Tschebull Antique Carpets  

Publications by Raoul Tschebull

Hali 1.3, 1978, The Development of Four Kazak Designs

(Please note that this oft-cited article is edited to eliminate the original picture references and content remains basically unchanged. Although more than twenty years old, the article is not as dated as I first feared it was.) RET
Many south Caucasian village and nomad rug designs are combinations of field and border patterns which do not reflect a long tradition. Up until at least 1920, however, the Kazak weaving area produced four common designs which closely followed a convention which dictated their general size, color, field pattern and borders. This convention tended to erode toward the end of the period and the development of the style was not always chronological.

These four patterns are best known by their inappropriate, but understandable, trade names Karachov, Triple Medallion, Sewan (the cruciform medallion) and Lori Pambak. In early times, they almost always appeared on rugs approximately 5 ft 5 in x 7 ft 5 in (1.65 x 2.26m), perhaps reflecting a traditional [cult?] use. All four patterns have probable pre-Islamic totemic sources which they share with the Turkoman göl. Over the period from which examples are available, the first three were, and basically remained, stable geometric designs, possibly because they were ancient adaptations of Central Asian totemic themes. The last type, however, changed substantially over a documented period up until about 1920, starting as a totemic göl form and ending as a geometric design. This last group also lost some of its original color range and spatial relationships.

The cruciform pattern is the one furthest removed from a central Asian totemic tradition, and, indeed, Joseph McMullan suggested that a rounded version of the medallion represented a water-basin. Moreover, there are no rugs known to me that show this medallion containing multiple, Turkoman-style animal forms; perhaps an example exists as a fragment in a museum storeroom. The angular shape does not change much between the versions of the early 19th and early 20th centuries, even down to the four tree forms. However, earlier pieces usually have a border composed of bug-like devices.

The rug shown in 'Kazak', 1971, plate 27, which is very similar to an example dated 1828, is very finely and loosely knotted, with long pile and a large number of colors including two shades of red. The guard borders are atypical of later pieces in this design. There are two clearly drawn animals in the medallion center, which probably are the remains of Turkoman animal forms mentioned above. Documented by dates (1902, 1909), later angular cruciforms are usually one third coarser, with shorter pile and uniform ivory warps. I know of only one very small angular cruciform rug. Since the rugs in the later group are very similar in color and construction, it is not unreasonable to suggest that they were made in a single area but whether both the old and new styles are from the same area is unknown.

At least two quite different rounded cruciform types are to be seen on rugs which are either shaggy and coarse or fairly fine, with tightly plied gray and brown warps, blue wefts and extremely high gloss wool (See 'Kazak', plate 29). The McMullan rug (Plate 49, 'Islamic Carpets', 1965) is clearly dated AH 1212 (1797 A.D.) and is in terrible condition. The similarities between these last two rugs include the number of elements in their borders, their sizes and their structures. Their colors, however, vary; the first rug has a green and a light purple beige, the latter obtained from iron-madder dye. The McMullan rug has the normal blue and a dark purple. The first rug is in much better condition than the McMullan example and its close similarity to the latter has led to it, too, being considered early. However, chemical analysis of the red in the tiny triangles in the centers of the upper hooked diamonds showed it to be a combination of madder and the Russian synthetic dye known as Ponceau 2R; this was not available before 1880 and, to my knowledge, its first documented appearance in a south Caucasian rug is in a Karabagh dated 1886.

The question is whether these two cruciform Kazaks reflect an extreme case of stability in design and construction in some south Caucasian villages over a 100 year period or whether the McMullan example is newer than its date indicates. I know other rugs from the same area which are of similar color and structure to the McMullan example, but with different designs. However, we should not ignore the dark purple color in the McMullan rug which seems to appear mainly in older Kazaks.

The development of the Triple Medallion Kazak pattern (see 'Kazak', plate 5) presents a number of problems. As far as I am aware, the pattern does not appear in western Turkey and is rarely found on dated pieces; the rugs upon which it does appear show no clear structural trends, other than a tendency towards larger sizes later in the 19th century. In fact, rugs with a large variety of structures and colors are found with this pattern; also, the gül centers are variable but a repeat of two blues and a white, or vice versa, on a red field is the norm.

The number of variations suggests that Triple Medallion rugs were knotted all over the south-west Caucasus. One example has the apparently early form with the typical accompanying border; another is a larger, later, piece but with native dyes and, judging by the wool quality and colors, probably from the Erivan area. A rug with the same design was illustrated by Mumford in 1900 and another - more finely knotted and with what seem to be vestigial animal forms in the white medallions - is from yet another area. Another coarse piece, with its complex white 'main' medallions, appears old and has an unreadable date inscription. It is difficult to pinpoint the difference between what may be early and late pieces; dyes change, borders become wider and there are changes in other spatial relationships. The Triple Medallion format made it easy for the weaver to design rugs of different lengths, yet there are few examples with other than three medallions and most of these exceptions have the hall-marks of 1910 weaving.

The so-called Karachov pattern was probably the most widely produced of the four Kazak designs, and comes in a great variety of weaves, colors and sizes. Again, the elements and their inter-relationships remain stable but the medallion contents change. Karachovs are not usually long format rugs and often they are dated, although usually not particularly early. A 'proto-Karachov' (McMullan, plate 98) from western Turkey shows four pairs of red stylized animals in the central octagonal medallion, surrounded by four minor medallions edged with typical Turkoman-style kotshak forms. Most probably it had a south Caucasian contemporary, but no example is known. The earliest Karachovs I know, like other early Kazaks, are finely knotted and not very large. They have either a red or green field, and a dark purple color is almost always present.

In a rug dated 1860, the kotshaks in the subsidiary medallions have changed to a few hook-forms and the border is less complex. The tendency to change both the design and the format at the end of the 19th century is shown in a coarse rug 5 ft (152 cm) square, which is dated 1902, and by another rug dated 1911. The latter has a substantially different minor medallion, one which is often seen on examples with synthetic dyes.

The 'Lori Pambak' pattern [may] be indigenous to the south-west Caucasus. Of the four common types of Kazak main carpets, it shows the most radical change in design and structure over the period from which examples are available. A considerable number of dated examples exist.

The green form in the main medallion has been called a 'cross-flower'. However, early examples show the central medallion as an ancestral Turkoman göl, with two palmettes above and below. The main medallions contain four opposed pairs of green animals, all those in one example having crests and eyes, but only the upper and lower pair in a second having eyes and attached crests. The comparison shows how the later version of the design came to be known as the 'cross-flower'.

The second rug is dated 1846 and is much coarser than the first one. It also has a more limited color range, with a much lighter purple. There are other significant differences. The first piece has a quartered medallion and on the left and right side appears a dark brown double headed Turkoman-style crested bird, which contrasts with the saw-edged pair of triangles in the dated rug. However, the borders on both rugs are quite similar and archetypal and they both have a large number of amulet-like forms. The yellow motifs on each side of the green palmette persist on rugs of this pattern made at the turn of the century but the other amulets, especially the wurma form on the first rug, disappear in later examples.

Both of these old rugs have their original selvedges which are attached in a similar fashion. The style of knotting and the wool quality are also alike, and although the colors differ, they are used in a similar manner. I think that the first rug, with its finer knotting and more marked animal symbolism, is the older of the two, at least in style, but how much older is a question which will have to await more sophisticated chemical tests.

Both rugs probably come from the same as yet unidentified village in the Transcaucasus. Ralph Yohe was in Kars, Turkey, in 1974 and took photographs of similar rugs through the window of a locked mosque. These pieces, made after 1900 (one is dated 1914), contain a faded green and the traditional border has been abandoned. Whether rugs of this style are only from around Kars or were also made in other areas is a question the answer to which will require visits to many of the mosques in eastern Turkey.

Following the early and intermediate (1846) styles of this pattern, two more rugs show the next two stages of its development in pieces dated 1897 and 1911. The main borders are traditional but their relationship to the guard borders has changed. In one rug, the medallion is strongly geometric and its elements have fused; it has lost its Turkoman shape and has grown kotshaks although the bird forms at top and bottom still have crude crests. The flanking medallions have retained a palmette shape and the yellow amulets are still present. On the second rug, the palmette shape has been lost and the animals show a greater stylization. Both rugs are as coarse as the late cruciform examples.

The next development is shown by a rug which has lost its main carpet pattern completely, although it is still of large size; the green dye is probably synthetic. Another pair of rugs are small, measuring approximately 5 ft x 4 ft (152 x 1.22 m); they both have the elements of a main 'bug' medallion and are attractive examples of a type which frequently has synthetic dyes, something which indicates a late date for the group as a whole.

Why did the 'Lori Pambak' pattern undergo the drastic changes I have shown while the other Kazak main carpet patterns remained relatively stable? Possibly the 'Lori Pambak' design was brought to the south Caucasus by a late migration of Turkoman nomads, a group which did not penetrate further west.

In general, little is known about the village and nomad rugs of the south Caucasus. Chemical tests, dates on rugs, accession dates in museums and observations made over the years all help to answer some of the questions. But many problems still remain; for example, there is a group of rugs [some with Lori Pamback medallions] that have a structure and color range I know well, and are, perhaps, from the same weaver and dyer. Their distinguishing characteristics include their [predominantly small] size and the use of a dark red and a chalk white. Hopefully, a dated rug of this type will appear, or some other hard data will be found. Kazak main carpet weaving was very conservative but the design [sources are likely more varied that we can imagine.] Much ethnographic research remains to be reviewed before we can answer questions confidently about the development of designs, the use of color, the [impact of commercialization on rug formats] and the effects of western civilization on the south Caucasus.


Tschebull Antique Carpets
903 Post Rd.
Darien, CT 06820