Question asked:

I noticed that the rugs you posted in the thread in which you gave the Harvard link and talked about religious influences are quite young. Is it your view that the "authentically Kurdish pieces" can include say early 20th century material?

I will simply answer YES because they "stay in the tradition" and will develop my thoughts here below.

First of all very few old, let’s say before 1880, Kurdish rugs are available as the market wasn’t organized to export great quantities of them and as they were produced almost for local use. By chance, I have recently examined an old Kurdish rug which I feel to be "authentic". I mean here, as the term was confusing, "being in the tradition".

OK, but what’s the "tradition"?

This small summary can help to give you a better idea of what I mean:

The Kurdish weaving tradition is:

1/ traditional large scale geometric design and few small scattered motifs to correspond with the height of the pile, inherited from more restrictive flatweave technique or borrowed from neighbor sources but adapted in their own way, or traducing old tradition like ancestral religious beliefs… look at the snake-like forms, symbol of the devil, drawn in the two Yezidis rugs, woven somewhere in the vicinity of Mosul, that I have showed in a previous thread.

2/ traditional weaving structure - to cite the most important:

In their pile rugs : a coarse weave without warp depression, an all 2ply wool construction with often goat hairs, and in some case cotton for white highlights or metallic wrapped cotton.

In their flatweaves; various kind of brocading (reciprocal, overlay-underlay), soumak and weftless soumak, slit tapestry weave and paired warps tapestry, warp faced and warp substitution weave.

In their end finishes: weft faced skirts adorned with rows of twinning, oblique interlacing or more distinctive oblique wrapping,

And last flat large selvages interlaced by the ground wefts and reinforced with additional colored wefts in bands of contrasting colors.

3/ traditional very saturated colors with a distinctive palette containing some colors cherished by Kurdish weavers like salmon, pink, orange, yellow and white outlining of the design.

4/ and last traditional long pile silky mountain wool, a heavy and floppy handle.

I am not allowed to publish the picture of this old Kurdish rug but it is closely related in design and structure (see below if it gets your interest) to one rug woven by the Goyan tribes, illustrated in Eagleton’s book (plate 72) , reproduced here below. Other examples also related to this rug are, a Goyan rug from Alberto Boralevi published in James H. Klingner article in Hali 105 page 84 and another Goyan rug is on display onto Cloudband in Alberto Levi’s gallery.

According to Eagleton, the Goyan tribe was probably once a section of the Hartushi, which is by far the most important tribe in the South East Turkey Hakkari region. They are located just above the Iraki frontier near the town of Uludere, situated midway between the towns of Cizre (Djezire) and Hakkari.

Please respect our policy and don’t discuss this rug which is for sale.

Is the old Kurdish rug I have handled an authentic rug staying in the continuum of old traditions? Yes.

Look how many of the characteristics summarize in the introduction to this topic are combined in this rug.

Use this photo to imagine it : Eagleton’s book (plate 72)

PHOTO

In this old "authentic" Kurdish rug, may be a divan cover, 5 1/2 concentric large scale hooked diamonds, may be coming out of more restrictive weave, are vertically aligned on the field surround by a simple main border containing herringbone motif. The colors are very saturated and the palette related to the Eagleton rug is typically Kurdish with various shades of madder red, indigo blue, yellow, brown, green, white and salmon.

It’s sizes, 269cm x 105cm are in the familiar long and narrow East Anatolian format and its handle heavy and floppy as one would expect for this type of rug.

It’s structure is in the continuum of the more early primitive filikli rugs

The wool is a mountainous silky wool and the pile height is about 1cm to 2,5cm. Thick (4 singles) symmetric knots are tied on four 2 ply tan wool warps, like jufti knots. The density is very coarse with only 18 knots per square inch and 10 to 14 wefts (single in various colors) separate the rows of knots.

As it is may be a divan cover the selvages haven’t any special treatment. At top and bottom there is a small skirt measuring 3" in weft faced plain weave with blue and red strips. End finishes are usual oblique interlacing..

Thanks, Daniel

Eagleton plate 72 (structure analysis adapted to Marla Mallett nomenclature)

Type: Van-Hakkari – Goyan tribe – circa 1940

Sizes: V345cm x H104cm, V11’6" H3’6"

Knots: symmetric, wool, 5’ x 5’

Warps: tan wool.

Weft: tan, red, white, yellow wool wefts, 4-5 picks

Selvage: flat interlaced selvage.

Ends: one type of oblique (interlacing or wrapping) end finish

Old Kurdish rug (divan cover) with 5 1/2 concentric hooked diamonds.

South East Anatolia - May be Goyan tribe – circa 1870

Structure analysis:

Sizes: V 269 cm x H105 cm;V9’ x.3’6" (end finishes included)

Yarn: spin Z

Knots: symmetric, wool, 4 singles tied on 4 warps, pile height is about 1.5 – 2,5 cm – density : H3,5pi V5/pi 18psi

Warps: two ply medium tan wool; no depression.

Wefts: wool singles dyed red, brown, dark ,blue wool, 10-14 picks, wefts don’t cross.

Selvage: no special treatment (divan cover?)

End: 3" in weft faced plain weave skirt with blue and red strips - oblique interlacing

It is almost frightening to see someone articulate your own thinking as Daniel has done above. In short, although it represents only the beginning of a longer journey, Daniel's thinking concurs with my own.

Personally, I always prefer weavings that appear to be old and the older the better. Some of the weavings Daniel has offered in this Salon push my own tolerance button for age. However, the critical point is that if one is collecting to understand and to form a comprehensive collection, what is the choice? I would argue that the Kurdish weaving tradition is a several thousand year continuum. In this light, splitting hairs over a generation or even several can be counter-productive if a weaving otherwise fits into the puzzle and meets you own criteria for selection. I am not arguing that one should be less selective or less discriminating. I am arguing that a collector should not eliminate a piece merely because it would not appear to be old enough.

Thanks to Daniel. Michael

Thanks for the Salon. So far, it has been a pleasure in reading and thinking about the subject.

The snakes you mentioned could be Christian.
The devil in Yezidis tradition is portrayed by the peacock.
The "Proto Kurdish" rugs could be Armenian.

Up to this moment, I did get some extra info about traditional Kurdish objects, I think, but it's the whole picture that makes the rug. Maybe I'll have to learn more.

Best regards,
Vincent

In my previous posting I referred to an old "authentic" Kurdish rug that I was not allowed to show on the board.

Michael Wendorf informed me that he had just purchased it and as it was no more in dealers hands that I was free to use it. Congratulations Michael.

Few rugs of this quality woven in the "authentic" tradition are available. So I am very glad to offer you the photos of it.

For those of you who would want to use my summary of "what is the tradition in Kurdish weavings", as a guide line, it would be necessary to be complete, to add that, Kurdish rug selvages are not limited to the reinforced examples I have showed. I forgot to cite the "split symmetrical wrapping" that Bruggeman and Böhmer (Rug of the Peasants of Anatolia) recorded in so many East Anatolian Kurdish rugs.

Here is what I said about this rug in my previous posting:

Is this old Kurdish rug, I have handled an authentic rug staying in the continuum of old traditions? Yes.

Look how many of the characteristics summarize in the introduction to this topic are combined in this rug.

In this old "authentic" Kurdish rug, may be a divan cover, 5 1/2 concentric large scale hooked diamonds, may be coming out of more restrictive weave, are vertically aligned on the field surround by a simple main border containing herringbone motif.

It’s sizes, 269cm x 105cm are in the familiar long and narrow East Anatolian format and its handle heavy and floppy as one would expect for this type of rug.

The colors are very saturated and the palette related to the Eagleton rug is typically Kurdish with various shades of madder red, indigo blue, yellow, brown, green, white and salmon.

The wool is a mountainous silky wool and the pile height is about 1cm to 2,5cm.

It’s structure is in the continuum of the more early primitive filikli rugs

Thick (4 singles) symmetric knots are tied on four 2 ply tan wool warps, like jufti knots. The density is very coarse with only 18 knots per square inch and 10 to 14 wefts (single in various colors) separate the rows of knots.

As it is may be a divan cover the selvages haven’t any special treatment. At top and bottom there is a small skirt measuring 3" in weft faced plain weave with blue and red strips. End finishes are usual oblique interlacing..

Thanks, Daniel

Structure analysis

Old Kurdish rug (divan cover) with 5 1/2 concentric hooked diamonds.

South East Anatolia - May be Goyan tribe – circa 1870

Sizes: V 269 cm x H105 cm;V9’ x.3’6" (end finishes included)

Yarn: spin Z

Knots: symmetric, wool, 4 singles tied on 4 warps, pile height is about 1.5 – 2,5 cm – density : H3,5pi V5/pi 18psi

Warps: two ply medium tan wool; no depression.

Wefts: wool singles dyed red, brown, dark ,blue wool, 10-14 picks, wefts don’t cross.

Selvage: no special treatment (divan cover?)

End: 3" in weft faced plain weave skirt with blue and red strips - oblique interlacing

The most striking feature of your rug, with a gabbeh-like structure, is that it shows an intermediary stage of evolution between the primitive Filikli rugs and the more conventional ones.

Any thoughts?

Best,
Daniel

My appreciation and understanding of this rug evolved a little different than you might imagine. My initial reaction was to the color and drawing. The color shouts out that the rug is Kurdish and in its careful drawing and execution (despite its coarse weave)I saw a well practised and powerful old tradition.

The old tradition I initially saw was a weft-float brocade tradition. While some may prefer a less well drawn design and call it archaic, thereby implying prototype or greater age, I compared it to a complete bag that I have displayed and referenced on these boards before - it is a Kurdish saddle-bag or khorjin that was pictured in Hali 81 page 65,plate 20.

The saddle bags are woven in a 2/2 weft-float brocade on a plain weave ground (one ground weft after two pattern rows) with the two-span floats alternately, and in places, vertically aligned and covering the ground weave. extra weft wrapping: (i)4/2 horizontal; (ii) 2/3 diagonal with the oblique spans on the front. To me, this is a kind of weft-float brocading which is very old and which entails interlacing warps and wefts. This more restrictive technique (compared to knotted pile)drives the design. If it is not well executed it cannot fairly be called archaic, it should be called sloppy or derivative, in my opinion.

As we have observed, Kurds seem to love such interlacing as we find even in weft-float brocading. In any event, the saddle bags referred to above also contain large concentric diamonds, in this case one in each half. In addition, I have rugs and a piled single detached face with a similar concentric diamond design.

When I saw this rug with colors so characteristic of Kurdish taste and evoking this weft-float brocading tradition that I know from other cherished pieces with 5 1/2 concentric large hooked diamonds, it struck a chord.

Of course, we examine rugs from many perspectives to understand them on their own terms. So I also sought to understand the structure as part of the process of evaluating this rug. Here is where I started to place this rug in some kind of possible continuum and as a possible link to very early or primitive knotted pile. This is because when I think of really early weaving, I think back to weavers working before the invention of sheds and heddles and of techniques like weftless soumak. Such simple forms of weft-wrapping take us back perhaps 8000 years in time. As John Wertime has written in Hali 100, long sleeping rugs such as the one in your salon and which you refer to as filikli may themselves have evolved out of extra weft wrapping. Indeed, Wertime wrote that "perhaps unspun lenghts of fibre were used at first, as they required no particular preparation or cutting given their limited, albeit often long, natural dimensions. In time, however, the use of spun yarn in hanks became standard because of the ease of handling, wrapping, storage and dyeing...the technique of making individual knots eliminated any trace of the original pile loops, and in time transformed the nature of knotted pile." See Wertime at page 93 of Hali 100 for more. As I think about the gabbeh like structure of this rug and the pattern I start to imagine that this is what rugs may have looked like when and if weavers used abbreviated wrapping techniques as an intermediate step in the evolution of knotted pile with what Wertime surmised were long, discrete lengths of fiber, or cutting yarn from a hank prior to the establishment of the short pile rugs with intricate designs that came to dominant over time.

Well it is a little embarrasing to carry on like this about a rug in one's collection, but I hope that perhaps I have shared a thought or two that could be relevant to your question. Perhaps you could let me know whether you share any of my interpretations or see it differently.

Thanks for your interest, Michael

Wertime doesn't specify the different way in handling the weft-wrapping structure.

The so-called simple-weft-wrapping with the overlapping wrapping-wefts slanted down to the right (in order to get the plain-weave ground as a symmetric-knot structure) is not convincing to me. The weft handling of the forward 2/back 1 weft-wrapping could well be a post-symmetrical knot invention and is different from the so-called simple-weft-wrapping. In short....I think it's very difficult to establish a very long historic episode (8000 years) in combination with Kurdish weft-wrapping structure, symmetrical knots and the limitations in design.

Best regards,
Vincent

I agree, at least in part, with Vincent. But since I have discussed the flaws in Emory's weft-wrapping/knotted-pile evolution argument at length in my book (p. 47), I won't do so here.

I have a couple of comments concerning the "tradition" that you say you see embodied in the reciprocal brocade saddlebag.

1. It is not at all certain that these saddlebags are Kurdish. They are from the Kars area of NE Anatolia where there is a large Circasian Turk population.

2. The reciprocal brocade structure has been used most prolifically by Turkmen/Yoruk nomads in Western Turkey, but also by both Kurds and Turks in SE Anatolia. The structure does not appear, to my knowledge outside of Turkey--in fact it is unknown to weavers in the rest of the world. The superficially similar-looking structure used by Kurds in Iran is wrapped, not interlaced.

3. We have no solid evidence to suggest that the reciprocal brocade structure has a long history. No archaeological examples have been found, nor examples which suggest great age.

4. Unlike all other brocading, the reciprocal brocade structure is NOT restrictive. Rather, it is the opposite, as it is essentially reinforced tapestry. In experimental examples, the ground wefts appear to have been added merely to allow the articulation of popular pile rug motifs like Memling Guls that have long verticals. When used for diagonal tapestry-style patterning, those ground wefts merely stabilize the weave. The reciprocal brocade PROCESS, however, does encourage narrow pattern units.

It's an interesting pile rug you have shown--with a pure slit-tapestry motif.

Best,
Marla

I take your valuable and interesting points, particularly the drawing supplied by Vincent. And, of course it is difficult to establish even a historical weaving continuum, much less a continuum between weft wrapping and symmetrical knots. Still there are disturbing gaps in the picture I am trying to see that seem to go back to weft-wrapping structures - for example and in particular, weftless soumak.

Marla, I regret that I am unable to lay my hands on your book at this time so that I can refresh my recollection of your response to Irene Emery's arguments in "The Primary Structures of Fabrics: An Historical Classification." Of course, John Wertime was a student of Emery and his thoughts reflect that influence. I think you are correct that this is not the Salon to discuss
your differences with Emery.

Responding to your specific comments:

1. I hope that you are not suggesting that the bags in the image above, which are related to a bag on your website, are Circassian Turk. If you are, we have to reevaluate the underpinnings of a lot of what the rug world commonly calls Kurdish and I would like to know the basis of your conclusion beyond the fact that other people live in the area they are typically found in. To me they certainly are Kurdish, all the way down to the warps and wefts.

2. OK, but again are you suggesting these bags are not Kurdish? The fact that other people, mostly Turks in eastern and western Turkey, also use reciprocal brocading means what - that these bags are not Kurdish? Is it not at least a little interesting that Turks who entered Kurdistan from the east and also moved further west use this structure while Turks elsewhere do not. What may we reasonably infer from this? That Turks learned this technique from Kurds in historic Kurdistan?

3. You are correct, I have no solid evidence and would be the first to admit that I could be wrong. The only evidence is what you have in your hands when you hold some of these pieces. To me, they suggest a long tradition. Of course, the same lack of solid evidence is basically true of weftless soumak, the archaeological samples are not wool. The same is also true of the sleeping rugs with the unspun wool pile that Daniel calls Filikli. The picture is incomplete. What to do?

4. I do not see the rug as a pure slit-tapestry motif, either in the field or border, but I understand your view. I hope we have a chance to handle both the rug and the bags together in the near future.

Thanks to both of you for the challenge to easy thinking.

Best, Michael

You and I have previously discussed the great difficulties involved in separating old Turkic and Kurdish brocaded storage sacks from the Antep/Maras area of southeastern central Anatolia. Since neither of us has done field work in that area, we simply cannot say with certainty who made which pieces. Unfortunately, in the literature, the various flatweave structures have been badly confused and misidentified. The area around Kars provides similar difficulties as there are several different populations in the area, so the "conclusion" that I stated above was simply that "it is not at all certain" that the reciprocal brocade bags from there (both yours and mine) are Kurdish. Similar pieces are not made elsewhere, to my knowledge. They represent one quite specific type. Kurds in the Malatya area do use the technique, but with different patterning on quite different objects. Kurds living south of the Kars area and on down into the southeastern corner of Turkey do not use the technique at all, nor do Kurds across the borders in Iran or Iraq. Such much for the technique's presence in the "heartland."

The areas in which we can make certain attributions for such work are in northwestern Turkey where the nomadic weavers are all Turkmen/Yoruk. Nearly all of these people's storage sacks incorporate reciprocal brocading.

We do have positive evidence that weftless soumak is ancient. A textile fragment with this structure was found at Neolithic Catal Huyuk in central Anatolia. The material was a bast fiber, but the structure was the same as in modern Kurdish work. There is no question about that one!

We can only guess at the history behind reciprocal brocading, but my instincts, after having studied these pieces for about 20 years, tell me that this structure is a fairly recent development--that it probably evolved some time after Turkic migrations into Anatolia. The purest forms appear in the Bergama area of northwest Anatolia, while bastardized versions are more common among the Antep/Maras area material. It's the experimentation that I've seen among the earliest extant work that suggests to me that the structure is a relatively recent development. Please realize that new woven structures and variations are evolving all the time, and a few transitional pieces do exist that display earlier reciprocal brocade structures. One unusual example is shown in Fig. 8.26 of my book. I have not seen other published examples.

I mentioned before that unlike other brocading, reciprocal brocading is not a restrictive structure, and the patterning on most reciprocal brocaded objects is derivative, having come from other media--primarily slit tapestry, but also overlay-underlay brocading and knotted pile. I seriously doubt that you want to read a long-winded discussion of this technical evolution and the resulting pattern developments!

Vincent,

Irene Emery's theory of symmetrical knotted-pile evolution (cited by J. Wertime) was based on a rarely used structure, "alternating soumak," not common soumak wrapping. The structure is not found among Anatolian flatweaves, only among occasional pieces from the Caucasus. There are very serious flaws in the argument.

Best,
Marla

Yes, you and I are in total agreement on weftless soumak.

My conclusion or certainty concerning the reciprocal brocaded bags being Kurdish is not predicated on point of origin or structure, though none of these are inconsistent with or rule out a Kurdish origin. These bags "feel" Kurdish to me - this extends to color, design, handle, wool, warp and weft. I think you sometimes underestimate the importance of color in this analysis. Alternatively, perhaps I over value color. But you are correct that I cannot make such an attribution with absolute certainty.

Only one more thing, I actually find your long-winded discussions of technical evolution and the resulting pattern developments to be very compelling and thought provoking. Do continue, please.

Best, Michael