The Joy of Fragments: Tim Hays, Ali Aydin, and Wendel Swan, Part 1, the Lecture

On October 4, 2014, Tim Hays,


Ali Aydin


and Wendel Swan

gave a Rug and Textile Appreciation on “The Joy of Fragments,” still in the Myers Room of The Textile Museum, at the old S St. location. 

Ali arranged for many of the in-room examples and Wendel produced the Powerpoint presentation for the program, but both had to be out of town on the day of the program and could not participate except in absentia. 

So Tim gave the lecture and led the session.


The lecture began by presenting you with an array of fragments all on a single slide, announcing the title of the session. 

WendelTitle slide

(click on image above to see a larger version)

Tim began by indicating that George Hewitt Myers himself collected fragments, and so this program was within one of the traditions of the TM since its founding.


Among the fragments that Mr. Myers collected are this wonderful pre-Columbian fragment (you can click on most images to see a larger version),


this Spanish assemblage of fragments that resembles a complete and wonderfully composed small rug,

Slide2lowerleftand the Mughal fragment below, which was cut so as to create a beautiful and artful composition in a very painterly fashion.


One of the debates in the area of the subject of “fragments” is what is to count as one.

Europeans tend to take a strict view and call something a “fragment” if any part of it is missing.  Like this mounted fragment of a mid-19th Century Cappadocian kilim from Central Anatolia.


We, in the U.S. tend to reserve the term “fragment” for instances in which a quite substantial part of a given piece is missing. 


However, many are reluctant to use the term “fragment” to describe a piece that is, basically, “all there.”


Tim, Wendel and Ali proposed what seems like a reasonable terminological middle ground.  First, that pieces that are only small pieces of a much larger textile or are substantial pieces from which quite a bit is missing should be described with the frank term “fragment.” 


But in instances in which a piece is missing minor parts but is still, basically, “all there,” the piece should be described as “fragmentary.”


There is also the question of why we collect fragments at all? (and there are those who don’t)

Here are some reasons that are often offered:

  • Because they can be inexpensive and therefore within collecting reach, even for those collecting on a budget.
  • Because in spite of what they are missing, they can be beautiful (in fact, Rugrabbit dealers seem to have discovered that details of pieces are often better “teasers” than are images of the complete pieces themselves).
  • For historical documentation of rare or otherwise unknown textiles (more about this, shortly, below).
  • Size.  Fragments are usually smaller and can be more readily displayed (think of the difference between a complete two-panel kilim, 1o feet long, and an attractive part of only one of its panels).
  • Fragments can be intellectually stimulating (for example, what did the designs on the entire piece look like?)


The important book, above, on fragments is also one of the most important books on early carpets. A Swede, C. J. Lamm, collected about 40 early fragments in Cairo, many probably from the Fostat area in Cairo. His collection included centuries of pile weavings that are otherwise unknown.

The cover, above, shows a fragment of the early 15th Century from Turkey or the Caucasus, while below is a fragment from Egypt of the 8-12th Century with a form of unknotted pile.



To follow and exemplify the last point in our tabulation above, the fragments in the Lamm collection are often very small.


But the fragment above, only 21 x 11.5 cm, has been the basis for the reconstruction below of the likely design of a larger area of this rug.  Note the indication of where this fragment “fits” in this reconstruction.


Below is a small fragment in a Washington area collection that is from 1,200 BC and is perhaps the oldest known pile weaving.


It is not published and this image was presented to the rug world for the first time anywhere. With multiple wefts, it would have served the same purpose for sleeping as the more recent rugs we call gabbehs or yataks.

With this fragment and the images that follow, we witness more than 3,000 years of weaving history.  There are other fragments of ancient or antique rugs that are structurally nearly identical to the 3,200 year old sleeping rug.

Below is a modern filikli,


while, below again, is a fragment of a sleeping rug from 200-400 AD.


They have nearly identical structures.

This very old fragment and its contemporary counter part let us see, as John Wertime said in his seminal Hali article on primitive pile rugs, that “the new does not always slay the old” and that “some most primitive method may survive for some special use.”  This ancient structure is still being woven today.

It is well known that designs travel widely, even between media.  The colorful fragment of a Central Anatolian rug (on the left in the composite image below),


uniquely demonstrates the dissemination and sharing of motifs over millennia.

This outer border motif


can been seen in this Anatolian or Gordian wooden stand from 800 BC


WendelGordian box detail 2

and as the outer border of a 19th Century Tekke engsi.


While the motif of the inner border of the Anatolian rug,



is reflected and doubled in the lower border,


after which it closely resembles the lappets that we see on Anatolian yastiks and rugs.


We learn from this fragment to not assign most motifs exclusively to certain weaving groups or even particular media.  There are lots of transfers of designs, for example, from architecture and pottery to textiles.

Another instance of design comparison over time is from further east. 


This framed fragment may be part of a large carpet from Eastern Turkestan or Western China, possibly from the 17th Century or even earlier.


It has the very thin warps of Ming Chinese rugs, but the brown is some kind of hair, perhaps goat or yak.

Focusing more closely on its border element,


we can compare it to the 19th Century descendant (below) that was probably made with synthetic dyes that have faded.

Slide10right(click image for larger version)

I am unaware of any other piece with a border design comparable to the early fragment, but this design seems to have traveled a long way in time, at least once.

The painting below, by Hans Memling, in the Prado in Madrid, gives rise to our present day use of the term “Memling gul.”

In this Photoshop reconstruction that Wendel did several years ago,


you can see how closely the Memling guls were juxtaposed in 1490.

Although the rug Memling painted was probably flatwoven, note that it has three horizontal guls grouped in almost a tile-like manner.

Subsequent yellow ground Memling gul rugs from the Konya area, such as this from the early 19th Century, display increasingly separated guls.

And nearly all of them have only two guls abreast.


This yellow ground pile fragment, from the Konya area, is one of very few which also has three guls across.


This fragment retains for us to the character of the Memling gul usages that Memling was looking at and painting in the 15th century.  It might even indicate that this fragment is from a rug older than most Memling gul rugs, and fragments from rugs, which have only two-guls on the horizontal.

The Memling gul is one of the most frequently and widely used rug and textile design devices.  And even a fragment with a single Memling gul can be attractive.

This Northwest Persian fragment


and this Eastern Anatolian fragment


both have isolated Memling guls. They are both attractive fragment examples for differing reasons.

One does not need more of the originals to appreciate them, and they illustrate how many different versions of the Memling gul motif have been fashioned.

As with an intact textile, fragments should have aesthetic appeal because of their composition, visual integrity and inherent quality.

Generally speaking, we expect a fragment to have superior qualities of age, color, wool or other qualities. This end, below, of a Talish rug


is good, but not exceptional. Its border elements are unusual, but the Persian fragment, again below, is much older and rarer and has a wonderful sense of composition and color.


Another dimension on which fragments can be evaluated is inherent quality.

Here are two fragments that are about the same percentage of what the intact weaving from which they were taken would have been.

The first is this fragment of a Kazak rug, approximately half of its original size.

Wendelmediocre Kazak fragment on black

Below is a visual reconstruction suggesting what the complete rug would have looked like.


The second is this Salor Turkman fragment.


Although the notion of “inherent quality” can be debated, most would agree that the Salor fragment, immediately above, is clearly more appealing than the Kazak, further up.

Often, a fragment of a great rug is better than the whole of a mediocre one.  Inherent quality matters.

The next dimension on which fragments are evaluated might be called “readability.”  Because fragments generally come from worn carpets, wear and holes can be problematic. The condition of a fragment becomes an issue, as it does with a complete textile.

However wonderful they may have once been, this Cairene


and this Persian fragment, are worn to the point of distraction for some.

Wendel15 C Persian border fragment 

The colors and drawing in this early Anatolian fragment


are a bit easier to see, ***** but the brilliance of the original is retained when holes and damage do not distract the eye.

Concentrated study is required to understand the patterns in some battered fragments,


whereas in others, in better condition, the splendor of the original weaving is preserved.


Although of two entirely different styles, both of these would grace anyone’s wall.

Here is a test. Examine the two pieces below.



Do you think that one or both or neither of these is significantly fragmented?

The upper one is the fragment. The lower one is a complete Northwest Persian or Caucasian bag.

Let’s examine the upper fragment more closely.  It is part of the center field of an Eastern Caucasus rug.


(click for larger image)

Notice that the bottom red line at the edge of the top the field continues all the way to both edges of the piece (and that the red edge line at the bottom border does not). 

If we had it in hand we might pretty quickly discover that a strip of the trefoil border was attached to the top.  The warps in this addition run horizontally.

Here is how this piece looks without this top border addition.


You can see that there is a sense of incompleteness.

Let’s go a little further with this fragment in a slightly difference direction.  We have said that it is part of a longer rug.  Here is a reconstruction indicating what this longer rug might have looked like.


Below is similar rug with an outstanding ivory ground.  Compare this piece with the reconstructed “rug” above.  Which to you prefer?


The filler in the reconstructed version is distracting and the ivory ground piece is clearly preferable. 

Interestingly, the problematic character of the fillers, in the full reconstructed rug, is less visible and troublesome in the fragment with the top border added.


The problematic nature of the fillers needed the longer vertical length to make itself more evident.

Sometimes a fragment of an object is preferable to what the whole might have been.


That is the case here.

We have often puzzle over what the whole rug, from which the fragment below came, might have looked like.


After years of speculation, the rug below appeared,


suggesting that the medallions repeat.

Examination of this second piece reveals that it is also is fragmented.  It is part of a longer rug, similar to the fragment further above. The piece immediately above was cut below the partial blue medallion and the bottom border was added.

Let’s look at a few more attractive fragments.


These Mughal fantasy carpet fragments are just incredible.


Below is a spectacular fragment of an Uzbek saf.


Below, sections of a silk velvet have been composed to create a well-balanced image that you would not see in a garment,


while, down again, is a silk Caucasian embroidery.


The silk Mughal fragment, below, is a wonderful artistic achievement, just by itself,


and the delightful 17th Century Central Anatolian fragment, down again, is well-composed.


Both are so delightful that we almost don’t have to wish to see more of them.

This untrimmed Anatolian rug fragment has its own special appeal,


as does the velvet embroidery, below, 


and the following Bakhtiyari fragment.


Size is another advantage fragments often have.  The frequent smaller size of fragments allows us to display them where we might not otherwise be able to show the entire piece.

Fragments often permit display flexibility.  Below are, in turn, a Turkish kilim and an early Persian rug border, displayed, first, vertically,



and, then, horizontally.


(click both horizontal images for larger ones)


 This Turkish kilim and early Persian rug border differ in style, but we could enjoy either of them on the wall, mounted either vertically or horizontally.

Some fragments, especially kilims, offer the opportunity to simply enjoy color.


Even when there are plenty of holes.


The Turkish kilim below,


and this Turkmen tent band fragment


are easily displayed even though it would have been difficult for most of us to display them as they originally existed.  The kilim might have been 10 to 12 feet long and maybe four wide.  Turkman mixed technique tent bands are often about 44 feet long.

The Persian kilim fragment, below, is small, but unusually attractive.


Whether the rest of it would have pleased us even more is largely irrelevant. We can love this as it is.

The Persian kilim fragment, below,


and the Salor main carpet fragment, down again,


could now easily be displayed to great effect.

Even fragments, such as this from a 19th Century Kirman, can have uses.


A similar use was made of this Turkish kilim.


Of course, if you have lots of old, worn carpets and some sharp shears, you too could upholster the furniture in your private railroad car.


With the Powerpoint presentation ended, Tim4Fetzen

Tim took questions and said that we’d now move to the fragments that had been brought in.

To see those you need to go on to Part 2 at this link:

R. John Howe

Comments are closed.

%d bloggers like this: