Here’s a photo taken by Sophia of a Yomut chuval detail that often puzzles people: diagonal outlines that are smooth slanting upward to the left, stepped slanting upward to the right (as seen from the back).

There is no offset knotting in this piece, but the contrast in the different diagonals might lead one to think so. Rather, this disparity is just the result of knots that are pulled consistently to one side as they are tied. Thus each is set on the warps at a slight angle. Line up a bunch of slanting knots along diagonals, and they make a smooth line going in one direction, while they separate visually going the other way.

Marla

Looks like a left- or righthanded woven rug.
No warpdepression here.

If warpdepression is the case, and the weaver is righthanded, the left warp of each knot is the upper one. Lefthanded: It's the other way around.

Only found one kurdish bagface with offset knots in the collection. No Caucasian, no Qashqaï. No Beloudch. So far Nada.

Question: The higher the knotcount, less need for offset knotting, because you can easy change colordesign within a specific pattern. It will not show at the frontside.

Best regards,
Vincent

PS. If any offset knots pop up, I'll let you know. (14.687 rugs to go, so some patience needed.)

We will be eagerly awaiting the results from those 14,687 remaining rugs!

I have one quibble. You said: “The higher the knot count, less need for offset knotting, because you can easy change color design within a specific pattern. It will not show at the front side.”

This is a good point, and certainly often true. The exception was when knot counts became very unbalanced. In Turkmen weavings, offset knotting seems to have been used more frequently as the knot counts increased, because sometimes there are twice or even three times more knots in the vertical direction. Then offset knotting was sometimes used to “square” motifs, so they were not so squashed. The ak chuval on my web pages has 10 knots horizontally, 36 knots vertically, and 360 per square inch!

Best,
Marla

Here is a scan of the back of a Yomut okbash that shows the same diagonal displacement pattern as the Sophia Gates chuval. The diagonals that in the photo run up from right to left are relatively smooth, while the diagonals going up from left to right are stepped. My guess is that without using offset knotting it is possible to skew or slant the knot in one direction to smooth out the line to suggest a neat diagonal. But this would work only in one direction, since skewing it in the opposing direction would distort the weaving by setting up extra tension. Does that make sense?

Best regards,

Ken

Yes, that makes sense. But actually, I oversimplified the issue when talking about Sophia’s chuval. The varied diagonals are partly shaped by the way each knot is tightened down on the warps, but the relationship between the knots and the wefts is also involved. In other words, a knot’s slant also depends partly upon which weft lies against which part of it. By changing the weft sequence, one can actually change the direction of the knot's slant, or at least straighten it a bit.

Other weavings—without any knotting—often display the same phenomenon. Here’s a double-woven band in which two quite different diagonal features are produced, with no change in the process used. Yet one diagonal is straight and smooth, while the other is stepped.

Actually, I was just trying to make sure that everyone realized that this feature had nothing to do with offset knotting... Let me know if I need to make a diagram showing how this works with knotting.

By the way, John, this issue plays a quite significant role in the Neff and Maggs distinctive back-side appearance thing...It's one technical reason for differences in otherwise similar weaves.

Best,
Marla

Here’s how changing a weft sequence can change the way rug knots are set on pairs of warps. Depending upon the weave balance and also how the knots are tightened, one side of each knot can sometimes slide down over part of the weft that precedes it.

Marla

Is this change of weft sequence intentionally used with any frequency in any particular group of pile weavings to convey a particular effect? Is is a feature that might be used for identification?

Many thanks. Best regards, Ken

I’m not aware of any weavers anywhere who have reversed plain-weave sequences for the express purpose of altering the character of opposing diagonals. (Various complex pattern weaves of course feature this kind of alternation all the time.) Nor have I seen a knotted-pile weaving in which this has been done consciously. The inherent differences in opposing diagonals have just been accepted by weavers as “the way it is” with certain weave set-ups. When the disparities have been disturbing, I think it’s more often been the design that has been altered, than the weave structure. Of course as weaves become finer, disparities like those in the weavings above tend to disappear.

Think of some possible variations on the knot/warp/weft combinations in the drawings above: In rugs with either one or three wefts between knotted rows, the knots may shift constantly in orientation, slanting slightly in one direction, then in the other. Senneh rugs, with their single wefts, offer perhaps the most dramatic example of this. With either of the knot orientations I’ve drawn, the knots can be tightened in either direction—either to accentuate the slant, or in the opposite direction, to tend to straighten the knot. Baluch weavings, with their very wavy but smooth, regular knot rows, may be the best example of the first. If alternate warps are even slightly depressed, because one weft is more taut than the other, still another variable is introduced. Finally, as weave balances change there are still more variations—changes in the relative sizes of warps and wefts, differences in their spacing, and also differences in their elasticity, stiffness, suppleness, etc. Stir all of these variables together, and differences in weaves result that we identify as distinctive.

Best wishes,
Marla