Jaff Rugs
Diamonds in the Pile   Jaff Kurd Bagfaces

by Mark Hopkins

I never gave them much serious thought until the day a respected rug dealer I knew made a wry face and said, "Why are you collecting those things?" Had I, I began wondering, made a serious mistake? In those days they were just called Kurd bagfaces, or "diamond doormats." People didn't pay much attention to the occasional book that attributed them to a tribal group called Jaff Kurds. They weren't important enough; nobody really cared. But I did. For one thing, there were a lot of them around. Most were pretty ugly, but there were great ones too. I could afford them. And I liked them -- a lot. I'm into other kinds of weavings now, but several things I can attest. They're intriguing from a structural standpoint because they use a peculiar pile knotting technique seen in few other weavingsi They just may be showing us a bridge to one of the oldest existing design motifs in the cloudy history of pile weaving. And perhaps most important, their aesthetic merits wear most comfortably over the years. I saw my first Jaff bagface years ago. It had been bought by a friend who at the time, like me, hadn't the slightest interest in Oriental rugs. But she had a certain bare spot on the parquet floor that her cocker spaniel favored. The little rug (we'd never heard of bagfaces) had been the perfect buy. It had just the right colors. And it was cheap. There was something else about it that riveted my attention. It emanated a sort of gentle, elusive air of mystery, a sense of shepherds and dunes and knives glinting in moonlight. Something that quietly swept me beyond visions of dark-eyebrowed merchants in polyester suits to a land of woodsmoke and dust whose horizons had serene caravans etched against their sunsets. This was the first Oriental rug I had ever seen that was clearly intended not for a parquet floor but for a camel. (It also smelled like a camel, but that had more to do with the cocker spaniel than its mystic origins.) Whatever it was about it, it had me quite captured. I didn't come to grips with the fact for many years, but I was hooked. What finalized the hooking, I think, was my eventual discovery that the aesthetics of these little pieces can be very, very rewarding. Granted that flies in the face of many rug lovers who consider them quite beneath their notice. They shouldn't. Bear with me through a slight digression while I explain that point.

The Line Between Boring and Beautiful

The problem is, what we're dealing with here is weavings with repetitive designs. And they, depending on how they are executed, can be very, very boring...or extremely wonderful to behold. One successful way to categorize rugs is by what happens in their fields. They quickly divide into two groups. One has a central design focus - medallions or other interesting things framed by borders. The other group uses the border to window what is essentially a small section of a bigger repetitive design. Numerous rug authors speak of this as a sort of "Islamic window on infinity"; perhaps one day this observation will be expanded in scholarly terms. Boring though they potentially may be, repetitive designs are a principal mainstay in the rug weaver's lexicon. They grace rugs from looms as far spaced as Izmir and Beijing. And the good ones are not boring because their weavers have ways of making them not boring. Some tribal groups admired for their repetitive motifs - Turkoman, for example - achieve artistic merit mainly by dint of clear drawing, rich dyes, expert use of weaving materials, and painstaking regularity of detail. Their focus is on symmetry of form and color. But there is usually a tile-floor-like sameness to their motifs. It is the presentation that counts. Other rug groupings respected for their successful repetitive designs -- Kurds, Afshars, and most of the Caucasian categories, for instance - -travel a very different road. Symmetry of form remains. But in most cases symmetry of color is banished. In these categories of weavings, the principal artistic merit lies in the "random" juxtaposition of identical design elements appearing in widely varying color combinations. We're talking here about Caucasian and Anatolian rugs with rows of memling gul motifs, Shahsavan weavings, Afshar rugs with repetitive boteh variants... that sort of thing. And, of course, not coincidentally, Jaff Kurd bagfaces. The quotes around the word random are very important. Because while the outwardly casual appearance of one of these asymmetrical weavings implies that the color combinations of its elements were quite casually conceived, in fact the opposite is true. The assemblage, at least in the good ones, has nothing random about it. It is painstakingly planned. In the best instances, it is pure artistry.

Take an example. Figure 1 is the pile face of a large Jaff khorjin measuring nearly 3x4 feet. The field design is typical, but the effect is quite magical. If these diamonds were ceramic tiles, there would be a graphic rigidity that here gives way to flowing, undulating lines. It is a field of diamonds in motion. No two places where the glance can touch are quite the same. It is a design to which the eye (at least my eye) can return again and again without tiring.

How Many Ways to Color a Diamond

What kind of artistry is involved here? To answer that, try a simple test. We'll supply you with seven crayons (that's all the colors there are in Figure l's 38 full diamonds: red, two blues, pale green, brown/aubergine, plus ivory and peach as accent colors). Then we'll give you a line drawing of a Jaff diamond field that outlines the shapes but is left empty of color. Your job: start coloring. There are three rules: 1. No diamond may butt up against another of the same color. 2. The results must be well balanced, interesting to look at, and aesthetically pleasing. 3. (And this is the hard part) You have to color from the bottom up with no skipping ahead... just the way the weaver would do it.

Believe me, it's not easy. In fact, it's one of those challenges that today we tend to resolve by buying the right software. Is there a formula passed from weaver to weaver? I've looked at hundreds of these weavings and have never seen a suggestion of random patterns emerging. No doubt there were unwritten rules of some kind passed down through generations. But a lot must have also evolved from each weaver's own intuitive sense of balance and design.

Sometimes it helps to compare the good with the not so good. Figure 4 is an example of a Jaff bag whose field design doesn't quite make it. It is a fine old piece, but there is a clumsy diagonal of predominantly red diamonds in its field that grinds any sense of motion to a halt. There is additional merit to the piece shown in Figure 4, by the way. As one of the few complete individual Jaff bags I have encountered, it suggests that some of the flatwoven backs missing from most all Jaff bagfaces may have been quite interesting in their own right. Complete double bags, which is the format for which most if not all extant Jaff bagfaces were originally intended, appear on the market even more rarely.1 Where do these weavings come from? Until a few years ago ithe stock answer was northwest Iran. That perception began to change with the publication of an article by William Eagleton in the catalog2 of a Kurdish weaving exhibit held at Northwestern University in early 1984. In it he suggested that "Iraqi Kurdistan may be the primary source of Jaff Kurd weavings." This is reinforced in his important new book, An Introduction to Kurdish Rugs and Other Weavings.

The Jaff Kurds are among the southernmost dwelling of the Kurdish people. They are a large tribe, inhabiting an area located at the extreme lower tip of that huge, triangular region that nips corners of Turkey, Syria, Iraq, and Iran called Kurdistan -- home of the Kurds. The Jaffs bridge the Iraq/Iran border, with their main concentration located in Iraq southeast of Kirkuk. On the Iranian side, where their numbers are smaller, they live in a large area northwest of Kermanshah and are called Javanrud Jaff. The Jaff are Sunni Moslems, and their numbers are now several hundred thousand strong. Once nomadic, they have more recently settled into a predominantly agricultural way of life. During their nomadic days, however, they were one of the recognized Kurdish weaving tribes.3 The Jaff bagfaces we see today were once the decorative pile faces of double saddlebags imported into Western markets by the rug trade. The repetitive Jaff diamond design itself is, of course, not necessarily limited to these. It also appears in coarse, shaggy rugs made by Jaff weavers4 and in other tribal accessories such as salt bags.5 And, as we shall mention later on, it is also found in the weavings of several other major tribal groups.

A Hit in the Roaring Twenties

Most of the Jaff bagfaces one uncovers in the marketplace today probably arrived in a flood of imports during the 1920s and 1930s. Charles W. Jacobsen, the New York State dealer whose once-definitive book, Oriental Rugs: A Complete Guide, was first published in 1962, wrote: "Until about 1936 thousands of (Jaff bagfaces) could be bought in New York or the European market. After removing the back and making them into small rugs by fringing the ends, they were a delightful and exquisite small rug.... I myself have sold tens of thousands of these, and yet practically none appear in used rug sales today. The reason is that the heirs do not include these but keep them for their own use. Most of them were placed in doorways and heavy traffic eliminated them. None are to be bought in Iran or any market today (ca. 1960)."6 The demise of the Jaff weaving industry is also reflected in the writings of the esteemed English dealer Cecil Edwards who wrote in his 1953 book, The Persian Carpet: "A few... Jafffi rugs still come into the market. But the output is certainly not more than a tenth of what it was in the early thirties."7 This isn't meant to imply that there are no older Jaff pieces to be found. The rush to export that Jacobsen documents surely swept up a lot of older pieces along with the newly woven ones. William Eagleton has published a Jaff bagface that was old when his grandfather purchased it in the 1920s.8 This suggests that older ones will continue to be found, although the hard fact that utilitarian weavings are usually doomed to short if eventful lives certainly diminishes the odds. The reasons why the flow of Jaff bagfaces into the marketplace gradually dried up are the same as those shared by most extinct tribal weavings. Modernization of travel replaced hooves with wheels, eliminating the need for many bags. The region's nomads were gradually settling into a village existence with no further need for the trappings of a traveling life. And the final death knell sounded when the price of wool headed skyward, allowing families to sell their raw harvest for as much as a weaving would bring. The good news, however, is that some of these Jaff pieces have somehow eluded the ravages of hard years and gritty overshoes. From time to time, a few wonderful ones find their way back to market, and they still constitute an easily collectible category.

Getting in Under the Pile

One of the more interesting parts of the Jaff bag story lies in the way each piece is woven. First, let's establish some basic technical parameters. Jaff bagfaces normally range in size between 2x2 and 4x3 feet. Their foundations are almost always wool. Warps are normally two-plied or three-plied ivory wool, though occasionally brown and ivory plies will be twisted together to create a candy stripe look. Their wefts are usually two-plied yarn. Colors will range from ivory through all shades of natural brown, and will include dyed colors in various red, blue, and red-brown shades. Undyed goat hair is also sometimes used. Some pieces will use several different weft colors in a sort of "grab-bag" way. Edges are almost always multiple warp cords overcast with a single color of yarn. Knots are always symmetrical (Turkish). The structure is flat, without any depression. Usually there are two weft shoots between the knot rows, although there are also bags with occasional weftings of four or more shoots. I have only encountered one bagface with single wefting, and it was a coarse and uninteresting piece of fairly recent origin. Considering the number of these pieces that were exported, it is conceivable that weavers of other tribal origins copied the design for commercial gain, which may have been the case here. Beyond these routine observations, however, there is a very interesting technique that sets these Kurdish pieces apart from pile weavings of most other provenances. In order to achieve the smooth diagonal sequences of their diamond designs, the knots in their fields are invariably tied on alternating warps instead of being stacked in vertical rows as in other weavings (Figure 7).

Like most tribal weavings, these Jaff bagfaces are coarsely knotted pieces. Most have between 4 1/2 and 6 1/2 knots to the horizontal inch and 9 to 16 rows to the vertical inch, creating a range from 45 to about 100 knots per square inch. So this special technique greatly facilitates the weaver's ability to "draw" the diamonds' diagonal structures. In Figure 5 its effectiveness can be compared with a conventionally woven Kurd-Baluch rug.

However, the sides of each Jaff Kurd field are framed by the vertical rules of its borders, in which the knots must be lined up vertically in order to create straight lines. So where does the weaver make the transition from stacked to alternating without skipping a warp? The answer is, a lot of times they simply do skip a warp (Figure 8a). On many weavings the pile is long and the structure fairly "meaty," making the practice of leaving an occasional warp uncovered a virtually unnoticeable one on the piled side. Even on the back it takes a careful look to find the bare spots.

The Jaff weaver also practices three other options. One is to cover the single exposed warp with a sort of Spanish knot (Figure 8d). This is normally used at the corners of figures where a single tuft of color is needed. Another is to bridge the gap by wrapping one side of the knot around two warps (Figures 8c), a sort of semi-jufti configuration. And a third (Figure 8b) is to tie an improvised knot that starts as a conventional Turkish knot but continues with an extra turn around the third warp. Usually two or more of these techniques will be used together. The alternating warp method is not exclusive to the rugs and accessories of the Jaff Kurds, although to my knowledge no other category uses it with consistency. I have also seen it in a very old Anatolian Kurdish rug attributed to the north-central Turkish town of Sarkisla.9 One restorer I know has encountered it in Chodor weavings. Another mentions seeing it often in old Chinese pile weavings, where the technique is used improvisationally to execute difficult details in these coarsely knotted pieces. He also found it recently in a Melas prayer rug, where it was used to achieve diagonal details in the mihrab. Beyond these informal observations, however, any useful comments on a broader prevalence of this interesting weaving technique will have to come from those more widely versed than I.

More Answers Are Still to Come

There is, in fact, much more to be learned about this modest classification of pile weavings. Are they, for instance, entirely the product of the Jaff Kurds? The catalog, Discoveries From Kurdish Looms, suggests that the answer is no. It notes that a supplemental weft patterning along the bottom of one of the exhibition's diamond-decorated bags is unusual in a Jaff piece, and postulates a possible Turkish Kurdish origin.10 Since the Turkish border lies more than 200 miles north of the Jaff heartland, and many other Kurdish weaving tribes dwell in between, it may well be that other groups have adapted this distinctive design into their pile weavings as well. Or perhaps it has belonged to them as long as the Jaffs have used it. Beyond that, there are other hooked-diamond Kurdish pile weavings that don't use the randomly juxtaposed diamond motif at all. Instead, some feature a single, multi-tiered diamond as a central medallion11 Others maintain the repetitive theme but organize the diamonds into completely symmetrical patterns.12 Are these Jaff or are other related Kurdish tribes responsible? There are many questions that remain. There is also the need to fine tune the classification of Jaff weavings themselves. The photographs on these pages show several different versions of the basic Jaff design. And there are more. Differentiation should probably be made on the basis of color combinations, major/minor border designs, and decorative end treatments. Size and weaving quality also appear to be factors. Most importantly, any future; classification will need access to first-hand knowledge of the various Kurdish weaving groups in order to assign specific provenance and give the subdivisions meaning, Finally, there is the intriguing question of design origin. The individual hooked diamond itself is clearly the legacy of an ancient design tradition. In fact, its enormous prevalence in Eastern tribal art suggests it may be one of the oldest basic motifs. It can be found in the weavings of all of the major tribal groups of South Persia, and in yüncü and other pile pieces from Anatolia. It appears everywhere in the output of the Caucasus. In its most basic form (Figures 10a and 10b), it can be found in Anatolian kilims, Luri rugs, Shahsavan mafrash, and pile rugs and flatweaves of the Kurdish and Baluch tribes of Khorasan, just to name a few. Beyond that, its reach emanates for thousands of miles. It appears, for instance, in flatweaves of the isolated Himalayan kingdom of Bhutan.13 It is seen as far away as the weavings of Timor,14 one of the Indonesian archipelago's easternmost islands. Current evidence suggests that it must have moved along ancient trade and migratory routes over several millenia to arrive where it is today. Where it came from, nobody knows. There are, of course, various theories. Max Allen, for example, proposes it is one of many versions of a birth symbol dating to the earliest days of civilization.15 James Opie traces it back to 2,500-year-old bronze totemic animal-headed figures unearthed in the Luristan region of Iran.16 (Surely the resemblance of the hooks to animal or bird heads is hard to miss.) What appears to make these little Jaff khorgins distinctive is their highly consistent use of repetitively spaced, randomly colored hooked diamonds, framed by traditional Kurdish border designs, in a pile format. The same motif does indeed appear in other container-type weavings -- most frequently in Khurasan Kurd and, to a lesser extent, Shahsavan bags. But as far as is known, these are virtually always flatwoven.

An Ancient Tradition of Diamonds

While history shows little inclination to reveal any conclusive evidence of the hooked diamond's origins, the use of this ubiquitous figure in repetitive, latticed weaving designs does have some historic visibility. There is a curious similarity, for example, between the repetitive Jaff Kurd diamond motif of today and the field design of an ancient Seljuk carpet fragment discovered in a mosque in Beysehir, Turkey. Dated by contemporary scholars to the l3th century, the piece is a fine example of the symmetrical, tile-like designs that characterize Seljuk, weavings. Figure 6 shows the motif; the actual colors are typical of the provenance: a medium blue diamond figure with red highlights on a deep blue field.17

Granted that Beysehir and the nearby ancient Seljuk capital city of Konya are separated from the land of the Jaffs by a good 800 miles, but the history of Anatolia is a history of conquering armies sweeping in from East and West, and the records document that Kurdistan was hardly left untouched by the Seljuks' passing. It is conceivable that the diamond motif was adapted by the Jaff or their predecessors nearly a millenium ago. It has also been suggested that the motif shown in Figure 6 may have already been an ancient one when it arrived in Anatolia. The Turkish scholar Oktay Aslanapa postulates that it may have originated as a Buddhist symbol brought West by the Uighurs.18 And, of course, there is always the possibility that the Kurds, or their predecessors, originated the design themselves. There is another old weaving whose design bears a striking resemblance to the diamond motif. Some 50 years ago the Swedish scholar C.J. Lamm uncovered a treasure trove of pile fragments in Fostat, Egypt. Two of them, subsequently dated to the 9th century, use patterns very suggestive of the hooked diamond, one in its field and the other in its main border. Addressing the issue of original origin, Lamm suggests: "The designs of these fragments are, undoubtedly, more closely connected with Iraqi (and, ultimately, Sasanian) than with Egyptian art."19 It is also interesting to note that one of these two fragments uses the alternating warp knot technique.20 Even the border designs of some Jaff bagfaces can sometimes closely resemble those of older provenances. For example, the outer border in Figure 1, a very common one, bears a striking similarity to an old Anatolian (Kurdish?) design used in the main border of a 16th/17th century Central Anatolian rug that now hangs in the Vakiflar Museum in Istanbul.21 Clearly these connections are all speculative, wanting much further thought and research. But the Kurds are an ancient people; it is said they are descendants of the Medes who inhabited eastern Kurdistan some 3,000 years ago.22 The nomadic tribes among them most likely stayed relatively isolated from the changing styles of the dynasties that swirled around them over the centuries. It could well be that the familiar hooked diamond design we call Jaff may bridge a span of several millenia in the lexicon of Middle Eastern tribal weavings. You won't find them in the great collections. They may not be as glitzy as Marasali prayer rugs, or as bumptious as Anatolian yataks, or as intricately subtle as Afshar carpets, but these little bagfaces have a special charm all their own. Unlike most weavings in the market today, they are truly tribal; until early in this century there is no indication they were touched by commercial demand. And they certainly yielded few if any changes when it did happen. Like their Shahsavan neighbors to the Northeast, the Jaff Kurds clearly took great pleasure in nurturing gentle, inviting designs that befriend the beholding eye. The few great examples of their weavings that remain should quickly be rescued from the kennels and doorways of the Western world and accorded the recognition they have long deserved.

Fig. 1. Large Jaff Kurd bagface, 3'11"x2'10". The diamonds in the field of this striking piece use only five main colors and two additional accent colors (not including the dark brown that separates them). The weaving's 38 full diamonds are presented in 20 different color combinations to achieve a highly successful polychromatic effect.

Fig. 2. Jaff Kurd bagface, 2'3"x2'1". This striking old piece is a fine example of how ingenious color juxtaposition can create interesting design. The border alone has 24 different combinations, all accomplished with eight colors plus dark brown. The decreasing angle of the diamonds toward the top was caused by reducing the thickness of the weft shoots, which are primarily brown goat hair. Note that the motif within the diamonds is atypical.

Fig. 3. Jaff Kurd bagface, 3'7"x2'1". Both the shape and the fine weave of this piece are unusual. The use of a centered ivory diamond is also interesting. With a limited palette of six colors, not including the delineation brown, the weaver has achieved its elegant design by artfully interspersing nine distinct color combinations. Exceptionally soft, glossy wool and blue-dyed wefts add to its interest.

Figure 4. Jaff Kurd bag, 2'x2'. It is rare to find a fine old Jaff piece with asurviving back panel, especially one as interesting as this.

The back of Figure 4. The geometric design on the back is achieved using a slit tapestry weave embellished with colorful extra-weft brocading at the top

Fig. 9. The Jaff Kurd homeland

Fig. 7. While the fields of Jaff Kurd weavings use the alternating warp method, the borders are woven conventionally. This shows how the transition is made.

Fig. 5. A comparison of two pile weavings, viewed from the back, illustrates the advantage of the alternating warp technique. The small Baluch rug (Fig. 5a.) is conventionally woven with a symmetrical knot, which gives its diagonal lines a stepped appearance. The Jaff weaving (Fig. 5b.) Also shown in Fig. 3, achieves cleaner, more effective diagonals by using the alternating warp technique.

Figure 6. The Kurdish diamond motif bears an interesting resemblance to the tile-like field of a 13th century Seljuk carpet fragment now in the Museum of Turkish and Islamic Arts in Istanbul.

Fig. 11. Although conventional in design, this piece is exceptional for its extensive palette of 11 soft colors. Surrounding the centralized (but not centered) ivory diamond are 32 other diamonds appearing in 27 color combinations. The chevroned skirt is not uncommon


1. For an example, see Reinisch, H., Satteltaschen, Graz, Austria, 1985, pl. 33.

2. Briggs, R.D., Discoveries From Kurdish Looms, Northwestern University, Evanston, III., 1983, pp. 36-43.

3. Eagleton, W., An Introduction to Kurdish Rugs and Other Weavings, New York, 1988, p. 21 and pp. 26-27.

4. For examples, see Housego, J., Tribal Rugs, London, 1978, pl. 61, and Eiland, M.L., Oriental Rugs, A New Comprehensive Guide, Boston, 1981, p. 10 (b).

5. For an example, see Herrmann, E., Seltene Orientteppiche VIII, pl. 86a.

6. Jacobsen, C.W., Oriental Rugs, A Complete Guide, Rutland, Vt., and Tokyo, Japan, 1980, p. 250.

7. Edwards, A.C., The Persian Carpet, London, I953, p. 126.

8. Eagleton, op. cit., pl. 16 and p. 62.

9. From the author's collection; for a similar piece, see MacLean, J.A. and Blair, D., Ballard Collection of Oriental Rugs, Herron Art Institute, 1924, pl. 45.

10. Biggs, op. cit., p. 88.

11. For an example, see Eagleton, op. cit., pl. 40.

12. For examples, see Hali, XI, No. 1, p. 80, and Benardout, R., Exhibition Catalog, 1979, Nos. 41 and 43.

13. Myers, D.K., "Costume and Ceremonial Textiles of Bhutan," The Texrtile Museum Journal 1987, Washington, D.C., p. 30.

14. Allen, M., The Birth Symbol in Traditional Women's Art, Toronto, Museum for Textiles, 1981, pp. 66-67.

15. Ibid. 16. Opie, J., "The Animal Head Design in Lori/Bakhtiyari Weavings," Hali, V, No. 4, pp. 450-461.

17. Yetkin, S., Historical Turkish Carpets, Istanbul, 1981, pp. 22-23, and pl. 8.

18 Aslanapa, O., One Thousand Years of Turkish Carpets, Istanbul, 1988, p. 26.

19. Lamm, C.J., Carpet Fragments, Sweden, 1985, pl.1 and p. 13.

20 . Ibid., p. 35.

2l. Balpinar, B. and Hirsch, U., Carpets of the Vakiflar Museum, Istanbul, Wesel, West Germany, 1988, pl. 17.

22. Eagleton, op. cit., pp. 9-10.

Jaff Tribe

By Dr.Sarwat Jaff
[email protected] [email protected]