buddhism, local traditions, and the urban met and compromised.

Postcript 2
In Gandhara art the Graeco-Roman iconography of velificatio, the billowing drapery above the head, is usually applied to female figures like Nereids on toilet-trays, Maenads (MMA access no. 2000.284.15) and to the goddess Selene (Mevissen 2011;Tanabe 1998; see also the male figure on sea-monster in AKD 64). To this group of figures, one should now add a male deity represented on a fragmentary frieze from Gumbat (GBK 34; [figs 21-22]) that we publish here with the aim of drawing to it the interest of more expert eyes. I hence limit myself to a plain description of the piece. This is a fragment of a small frieze (0.125 l., 0.10 h. [thickness not indicated in the inventory book]), broken on three sides and heavily corroded and defaced, found in the refilling of a recent pit dug to the east of stupa 3. From the data available we are not able to infer the provenance of the piece which however, in terms of dimensions and physical features, does not seem to be a stair-riser.
Despite its fragmentary state of preservation, a few observations can be made on the figurative group here illustrated. The central figure of the group is a standing male deity, as indicated by the halo, wearing a short tunic with trousers and holding with both hands a billowing drapery that covers his head. The figure looks towards another standing male figure, wearing horseman (?) trousers with a belt, almost completely defaced, that seems to offer something to the deity. At the back of the deity is a third male figure wearing trousers and shorts with a frontally hanging pointed hem. Although his upper part is completely lost, we can still note the joined hands of the figure, which has its right leg slightly bent with the heel raised upward.
The comparison for the central figure comes from the numismatic repertoire. In fact, the deity in Iranian dress holding a drapery over his head immediately brings to mind the iconography of the windgod represented on the reverse of Kushan coins bearing the Bactrian legend OAΔO. On these coins he is usually represented running left (in one case, right) and holding up a billowing drape to represent the wind. While we cannot say much about the figure on the left, re-Annali di Ca' Foscari. Serie orientale e-ISSN 2385-3042 57, 2021, 197-240 234 garding the figure on the right we cannot refrain from noting that both gestures and dress are reminiscent of the dancer from Butkara performing the 'Persian snap' (Lo Muzio 2019, fig. 4.3). The arms of the figure are unfortunately heavily corroded but the impression that the man is actually performing the 'Persian snap' rather than an añjalimudrā is solid, especially if we consider both dress and position of the legs.
If this interpretation is confirmed, we may have here a composition where the Iranian wind-god, celebrated by a dancer performing the 'Persian snap', is receiving something (an offer?) from a male figure (wearing the horseman trousers). A strong Iranian component then, which counterbalances the 'Western' imagery discussed above, and illustrates, better than anything else, the cultural complexity behind the Gandharan artistic phenomenon in Swat.