Posted by R. John Howe on 10-18-2003 07:15 AM:

"Twining" is not "weaving." Agreed? Why or why not?

Dear folks -

In the first edition of her book, "Woven Structures," Marla Mallett makes the follwing statement:

"...In the strictest sense, twining is not weaving, and yarns are not wefts. Crosswise twined elements do not interlace; instead two or more spiral around each other as they enclose first one warp yarn and then another."

Marla objects to the use of "weft" in the expression "weftless sumak" on similar grouonds. Here's her sentence about that usage. "...In the strictest sense, it is improper to call a soumak patternig yarn a 'weft,' since it does no interlacing. If the term 'weft' is justified merely because the element is transverse, 'vertical extra-weft wrapping' becomes a particular contradiction in terms."

I have been puzzling about these statements, trying in particular to get clear in the case of weft twining, why the horizontal movement of the weft yarns as they encircle the warps and then twist slightly (note not completely) before moving to encircle the next warps, should not be counted as an instance of "interlacing."

Here is one of Marla's drawings of weft twining.

Now I am trying to look very carefully at this drawing. In plain weave, the wefts simply go over and under on set of alternate warps moving in one direction and then over and under a different set of warps on return. It would seem that it is this over and under movement that counts as "interlacing."

In the twining drawing above, it seems to me that the horizontal yarns do actually move over and under the warps as they (the horizontal yarns) cross them. So that characteristic would still seem to be present in twining. But one more feature is added in the case of twining: the horizontal yarns also move over and under one another as they encircle a given warp or set of them. It would seem that it is this partial twisting about one another serves to disqualify them from Marla's notion of "interlacing." I wonder if this is not too stringent a standard for "interlacing," one that seems to require us not to notice that the horizontal yarns are in fact meeting the main standard for "interlacing" by moving over and under alternate warps. We are asked to focus entirely on the distinctive twisting feature.

It is true that "weft twining" entails the simultaneous use of three yarns and weft "interlacing" employs only two, but that is different from claiming that "twining" does not achieve "interlacing."

Marla, as she was writing the first edition of her book, was patient with me while I debated with her about the proper application of the word "knot," and so, maybe, she will be similarly so here.

I am not as clear as I would like to be about why "weft twining" does not achieve "interlacing," and must therefore be excluded from "weaving" closely defined.

What do you think?


R. John Howe

Posted by Marla Mallett on 10-18-2003 07:02 PM:

Dear John,

This all might seem like just a matter of semantics, when the differences are actually among the most substantive in the textile field. Just ask anyone interested in mechanizing these processes! WEAVING, has, for at least hundreds of years, been a term that describes a construction process—actually a wide group of processes, some simple, some extremely complex. A WOVEN STRUCTURE is a fabric or object made by one or more of those processes. WEFT is a term that for hundreds of years has described one of the elements used in those processes. In each case, INTERLACING is the PROCESS in which a WEFT element is put through a shed, either picked by hand with a pick-up stick, or made mechanically, through a set of WARPS. TWINING, instead, involves twisting two or more elements around individual “warps” or pairs (or more) of warps—a very different and extremely laborious procedure. One need only try the simplest versions of each process to be immediately aware of the radical difference between WEAVING and TWINING. After just a little experience with the two very different procedures, anyone should understand why twining was abandoned for fabric construction in most parts of the world when heddles were invented and sheds could be opened with ease—for the insertion of WEFTS.

The term WEFT is badly corrupted if it is applied to elements in fabrics made by other processes. It has no place in descriptions of fabric structures that are knit, looped, knotted, wrapped, stitched, crocheted, or twisted (twined). The terms are frequently confused when combinations of processes are used to produce a single object—when woven Uzbek covers are enhanced with embroidery, when Caucasian woven horse covers are decorated with images made by wrapping selected groups of warps (soumak), when woven Moroccan rugs are decorated with intermittent bands of twining, or even when pile rugs are constructed by combining rows of weaving with rows of knotting. Yet we tend to casually refer to the finished objects as “weavings” for the sake of convenience, and to the artisans as “weavers.”

There is, for example, no way to logically call the knotting part of making a pile rug “weaving.” Nor can we call the small cut-pile attachments WEFTS. Kinds of pile do exist that ARE made by weaving: INLAID wefts may be interlaced and pulled up to make pile. Or long WEFT FLOATS can be cut to make pile, or EXTRA WARPS can be pulled up and cut to make pile (as with velvet). Pile made by these processes can indeed be identified as WOVEN. Yet for convenience, we also call knotted-pile rugs HANDWOVEN, because weaving makes the basic part of the structure.

I think terminology should rarely be dependent upon the particular kind of equipment used. We may think of knitting as a technique that requires knitting needles, but indeed, knit structures can be produced instead with the fingers. Or knit tubes can be made with a spool that has nails around the hole on one end—Remember those scout camp crafts sessions? Woven structures can be produced either on a loom or not—although having one set of yarns held taut definitely simplifies the process. Likewise, twining can be done on either taut or loose yarns or “warps,” and just because pairs of yarns are twisted across a series of yarns suspended from a heading cord we cannot automatically label the construction a WEAVING. The fact that the set-up looks vaguely similar to a conventional warp-weighted loom is not sufficient. If there is no SHED produced (mechanically or not), there is no WEFT, and no INTERLACING. The process has more in common with net-making than weaving.

As for your statement, John, that I have objected to the term WEFTLESS SOUMAK, that is incorrect. It seems completely logical to call fabrics made with wrapping but with NO ground wefts, WEFTLESS SOUMAK, and I have done so. What I object to is calling the WRAPPING elements in conventional soumak fabrics WEFTS. What you quoted is: "...In the strictest sense, it is improper to call a soumak PATTERNING yarn a 'weft,' since it does no interlacing.”

John, you ask why "weft twining" does not achieve "interlacing," and must therefore be excluded from "weaving" closely defined. From my perspective, INTERLACING is what a weft does during the construction process. How can one legitimately use the word to describe one feature of the structure that is detected afterwards, while not acknowledging the twisting of paired crosswise yarns and their movements in opposite directions as a critical element? Those twists that seem so insignificant to you produce a radically different structure from an interlaced one! Twining is a far more stable and sturdy structure, as well as a painfully slow one to produce. I don’t believe that anyone who ever gave it a try would choose to categorize a twined structure along with an interlaced (i.e. “woven”) construction.

John, you’ve said: “I have been puzzling about these statements, trying in particular to get clear in the case of weft twining, why the horizontal movement of the weft yarns as they encircle the warps and then twist slightly (note not completely) before moving to encircle the next warps, should not be counted as an instance of interlacing."

In fact, the crosswise yarns DO twist completely each time that they ENCIRCLE a “warp,” (half of the twist is before the individual warp, and half is after; if the “warps” were removed, each pair of crosswise yarns would retain the series of twists across their length!

It may be easy to confuse twining with LENO (gauze) weaves that feature twisted WARPS, rather than wefts. In these fabrics, however, the crosswise elements—the WEFTS--are interlaced through a shed, made either by hand with a pick-up stick, or with doups—heddles that shift the warps into new positions and create the twist when the counter shed is opened. In LENO fabrics, the warps twist and then subsequently untwist—unlike in warp twining or conventional twining done with crosswise elements.

Best wishes,

Posted by R. John Howe on 10-18-2003 09:21 PM:

Marla -

Thanks for this very fulsome reply. I do see the distinction that you are making and it is of your usual sort: the perspective of the weaver and the various processes of "weaving" must be honored.

And it is true (I have experienced it a little myself) that DOING weft twining is very different from WEAVING, if "weaving," narrowly construed, is restricted to putting a pick of weft over and under alternative warps, however that is done. (Is it appropriate still to speak of a shed, in the case of true weaving on a weighted warp setup? That is, warp ends not attached at the bottom and no heddle.)

This distinction (between "twining" and "weaving") seems to illustrate again the danger of describing structures on the basis of what one sees after they have been made. Such descriptions are open to error because they often do not carry along and acknowledge the processes required to make them.

I am sorry for my inaccuracy about your statements on "weftless soumak."

Thanks again for taking the time to make this informative post.

I know in a man-on-the-street way what "gauze" is, but have not, I think, previously encountered the term "LENO."


R. John Howe

Posted by Marla Mallett on 10-19-2003 08:34 PM:


Yes, it is entirely appropriate to speak of a shed in connection with any weaving done on a warp-weighted loom. Most warp-weighted looms used over the past 3000 years or so have included both a heddle bar and shed stick and so could mechanically open two sheds. ALL, though, from the very earliest times, have used at least a shed stick and so held open a primary shed constantly. On such looms, (before the invention of heddles) the weaver would have picked the secondary shed with a pick-up stick which was then turned on its side to open a shed. Twining, on the other hand, uses NO shed at all.


Posted by Sue Zimmerman on 10-19-2003 10:50 PM:

Hi John,

Some people think that using a loom for twining is a nuisance. They prefer to work freely with the piece on their lap. Sue