Identifying Several Visual Types in Gandhāran Buddha Images
How are we to define Gandhāran Buddha images in visual terms? Or how are they visually constituted? We can generally recognize and characterize them—from their distinctive materials (e.g., particular stone types of schist or phyllite); their generic differences from images created in neighboring areas of India proper; and their apparent affinities, in varying degrees, to images from the Classical traditions of the Mediterranean and eastward to the borders of Iran. Beyond these generalities, however, we remain astonishingly inarticulate, despite more than a century of serious investigations into the subject. This may be attributable partly to the extreme paucity of securely datable materials, which would allow us to lay the images out, as it were, in clusters of dots along a timeline. As is well known, out of thousands of extant images from Gandhāra, merely five carry inscribed dates. Even of those five, only one (year 89) is unequivocally datable in the Kaniṣka era (Fig. 35), whereas the Common Era equivalents of three others, dated in the years 318, 384, and 399, are still debated.1 Our problem also—and probably to a greater extent—stems from the dominant concerns that have formed major lines of research on visual problems in previous scholarship.
From the very outset, many scholars in the field were understandably preoccupied with identifying external sources—Hellenistic or Roman, and later, Iranian—for the Gandhāran art tradition, which may be deemed essentially derivative.2 Although this approach did clarify important aspects of this tradition, it had grave limitations: for in Gandhāran art, "influences" from the West came from multiple sources and at multiple moments, and these "influences" were continually mixed with preexisting or other heterogeneous elements, and thus most often transformed into highly complex forms. Attempts to identify specific external sources were thus limited mostly to a small number of isolated examples, and scholars tended to note visual features identifiable with extraneous traditions already familiar to them; those who were conversant with Roman art (e.g., Mortimer Wheeler, Benjamin Rowland, and Alexander Soper) first saw Roman elements in Gandhāran objects, and those who were originally Iranian experts (e.g., Daniel Schlumberger and Roman Ghirshman) were more inclined to find Iranian elements. Beyond presenting a rough picture, this approach largely failed to consider aspects potentially more significant, such as the internal grouping of images.
Another prevalent approach was to create a chronology of images. Although this is indeed a valuable rudimentary step in constructing knowledge of any material culture, it has been significantly hampered by the dearth of dependable evidence from either inscriptions or stratified excavations, and moreover, the chronology of history in the region itself is a notoriously controversial subject.3 Stratigraphic information has been accumulated from excavations conducted with advanced techniques since the second half of the twentieth century, but evidence that could be useful for understanding the chronology of Gandhāran art is still scarce.4 Naïve comparisons with parallels in the West were also not much help, because attempts to correlate the development of Gandhāran art with external artistic developments cannot provide a sound chronology. These conditions cast doubt on the validity of any detailed chronology of Gandhāran images yet attempted.5
A notable problem with the various chronological approaches was that they commonly treated the Gandhāran tradition—or, more correctly, "traditions"—as a single chronological continuum subject to a unilinear developmental scheme.6 In the Greater Gandhāran region, however, a number of regional units—such as the Peshawar valley, Swāt, Taxila, Jālālabad, and Kāpiśī—flourished side by side, each producing visually distinctive images; again, within each unit diverse smaller units are discernible—smaller units which reflect individual subregions or workshops, sectarian affiliations, economic capacity, or the taste of local patrons. In applying a unilinear developmental scheme, such multifarious factors, whose acknowledgement would have helped reconstruct internal relationships among Gandhāran images, were often ignored or even suppressed.7
This paper attempts to clarify an aspect of the internal configuration of Gandhāran art, with a focus on Buddha images. More specifically, it aims to identify [End Page 43] several visual types8 among Buddha images in stone from the Peshawar valley. Regionally, the Peshawar valley, "Gandhāra proper" in a more restricted definition, was the central area in the Greater Gandhāra, and it was in this area that the majority of larger-scale stone statues, which are commonly equated with Gandhāran art, were produced. Among the statues, those that portray the Buddha were indisputably the most thematically significant as well as the most numerous, since the visual tradition of Gandhāra essentially served Buddhism.9 Although stucco was widely used, it was stone that was most favored for image making in the Peshawar valley. Ever since I began studying Gandhāran art, I have found it intriguing that certain visual types stand out conspicuously among Buddha images. Remarkable in quality and dominant in quantity, these visual types were no doubt major constituents of the genre we call Gandhāran Buddhas. I believe that their noteworthiness did not entirely elude the attention of previous researchers, who occasionally remarked in passing on some of those visual types.10 But they have never been treated in a comprehensive, systematic manner. Five of these types, the most outstanding ones that I have noted, will be discussed here.11
I present each of these five types as a cluster of images comprising a small number of especially characteristic primary objects and a large number of related objects. The primary objects in each group were most probably made in the same workshop or in a handful of closely associated workshops and within a limited time. Most of the related objects are variations of primary objects—imitations, adaptations, or derivations—produced over a more extended period, or parallel off-shoots that originated from a common source. Rare instances [End Page 44] could have been prototypes of the primary objects. My conception of the visual type was influenced by ideas that George Kubler put forth in many of his writings, especially in The Shape of Time (1962). The binary terms I use here, "primary" and "related" objects, are modelled on a pair of Kubler's key concepts, "prime object" and "replication," but are modified considerably to suit my purpose.12 I find it more practical to define a prime object not as a thing in which an innovation first took place, but as one in which it became manifest and established.13 The term "primary object" in this paper is used in this modified sense. I have also avoided the use of Kubler's term "replication," because the temporal precedence of a thing over another cannot always be ascertained either in reference to absolute time scale or to relative formal sequence. The term "related object" is used instead.
The terms "primary" and "related" as I apply them to the five types are stated as a conceptual outlook underlying this undertaking, and thus specimens in each type will not be rigidly labelled as either "primary" or "related" objects. The selection of primary objects for each type and the consequent clustering of each of the five types are essentially based on my own cumulative visual experience with extant Gandhāran Buddhas. While visually registering affinities and differences, on the basis of intuition and visual analysis I realized that hair and facial types were the most crucial elements. Encouragingly, typological distinctions based on these features coincide with what we observe in the execution of drapery. I concede that the formulation of a type or types can vary depending on the identification of primary objects and [End Page 45] the definition of the relationship between various groups of primary and related objects. Thus, my five types are intended as flexible and loosely demarcated categories, which can well be modified or refined. And that is why I designate my undertaking as "identification" rather than the more familiar term "classification."
A visual type may be viewed as a small formal series within the larger, imaginary series of Gandhāran art. Primary and related objects within a series are not simply static dots distributed on a two-dimensional plane but active and dynamic participants in complex relationships with one another. The five series or types, however, are presented here initially as synchronic units, without consideration of their chronological relationship. I suspect that the principal element enabling me to discern each type was distinct visual consistency maintained by closely related workshops or formed within the smaller regional units that constitute the Peshawar valley. Differences in the origin and duration of various types will be noted where plausible and relevant. The five types are designated I to V, but these numerals do not connote either chronological ordering or qualitative hierarchy.
A primary example of Type I is a standing statue in Peshawar Museum (Figs. 1, 2).14 This image is made of light gray schist and stands 1.7 meters high, the most common size among Gandhāran Buddhas. It is characterized by a plump, round face and by distinctive features such as wide-open eyes with clearly marked irises and pupils (shaped like narrow crescents), and by a conspicuous mustache and a broad uṣṇīṣa tied with a string. Interestingly enough, the top of the uṣṇīṣa is flat and was left uncarved except for a circular groove, which seems to have been designed to accommodate an insertion. I have suggested elsewhere that the present uṣṇīṣa must have held an upper part currently missing, making the original uṣṇīṣa much higher than it is now.15 Above the string, the hair is in S-shaped waves, and below the string, framing the face, the hair is arranged in a series [End Page 46] of upward-flowing, symmetrical arcs. The rendering of the body and drapery is also distinctive. A relatively thin garment reveals the slender body and the contours of the shoulders and the chest. Even the nipples are visible beneath the garment, an extraordinary feature among Gandhāran Buddhas. Also, unlike the majority of Gandhāran Buddhas, the drapery is not upturned or folded thickly around the neck. The drapery folds covering the body consist of the common high and low ridges, but the high ridges are placed at much wider intervals than is typical of Gandhāran Buddhas. Since the thin garment reveals the contours of the body beneath, this Buddha appears more naturalistic than other, more common, types (Types II and III in this paper). Although both hands are lost, the remaining parts indicate that the right hand was originally in abhaya mudrā, and the left hand was at chest level, holding the garment.16 The left hand held up in this manner is unusual among more common Buddha statues of Gandhāra (Types II and III again), in which the left arm is usually held straight down with the hand holding the garment at what might be called skirt level. But the position of the left hand in Figure 1 is frequent among the finds from Swāt as well as in Buddhas from Mathurā that appeared beginning in the middle of the first century of the Kaniṣka era.17
A statue in the Lahore Museum is similar to that in Figure 1 (Figs. 3, 4).18 Although badly damaged, this statue is almost identical to the Peshawar Buddha in the shape of the head and the facial features.19 Even the top of the uṣṇīṣa is carved with a similar circular groove.20 The overall shape of the body, the configuration and folds of the drapery, and the noticeable nipples are also reminiscent of the Peshawar Buddha. Yet the drapery is upturned around the neck, and what remains of the broken left arm suggests that it originally hung straight down. Despite these differences, which do not seem to [End Page 47] have chronological significance, the two images appear extremely close in origin.
A statue in the British Museum is another example of Type I (Figs. 5, 6). It is much smaller than the aforementioned statues, standing only 92 centimeters high, the feet having been broken off. The shape of the head and the facial features are almost the same; the irises and pupils are delineated in incised lines. The left hand hangs straight down at the side as in the more common Gandhāran Buddhas (Types II and III in this paper). Under its clinging drapery the slender body seems even more naturalistic than the bodies of the preceding two Buddhas (Figs. 1-4); otherwise, the three images share most characteristics.21 The top of the uṣṇīṣa is similarly flat and uncarved but lacks a circular groove.22 Perhaps an upper part had not yet been installed; or the upper part was attached in a different manner. According to the museum record, this Buddha originated at Takht-i-Bāhī, a famous monastery site in the Peshawar valley.23
A statue in the Museum für Asiatische Kunst in Berlin shows affinities with the preceding three Buddhas in the wide-open eyes with incised irises and pupils, the mustache, and the raised position of the left hand (this last feature resembles the Buddha of Figs. 1, 2) (Figs. 7, 8). There are differences as well. The face is more oval, with slightly protruding cheekbones and slender jaws.24 The hair is arranged in thick strands, alike in symmetry to that of Figures 1-6. Regarding the drapery, the high-ridge folds are less three-dimensional, and the folds between them are represented by lines incised in a somewhat disorderly manner. Particularly notable is the uṣṇīṣa, arranged in two distinct parts, with its upper part having a hole in its center; this suggests how the uṣṇīṣas of the Buddhas in Peshawar and Lahore (Figs. 1-4) would have looked originally.25 Despite somewhat different features, this image seems related to the preceding three Buddhas and could be loosely identified as Type I, although its exact relationship with these other three Buddhas remains a question for future research. The Berlin Buddha is known to have originated in Swāt, but its exact provenance is unknown.26 Quite a few images with similar features have come from the Swāt or [End Page 48] Mālākand areas.27 A seated Buddha from Andan-Ḍherī near Chaṭpaṭt (h. 54 cm) also seems to be related, although the strands of its hair, instead of arcing upward toward the uṣṇīṣa, curl downward on either side of the center (Fig. 9).28 Seated images loosely connectable to Type I show only the abhaya mudrā, never any other hand gestures.
Buddhas similar to Type I images also appear in narrative reliefs, such as a relief representing a standing Buddha and Vajrapāṇi, supposedly from Swāt (Fig. 10). Here the uṣṇīṣa is disproportionately high and is divided in two horizontal tiers, although it lacks a hole as is usual in such narrative reliefs. Buddhas of similar forms are also seen in the reliefs depicting the Offering of Grass and the Visit to a Teacher from the so-called Mardan group, which is also known to have been recovered from the Swāt area, during the Mālākand campaign.29
In the group Type I, the statues in Peshawar and Lahore are almost identical, and quite probably were carved in the same workshop, if not by the same sculptor; 30 the statue in the British Museum is closely related to these two.31 The remaining examples seem visually rather loosely associated, possibly because they originated in different workshops, regions, or times. The statues in Peshawar and Lahore, together with the reliefs of the Mardan Group, have often been regarded as exemplifying the early phase of Gandhāran art. John Marshall dated them to the period he called the Adolescence (dated ca. 60-100 CE in his chronology) and classified them among the earliest of standing Buddha statues in Gandhāra.32 In type of stone, size, and format, however, the Peshawar and Lahore Buddhas (and the British Museum Buddha) are not much different from specimens of Types II and III, which constitute the majority of Gandhāran Buddhas, and one wonders whether the differences between Type I and the other, [End Page 49] more common, types do reflect a difference in chronology. Even if Type I actually predated the other types, the time difference may not have been great.
A number of objects associable to this group were from Swāt or Mālākand. This tempts one to attribute the origin or prevalence of Type I to the Swāt area. But the statue in the British Museum is known to have come from Takht-i-Bāhī in the Peshawar valley; the two statues in Peshawar and Lahore, whose provenances are unknown, are not essentially different in size or stone from statues that originated in the Peshawar valley. The current stage of our knowledge does not permit us to decide whether a visual type created in the Swāt area spread to neighboring areas, including the Peshawar valley to the south, or the sculptors in Swāt adopted a visual type transmitted from the south.
For the second type the best example is one of the largest extant Gandhāran Buddhas, now in the Peshawar Museum (Figs. 11, 12). Measuring 2.6 meters, this statue, of dark brownish schist, comes from the Sahrī-Bahlol Mound B. Its form and presence, even more than its comparatively large size, make it a truly monumental work. A string, such as discussed earlier, ties off the high uṣṇīṣa. Its coiffure is markedly similar to that of a Peshawar Museum Buddha of Type I (Fig. 2), except over the exact center of the forehead, where the coiffure forms a distinctive pattern that has been characterized as "almond shaped."33 The flat-topped uṣṇīṣa with a circular groove that we saw in some Buddhas of Type I is absent from the Buddha in Figure 11 and from [End Page 50] all other Buddhas of Type II. Also, no hole is carved in any examples of Type II that I have examined.34 As shown in Figure 11, the image's face is a slightly rounded oval, and the rendering and composition of its features are extremely refined, with the exception of a few features that are angled, such as the ridge of the nose and the demarcation between the eyebrows and eye sockets. The statue's eyes are not fully open like those of Type I; barely visible incised lines delineate the irises. The mustache is clearly discernible, and far more elegantly shaped than those of Type I. Yet in other respects the features of the lower face do not differ significantly from those of the Type I Buddha at Peshawar Museum (Figs. 1, 2).35 The left hand hangs straight down and holds the garment as usual. The garment is thicker than those of Type I images, and the drapery folds are executed in alternating high- and low-ridge lines common in Buddhas of Gandhāra. Though the carving is highly skillful, it is somewhat schematic, and the thickness of the garment creates an effect less naturalistic36 than the garments of Type I images.
Another Buddha from the same site and virtually the same spot is identical in size and format (Figs. 13, 14),37 with minor differences in the face as well as in the execution of the body and drapery. Especially, the lower part of the face is slightly narrower, the mustache a bit less crisply articulated, the eyes are not fully open (although with clearly delineated irises), and the distance between the eyebrows and upper eyelids is greater. This Buddha was probably carved in the same workshop but definitely by a different hand. A gigantic bodhisattva head (52 cm high) from the adjacent Mound C at Sahrī-Bahlol exhibits a face similar to this Buddha's, and is thus attributable to the same sculptor (Fig. 15).38 Evidently, sculptors affiliated with a workshop in the Sahrī-Bahlol area were quite active in producing large statues for nearby monasteries.
Extant images of Type II are far more numerous than those of Type I. One example is a Buddha in the [End Page 51] Lahore Museum (Fig. 16). Although only its upper torso and head remain, it was originally a standing statue, slightly smaller than the two gigantic Sahrī-Bahlol Buddhas. It shares most features with them, particularly with the first (Figs. 11, 12), except that the eyes are not wide open and the mustache is less crisply articulated.39 Another standing Buddha in the Peshawar Museum (Figs. 17, 18) is smaller (1.57 m) but remarkably close in face and drapery to the second, much larger Buddha from Sahrī-Bahlol Mound B (Figs. 13, 14). The statues in Figures 16 (Lahore) and 17 (Peshawar), although of unknown provenance, are likely to have come either from the same workshop or from a handful of workshops all working in an identical style.
Two standing statues, one in the Taxila Museum (Fig. 19) and the other in the British Museum (Fig. 20), fit broadly within the same visual series.40 They are much smaller than the Type II images discussed above (95 and 104.2 cm high, respectively) and are made of gray schist, a common material for images of this size. In form and proportion they are less imposing than the preceding examples. The facial features of the Buddha in Figure 20 are similar to those of the second very large Buddha from Sahrī-Bahlol (Fig. 14). But the face is square rather than oval, and the execution seems more delicate, perhaps owing to the size and the material;41 [End Page 52] the mustache is not explicitly delineated but suggested by a slight excrescence above the upper lip. The two Buddhas may have originated at different workshops or at slightly different times. Many Buddha heads, missing their bodies, are of Type II.42
Seated statues of Type II almost invariably hold their hands in dharmacakra mudrā and have the right shoulder bare. Two seated Buddhas, one in the National Museum in Karachi (Fig. 21) and the other in the Peshawar Museum, are the closest examples to the two great Buddhas from the Sahrī-Bahlol mounds (Figs. 11-14).43 In both images the coiffure is somewhat wavier above the hair string than in the Sahrī-Bahlol Buddhas, more closely resembling that of the statue in Figure 16. Another seated Buddha in the Peshawar Museum (from Sahrī-Bahlol Mound B) holds its hands in dhyāna mudrā and has both shoulders covered (Fig. 22). But the coiffure is somewhat different in detail, and so is the face; thus, this statue may be regarded as a distant variant. No known seated statue of Type II holds its hands in abhaya mudrā.
Seated Buddhas of Type II appear quite frequently in the triad format. One triad from Mound A and another from Mound D at Sahrī-Bahlol are well-known examples.44 The one from Mound D (Fig. 23) exhibits close affinities with the independent seated statues (e.g., Fig. 21). In the other (from Mound A) the mustache is barely visible, the face narrow, and the body markedly slender.45 Another important triad with seated Buddha (Fig. 24) has received considerable attention for its inscribed date, "year 5," which is usually equated with the 5th year of the Kaniṣka era or sometimes with the 105th year of the Kaniṣka era, the "hundred," presumably, having been omitted.46 The first date would coincide with the common assignment of stone images of the most refined craftsmanship to the Kaniṣka reign, and it could provide a clue to the chronological position of the gigantic Buddhas of the same type from Sahrī-Bahlol. [End Page 53]
[End Page 54]
[End Page 55]
But it is particularly difficult to precisely date objects such as the triad, since it is questionable whether the triad format was already established by the fifth year of Kaniṣka. This piece, which was first reported in the early 1970s and has not been exhibited since 1986, requires more systematic study.47
Another Buddha with this distinctive facial type appears in representations of the Buddha displaying a snake bowl to Uruvilvā Kāśyapa (Fig. 25). Interestingly, three out of four extant examples I have recorded were unearthed at the same site, Mound B of Sahrī-Bahlol, and the fourth was unearthed at the nearby village of Koṭ.48 Since some of these statues show slight differences in the execution of face and body, it does not seem that they were all made in the same workshop; but Sahrī-Bahlol emerges again as important source of this type.49
The extent of visual variation in Buddha images of Type II is limited. The number of variants from what I call prime objects is small, and the degree of variation within the variants is limited. This seems to indicate that most of these statues were made in the same geographical area and within a rather short time span. Note that the majority of such images were excavated at the sites in Sahrī-Bahlol, and some at Takht-i-Bāhī.50 Sahrī-Bahlol and Takht-i-Bāhī are located merely one kilometer apart in the northwestern part of the Peshawar valley, and Type II was most probably created, or at least established, in a workshop in this locality and was supplied in large numbers for dedications at nearby monasteries.51 Considering the fairly limited stylistic variations among Type II images, I would presume they came from one major and a few minor workshops. [End Page 56]
Selection of a single representative example, as was done for Types I and II, is difficult for Type III. Although numerous images that I have identified in as Type III display apparent typological homogeneity, they seem to be "related" objects—"replications" in George Kubler's terminology—and there are few that might be considered "primary" objects. Furthermore, among those possible primary objects, none is preserved in a complete state. Therefore we start our characterization of this type with several Buddha heads.
In the Peshawar Museum is a head about 45 centimeters high (Fig. 26). Because all statues of this size made in Gandhāra were, without exception, standing, this head must have been part of a standing statue whose original height must have been well over 2.5 meters.52 The head is made of dark brownish schist, which was commonly used for statues of Buddhas and bodhisattvas in the Sahrī-Bahlol and Takht-i-Bāhī area, including the two heroic-sized Buddhas of Type II from Sahrī-Bahlol Mound B (Figs. 11-14). This head, excavated at Mound D of Sahrī-Bahlol, bears a high uṣṇīṣa, which is not tied off with a string. Since among statues of Type III the string is equally present or absent, it seems not to have been deemed significant. The hair is arranged in a wavy pattern, broadly undulating over the entirety of the head, unlike the hair of Types I and II. In Figures 26 and 27, the latter discussed later, low, flat ridges form the waves of hair; in other images in [End Page 57] this formal series the hair is further reduced to incised lines. As in Type II, the peculiar uṣṇīṣa design of Type I is never found in Type III, and a hole at the top of the uṣṇīṣ is extremely rare in images whose wavy hair was executed in the manner described. In statues of significantly large size, a hole in the uṣṇīṣa never appears. Prominent in the oval face are a heavy jaw and long nose. The half-closed eyes convey both meditation and aloofness. Such half-closed eyes appear in some examples of Type II, but in Type III images they are an invariable feature.53 The mustache is not manifestly carved, but probably was delineated by paint on the gentle swelling above the mouth.54 The rendering of the hair and facial features appears somewhat schematic, and the carving is not as refined as in the representative examples of Type II. Another gigantic head in the Peshawar Museum, which also measures 45 centimeters high and is made of the same dark brownish schist, is stylistically similar except for the string tying off the uṣṇīṣa and the somewhat different face with broader jaws (Fig. 27). It was almost certainly made in the same workshop. Although its provenance is unknown, it most probably came out of one of the mounds at Sahrī-Bahlol or Takht-i-Bāhī.
Two smaller heads in the Peshawar Museum (from Mounds A and B at Sahrī-Bahlol), 36 and 34 centimeters high (Fig. 28), are quite visually close to the large heads mentioned earlier.55 Their facial features are more delicately executed than those of the large heads. It is hard to judge whether Figures 26 and 27 were enlarged replicas of these small heads or if the small heads in Figure 28 appear more delicate simply because they are smaller. In any case, all four heads are most probably from the same workshop. A large number of Buddha heads with similar features are extant, and many of [End Page 58] these are also attributable to the same workshop or to closely related ones near Sahrī-Bahlol and Takht-i-Bāhī.56 Even with only the heads extant, it is evident that these four images form a distinctive type. They exhibit significant differences from the heroic-sized statues of Type II from Sahrī-Bahlol Mound B, although one wonders whether these variations can be attributed to different production dates.
For the body and drapery associated with the Buddha heads discussed above, we have to consult other examples. A standing Buddha in the Peshawar Museum (from Sahrī-Bahlol Mound A;57Fig. 29), whose upper half only is preserved, has a face identical to those of the two smaller Buddha heads in Figure 28. This standing Buddha's head is also about the same size, and the total height of the complete image (including the halo) would have been well over 2 meters. Although we only have the upper half of the statue to judge from, it seems similar, in body and drapery, to the two heroicsized Buddhas of Type II. The variations are limited: the statue's drapery is somewhat less three-dimensional than those of the two great Buddhas, in particular the high-ridge folds are not as high. Similar drapery is visible in a better-preserved, smaller statue of the same type (from Takht-i-Bāhī) that represents the Buddha receiving dust from a previous incarnation of Aśoka (Fig. 30).58 In a standing Buddha in the Peshawar Museum (from Mound A at Sahrī-Bahlol; Figs. 31, 32), whose face is quite similar to one of the large heads of Type III (Fig. 27), the drapery is more schematic, and the proportions and posture are somewhat awkward.59 An excellent carving in the Lahore Museum (Fig. 33), although badly damaged, is more three-dimensional and naturalistic, thus closely resembling the heroicsized statues of Type II. This presents two possible scenarios: [End Page 59]
[End Page 60]
possibly the production periods of Types II and III over-lapped, if only briefly, or possibly the stylistic rendering of the body and drapery changed at a slower pace than that of the head or face. Figure 33 was found at Jamālgaṛhī, another important site on the northern side of the Kabul River in the Peshawar valley, not far from the Takht-i-Bāhī and Sahrī-Bahlol area.60
The majority of seated images of Type III hold their hands in dhyāna mudrā and have both shoulders covered (Fig. 34);61 a smaller number display dharmacakra mudrā and have the right shoulder bare.62 One of the rare images with inscribed dates from Gandhāra, the Buddha in the Indraśailaguhā of year 89, is of this type (Fig. 35). Most scholars find no problem in ascribing it to the eighty-ninth year of the Kaniṣka era. If we date the Kaniṣka era from about 120 or 127 CE, the image is attributable to the early third century (208 or 215 CE).63 This image came from Mamane-&1E0Cherī near Chārsada, again not far from the Takht-i-Bāhī and Sahrī-Bahlol area.64
Type III comprises by far the largest number of Buddha images of Gandhāra. Unlike Type II, Type III consists of numerous variations, or related objects, of diverse quality, and thus Type III shows a greater range of stylistic variation. For instance, in a statue in the Peshawar Museum (from Takht-i-Bāhī), the uṣṇīṣa is higher and wider, and both the face and body are rather corpulent (Fig. 36).65 The drapery' high-ridge folds are shallow, and the lower part of the body and drapery are particularly flat. A number of modified sub-types—both visual and iconographic—seem to have derived from Type III. A gigantic statue in the Lahore Museum (from Takht-i-Bāhī) would look not much different from the primary pieces of Type III, were it not for its snail-shell curls in place of the wavy hair common on Gandhāran images (Fig. 37)—even though the corpulence [End Page 61]
[End Page 62]
and the flatness of this image are more marked than in the preceding image (Fig. 36).66 Snail-shell curls, which probably appeared in Gandhāran images under the influence of Mathurā, are quite rare among Gandhāran Buddhas, but the majority of the images on which they do appear correspond loosely to Type III in all other respects.67 Although these images might be considered a separate category, I have the impression that the snail-shell coiffure was an optional variation, adopted in limited instances in this visual category.
Overall, Type III encompasses a large number of images of quite inferior quality. One has the impression that such images were produced in enormous quantity at various levels of craftsmanship and that a larger number of workshops—far larger than for Types I and II and at various levels of craftsmanship—were active across a vast area. The leading workshops that produced the primary and highest-quality Type III pieces seem to have been concentrated in the northern part of the Peshawar valley around Sahrī-Bahlol, Takht-i-Bāhī, and perhaps Jamālgaṛī. The duration of Type III as a formal series was probably much longer than that of Type II.
Type IV is exemplified by a seated Buddha in the Indian Museum, Kolkata (Calcutta) (Fig. 38). Its hands are in dharmacakra mudrā, and the right shoulder is [End Page 63] bare, the format common for Type II. The hair, however, is arranged in the wavy pattern characteristic of Type III, but carved in shallow incised lines on a considerably flattened surface. The head is big and square and the body significantly rounded, shorter than in Types I to III, and fleshy, with little indication of muscles or bone structure. No mustache is visible. Although the drapery consists of alternating high- and low-ridge folds as usual, the high ridges are much less prominent than those of Type III; they generally look like strings schematically arranged. A cylindrical knob, protruding from the top of the halo, probably supported the chattra (parasol). This knob occurs in no other types of Gandhāran Buddha images; rather it reminds one of the chattra on a long staff held behind some kapardin type statues from Mathurā, such as one dedicated by the monk Bāla at Sārnāth in the third year of Kaniṣka (ca. 122 or 129 CE), which seems much earlier in date than Figure 38. Beneath the crossed legs of the Kolkata Buddha (Fig. 38) is a round, tapering knob, which was originally inserted into a lotus seat, another feature peculiar to Type IV. Several similar examples are extant in Kolkata, Lucknow, and Karachi (Figs. 39, 40).68 All the images in Kolkata were excavated at Loriyān-Tangai.
Many variants of Type IV exist. A piece from Takht-i-Bāhī in the Peshawar Museum is related to this group in pose and overall body form (Fig. 41).69 This piece, however, is markedly flat, particularly in the lower part of the body, where the organization of the drapery also differs slightly from that of Figures 38 and 39. Furthermore, it has a snail-shell coiffure, and is seated on a cushion supported by a square pedestal.70 A piece from Jamālgaṛhī in the British Museum wears the robe covering both shoulders, but otherwise it is the same as Figures 38-40, including its seat on a cushioned pedastal.71 Several images, such as one from Loriyān-Tangai (Fig. 42), hold their hands in dhyāna mudrā and have both shoulders covered.72 Type IV includes images frequently employed in the triad format, such as the two triads in Kolkata, both from Loriyān-Tangai (Fig. 43).73 There are a number of similar triads in other collections.74
Most extant images of Type IV are in the seated pose, but a few are standing images. A small standing image in the Lahore Museum, only 63 centimeters high, is a rare example associable with Type IV (Fig. 44). The overall qualities are similar, particularly the plump body and the string-like drapery folds. But the square head is somewhat elongated, and the mustache is carved. The hair seems a variant of Type II rather than Type III. The robe is worn in a pseudo-himation style, a peculiar feature seldom seen elsewhere among independent statues. This statue's provenance is sometimes given as Shāhjī-kī-Ḍherī in the vicinity of Peshawar, but this information is under question.75
Many extant images are loosely related to this group. The central Buddha in the famous Mohamed-Nari stele in Lahore is an example (Fig. 45). Not only is this image's iconographic format identical to those of Figures 38, 39, 41, and 43, but also the configuration of the body is similar. But the hairstyle was derived from Type II, as indicated by the characteristic almond-shaped pattern at the center of forehead.76
Although I try hard to avoid unwarranted diachronic presumptions, Type IV seems, in every aspect, to be later than the primary pieces of Types II and III. Most examples appear to be derivations from Type III, although the hairstyle is less consistent. Several pieces could be deemed central in this group, and among the numerous variants, the degree of divergence is small. The majority came from Loriyān-Tangai, which emerges as an important site for this type. This site is known to [End Page 64]
[End Page 65]
[End Page 66]
[End Page 67]
[End Page 68]
have been located near the Shāhkoṭ Pass, beyond the Mālākand Range in the north of the Peshawar valley, thus in a different area from Sahrī-Bahlol or Takht-i-Bāhī. In 1896 J. E. Caddy excavated numerous Buddhas and bodhisattvas at this site,77 and in an old photograph of the site, taken by the Archaeological Survey of India, we can see several Buddhas of Type IV as well as Type III.78 Although images of this type are not entirely absent at Sahrī-Bahlol or Takht-i-Bāhī, we should note that the absolute majority originated at Loriyān-Tangai.79 It seems reasonable to suppose that Type IV was established in a workshop in the nearby area, perhaps at a date somewhat later than the major activities around Sahrī-Bahlol and Takht-i-Bāhī.
A primary example of Type V is identifiable in a standing statue in Lahore Museum (Figs. 46, 47). It has a peculiar hairstyle: arranged on the scalp in undulating waves as is Type III but less angular, and over the uṣṇīṣa in vertical rows of small curls at first glance resembling snail-shell curls but clearly distinguishable from them. In the top center of the uṣṇīṣa is a hole, which I presume once contained a relic. In fact, the majority of Buddha images with a hole in the uṣṇīṣa wear the peculiar coiffure just mentioned.80 Compared with the Buddhas of Types II and III, this Buddha's forehead is narrower, its chin shorter, its jaws rather narrow, and its mouth small. Each side of the upper lip is marked by a gentle protrusion, but it is unclear whether a mustache was originally painted there. Overall, the facial features give a more youthful impression than do the features of the other types. The head seems slightly larger in proportion to the body, significantly reducing its majestic presence compared with that of the gigantic statues of Types II [End Page 69] and III. The drapery folds are evenly distributed over the body at regular and narrow intervals, the high-ridge folds are less pronounced, and the low-ridge folds have almost disappeared. In most examples of Types II and III, two or three folds flow down in steep diagonal lines from the left shoulder but end at the abdomen, thus highlighting its protrusion. By contrast, in Type V such folds are barely visible or are reduced to a single insignificant-looking line, which enhances the plainness of the body.81 A scene of the veneration of a reliquary, an extremely rare feature among Gandhāran Buddhas, is carved on the pedestal.82
Another standing Buddha in the Lahore Museum shows similar features in most details, even to the rare veneration of a reliquary carved on the pedestal (Fig. 48). The face looks somewhat broader, but this impression may be due to its badly eroded state. Other notable differences are the absence of upturned drapery around the neck and the drapery folds of a second garment visible on the lower right thigh. This second garment is either a nivāsana, an undergarment worn on the lower part of the body, or an uttarāsaṅgha, a less formal robe worn in place of, or under, a saṃghāṭī (formal overgarment). A Buddha from Sikri with an almost identically formed face and body, but made of gray schist, is in the Chandigarh Museum (Fig. 49), evidently having been transferred from Lahore at the time of the Partition.83 Although the provenances of Figures 46 and 48 are unknown, these two Buddhas and the Sikri Buddha most likely came from the same workshop.84
Two standing images in the Peshawar Museum are [End Page 70] also notable for affinities to the preceding examples.85 One is particularly close (Figs. 50, 51), the other slightly different in the face and in the drapery, especially in the necklines. Both are known to have come from the same site, near Daulat in Mardan. A standing image in the National Museum in New Delhi seems to be a somewhat degenerated example of the kind, with conspicuously short proportions, and close parallels are in the Lucknow State Museum and the Musée Guimet.86 Two other standing Buddhas in New Delhi can be grouped in the same category (Fig. 52).87 But their faces are distinctly different from those in Figures 48-51. Many other images are loosely associable to Type V.88
Examples of Type V mostly appear in standing pose; seated ones are extremely rare. A seated Buddha in the Indian Museum, Kolkata (Fig. 53) must be the closest parallel to the standing images we have seen.89 It is, however, patently different in dress and pose among Gandhāran seated Buddhas. The robe is worn with the right shoulder bare, but an undergarment, which is invariably revealed in other seated Buddhas with the right shoulder bare, is not depicted. Furthermore, the left hand, by holding an end of the robe, indicates that the broken right hand could only have been in abhaya mudrā. Another Buddha in Kolkata (Fig. 54) shows a similar head and facial type, although its robe covers both shoulders. A seated Buddha in New Delhi is another related example, but the head is somewhat square.90 [End Page 71]
[End Page 72]
The examples discussed so far exhibit marked visual consistency, and the majority of them display so strong a homogeneity in visual style that they can be securely attributed to the same workshop. We can visualize a workshop prominent in producing images of this type, perhaps along with a few more associated workshops.
Other images related to those of Type V require special remarks, even though I am reluctant to allocate them to this type. A standing Buddha in the Peshawar Museum (Figs. 55, 56) is much bigger than the first Buddha discussed in Type V (Figs. 46, 47); the proportions are also somewhat more elongated. Its stern face conveys a look of majesty. The hairstyle is similar to that of the preceding images of Type V, but slightly more three-dimensional, and the uṣṇīṣa is much higher. The face is oval, thus markedly different from the face in Figure 46. Although the distinction between high- and low-ridge folds is not as pronounced as it is in Figure 46, the overall body and the drapery seem visually closer to Type III. Thus, this image may stand between Type III and the ordinary examples of Type V. Its relationship to Type V presents an intriguing question: Does this image prefigure a trend within the larger formal series of Type III that later developed into a series I have defined as Type V? The conventional art history perspective may presume that abstraction or schematization is subsequent to naturalism. From this perspective, one may be tempted to assume that the uniquely abstract hairstyle of the Type V images was derived from the similar, yet more naturalistic, hairstyles of Type III. Rather than reaching a conclusion in haste, however, I prefer to leave this question open, as a useful starting point for future research. In any case, examples that show close affinities to the Peshawar Buddha are few, and even among them visual consistency is not so marked. Another Buddha in Peshawar has similar coiffure and drapery, but its face and its bodily proportions are considerably different (Fig. 57).91
Another image in Peshawar deserves to be noted in regard to the connection between Type III and Type V images (Figs. 58, 59). The significance of this image, once lauded by Alfred Foucher as "the most beautiful, and probably also the most ancient of the Buddhas, which it has ever been granted to me to encounter," was bluntly dismissed by John Marshall, who remarked, "It cannot be pretended that there is anything particularly spiritual in its expression, nor, as regards the rest of the statue, does it show any superiority over many other images of this period, either in the general proportions of the figure or in the technical skill displayed in the draping and finishing of the monk's robe. Indeed, the modeling of the ankles and feet is more than usually [End Page 73] clumsy. . . ."92 Nonetheless, this Buddha occupies a unique stylistic position among the Buddha images of Gandhāra. Examined carefully, however, the drapery seems essentially similar to that of Buddhas of Type V, except for double lines flowing from the left shoulder and ending at the stomach. Also, the face is markedly different from that of Type V. The hairstyle differs only slightly: the wavy hair is rendered more three-dimensionally and arranged more horizontally, and the curls over the uṣṇīṣa call to mind those on Type V images.93 Here it is intriguing to speculate about the relationship between this Buddha, examples of Type V, and the two Buddhas shown in Figures 55-57. One hypothesis could be that the images in Figures 58-59 were sources for the images shown in Figures 55-57 or for those of Type V. Although this hypothesis would require more corroboration, it seems certain that the hairstyles of Figures 58-59 produced derivations in which small curls were arranged in vertical rows covering the entire head. These derivations formed another distinctive formal series, although not a prominent one.94
The five types we have seen in this paper are mostly products of the Peshawar valley. Although the production of Type IV was probably located in the neighborhood of Loriyān-Tangai at the border of Swāt, images from this site clearly show close connection with the traditions [End Page 74] that flourished in the northwestern part of the Peshawar valley, centered around the Sahrī-Bahlol area. Type I may have had some association with the Swāt valley or its southwestern end in Mālākand, but its primary examples were presumably made in the Peshawar valley. Thus, the five types addressed in this article represent dominant formal series in stone statues of the Buddha produced in the Peshawar valley—and thereby in Gandhāra, as the Peshawar valley was undoubtedly the major center of creation of independent stone statues in Gandhāran art. Obviously, there were many other, smaller formal series in the Peshawar valley, but none comparable with these five in either size or consistency. Most other formal series may well be regarded as variants of Types I-V at divers levels or as marginal series.95 It is unsurprising that the majority of stone statues of the Buddha from Gandhāra belonged to, or were affiliated with, one of these types. Even clever forgers, who unfortunately have become quite knowledgeable in this regard, have been attempting to reproduce one or another of these types, creating increasingly better modern—but undeclared—replications in recent years.96
Of the five types addressed, Type III stands out most prominently in number of extant images and in the extent of variation. Images of Type III appear to have had the widest regional distribution, greatest number of workshops, and longest duration of production. Therefore, Type III can potentially be further divided into subtypes. Type II seems the most visually consistent, with the fewest workshops and a short production period. Type I seems to have been rather loosely organized save for a few primary pieces, but this may be an impression owed to the small number of extant specimens. Types I and II seem related, but it is difficult to tell definitively, at this stage, whether the latter derived from the former or both types were variations of the same prototype. Types II and III also share features. Images from Types IV and V seem fairly consistent and were probably later than the other types. I suggest Types IV and V may be derivations of Types II and III.
Together, the above five types constitute the Buddha images most commonly encountered in Gandhāra; of the five, Types II and III represent the most productive period in the history of Gandhāran Buddhist stone sculpture. Mainly, these statues were centered in the area northwest of the Peshawar valley, comprising such important sites as Sahrī-Bahlol, Takht-i-Bāhī, Jamālgaṛhī, Sikri, and Thāreli. It is no coincidence that multiple aligned chapels surrounding a stupa court were a distinctive architectural feature of monasteries in that area.97 [End Page 75]
The duration of each type and its implication for the overall duration of Gandhāran art, are fascinating lines of research. Scholars' estimates of the production period of Gandhāran art, specifically in stone sculpture, vary considerably: some consider the period to be about 200 years long, some even shorter. For instance, John Marshall dated the Adolescence of Gandhāran art to about four or five decades from 60 CE, the early Maturity of Gandhāran art to about a few decades until 140 CE, and the later Maturity of Gandhāran art to three generations (up to 215-230).98 Harald Ingholt assigns the period of independent statues a span of more than 200 years, from the latter part of the third century through the fifth century.99 Considering the fair degree of homogeneity among Gandhāran stone images, images of all five types may have been produced in a shorter time span than previously assumed.
The temptation to align these five types diachronically is strong. At present, however, establishing their temporal sequence is not my priority. I assume that the distinctiveness of the five types is primarily due to the presence of workshops, or groups of workshops, in the northwestern part of the Peshawar valley, each of which produced images of fairly homogeneous and idiosyncratic quality. At the same time, I do not believe that all the workshops, or the groups of workshops, existed simultaneously. For instance, primary pieces of Type I possibly predate any specimens of the four other types. Also, primary pieces of Type IV could not have been produced earlier than those of Types II or III.100 Still, [End Page 76] the chronological relationships between these types is far from linear, but rather an intricately overlapping process of transmission and borrowing of individual features. Therefore, I wish to treat these types first as synchronic units, attempting to refine our understanding of them in light of their mutual relationships and in conjunction with numerous smaller formal series. My next step will be to search for equivalents of the five types among bodhisattva images and then to extend the same approach to narrative reliefs.101
The first published version of this paper, which discussed Types I to IV, appeared in Korean in 1998 with the title "Kandara pulsang ŭi myŏt kaji yuhyŏng" (Several stylistic types of Gandhāran Buddha images), Misulsahak yŏn'gu, no. 219, pp. 5-40, and was presented in English at a biennial meeting of the American Council for Southern Asian Art in Philadelphia in 2000. A supplementary paper in Korean that treated Type V was published separately in the festschrift for Professor Kwon Young-pil in 2007. The present paper is a substantially revised and expanded version that integrates two earlier papers in Korean. I am grateful to Professors Joanna Williams, Lothar von Falkenhausen, and Maribeth Graybill for helping significantly to improve this paper, and also to the museums and their staffs in Pakistan and India for extending generous assistance during my numerous visits.
1. See Harald Ingholt and Islay Lyons, Gandhāran Art in Pakistan (New York: Pantheon Books, 1957), pp. 22-23; Wladimir Zwalf, A Catalogue of the Gandhāran Sculpture in the British Museum (London: British Museum, 1996), vol. 1, pp. 357-58. Recently B. N. Mukherjee argued three more pieces into the list of dated images; Mukherjee, "The Dated Gandhāra Sculptures—An Approach to Dating Gandhāra Objects," in Gandhāra Sculpture in the Government Museum and Art Gallery, Chandigarh, ed. D. C. Bhattacharyya (Chandigarh: Chandigarh Museum and Art Gallery, 2002), pp. 9-29. But according to Richard Salomon, an inscription on a standing bodhisattva cited by Mukherjee was apparently forged, and I doubt the authenticity of the image as well; Salomon, "The Name of Taxila: Greek Tαξιλα, Gāndhārī Taḳsaïla, Sanskrit Takṣaśilā, Pali Takkasilā," East and West, vol. 55 (2005), p. 270, n. 7. The readings of the other two inscriptions are also questionable (personal communications with Salomon and Andrew Glass, September 2007).
2. Diverse expositions of Gandhāran art put forward on the basis of its external sources are comprehensively discussed in Rekha Morris's unpublished Ph.D. dissertation, "A Prolegomena to a Study of Gandhāran Art" (University of Chicago, 1983), which is, unfortunately, already much dated. The most recent attempt at this approach is found in Lolita Nehru, Origins of the Gandhāran Style: A Study of Contributory Influences (Delhi: Oxford University Press, 1989).
3. As regards the controversial dating of the Kaniṣka era, undoubtedly the most crucial issue in the history of the region, some progress has been made recently with the advent of new evidence. On the basis of the new Kushan genealogy established through the Rabatak inscription from Afghanistan, Joe Cribb proposed the date of 100 or 120 CE for the beginning of the Kaniṣka era. Nicolas Sims-Williams and Joe Cribb, "A New Bactrian Inscription of Kanishka the Great," Silk Road Art and Archaeology, vol. 4 (1995/1996), pp. 74-142, esp. pp. 105-7. With a fresh analysis of astronomical sources, Harry Falk reached another date, 127 CE, which seems to be gaining wide acceptance; Falk, "The Yūga of Sphujiddhvaja and the Era of the Kuṣāṇas," Silk Road Art and Archaeology, vol. 7 (2001), pp. 121-36; "The Kaniṣka Era in Gupta Records," Silk Road Art and Archaeology, vol. 10 (2004), pp. 167-76.
4. For the scanty information available regarding the stratigraphic context of Gandhāran art, see a much dated work, K. W. Dobbins, "Gandhāran Art from Stratified Excavations," East and West, vol. 23 (1973), pp. 279-94; also Morris, "A Prolegomena," pp. 278-356. Attempts are also being made to reach a chronology based on numismatic evidence; cf. Joe Cribb, "Numismatic Perspectives on Chronology in the Crossroads of Asia," in Gandharan Art in Context, ed. Raymond Allchin et al. (Cambridge: The Ancient India and Iran Trust, 1997), pp. 215-30; Cribb, "Dating the Bīmarān Casket and the Development of Gandhāran Buddha Images," paper presented at the 14th Conference of the International Association of Buddhist Studies, August 2005, School of Oriental and African Studies, London.
5. The most influential among the various chronological theories proposed for Gandhāran art is probably John Marshall's in his Buddhist Art of Gandhāra (hereafter BAG; Cambridge: Cambridge University Press, 1960). Many scholars seem still to espouse Marshall's theory, whether explicitly or unconsciously. Cf. Daniel Schlumberger's review titled "Sir John Marshall and Gandhāra Art," Antiquity, vol. 35 (1961), pp. 176-80. For a most recent attempt at a chronology, see Zwalf, Gandhāra Sculpture in the British Museum, vol. 1, pp. 69-72.
6. About the time when the Italians started excavations in Swāt, Giuseppe Tucci remarked, referring to the need of the work in the region, "From those places come some of the finest specimens of Gandhāran art, found by clandestine diggers and now dispersed among private and public collections the world over; carefully conducted excavations made in these same territories offer good means for acquiring positive knowledge on the disputed questions of the successions of styles and stages, and the origins of the art of Gandhāra" ("Preliminary Report on an Archaeological Survey in Swāt," East and West, vol. 9, no. 4 , p. 284). Apparently he unjustifiably equated Swāt [End Page 77] with the rest of the Greater Gandhāra. I believe that such optimism is no longer shared in the field, but I wonder how much progress has been made, to date, in visually understanding each regional unit.
7. Criticizing Marshall's chronology of Gandhāran art based on stylistic reasoning, Schlumberger refers to the need for arranging objects "according to petrography, to geography, and to style as well as iconography (which was done by Foucher)" and continues, "The fragments could be grouped from the kind of stone they are made of, and from provenance (whenever this known), two secure bases; then from style, a basis much less secure, but none the less important; the ultimate problem remains to decide whether or not such a classification would allow some historical conclusions to be drawn, at least tentatively"; Schlumberger, "Sir John Marshall and Gandhāra Art," pp. 179-80. Although I do not entirely agree with Schlumberger on the unqualified security of petrography and geography, his comment contains an important point. The sorting of images according to such technological information, however, progresses very slowly. A positive development in recent years has been attempts to clarify the site provenance of various images, although these attempts do not seem to have been extended yet to visual aspects of images. Particularly notable are: Elizabeth Errington's works, "Western Discovery of the Art of Gandhāra and the Finds of Jamalgaṓhī" (Ph.D. diss., University of London, 1987); "Towards Clearer Attributions of Site Provenance for Some 19th Century Collections of Gandhara Sculpture," in South Asian Archaeology 1987, ed. Maurizio Taddei and Pierfrancesco Callieri (Rome: IsMEO, 1990), pp. 765-81; "Addenda to Ingholt's Gandhāran Art of Pakistan," Pakistan Archaeology, vol. 26, no. 2 (1991), pp. 48-70. A series of works by Francine Tissot on Sahrī-Bahlol are also useful: "The Site of Sahrī-Bahlol in Gandhāra," in South Asian Archaeology 1983, ed. J. Schotsmans and M. Taddei (Naples: Istituto Universitario Orientale, 1985), pp. 567-614; "The Site of Sahrī-Bahlol in Gandhāra, Pakistan: Further Investigations," in South Asian Archaeology 1985, ed. K. Frifelt and H. Sorenson (London: Curzon Press, 1989), pp. 417-25; "The Site of Sahrī-Bahlol in Gandhāra (part III)," in South Asian Archaeology 1987, ed. Maurizio Taddei, pp. 737-64; "Sahrī-Bahlol (part IV)," in South Asian Archaeology 1993, ed. A. Parpola and P. Koskikallio (Helsinki: Suomalainen Tiedeakatemia, 1994), pp. 733-44.
8. In this study I deliberately use the term "visual type" instead of the more conventional "stylistic type." Although style is one of the central concepts in arthistorical research and discourse, it has been endowed in the long history of the discipline with too multifarious, complex, and thus ambiguous—both semantically and pragmatically—meanings, and its significance and validity have often been questioned in modern scholarship. See Berel Lang, ed., The Concept of Style (Ithaca and London: Cornell University Press, 1979), particularly, Svetlana Alpers, "Style Is What You Make It: The Visual Arts Once Again," pp. 137-62; cf. Philip Sohm, Styles in the Art Theory of Early Modern Italy (Cambridge: Cambridge University Press, 2001), pp. 19-80. In archaeology the word "style" is still commonly used and has continued to be refined theoretically; it is usually applied at a micro level in relation to artifact analysis, or with an interpretational value associated with human behavior. See Margaret W. Conkey and Christine A. Hastorf, eds., The Use of Style in Archaeology (Cambridge: Cambridge University Press, 1990). In this work the word "style" is kept to a minimum and used in the sense of visual constancy.
9. Bodhisattvas are another subject commonly represented in statues. Unlike statues of the Buddha, which have a simple iconographic form, bodhisattvas wear elaborate dress, a turban, and a variety of jewelry, which are potentially advantageous for the creating of a detailed typology. An earlier attempt at the typology of the headdresses and hairstyles of bodhisattvas is notable in this regard: Carolyn Woodford Schmidt, "Bodhisattva Headdresses and Hairstyles in the Buddhist Art of Gandhāra and Related Regions of Swāt and Afghanistan" (Ph.D. diss., Ohio State University, 1990). Narrative reliefs were another important format in which the Buddha was presented. The continuity between statues and narrative reliefs may be presumed with some justification. But figures of the Buddha represented on a small scale in narrative reliefs are often hard to compare with larger statues. Furthermore, narrative reliefs present their own problems in visual analysis, such as the composition, the placement of figures, the suggestion of illusionary space and the depth of carving, which would complicate an approach of this kind at a preliminary stage.
10. For instance, Marshall, BAG, pp. 61-62, 83-84, 100, discussed the equivalents to my Types I, II, and V in his chronological schema, and Zwalf, Gandhāran Sculpture in the British Museum, vol. 1, pp. 80-81, nos. 2, 4, commented on the distinctiveness of objects similar to my Types I and V.
11. Typology is perhaps one of the least popular methods in art-historical research these days. In many fields art history has outgrown conventional explorations of visual styles and simple classification of visual types, and in Asian art as well a number of scholars are emulating the dominant practices of their colleagues in more advanced fields. But in some areas, such as Gandhāran sculpture, the rudimentary work of visual assortment is far from completed, and vast quantities of material still await such analysis. I hope in this work to incorporate more sophisticated considerations than have heretofore been applied in such approaches of the conventional mode.
12. Kubler, The Shape of Time (New Haven: Yale University Press, 1962), pp. 39-40.
13. Kubler defines prime objects as: "principal inventions," which "resemble the prime numbers of mathematics . . . have no divisors other than themselves," "differ(s) [End Page 78] from an ordinary object much as the individual bearer of a mutant gene differs from the standard example of that species." Thus a "prime object" means a thing in which a fundamental change or innovation in visual form is achieved. This term is quite useful in understanding the mechanism of development of form and the relationships among various material things in a dynamic process. Obviously, prime objects are more abstractions than realities, and it is futile to try to identify them in actuality. Understandably, therefore, one of the major criticisms of his theory was directed at his definition of this term. As Joyce Brodsky rightly pointed out, "an innovator is one who comes closest to revealing the pattern and rarely one who creates a new pattern"; "Continuity and Discontinuity in Style: A Problem in Art Historical Methodology," Journal of Aesthetics and Art Criticism, vol. 39, no. 1 (1980), p. 33.
14. This statue (acc. no. 2856), whose provenance is unknown, is currently displayed in the central hall of the Peshawar Museum. It was first published in Marshall, BAG, fig. 85, which locates it in the Lahore Museum; Saifur Rahman Dar, the former director of the Lahore Museum, informed me in a personal communication (January 2007) that this must be a mistake. This important image was not included in Ingholt and Lyons, Gandhāran Art in Pakistan. It was on view in the exhibition "The Art of Gandhāra, Pakistan" in Japan in 2002-2003, and was reproduced in a good color photo in its catalogue with the same title (Tokyo: Tokyo National Museum, 2002), pl. 1.
15. I suspect that a hole to contain a relic was carved in the missing upper part of the present uṣṇīṣa, which was regularly removed for display. For a detailed discussion of this matter, see Juhyung Rhi, "Images, Relics, and Jewels: The Assimilation of Images in the Buddhist Relic Cult of Gandhāra—or Vice Versa," Artibus Asiae, vol. 65, no. 2 (2005), pp. 175-77, figs. 6, 7, 9.
16. In coiffure and facial type this Buddha shows affinities with a Buddha carved on the Bīmarān reliquary. But given the position of the remaining right arm, the right hand is unlikely to have been positioned in front of the chest as in the Bīmarān reliquary. Cf. Zwalf, Gandhāra Sculpture in the British Museum, vol. 2, pl. 659: group A (1).
17. As regards the finds from Swāt, see, for example, those from Butkara I, reproduced in D. Faccenna, Sculptures from the Sacred Area of Butkara I, IsMEO Reports and Memoirs, vol. II,
19. The only notable difference in the face is that irises and pupils are not apparent. When I examined this Buddha—which had been tightly wrapped for more than two decades, since its return from an exhibition in Japan in 1988—in the dark storage of the Lahore Museum in January 2007, the irises and pupils were not clearly visible. As I recall, the incised lines that delineated them seem to have been eroded. Or they might have been painted in and, over the centuries, flaked off.
23. Zwalf, Gandhāra Sculpture in the British Museum, vol. 1, pp. 79-80 (no. 2). Zwalf mentions this image, as well as Figures 1-4, as an "early type," and classifies as similar several more examples, including three pieces in the Museum für Indische Kunst, Berlin (now, the Museum für Asiatische Kunst, Berlin) (acc. nos. I44, I114, I527; "I14" referred to by Zwalf must be a mistake for "I114," which will be discussed in n. 24), and another one in the Cleveland Museum of Art. Of the four comparable pieces mentioned by Zwalf, Berlin I44 is significantly different from the others and associable to a separate group that I attribute to the Barikot area in Swāt (see n. 95). The Cleveland piece may be loosely included in this type; see Stanislaw J. Czuma, Kushan Sculpture: Images from Early India (Cleveland: Cleveland Museum of Art, 1985), p. 196 (no. 107). It measures 1.2 m high, more or less similar in size to the British Museum Buddha. The rendering of the body and drapery is also quite similar. But the facial features are somewhat different: most notably, the eyebrows are connected, a frequent feature of Buddha images from Swāt. According to a personal communication from Czuma (2003), the top of the uṣṇīṣa is drilled in the center. In the catalogue cited above, he dates the image to the second half of the second century, but without any supporting evidence.
24. A head in Berlin (acc. no. I114) shows almost identical facial features except that its uṣṇīṣa has a snail-shell coiffure. Despite this difference, they may be regarded as products of the same workshop.
26. Albert Grünwedel, Buddhist Art in India, trans. Agnes Gibson and revised/enlarged by James Burgess (London: Bernard Quaritch, 1901), p. 83, n. 3 (a note by Burgess). By 1901 the Museum für Völkerkunde, Berlin, the predecessor of the Museum für Indische Kunst and the present Museum für Asiatische Kunst, owned sixty-three sculptural pieces from Gandhāra, of which the majority came from Swāt. A number of them were purchased from Gottlieb Leitner, who once lived in Lahore. See Errington, "The Western Discovery," p. 162. In Berlin there is one more head that could be considered Type I; Ernst Waldschmidt, "Die Entwicklungsgeschichte des Buddhabildes in Indien," Ostasiatische Zeitschrift, vol. 6 (1930), fig. 34b. It has a large hole in the center of the uṣṇīṣa. [End Page 79]
27. Faccenna, Sculptures from the Sacred Area of Butkara, vol. 1, pls. CXV, CXVI; (also for a bodhisattva or deva head with a similar face) pls. LXXXIX, CCLI-a, CCLVIII.
28. Ahmad Hasan Dani, "Excavations at Andandheri," Ancient Pakistan, vol. 4 (1968-1969), pl. 12; The Exhibition of Gandhāra Art of Pakistan (Tokyo: NHK, 1984), pl. I-19. For other examples from Andan-Ḍherī, including those in narrative reliefs, see Dani, "Excavations at Andandheri," pls. 13a (probably from another seated image), 28b. A similar seated Buddha is in the collection of the Royal Ontario Museum, Toronto; David Jongeward, Buddhist Art of Pakistan and Afghanistan: The Royal Ontario Museum Collection of Gandhara Sculpture (Toronto: Centre for Asian Studies, University of Toronto, 2003), no. 4.
29. For the two reliefs from the Mardan group, see Ingholts and Lyons, Gandhāran Art in Pakistan, nos. 54, 59; for the Mardan group in general, Marshall, BAG, pp. 40-41; Errington, "Western Discovery," pp. 210-14.
30. Marshall is of the same opinion (BAG, p. 62): "These two statues are so alike that we may well believe that they came from the same atelier, if not from the hand of the same artist."
31. A standing Buddha in Cleveland (see n. 23) could be grouped in the same category as the British Museum Buddha.
32. Marshall, BAG, pp. 40-62, esp., pp. 61-62: "For these differences [from the reliefs of the Mardan group] which are, no doubt, due to improving technique and changing taste, I do not think that more than a couple of decades need be postulated" (the brackets are mine). Thus, he places the two statues chronologically between the Mardan reliefs at the beginning of the Adolescent period and the Sikri reliefs at the very end of the period. This proposed chronological sequence, not to mention the presumed interval of a couple of decades, is pure speculation corroborated neither by substantial evidence nor by sound reasoning.
33. Ingholt and Lyons, Gandhāran Art in Pakistan, p. 111 (no. 210).
34. Perhaps this suggests that installing a relic in an uṣṇīṣa was not in fashion in the Sahrī-Bahlol area or in monasteries where images of this type were popular dedications.
35. The differences are not of sufficient extent to suggest a significant chronological gap between the two.
36. A Buddha similar to this one, formerly at the briefly existing Islamabad Museum, is currently on view at the Taxila Museum; see M. A. Halim and Sarwat Baig, Islamabad Museum, A Guide (Islamabad: Islamabad Museum, 1994), photo on p. 9; this piece is not included in the new comprehensive catalogue of the Taxila Museum, Muhammad Ashraf Khan et al., A Catalogue of the Gandhāra Stone Sculptures in the Taxila Museum, 2 vols. (Department of Archaeology and Museums, Government of Pakistan, 2005), probably because it was moved there after the publication of the catalogue. From its broken halo to the ankles, it measures 2.34 m high, slightly smaller than the Sahrī-Bahlol Buddha discussed above (Figs. 11, 12). It is known to have been confiscated from an illegal transaction in Mardan. That an image of such size had been previously unknown is curious. Its stern face seems to belie its similarities to Buddha images identified as belonging to Type II. Also curious is the absence of dust and traces of wear from parts of the image where we would normally expect to find them. These anomalies lead us to question its authenticity.
37. These two Buddhas were uncovered during the excavations of Mound B at Sahrī-Bahlol during the 1909-1910 season. They were discovered lying face downward, flanking a square stūpa base in the northern area of the main stupa court. Therefore they must have been installed one on either side of this stūpa, but it is doubtful that this was their original placement; see D. B. Spooner, "Excavations at Sahrī-Bahlol," Archaeological Survey of India, Annual Report, 1909-1910 (1914), pp. 48, 55-56, pl. XXII-b, c. (The Archaeological Survey of India, Annual Report is hereafter referred to as ASIAR.)
38. Regarding the discovery of this bodhisattva head, see M. A. Stein, "Excavations at Sahrī-Bahlol," ASIAR 1911-1912 (1915), fig. 19; cf. Juhyung Rhi, "A Bodhisattva Head in the Peshawar Museum," Ancient Pakistan, vol. 16 (2007).
39. When intact, it would have been slightly over 2 m high. Marshall assigns it, together with the Taxila Buddha that will be discussed below (Fig. 19), to the Early Maturity period (100-140 CE). He remarks that its affinity to two Buddhas in Peshawar and Lahore (the primary pieces of my Type I, Figs. 1-4) is manifest; Marshall, BAG, p. 83 and fig. 100. An almost identical Buddha head is in the Chandigarh Museum (acc. no. 235A, provenance unknown; Bhattacharyya, ed., Gandhāra Sculpture in Chandigarh, no. 72). It measures 28 cm high, about the same as the head of Figure 16, and it most probably came from the same site and workshop. Before the Partition, it was also in the collection of the Lahore Museum.
40. The Buddha in the British Museum was acquired in Peshawar, but its provenance is unknown; Zwalf, Gandhāran Sculpture in the British Museum, vol. 1, p. 79 (no. 1). The one in the Taxila Museum was excavated at Dharmarājikā, Taxila; Marshall, Taxila (Cambridge: Cambridge University Press, 1951), vol. 2, p. 724, vol. 3, pl. 225 (no. 158); Ashraf Khan et al., Gandhāra Stone Sculptures in the Taxila Museum, vol. 1, p. 113 (no. 52). Marshall states, "[between this Taxila Buddha and the broken standing Buddha in Lahore (Fig. 16)] there is little apparent difference except in the eyes and in the quality of the workmanship, which is hardly up to the standard" (BAG, p. 84 and fig. 111; the brackets are mine). He notes the difference between open and half-closed eyes, but remarks, "Throughout the first phase of the Maturity Period the treatment of the [End Page 80] eyelids was a matter for the personal discretion of the artist, not one canonical rule" (BAG, p. 83).
41. A Buddha head in Chandigarh (Bhattacharyya, ed., Gandhāra Sculpture in Chandigarh, no. 128, provenance unknown) is slightly bigger, but otherwise almost identical to the British Museum Buddha.
42. Other comparable examples reproduced in publications include: (1), (2) two heads from Takht-i-Bāhī, collection unknown; D. B. Spooner, "Excavations at Takht-i-Bāhī," ASIAR 1907-1908 (1911), pls. XLIXI-1, 7; (3) provenance unknown, in the National Museum, Karachi, Gandhāra Sculpture in the National Museum of Pakistan (Karachi: Department of Archaeology and Museums, 1964), pl. XVIII; (4) from Sikri, in the Lahore Museum, Ingholt and Lyons, Gandhāran Art in Pakistan, no. 273 (the top of the uṣṇīṣa is peculiarly cut); (5), (6) two heads, provenance unknown, Bhattacharyya, ed., Gandhāra Sculpture in Chandigarh, nos. 127, 131; (7) provenance unknown, in the Lucknow State Museum, N. P. Joshi and R. C. Sharma, Catalogue of Gandhāra Sculpture in the State Museum, Lucknow (Lucknow: Lucknow State Museum, 1969), fig. 40; (8) provenance unknown, in the Kyoto National Museum, Gandāra no chōkoku (Nara: Nara National Museum, 1985), no. 47. Marshall places a Buddha head from Mohṛa-Morādu, Taxila, in the same category as the two Buddhas in the Lahore and Taxila Museums (Figs. 16, 19), although it seems considerably different even in a poor black-and-white photo published in his BAG, fig. 109, and in Taxila, vol. 2, p. 724, no. 159; it may well be a variant. For an unknown reason, this piece is not listed in the most recent Taxila Museum catalogue (2005).
43. The Buddha in Peshawar, not illustrated in this paper, is reproduced in Ingholt and Lyons, Gandhāran Art in Pakistan, no. 248, and more recently in an excellent color photo in a Japanese exhibition catalogue, Buddha, The Spread of Buddhist Art in Asia (Tokyo: NHK, 1998). The Buddha in the Karachi Museum (Fig. 21), which was originally in the Lahore Museum, must have been of an impressive size, as it measures 79 cm high even without the missing lower part of the body. Besides, another seated image in the Lahore Museum (h. 1.1 m; Ingholt and Lyons, Gandhāran Art in Pakistan, no. 251), although it lacks a head, is quite close in dress and body to the above two pieces in Karachi (Fig. 21) and Peshawar and may well be related to them. Several more examples in various collections belong to the same category: another Buddha in Karachi (Ingholt and Lyons, Gandhāran Art in Pakistan, fig. XVI.2); one in the Asian Art Museum, San Francisco (Ingholt and Lyons, Gandhāran Art in Pakistan, fig. XVI.1); one in the Baltimore Art Gallery; one in the Tokyo National Museum; two more Buddhas in Lahore (J. Ph. Vogel, "Inscribed Gandhāran Sculpture," ASIAR 1903-1904 , p. 249; Ingholt and Lyons, Gandhāran Art in Pakistan, no. 247). The last two seem significant variants of this type.
44. Triad from Sahrī-Bahlol Mound A, in Spooner, "Excavations at Sahrī-Bahlol," ASIAR 1906-1907 (1909), p. 114; from Sahrī-Bahlol Mound D, in Stein "Excavations at Sahrī-Bahlol," ASIAR 1911-1912 (1915), pl. XLVI, fig. 28.
45. See Ingholt and Lyons, Gandhāran Art in Pakistan, no. 254.
46. J. C. Harle, "A Hitherto Unknown Dated Sculpture from Gandhāra: A Preliminary Report," South Asian Archaeology 1973, ed. J. E. van Lohuizen-de Leeuw and J.M.M. Ubaghs (Leiden: E. J. Brill, 1974), pp. 129-35; Czuma, Kushan Sculpture, no. 109. Joe Cribb suggests that it should be dated to the 105th year of the Kaniṣka era (Cribb, "Numismatic Perspectives," p. 226).
47. This piece, which was first reported in the early 1970s, has raised a number of complex questions about its date, iconography, and even authenticity. Most epigraphy specialists, including Richard Salomon (personal consultation) seem to find this inscription dependable; also see Gérard Fussman, "Documents épigraphiques Kouchans," Bulletin de L'École d'Extrême-Orient, vol. 61 (1974), pp. 54-58. Its facial features are close to those of the British Museum Buddha (Fig. 20), and many other elements in the triad seem to fit well with what we can expect in Type II, although the unbelievably immaculate state of preservation, except for insignificant breakages, still troubles me.
48. One reproduced in Fig. 25 came from Koṭ, a village "located some two miles northeast of the main group of monuments at Takht-i-Bāhī"; ASIAR 1922-1923, p. 98 and pl. X-c. Takht-i-Bāhī is near to Sahrī-Bahlol. The Japanese catalogue of a Silk Road exhibition held in Nara in 1988, The Route of Buddhist Art (Nara: Nara National Museum, 1988), gives the wrong provenance, Taxila. All other three examples are from Sahrī-Bahlol Mound B and are currently in the Peshawar Museum; Ingholt and Lyons, Gandhāran Art in Pakistan, nos. 87, 88, 89. Two of them (nos. 88, 89) are closer to the major examples of Type II, while one (no. 87) seems much degenerated in form. Thus, the theme of the Display of a Snake Bowl to Uruvilvā Kāśyapa was almost invariably represented in this type, except for the famous "Apollo Buddha," which I believe also represented the theme; Benjamin Rowland, The Art and Architecture of India (Harmondsworth: Penguin, 1977), fig. 66. This Buddha, which used to be in the Guides' Mess, Mardan, during the British colonial period (provenance and present whereabouts unknown), shows the wavy hair of Type III.
49. In narrative reliefs, Buddha figures with a similar hairstyle and a mustache are common. But few are comparable in quality with independent statues of this type, and affinities are merely approximate. Although a number of Buddha figures in reliefs on the Sikri stūpa in the Lahore Museum (Ingholt and Lyons, Gandhāran Art in Pakistan, nos. 60, 68, 70, 96, 104, 126) may be identifiable with this type, they differ considerably in the drapery and the proportions of the body. [End Page 81]
50. For Sahrī-Bahlol, see a series of works by Francine Tissot cited in n. 7.
51. I have noted monasteries at Takht-i-Bāhī and Sahrī Bahlol as possible Mahāyāna bases in the Peshawar valley; Juhyung Rhi, "Early Mahāyāna and Gandhāran Buddhism: A Reassessment of the Visual Evidence," The Eastern Buddhist, vol. 35, no. 1/2 (2003), pp. 152-202.
52. If importance was commensurate with size, standing images must have been accorded far greater importance among Gandhāran Buddhist statuary.
53. It is well known that in Kushan Mathurā half-closed eyes gradually replaced wide-open ones. If this transition also occurred in Gandhāra, then Gandhāran Buddhas with wide-open eyes would predate those with half-closed eyes. Francine Tissot once noted this remarkable difference in the depiction of eyes; Tissot, "The Problem of a Stylistic Vocabulary for Gandhāran Art," in Investigating Indian Art, ed. Marianne Yaldiz and Wibke Lobo (Berlin: Staatliche Museen, 1985), pp. 363-68; cf. Rhi, "Images, Relics, and Jewels," pp. 203-6. Also see Marshall's assessment, cited in n. 40.
54. Personal communication with Wladimir Zwalf at the British Museum in 1989.
55. For the head from Sahrī-Bahlol Mound B (Fig. 28), also see Ingholt and Lyons, Gandhāran Art in Pakistan, nos. 268, 269. The other head, from Sahrī-Bahlol Mound A, is reproduced in ASIAR 1906-1907, pl. XXXIV-d and in Ingholt and Lyons, Gandhāran Art in Pakistan, no. 270.
56. For other notable examples, refer to: (1), (2) two more heads from Sahrī-Bahlol Mound A, collections unknown, ASIAR 1906-1907, pls. XXXV-15, XXXIV-c; (3)-(7) five heads from Takht-i-Bāhī, collections unknown, ASIAR 1907-1908, pls. XLIX-2, 3, 4, 9, 11; (8) provenance unknown, in the British Museum, h. 38.7 cm, Zwalf, Gandhāra Sculpture in the British Museum, vol. 2, no. 39; (9) provenance unknown, in the Lahore Museum (Benjamin Rowland, Gandhāra Sculptures from Pakistan Museums (New York: Asia Society, 1960); (10) from Pālāṭu-Ḍherī, collection unknown, ASIAR 1902-1903, pl. XXV-d. Pālāṭu-Ḍherī is near Charsada, which is only 20 kilometers as the crow flies from Takht-i-Bāhī and Sahrī-Bahlol.
57. Ingholt states that the image came from Sahrī-Bahlol in the 1906-1907 season. Ingholt and Lyons, Gandhāran Art in Pakistan, no. 223. If this is correct, it came from Mound A, the site also of the Buddha head of Figure 28. But I was not able to locate the image in Spooner's excavation report of the site in ASIAR 1906-1907.
58. ASIAR 1907-1908, p. 147. The head of this Buddha is identical to a Buddha head in the Royal Ontario Museum; Jongeward, Buddhist Art of Pakistan and Afghanistan: The Royal Ontario Museum, no. 2. The two were most probably made by the same sculptor.
59. Cf. ASIAR 1906-1907, pl. XXXVI-b.
60. Cf. Ingholt and Lyons, Gandhāran Art in Pakistan, no. 202.
61. For the image reproduced here, see Ingholt and Lyons, Gandhāran Art in Pakistan, no. 234, which states that it originated at Takht-i-Bāhī. Of numerous other examples, see some notable ones reproduced in publications: (1)-(3) three images from Takht-i-Bāhī, ASIAR 1907-1908, pls. XLVII-a, XLVIII-a, b (the first two images are from the same workshop, possibly by the same hand; the third one is somewhat different); (4) from Sahrī-Bahlol Mound D, in the Peshawar Museum, ASIAR 1911-1912, pl. XLVIII, fig. 33; (5) provenance unknown, in the Royal Ontario Museum, Jongeward, Buddhist Art of Pakistan and Afghanistan: The Royal Ontario Museum, no. 3; (6) provenance unknown, in the Aurora Art Museum, Taipei, Lidai fodiaoyishu zhimei (Taipei: National Museum of History, 2006), pl. 1 (specimens (1)-(6) are most probably from the same workshop); (7) provenance unknown, in the Seattle Art Museum, Ingholt and Lyons, Gandhāran Art in Pakistan, fig. XIII.3; (8) provenance unknown, in the Royal Scottish Museum, Edinburgh, Ingholt and Lyons, fig. XV.2; (9) from Takht-i-Bāhī, in the Peshawar Museum, ASIAR 1907-1908, pp. 146-47, fig. 7, or Ingholt and Lyons, no. 235 (in this Buddha, the hair flows over the uṣṇīṣa without clear demarcation, unlike other Buddhas of this type; parallels are also found in standing images from Takht-i-Bāhī in the Museum für Asiatische Kunst, Berlin, and in Worcester Art Museum, cf. Grünwedel, Buddhist Art in India, fig. 122, and Benjamin Rowland, "Gandhāra and Late Antique Art," American Journal of Archaeology, vol. 46 , fig. 3); (10)-(11) two images, provenance unknown, in Chhatrapati Shivaji Museum (formerly Prince of Wales Museum) in Mumbai, Pramod Chandra, Stone Sculpture in the Prince of Wales Museum (Bombay: Prince of Wales Museum of Western India, 1974), nos. 45, 46; (12) provenance unknown, in the Indian Museum, Kolkata, Ingholt and Lyons, fig. XV.4 (beneath this image is a wide knob, which was originally inserted into a lotus pedestal; this assemblage is a distinctive feature of Type IV); (13) provenance unknown, in the Allahabad Museum, Pramod Chandra, Stone Sculpture in the Allahabad Museum (Poona: American Institute of Indian Studies, 1970), pl. XLIV (seated on a lotus).
62. See for example: (1) from Sahrī-Bahlol Mound B, in the Peshawar Museum (acc. no. 3083), publication unknown; (2) provenance unknown, in the Peshawar Museum, unpublished; (3) from Amlūk, Swāt (probably not made there), Evert Barger and Phillip Wright, Excavations in Swāt and Explorations in the Oxus Territories of Afghanistan, Memoirs of the Archaeological Survey of India, no. 64 (Delhi: Manager of Publications, 1941), pl. IV-2; (4) from Kadam-kuki Khel in Swāt (again, probably not made there), formerly in the Museum für Völkerkunde, Berlin, Grünwedel, Buddhist Art in India, fig. 137. In (3) and (4) the drapery appears quite degenerated.
63. See n. 3. [End Page 82]
64. ASIAR 1928-1929, p. 142.
65. Cf. ASIAR 1907-1908, pl. 48d; Ingholt and Lyons, Gandhāran Art in Pakistan, no. 221.
66. A Buddha head from Ghaz-Ḍherīnear Chārsada shows a face identical to the Type III heads discussed above; see ASIAR 1902-1903, pl. XXV-c. In Mathurā, Buddha images with snail-shell curls appeared in the early second century of the Kaniṣka era, and this may provide a terminus ad quem for the appearance of the hairstyle in Gandhāra. The Lahore Buddha (Fig. 37) and the Ghaz-Ḍherī head are possibly among the earliest of the images with snail-shell curls in Gandhāra.
67. Some notable examples include several seated images in dharmacakra mudrā (1-4), a seated image in dhyāna mudrā (5), two standing images (6-7), and a head only (8): (1) from Sahrī-Bahlol, in the Indian Museum, Kolkata, Alfred Foucher, L'Art gréco-bouddhique du Gandhāra (hereafter AGBG), vol. 2 (Paris: Leroux, 1922), fig. 483 (Foucher wrongly places the image in the Peshawar Museum; the facial features are quite similar to those of Fig. 37); (2) from Takht-i-Bāhī, in the National Museum, Karachi, ASIAR 1907-1908, pl. XLVII-b (similar to the first one); (3) from Takht-i-Bāhī, collection unknown, ASIAR 1907-1908, pl. XLVII-c (on a lotus seat, somewhat decadent); (4) from Takht-i-Bāhī, in the Peshawar Museum, Ingholt and Lyons, Gandhāran Art in Pakistan, no. 249; (5) from Sahrī-Bahlol Mound C, two standing images in the Peshawar Museum, Ingholt and Lyons, no. 233; (6) from Takht-i-Bāhī, in the Peshawar Museum, Ingholt and Lyons, nos. 204-5; (7) from Sahrī-Bahlol Mound B, in the Peshawar Museum (acc. no. 2854), unpublished; (8) from Sahrī-Bahlol Mound B, in the Peshawar Museum, ASIAR 1909-1910, pl. XXI-c, or Ingholt and Lyons, no. 272. Images (4), (6), (7), and (8) have in common rather plump bodies, which may be closer to Type IV rather than to Type III, although differing from Type IV as well. Images with such plump bodies are usually regarded as late works. The head of the standing image (6) is quite similar to the head (8), and could have come from the same workshop. Again, the head (8) is identical, except for the hair, to another head with wavy hair also from Takht-i-Bāhī and now in the Peshawar Museum (acc. no. 2868), no doubt from the same workshop. It seems unlikely that different hairstyles signified different Buddhas; probably the hairstyles were simply designs, without iconographical significance.
68. Of the few pieces in Kolkata, one more is reproduced in D. G. Majumdar, A Guide to the Sculptures in the Indian Museum, part 2 (Delhi: Manager of Publications, 1937), no. 274 and frontispiece. For two pieces in the Lucknow State Museum, see Joshi and Sharma, Catalogue of Gandhāra Sculpture in the State Museum, Lucknow, figs. 34, 36; the former is quite similar to images in Kolkata, but Figure 6 differs in its drapery, which consists of double-fold incised lines.
69. Ingholt and Lyons, Gandhāran Art in Pakistan, no. 249.
70. A Buddha from Loriyān-Tangai in the Indian Museum, Kolkata, also has a snail-shell coiffure; Majumdar, A Guide to the Sculptures in the Indian Museum, part 2, no. 266; Huntington Archive, scan no. 0004870. It too is seated on a cushion, and it is exceptionally large (h. 127.6 cm) among seated Buddhas of Type IV, which usually measure 60-80 cm high. Another image, formerly in the Gai collection in Peshawar, is also seated on a cushion, whose lower part is no longer extant. In the drapery over the lower part of its body it resembles the Kolkata Buddha (Fig. 38), but the hairstyle is a schematized variation of that in Type II; see Ingholt and Lyons, Gandhāran Art in Pakistan, no. 245.
71. Zwalf, Gandhāra Sculpture in the British Museum, vol. 1, no. 24; Foucher, AGBG, vol. 2, fig. 456.
72. This image from Loriyān-Tangai is cited in Grünwedel, Buddhist Art in India, fig. 112. Also see (1) another piece from Loriyān-Tangai in Kolkata, on a cushioned square pedestal under a tree; reproduced in ASIAR 1903-1904, pl. LXVIII; (2) an image of unknown provenance but most likely from Loriyān-Tangai (with a cylindrical, tapering knob beneath) in the Yale University Art Gallery, Ingholt and Lyons, Gandhāran Art in Pakistan, fig. XIV; and (3) one in Lahore, on a cushioned, square pedestal, Ingholt and Lyons, no. 232.
73. For another triad, see Grünwedel, Buddhist Art in India, fig. 147, or Foucher, AGBG, vol. 2, fig. 408. In this one the Buddha's body looks slightly heavier.
74. Also see triads reproduced in Miyaji Akira, "Gandāra sanzon keishiki no ryōkyōji bosatsu ni tsuite" (On two attendant bodhisattvas in Gandhāran triads), in Indo Pakisutan bukkyō zuzō chōsa (Hirosaki: Hirosaki University, 1985), pls. IV-1, VII-2; Isao Kurita, Gandhāran Art I: The Buddha's Life Story (2nd ed., Tokyo: Nigensha, 2003), figs. 406, 407, 411, 412.
75. Errington points out that the provenance of this Buddha, referred to as "Shāh-jī-kī-Ḍherī" in Ingholt and Lyons, Gandhāran Art in Pakistan, no. 211, was originally written in the Lahore Museum records as "Shah-ki-Dheri," which probably meant Taxila, not "Shāh-jī-kī-Ḍherī," which was discovered in the early 20th c.; see Elizabeth Errington, "Addenda to Ingholt's Gandhāran Art in Pakistan," p. 56. A quite similar image, whose upper part only remains, is in Chandigarh (Bhattacharyya, ed., Gandhāra Sculpture in Chandigarh, no. 155, provenance unknown). Also wearing the robe in the himation style, it most likely came from the same workshop.
76. A similar image is found on a stele in Chandigarh; see Alfred Foucher, The Beginning of Buddhist Art (Paris: Paul Geuthner, 1917), pl. 27; Bhattacharyya, ed., Gandhāra Sculpture in Chandigarh, no. 153.
77. Elizabeth Errington, "Loriyan Tangai," Dictionary of Art (London: McMillan, 1996), vol. XIX, p. 690.
78. Reproduced in a full view in Kurt Behrendt, The Buddhist Architecture of Gandhāra (Leiden: Brill, 2004), [End Page 83] fig. 88; this photo was originally kept in the India Records at the British Library (Punjab Photographs # 1042).
79. A female statue of unknown identity excavated at Sahrī-Bahlol Mound B has the exactly same face as the seated Buddha in Kolkata discussed above (Fig. 38). Unquestionably it came from the same workshop, which apparently supplied statues both to Loriyān-Tangai and to the monastery at Sahrī-Bahlol, and probably to many others as well; see ASIAR 1909-1910, pl. XXII-a; Ingholt and Lyons, Gandhāran Art in Pakistan, no. 400.
80. Rhi, "Images, Relics, and Jewels," p. 179.
81. Marshall (BAG, p. 100) contrasts this image with a standing Buddha of Type II from Taxila (Fig. 19): the style of the latter is "free and unconstrained—the work of a sculptor with an innate sense for simple rhythmic form"; that of the former is "tight, laboured and self-conscious—the work of a craftsman rather than an artist, who, like many of his contemporaries, was showing off his skill in the mechanical excellence and precision of his carving." I would not object to his statement about the Lahore Buddha (Fig. 46), a characterization which essentially fits the majority of Gandhāran images, but I can hardly agree with his overestimation of the Taxila Buddha (Fig. 19), which seems to me no better than the former. He concludes, "Such figural reliefs on the pedestals of statues did not come into fashion until the middle of the second century A.D., and to judge by its style in this particular relief can hardly be earlier than the last quarter of that century." This also seems to me a dubious argument.
82. I believe that this feature is related to the installation of a relic in the uṣṇīṣa.
83. Cf. Bhattacharyya, ed., Gandhāra Sculpture in Chandigarh, no. 542. It has a hole behind the uṣṇīṣa.
84. A head in Chandigarh (Bhattacharyya, ed., Gandhāra Sculpture in Chandigarh, no. 100; provenance unknown) is, although slightly smaller (19 cm high), almost identical to Figure 47. The hole for a relic is at the back of the uṣṇīṣa. In Chandigarh are a few more heads of the same type (Bhattacharyya, ed., nos. 134, 152, 571, the provenance of the first two unknown, the third from Sikri). In Bhattacharya, ed., no. 152, small curls cover the whole head, not just its upper part, probably a variation of the more common hairstyle in Type V. A head from Jamālgaṓhī in the British Museum, 22.6 cm high (Zwalf, Gandhāra Sculpture in the British Museum, vol. 2, pl. 41), is quite similar to Figure 47 and about the same size, but it is made of light gray schist, a different material from the Lahore Buddha, and its facial features are more crisply defined. It has a hole "behind the uṣṇīṣa and against its base"; Zwalf, Gandhāra Sculpture in the British Museum, vol. 1, p. 94. The torso of an image in the Ram collection in Delhi also shows great affinities with Figure 46. The Museum für Völkerkunde, Berlin, had two heads of Type V; one of them was lost during World War II; Staatliche Museen zu Berlin: Dokumentation der Verluste, band III (Berlin: Museum für Indische Kunst, 2002), p. 38 (IC 34940). Numerous extant examples comparable to the two Lahore images (Figs. 46, 48) indicate the popularity of this type.
85. Ingholt and Lyons, Gandhāran Art in Pakistan, nos. 215, 206. Zwalf was also aware of stylistic similarity between the British Museum head cited above (n. 84) and Figure 50; Zwalf, Gandhāra Sculpture in the British Museum, vol. 1, p. 81, no. 4, n. 1. Two statues, one in the National Museum, Karachi, and the other in the Newark Museum (Archives of Asian Art 37 , p. 121, fig. 33), and a head in the Rijksmuseum, Amsterdam, are other similar examples. Another standing statue, in the Royal Ontario Museum, is also similar to Figure 50, but somewhat heavier; Jongeward, Buddhist Art of Pakistan and Afghanistan: The Royal Ontario Museum, no. 1.
86. For the New Delhi Buddha in Figure 52 (acc. no. 48.413), see the American Institute of Indian Studies Photo Archive, no. 626. The standing Buddha in Lucknow (Joshi and Sharma, Catalogue of Gandhāra Sculpture in the State Museum, Lucknow, fig. 30) is a loan from the Chandigarh Museum (Bhattacharyya, ed., Gandhāra Sculpture in Chandigarh, no. 541). The New Delhi Buddha is also from Chandigarh. Both originally came from the Lahore Museum. For the piece in the Guimet (acc. no. AO 2902), which is known to be from Shahbaz-Garhi, see Alfred Foucher, "Sculptures gréco-bouddhiques (Musée du Louvre)," Monuments et Mémoires de la Fondation Eugène Piot, tome 7 (Paris, 1900), p. 41, fig. 1. I am grateful to M. Pierre Cambon of the Musée Guimet for providing this information.
87. National Museum, New Delhi, acc. nos. 49.24, 87.1153. For the latter, see the American Institute of Indian Studies Photo Archive, no. 625.
88. Zwalf cites three more pieces in the British Museum collection as similar (Gandhāra Sculpture in the British Museum, vol. 2, pls. 4, 5, 40); although they may also be loosely grouped in this category, their similarities are much more limited than those discussed above.
89. Majumdar, A Guide to the Sculptures in the Indian Museum, part 2, no. 336 and pl. IIa; Foucher, AGBG, vol. 2, fig. 481.
90. This Buddha, a loan from Chandigarh, is known to have come from Sikri; Bhattacharyya, ed., Gandhāra Sculpture in Chandigarh, no. 548.
91. Peshawar Museum, acc. no. 2852.
92. Foucher, BBA, p. 119; Marshall, BAG, p. 101. This Buddha, whose provenance is unknown, was acquired by the Guides in Mardan and installed in their mess in the late 19th c. It is still sometimes referred to by its previous collection, but has long been in the Peshawar Museum. Peculiarly, it was not reproduced in Ingholt and Lyons, Gandhāran Art in Pakistan.
93. A seated Buddha, quite parallel to this standing image, was recently purchased by the Virginia Museum of Fine Arts (acc. no. 2002.556). Although I have seen it only in a photo, I find these two images visually very close and probably made in the same workshop. [End Page 84]
95. Outside the Peshawar valley, several formal series are identifiable. For instance, Buddha images from the Kabul valley centered around the Kāpiśī area show quite consistent visual features and thus are easily recognizable (e.g., a standing Buddha from Paitava in the Musée Guimet, Musée national des Arts asiatiques—Guimet [Paris, 2001], p. 22). The wavy hair spreading out from the center of the hairline is divided in vertical rows, which somewhat resembles the hairstyle of Buddhas from the Peshawar valley (Figs. 58, 59). Another notable type is characterized by a high forehead, a low uṣṇīṣa, the hair rendered in series of small semicircles arranged in multiple tiers, and a sensuous-looking face (e.g., a Buddha head from Amluk; Barger and Wright, Excavations in Swat, pl. IV.1), reminiscent of images from India proper, particularly from Mathurā. This type seems to have flourished in the middle of the Swāt valley centered on the Barikoṭ area. Other distinctive types, produced in the Mingora area and the Chakdara area in Swāt, can be distinguished, although they are generally of crude craftsmanship. In a previous study, which has yet to appear in English, I examined several groups among sculptural finds from Swāt: "Swat chogak yangsikkwan chomyŏng" (An overview of sculpture from Swāt in visual style), Misul charyo, no. 60 (1998), pp. 71-98.
96. It is common knowledge that a distressingly large number of fakes are already in various collections even in reputable museums in the United States, Europe, and Japan. With significant advances in technical knowledge and skills in recent years, the forgers seem capable of producing images of increasingly higher quality, which convincingly recreate typical features of the five major types. It is becoming increasingly harder to detect forgeries by visual examination alone.
97. I have discussed such architectural structures in detail in a paper, "Gandhāra pulgyo sawŏn ŭi chosang pongan yangsik kwa ŭimi" (Modes of installation of images in Gandhāran monasteries and their implications), Misulsa yŏn'gu, vol. 8 (1994), pp. 165-73.
98. Marshall, BAG, pp. 40, 63-67.
99. Ingholt and Lyons, Gandhāran Art in Pakistan, pp. 13-46.
100. Among dated images from Gandhāra, we only know for sure that a specimen in Type III (Fig. 35) is dated "year 89," and the type was well established by the early third century CE (see p. 61 and n. 46). A triad dated "year 5" (Fig. 24) is perhaps attributable to the second century of the Kaniṣka era, and if so, is datable sixteen years later than the year 89 image. As regards two Buddhas dated in the 300s series, the one dated "year 318" in the Indian Museum (Ingholt and Lyons, Gandhāran Art in Pakistan, fig. II.1) may be related to Type III, but is of much degenerated quality; although the lack of low-ridge folds is notable in its drapery as in Type V, this image looks much heavier than any examples of Type V. The current whereabouts of the other one, dated "year 384" (Ingholt and Lyons, Gandhāran Art in Pakistan, fig. II.2), is unknown; in the single extant photograph, it most likely belongs to Type III; but that one photo is much blurred, making it hard to tell anything more specific. And of course the dates of the era to which the images belonged remain problematic. It seems a reasonable guess, but still only a guess, that the two images are datable somewhere in the third century CE.
101. My preliminary research shows that the majority of extant bodhisattva statues from Gandhāra correspond to Types III, IV, and V, with very few in Type II and almost none in Type I. But this requires corroboration in a separate paper after more extensive and meticulous visual examination. [End Page 85]