On October 4, 2008, John Wertime
and Wendel Swan
gave a Rug and Textile Appreciation Morning program here at The Textile Museum in Washington, D.C. on “Long-pile rugs.”
This program was explicitly based on Wertime’s important “Back to Basics” article in Hali’s issue 100.
Wertime took the lead saying that the earliest rug-like objects were likely animal skins. And it was with a couple varieties of such pieces that he began.
He said that whole animals skins were useful but that with the emergence of the needle and thread, humans could make animal skin/fur objects of specific sizes and shapes. The first piece he referenced was such a pieced animal pelt “rug” made by sewing strips of differently colored but undyed pieces together to make a “blanket” of a given shape and size and with a concentric square design. The image below is from his Hali, 100 article.
When humans discovered dyes, such animal pelt strips were “dip-dyed” in particular colors. “Dip-dyeing,” Wertime said, is accomplished by immersing an entire piece of animal pelt into a dye pot. Again, he referenced such a piece from his Hali, 100 article.
Wertime next rehearsed the early history of emergence of textiles and weaving. He said that plant fibers were used first. Flax is often cited, but other plants are likely early sources as well in particular parts of the world. The earliest, domesticated sheep, he reminded us, had short, colored coats. More like those of deer. Such fibers were too short to spin and colored wool does not, when dyed, produce the attractive shades that are possible if one dyes white wool. So the use of dyed wool in woven textiles likely “came into its own” only after longer-haired, white sheep had been developed.
The early techniques used to produce wool textiles included weft twining, tapestry and various kinds of wrapping, for example, sumak. These techniques result in textiles that are flat-woven. Wertime suggested that the earliest “pile” rugs were likely a variation on tapestry. Such pieces he said were made with a structural set of wefts interlacing the warps but an additional set of weft was also inserted, perhaps in their own sheds. This latter weft was then “pulled up” with a hook to form a “faux” pile (“faux” because there are no knots as there are in actual pile rugs).
One of the points that Wertime made in his “Back to Basics” article is that some ancient techniques can persist long after others have been developed. Old methods do not always die out. This method of creating “faux” pile pieces from flatwoven fabrics is one such. Likely the earliest form of creating pile rugs that we know, it is still used today in such areas as Siirt in eastern Anatolia.
Wertime and Swan had three of these Siirt “faux pile” pieces in the room. The first used horizontal stripes of different lengths to create a diamond form in negative space.
A second piece had a design of rich colored stripes.
Wertime said that such Siirt pieces always have only undyed wools of different colors. The wool is from a kind of mohair goat. He added that these Siirt pieces seem also always to have cotton warps.
A third Siirt rug had a niche design, again defined negatively.
This piece has a second level of design in the gray area under its niche. The fibers in this area have been pulled deliberately in particular directions. This “colic” effect works to create a series of subtle diamond devices.
Despite the fact that you can encounter such Siirt pieces in flea markets, folks like Wertime and Swan, who are very interested in the earliest weavings, do not write them off since, although frequently very recent, they exhibit perhaps the oldest “pile” structure we know.
Wertime and Swan now moved to a second way in which a kind of longish pile can be fashioned in a woven rug. Again there is a basic structure of warps and interlacing wefts, but in addition, there is a second set of wefts in separate sheds, that are pulled up at regular intervals to form loops. Sometimes the loops are formed over horizontal rods. Such loops can be either left in their loop form or cut so that their ends are like those of knotted pile rugs. The knots in Tibetan pile rugs are often tied over a rod and then cut before the rod is with drawn. Here is a drawing of such a Tibetan loop usage with the rod in place.
Wertime referred to two distinctive loop pile rugs. The first is one in which the loops are firmly anchored in a plain-woven structure. He had two rugs of this type in the room.
The first was the piece below.
Wertime said the loops were anchored by interlacing them with the warps.
Here is a second rug with extra weft loops on a plain-weave ground.
Wertime said that this piece was dip-dyed after weaving and he is confident that its rich golden color is from natural sources. The detail of this piece below lets you see something of its back and provides a closer look at its loops.
A third rug with looped pile is the one below. It is of a second type of looped structure.
This rug begins to look more like tribal pieces that are more familiar, but it has a very distinctive structure. It has a looped pile, but different from that of the golden Kurdish rug above. In the golden rug the loops are anchored in the structure. The loops in this colorful piece are not anchored and can slip. That is, you can pull on one end or loop and the entire row of loops will come out. This is why this structure is called “slip-loop pile.” In Hali 100, Wertime reports that the attribution of this piece is not certain. Perhaps Afghanistan.
At some point in time Wertime said pile knotting emerged from the flatwoven tradition. While it is not entirely clear how this occurred Wertime said that there are some indications that allow a plausible argument. First, he noted that the pile knots in the oldest pile rugs we have are always symmetric. Here is a drawing of a symmetric knot.
Wertime thinks that the symmetric knot is an adaptation of some species of wrapping. Here is a drawing of two short sections of sumak wrapping in which the weft is first taken over two warps and then brought back under one warp before moving forward again.
In the variety of sumak in which we are interested there would be two rows of structural wefts between the two rows of wrappings in the image above, but that is not our concern here. Here we want to focus on the character of the wrapping as it goes around any two sets of warps. Below is a detail showing the wrapping on just two warps and below I have repeated the drawing of a symmetric knot for ease of comparison.
You can see in the sumak detail that one has only to move the left-hand loop so that its end comes (as the right end does) in front of the wrapping “collar” to convert this section of sumak wrapping into a symmetric pile knot.
Weavers using sumak knew that one could create designs by varying the length of a given horizontal line of such wrapping. Some weaver may have decided that she wanted to have a VERY short section of a given color, created a section of sumak that was only two warps wide and then decided that this unit would be more stable if she brought the left hand end in front of the wrapping collar as described above.
In this short step she had made two major advances in weaving. First, she had created the equivalent of the symmetric pile knot. And she had made more explicit something that was present, but not accentuated in the sumak structure, the very great potential of digital design.
Now this is all conjecture, but it is plausible conjecture. One can see how it could happen, as well as the likelihood that it would…repeatedly. Anyway, that is one view of how pile knots emerged and why the early version seems always to be symmetric.
From this point forward, Wertime said, all the pieces shown have symmetric pile knots.
The next piece was the green piece below.
This piece, Wertime said has single rows of long-pile knots separated by multiple rows of weft. Such narrow pieces were often sewn together to make larger formats. The narrowness likely signals a nomad source since narrower looms are easier to move from place to place. Wertime added that this piece is one of those woven first and then dip-dyed to produce this green.
The next piece had the same structure with lots of weft between rows of very long pile, but is wider, is in one piece, and has large red cruciform device at its center.
It is very coarse with large knots, having only one knot per square inch. Its owner reported that Harold Boehmer examined it recently and opined that its red is natural. This piece is attributed to Karapinar, just east of Konya area where strong natural yellows are also not infrequent.
The next two pieces were sleeping rugs from Eastern Anatolia. Such rugs are long enough that one can sit down on the lower part of one half and draw the other half up and over to cover yourself while sleeping. Here is the first one.
Again, rows of knotted pile separated by a number of rows of weft. There is brocade at the sides
and also on a central cross panel.
The second of these two sleeping rugs is the one below.
This time the brocade is at the side but is also used more frequently to decorate flatwoven areas between rows of knots.
If you look at the preceding overall image of this rug above you will be able to see traces of this more general use of brocade decoration.
Wertime now moved to long-pile rugs from Central Asia.
The first of these was done in long, narrow strips and then sewn together.
Wertime said that this is another rug the strips of which were dip-dyed after weaving.
A second Central Asian piece was the one below.
This piece was also woven in strips but with a more complex design.
Here is a closer corner of this piece.
Below is a detail of its back after it had been taken down from the front board.
These long rugs are woven with symmetric knots tied on alternate raised warps. One result of this is that the designs done in pile on the front are almost or entirely obscured on the back.
Another of these Central Asian pieces is the one Wertime is standing in front of above. Here is an unobstructed view of it.
Here is a closer detail of the piece above.
Because they are woven by tying knots on alternate raised warps (that is, there is always an “untied” warp running behind each knot but not part of it) this pile structure is “offset knotted” by nature. This may be why diagonals are sometimes stepped as in the design above. This stepping reduces the steep angle of the offset knotting would otherwise produce.
Wertime indicated that at least some of these long sleeping type rugs are thought to have been produced by Central Asian Arabs.
Wertime turned to western Anatolian with his next example.
This rug is unusual, Wertime said, in that rows of long pile knots alternate with narrow strips of short knotted pile.
Here is a closer look at the pile-side of this piece.
The next rug was a contemporary Turkish “tulu.”
The colors in the piece are from undyed brown and white wool.
The next two pieces were attributed to Central Anatolia. First was this sizable rug.
Wertime described this piece as a yatak or sleeping rug. He said that they are sometimes mistaken for Kurdish weavings, but are not. Here is a closer look.
Below is a second rug of similar design, but possibly a little small to be a yatak or sleeping rug. This particular example was probably woven by nomads in the Konya area, but not Kurds in all likelihood, according to Swan. Although these rugs have longer pile, they are not as heavily wefted as gabbehs from Iran.
And here is a closer detail. The “tile” design in this piece is seen by some to be sourced in “cloud collar” designs from east Asia.
The wool in this rug is very soft.
The last two pieces were described as having longer pile than most rugs but not as long as that of the sleeping rugs treated above. They were probably not intended as sleeping rugs, but regional differences might indicate otherwise.
The first of these was the piece below from Zakatala in northern Azerbaijan.
This handsome rug is a type only identified: a “Zakatala” within relatively recent years. Their designs often resemble those of rugs from the Konya area.
Here is a closer corner. This rug may have goat hair edging.
And a closer detail of the field.
Most rugs have wool pile that is “Z-spun” and “S-plied.” Rugs with “S-spun” and “Z-plied” are rare-ish. Mamluk rugs, for example, are famous for having this latter structure. Some Zakatalas do, too, including, Wendel pointed out, this example.
This is one instance in which a marked structural distinction can aid in making a firm attribution of a Caucasian rug.
The last rug was the Kurdish piece below, woven in eastern Anatolia.
This rug is said to have a “baklava” design because of the similarity of its devices to that pastry.
Here is a closer corner.
And here is a closer detail of the field.
The reds in this piece exhibit both cochineal and madder hues.
Wertime ended this session by returning to the issue 100 of Hali in which his “Back to Basics” article appeared. He noted that in this same issue Murray Eiland, Jr. had written an article on Kerman rugs.
Kerman, Wertime reminded us is often seen to be at the acme of Persian weaving. Kerman wool, and design and color usage are widely admired.
Here is a straight-on look at the piece on the right side in Eiland’s article page that Wertime has open above.
It is hard, Wertime said, to think of the wonderfully sophisticated Kerman rug as related to the relatively simple, even crude pieces we have examined in this “rug morning,” but that is the case.
The wonder of this great Kerman rug, and that of others like it, have their roots, precisely, in such primitive pieces.
The session adjourned and folks came eagerly to the front to get their hands on these interesting pieces.
My thanks to John and Wendel for permitting me to share this virtual version of their fine program with you, and for their editing assistance. Thanks, too, to Amy Rispin, who took an excellent set of notes for me.
I hope your enjoyed this virtual version of the Textile Museum Rug and Textile Appreciation Morning program.
R. John Howe