Turkmen Shemle Gul Torba

Hi People

The shemle gul design isn't exactly rare, but is pretty uncommon. I've seen examples attributed to the Salor, Saryk, Tekke and various Ersari subgroups, although I don't recall any with Chodor or Yomud attibutions. Anyway, it seems to have crossed Turkmen tribal lines fairly widely. I don't recall seeing it on any format except the torba (or jollar; it's hard to know which unless the back is still on a torba).

There are two excellent Salor examples in the Russian collections. Here's one of them, plate 13 in Tsareva's Rugs and Carpets from Central Asia:

The "bow tie" main border is present on most shemle gul torbas. The field design comes in two variants. One, which can be seen in the piece above, has the lattice colors arranged more or less at random. The other has them in diagonal lines. I was fortunate enough to acquire one of that type not too long ago. Here is is:

Like most of the published examples, the colors are very saturated and intense, and don't photograph well. Here are a couple of details,

The reds in the first detail shot look much more aggressive on my monitor than they do in person, and I think the dyes are all natural. The handle is about like that of a wet rag. It is knotted asymmetrically, open to the right, about 8 to 9 per inch horizontally and about 12 per inch vertically; around 100 knots per square inch. I would attribute it to some Ersari group, but am reluctant to be any more specific than that.

Incidentally, both the pieces shown here are about the same size. The difference in detail is obvious, and reflects the difference in knot density. The Salor example has about 260 knots per square inch.


Steve Price

Posted by Richard Larkin on 03-08-2008 10:01 AM:

Beautiful piece, Steve. Mabrouk! Any chance of a close shot of the back?

Rich Larkin

Posted by Steve Price on 03-08-2008 10:55 AM:

Hi Rich

It's hanging on a wall, so the back can't be conveniently photographed at the moment. But I did roll back a section, so you can see a little of the back here:

There's no warp depression, the foundation looks to be all wool. Two ply two ply wefts, ivory. The warps are also two ply. Those on the left hand six inches pretty coarse; some all brown, some brown and ivory plied together. For the other 3'6" or so the warps are quite fine, brown and ivory. I don't know what to make of that change; probably just an oddity.


Steve Price

Posted by Jim Allen on 03-08-2008 11:57 AM:


The first thing that struck my attention is the fact that the design is upside down. The knotting seems to be very neat. It is obviously not a commercial piece and for Ersari work that is pretty rare. The Ersari and the Salor were closely related during the classical period so the fact that the Ersari later copied Salor designs isn’t surprising. I have never understood the shemle design but I wonder if it was inspired by the family crests of knights or their shields? I have only recently been made aware of communications between old Armenia and central Asian steppe nomads via Dr. Manuelian of Tufts University. As time goes along I think more and more linkages connecting diverse Asian groups will come to light. One thing I learned from reading O’Donovan was that the Turkomen highly valued females from competing peoples. The Turkomen were matriarchal and both names and property passed on the maternal side of families. This might go a long way in explaining some of the more obtuse Turkoman iconographies, like the shemle gull. One might even suspect that really old shemle gull trappings were woven by descendants of Armenian’s. We all know about the nomad’s penchant for stealing or procuring fine Chinese brides so why not Armenian ones?

Jim Allen

Posted by Steve Price on 03-08-2008 04:01 PM:

Hi Jim

Although the orientation of the guls differ by 180 degrees in these two examples, I don't know of any way to determine which way is "right side up" (or even if there is such a thing as "right side up" for this motif). There are published examples of presumably old pieces showing it both ways. For example, a Salor in Mackie and Thompson's Turkmen has the same orientation as mine, as does a piece attributed by George O'Bannon to the Kizyl Ayak in Vanishing Jewels. Another in Tsareva's book, from the Dudin collection, has the opposite orientation.

Steve Price

Posted by Jim Allen on 03-08-2008 04:29 PM:


It is interesting but the same can be said about the red and white orientation of a great many Turkoman chuval gulls. I haven't compiled any statistics but I feel like the mirror orientation of the red and white quadrants of Tekke chuval gulls varies by generations and not by individual weavers. I have always thought that such manifestations of biased conjugate polarity was indicative of copying, one generation after the other or about every fifteen years. Wouldn't it be interesting if this red and white flipping counted Tekke generations like white marker knots do the time it took to weave any good Yomud dowry piece.

Posted by Yon Bard on 03-08-2008 10:22 PM:


[How do I get pictures into this thing?]
By a strange coincidence, this shemle-gul torba appeared within a couple of weeks of two other pieces of this ilk. One was sold at Northeast Auctions on 2/24;

the other was shown recently at a Textile Museum Rug Morning.

Apart from their similar designs, these two pieces share an attribution: the Northeast piece was cataloged as Salor, and the TM piece (I’ll call it that though it doesn’t actually belong to the Museum) was judged to be a Salor by some of the experts present (so it’s been reported; I wasn’t there myself). So this gives me a chance to raise the question of Salor attributions to pieces that are not “classical” Salors.

"Classical" pieces, presumably those predating the Salor defeat in the 1850s, are instantly recognizable by means of the following characteristics, and perhaps some others:

1. Open to the left knots (it's been claimed that perhaps 10% of the pieces are open right; so be it, but I have never personally seen one of those; no matter, the other characteristics are sufficient in themselves)
2. Deeply (but not necessarily completely) depressed alternate warps
3. Exquisite workmanship
4. Knot density more than 200 per square inch (except in main carpets)
5. Fairly strict adherence to prescribed design elements for each type of piece.

Non-classical Salors, whose existence is controversial, are presumably later, perhaps woven by Salor women in their Diaspora. The attribution is based on the presence of some, but not all, of these criteria, and the lack of obvious alternative attributions.

Coming to the pieces at hand, what can we say about them relative to the classical Salor attributes? (I realize that Steve makes no claim for Salorhood, but I’ll include it anyway since it’s so similar to the others)

1. The TM piece has open to the right knots, which does not absolutely preclude Salor attribution. The Northeast piece is open left. Don’t know about Steve’s piece.
2-4. Warp depression and workmanship: for the TM piece we don't know. We also don’t know about Steve’s piece, except for the knot density of 100. Perhaps the missing information can be provided by the owners. The Northeast piece has slightly but irregularly depressed warps, sloppy workmanship, and a knot density of about 110 – all of which argues against Salor origin.
5. Adherence to classical design: all three pieces are definitely not classical Salor in design. Just about all Salor torbas have a narrow band of rams' horns just below the top; our pieces don't. All the Salor shemle-gul torbas that I have seen in the literature (Tsareva 13, 14; Jourdan 12, 13; Mackie & Thompson 11) have a somewhat vertical shemle arrangement, i.e., similarly colored shemle guls are often stacked one above the other, rather than the strictly diagonal scheme of all three of our piece, which is characteristic of Saryk (Jourdan 30) and other tribes.

I have examined the Northeast piece at length. It has much corroded silk and some striking colors, particularly an almost-black blue. While it is easy to say “some unknown Amu-Darya group, possibly related to the Salor,” I feel that this would be a cop-out. I left feeling that I really couldn’t clearly place it in any known category.

I'd be interested in hearing other people's thoughts on these issues.

Regards, Yon Bard

Posted by Steve Price on 03-08-2008 10:53 PM:

Hi yon

Welcome back - it's been awhile since I've heard from you.

If you have images to post, just send them to me as e-mail attachments and I'll send you back instructions on how to put them into a message, or I can put them into your messages myself if you prefer.

I've seen some of those nonclassical Salor attributions, and am kind of baffled by them. If the pieces don't fit the customary criteria for Salor, why call them that? It's uncertain enough when the traditional characteristics are present.

My piece is knotted asymmetric, open right, no warp depression, about 100 knots per square inch. Also, no silk, none of the additional borders generally associated with Salor shemle gul torbas. Do I know who wove it? Nope. But the best fit is somewhere within the Ersari universe. Is it possible that it's Salor? I guess so, but unless someone can trace the provenance back to some Salor woman who wove it, that wouldn't rise very high on my list of likely attributions.

Do you happen to know the basis on which the experts at the TM gave the Salor attribution to the piece shown there?


Steve Price

Posted by Filiberto Boncompagni on 03-09-2008 04:16 AM:

The images are now in Yon's post (Sorry Yon, I thought you knew that you had to copy the links and paste them into your text).


Posted by Yon Bard on 03-09-2008 11:18 AM:

Steve, I feel very frustrated - I agree with everything you say. Nothing to argue about (remember the fierce arguments we used to have about internal elems, etc.?).

No, I don't know what the experts' thoughts were that led them to the Salor attribution. I'd like to find out myself, but I don't even know who they are.

Regards, Yon

Posted by Steve Price on 03-09-2008 12:30 PM:

Hi Yon

It's good to have you back. I'm sure we'll come to something about which we disagree soon enough.

The two photos you posted remind me of one other reason I don't think mine is Salor: the color. The piece from the NE auction has a ground color that looks more Yomud-like than the clear red I'm accustomed to seeing on Salor weavings. Mine also has a ground that's more of a Yomud rust-red than Salor. I also notice that the guls in the two you posted are oriented 180 degrees differently. My informal impression is that about half the ones I've seen are one way, about half are the other.


Steve Price

Posted by Yon Bard on 03-09-2008 12:45 PM:

Steve, I doubt there's any significance to the orientation.

Posted by Steve Price on 03-09-2008 01:06 PM:

Hi Yon

I'm with you on this one, too. I only mentioned it because of an earlier post by someone implying that one orientation is rightside up and the other is upside down.


Steve Price

Posted by Jim Allen on 03-09-2008 01:12 PM:

flipprd gulls

It would be nice to know if these torbas were woven in pairs on one loom and the "gull" orientation was flipped by weaving its mirror image. This might also explain the red and white oscillation around the main gulls centers in pieces with chuval gulls. Jim Allen

Posted by Steve Price on 03-09-2008 01:17 PM:

Hi Jim

That seems reasonable to me. I suspect that any torba-shaped object was woven in multiple copies in the vertical direction (one face, its back, another face, its back, etc.). Why set up a loom to make something as shallow as a torba all by itself?

Inversion of motifs on one of a pair is something I recall seeing in intact khorjins, so the same might be the case here.

Steve Price

Posted by Yon Bard on 03-09-2008 03:17 PM:

Jim, If weavers were consistent in which orientation they used first, then one could verify your hypothesis by seeing whether there is a correlation between the orientation of the diagonals and the direction of weave (from the bottom or from the top). Unfortunately I don't have this information for any of the pieces undr discussion.

Regards, Yon

Posted by Richard Larkin on 03-09-2008 04:21 PM:

Hi all,

Steve, I agree with your comment about the color of the Northeast example being unlike what one expects in "classic: Salor pieces, although at least one of the examples in Ms. Tsareva's catalog looks pretty dark. Of course, it begs the whole question of what qualifies as "classic" Salor, and who wove them when. I agree with Yon that the "real" ones seem to be set apart from other weavings, and have very distinctive weave, structure, handle, and presence. I haven't handled that many, nor had them in hand very long, but they always suggested to me that the weavers were very sophisticated, and knew exactly what to do to achieve a highly distinctive fabric. By comparison and contrast, other pieces, including some of high quality, seemed more to have achieved their end character more by accident. That is, reasonable choices of materials, technique, etc., were made, and the pieces turned out the way they turned out. Perhaps I'm reading too much into those Salors.

If there was a special level of skill in that way among pre-1850 Salor weavers, one wonders what happened afterwards. Notwithstanding devastating military and social setbacks, there must have been continuing day to day weaving among the survivors.

Posted by Jim Allen on 03-09-2008 04:43 PM:


I have examined only a few pairs of Turkoman bag faces. I have never seen a true antique bag set still joined in the middle. I would like to know what the loom setup was like for weaving Turkoman bags. If they are woven on looms with warp ends staked in the ground like a rug then a weaver moving from one end of the loom to the other end or one side to the other would be predisposed to weave a general mirror inversion of diagonally arranged colored design elements.

The few pairs I have seen showed quite a bit of difference between their designs. They looked like individual creations and not one copied after the other. I must admit I have never consciously looked for the kind of color and form inversions I am proposing here. I do think this is interesting and I should have been more observant.

I don’t live anywhere near a body of rugs to examine but this might be a good enough question the TM or the MET might be interested in looking into it. The MET has perhaps the most famous of all chuval pairs, the Ballard Arabatchis. In Turkmen, image #54, the inventory number is 22.100.40a. The “a” at the end means there is a “b”. I have seen the pair and they are both top of the pyramid pieces but they show quite a number of differences. Again I can’t quantify my statements but it is something that could be investigated.

I would tend to consider the details concerning this one pair to be exemplary of classical Turkoman bag weaving in general. Everybody who is deeply interested in Turkoman weavings should make an appointment and go see these two chuvals for themselves. I guarantee that you will come away with a new bench mark in your mind for what constitutes the absolute best in Turkoman weaving.

Jim Allen

Posted by Rob van Wieringen on 03-10-2008 07:12 PM:


Recently I have handled an early, matching pair of Yomut chuvals. The design was identical and they both had the same pile direction.
Both the main gul and the minor gul had this color inversion, when comparing the two chuvals.
In the main gul there was not only the red-white inversion, but also total colour inversion in the smaller details of the main gul.

This color inversion seemed to be done pretty on purpose and I can imagine when hanging "on the flanks of a horse" would give a more balanced impression, then without this inversion.



Posted by Yon Bard on 03-10-2008 10:57 PM:

Rob, what do you mean the pile was in the same direction in both chuvals? Do you mean that both were woven the same way, either from the top or from the bottom? This would fly in the face of the wide-spread belief that bag pairs were woven like saddle bags, i.e. in the following order: first face, first back, decond back, second face, which implies that in each pair the first face is woven top-down and the second bottom-up.

Regards, Yon

Posted by Rob van Wieringen on 03-11-2008 04:43 AM:


Both were woven from bottom to the top.
Apparently the Turkmen were not mainstream and had a different way of making their bags.


Posted by Steve Price on 03-11-2008 04:53 AM:

Hi Yon

A "matching pair" (by which Rob appears to mean, having nearly identical designs and colors) may have been made separately (that is, not along the same warps). Juvals are about twice the depth of torbas; many may have been woven individually. Including the back, most juvals have warps around 5 or 6 feet long before being removed from the loom.


Steve Price

Added note: This subject got some attention here nearly 10 years ago in the discussion associated with Salon 16

Posted by Yon Bard on 03-11-2008 09:40 AM:

If matched pairs were made either separately or not in the saddlebag sequence, then that could explain the phenomenon that I mentioned in Salon 16, namely that according to my count bag faces woven bottom-up outnumber top-down by 3:2. This is contrary to expectations from the saddlebag sequence.

Regards, Yon

Posted by Sue Zimmerman on 03-11-2008 11:13 AM:


Hi Yon,

Have you ever seen a Turkmen torba, which was woven with fringe, which wasn't woven upside down? I haven't.
I have just two pairs of Turkmen weavings. Neither set even shared the same loom time with their mate. Different warp setts. I find studying pairs of weavings very interesting. Very expensive, though. Sue

Posted by Steve Price on 03-11-2008 12:13 PM:

Hi Sue

The torba with which I opened this thread had a fringe once upon a time (you can see remnants of it near the bottom). It was woven from the bottom to the top.

I'm reluctant to say that nearly every Turkmen torba had a fringe originally, but an awful lot of them did. I'm not aware of any reason to associate the inclusion of a fringe with the direction of the weave. Nor do I understand what you mean by pairs of Turkmen weavings if they come from different warp setts.

Steve Price

Posted by Sue Zimmerman on 03-11-2008 02:11 PM:


I have said nothing to imply all torbas were woven with fringe, if that is what you are implying or thinking. I am just pointing out I have never seen one which was woven with fringe that wasn't woven upside down, or as you put it, woven from the bottom up. Nor am I saying I have seen every torba with woven fringe. I'm curious whether Yon, or others, have seen a torba with fringe woven from the top down, as you would put it. I've been looking for one like that, unsuccessfully. I'd like to see one if one exists.
By ''pairs'' I mean Turkmen weavings which were traditionally woven in pairs. Had I simply seen the obviously paired pairs I have, just in photos, I would probably have assumed they had the same warp sett. They do not. I was surprised to find that out. That's one of the reasons I like to study pairs of weavings. I hope that is clear enough now for you to understand. I bought them from a very experienced and deservedly respected dealer as pairs, if it's my ability to know a pair when I see one you question or object to. Sue

Posted by Steve Price on 03-11-2008 05:47 PM:

Hi Sue

I'll just say that I'm unable to conceive of any reason why a weaving should be considered to have been woven upside down if the weaver's position has her seeing it in its preferred orientation. For a torba, of course, that's with the bottom at the bottom and the top at the top.

Setting that aside. The torba with which I opened this thread was woven from the bottom up, and had fringes once upon a time. The one on the wall in the next picture was woven from the top down, and also had fringes at the bottom.

Both are within view from where I'm sitting, which makes me think that either weaving orientation is likely to occur more than once in a lifetime.

Steve Price

Posted by Chuck Wagner on 03-11-2008 08:00 PM:

Hi Steve,

This image is from 'Turkmenskoye Narodnoye Iskusstvo' (Turkmen Folk Art) by Medzhitova, Dzhumaniyazova, and Grishin :

The caption says it's Ersari, from Kerki, and late 19th-early 20th century.

Chuck Wagner

Chuck Wagner

Posted by Sue Zimmerman on 03-11-2008 08:03 PM:

Steve, I was talking only of torbas with fringe. Just torbas. Sue

Posted by Steve Price on 03-11-2008 08:21 PM:

Originally posted by Sue Zimmerman
Steve, I was talking only of torbas with fringe. Just torbas. Sue

Hi Sue

So was I. The pieces I referred (and showed) to are both torbas, both have the remnants of fringe. I have at least one more, a ca 1900 Yomud, with fringe still more or less intact. Woven from the top down. Oh, I forgot to mention - that one's a torba, too.

Steve Price

Posted by David R E Hunt on 03-16-2008 02:09 PM:

Hi Steve

Not that I want to change change the subject, but that white ground torba in the photo you posted above is beautiful.
I'm sure it has been discussed here on turkotek before, but I can't seem to locate the thread. Could you direct me to the thread? What is the torba's provenance? Tekke? Yomud?


Posted by Steve Price on 03-16-2008 02:21 PM:

Hi Dave

It appears in a thread in Jerry Silverman's Salon, entitled "When a rug is actually a rug?" . I think it's probably Igdyr, on the basis of the palette (especially, the green). Here is a link to the Salon essay. This piece shows up in a photo in this thread.


Steve Price

Posted by James Blanchard on 03-16-2008 02:24 PM:

I second David's motion.... Steve, any chance of a closer look? It looks very appealing to me.


Posted by Steve Price on 03-16-2008 02:47 PM:

Hi James and Dave

I opened another thread with it. I'd like to keep this one from losing focus much more than it has already.


Steve Price

Posted by David R E Hunt on 03-20-2008 10:28 PM:

Hi Folks

Find below a couple images of a torba in the shemle pattern with the more traditional (open left) structure and some synthetic dyes. It was reputed to be a late Salor weaving, and while the final bid was much less than what you would expect for a classic Salor weaving, it wasn't inexpensive. I can't recall the knot count.

While it may have the traditional knot structure, the kont count seems low, to judge from the drawing of the design elements. I suppose the warp depression within the Salor range. Some of the design elements seem distorted, and there is a diagonal orientation of the design.

Is this some kind of later hybrid? The texture, and knotting as seen from the back remind of an Ali Eli chuval that I own, which is also woven open to the left. I remember seeing a Salor chuval being offered for sale by a well known dealer, which sported a gul kindered to the Ali Eli, and purported to proceed from the close spatial relationship the two tribes shared at some point in time.


Posted by Richard Larkin on 03-21-2008 09:15 AM:

Hi Dave,

I barely know who the Ali Eli are, so you're probably way ahead of me on this. I would say, however, that the close-up of the back you posted looks very Salor to me, though degraded. I find it very interesting, as one wonders where the weavers of the great earlier pieces went.

Rich Larkin

Posted by James Blanchard on 03-21-2008 11:19 AM:

Hi Steve and all,

Not much has been mentioned in this thread about the possible origins of the somewhat enigmatic shemle gul. Peter Stone relates it to the internal instrumentation of the "kejebe" design seen on many Salor and Saryk torbas, and in older engsis, including those attributed to the Arabatchi. The design similarity makes sense to me, considering the following example (from JBOC's site).

If that is the design lineage, then there seems to have been quite a bit of design re-use and degeneration in later weavings. Could the following "diamond and triple-cross" usage in a Herati-design Ersari-Beshir be another later version of the basic design?


Posted by Steve Price on 03-21-2008 11:55 AM:

Hi James

Stone's speculation might be right, might not be. If it is, yours might be right, might not be.

Any pattern can be morphed into any other pattern in very small steps. It may take a lot of them, but they can each be small enough to be plausible. I have no solution to the difficulty this presents in figuring out a pattern's evolution, and for this reason I generally sidestep design evolution.


Steve Price

Posted by Richard Larkin on 03-21-2008 12:23 PM:

Hi all,

I'm sort of straddling the Steve and James positions on this. On the one hand, it is difficult in my opinion to avoid the conclusion that the so-called "shemle" design element in the torba is somehow related to the element James illustrated in the kejebe patern (or vice versa!). What is more difficult is to assess the attitude and intent of the weaver in implementing it. It could be something as simple as choosing and modifying a suitably interesting design element that happened to be in the weaver's repertoire to fill a repeating hexagonal space; or it could involve the studied implementation of an important symbolic motif. Furthermore, the history of how the one got to be the other could be highly complex, and involve many venues. That's what we don't know. At least, I don't know.

Posted by Jim Allen on 03-21-2008 03:32 PM:

shemle gull?

Dr. Lucy Manuelien, Tufts University, has recently published material pointing to the existence of cultural exchanges between ancient Armenians and the Turcoman people. See her introduction to Merchants, Weavers, and Kings. The decidedly non Turcoman appearance (to me) of the Shemle design can be easily explained by the aforementioned cultural exchanges. I am specifically thinking about marriages, like those of young Armenian women to Turcoman men. The Armenian women would have grown up experiencing arched church windows, stone archway infill designs, and crosses of all kinds. In earlier Salor examples (see Tsareva, image 13) the inner designs remind me of a coat of arms, a medieval shield form, or an architectural element of design. I also would like to point out that small five knot crosses are used as infill designs within the hexagons of the Tsareva example and crosses are also used in more propitious places like the central apex for some of the inner “shemle” designs. This is very reminiscent of Armenian Church architecture where many design complexes are topped with crosses. Were the weavers of Shemle design wedding weavings keeping alive a tradition reaching back into old Armenia? Were later shemle gull pieces woven for reasons of fashion or heritage? I believe this is an avenue of inquiry worth observing. Jim Allen

Posted by David R E Hunt on 03-22-2008 12:36 PM:

Hi Richard

I checked with the TM's "Turkmen", and Thompson suggests that the shemle pattern is from other textiles. Seems to follow in the tradition of using Ikat as a pattern source in Turkmen carpet weaving. Jourdan, in his "Turkoman" suggests that numerous kindered examples are found in the lattice patterns of both Persia and Anatolia.


Posted by David R E Hunt on 03-23-2008 10:46 PM:

Hi Folks

This assertion of Thompson's, citing a textile as a possible source for the shemle pattern, inspired me to look around, and came up with a couple kindered patters of ikat. But first, the Salor Torba with shemle pattern, once again.

The two following ikat samples seem of kindered design.

The first is is about a 3' x 5' sample of modern silk ikat, and the second is a detail of another sample, of unstated dimensions. While neither is a dead ringer for the shemle pattern, I believe the common features are obvious. The vertical arrangement of both the design and colors are found in both ikat and the pile work of the shemle torba. Some of the design features of the ikat patterns are roughly analogous to the "lamp" elements in the shemle (are they both stylized floral forms?).And while it might have no significance, I noticed that the design elements,(the larger two on the sides), bear a rather strong resemblance to the floral forms we know and love so well, found in the elem panels of Turkmen weaving, when the smaller detailed ikat sample is inverted. Also, note these small green lozenge shaped elements, metered about the perimiter of the ikat floral form and surrounding the purple, triangular central element. Are their analogs to be found in the floral forms of the following detail from a Tekke engsi?