I am happy that Central Asia carpets are the topic of John Howe Salon.
One of the reasons for that is that, few years ago, I bought one main carpet that I believe belongs to the Central Asian production. I thought that adding another example of this rare group of carpets can be useful for the overall discussion.

The carpet is quite large (about 2m x 3.40m), it is in good conditions, the sides have been remade but it retains its original ends/kilim. One particular structure is that the warps are in cotton. The warp is made by 5 strands of cotton. According with Murray Eiland book, this feature may indicate that the warp is hand made.
The field of my carpets contains rows of rosettes that are organized diagonally for almost all part of the field. The range and the richness of the colors of this carpet still impress me (see photo 3). The border design of my carpet is similar to the border of carpet 4 of Howe's Salon.

I have no exact idea about where this carpet comes from and what is its age.
My feeling, based on the state of conservation and the wide range of colors of this carpet, is that it was made at the beginning of 1900. A well-known Italian dealer-scholar that professionally cleaned this carpet, though, believes that the all the colors are natural.

The (bad) photos I posted were taken before that carpet was washed and I did not alterate them.

References:

1) Richard Purdon Carpet, Hali 84, page 130 (1.98m x 3.08m), attributed to Kazakh, Central Asia, circa 1850.
For field design, color palette and size, the Purdon carpet is very close to my piece.

2) Galerie Koller Zurich, sale on 13 Sept 1986, adv. in Hali 31, back cover, (2.10mx 3.40m), attributed to Uzbek. The elements between the rosettes are similar to those in my piece.

3) Lot 71 Sotheby's NY, 14 Dec1995, (1.37mx 2.97m), Central Asian Carpet, early 19th century.
The entry for this carpet refers alos to lot 42 of the Jon Thompson sale in Sotheby's NY, 16 Dec. 1993, Yomud Turkmen Carpet, 18th century.

Finally, there is an article of George O'Bannon in Hali 89, pages 74-75 that discusses the SNY and Purdon pieces.

Thanks,

Guido

Thanks for putting up this interesting rug.

I have no particular expertise about such pieces but have consulted the references you cited and have been considering things a bit.

First, I think it is quite unusual for non-Turkmen Central Asian rugs to have cotton warps. And the fact that this piece has them made me think of East Turkestan, since a number of rugs from that area have cotton warps.

On the other hand, I looked through Hans Bidder's book "Carpets from East Turkestan" didn't see anything really close. If you have a copy look at Plate XIV, 1 and at plate XIX. The major field element in plate XIX is close outline of that in yours. (Bidder provides an interesting set of line drawings on page 69 where he conjectures about how devices like your main "gul" may have developed from Chinese forms.)

Another thing I noticed about your piece is that while there are instances of the diagonal use of color in both the borders and in the field,(O'Bannon treats that as an indicator of Turkmen Uzbek weaving) the use in the field is inconsistent. So I'm not sure we could lean heavily on diagonal use of color if we wanted to argue that this piece might be Turkmen Uzbek, which, as you know, is what Bannon thought was the most likely attribution for the Purdon rug and the Sotheby's piece.

The minor field element in your piece is different from those on either the Purdon or the Sotheby's rugs but is among those listed under Turkmen Uzbeks in O'Bannon's translation of and commentary on Moskova. Versions of it, of course, occur widely in Turkic weaving.

I have sent Jim Blackmon a link to your post and have encouraged him to comment if he feels so moved. We'll see what happens.

Thanks again for letting us see this lovely, colorful piece.

Regards,

R. John Howe

You did it again, what an impressive carpet.

When I see these rosettes or "roundrels" (as Ellis called them), one of the first impressions is of a carpet adapted and modified from Persian prototypes. But it even evokes old Caucasian traditions such as early "Afshan" or "Avshan."

Upon further reflection, I think it plausible that the model for these roundrels or rosettes goes back much further to earlier brocaded silks. Possibly your carpet is a Central Asian adaptation of such textiles? Brocaded silks and related textiles seem to have provided models for carpet weavers in many places. Given the trade along the silk route, it seems plausible that local weavers could have adapted such textiles into a carpet design.

I think this adaptation is quite successful.

You mention the Thompson Yomud carpet only in passing but the relationship probably helps put your carpet in context. That carpet was lot 42 in the December 1993 sale at Sotheby's. (I hope someone can post an image of it into this thread.) I find the drawing of the rosettes or roundrels much more refined in your carpet than in the Thompson carpet though I think it plausible that they are adaptations of the same thing.

For the record, the Thompson carpet was 2.85 m x 1.30 m and had wool warp, Z2S and wool wefts, 2Z, 2 shoots. Oddly, the left half had symmetrical knots and the right half had asymmetrical knots open to the right. In describing his carpet, Thompson wrote: "unlike a number of Yomud carpet designs which were adapted and modified from Persian prototypes, the field design of this carpet could be described as mainstream. It shares its secondary ornament with a number of other tribes and the main design has a generic relationship both in shape and internal structure to other Turkmen motifs, even though its exact form occurs on no other piece."

As stated above, we may be able to go even beyond the Persian prototypes Thompson refers to and look at earlier brocaded silks to find the prototype of these adaptations.

Thanks for posting the image and your references to related pieces.

Best, Michael Wendorf

I agree with you that the rosettes of these beautiful and interesting carpets are very very near to one Afshan-element and I’m curious for the reason to abandon this track so readily.

Regards,
Christoph

Here is the image of the Thompson Yomut main carpet from his Sotheby's sale, that Michael Wendorf wished for.

Regards,

R. John Howe

I do not wish to abandon the idea that the multi-foiled rosettes or roundrels on these carpets could be related to the very similar device found as part of the all over floral Avshan pattern. The drawing is so close as to make it difficult to abandon this thought.

On the other hand, what is the source of the rosette/roundrel device found within this Avshan pattern and why do we not see more of the pattern suggested in either Guido's or the Purdon carpet? Put another way, why only one element?

I have no easy answer to this question. Charles Grant Ellis, for one, seems to have concluded that the Avshan pattern itself was derived from brocaded textiles and that this idea is evidenced by the appearance of the Avshan design in carpets made over many areas - even India. In Oriental Carpets, Philadelphia Museum of Art he observed that "thin textiles, light in weight and highly portable, can be moved for great distances far more rapidly and easily than carpets and, if their designs prove fashionable, offer ready models for weavers." See page 148. If this is so, then it is plausible that the Avshan pattern and the pattern on these carpets could have had a similar prototype yet be completely distinct adaptations by weavers with completely distinct weaving traditions.

Adding the unusual Thompson Yomud, one might even go further and infer that many Turkmen guls have have their origins in similar but equally distinct adaptations of roundrels/rosettes found on earlier brocades.

What is your thinking on these questions?

Best, Michael

You asked: Why only one element? If you look for example at Eiland’s articles about the development of the Harhang (HALI 4/4 p. 338ff) and the Afshan design (HALI 104 p. 59) you see that while we’re always speaking of ‘development’ ‘deterioration’ or ‘disintegration’ often would be more accurate.
By the way look at Fig. 7 of the above mentioned article about the Afshan design (HALI 104, p. 62). If you replace the secondary ornaments and turn the secondary border into the main border you will be very near to Guido’s carpet.
I agree with you that textiles are a likely inspiration for this kind of patterns (and some others too, including some Vase- and Dragon carpets) but I think we shouldn’t “go even beyond the Persian prototypes“ (what does this mean anyway?) and search for Safavid and/or Mughal models.

Best regards,
Christoph

We've debated (not you and I but the group generally) here on Turkotek what goes on in this kind of design analysis several times before.

As far as I can tell there are two broad tendencies. The first, is comparison of similarities in design. Pretty pure taxonomy and entirely unobjectionable.

But often there is a tendency to slide into a language of development without some demonstation actual developmental connection. I have sometimes used the word "illicit" about these latter inferences to the consternation of some my friends who are lawyers. (It may be a point on their side that, Steve, who is trained as a biologist and so is familiar with such argments in his own field does not, I think, find such "evolutionary" language objectionable.)

In any event, it has become clear to me that some enjoy this kind of analysis and conjecture so much that I have largely given up criticizing "genetic-seeming" descriptions when they occur.

But I too think "proto-type" is a term we are not in position to use unless the intent is purely descriptive and doesn't carry any implications of "development."

But it's an old debate and it has not gone anywhere in particular when we've taken it up, excepting that some feelings have occasionally been momentarily bruised.

Regards,

R. John Howe

Fair point and the example illustrated by Eiland in Hali 104, page 62 is as you describe it. Indeed, Eiland does argue that a case can be made for the Avshan design rapidly evolving from Caucasian and Persian roots in the 19th century in a predictable manner -- that is a steady simplification and a reduction of the number of design elements. It does not take much effort to imagine that some of these carpets might have reached Central Asia and been considered fashionable enough to imitate. But this is only one case that can be made, there are others too.

In this regard, it is always difficult to draw any hard conclusions from a handful of carpets. Though Eiland's observations are keen, they are used to support his own Caucasian/Persiancentric theories. He goes to fairly considerable lengths to diminish the importance of the Keir fragment, illustration 1, page 58, presumably because it is Indian, not Persian or Caucasian. Of course, Ellis thought this fragment (another or closely related piece is in the Textile Museum)to be an earlier and purer design than any of the Caucasian examples. And while the process of simplification seems to generally be true, you and I could both reference any number of late 19th or even 20th century carpets that faithfully reproduce the entire Avshan pattern. A number of Bijars come immediately to mind.

My point, I think, is that despite the apparent relationship between one of the classic Avshan design elements and the design of these two carpets, we do not have to assume that these Central Asian weavers were adapting or simplifying this pattern. It could also be that they were adapting or interpreting the same or similar designs that originally begat the Avshan pattern quite separate and distinct from what weavers in the Caucasus and Persia were doing. In this regard, I think we are in agreement that there are brocades that could well have served as a design source of all of these adaptations. If this is plausible, then it seems to me we can and probably should at least consider these models and not content ourselves to rely on what we are describing as the Persian or Caucasian prototypes.

John, if it makes you feel better, you may make now a list, a checklist and a flowchart of these various design models so that we conduct a proper evaluation.

Best, Michael

No, I'm glad to assist in providing a concrete image for your analysis, which I acknowledge is congruent with the majority of the rug literature that has been written to date.

I am in my administrative capacity making one small revision in your post above.
You put a page reference in paretheses the end of which was an "8" followed by a "closed paren." I learned in a recent post of my own that these two ending characters in that order are apparently the code for one of our "emotions" faces, which I think you did not intend. I have dropped the parentheses and the little face has gone away.

That's something folks might generally want to know: "8" followed by ")" apparently produces an emotions face. This one

Regards,

R. John Howe

You are correct. Although sometimes sentimental about carpets and weaving traditions, I made my post without any emotion at all.

Thank you for the Salon, Michael