Posted by Filiberto Boncompagni on 01-05-2002 11:44 PM:

Of Carpets and Paintings

Dear All,

There are a lot of question marks on the origins of the Caucasian design or even of the whole Oriental Rug design…
I have come to the thinking that Gantzhorn is fundamentally right in attributing the paternity of it to the Armenians - although he did it in a very fundamentalist way and he left out of the equation another very important player: the Kurd people. But I don’t want to go through that, at least not on this thread.
Let’s just only assume that Armenians are part of the Caucasian weavers and there is historical evidence that they were making rugs since at least 2000 years.
The fact is that no specimens of those ancient textiles survived (besides some "dragon" rugs), so we have to turn to evidence in ancient pictorial art.

Early Armenian miniatures show some rugs, but with insufficient detail. European paintings from the early 15th century show a lot of paintings with geometric design, but "the majority of scholars take the view"…that they "are most certainly Anatolian and not Caucasian; this view, of course, does not rule out the possibility that they may very well be the work of Armenian weavers…Thus it may be that a good number of the weaving seen in 15th century European pictures, while Anatolian geographically, are the work of Armenians, who inhabited a large area of eastern Anatolia." (Ian Bennett, in Oriental Rugs - Caucasian, page 9).

Now let’s talk about paintings and para-Mamluks rugs. Before that, to get some familiarity with the subject, I advise you to visit one of Wendel Swan’s articles on our web site:

Looking in my vast collection of Hali issues (two numbers ), I found an interesting essay by John Mills: "The Chihl Sutun Para-Mamluk Prayer Rug" (Hali # 93).

In short, this rug

which, incidentally, appears also on Ganzhorn’s book (plate 328), was seen by some scholars (Erdman and Ellis) "as a late 19th century Turkish product imitating earlier designs". That is, until Mr. Mills discovered the similarities between it and the lower part of a carpet appearing in a painting of an Italian Artist, dated 1501.

It is plausible to think, given the resemblance of the visible part of the rug in the painting with the real one, that "the painting shows an exact reproduction of the lower part of a 15th century rug resembling the Chihil Sutun rug, which may itself reasonably be accepted as of similar date."
Thus upgrading it at once with 4 additional centuries - such are the winds of changes in rug scholarship!
Notice the composition with the field divided in two compartments by the two "minbar" (minbar: a pulpit that is entered by a flight of steps and stands next to the mihrab). Mills remarks that: "Is the minbar in fact the origin of the hooked (stepped?) triangular corner pieces often seen on early prayer rugs?"

Anyway, I was looking in my books on paintings, searching for possible examples of early Caucasian rug design, when I found this:

detail of the Witchcraft (Allegory of Hercules), by Dosso Dossi,
c.1535 Galleria degli Uffizi, Florence

By the way, if you are interested in old paintings there is a wonderful web site, a huge virtual gallery, at this address:

This link will take you to a page with the whole picture of the above painting and some information about it:

Clicking on the picture you’ll be able to see it in full size and even to zoom in and out.

At first sight I thought it was another Para-Mamluk prayer rug with a coarse make (village production?). Like the other, it has a Kufic-sque border (mostly hidden, but recognizable). The field is divided in two sections by two triangular devices (hooked or stepped?), the lower occupied by an octagonal medallion. The medallion border and the rug secondary border appears to be the same - you have to trust me on that: the picture on my book, albeit small, is clearer than the one you see here. The aforesaid border is in the shadow, so one can distinguish its colors only when it departs in full light to form a good, old CAUCASIAN PRAYER ARCH!

Kaffel: "There are almost no documented early (pre -1800) Caucasian prayer rugs" (page 8). On page 22: "…former curator of textiles at the De Young Museum in San Francisco, conjectures that ‘as no Caucasian Prayer rugs have been discovered that pre-date the 19th century, this style of weaving could represent a late development’. It is, however, difficult to accept such a premise". Right. Please, look at this from Kaffel’s book:

Karabagh prayer rug, dated 1807. Kaffel: " Various provenances have been suggested for these pieces…". "However, it is most probable that this group of rugs…is the work of a single workshop of Armenian weavers from Karabagh. Although the design is related to the 17th and 18th centuries embroideries of Azerbaijan, this design has no clear antecedents among Caucasian pile weaving". But he also noted "an Anatolian influence, specifically the 16th century Holbein design group was suggested".

In my humble opinion, this design is a derivation from the one depicted in Dosso’s painting in 1532. Which is an ARMENIAN or CAUCASIAN (as you like) interpretation of an almost contemporary design illustrated in Hali’s article (before breaking the news to the world of this astonishing discovery I prefer to wait for your opinions, though).
I also suggest that the rug on Gantzhorn’s plate 695

attributed to the end of 15th century, showing a very similar prayer arch border to the one of Dossi, has a lot of Caucasian features too.

Well, that’s all for now.

- Gantzhorn’s book : "The Christian Oriental Carpet" by V. Gantzhorn, Cologne1991
- Kaffel’s book: "Caucasian Prayer Rugs" by Ralph Kaffel, London 1998.

Posted by Marvin Amstey on 01-06-2002 12:13 PM:

Good morning (afternoon), Filiberto. I humbly disagree with the origins of the Kaffel rug and the re-entrant carpet shown above. The latter has all the earmarks of Anatolian work - albeit probably by Armenian weavers. The Kaffel rug is very similar to a rug that was advertised in Hali nearly 20 years ago by Gustave Tasch from Toronto. I had the opportunity to handle that rug, and I agreed with his deduction that it was 18th c. Anatolian - probably from the west or southwest, i.e. Bergama-like. In fact, I believe that most Caucasian design came from Anatolia simply as a progression of rug design from East to West - or in this case East to Northwest.
Best regards,

Posted by Sophia Gates on 01-06-2002 12:45 PM:

Just a thought - probably not original - I'm wondering if the Kufic borders might have derived from caligraphy?

Posted by Filiberto Boncompagni on 01-06-2002 01:02 PM:

Good evening (morning) Marvin,

Forgive my ignorance: what are exactly the earmarks of Anatolian work?


Posted by Steve Price on 01-06-2002 01:21 PM:

Hi Sophia,

I recall reading the suggestion that Kufic borders originated with calligraphy (one of the incredible art forms of the Moslem world, but that's another topic for another day), but I don't remember where I read it or who wrote it.


Steve Price

Posted by Marvin Amstey on 01-06-2002 04:14 PM:

Hi Filiberto,
I would say that the re-entrant design, the widely spaced, 8-pointed stars, the wide meander on a yellow ground as a narrow border, the hanging plants (pomegrantes?), the red ground, the green mirhab outline; all these say Anatolian. I wouldn't be surprised if the wefts are red wool and from 2-4 per row of knots. Lastly, the multicord selvedge also fits that picture.
Best regards,

Posted by Patrick Weiler on 01-06-2002 05:55 PM:

Kufic Script


I believe there is little doubt that the origin of the Kufic borders is the Kufic calligraphic script. Specifically Knotted Kufic. You will find a link here that shows the different styles of Kufic:

Kufic writing is said to have been developed and achieved its peak of perfection in the 8th century. It is named after the town of Kufah, where it was first used in official documents.
The question of when it began being used on rugs is open to debate. It was often used as writing in borders. Numerous old rugs show it in cartouches. There is such a rug in the Gantzhorn book, Illustration 21. It is a "Moslem prayer rug" with Kufic script in square cartouches in the border. There are over 30 examples of the "square Kufic" script in cartouches in the borders of similar prayer rugs shown in the book of The Topkapi Saray Museum Carpets.

Seven of the first eight rugs in the Ellis book Oriental Carpets of the Philadelphia Museum of Art have "Kufesque" borders.
Filiberto may be interested in figure 26b of this book. It shows an Anatolian re-entrant prayer rug, similar to the one shown above, with a "Kufesque" border.

The Kufic border became stylized by either purposely being transformed into a repeated non-sensical design, or by being copied by illiterate weavers/cartoonists as a geometric design which has subsequently been used for many years, particularly in Caucasian and Turkish rugs. The Kufic script is certainly unintelligible to readers of Western writing.

Gantzhorn, contrarily, says that the Kufic lettering source of this type of border "proves to be without support." and that it is actually derived from Christian sources. He generates "the suspicion that the influence may possibly have run in the opposite direction."
Gantzhorn, in his inimitable convoluted fashion, says that one rug with Kufic inscriptions, Illustration 16, must have been produced in the 16th century by Armenians who had converted to Islam.
I have a personal theory that the Kufic border even developed further in tribal areas and will try to post a couple of photos in the next day or so.

Patrick Weiler

Posted by Filiberto Boncompagni on 01-07-2002 05:09 AM:

Dear Marvin,

Well, well. You say "I believe that most Caucasian design came from Anatolia". That is a step forward in trying, as Daniel asks, "to determine what can be considered as typically Caucasian in both structure and design".
Ancient Armenia was between Anatolia and Caucasus. Armenians wove Anatolian and Caucasian rugs - although in the first case rugs were labeled only Anatolian. The only early Armenians - labeled as such -we know are the dragons and palmettes rugs.
Next question is: what are the differences in structure between dragon rugs and Anatolian rugs?
This is not a rhetoric question. I really don’t know.

Sophia, you are right. The use of calligraphic ornamentation is a dominant feature of Islamic art since its early days. By the way, the script in the spandrels of the Chihl Sutun Prayer Rug is Kufic.


Posted by Patrick Weiler on 01-07-2002 09:52 AM:

Whither Kufic

Dear Kuficcianado's,

The Kufic border consists of a rosette surrounded by brackets.
I have long seen a resemblance between the Kufic, or Kufesque, border:

and this more geometrically severe or simple Persian tribal version:

You will note that the tribal version has taken the basic components and reduced them to their bare minimums.

It is likely that this theory will be contentious. I do not have years of rigorous research showing the process of design transmogrification, but I am willing to withstand the withering heckling, the humiliating villification, even the elimination of potential tenure at a major university research center to bring this groundbreaking, earthshattering speculation to YOU, the viewers of Turkotek.

A possible transfer mechanism could be related to the theory that the Qashqai migrated south from the Caucasus hundreds of years ago.

There are no hidden animals or religious inferences in this design change, just a common reduction in complexity as the design transferred from urban to the tribal milieu.

Transmogrifically yours,

Patrick Weiler

Posted by Marvin Amstey on 01-07-2002 05:19 PM:

Hi Filiberto,
I did a little searching to see if I could find any differences between dragon carpets and 15-17th c Anatolian carpets as you asked. The best I could find in terms of differences is that the selveges of bird and lotto rugs are described as 4-cord by Charles Ellis, and the dragon rugs as 1 or 2-cord selvedges. These analyses are from rugs in the Philadelphia Museum and others mentioned in that book that he reviewed. As far as wefting (number of shoots), weft color, kilim end finishes, and knot density are concerned, there are no differences. The use of color, however, is different. The dragon rugs are the usual Caucasian red, white and blue. The Anatolian carpets are predominately red - except for the white-ground bird and Chintamani rugs - with much more yellow in the designs. This is not to say that dragon rugs don't have yellow; they do, but used differently. In summary, there is precious little difference, and my guess is that they were all made by Armenian weavers.
Best regards,

Posted by Steve Price on 01-07-2002 05:30 PM:

Hi Marvin,

Am I correct in assuming that when you say that the 15th-17th century Anatolian carpets under discussion were probably made by Armenian weavers, you mean ethnic Armenians living in Turkey? People often forget how many Armenians there were in Turkey before the 20th century, and how prominent some were in the world of Turkish rugs.


Steve Price

Posted by Marvin Amstey on 01-07-2002 05:38 PM:

Yes; Armenians in Turkey and Armenians in the Caucasus. My old friend and mentor, Vartan Deverian, is convinced that all Turkish rugs west of Konya, and all Caucasian rugs were made by Armenians. While he "hates" Turks, he won't sell any of his Turkish rugs, unless they were made by kurds, because of their "heritage".
Best regadrs,

Posted by Sophia Gates on 01-07-2002 08:17 PM:

Given the multiplicity of ethnic groups in Anatolia and the Caucasus, and the probability of overlapping influences in design and even technique, doesn't it seem a bit of a stretch to attribute the whole of the huge production in these areas to just one group? Especially since virtually all of the tribal and village peoples in the regions are known to have been gifted weavers? Or are we limiting the discussion only to commercial rugs?

Even it that were the case, I'd still be surprised if one could make such an attribution - if you'll forgive me for a Patrick-style "contentious" simile - that would be like attributing all Renaissance painting to Florence.

As for Patrick's conclusions about the possible geometrization of Kufic borders in tribal art, I don't see why that couldn't be possible. On the other hand, if one goes back far enough it can be seen that the script in fact evolved from the angular symbols of ancient calligraphy - which had originally been carved into stone. And further - what of the possible derivation of these positive-negative geometric forms, as shown in the tribal example Patrick has provided, from even older kilim designs?

I have a theory that - like other human endeavors - art is limited by two variables: what is necessary, and what is possible. In the visual arts the "necessary" could include the function of what one is designing - be it building or blanket - and theoretically one would also include whatever iconography is to be carried with the design. The "possible" of course are the limiting factors of the materials and technology employed. Stone will behave one way, warp and weft another, and while it is of course possible to make stone look like silk and wool like watercolor, a great deal of modernist thought revolves around the ideal of form married to function - as well as honesty in the use of materials. I think the brute force many of us respond to in the carpets and kilims of Northwest Asia are directly related to the immediacy and simplicity of their form: geometric designs rendered in a straightforward manner - or, in the case of dragon rugs and "eagle Kazaks", for example , formerly ornate and curvilinear ideas rendered more directly. In any case it seems to me that the majority of design ideas both in Anatolian and Caucasian carpets possess a great deal of universality and may have been - I believe probably were - derived from a variety of sources by a variety of people.

Fire when ready!

Posted by Filiberto Boncompagni on 01-08-2002 05:07 AM:

Yes, not all Renaissance paintings come from Florence - where Renaissance started - only the best ones!
Joking apart and before we sink I an ethnic-political discussion, I’d like to compare art, especially folk art to a language.
Take modern Italian, for example. It evolved from Latin, but Latin had a lot of influences from Greek and probably other precedent languages. So, in today Italian most of the words come from Latin first, then Greek, then all other languages of people we came in contact with wars, invasions, commercial exchanges…Like: Arabic, other neo-Latin languages, German, English, whatever.
Languages are in constant evolution, but have very defined origins. In my humble opinion, the Anatolian and Caucasian (rug) languages must have a common origin from the oldest ethnic groups living in the Anatolian-Caucasian region. The ones coming to mind are Armenians and Kurds, for we know they lived there for thousands of years and had a strong tradition of rugs weaving. Of course, then the language evolved coming in contact with a lot of other ethnic groups AND with the influence of city-court workshops, new fashions, commercial influences and so on.
It’s that making sense?


Posted by Steve Price on 01-08-2002 06:12 AM:

Hi Filiberto,

I like the linguistic analogy, which is a very useful model for all kinds of evolutions. Darwin used language evolution as an analogue for biological evolution, with considerable success in persuading his peers.


Steve Price

Posted by Filiberto Boncompagni on 01-08-2002 10:35 AM:

Thanks Steve,
Now, let’s resume.
As you see above we have some examples of early Anatolian rugs, probably woven by Armenians that - at least to my eyes - have some Caucasian elements.
On Rudolf Hilbert’s "two one two design" thread we have also two rugs probably Kurdish Anatolian very Kazak looking - unfortunately they are not too early: their estimated age is only 19th century.
No early Caucasian rugs are known.
No significant differences in structure between Armenian Dragons and Anatolian rugs of the same age.
Next question: Where are the carpets woven in the Caucasus in 15-17th centuries and why they don’t appear even in 15-17th c. paintings - as Anatolians do?
My answer is simple: perhaps Caucasians were labeled as Anatolians too - and incidentally, a lot of them (Anatolians and Caucasians) were probably of Armenian make.
So, the rug in Dosso’s painting COULD be an early Caucasian prayer rug?


Posted by Sophia Gates on 01-08-2002 11:03 AM:

Art/Design as Language

Yes, Filiberto - that's a persuasive argument.

Hmmmmm. Will have to ponder awhile.

As far as no early Caucasians being found - that's a problem, isn't it - perhaps they wove primarily flatweaves, for their own use? Other writers have pointed out the apparent fact that we have precious few Southern Persian rugs that can be confidently dated
to before the mid-19th century - same rationale?

Ideas, anybody?

Posted by Rudolf Hilbert on 01-08-2002 01:56 PM:

dear caucasophiles,

after browsing through the books "Early Caucasian Carpets I/II" by Serare Yetkin I must confess that almost all carpets shown may be put into the categories Caucasian dragon or floral rugs.

That's for most of the pre 18 th ct. caucasian carpets.

But for me the interesting thing begins in 18 th ct.
Look for example at this well known rug:

The small group of known rugs of this type for me is a true caucasian (re)invention.
The highly fascinating caucasian Bijov-group may be derived from there.

What do you think ?


Posted by Vincent Keers on 01-08-2002 08:09 PM:

Dear Filiberto,

Why all rugs are Anatolian?
Because before 1900 we, the west, imported most rugs by ship and the main port was Venice.
Rugs where traded as from Turkey, because that's where they came from.
Most Persian rugs came from Anatolia to, because of this.
This is a very simple explanation, but why make it more difficult?

One main feature that puts Caucasian, Anatolian in the same melting pot is Sunnit Islam.
That's why the rugs share a lot of design elements.

And I do think Armenians, especially the poor, converted to Islam from time to time.
Same God, different laws but.........more food
As lang as they could see mount Arat, everything was o.k.

Best regards,

Posted by Filiberto Boncompagni on 01-09-2002 04:34 AM:

Hi Rudi,

That rug could be very well the transition between the Dragon-palmettes design to the Bidjov group. No other useful information from Yetkin’s book? Bennett says that Yetkin suggested that the "floral" motif derived from the Dragon one. Bennett adds that the Gohar carpet (see Gantzhorn pl. 117) with Armenian inscriptions and dated 1699 could be an example of such transition.

Vincent, you say that converted Armenians shared the design with Sunni Muslim. I disagree: no need for conversions. Rugs were woven "as luxury goods for the rich, for princes and their court, for mosques and churches, for gifts or simply for the home. Rugs for ‘everyday’ purposes would be woven at home by the women of the village, following more or less predefined patterns, while teams of master artisans would be commissioned for the more prestigious pieces….’Master’ weavers were in fact mostly women, while master designers were mostly men. These masters were in great demand and moved around from place to place." (Kaffel, page 19).
You see, those master probably had to adapt their design to the fashion and/or to meet their Muslim or Christian customer’s taste. Going around (say, from Caucasus to Anatolia or vice versa) they could see different designs from other weavers. Coming back home they could make variations for their own use and other weavers of the same village could copy them…


Editor's Note: This topic continues on another page. Clicking your mouse on the arrows >> at the lower left will bring you to it.

Posted by Bertram Frauenknecht on 01-09-2002 05:32 AM:

Hello All,
I have one suggestion as I have the feeling that most of you are not familiar with the map. Azerbajdschan was till 1828 Persian. Which means that the Khans of Gendje, Shemaka, Baku, Karabagh, Talish were part of the tribal groups that were called in the area by the Safavids around 1600.
(The Safavids came from Ardabil) From 1630 to 1720
the area was rather quiet and therefore prosperous
to an extent, that the tribes which were supposed to form an army simply refused. That was one of the reasons why the Safavids went down. Most of these tribes rooted in Anatolia and were of Turcoman origin. There are plenty of examples of village or tribal rugs both from Anatolia or the Caucasus that carry related designs. And certainly there are many Armenian rugs (I believe that Erzurum was a major center for big carpets like dragon and all the other stuff) made in the whole area. Many of the fragments that I have radiocarbon dated came up between 15th to 17th c. and show features that do not allow a distinct attribution. There was most probably a trading route in that period that went East rather than West. That would explain why these designs are not
known here, but survived in monasteries in Tibet.

Posted by Vincent Keers on 01-09-2002 06:34 AM:

Hi Filiberto,

You're painting a romatic picture. But you can't disagree with the hysterical, history the Caucasus and Anatolia has been in for thousands of years.
The bread one eats, the word one speaks.
In short, the only division that can be made is: Christian or Sunni Islam. These two formed the Law the society was based on and nothing else. No United Nations, no Human rights, no political systems.
That's why mr. Gantzhorn could write his book.
And he made all, pre-1920 Armenians Christian because this was the only way he could give them back a shred of the lost culture.
So, if you're looking for Armenian rugs, maybe the Christian designs are a lead. But I suspect some non-Armenians were Christian to.

Best regards,

Posted by R. John Howe on 01-09-2002 06:55 AM:

Roots, Purity and Historical Perspective

Roots, Purity and Historical Perspective

Dear folks -

Two days ago I tried (unsuccessfully) to make a version of this post as a separate thread. But given the way this one is proceeding, perhaps it belongs here instead. It, in part, echoes a point that Mr. Frauenknecht
has made above.

First, my congratulations to Daniel for another thoughtfully constructed and beautifully produced salon essay.

Some thoughts.

First, I think it is unlikely to look for or to hold that the sources of a given rug design or group of rugs come from a particular direction. East to west or west to east is one of the old debates and it seems to be that "both" is the most likely answer so far.

I was in the room at the 1996 session in which Wendel Swan made his "wine glass and calyx leaf" paper and at the end P.R.J. Ford stood up and said, a little impolitely I thought, given the care with which Wendel had made his argument and the unmistakable similarities he identified, that it was a perfectly marvelous presentation but of course it was dead wrong since Caucasian designs owed much more to Persian influences that to Turkish ones.

Daniel has also chosen Turkish sources for the designs he has examined, and although he has not claimed that Caucasian designs are derived mostly from Turkish models someone else in this thread has.

It might be useful to note that there are those, in addition to Mr. Ford, who disagree with this latter claim. In Ghereh, 14, pp.7-21, Christine Klose argues that the "Eagle Kazak" design originates in 17th century Persian "vase" carpets. And in Ghereh 15, pp. 7-21, Azadi makes a similar argument about "Azerbaijan-Caucasian" rugs. Here's is Azadi's ending paragraph.

"…This historical overview was designed to demonstrate and highlight to what extent this region was part of the Persian culture and can therefore not be regarded as separate. This is true of the knotting and the weaving tradition of Azerbaijan and the Caucasus. It is therefore, not correct, even absurd to refer to these rugs as Russian."

Daniel also asks the question of whether there is something like a "pure Caucasian design." This sort of question invokes in me an attitude similar to what Murray Eiland, Jr. displays with regard to such notions as "authentic" and "non-commercial." The mixing and the trading has gone on for too long. At one point Daniel appears to speak critically about the later 19th century "kustar" Caucasian production. I think it would be difficult to demonstrate that the weaving of the pre-kustar area was not subject to its own set of external influences and interferences. This may be a distinction, that is largely without a difference.

Last, I am always stuck by the rather limited historical perspective we adopt when we begin to speculate about possible sources of rug designs. Mostly, probably, it is the fault of the rugs woven before the 15th century for not having stayed around much, but one rarely has one's attention drawn to similarities to Greek and Roman designs in other media (say mosaics, some of which served rug-like decorative purposes on floors). I had the occasion a couple of years ago to examine briefly some of the serious literature on Roman mosaics (the Timgad collection in northern Africa) and began to think that I might be able to find almost anything there if I looked persistently enough. It seems to me that there is a kind of surprising ahistoricism in most of the rug design literature, despite its being enamored of the 16th centuy.


R. John Howe

Posted by Vincent Keers on 01-09-2002 06:57 AM:

Dear Bertram,

Thank you for putting it in historic perspective.
I suspect a trading route that goes west-east, goes east-west as well. One silk-route crossed the Caspian sea as well.

Best regards,

Posted by Steve Price on 01-09-2002 07:19 AM:

Hi John,

Your remark about the absurdity of rugs being called "Russian" caught my eye. That one has very odd historical roots. The basis is the widespread confusion that existed in the US about the USSR, which we generally referred to simply as "Russia". Hence, any rug woven within the boundaries of what was the USSR - which included Uzbekistan, Turkmenistan, all of the Caucasian states, and other so-called republics, was called "Russian". This, happily, seems to have finally run its course.

One of the ironies of it is that most citizens of the Soviet "Republics" resented the Russian control, and would have objected to and felt insulted by being called Russians. It was bad enough that they were Soviets!


Steve Price

Posted by Filiberto Boncompagni on 01-09-2002 09:32 AM:

Dear All,

OK. Now I have to quote myself. On the first page of this thread I wrote:
"Where are the carpets woven in the Caucasus in 15-17th centuries and why they don’t appear even in 15-17th c. paintings - as Anatolians do?
My answer is simple: perhaps Caucasians were labeled as Anatolians too - and incidentally, a lot of them (Anatolians and Caucasians) were probably of Armenian make.
So, the rug in Dosso’s painting COULD be an early Caucasian prayer rug?"

I’m not excluding Turkic or other influences, nor saying Armenians are the ONLY responsible for the Caucasian design. I say they are ONE (although very important) element at the origin of the Caucasian lexicon. Anyway, my point is just this: IF we search for early Caucasian examples, let’s look among early Anatolian rugs, both real and painted.
And…excuse me if I insist: instead of speaking of general theories, what about the rug in Dosso’s painting? That is the main subject of this posting and the clue to my reasoning.
It looks Caucasian to you?
It is similar to the one in Kaffel’s book?
Could you agree or exclude the influence between the HALI, Dosso and Kaffel rugs?


Posted by Deschuyteneer daniel on 01-09-2002 10:12 AM:

Dear all,
This other posting to answer more personally to some questions raised in this thread:

1/ Filiberto,
The Dosso Dossi’s rug you present, is certainly not, a para-mamuk rug. It lacks the floral (flowers and cypress) spray on the field and what you think to be an octogonal medallion at the bottom of this rug (left side of the photo) is in fact the upper part of the “reentrant keyhole” of this typical 16th century Anatolian Bellini prayer rug. The fact that we have the same design in the border surrounding what you think to be an octogonal medallion and the minor border is also in favor of my hypothesis.

Which historical facts … do you have to assert that Armenians were the weavers of the Dragon rugs and their relatives?
Karabagh where these rugs were most probably woven was, I agree, inhabited by Armenians, but I think Armenians were more traders than weavers and that Azeri Turks played a more important role as weavers.

2/ Marvin,
I agree with your analysis of the Anatolian “earmarks” you see in the Gantzhorn’ s plate 695- presented by Filiberto, in his first posting. I would add that the two hooked motifs emerging from the top of the keyhole are typical of a lot of Karapinar rugs.
I disagree that there are so few differences in structure between early Anatolian and Caucasian rugs. The cable wefting of the early Caucasian rugs is at least one distinctive structure characteristic which don’t appear in Anatolian rugs.

3/Patrick and Sofia,
I agree that the origin of the Kufic borders is most probably the Kufic calligraphy script. I can’t accept that the “tribal version” seen in the South Persian border illustrated in Patrick’s posting has the same origins, without seeing intermediary stages of evolution (or devolution).

Posted by Filiberto Boncompagni on 01-11-2002 04:19 AM:

Dear Daniel,

I start from your question first. You ask:
Which historical facts … do you have to assert that Armenians were the weavers of the Dragon rugs and their relatives?

Well, I do not have historical facts (not even an Armenian Grandma), only a few books. Let’s forget Gantzhorn.
Both Bennett (Caucasian) and Gans-Ruedin (Le Tapis du Caucase) write more or less the same: in origin dragon carpets were attributed to Armenia, then some writers contested the attribution. Then they both refer to Maurice Dimand, the former head of the Islamic department at the Metropolitan Museum of Art, New York. "As to who wove the dragon rugs, we return to the theory put forward by Maurice Dimand that the Armenian national, as opposed to the geographical, provenance is a strong possibility." This is the quotation from Bennett, I spare you the similar one from Gans-Ruedin.

You say:
the rug you present…is certainly not, a para-mamluk rug. Well, never said it was!
what you think to be an octogonal medallion at the bottom of this rug (left side of the photo) is in fact the upper part of the "reentrant keyhole" of this typical 16th century Anatolian Bellini prayer rug.

Keyhole? Yes it may be. Or may be not - but let’s say you are right:
1- this does not seems to me a typical 16th century Anatolian Bellini prayer rug, although it is a contemporary. The typical Bellini rugs, as shown on Wendel’s article, have an almost empty field and a much sober design.
2- Where else can we find prayer rugs with "re-entrant keyholes"?

But in Caucasian prayer rugs, of course.

Some more quotations? (I am sick of typing quotations, you know, but it has be done).
Kaffel, page 42: "Prayer rugs from the mountainous region of southwestern Caucasia …are profoundly influenced by older Anatolian examples in which the mirhab, often with a re-entrant motif, take most of the field."
Page 44 (referring to a sort of "Tree of Life" device): "most likely evolves from the crosses in medieval Armenian churches….Many Armenian symbols and design are incorporated into Caucasian weaving, and it is possible that many of these rugs were woven by Armenians".


Posted by Guido_Imbimbo on 01-11-2002 08:28 AM:

Dear Filiberto,

for Bellini carpets please refer to John Mills, "The Bellini Keyhole or Reantrant Rugs" in Hali 58, pp. 86-103.

At page 91 you would find your Dosso Dossi painting "The Allegory of Hercules". According to Mills, this painting represents a "Bellini" rug and in particular it "shows the 'Ghirlandaio' medaillion well" (p. 91)
If you any doubts left go to the first page in the article that reproduces the Topkapi Re-entrant Prayer Rug that is the same rug painted by Dossi.

Best regards


Posted by Filiberto Boncompagni on 01-11-2002 08:55 AM:

Thanks, Guido.

I trust your words. Besides, I have only TWO numbers of Hali and # 58 is not among them.
Still, I’d like to see the differences or similarities between Kaffel’s rug and the Dosso-Topkapi rug. Could you post a scan?
Thanks again,


Posted by Guido Imbimbo on 01-11-2002 09:19 AM:

Dear Filiberto,

here is a (bad) picture of the "Topkapi Re-entrant Prayer Rug", West(?) Anatolia, 15th or 16th century, 125x192 in the Topkapi Saray Museum, Instambul,inv. no. 13/2043.

The similarity of the Dossi painting is impressive. Same border, same medallion (called by Mills "Ghirlandaio"), same tringular devices in the field, similar sprandels, etc.

There is another Turkish carpet "The Sion Symmetrical Re-entrant Rug" rieviewed by Mills in the same article (p. 102) that has very high similarity with the Dossi carpet.

This rug was already analyzed by Marino and Clara Dall'Oglio in an earlier article in Hali 27 (p. 36-39). In that article ("A Discover at Sion") the authors noticed "the same ornaments, in particular the medallion, sorrounding comb-like motifs and the re-entry band design, appears in a painting by Dosso Dossi" (p. 37).



Posted by Filiberto Boncompagni on 01-11-2002 11:28 AM:

Thanks Guido,

Yes, there is no doubt the two carpets are related, but Dosso’s rug seems a more rustic version of the Topkapi. I put here again the painting detail, for more convenience:

Same medallion, yes - but the border system seems much simpler in the painting. The main border, for what we can see in Dosso’s rug, is Kufic-sque, the secondary much more similar to the one in Gantzhorn. Again, a coarser version of the Topkapi. The triangular devices in Dossi are sort of transition between the Topkapi and Kaffel’s 31 (and similar).The spandrels? Hard to see, hard to say.

An important characteristic of Dossi’s rug is that the inner border departs first perpendicularly then diagonally to form the mirhab, exactly like in most Caucasian prayer rugs. The inner borders in the Topkapi and the Bellinis, instead, depart only in diagonal.
Sorry if I insist, but the rug in the painting does not look a very typical Bellini to me.
It seems rather a village version of the Topkapi - with quite Caucasian borders and a Caucasian prayer arch. Oops.
Other opinions?


Posted by Guido Imbimbo on 01-11-2002 10:53 PM:

Dear Filiberto,

You asked for other opinions regarding the rug depicted in the Dossi's painting.

This is a list of references to the Dossi's painting in carpet literature.

All the authors relate the Dossi's Rug to the 15th century Anatolian carpets known as "Keyhole" or "Re-entrant" rugs.
Note also that Michael Franses mentions that the upper niche in the Dossi's Rug is also "found in later Caucasian prayer rugs".

Dossi's Painting Witchcraft (The Allegory of Hercules) in Carpets Literature

1) Johanna Zick, Eine Gruppe von Gebetsteppichen und ihre Datierung, Berlin 1961, p. 14 (cited).

2) Michele Campana, 1962, pl. 21

3) Marino and Clara Dall'Oglio, 1985, op. cit., Hali 27, p. 37

4) John Mills, 1991, Hali 58, op. cit., p.91, fig. 9

5) Michael Franses, "The Bellini Keyhole", in Orient Stars, 1993

"There are two paintings depicting with a niche with 'shoulders', one by Carpaccio, circa 1490, the other by Dosso Dossi, circa 1535. … The rug in the Dossi painting has the eight-pointed star medallion with hexagonal lozenge extensions and 'hanging lamps', motifs which can be found on several Anatolian rugs from the 15th century" (p. 279).

"Dossi's rug has several features that relate it to the Turk ve Islam Eserleri Museum example, inv. No. 357, and the rug in Sion" (note 413, p. 378).

"Some paintings Keyhole rugs of type B: ….. 4) Dosso Dossi: Witchcraft (The Allegory of Hercules). c. 1535. Uffizi Gallery, Florence. …. Carpet: most visible; unusual upper niche, similar to those found in later Caucasian prayer rugs" (note 392, p. 376).

I hope this helps you.

Best regards


Posted by Filiberto Boncompagni on 01-13-2002 01:50 AM:

Dear Guido,

Ancora grazie for your work of research.
So, as it should be expected, other people remarked the Caucasian prayer niche.
One could say, as a swallow doesn’t make a summer, a prayer niche with shoulders doesn’t make a Caucasian prayer rug… but I think it’s a step in the right direction.


Posted by R. John Howe on 01-13-2002 08:45 AM:

A Little More on "Purity"

Dear folks -

I mentioned in a post above the Roman Timgad masaics collection in Algiers.

One further thing about this collection is that it is especially admired by students of Roman mosaics because the Timgard mosaic artists reputedly did not respond much to external influences, while mosaic makers in other areas did.

So this "purity" business has gone on for some time and in other settings.


R. John Howe

Posted by Filiberto Boncompagni on 01-13-2002 11:54 AM:

Dear All,

A very good friend of mine works in the Galleria degli Uffizi,
I was curious to know if the octagon or keyhole in Dossi’s rug was empty or it had a motif in it and I asked him to have a close look.
He just answered me: the octagon internal field is red, like the rest of the rug field.

I feel I have to mention another "curiosity": my friend’s name is - guess what?


Posted by Filiberto Boncompagni on 01-14-2002 09:56 AM:

Design Migration

Dear All,

There is something intriguing me. Look again at the pictures in this posting.
The Chihil-Sutun rug, surely a court or city workshop product, appears in a 1501 painting, thus is dated to the same age.
Take this design, add a key-hole inner border wrapping the lower medallion, transform the minbars in triangular stepped or hooked devices, simplify it removing the floral spray design and you will get more or less the Topkapi rug (very approximately dated as 15th/16th century).
Simplify again the design (simpler border system) empty the keyhole, add shoulders at the mihrab and you have the rug in Dosso Dossi’s painting (dated 1535).
It seems a classical migration of design from workshop to village weaving. I was used to think that, in this migration, time was also involved but those rugs should be contemporary.
This suggests a strong activity in design evolution at the end of 15th century in the region.
We do not know, of course, the real age of the carpets depicted in the paintings.
What do you think?


Posted by Greg Koos on 01-16-2002 11:13 PM:

Mr. Deschuyteneer,

I hope this isn't too far off the thread. I'm interested in the "triangular devices." The Adlai Stevenson House collection contains a Tulip Ladik which features four of these devices - butterflied - in the mirab. Are you aware of a terminal date for the use of the device?

You stated that the use of the device is typicaly found in early rugs. I assume that the rug is mid-19th centry, although it carries an english language tag - also 19th century reading "150 years old" I believe the tag to be an old dealers exageration/lie.
Greg Koos
Greg Koos

Posted by Steve Price on 01-17-2002 06:38 AM:

Hi Anyone,

I've long found it interesting that seeing a rug of a particular design and color in, say, a 16th century painting, is used to justify the notion that any rug that looks like that one is of similar age. It's obvious enough that the rug design and colors must have been used at least as early as that date, but that's about as far as I get when I try to reason it out.

Any comments to help clarify my addled brain on this?

Steve Price

Posted by Marvin Amstey on 01-17-2002 08:38 AM:

Not addled; right on! Marvin

Editor's Note: This topic continues on another page. Clicking your mouse on the arrows >> at the lower left will bring you to it.

Posted by Filiberto Boncompagni on 01-18-2002 02:27 AM:

Hi Steve,

You find it interesting.
I found it plain odd, enforcing my conviction that we have to take rug scholarship with a pinch of salt and use our own judgement, when it is possible.
And take that one with a pinch of salt too…