A number of the rugs in Jim Blackmon's presentation have a distinctive structure.

They are made as follows:

"The weaver opens a plain-weave shed, automatically selecting half of the warps. She ties 'symmetrical' knots on just those warps that are raised. With the shed still open she inserts a weft selvage to selvage. Then she closes the shed and beats everything into place. Opening the opposite shed she ties knots on a new set of raised warps and lays in another weft..." source: Marla Mallett, "Woven Structures, p. 58

I notice that this structure is described differently by different folks. I have most often heard it described as "knots tied only on raised alternate warps." Jim Blackmon uses the expressions "warp-predominant plain weave" and "warp-faced back."

There are some aspects of this structure worth noting.

First, the pattern on the front of such pieces is either invisible or only faintly visible on the back. This is precisely because the pile knots are thrust somewhat to the top of the piece and the warps not used in each row of knots and the wefts combine in a kind of plain-weave on the back of the rug that either partially or entirely obscures the pattern on the front. So such peices are very easy to identify simply by looking at their backs.

A second thing to note is that rugs with this structure are automatically "offset knotted" something we don't automatically notice when we examine them.

Third, Marla Mallett's description above of how this structure is made provides an instance show why the expressions "single-wefted" and "double-wefted" that we seem unable to avoid are unsatisfactory (Marla argues that they are so for several reasons).

One of these is that they are ambiguous because they do not indicate whether the wefts are in the same shed with either the knotted rows or with each other. In the case of this structure, each weft is in the SAME shed as the preceding row of knots. But that would not have to be true in other fabrics and in fact the more usual practice is to have rows of knots and the intervening rows of weft each in their own sheds.

I think Marla suggests that we should refer to the number of wefts in each "pick" to avoid this particular ambiguity. In this case each row of knots and a single weft are in the same shed and therefore there is a row of knots and a row of wefts above it in each "pick."

Marla also points out that, although knots are not tied on contiguous warps in this structure, this is not a species of "jufti" knotting since knots are tied on only two warps and on every two warps. No warps are "skipped" so to speak as the two center ones are when a jufti knot is tied over four warps.

Marla's suggestions for describing this structure include: "symmetrical knotting on alternate warps," or "symmetrical knotting on a warp-faced ground," or "symmetrical knotting on an open shed." She sees no basis for preference but sorrows that there isn't simpler accurate language for such description.

Perhaps she and others will elaborate on or correct my extensive drawing on her work in what I have written above.

Regards,

R. John Howe

This makes a lot of sence. (The backs aren't the result of poor quality images)
Now I understand why the backs seem machine made. Next time I'll have a better look at the rugs that pass me by. Hwo knows how many have disapeared in the rubbish.
This undermines my question in the previous "New/Old" posting.
An idea...asymmetric, symmetric.
Triple symmetric because of three warps?

Thanks for the explanation.

Best regards,
Vincent

One more item from Marla Mallett's discussion of "Knotting on Warp-Faced Grounds."

When we see narrow strips woven we often assume that this is the result of the state of loom technology available to the weaver and that wider looms were not available.

While the state of loom technology in a giving weaving area could still be a factor limiting the width of textiles woven there, Marla provides another reason why some of these non-Turkmen Central Asian pieces have been woven in narrow strips and then sewn together.

Here is her paragraph from p. 58:

"With a dense wool warp, it is difficult to open a clear shed. The wider the weaving, the more difficult this is. Thus nearly all warp-faced weavings are narrow. Some Central Asian weavers made coarse, shaggy pile rugs called "julkhirs" on warp-faced grounds by knotting on open sheds but they did this by weaving long, narrow strips which they later sewed together."

So loom techology is not alway the only variable affecting the width of a weaving.

Regards,

R. John Howe

Many thanks for this beautiful Salon. I like this ‘distinctive structure’ very much and would like to seize the opportunity for a bit of “statistics”:
On a textile which was knotted on the open shed, there are several possibilities for the arrangement of the knots, depending on the number of wefts (actually the shed-changes) and on the decision of the weaver.

"a" If there is an even number of wefts all knots are vertically aligned and are tied on only one set of warps (on my diagram on the blue one).

With an uneven number of wefts both sets of warps are holding knots (although in each row still only half of the warps are “used”) and as an obvious consequence there is a certain offset-effect. There are two kinds:

"b" On alternate rows the knots are tied to the same pairs of warps.

"c" The weaver “extra-offsets” the knots, i.e. the right warp of a given knot in the first row is the left one of a knot in the third row. So the relation of knots of alternate rows is exactly the same we know from “normal” offset knotted carpets.

I suppose that most julkhirs are knotted as on my diagram “a”.
No 11 which you describe as having “three wefts between rows of knots“ would be either “b” or (less likely) “c” - if the weft which shares the shed with the row of knots is one of the three wefts.

I would be very interested to hear of any julkhir of type “c” and I think it could be nice to know how much of the julkhirs have their knots aligned (“a”) and how much are offset (“b” and “c”).

I also would be very interested to hear how diagonals are drawn on type “b” julkhirs. Are they as on diagram “b1” where the knots forming the diagonal are alternately tied on both sets of warps or as on “b2” where they are all tied on the same set?

By the way, A. Felkersam wrote in 1914/15:
“All Uzbek are producing carpets, regardless whether they are nomadic or settled. In Turkestan only those pieces are called Uzbek-carpets on which on the back side the ornament of the front isn’t to be seen at all (S.M. Dudin).” (after the German translation: A. Felkersam, Alte Teppiche Mittelasiens, Schletzer, Hamburg, 1979, p. 87)
“In Turkestan” is rather important in this paragraph because later he goes on saying that:
“While in the case of the Kirghiz and Turkmen only the nomadic tribes produce carpets, in the case of the Uzbek especially the settled ones concern themselves with this craft - or, better said, the carpets of the settled Uzbeks are those which deserve our attention, because they are of high quality. Thereby it can be explained why these carpets in Turkestan and even in the Caucasus aren’t ascribed to the Uzbeks, but usually named after the place of manufacture, i.e. “Beshir-” and “Kerki-carpets”. However sometimes these carpets bear the name of an Uzbek tribe, which made them, for example “Kizilayak” (S.M. Dudin). But one has to pay attention that one Uzbek tribe is called “Beshir” and that there are also Turkmen tribes (see for example Bogolyubov) which bear the names “Kizilayak” and “Beshir”. Hence it appears that these questions can’t be considered as solved.”
He labels the famous “Ersari-Namaslyk”, Tzareva No 98, as “Uzbek Bashir-Namaslyk, 17th century”...

Best regards,
Christoph

Thanks for this nice clarifying graphic display again.

It is my impression that all of the pieces with symmetrical knotting on alternate warps are of the "b" variety.

And I must be a little "thick" tonight because I'm not fully taking in how the number of wefts necessarily affects whether a weave is type a, b or c.

I think too that perhaps I have left an over-simplified impression in my indication of which rugs in this presentation likely have the "symmetric knot on alternative raised warps" structure.

For example, I included rug 18 in my listing of those with this structure but Wertime is at some pains to say that although the structure on rug 18 has a faint back, its structure is quite different.

Here is how he describes the structure of rug 18:

"...Wool and goat hair slip-loop pile on wool and goat hair weft-faced plain-weave ground....The reverse detail show a striped weft-faced plain-weave ground compounded by slip-loop pile..."

Rug 18, Wertime says has "widely-spaced rows of pile held in place by a weft-faced plain-weave ground." So there are often more than three picks of weft between rows of knots and the pile is made with the slip-loop technique. A very different structure, indeed, from the "symmetric knots on raised alternate warps" sort.

One thing more I have not passed on about the "julkiyrs" is that both Wertime and even Murray Eiland seem agreed that they were woven by Central Asian "Arabs." This is one place where attribution has apparently moved beyond the Turkmen, non-Turkmen dichotomy.

Regards,

R. John Howe

Regards,

R. John Howe

The only julkhir I’ve access to at the moment is of type “a”, although there are occasional “errors”. Looking again at the back of Fig. 16 in HALI 100, p. 95 there seems surprisingly to be a random change between three and four wefts which doesn’t pose any problems since we’re dealing here with a weaving without any knotted ornaments.
On the number of shed-changes (which are done after the insertion of a weft) depends which set of warps is risen, on which set of warps the knots are tied. With an even number of wefts always the same set of warps is risen and therefore the knots are vertically aligned. An uneven number of wefts results in either “b” or “c”, depending on the decision of weaver.
Is there anyone who could report the structure of his/hers julkhir(s) to build up a little “database”?

Best regards,
Christoph

I am inserting a couple of photos showing a Turkmen tend band. You may be able to see from the shot of the front that the knots visible at the top of the pile sections seem to have spaces or gaps between them, similar to David Letterman, Alfred E Newman or Terry Thomas. This hints at the construction.

From the back, the distinctive faint appearance of the pattern is only visible where the knots peek from between the warp threads that are "depressed", since the knots are all tied on the warp threads that are above.

Hope this helps.

Patrick Weiler