Home > LACMA's Questionable Rug Purchase >RK Looks at Mills's "Bellini" Theory
email: jc@rugkazbah.com
Sun, Jan 1st, 2006 12:08:38 PM
Topic: RK Looks at Mills's "Bellini" Theory

(ed. December 30, 2005 Once again and for the second time some comments on John Mill's re-entrant rug theories written and published on RK.com several years ago.

We believe they are still pertinent for many reasons, the perspective they lend the LACMA/dodds rug fiasco not the least of them.

What immediately follows is the intro to the first republishing, this being intro to the second.)

Several years ago, RK.com illustrated several “re-entrant” type rugs along with John Mills’s views about this group of early rugs.

Although these re-entrants were all ‘prayer’ rugs, ie. single niche, we feel it is pertinent to have viewers read them now this topic – the re-entrant - has again come to the fore.

These, as well as all of the other previous posts that have appeared on RK.com and are not currently in the various topic areas, have been moved to the “Archive” area where they can be still perused.

Our hard working, and very talented webmaster, located them and will repost them along with this short re-introduction.

Entitled “A Tale of Two Prayer Rugs”, this 4 part series provides a rather in depth look at some masterpiece, and some lesser “re-entrant” style weavings.

They have much to say in line with the dodds/LACMA rug, which by the way is so inferior to even the worst one chosen, it wouldn’t have stood the chance of being included, even as the runt of the litter.

Without further ado here are those posts in their entirety. We have added some bold type in a couple of place for emphasis and added a few new comments, marked by (ed.).


The majority of antique oriental rugs are copies.

Of course, that does not imply reproduction, in the sense of fakery – a topic that has recently come to the fore, as more and more blatant contemporary ‘fakes’ enter the marketplace as ‘antiques’.

This is not the meaning the opening sentence of this post is intended to convey.

The copying mentioned above refers to the practice of most weavers to use an existing, i.e. older, example as a model for their work.

This convention exists in all art areas. For instance in paintings and sculpture “school of” and “after” are the descriptive terms used to describe this practice.

In dealing with carpets, unlike paintings and sculpture which are signed and dated by artists, the pervasive anonymity precludes an easy road to understand the age relationship similar examples maintain.

As readers of RK.com know, there are a number of complaints and pet-peeves with way the rug world revolves that have been voiced in my writing and while some, like professor steev price=clown might chose to characterize as personally directed at certain individuals and trivial, they surely are not.

One of the most disturbing is something I have characterized as “over-dating”. For those readers who might not understand this, you can read more about this topic in a series of posts entitled “Over-dating – A Good or Bad Practice” which also appears here in JC’s Corner. (ed. This series of posts is in the Archive for any readers desirous of reading them in full)

The over-dating described therein deals with one particular aspect – the use of over-dating as a sales tool. In theory the older a carpet is the more valuable. In theory this is something no-one would deny.

However, falsification of age determination, which is rampant in the rug world, is not the issue “The Tail of Two Prayer Rugs” is meant to address.

In tandem with purposeful over-dating, another types also exists. This we can call, for lack of a better term, over-dating by ignorance. Whether through lack of expertise or just plain ignorance, it is all the same in the end.

Ok, let’s now look at these two prayer rugs:

Both these prayer rugs (the one on the left from the east Berlin Museum for Islamic Art and the other from the Ballard Collection/Met Museum in N.Y.), and a number of other with the “re-entrant” design, appeared in an article written by John Mills, which appeared in hali #58. What is remarkable is that he dates both of them to the first half of the 16th century.

While Mr. Mills’s insight and knowledge about carpets in paintings is well proven, his ability to determine their age is surely not.

One of these is, in my opinion, much older than the other. How much? Well that impossible for me or anyone else to positively state but be assured these two pieces are separated by more than a generation. In my estimation at least 100-150 years.

In part two of this discussion we will look at some of the reasons for my assessment.


For the past 15 years “It’s all about Color” has been the predominant mantra antique carpet collectors have voiced. Before then it was “It’s all about Condition”.

While both of these are inarguable and entirely valid, depending on ones predilections, neither is omnipotent. Color and condition are important for obvious reasons, however, when referring to historical carpets, i.e. those that are prototypes and archetypes, they take the back seat in lieu of design, materials and structure.

The re-entrant prayer rugs examined by John Mills are in such a category and Mills’s thesis and its conclusions are based solely on design, as no technical aspects of these rugs are discussed.

His lack of attention to these details is understandable from a convenience viewpoint, as locating and confirming these details is quite difficult and time-consuming. However, their absence blows a big hole through his exercise. Why he and hali’s editors did not attempt to include this information is, in my opinion, unforgivable and, at the least, diminishes the value of this work.

But more importantly, it demonstrates the frailty of Mills’s premise and it’s conclusions.

Examining carpets solely on the basis of design and then drawing valid conclusions from them is possible but fraught with pitfalls, especially when the examiner is not aware enough of the intricacies of the examples he has chosen to examine.

Design is the most easily reproduced aspect of any weaving and unless the observer has a deep and thorough understanding of the multitude of examples in any group or cluster, those judgments are nothing more than ‘opinions’. They surely cannot be considered facts.

By comparing the relative ages he assigns to these weaving, which is the basis for his conclusions about them, it is very apparent to me that Mills may know a lot about paintings picturing carpets but he doesn’t know much about the carpets themselves.

In part I two of those rugs are pictured. Here are two more:

Again, I have strong reservations about the dates Mills assigns to these rugs. The one on the left, from the West Berlin Museum for Islamic Art, he dates as 15th-16th century and the other, from the Wher Collection, 16th century. But in this comparison at least Mills got their relative ages right – the west Berlin piece is older than the Wher example. As such, I would prefer to see the west Berlin example dated at least 50 years later to the mid- 16th and the Wher to the early or mid-17th.

Alright then, on what basis do I make such statements, considering I, too, am using only design as a basis?

Simply put, the only mantra which has any credibility in dealing with historical carpets compared on the basis of design is “It’s all about Proportion".

Nuance of design is the foundation Mills, and I must say all others who delve into these types of rug comparisons, use to hang their conclusions upon. However, as even the most novice student of carpets soon realizes, weavers are good copyists and the little zigs and zags of a design were not lost on some weavers who had the perceptions and skills to faithfully reproduce the designs formulated by previous generations.

But getting the proportions right and exact was far more difficult and demanding.

Duplicating the intricacies of the internal proportions of each design itself and the even more demanding chore of the proportional relationship between all the designs as a whole, while appearing simple, actually is an incredibly difficult task to accomplish -- it was achieved by very, very few weavers.

So while the designs might ‘look’ the same, in reality when they are minutely examined, they almost always prove to be quite dissimilar. And these differences are the significant clues that allow a relative age determination. In short: Proportion is everything in design comparisions.

Of the four pieces so far illustrated, two in Part I and two here, the east Berlin example is, without a doubt the earliest and its proportions show the epitome for this type of re-entrant prayer rug. Go study them and you’ll see.

Personally, I am not very interested in Classical Carpets and even though I appreciate them and have a strong working knowledge of some types, particularly Ottoman examples like these, they don’t ring my bell as loudly as early village and other low-culture weaving types.

That said, some the earliest Classical Carpets can demonstrate facets of those qualities I find fascinating in non-classical types. The east Berlin piece is one of these.

Its design is powerful. Why? Because the design elements are boldly articulated and clear; the proportions dynamic, each element not only exists on its own but more importantly interacts with all the others – there is a synthesis of design.

Like a fine piece of machinery, all the parts work together – nothing is unnecessary. There is no design for design sake, nothing extra or superfluous – no fussy do-dads to muddy the picture. Border and field balanced, neither overpowering the other.

Again, while this might seem easy to do, it’s not and the weavers of the three other examples prove how difficult it is.

One last comment, notice the presence of two large lazy lines just above the medallion of the east Berlin piece. There are also two similar ones that begin at the center of the re-entrant design.

I would bet none of the others have this structural aspect.

When technicalities like this - just like other structural, dye and material differences that I am sure could be shown to exist in these four rugs – would be compiled, they would validate the assessments of relative age made here and not those forwarded by Mills.

Ok then, that’s it for Part II. Stay tuned for Part III when the antecedent examples for the re-entrant prayer rug design will be illustrated and discussed.


According to Mills “…arguably the oldest and finest of surviving re-entrant prayer rugs” is an example from the Topkapi in Istanbul. Here is the photo:

Sandwiched into the re-entrant ‘keyhole’ is a “Ghirlandaio” medallion, which draws its name from the appearance of a similar medallion used in a painting by a 15th century Italian artist whose name was Ghirlandaio.

This medallion bears a striking similarity many similar ones found on early Turkish, Mamluk and Turkmen rugs and while Mills seems to equate its appearance in the re-entrant rugs as a significant factor denoting age, I don’t necessarily agree, especially when it is squeezed and shoe-horned into that space, as it is here. (ed. The dodds/LACMA rug is even worse in this respect. But instead of a ghirlandaio sandwich, there the medallion is sandwiched within its own confines, a far worse fate)

The Topkapi re-entrant holds little fascination or interest for me. Why? Because, save the drawing of the main border, all other elements have that contrived and cutesy look of later 18th and 19th century Turkish weaving.

I am in no way implying it is from such a late period, though I’d be hard pressed to date it any earlier than the later part of the 17th century, but the15th or early 16th century date Mills gives it doesn’t hold any water in my estimation.

I do agree with his appreciation of the main border, which I believe has its source in architectural decoration.

Here is a decorated column from the Mosque in Dandanakan, South Turkmenistan, that is dated to the 11th – 12th centuries:

The open fret-work and stylistic conventions of the column bear strong witness to the open fretwork border design and to the strong link the Ottomans maintained with their central Asian heritage. The design also relates quite well to the field pattern of the Pazyryk carpet, which of course is quite understandable considering many believe it had a central Asian origin.

Surprisingly, Mills illustrates a rug that actually is the “oldest and finest” of this type of re-entrant rug and why he mis-judged the Topkapi piece as ‘the great one’ can only be ascribed to his lack of real rug expertise and knowledge. Commonly called Mamluk but, like all the others given this attribution, it is still of an unproven provenance. This prayer rug is from the east Berlin Islamic Museum.

Here is the photo:

In common parlance this is a ‘pisser’ of a prayer rug, one that’s got it all. Granted it is somewhat condition challenged but, aside this fact, the colors, design, articulation, and proportion are perfect. Almost too perfect.

In my estimation, it is the prototype for all the others illustrated in parts I and II of this discussion, as well as, all of those Mills illustrated in his article. (ed. And it makes the dodds/lacma rug look like the inferior, late period reproduction we claim it is)

No cutesy hanging lamps, scrunched “Ghirlandaio” medallions or contrived designs – this piece is a powerhouse, a freaking tour-de-force of weaving.

Look carefully to notice the light blue plant form growing from atop the ‘keyhole’- this is the archaic source for the medallions that float in the fields of the four pieces illustrated in Parts I and II.

The magnificently conceived crenellated mirhab and mysterious forms in the panel above it speak volumes as to the fact that this weaving could really be attributed to the 15th century. The sinuous dragon-like cloudbands of the red inner border are wonderfully life-like.

But it is the open, proto-fretwork green ground outer border that seals its position as the prototype of the Topkapi rug.

One more point, notice the two proto-‘Ghirlandaio” medallions flanking the mirhab in the two green-ground spandrels. And if you look carefully, they are each suspended within a red fretwork ground pattern that is both complex and unique. So much for Mills’s Topkapi choice.

But there is one other Turkish rug of the re-entrant category deserving mention here, although, again, Mills missed this one, too.

Here it is:

This amazing survivor, from the Turk ve Islam Eserleri Museum in Turkey, almost defies words to describe it.

If the Mamluk was the prototype for all the other re-entrant prayer rugs, this piece is the archetype. Some readers might remember it from another issue of hali where it was published as one of the connoisseur choices.

Unknown to most people before then, this was the first publication but I had a black and white photo of it since the mid-seventies in my photo-file collection. I still marvel at it even after knowing it for so many years and having seen it in the flesh in Turkey. To say it’s a great rug is an understatement of the greatest degree – it’s a supreme rug, one of my favorite pile-woven Turkish weavings.

Here all the elements of the re-entrant style are delineated in their archaic form – the ‘key-hole’, the “Ghirlandaio” medallion, the crenellated mirhab and the open fret-work border.

In the text that accompanied its connoisseur choice publication debut, gary muse, to his credit, recognized the relationship and association to the re-entrant style rugs. But his waxing on about its Neolithic connections, its iconographic “…interrelationship of human and animal worlds…(and) the concept of ‘guardianship associated to the Neolithic period”, animal skin prayer rug relationship and references to “…the ancient language of prehistory” are as foolish as Mills’s ignoring it in his survey of re-entrant prayer rugs.

This is a great rug indeed, and as muse surmised it may pre-date even a 15th century attribution (personally I am sure it does).

But what is even greater is the light it sheds on the proscribed nature of early Turkish Village weaving traditions.

Studying a rug like this and comparing it to other examples, not only re-entrant type weavings, is more instructive than reading dozens of rug books. Try it and see for yourself:


We have received several emails questioning the choice of the title chosen for this discussion: “A Tail of Two Prayer Rugs”. This is not surprising as more than two prayer rugs are presented and there are no tails in sight.

Ostensibly this can be explained by stating it was a take-off of the title John Taylor used for his post “A Tale of Two Safs”, which was published last week in the “Pictures for Discussion” Topic Area. But there is another, more complicated, one as well.

Collecting Antique Oriental Rugs remains one of the tiniest, if not the tiniest, corners of the Art World. This situation, which has existed for the hundreds of years they have been objects of interest in the West, results from the extreme rarity of the best examples, the difficulties inherent in their identification and provenance and, most importantly, the lack of real information about them that is available for reference.

Today things are no different even though literally hundreds of books have been published in the past 35 years. These factors have engendered this circumstance but, as significantly, is the somewhat unhealthy level of ‘packaging’ and control exerted by a few individuals, who are perceived almost as royalty in rugdom.

While defacto hierarchies exist in most other art areas, their larger constituencies and superior levels of reference material act as a checks-and-balances.

However because these factors do not exist in the rug world, the small group of ‘scholars’, like thompson, mills, etc., ‘dealers’, like franses and herman, and the ubiquitous hali magazine can dictate taste, prices and fashion according to their own agendas or personal opinions.

This is bad enough but it is exacerbated by the complete lack of any dissention or critique whether from within the rug community or from outside it. Added to this is the collusive nature of relationships these groups, ‘scholars’, dealers and hali maintain – the old one hand washes the other adage sums it up to nicely.

The rug public at large isn’t very large and their sheepish acceptance and eyes wide shut attitude allows this status quo buddy-ism to flourish and perpetuate.

It doesn’t encourage change and, what is more alarming, no one seem to want it.

So, while these cliques clack together for short-term mutual and personal benefit, the opportunities to create wide and lasting appreciation for the art historical value and importance of Oriental Rugs languishes by the wayside.

There is a tremendous amount of hard work to be done to stimulate and affect change. But because it’s easier to ‘sell’, both literally and figuratively, the accepted old ideas and half-truths that imply classical court carpets and some types of ‘decorative’ ones are the pinnacle of rugdom, those ‘scholars’, holier than thou dealers and lazy hali editors make no effort to investigate these "false Gods" of rugdom or to look carefully in other directions.

The result is the old tail wagging the dog and that is the rest of the story. (ed. How true these words ring in talking about the present-day LACMA fiasco.)

Home   Buy/Sell at the Kazbah   Terms Of Service