Knitless knitting

1 April 201922 November 2021 stringbedLeave a comment

NOTE: This post initially appeared on April 1st and complies with the guidelines for loop-related humor issued by the Coalition for Responsible Loopography.

* * *

The term ‘slip stitch’ has figured prominently in the preceding suite of posts, designating the definitive element of plain crochet. The same label is used for something quite different in knitting, where a slip (or slipped) stitch is a loop that is moved from the holding needle to the working needle without knitting a new loop into it.

I’ve therefore taken care to refer to a ‘crochet-type’ slip stitch whenever the distinction was not clear from context. Thus far, that has only been needed in the discussion of fabric that slip-stitch crocheters would immediately recognize as a product of their craft but which has been described as nalbinding in a few archaeological publications (but noted as highly atypical of that craft by the two authors who have done so, also commenting on its resemblance to crochet).

Another cluster of posts examined the confusion that once attached to the relationship between ‘cross-knit nalbinding’ and ‘closed-loop knitting.’ These are also structurally identical and can only be differentiated if fabric that can have been made by either technique includes further detail specific to only one of them.

Open-loop knitting can also be produced using different tools. Those most widely employed for hand knitting are knitting needles and peg looms, while both home and industrial knitting machines use hooks. The eyed needle of nalbinding is not part of this array since the intermeshing of loops by pulling the free end of the yarn successively through those loops inherently crosses their legs and closes them.

By definition an eyed needle can pull a single strand of yarn along any path it can physically traverse. However, turning a meandering length of yarn into stable fabric requires some form of underpinning where the curve inflects, until the loops are fixed into stitches. Beyond the need for an initial foundation, closed loops can be self supporting but open loops cannot to any practicable degree. Working them requires the additional support of, say, knitting needles.

One might therefore suspect that there is a fundamental flaw is this drawing of what is presented as “needle knitting” in Odhams Encyclopaedia of Knitting from 1957, by James Norbury and Margaret Agutter.

The cited source of that term is an article on “Peruvian ‘Needleknitting’” by Lila M. O’Neale, published in an issue of the American Anthropologist from 1934. Ongoing controversy about the appropriate designation for what at least in craft contexts is now widely called calling nalbinding, was fueled by Daniel S. Davidson in the same journal a few months later with an article titled “Knotless Netting in America and Oceania.”

One of the weaknesses of the term ‘needle knitting’ is that it also designates true knitting done with needles in contrast to work on a peg loom. Another is that true knitting involves the working of one loop into another with the tool(s) positioned between the fabric and the yarn supply. The yarn is worked into the fabric loop by loop in what might be called a ‘loop-led’ technique. In contrast, when using an eyed needle the yarn is interposed between the tool and the fabric and the entire working length of the yarn is pulled through each loop. Techniques doing this can similarly be termed ‘end-led.’

A core problem with ‘knotless netting’ is that leading the end of a piece of yarn through a loop that it has just formed creates a knot by any conventional definition of that term, even if it isn’t pulled tight. There is also a form of netting that is truly knotless and commonly termed knotless netting in industrial contexts. This suggests ‘loose-knot netting’ as a more precise alternative, assuming there is good reason for regarding it as netting to begin with.

It is hardly an appropriate descriptor for the dense fabric that characterizes the Nordic nalbound mittens that provide yet another generic designation for both the technique and the family of stitches produced by it — vantsöm — literally meaning ‘mitten stitch’ in Swedish. That label appears frequently in museum catalog records for socks made in the Nile valley which are commonly, although often questionably, associated with Coptic Egypt.

As a general principle, it is best to avoid categorizing something in terms of what it is not. Unfortunately, labeling end-led looped structures as knotless netting has become too entrenched for it simply to be waved off. However, since all but one of the variant forms sharing that designation can’t be knitted either, the concept of ‘knitless’ would be equally applicable. (The exception is the doubly inappropriately labeled ‘Coptic knitting,’ which procedurally is neither knitting nor historically particularly Coptic.)

This suggests the euphonious albeit ambiguous ‘knitless knitting.’ It might suffice to leave that as a mirthful curiosity. As it happens, however, there is substantive evidence of that very practice in New Kingdom Egypt, three millennia before the first description of what has thus far been assumed to be a purely abstract fabric structure.

It is illustrated with particular clarity by Montse Stanley in The Handknitter’s Handbook, from 1986, as an unknitted precursor to a true knitted structure.

She describes it somewhat circuitously as a “non-interlocked succession of yarn waves,” avoiding the clearer alternative of knitless knitting for unstated reasons that are presumably rooted in an aversion to Davidson’s earlier knotless netting.

Whatever the explanation for her labeling may be, the structure itself also appears in a painting of the goddess Imentet on a mummy case from Luxor, Egypt, dated 1000-970 BCE, on display at the National Museum of Denmark, in Copenhagen.

This detail is from Imentet’s torso and the following one shows her head proximal to archetypal forms of open and closed looping.

The one on the left unequivocally represents the basic element of true knitless knitting. Subject to the considerations discussed above, such fabric cannot be produced with an eyed needle or otherwise end-led. The one on the right is a loop-and-twist structure, which unlike simple closed loops, cannot realistically be worked into fabric with knitting needles. It therefore either provides evidence of peg-loom-based knitless knitting, or what in extension of Harley Davidson’s terminology might be called ‘knitless knotless netting.’

In either case this painting provides the earliest illustrations of both unknitted knittable structures and unbound nalbindable structures yet noted. Their juxtaposition in a single image provides concrete evidence of the contemporaneous practice of the two forms of loopcraft at a far earlier date than has yet been recognized.

* * * *

The photographs of the mummy case shown above were taken through a glass case in a public exhibition and I wasn’t able to find a camera angle that permitted a view of the entire object without a reflection of the display lighting. The museum doesn’t have a digital image available online and here’s the best I could do on site.

Looped Fabric

Methods for knitting metal tubes

31 October 201722 November 2021 stringbed1 Comment

Tubular open-stitch knitting of the previously described type is a common find at Viking sites. This discussion began with it because there is little question about it being stocking stitch in the present-day Western sense, albeit with a compound structure. Comparable specimens with twisted stitches have also been found, as has the cross-knit looping that can often be difficult to distinguish from it.

Correctly differentiating the knitted and looped forms is not eased by their often being conflated in the literature of the modern craft known as “Viking knitting.” Despite calling this technique knitting, it is straightforward cross-knit looping. The underlying issue is that twisted-stitch knitting and cross-knit looping share the same basic structure. An archaeologically-recovered metal tube made in either manner may not reveal the secondary structural detail needed to tell them apart, especially if only a fragment remains. Other indications of the production method might also help but we lack specific knowledge of the techniques used by the Vikings to make any such wirework. It is possible nonetheless, to test the suitability of more recent methods and list conceivable options.

The post linked to above cites a remark by Richard Rutt about the equivalent Irish tubular knitting: “…the Celtic work was done with a knitting nancy. Complex knitting is much easier on a knitting nancy than it is on knitting needles.” (A knitting nancy is a small peg loom, also called a knitting spool.) This judgment would be entirely correct for work with yarn and is consistent with what other authors have suggested about the use of a peg loom for the Egyptian knitted tubes. However, metal has a number of mechanical properties that yarn does not. It is vulnerable to kinks and dents, breaking easily at either, and hardens when worked. All of the methods for open-loop knitting on a peg loom put the stitches under significant tension, as does compound knitting. This can become critical with wire, particularly at small diameters. The method that strains it the least, minimizes the risk of damage.

The stitches on a peg loom are commonly worked with a hook. Although a peg loom can be used for knitting both yarn and wire, a hook alone is adequate for the latter. Wire is rigid enough for a loop formed of it not to require mechanical support pending its being secured to another loop. The process is illustrated in this video where it is named for the tool, but the parent website also uses the more analytical designation “Invisible Spool Knitting” — ISK.

This basic technique will be familiar to any knitter who has repaired dropped stitches by re-knitting them vertically with a crochet hook. It is easy enough to see how it can also produce a twisted-stitch structure simply by rotating each loop 180° after it is pulled through the corresponding loop in the previous row. If one similarly envisions a loop of the one color wire being drawn over the loop of the other color in the next row, and then secured to the loop of its own color in the row after that, the result is compound knitting. (The strip of 8th-century Irish flat knitting shown in the preceding post displays precisely such a two-colored compound structure.)

The drawplate is another important implement specific to wirework. It allows a tube to be knitted at a comfortably large diameter which is then reduced to whatever is desired for the finished tube. This means, at least in principle, that knitting needles can be used to form a looser version of the same structure, which is then drawn down to size — but also elongated in the doing. Ensuring adequately dense stitching in the final tube is one of the effects of compound knitting. It also appears for that purpose in the tutorial literature of modern Viking knitting as double knitting, with triple knitting also described.

Again, though, none of this establishes how the Viking and Irish tubes were actually knitted. Nonetheless, of all the plausible alternatives, using a hook in the manner shown in the video affords the greatest economy of motion and subjects the wire to the least stress. Minimizing the number of contact points that can potentially injure the surface of the wire might similarly explain a preference for an open-loop structure.

Looped Fabric

Looped tubes from Ancient Siberia

16 October 201722 November 2021 stringbedLeave a comment

Sergei Rudenko published a book in 1953, titled Culture of the Altai People in Scythian Times. It includes photographs of the structural detail of two pieces of “woolen lace fabric.”

They were taken from two tubular “braid covers” (shown fully below) with the one detailed on the left being an inner lining to the one on the right. The composite object was used as a sheath for the braided hair that had been removed from the skull of a woman in preparation for her interment in the grave where they were found.

An English translation by Michael W. Thompson, Frozen Tombs of Siberia, was published in 1970. This includes Rudenko’s revisions to the initial text and notes a subsequent radiocarbon dating of the tombs, setting their average age to 430 BCE. Thompson calls the tubular objects the “inner cover of pigtail” and “outer cover of pigtail,” and their structure “lace weave.” It is not clear why he regards this fabric as woven, which it clearly is not.

In her review of the history of nålbinding from 2012, Nalbinden – Was ist denn das?, Ulrike Claßen-Büttner notes that Thompson’s ascription was incorrect and says that the structure of the fabric “was presumably [vermutlich] twisted looping.” The English translation published three years later strengthens this to “obviously made by twisted looping.” The reasons for the increased surety aren’t explained but the photographs certainly do appear to be of loop-and-twist fabric, which is among the earliest archaeologically recovered forms of looping and has been found at a Neolithic site in Denmark.

This appraisal was not shared by Lyudmila L. Barkova in her article titled A technological characterization of woolen textiles from the Great Altai kurgans, in the 2013 volume (nr. 39) of the Archaeological Digest published by the State Hermitage Museum in St. Petersburg, Russia. (There is an English summary at the end of the article and a clickable list of its references here.) It includes full photographs of both objects.

The text describes the structure and method of its production as crochet in the modern sense, but if for no reason other than a single structural detail visible in both her and Rudenko’s photographs, this is not possible. The joins at the corners of several squares in the openwork portion of the fabric are strained, revealing a single strand of two-ply wool (the material noted in the description) stretched across the gap. In crochet, the basic element of such work is a row of chains into which the vertical stitches are worked. It is not possible to pull this in any manner that would expose a single stand of yarn bridging adjacent squares.

The openwork outer sheath was on display at the British Museum from September 2017 to January 2018 in the exhibition Scythians: warriors of ancient Siberia, produced together with the State Hermitage Museum. The text accompanying the object said it was made at some time during late-4th and early-3rd centuries BCE and called it a “crocheted wool hair-case,” further explaining that “…other finds suggest that crocheted hair-cases were common among Scythian women in this region.”

The exhibition catalog describes the object in greater detail, but calls it a “woman’s knitted hair-case” and additionally referred to “crocheted netting” in one of the comparable other finds. The Russian word vyazaniye (вязание) designates all three crafts, with an additional qualifier used to distinguish among them. Barkova’s text, which is cited in the catalog, unequivocally describes the use of a crochet hook to make crochet stitches.

The structure of the object resembles crochet closely enough for it to be seen as such under the viewing conditions in the exhibition. While in front of the display case, I overheard other visitors comment with surprise about crochet having been practiced so long ago. This is unfortunate, given the opportunity to offset — rather than propagate — a misconception that has been sustained by many other examples of loop-and-twist (as well as simple looped) fabric in museum collections being misidentified as crochet. (This piece shares the additional distinction of having been identified elsewhere as sprang, a plaiting technique often enough applied to hairnets — but not to this one, as seen in the same structural details that preclude its having been crocheted.)

The decisive attribute noted above is readily apparent in the catalog photograph.

The details of close-worked loop-and-twist are seen here:

and square mesh here:

The individual stitches in the Scythian piece appear to have one or two fewer twists but the basic structure is identical. The tool used for its production would almost certainly have been an eyed needle, and such work remains a basic element of needlecraft (with the illustrations here taken from a text on needle lace from 1870). Considering both their structural detail and current trends in the ascription of historical craft identity, the braid sheaths would reasonably be categorized as nalbinding. In the doing, however, a spotlight is cast on the need for a nuanced terminology that distinguishes clearly between the various simple and compound looped structures that nalbinding subsumes.

Looped Fabric

Loop-and-twist

28 November 201622 November 2021 stringbedLeave a comment

Irene Emery describes two variations of the simple looping that was discussed in the preceding post. The first is immediately applicable to purse making on a cup mold and is seen in extant looped purses. (The second variation is also relevant to the knitting loom and will be considered separately with that implement.)

VARIATION: Loop-and-twist

The mesh is somewhat elongated and the stitch made firmer when the simple loop is elaborated by the addition of one or more turns of the element about itself. This is usually referred to as loop-and-twist (if there is more than one twist the number can be stated), occasionally as twisted loop or twisted buttonhole stitch, or differentiated as having a ‘full turn.’ In lace-making it has a variety of other names.

Rather than illustrate this with Emery’s photograph, here is the structure as seen in the chapter on needle lace in Beeton’s Book of Needlework from 1870 where it is termed Point d’Espagne (Spanish stitch):

A variant of this structure — Treble Spanish Stitch — is also shown,

with a third variant — Close Spanish Stitch — all appearing in the Dutch purses:

The single Spanish stitch is twisted once and each of the stitches in the triple and close forms are twisted three times, despite the accompanying text stating all are “worked in exactly the same way.” Beeton also illustrates a simple looped stitch but gives it the separate name, Point de Bruxelles (Brussels lace stitch). However, the Brussels stitch appears by that name in other illustrations where it is twisted.

No latitude for modification appears in the instructions for the first of the Louisiana purses from 1823 but those for the second pattern describe adjustments to the stitch count to accommodate thread size, and the third has implicit need to allow for beading. Regardless of any elective twisting (which the Beeton’s illustrations suggest was applied as desired without separate note), the tripling of the stitches in the first pattern produces a filet mesh of the same basic form as is illustrated here. The height of the open spaces relative to their width is determined by the number of twists.

The second and third Louisiana purses are made with the close stitch, which Beeton says should be worked “so closely as to allow only the needle to pass through in the next row.” If the first purse is seen as intrinsically openwork, the second and third are closed work. The last two share the same basic pattern but the third also includes beads. Twisting is an obvious means for balancing the loops adjacent to a bead with the shape of that bead, where they are not closely matched. If the loop is already larger, multiple beads can be stacked but twisting may still be beneficial. It is a nimbler expedient in any case, than is restricting the selection of beads.

Finally, although the stitch patterns illustrated above are all directly relevant to the purses, they show needle lace being worked back and forth in rows, with the direction of the twist alternating in each. The purses are worked in the round and all twisting is in the same direction.