Archive for October, 2014

Frankly Fragmentary

Posted in Uncategorized on October 17, 2014 by rjohn

On Saturday, November 12, 2006, members of the Philadelphia Rug Society, and a couple of guests, met at Craig Wallen’s Gallery 51 to consider the subject of rug and textile fragments.

fragpeople1.jpg fragpeople2.jpg

Although the session was billed under the question “When is a Fragment Collectible?,” the organizers prepared a broader list of questions to guide and stimulate discussion.  These were:

  • What is a fragment?  Why fragments?  Fragmented vs complete?
  • Primary considerations: Age/history?  Rarity/provenance?  Design/graphics?
  • Secondary considerations: Size? Availability? Cost? Can it be displayed readily?
  • How will it add to your collection and to your understanding of rugs?
  • How will it be displayed?  As is? On a backing? Framed?
  • Curatorial and conservation considerations?
  • Where to find a good fragment?  When is a fragment not “honest?”

We did not systematically address all of the issues reflected in these questions.  Nor did we restrict ourselves entirely to them.  But the preparation was functional for our discussion.

What follows is likely best described as a series of snapshots of this conversation, supplemented when possible with images of related fragments mostly brought by the participants.

Craig began the session with the question “What is a ‘fragment’?”


Well, it is a “part of a larger piece or of a whole textile.”

But Dennis Dodds quickly noted that the notion of what a fragment is is likely broader in Europe than in the U.S. Seemingly very complete pieces are published in Hali as “fragments,” with meticulous notations of what is missing.  In the U.S., it seems, if it’s mostly there, it isn’t a fragment.  This U.S. tendency, by the way, was reflected in our discussion.  Some were unwilling to call a given item offered a “fragment.”

Sometimes, this unwillingness was explicitly tied to the notion that a given piece was “basically all there,” despite its missing, say, a border or corner.  But another aspect of “completeness” was explored as well.  That is, that sometimes particular traditional formats are meant to be “composed.” That is, they are made up of two or more pieces, each of which was woven as a “complete” item despite the intention to combine it with others to form a larger array.

A first order example of this latter sort of “fragment” was an 18th century Anatolian striped kilim that Craig brought out.


People in this session were willing to call this piece “half of a two-piece” kilim, but not a “fragment.”

John Howe had a piece the “fragmentary” nature of which is similar but more subtle.


He said that neither he, nor the dealer he bought it from, knew what it was when he bought it. This piece is visibly complete as it came off the loom. What could it be? A number of experienced rug and textile people could not say. But Wendel Swan recognized it as one front chest tab from a Persian (likely Kurdish) horse cover.

Later in this session Howe showed a Turkman piece that was part of a head decoration for a horse or camel.


He said that, while it is clearly only one element of a much larger assemblage, part of its appeal for him was that it projects an unmistakable “wholeness” by itself.

So we need to decide whether we include in our notion of “fragment” something like these pieces above: complete as they were woven or embroidered, but meant to be sewn together with some other separate parts to form a larger assemblage. Conceptually, this horse cover tab and this head decoration component are the same as the sections of any textile (kilims, jajims, etc.) woven in parts, but sewn together to make a particular “complete” item of a given format.

A next part of this discussion centered on why it is that sometimes a very good fragment will be found, say in a flea market or country auction in the Philadelphia area, but can’t command the price of something (let’s say that the two pieces are very similar, even the same one) say, found by someone like Michael Franses? A second version of this question was “How does a piece go from being an item of near “trash” to something seen as rare and valuable and that can draw a big price on the international market?” (This question isn’t fragment-specific, but is likely more acute in the case of a fragment.)

The seeming conclusion took two forms.

First, Dennis Dodds pointed out, dealers function in part to research given items, to discover what they have sold for recently.


This has an “auction” side and a “private sale” side.  Dealers also collect the qualities that seem related to price.  He acknowledged that such research and comparison is more difficult for fragments.  But the thrust of this argument was that dealer research can show what the likely current market value of a piece is.

A second aspect of this difference in value has to do with the fact that there are different “levels” in the market.  Price will also be affected by the level(s) of the market available to a given individual or dealer.  Howe said that his brother, who is an antique tool dealer, once found a rare early American tool, bought it for about $100 and sold it for about $2500.  In about three turns in the market it subsequently sold for $10K.  But Howe’s brother said that he felt no resentment about not getting the $10K.  He had made good money at the level of the market at which he could operate.  The subsequent sales happened at levels of the market to which he simply did not have access.  This offers another explanation of why a given piece might sell for very different prices.  There are many levels of the market and not all of us have ready access to a particular market level.

There was another aspect of this discussion, not all of which occurred at one point in it.  It was that there was some puzzling about how to determine what a fragment is worth. Here, there seemed some call for auction houses to become more active in their research on fragments.  But it was also acknowledged that there has been a great increase in the flow of information about textiles, in general.  This  increase has been partly due to an expansion in the volume of textile publications since about 1970, but is also, more recently, the result of the internet.  The information available to the interested rug and textile collector today is dramatically greater than it was 40 years ago.

We next talked about “Why buy a fragment?”

“Well,a fragment can be beautiful,” someone said.  There can be something about its colors, patterns, textures that make it stand out despite its fragmentary nature.  It can, on occasion, be more interesting than a more complete, even an entire, piece. The Textile Museum’s “Pieces of a Puzzle” exhibit provided an example.

The links to this exhibition no longer work, but I can describe this instance.  This exhibition presented some large fragments of a very large rug.  One of these large fragments included the border shown below.

The large fragment on which this border occurred is, itself, only part of a huge palace-sized carpet with wide areas of an open red field. This field is unattractive, in part, because it is not in good condition, but the important aspect to notice here is that this huge, red field works in the larger fragment to overwhelm the narrow border.

An isolated image of this narrow border, treating it as a fragment separate from the larger piece on which it occurs, is, for me, the best piece in this exhibition. It is, precisely, its separation from the larger piece and its presentation in the exhibition as a discrete item that lets us see its real visual merits.

Now I do not want to argue that the point above is general, even frequent, but this isolated border fragment demonstrates to me that it can occur.  Internet textile dealers have discovered this and often present an initial, tantalizing, detail image (a kind of digital “fragment”) of a larger piece they are selling.  Sometimes one is disappointed to find that the whole piece is less interesting than the initial detail shown of it.

“One can learn a lot from a fragment.” It’s not frequently the case any longer, but there were once collections largely composed of what were called “study pieces.”  Their fragmentary nature often made some of their aspects (e.g., their structure) more accessible.

In the TM “Pieces of a Puzzle” exhibition on classical Persian carpets from the 16th and 17th centuries, where fragments were drawn from carpets that were very large, Dan Walker pointed out that the fragmentary character of the available pieces actually made side-by-side comparison of them more possible than would have been the case between the original large carpets.

And, sometimes, one can determine important aspects of a very large piece on the basis of a fragment alone.  Again, the illustrating links no longer work, but if we looked at three fragments of a very large carpet fragment in this exhibition, we could see that the left side of the two larger fragments are selveged without a border.  This fact, and the further one that the field medallions are interrupted by this selvedged left edge precisely at a halfway point, combine to suggest that these fragments were parts of a much larger assemblage.

These two features demonstrate that important aspects of a much larger piece can sometimes be suggested by examining a quite small fragment alone.

Some feel that they can put a fragment up on a wall and reconstruct the entire piece.  Christopher Alexander’s efforts come to mind.  True, he claims that his reconstructions are importantly aided by his “theory,” but the example of a 13th century fragment from which he claims to have derived a complex endless design that he rendered in a series of paintings is dramatic. (See his A Foreshadowing of 21st Century Art, pp. 130-137, 351.)

Collectors sometimes find a fragment attractive because they “can afford it.” Fragments can be (but are not always) priced a lot lower than complete pieces of a given sort. This aspect has lots of facets but we treated only one.

First, fragments are often pretty frequent when one begins to look for rugs, especially if one is looking in places like flea markets or country auctions.  And while fragments may not be the best place for a novice to begin, it is likely that rugs in poor condition, some of which will be fragmented, are likely to be part of what a new collector comes onto frequently.

Extensive praise of fragments sometimes has unintended consequences.  There was, awhile back, in Germany, an exhibition of striking fragments.  The catalog praised them so effusively that, it is said, some dealers were led to cut up some complete pieces to produce some.

At this point Howe passed around a full-pile fragment of an approximate quarter of a Middle Amu Darya chuval with a mina khani design, very soft wool,  good color, and a lavish use of silk.


We do not know the condition of the rest of this piece when it was cut, but it seems possible that this is an example of a large piece cut up to produce several attractive fragments.

One reason why there might be more temptation to “produce” fragments in the rug world is that the market is more forgiving of rug and textile fragments than it is about fragments in other areas.

David Weiss, associated with the Philadelphia auction firm, Freeman’s, was a member of our salon group and was asked about this.


He agreed that a fragmentary condition is sometimes disqualifying in other areas of art and antiques.  The famous illustration is that an antique chair, with one leg missing, is not worth much, but this rule applies more generally.  For example, someone suggested “A fragmentary Picasso would not have much value, would it?”  David seemed to agree.

“Fragments are often easier to display.  They are often smaller and fit into the available space more readily.”

The snapshots of this conversation above do not do it justice.  It was focused, intelligently exploratory, and vigorous enough that it went on for over an hour before we began seriously to look at the pieces participants had brought.  But, finally, we did.

Craig began with this Coptic fragment.


He said that this is the third piece like this he has owned.  They all feature various numbers of similarly drawn birds and camels and he has wondered whether they are from the same piece.  He said it is no longer possible to assemble them side-by-side.  Coptic pieces are variously dated from the 4th to the 7th centuries.

John Howe had brought a Coptic piece.  He said that he bought it entirely on impulse from the window of an antiquities shop in Georgetown.


He said that this piece is a “little rug” composition made up of parts of Coptic garments.  The field is densely patterned with a variety of designs, some of them, perhaps, of people.  The “borders” are likely strips from the edges of garments that have good spatial qualities and recognizable bird designs.

Pieces composed of fragments are encountered in other areas as well.  Harold Keshishian has an impressive “prayer rug” design


that on closer examination turns out to be entirely composed of fragments of Greek Island embroideries.

We moved next to Anatolia with Craig’s Karapinar fragment that some of you will know.


He said that this piece impresses him with its definite color. He estimates it as first half of the 17th century.

The presentation of this fragment also shows how good mounting can enhance one’s ability to “see” what is there.

Craig’s third fragment is from a 17th century “Caucasian dragon” carpet.


It has great color and although “palmettes” and “dragons” can be discerned, this seems to be a fragment that is better appreciated if one can visualize, while looking at it, the larger composition from which it is taken. The drawing is attractively spacious.

Craig’s next piece was this West Anatolian “re-entrant” prayer rug fragment, 17th/18th century.


He said that the drawing suggests that this is a “village” rather than a “workshop” piece.

Here is a closer look image of its border system.


In informal discussion after our meeting proper was over, Craig brought out two additional fragments. The first of these was an East Anatolian Kurdish fragment, 18th century, with an unusual “zipper” border.


A closer look at this unusual border.


A second “brought-out-after” piece was the border of a Konya rug below.


Good, rich colors and a spacious major border design.

Again a closer look.


Craig estimates this piece as early 19th century.

Dennis Dodds had brought an Anatolian piece.


This fragment has great pile condition and graphics.


Its design is likely sourced in the Caucasian “shield’ carpets, but it is Anatolian. It is estimated to have been woven in the 18th century.

A closer look at the shield device.


Samy Rabinovic had three Anatolian pieces, all from Karapinar.  He said that he thinks all three of these pieces could be 18th century because the designs in them seem older than similar 19th century published examples.

He began with the one below.


There are colorful lappets on the upper side.


Using the approximate half here, one can imagine the full design.  Width: 47 inches, Length: 40 inches.  Strong saturated colors, mostly indigo, madder, madder-derived purple and yellow.

A closer detail below.


A second Anatolian piece may be the oldest and most interesting of the three, but does not show itself as well in the image below as it does “in the wool.”  It has wonderful, clear, old colors.


The design features a white hooked octagon on a white background inside a larger rectilinear “kaikalak” design with only two rams horns placed opposite. The connecting device is very well drawn and visible. Width: 44 inches, Length: 38 inches.

A third Karapinar, perhaps Karaman, fragment has very strong, saturated colors.


The connecting device used in 19th century pieces is the main part of the top medallion, so this piece might well be 18th century, but most probably is early 19th. Width: 36 inches, Length: 27 inches.

John Howe had a large Anatolian village rug fragment.  It has nice old, colors, including a vivid lighter blue and a large central medallion that provides a lot of graphic impact.


The most frequent attribution of this piece has been Konya, but Dennis placed it further west on the basis of the colors and one border system. This piece is very heavy and coarse with only 25 kpsi. Some examiners have estimated it as 18th century. It is approximately 4.5 X 7 feet.

Howe said that he feels that this is an example of a piece better mounted on a backing without a frame, since a frame would make it difficult to move and store readily.

Here is a closer look at the border system on this piece.


And of one niche in its central medallion.


The next set of Anatolian fragments we examined was from three kilims brought by a member of this salon who prefers to be anonymous in this report.

Here is the first of his three pieces.


He said that he was impressed when he got it by how similar elements in this design are to those in the piece in the book below.


As is evident, this piece features the “elibilinde” device, described as “hands on hips.”


Our anonymous participant’s second kilim fragment was the one below.


I didn’t catch the attribution, but it has a wonderfully archaic feeling about it.

Here’s a closer look.


It is estimated to have been woven in the 18th century.

His third example featured cross-panels, dramatic drawing and an unusual use of color.


The top panels have deeply saturated red and blue grounds, while the lower panels have much milder shades. He said that he’s not seen the ground shade of the bottom panel before.

Here’s a closer detail of the upper panels.


There was some question about whether this piece should be seen as a fragment.  It has both halves.  It may qualify because it is missing one border outside the white ground at its bottom.

I don’t have attributions, as I write, beyond “Anatolian” for any of our anonymous participant’s three pieces.

Bob Kent (on the right in the photo below)


had brought a rug with a “Transylvanian” design.


Although faded, this rug exhibited the characteristic spacious wide borders often seen in “Transylvanian” pieces. Bob said that while examining this piece he discovered for the first time that “lazy lines” can occur on pile rugs and what they looked like there. Here is a section of the back of this piece showing them.


Bob had also brought two Balouch pieces. Here is the first of these.


The field of this piece is a large-scale mina-khani-influenced design.  Bob said that he admired the two reds in its borders.  A piece, possibly, with good age.

Bob’s second Balouch piece is this one.


(Note: the “white” spot on the lower part of this image is a malfunction of my camera and is not on Bob’s rug.)

This piece seemed finer both in design and texture. It has a spacious, large-scale border. The lower left quarter of the field showed something of the color that this piece must have projected.


Some in the group said that while Bob’s three pieces were worn, they did not consider them fragments, since the entire outline of all three pieces was still there, including, in the cases of the two Balouch pieces, the end kilims.  Bob was able to point to considerable side selvedge damage on one Balouch and to intrusion on the end kilim of the other as aspects that might still qualify at least two of his pieces.  This discussion reinforced the notion that the question of what should be described as a “fragment” is not straightforward.

We were now on edge of Central Asia and Dennis Dodds had brought two Turkmen fragments, both from Chodor rugs.

The first Chodor pieces seems from an end.


The second Chodor piece is a portion of a field section but includes one border.


Here’s a closer detail of the second one.


Dennis said that although some are less interested, nowadays, in Chodor pieces, both the quality of the materials in, and the weaving of these two fragments indicate that they are worthy of notice.

Howe had Central Asian fragment as well. Again, a largish piece, maybe a little more than 4 X 6 feet.


It features two versions of Memling guls and somewhat over-sized, almost Caucasian-like, white ground borders. It is seen to be non-Turkmen Central Asian, perhaps Uzbek (Dennis thought maybe Kyrghyz).

The smaller Memling guls (in the minor ornament position) are an interesting instance of a device based on negative space.

The piece also demonstrates an important aspect of color choice in mounting.  The conservator chose a color close to that of the ground color of this piece and its use tends to minimize the distracting effect of the fairly large holes in it.  This may be a general rule: that if you are conserving a piece with holes, a backing close to the ground color will make it look better. (Howe’s large Anatolian village rug fragment, treated earlier, the piece immediately following here, and one later on, provide further opportunities to consider aspects of this backing color choice issue.)

Howe had also brought a Central Asian ikat fragment.


This piece is likely from the edge of a garment. Both of its parts are ikat: the diagonally-striped edge, as well as the dark-ground area. The dark part is so thin that if mounted on a black ground, the ground washes out the color of the piece entirely. So a red ground was substituted. This small, delicate piece is of the sort that needs a framed mounting (which this piece has).

Howe had another Turkmen piece, likely a section taken from the center-back of a gold-ground chyrpy.


He said that the thing to notice about this fragment is that while it is clearly part of a larger, heavily embroidered garment, the selection of the part of the design to be featured is of the sort capable of a graphic wholeness and independence.

Dennis had a framed silk kesi (slit-weave tapestry) fragment from the borderlands between China and East Turkestan.


He said that this is a piece from a garment, probably for a high ranking ruler. It is estimated to the Song dynasty. 12th-13th centuries.

This is another example of very skillful and successful mounting. Notice that the ground color of the backing on which the piece is mounted is very close to that of it predominant colors. Dennis noted that this piece has been mounted under glass so that it can “breathe.”

The last piece was an Italian velvet that Dennis had brought.


He said this is probably part of a cope or religious vestment, but may have begun life as a simple, but elegant, furnishing fabric, estimated to the late 15th or early 16th centuries. One can see in it the clear connection with Ottoman velvets. Not only did the Ottomans buy Italian velvets, Harold Keshishian reports that the Ottomans licensed Italian factories in their palaces to produce such things.

The session came to an end with folks wondering how 1 pm had come so quickly.

This was the first such effort by members of the Philadelphia club and the consensus was that it had been a pretty successful experiment.  Club members participating said that they would be considering what the focus of their next salon should be.

I was particularly pleased to have been alerted about it and invited to it.  My thanks to the Philadelphia Rug Society for permitting me to share this enjoyable session with you.


R. John Howe

Dear folks –

After I finished drafting the post above two things happened that should be included here.

First, Brian Morehouse has wrote me on the side with a comment on fragments.  He has permitted me to quote him.

Morehouse quote:


“…Fragments broaden our knowledge especially when filling in the various voids in carpet history. However, I would certainly not want to overestimate their visual importance, even though they often transmit a great deal of the visual vocabulary.

“The problem is twofold:
One: Carpets are often judged by their impact which is often lost in fragments.
Two: Carpets have a personality that is often developed in the filler ornaments. It allows the weaver self expression to come out; however, when much of the rug is lost much of this area to display her own unique design perception is lost and therefore the personality. “Even with rugs of repeating motifs, like Turkmen, often the nuance of an animal or unique rendering of a naturalistic flower, compared to a stiff geometric one, is vital to its personality. Something of the weaver is lost.
While I find fragments useful and often beautiful, I am wary that much of both the impact and personality have been lost.”
End of Morehouse quote.
A second thing happened, as I was in the last two weeks preparing this post for publication.  I had an email exchange with Alberto Boralevi, the textile scholar and dealer, in which I asked him some questions about some fragments of his I had found on the internet.
Most readers will know that Mr. Boralevi has a long-standing interest in textile fragments.  He has given me permission to use his indications to me as well as some links he included on fragments.  There are also some photos of his stand at a textile event.
Boralevi email:
Dear John,
Thanks for your message. You are right, I have always been interested in fragments as a mean for studying carpets, but also as decorative objects. I often tell to my customers that it is better to hang on the wall a nice and authentic rug fragment than an ugly or fake painting.I bought my first fragments in 1981 by an Italian antique textile dealer who had some interesting examples and later I got more pieces from old carpet restorers who often found them as patches used in repairing carpets…
If you are interested I can tell you several anecdotes about fragments, like for instance the reaction of my wife when I went back home with the five pieces I had bought in 1981 (I didn’t have a gallery at that time). She told me that I was crazy wasting money in such odd items, while I had to concentrate myself on things that I could sell more easily. Actually, I was not thinking to re-sell them but six or seven years later I sold that group (they were all classical, including 17th c. Isfahans, Lottos, Ushaks and similar) to an important international dealer making three times the price I had originally paid. Well, my wife then said that I had sold them too cheap!
In 1998 I made my first just fragment exhibition the first held at the same time in my Gallery and on the internet. At that time I was still using my old style film camera, not a digital one and I have prints of all the exhibits, while the online show was hosted by a friend of mine because I didn’t have any website on my own. It was a very rich exhibition with about 90 pieces, but some of them were really small. I must say that it was also very successful and I sold quickly almost all of them.

Ed: Color changes due to coding problems. RJH

I have small scans of those fragments that you can find at this link:
My second exhibition was in 2006 in my new (smaller) gallery (not on internet) and here you can find some pictures here:
And here for more examples:
I am still collecting and selling frags and last September my stand in Sartirana was almost entirely devoted to them (see the attached pictures)

Of course, you have my permission to refer to me and to take as many pictures as you like.Kindest regards,


Alberto Boralevi
Antique Rugs and Textiles
First Floor Palazzo Frescobaldi
Via S.Spirito, 11
50125 Florence – Italy
tel./ fax +39 055 211 423
mobile +39 335 542 7107

End of Boralevi quote.
Here are some photos Boralevi included of some of his fragments on display.
BoralleviStand AB 2
BoraleviStand AB 4
My thanks to Brian Morehouse for his comment and to Mr. Boralevi for this wealth of additional material on fragments.
As you can see, this started out to be a report on a nice salon on fragments organized by members of the Philadelphia Rug Society, but has turned out to include much more.
As I said at the beginning, there has been a recent Rug and Textile Appreciation Morning program on fragments, that I will report on, eventually, on my Textiles and Text site.

I hope you have enjoyed what has turned out to be a rather fulsome look at textile fragments.

R. John Howe