Collectors (all kinds, not only rug collectors) like to be able to attach reliable labels to their treasures. Because of this, we find ease of attribution to be important and are much more often attracted to items we can name than to pieces we can't identify with certainty. Art collectors have an advantage that greatly simplifies attribution: paintings, prints and sculptures often include the artist's signature. With very few exceptions, rugs are not signed and even the most gifted weavers remain anonymous permanently. Structural and design similarities can be interpreted as evidence that two rugs were made in the same area at about the same time, but there are no well accepted criteria for attributing two pieces to the same weaver. Still, it is usually not very difficult to determine where a rug was made.
| Figure 1. Detail, Karabagh rug, 3'5" x 9'0". This is the end of the rug that was woven last and illustrates how the weaver has made the next-to-last stepped diamond in the main border smaller than the others in order to achieve smooth transitions in the border corners. |
Many of the rugs from the Talish district are runners about 3'6" x 9'6", with a characteristic wide main border having a white background, and a narrow field, often blue, that sometimes includes only the abrash as decoration (the met-hane design). In those with more than a few small, isolated motifs in it, the field is usually nearly filled by an overall repeating design (for example, rosettes or botehs).
Those from the district capital, Lenkoran, are unlike other Talish rugs. Their borders are sometimes quite narrow, and the field is generally dominated by what are known as Lenkoran medallions. Most often there are three, and there can be as many as five medallions in runners, one or two in smaller pieces. The medallions are so typical of Lenkorans that they almost define the type. Nearly any rug on which they appear would be called a Lenkoran, and few rugs that would be attributed to Lenkoran lack them. Of course, this may simply be an error perpetuated by its repetition. An example of a published rug being attributed to Lenkoran although lacking such medallions is Bennett's Plate 189. The attribution is based on the layout, borders and dimensions being virtually identical to those of another (Plate 1901) featuring three typical Lenkoran medallions, but with the medallions in Plate 1901 replaced by St. Andrew's crosses in Plate 1891. Bennett notes that the similarities between the two are sufficient to make it likely that both were made in the same place at about the same time. The medallions on Plate 1901 indicate that place to be Lenkoran. In fact, some Karabagh rugs have enough in common with Bennett's Plates 189 and 190 to invite speculation about their relation to the Lenkorans.
Figure 1 shows one end of a rug that I believe can be confidently attributed to Karabagh. It is nearly identical in size, overall design, and some very peculiar details to a Karabagh belonging to the Armenian National Museum, published as Plate 93 in Stepanian. There are so many similarities between the rug in Figure 1 and that in Stepanian2 that they were almost surely made by the same weaver. Both have the same horizontal and vertical knot counts and both have fragile warps near the edges, manifested as numerous slits along the sides. But the most striking evidence of a common origin is the solution to the problem of carrying the main border design around the corners, which is resolved in the same unusual manner in both. By making the next-to-last stepped diamond in the main border smaller than the others, the weaver neatly changes the border pattern at the corners from vertical to horizontal. Discontinuity of the border pattern at the corners, especially at the end woven last, is practically a hallmark of village rugs. When two pieces of nearly identical design avoid this in identical ways, the likelihood that this occurred by happenstance seems pretty remote.
The only significant difference between the rug in Figure 1 and the one in Stepanian2 is that the latter has some trees and small stepped diamonds along the edges of the field. It is attributed to Shusha, the Karabagh district capital and listed as having been made in 1886, although no date is woven into it. Except for a few that have dates inscribed, it is the only one in Stepanian2 to which a specific year is assigned. Indeed, I know of no other published antique rug lacking an inscribed date to which any author specifies a specific year of production. Therefore, I assume that Stepanian (or the Armenian National Museum) had detailed, reliable information about this rug's provenance. The late Liatif Kerimov, who edited the book that contains the chapter by Stepanian2, was born in Shusha in 1906. His mother was a weaver; perhaps he knew other weavers in the city and had learned the history of the piece. The one illustrated by Stepanian2 has a few more motifs than does Figure 1. Weavers generally added small modifications with successive rugs. If that is the case here, the presence of the trees and the row of small stepped diamonds in the piece illustrated in Stepanian suggest that the rug in Figure 1 is the older of the two. Since the rug in Figure 1 includes what appears to be a synthetic orange dye, it cannot have been made much before 1870. I believe 1875-1885 is a reasonable estimate of when it was woven, although it may actually be more recent than the one illustrated in Stepanian2.
Very few published Caucasian rugs show borders that turn the corners without a discontinuity in the pattern. One is Bennett's Plate 1891. Another, Schurmann's Plate 53, is a Talish with a typical main border of large blossoms alternating with squares formed from four small motifs, the squares and blossoms widely separated on the white background. The wide spacing makes it relatively simple to carry the design around the corners neatly. But this is rare even on Talish rugs, and Schurmann notes the careful treatment of the corners in his commentary on this one.
I emphasized the basis for attributing Figure 1 to Karabagh because the design and dimensions would otherwise be consistent with it being from the Talish district or even being a Lenkoran without the usual medallions. In fact, all the borders are identical to those on the two that Bennett attributes to Lenkoran, on Schurmann's Plate 584, also a Lenkoran medallion runner and a Lenkoran recently sold by Skinner's (September 1992, Lot 164). Interestingly, in Bennett's Plate 1891 (the Lenkoran with the medallions replaced by St. Andrew's crosses) the problem of turning the corners in the main border is resolved in the same way as it is in the rugs shown in Figure 1 and in Stepanian's Plate 932. Another peculiarity common to all of the pieces under discussion (Figure 1, the two in Bennett, the one in Stepanian and the one in Schurmann) is that the edges and ends of the fields have lines of small motifs, blossom heads in some and stepped diamonds in others. These are nearly continuous and form what almost amount to additional borders within the fields.
These characteristics suggest common origins or, perhaps, a common ancestor for the two cited examples in Bennett's book, the one in Schurmann's, the one in Stepanian, and Figure 1 here. The similar idiosyncracies of the rug in Figure 1 and the one in Stepanian are so striking that I think it is safe to consider both to have been woven by the same person. Bennett's Plate 1891 has identical borders and makes use of the same creative solution to having the borders turn the corners neatly. It is tempting to speculate that it, too, was made by the same weaver. On the other hand, the field design is different, and is so closely related in layout to Bennett's Plate 1901 that it is hard to disagree with his suggestion that both were made in the same area at about the same time. Perhaps all four pieces have a common historical thread: a family of weavers, or weavers related as master and apprentice.
What about the Lenkoran medallion? If it rarely appears on anything except Lenkoran rugs, are there related motifs that do? The central elements in Lenkoran medallions are some variant of an elongated hexagon, often with hooked lines projecting from it. To this occidental eye, they suggest stylized insects. They have some similarity to the central element in the peculiar medallions in Figure 1 (and in Stepanian's Plate 932). Perhaps this should not be surprising; the similarities between the rug in Figure 1 and some Lenkorans has already been emphasized. On the other hand, I have found nothing else in 19th century Caucasian rugs that seems related to Lenkoran medallions, particularly the elements that form the lateral enclosures to their octagonal fields. Bennett suggests that they are derived from the dragons on late 16th to early 18th century Caucasian rugs. This seems reasonable, although the white background of these figures typically have motifs upon them that look far more dragon-like than do the figures themselves.
There is a group of motifs that resemble either the Lenkoran medallion outlines, central hexagonal elements, or both. It might be expected that these would be on weavings from the Caucasus, the areas closest to Lenkoran. Interestingly, they are not found on Caucasian weavings, but on certain products of nomadic Turkomans, geographically relatively remote from the Talish region, separated from it by the Caspian Sea. Some motifs bearing striking similarities to Lenkoran medallions occur in the lower elem panels on some Turkoman ensis. Pinner illustrates a number of them and shows that most are probably stylized birds. One, from a Yomud ensi, has the outline of the Lenkoran medallion's lateral enclosure. Several others can easily be imagined to be related to the hexagon that forms the Lenkoran medallion's center (the "insect legs" very clearly being stylized wing feathers in the Turkoman motifs). A significant difference between the Turkoman and Caucasian uses of the motifs is that the medallion in Lenkoran rugs is the dominant feature of the rug, while the related forms in Turkoman ensis are much smaller, appearing only as subsidiary motifs in end borders.
We are left with unanswered questions. If the lateral enclosure that gives the Lenkoran medallion its characteristic form is descended from the early Caucasian dragon motifs, why is it so rare in other rugs made in the nearby regions of the Caucasus? Is its hexagonal central element related to that of a medallion found in some Karabagh rugs that may represent a "school," perhaps a family, of weavers? What is the explanation of its presence as minor figures in ensis made by several Turkoman tribes, nomads in areas separated from Lenkoran by great distances and a substantial body of water, when it doesn't occur in places between Lenkoran and Turkmenia? Or does it, occasionally, in runners from Karabagh?
Figure 1. Detail, Karabagh rug, 3'5" x 9'0". This is the end of the rug that was woven last and illustrates how the weaver has made the next-to-last stepped diamond in the main border smaller than the others in order to achieve smooth transitions in the border corners.
1. Bennett, Ian, Oriental Rugs. Volume 1. Caucasian. Oriental Textile Press Ltd., 1981.
2. Stepanian, Nonna, "Carpets of Armenia". In, Kerimov, Liatif, Nonna Stepanian, Tatyana Grigoliya and David Tsitsishvili, Rugs & Carpets from the Caucasus, The Russian Collections, Penguin Books, 1984.
3. Hali, No. 61, February 1992, p. 159.
4. Schurmann, Ulrich, Caucasian Rugs, Old 99 Associates, 1990.
5. Lot 109 at Sotheby's (New York) sale of September 24, 1991, is a runner with similar borders and a field covered by diagonal rows of large botehs. It is attributed to Talish in the sale catalog.
6. This combination of borders is actually fairly common in Lenkoran rugs.
7. Pinner, R., "Salor ensis", Hali, No. 60, December 1991, pp. 86-96.
8. This has been commented upon briefly (Steven Price, Hali, No. 64, August 1992, p. 89).
