Posted by Lee Koch on 08-03-2007 11:36 AM:

Note: Part of this thread consists of posts that were originally in another forum, and were moved to this one for reasons that will become obvious. The sometimes awkward breaks in the topic are a result of this.

Steve Price

Folks, if we can take a break from the onslaught of Cassin invective for a moment to actually discuss the rug, my first question to the group is, what's the authority behind the statement that a finer weave combined with a higher vertical to horizontal knot ratio is a fail safe method of identifying only a late tekke rug?

This morning I took a quick glance through my copy of Tsareva "Rugs and Carpets of Central Asia" and of the torbas, mafrashes and chuvals that she identifies in the tekke section as having been woven in the mid nineteenth century or before, the v:h ratio ranges from about 1.4 to 2.2. (If I'm off by a decimal point because I'm doing this from memory and don't have access to the book at the moment, I ask your indulgence.)

Furthermore, the finer=later proposition hasn't been borne out in my personal experience either. I own about 15 pre-synthetic tekke chuvals and their ratios vary in that same range. The torbas in my collection follow a similar pattern.

So is it possible that the judgements being made about the subject chuval (not those about Cassin personally), are based on an unverifiable premise? What I mean is, while it seems probable that late tekke material invariably has a fine weave and a high v:h ratio, I don't think it necessarily follows that no non-late tekke production can't have the same characteristic. It's like when your doctor tells you your ok, what he's really saying is that he finds no evidence of disease in your body. But that's far from being the same thing as him telling you he has evidence or proof that you're disease free.

As for the number of guls and their size, other pieces have been identified as being relatively early with a similar number of guls. David Reuben published a piece in Gols and Guls II with 36 guls in a 6x6 configuration that he characterized as mid-nineteenth century. And when he ran the rug department at Woolley and Wallis, Ian Bennett identified another tekke small gul chuval with the same layout of 6x6 guls as having been woven in the third quarter of the nineteenth century. That bag is pictured on Barry O'Connell's website in the tekke chuval section.

So if there's no proof that the ebay chuval is a product of late tekke manufacture, the only basis for evaluating it is aesthetically, which basically comes down to personal preference. My comments about the subject bag's aesthetics relate to the depth of its field and the primary border. Anecdotally, my reaction to the size of the field was that it's unusually deep for a tekke. It's as if the weaver was determined to fit vertical columns of seven guls in the field. I haven't measured the depth to width ratio of other tekke chuvals so this is just the impression I had when I first saw the rug.

The primary border is another thing entirely. It's relatively rare in tekke work and appears to be a repetition of a figure in an uncommon carpet border seen in a few tekke and yomud rugs. Check out Leigh and Sally Marsh's small tekke rug in the Atlantic Collections book for an example. Looks like a curled leaf variant. To my knowledge, there are no published tekke bags with this border in any of the widely available turkmen reference books though I'm happy to be corrected if anyone knows of any. Klaus Troost published a yomud chuval with this border in his exhibition catalogue, Die Yomut. And Hans Sienknecht produced a home-made photo album of his turkmen bags in the 1990's in which he displays three chuvals with this primary border. Two he calls eagle group II and the third he calls tekke. Interestingly, the v:h ratios in all three Sienknecht bags are in the 2:1 range and the knots counts are all over 200 kpsi. The iconography is similar in all three pieces and it'd be interesting (to me, anyway) to hear his rationale for the attributions he made.

Lastly, I'll close with the comment that in my experience, presynthetic tekke chuvals are not all that common and I don't think the ebay bag has any bad dyes. Other tekke chuvals may evince a more spacious design with fewer ornaments that appeals to mainline turkmen orthodoxy more. But even if the subject bag tends toward the later end of the non-commercial period, it's a piece with enough merit to take it out of the realm of what passes for turkmen material on ebay these days.

Let the flaming resume.

Posted by Jack Williams on 08-03-2007 12:34 PM:

Gentlemen and ladies:

Focusing on the obvious deception and errors in knot count and ratio, tends to cause us to overlook other questions about this item.

Range of color - the lack of the usual number of Tekke colors and the almost randomness of the use of what colors there are seems strange. Imho, these newer “Tekke” items also seem to somehow give an impression of an over-use of "white." Compare the aura of the Thompson to his chuval.

Red color - I have some reservations about that red field. Its mottled look in the photos resembles the artificially dyed, chemical alizarin reds I've seen from 1920s and later maybe even lightly chemically washed. Compare the red of his chuval to that of the Thompson, you may see my point.

Blue color - I noted that the light blue in his posted pictures is too brilliant. It is a frequent result of a close-up taken under strong light. But true impressionistic color, notice you can hardly tell the difference between the blues in the picture of Cassin's whole rug. Notice that you can obviously see the different blues in the Thompson chuval.

Selvedge, border, and reduced size - he didn't mention anything about the selvedge, even after I asked specifically about that blue wrap. His rug was cut down and the selvedge re-worked. But... when was it done and how much more of the rug was repaired? Also, the border and elem..something is strange. They have almost a yomud look... In my experience this is found found mostly along the Amu Darya, the river-Turkmen, later a stew of many different tribal groups.

Knot, Asy open rt.? Given Cassin's clear incompetence or deceptions, what are the chances that the open right knot is correct?

Drawing - In my opinion, the chemche gul is where weavers usually applied special individuality and creativity. And imho the chemches are a key to visual artistic contrast.

In Cassin’s rug, the chuval guls are flattened and internally simplified almost to a cartoon. But also, the chemche guls seems to have no unique interpretation, lacking any contrast with the chuval gul. In the Thompson chuval, notice the interesting visual contrast between chemche-chuval gul. Notice the color changes in the chemches. Cassin's chemches and the mechanical sameness of the whole blends the composition into one jarring repetitous impression. I think this is more than just asthetic...this could/might be a cause for one to wonder about who, when, and where.

At one time, I occasionally considered Cassin’s opinions (mostly dismissing them). Perhaps I thought that despite his nuttiness, his long experience must count for something, and perhaps others felt the same way. But, I suspect this open display of his incompetence will just about end any attention previously given to him or his opinions. He is certainly now off my "watch" or "care" list.


Jack Williams

Posted by Lee Koch on 08-03-2007 04:50 PM:

Jack, hi –

As far as your concerns with the color representation of the ebay chuval, I believe they arise from the inaccuracies inherent in color transmission over the internet. With respect to the range of color in the subject piece, the basic tekke palette is undyed brown, ivory, two reds and a blue and a green or second blue. Exactly what the subject piece contains. As far as the field color is concerned, I’ve seen the same reflective qualities in a tekke small gul chuval I bought years ago from Marvin Amstey which is also a piece where the pile has survived intact. I don't believe the color irregularity you notice in the ebay images arises from the use of an unstable synthetic red but rather from a similar high density weave combined with the flawless surface of light reflective wool pile saturated with color. As far as the second blue goes, I was fortunate to collect the Pinner small tekke rug, lot 5 in the Rippon Boswell sale of the Pinner collection. The piece was unsold at the sale but I was in right place and right time later to acquire it from a dealer who bought it post-sale. This rug was originally purchased at auction in 1980 in London and in its Auction Price Guide section in the corresponding issue, the Hali editor commented on the piece's apparent age, saying it was a very early tekke item. I mention this because it in actually shares an electric light blue like the one you noticed in the ebay chuval, a fact which caused me to believe the subject piece could indeed contain a color of such brilliance.

With respect to the simplified major and minor ornaments, I would direct your attention to lot 9 of the Jon Thompson sale at Sotheby’s New York, December 1993. I believe Cassin erred in comparing his piece to the Hoffmeister chuval and the related piece in Turkmen. You’ll notice many more similarities with lot 9. Like the ebay piece, it contains the same simplified major and minor ornaments and the same tree designs in the elem featuring serrated leaves and diamond shaped “heads”. This chuval is representative of the group that the ebay bag clearly fits in. Another example is found in the Wiedersperg Collection at the DeYoung Museum in San Francisco. I own a third example and there are others. If you check these pieces out, you’ll notice that they all have the same basic drawing of the chuval guls and chemches, both simplified in their forms as you correctly point out. But rather than being emblematic of late production, the simplified major and minor ornaments are among the identifying factors signifying members of this group of chuvals. I do agree that the ebay piece chuval guls are more compressed than on the bags where there are fewer major guls in the field. I attribute this compression to the weaver's desire to incorporate columns of guls seven deep in her design for the field of her bag.

If you’re interested in learning more about classifications of small gul tekke chuvals, I’ll refer you to Ned Long’s pioneering and interesting summary which was published in Oriental Carpet and Textile Studies many years ago. Ned is a collector who took a special interest in tekke small gul chuvals and he developed a classification system based on the repetitive occurence in numerous chuvals of like elements, i.e. the drawing of major and minor guls and elem tree figures that we’ve been discussing here.

As for your observations about the primary border, I’ve addressed those concerns in my initial post in this thread. I think the primary border is a positive, rather than a negative feature of the ebay chuval.

Thanks for indulging me by reading this lengthy post.

Take care!

Posted by Jack Williams on 08-03-2007 05:08 PM:


Thanks Lee. Very informative and I shall look up those references. I've bought a number of things on line and have learned some rules of thumb about judging computer generated colors. Your information is well organized and presented...such is always welcome and appreciated.

Regards, Jack Williams

Posted by Steve Price on 08-03-2007 05:16 PM:

Hi Lee

Some time back, I presented a Salon examining the bases for rug attribution. A link to it is here. To make a long story short, my position was (and still is) that reliable criteria for age attribution of a particular genre of rugs can only be derived from a substantial database of rugs of that genre with documented ages covering the span of interest.

I doubt that there are as many as a dozen Turkmen pile weavings that can be documented to predate 1875. Attribution criteria before that time is conventional wisdom. Like most conventional wisdoms, there is probably more than a grain of truth in them. But they are still subject to large, often unappreciated uncertainties.

Within the conventional wisdom criteria for Turkmen rugs and bags is this: vertical knot density began increasing around the middle of the 19th century, becoming greatly increased by its end. It is believed (rightly or wrongly) that vertical knot counts were not much greater than horizontal knot counts before about 1850, and became more than twice as great by 1900. One consequence of this that is obvious when looking at pieces is that motifs become flattened and more crowded.

Regardless of whether the conventional wisdom is correct is beside the point here. It was applied by Cassin in the description of his juval, and the relevant information about his piece was wrong. The vertical/horizontal knot ratio is more than 2, not 1.2 as he claimed, and using the criterion that he applied, this suggests a late weaving, not an early one.

To my eye the only thing about it that would cause anyone to place it into the 19th century is that none of the colors are obviously synthetic. None of the comparison pieces that are attributed to pre-1850 dates (again, rightly or wrongly), cited by Cassin, have the flattened motifs and high vertical/horizontal knot ratios of this one. And that is the ONLY criterion Cassin brings to bear in attributing his juval to 1850 or earlier.


Steve Price

Posted by matthias wohlgemuth on 08-03-2007 06:14 PM:

hi everybody,

don't you feel - like i do - that it was a great relief to read the sensible and obviously well founded thoughts and observations of lee koch? and all and just about the subject of our common passion: the innocent RUG!

matthias wohlgemuth

p.s.: i own still another tekke (?) chuval with that same main border design, but featuring 4 x 4 diamond güls and a different minor güls (like two of the ones published by sienknecht and mentioned by lee). i don't think it's 20th c.;-)

Posted by Lee Koch on 08-03-2007 07:01 PM:


In my original post I stated my intention to bring the discussion back to the rug in question and away from the (mostly) valid criticisms of Mr. Cassin. I don’t believe that I have ever argued against any of those criticisms, nor would I. Like you, I am aware of the conventional wisdom involving the degeneration of Turkmen iconography based on market preferences for increasingly finer production that occurred mostly after the tekke were overrun by the Russian army in the 1880’s. The conventional wisdom also holds, with greater validity in my opinion, that the aforementioned compression of the weave pattern was also accompanied by the rapid multiplication of field elements and borders. While it does show an increased number of major and minor guls, the ebay piece does not show the numerous borders apparent in most obviously late tekke production. In fact, if you refer to the cited analogs from my previous post, you’ll see that the Cassin chuval has the same number of borders as the comparison examples, a fact that argues against it being 20th century work, in my opinion. The border systems in question feature a primary border flanked on both sides by a secondary border which is itself flanked by two guard borders. However, it’s undeniably illogical to assert a conventional wisdom that argues in favor of certain trends that were occurring in the 1850’s when you yourself assert that the sample of verifiable pre-1875 pieces is too small to give a meaningful understanding of that era of production.

And the fact is, I never opined that the ebay chuval was circa 1850. I have asserted that this chuval is more aptly compared to published examples that Cassin failed to mention in his description. And I am also asserting that because the subject bag has no apparent synthetic dyes, there is no proof that it is not of the pre-synthetic era, especially when other late production identifying factors like multiple borders that crowd the field are not present in the subject chuval.

To make sure that there is no misunderstanding of my assessment of this piece, let me recap that in closing. My argument has never been about Jack Cassin. I do not know him and I have no reason to believe I ever will. I look upon all ebay rug descriptions as being potentially full of mischaracterizations and errors with all of their inherent risk and make judgements about material presented on that venue based on my own understanding of rugs aided by whatever images the seller provides. That said, to me the Cassin bag features apparently natural dyes because nothing in the images shows me any evidence to the contrary. This is a matter of opinion that can only be argued fruitlessly in the absence of our ability to examine the piece in person. Second, the bag’s iconography is faithful to the tradition expressed in the analogous pieces I’ve already mentioned in my second post to this thread. By that I mean that the drawing of the borders and major and minor guls is not markedly different than those 19th century comparison pieces from the Jon Thompson and Wiedersberg collections. I have conceded that the ebay chuval suffers from some observable but (again in my opinion) not dramatic compression of the major guls. I believe I acknowledged that by saying the piece was probably produced later in the pre-commercial era. And last, I asserted that the subject bag falls neatly into one of Ned Long’s tekke small gul chuval classification groups and supported that assertion with references to several published and unpublished 19th century analogs from the same group.

Unlike most of the other participants in this discussion, I’m not willing to throw the baby out with the bathwater because of any personal antipathy I may harbor toward Jack Cassin. In fact, I harbor none. But I can understand how you and the others came by your antipathy and I also understand the group’s consensus issues with the errors in his description of the piece accompanying the auction. In fact, I pointed out the serious mistake Cassin made when he failed to note the most obviously comparable published analogs in his ebay listing.

Having said all that, my interest is in rugs and increasing my understanding and my appreciation of them. To that end, I fail to make progress in those aims by focusing on personal issues unrelated to rugs. If my attempt to separate a judgement of the rug from a judgement of the seller is upsetting to members of Turkotek, I apologize. But my own progress as a student of the art form will not be enhanced by a diversionary excursion into arguments over personalities, however disagreeable they may be. My only aim is to satisfy my desire to know more about rugs. The rest is irrelevant to me.

Posted by Jack Williams on 08-03-2007 08:58 PM:

line is about Cassin...see title.

Sorry Lee. I've looked further, following your advice.

1. I am having trouble understanding exactly your position about this chuval. Is it or is it not pre 1850 in your opinion? Yes or NO...?

Regardless of your citations, many hundreds of later examples seem to exist that are confirmed dated turn of century or later. You cite one or two that are ealier, but without posting the evidence. Most of us would probably agree that if that chuval were placed in a group of known mid 18th C. Turkmen weavings it would be the one declared recent. The burden of proof is for those who think it early, which is an intimate group of one at this point. You may be right...but....

2. This chuval was presented on eBay as having certain attributes. It does NOT have the attributes it was claimed to have....period. That is a fact. To say that...."well it has other attributes...." is to begin a different conversation entirely.

3. Whatever conventional wisdom is worth, most people who have discussed the characterization of Tekke weavings have listed the characteristics of this one as being common in later commercial weavings. I would have a hard time accepting those packed wefts, stunted designs and crowded conditions as being early Tekke.

4. Dyes are notoriously difficult to judge on line. I have been formost in noting that on this board. That said, the most common artificial dye used in Turkmen rugs is "red." The reds in this chuval look suspect to me. The "sheen" glossy pile, etc. you mention, I've long been accumstomed to make allowences for and I've bought a lot off the internet, more than most people I suspect.

There are so few other colors in the rug, basicaly only blue, that it would be hard to find another artificial color. But the red in this chuval does NOT have the look of other early Tekke weavings, to my eye. Compare the look of Cassins bag to the Thompson one. Again the burden of proof is on those who think it older.

5. If this line were simply about a weaving, I could reasonably say that opinions can be varied. It is not about that weaving. This line is about abuse, cursing, harassment, false claims, libelous statements, ad naseum, etc. The only reason this completely pedestrian bag is getting the attention it does not deserve is because it was so totally mischaracterized by Jack Cassin who has done some really terrible things to many people here.

I suggest you cease trying to turn the conversation into something it was never about. There are plenty of excellent Turkmen products posted from time to time on the regular lines that the attention of your expertise and academic nature would be most appreciated and could help us all. There is an interesting Ersari carpet in the show and tell section that I for one would welcome your attention.

Regards, Jack Williams

Posted by Steve Price on 08-03-2007 11:48 PM:

Hi Lee

You raise interesting and relevant questions about criteria for attributing dates to Turkmen pile weavings. Unfortunately, though, the questions are raised in a forum dedicated to examining the words and works of Jack Cassin. The title of the forum is You don't know Jack? Learn about him here, with a subtitle/description, Opinions and comments about, and quotations from, "The Mouth That Roared". Not for the squeamish.. Serious discussions of rugs, like the one you've introduced, have places on Turkotek but this forum isn't one of them.

I would be very happy to see your topic raised with one or a few examples (including the one Cassin just sold, if you like) on Show and Tell, with a somewhat more extensive introduction and a few examples as a Mini-Salon, with a fairly extensive introduction and at least 8 or 10 examples as a Salon, or without specific examples in Miscellaneous Topics.


Steve Price

Posted by Steve Price on 08-04-2007 09:44 AM:

Age Attribution: Tekke Juvals

Hi People

There's some discussion on another of our forums concerning age attribution of a subset of Tekke juvals. That topic is more appropriate to this forum

Here are two juvals. One is plate 30 from Mackie and Thompson's TURKMEN, the other was recently sold on eBay. There current or former ownership is irrelevant to this forum, their age and aesthetic qualities are relevant. More important, in my opinion, is the criteria on which their dating and aesthetics are judged. Most important (again, in my opionion), is the foundation on which those criteria rest.

Finally, a closer view of the second one.


Steve Price

Posted by matthias wohlgemuth on 08-04-2007 12:13 PM:

hi Lee,

i did go into my books and thought to let everybody know that the rare main border of the ebay chuval is to be found as well in: Jourdan: Oriental Rugs 5: Turkoman, pl. 153 (=yomut=Troost, pl. 21, a rather late piece) and in: Vanishing Jewels... (Amstey collection), pl. 28 (yomut? the Sienknecht eagle group attribution? a rather early piece).

when checking about my own chuval, which has tekke structure, fineness and colors like the ebay chuval (but "yomut" fastening marks at the top;-), i never noticed the existence of this very design in the main border of those two Tekke small rugs you mention. (congratulations for securing the ex-Pinner piece, i was after it, too, but was too slow after sale...) There seems to be some link with the well known "flag" border (Mackie: Turkmen..., figs. 83, 84) which basically features the same four triangle-lay-out (minus dot-details). Your thoughts would be of interest, your unbiased reasoning was a treat - even if it appeared in the wrong thread...

greetings from switzerland!

Posted by Steve Price on 08-04-2007 10:33 PM:

Hi People

I'm very skeptical about generalizations concerning age attribution of Turkmen weavings. I believe that the proliferation of borders and increased vertical knot density (resulting in flattening of nearly every design element) is pretty well documented as a response to commercial pressures to produce finer, more detailed rugs, and characterizes ca 1900 pieces.

Drawing the line from that to the notion that flattened motifs and high vertical knot density were exclusively late 19th century properties isn't so easy, though. In the absence of this, it's probably reasonable to consider it to be suggestive (but far from conclusive) evidence that pieces with high knot densities and squashed design elements are younger than analogous pieces with vertical/horizontal knot ratios closer to 1.0 than to 2.5.

The two juvals shown here appear to be analogous. They have the same major and minor guls, and skirts that can easily be imagined to be of the same genre as expressed at two fairly widely different times. If the above considerations are correct, the question becomes, which one is older, and by how much.

The "flattening" criterion clearly places the TURKMEN example at a much earlier date than the eBay example. The question of how reliable this criterion is remains unanswered. This, of course, doesn't lead to the conclusion that the opposite is true.

Aesthetically, I find the TURKMEN piece far superior to the eBay example. The flattened, crowded eBay juval doesn't suit my tastes at all. Having said that, I add that my preference for its characteristics is almost certainly an acquired taste associated with the conventional wisdom about Turkmen rugs.


Steve Price

Posted by James Blanchard on 08-06-2007 09:04 AM:

Hi all,

I had posted some general comments about Tekke chuvals, including the one recently sold by Mr. Cassin on EBay in another thread, but Filiberto thought it more relevant to post them here.

First, I would say at the outset that I think there are some positive features to the "Cassin Tekke chuval". The colours look good, and I find the border and elem interesting. Still, it is not really the type that I go for, particularly because of the crowding of the field with too many guls. The main problem with that is that this crowding of the field results in the absence of what I consider to be an important feature of a good Tekke chuval... a sense of "depth". If you compare these three examples (including Cassin's), I think many will agree that the other two examples give a much more satisfying sense of depth, with the perception that some elements "float" above the background. To me, this aesthetic issue is quite important, perhaps as important as colour, even if it is not a definitive indicator of age.

The first example is from Barry O'Connell's site and is described as being from the Jon Thompson collection and auction. The second was already shown in this thread, and referenced as "Thompson's". The third is Cassin's. Note also how the rectangular elements in the minor borders of Cassin's and the first Thompson are similar, but much more "squished" in Cassin's.

I think I like the first Thompson chuval the best, followed by the other one, but I like both of them a lot more than Cassin's.

I won't pretend to give a definitive assessment on the relative ages of these three, but in all of my reading and observations the older Tekke chuvals have a more even vertical:horizontal ratio, whether one estimates this by knots or by counting warps and wefts. To me, it is very obvious from the picture of the back of Cassin's rug that the knots are much flattened and therefore there are many more vertical knots than horizontal knots in an inch of weaving. As Filiberto has pointed out, in a double-wefted weaving such as this one, the ratio of wefts:warps is the same as the vertical:horizontal knot ratio, so the alternate way of describing the structure is a bit of a red herring in this context.

I am still very much in the learning mode but had am aware that this vertical:horizontal knotting ratio has been widely proposed as an age indicator for Turkmen weavings, with a higher ratio being a later feature. It has always been my impression that this high level of knot compression is a feature of somewhat later weaving practices when "fine knotting" was desired and achieved by vigorous hammering down on the wefts and knots that caused this vertical compression. In fact, in my other post I had asked if someone could provide at least one example of an "early" chuval with a compressed knotting ratio. Well, I think I might have found one myself. Below is a "Salor gul" Tekke chuval published by Jourdan (plate 63). Jourdan dates it to the "first half of the 19th century", and although he doesn't provide structural details mentions that "the much higher knot density along the warps than along the wefts gives the guls their flattened look". Does this example and Jourdan's commentary call into question the assertion made by Cassin and others about the vertical:horizontal knot ratio being an important indicator of early age? I would appreciate comments and seeing any other examples of presumably early examples that have a compressed ratio.


Posted by Steve Price on 08-06-2007 10:12 AM:

Hi James

One easy way to estimate the ratio of vertical to horizontal knot counts is to just look at the steepest diagonal lines in the designs. It is immediately obvious that they are much steeper in the Jon Thompson piece and the second one, which belongs to Jerry Thompson and was published in TURKMEN, which Jon Thompson co-authored.

There's no doubt that increased vertical knot density and corresponding flattening of motifs happened in the latter part of the 19th century, although this doesn't eliminate the possibility that it also occurred in some earlier pieces. I think Jourdan's attribution of the turreted gul juval to early 19th century is pretty aggressive. Judging from the photo I'd have guessed it to fall within the range of things usually attributed (rightly or wrongly) to the second half of the 19th century. The skirt design suggests earlier work, and if I were using this as my sole criterion for age attribution I'd place it between the two Thompson pieces and the Cassin juval.

There is a strong tendency to make age attributions by comparing pieces to published similar pieces in which the author, presumably expert, has made an attribution. The problem with this is that being an author doesn't make somebody an expert (although it's a good credential for an expert to have), and even experts in Rugdom often simply apply conventional wisdom without knowing or challenging its foundation. The result is that some errors become embedded in the literature by repetitive use. The challenge is to root out the bases of various criteria to find out which ones are on firm footing.


Steve Price

Posted by Steve Price on 08-07-2007 08:54 AM:

Hi James

It's the pieces that send mixed messages that present the most difficult attribution problems. What this means in practical terms is that we have to bear in mind that every attribution is only a probability statement and that some are more likely to be correct than others.

It's not unusual to see someone using flattening of motifs, profusion of details and increased ratio of vertical to horizontal knots as though those were three criteria and, therefore, should count three times in tallying up evidence. In fact, all three are the same thing.

In the case of the juvals presented here, there are also what appear to be differences in color (can't really be sure by looking at a computer monitor), but I don't think they tell much about age except to suggest that all are done with natural dyes. After that, we're left with flattening and general spaciousness. The eBay juval finishes third in its group on both criteria, the other two are approximately tied in flattening; the Jon Thompson piece is clearly the most spacious. If we adopt the often used approach that says, more or less, "I like old rugs, so the ones I like the most must be older than the ones I like less" (and I'm not defending that approach, just pointing out that it's widely used), I come out with the Jon Thompson piece as best (therefore, oldest ), the TURKMEN piece next, and the eBay piece as youngest. Absolute ages? My guess: Jon Thompson's and the TURKMEN piece, 1850 or earlier; the eBay piece, last quarter of the 19th century.

I'd treat the turreted gul juval separately, since it's not really analogous to the others. The flattening criterion suggests late production, the absence of silk does, too. Just about every turreted gul juval I've seen with an early attribution (pre-1850, rightly or wrongly) has lots of silk. The design in the skirt is different than most late 19th century Tekke juvals, but that would still be my best guess of when it was made.


Steve Price

Posted by Itzik Mordekhai on 08-07-2007 12:09 PM:

Perceived age of rugs

Hi all,

I am following this thread with great interest and a nagging question comes to mind.

It is generally accepted that the bolder the design and the lesser the number of borders, the earlier is the rug. Can someone please enlighten me as to the reason, or reasons, behind this well-established concept? How is it that the weavers of the earlier rugs KNEW how to make them with a bold design while later weavers FORGOT this age-related feature?

My second, seemingly paradoxical, question that comes into mind is: How is it that with some rugs (18th-19th century) this formula works (the bolder the earlier) while with other rugs, for instance 16th century Safavid carpets, with their seemingly very crowded design and a high number of borders, the reality seems to suggest that the more crowded the design the earlier the carpet?

I am curious to know what you think!


Posted by Steve Price on 08-07-2007 12:29 PM:

Hi Itzik

I hope you don't mind - I moved your post into this thread rather than leaving it as a new thread. It seems to belong here.

The tendency for vertical knot density to increase during the second half of the 19th century isn't a general phenomenon, but a specific one in Turkmen pile weaving. The usual explanation, which makes sense to me and for which I believe there is some documentation is this: Around 1850 or so, European rug dealers and merchants sort of discovered that Turkmen made high quality things, and began buying from Turkmen sources. Their clientele held the notion (still current among many rug buyers today) that finer, more detailed designs were more desirable than simpler, more sparsely decorated weavings. So the merchants encouraged their Turkmen suppliers to get them the pieces with the most decoration in their designs and layouts. The demand for greater detail and fineness of weave was transmitted to the weavers, and this commercial pressure resulted in their increasing vertical knot density, numbers of borders, and generally leaving less open space.

The result? Turkmen weavings, which tended to be spacious and to have vertical knot densities only a little higher than the horizontal knot densities (at least, for Tekke during the period from about 1800 to about 1850), fairly rapidly evolved to the easily recognizable characteristics that we associate with circa 1900 Tekke work.


Steve Price

Posted by Itzik Mordekhai on 08-08-2007 03:48 AM:

Boldness across cultures and eras

Hi Steve,

I understand your explanation about the apparent degenaration of the Turkoman rug, mainly the Tekke, over the years.

As to the second part of my question, I still don't fully understand how some 16th century carpets, mainly Persian, are still being considered, and justly so, masterpieces of design and coloration even though they look very busy and crowded vis a vis the current understanding that bold and spaciousness spell early age.
Is this a phenomenon limited to some Persian carpets? I know the earliest Anatolian rugs were quite bold(Seljuk rugs) and so were the 16th-17th century Chinese rugs.

I think what I'm really asking is whether in some cultures beautiful and desirable means crowded and in other cultures it means bold and spacious.

What is your opinion on this?


Posted by Steve Price on 08-08-2007 06:01 AM:

Hi Itzik

I think the easiest explanation for why some genres are considered more beautiful if more detailed, others if more sparse, is, as Winnie-the-Pooh says so often, THE WAY THINGS ARE (I'm not shouting, he always speaks in uppercase letters in Milne's books). Collectors apply different aesthetic criteria to different kinds of rugs. A Shirvan or Kuba is preferably fine and with lots of detail, a Kazak is preferably sparse and open, for example. We make similar aesthetic distinctions in all sorts of art, it isn't peculiar to rugs.

I don't think anyone knows what the weaver's culture thought was most beautiful, especially with early tribal weavings, which included visual properties that were believed to be functional (distraction to the Evil Eye, for example). The notion that our aesthetic preferences are the same as those of the long-gone cultures that made the pieces is romantic, but not supported by much evidence.


Steve Price

Posted by Lee Koch on 08-19-2007 09:04 PM:

knot counts in tekke chuvals

Well, I've just returned from a two week vacation and am happy to report that there's life beyond the world wide web. I noticed this thread and wanted to add a coda to what I'd already said in the Cassin thread.

I own ten tekke small gul chuvals. Nine are clearly without synthetic dyes and the tenth is clearly not. Just for fun, I pulled them out this evening and did a quick knot count of each. I counted the knots in the center of a primary gul on every chuval so as to sample in similar areas on all of them. Here are the results with the horizontal count given first: 10x21, 12x20, 10x20, 10x20, 12x29, 10x17, 12x24, 11x19, 11x19, and 11x20. Now would anyone care to guess which knot count correlates with the later piece?

If you guessed the 12x29 knot count, you'd be correct insofar as the conventional wisdom goes but incorrect in the real world. In fact, the late piece is the one I gave first, 10x21. The piece with the 12x29 knot is the bag that I mentioned in the other thread that I bought from Marvin, and anyone who knows Marvin knows he would never have acquired a late turkmen bag while he was forming his collection of central asian material. It displays four vertical columns of five primary guls and its elem is decorated with the the same floral or tree figure as in plate 9 of Tent Band Tent Bag and two of the Rippon Boswell Pinner sale tekke chuvals. I'd cite the lot numbers of the Pinner pieces but can't put my hands on the sale catalogue at the moment.

As for the chuval with the synthetic dye, it shows a field arrangement of 5 columns of 5 primary guls and traditionally correct border systems. So conventional wisdom about the knot count and the iconography led me to conclude that it should have been "right" from an internet image when in fact, it wasn't.

I suppose that the moral of this story is that relying on generalizations and conclusions drawn by authors in 20 year old rug books is not as reliable as learning from a pile of rugs in your lap.

Rug books are a good way to begin to gain some familiarity with attribution and iconography but they'll always be a poor substitute for the kind of "boots on the ground" intel that comes from handling the material in real time.

Posted by James Blanchard on 08-19-2007 09:55 PM:

Hi Lee,

I agree that your approach is exactly the type of data analysis that is too often lacking in rugdom, even if the sample size is only 10.

Now you have raised a few new questions for me. First, what would you estimate is the age range of the 10 chuvals you have examined? Would you estimate that they are clustered within say 25-50 years, or do you think they might be generations apart? I think that "synthetic" vs. "non-synthetic" dye dichotomy results in a rather wide time range on the "non-synthetic" end (ranging from the late 19th century all the way back to the 18th century and perhaps earlier). Second, if the vertical:horizontal knotting ratio is not a valid criterion for assessing the relative age of these chuvals, what other criteria would you recommend as being more valid? Third, given your own experience and collection, where would you place the three chuvals illustrated in this discussion in the age continuum (see my earlier post)?

Finally, I for one would find it very interesting and educational to see more examples of this type of chuval. Would you consider posting pictures of some or all of your chuvals to that end?



P.S. I appreciate your admonition to not rely too much on rug books, with the implicit message being that those who handle and study rugs are in a better position to comment on these issues. What puzzles me, as a rug novice, is that different experienced folks often seem to have contradictory experiences. A large portion of this discussion emanated from the fact that Mr. Cassin, who has reportedly handled a lot of old Turkmen weavings, explicitly mentioned the lower vertical:horizontal knotting ratio as a sign of earlier weaving. This contradicts your experience, so what is one to think?

Posted by Steve Price on 08-19-2007 10:08 PM:

Hi Lee

Welcome back. Date attribution of Turkmen material is always subject to a good deal of uncertainty, and I wonder if we wouldn't be better served by just acknowledging it and living with that as a fact of life.

Nobody - not me, not you, not anyone else - really knows the ages of any of your 10 juvals, although if the obvious synthetics really are synthetic dyes, that one can't have been made before 1860. Can we know with reasonable certainty that any of the other nine predate 1860, or that the one with synthetics is much later than 1860? I doubt it, although Tekke juvals typical of those woven around 1900 are not hard to recognize, nor are those of the Soviet era.

Ratio of vertical to horizontal knots is an indicator, but only an indicator, that's associated with earlier Tekke things when it's around 1.2 to 1.5, associated with circa 1900 pieces when it's over 2.0, but not definitive evidence for age in any case. Of your 10 pieces, two have vertical to horizontal knot ratios over 2.0, and one of those also has synthetic dyes. None of the other 8 have ratios that suggest earlier dates, although they don't eliminate that possibility all by themselves.

I have the greatest respect for Marvin and his taste, but the fact that he owned one of them isn't very compelling evidence that it was woven early. Even without seeing it, I feel confident that it is beautiful, though.

You're right, of course, that authors of 20 year old rug books can be mistaken. So can authors of rug books written yesterday. In fact, they will have mostly the same mistakes in them, because the older books are their sources of information. I disagree with the notion that having a lap full of rugs (say, 10 Tekke juvals) makes it easier for you (or anyone else) to learn how to make accurate age attributions, though. In the absence of a reasonable database of specimens of documented ages, the criteria we learn from our examples hardly rise above the I like old rugs. I like this one more than that one. Therefore, this one is older. paradigm.


Steve Price

Posted by Richard Larkin on 08-20-2007 08:02 AM:

Hi Folks,

Another variable we have to keep in mind is that we can't be sure the ten juvals we have in our lap can be effectively compared with one another meaningfully in order to estimate age simply because they are Tekke. I doubt the Tekke were a monolithic weaving bloc, producing homogeneously across the board at any given time. Geography and other considerations must have figured into the mix. It's very possible that Lee's juval with the synthetic dye was woven before one or more of the others, in another place; or during the same year, in another tent. Similarly, indicators that are apparently age related (e. g., vertical/horizontal knot ratio) were probably persistent in specific areas against the developing norm, perhaps among more isolated groups.

Rich Larkin