By Katherine Larson

While experimenting with methods for double weave on the warp-weighted loom, I began to think about double-notched heddle-rod supports, wondering if they might provide any insight. These implements, discovered in the 1970s by Norwegian archaeologists excavating medieval layers in Trondheim, have been interpreted by today’s researchers as being attachments for the warp-weighted loom, possibly used for weaving twill.  I reasoned that anyone who knew anything about the warp-weighted loom (as the weavers of 17th century double-weave coverlets obviously did), would likely have known how to weave twill, a durable and dense fabric commonly used for clothing in northern Europe.  Perhaps the role these implements played in weaving such a standard cloth had carried over into double weave? 

If there was any connection to be found between double-weave and double-notched supports, I never discovered it. Instead, my experiments with these implements led in a completely different direction. 

To provide a little background, those who study the warp-weighted loom owe a debt of gratitude to Norwegian researcher Marta Hoffmann, who in the 1950s suspected there might still be women in her country who knew how to use this ancient and seemingly extinct loom. To her credit she searched for these weavers and found two traditions, both of which used the loom to produce plain-weave coverlets: a single family in western Norway, and a small community of weavers in northern Norway. Interest in their coverlets had almost disappeared after World War II, but these women still knew how to warp and weave on their looms, and by documenting their methods Hoffmann was able to record this last link in a chain of weaving skills that had passed from mother to daughter for thousands of years.

Missing from Hoffmann’s observations was the method by which twill was once woven on the warp-weighted loom. Twill had developed into an essential fabric in northern Europe since its appearance in the late Bronze Age, but the weaving of twill had been transferred from the warp-weighted loom to the more efficient floor loom long before the 20th century in Norway. Fortunately for textile historians, several sources of information survived in Iceland, in 18th century drawings that portrayed the loom set up to weave twill, and in descriptions from two elderly Icelandic women who remembered weaving twill on this loom in their youth.  Hoffmann described what was known of this Icelandic tradition, including intriguing details such as the slanted rods (meiðmar) used to hold forward selected heddle rods, and the long warps the Icelandic weavers wove as lengths of cloth for trade.  But the basic premise of the Icelandic technique was the practice of creating sheds for 2/2 twill by using pairs of neighboring warp threads, first to form the forward and back layers of the natural shed, and following this principle to form the three remaining heddle-rod controlled sheds with heddles that enclosed pairs of neighboring warps.

The Icelandic information was first published in 1914, but during the 20th century other ideas were suggested for how twill might have been woven on the warp-weighted loom. These methods fell into two main camps. In the method first proposed by German textile historian Karl Schlabow, the Icelandic tradition of treating the warps as pairs was dispensed with, replaced by heddles enclosing single warp threads, and the warp was no longer divided in half by a shed rod but hung in four individually weighted layers, manipulated by four heddle rods. This loom setup was followed by a number of subsequent researchers. Meanwhile others studying twill followed the basic premise of the Icelandic method, setting up the loom with two warps per heddle, but choosing to utilize the heddle rod supports known from the Norwegian tradition rather than the slanted rods used in Iceland. This method also had a number of adherents. 

Drawing elements from both of these school of thought was yet another idea proposed by A. E. Haynes in an article published in 1975. Following the description given by Hoffmann, Haynes had noted that warp tension was uneven when weaving with the Icelandic method, resulting in unclear sheds. This problem was ascribed to pairs of warps being tied to the same weight when half of the pair enclosed by a single heddle was drawn forward. To solve this problem Haynes, like Schlabow, suggested four separately weighted warp layers.  Unlike Schlabow, however, the loom’s fixed shed rod was retained but in a reduced role: only the foremost of the four layers passed over the shed rod, and from this position those warps could be joined by the warps of two other layers, pulled forward consecutively to form two of twill’s four sheds. Since these sheds were to form behind the foremost layer they required another innovation, a heddle rod support with a notch at the halfway point: a double-notched support. (The remaining two sheds were formed in front of the foremost layer in the normal manner, by pulling the final two heddle rods to the outermost notches.) At the time when this proposal was made, double-notched heddle-rod supports had not yet been discovered in Trondheim, and therefore this was quite an imaginative proposal.

If this system sounds a bit confusing, it became slightly more so after the discovery of actual double-notched supports. At that point Danish weavers at the Historical-Archaeological Research Center in Lejre gave the Haynes method further consideration, adding several refinements to improve the system. Retaining Haynes’s novel idea of forming sheds in different ways (using variously one or two heddle rods and forming sheds either behind or in front of the foremost layer), they realized that the heddles on each rod needed to be of graduated length, since the three back warp layers each hung at different distances from the front of the loom. They further revised the order of heddle rod placement, with heddle rod II on the top supports and I and III on the lower supports. Despite the somewhat involved nature of this revised system, the researchers concurred with Haynes, reporting that many different weave structures were now possible with this improved method: “tabby, basket-weave, 2/2 and 3/1 twill and different floating lacings and pattern effects.” This method is usually referred to as the “four-weight-row” method, and has been adopted by numerous textile historians.

After studying all the ins and outs of the four-weight-row method for using double-notched supports, I decided to begin my twill experiments with the more straightforward description provided in the recent book, The Warp-Weighted Loom, Clinking Stones (2016). I hoped to get a feel for how twill worked on the warp-weighted loom before trying the double-notched supports method. However it became immediately apparent when weaving with the two-warps-per-heddle method from Clinking Stones that there was a problem with heddle jamming. This was especially noticeable whenever the middle heddle rod was drawn forward, an observation that brought to mind a Faroese saying described by Hoffmann, but first noted by Margretha Hald in 1935, that problems with the middle shaft were comparable to disagreements between neighbors – obviously others before me had noted this difficulty! 

Although heddle jams can be cleared fairly easily by tugging on the affected heddle rods, this requires extra effort that not only interferes with the smooth flow of weaving, but would likely wear on the affected warps and heddles if continually repeated. Since the jamming appeared to be caused by slack heddles when manipulating the middle heddle rod, it occurred to me that here was a problem begging to be fixed by double-notched supports, which I just happened to have waiting in the wings.  With this idea in mind, I began to wonder if the two-warps-per-heddle idea, the only twill-weaving method that is actually attested in historical records, might warrant further consideration.

After promising results from some initial experiments, I proposed the idea of using double-notched supports for weaving twill to Monika Ravnanger and Marta Kløve Juuhl.  Monika is a curator at Osterøy Museum near Bergen, an institution central to the study of the warp-weighted loom, and Marta, one of the authors of Clinking Stones, was a proponent of the two-threads-per-heddle method described in her book.  They were both intrigued by this idea and agreed on a program to test its effectiveness. We were especially interested to find how the use of single notched supports compared to using those with double-notches, and also to consider how this new method for using double-notched supports compared to the four-weight-row method.

Our tests, described in detail in our article, provided positive support for our idea. As a result we believe our proposed use of double-notched supports to be simpler and more intuitive than that proposed in the four-weight-row method: Our suggested method employs only one heddle rod at a time to open all but the natural shed; the heddle-rod sheds are always formed in front of the foremost layer; heddles are all tied at the same length; and heddle rods are ordered and utilized in a logical fashion from top to bottom.

We further note that in proposing what became known as the four-weight-row method, Haynes was attempting to solve a problem that did not actually exist.  As it turned out, several factors were overlooked in Hoffmann’s description of the Icelandic method, and a more careful reading of the original Icelandic reports some years later revealed details that solved the tension problem at the heart of the Haynes argument. Seen in this light, the Haynes method could be viewed as an inventive rethinking of how a woven structure can be achieved. Yet perhaps not enough consideration was given to the fact that this innovative method strayed from what may have been a foundational element for weaving twill on the warp-weighted loom: using neighboring pairs of warp threads as the basis for both dividing the warp over the shed rod, and pulling forth neighboring pairs of warp threads to form subsequent sheds. 

While no one can say for sure how twill was woven on the warp-weighted loom, if double-notched heddle-rod supports were used, a method similar to that proposed by our research would seem to be the most likely. For more details and complete citations, the full article inArchaeological Textiles Review No. 67 can be read here: “Twill on the warp-weighted loom: reconsidering double-notched supports” by Katherine Larson, Marta Kløve Juuhl and Monika Ravnanger.

For a translation of the original Danish report describing refinements that became known as the four-weight-row method, read: “The Warp-Weighted Loom, New Experimental Observations” by Anne Batzer and Lis Dokkedal, translation by Katherine Larson.

For a synopsis of Larson’s double-weave research, read: “Go Big or Go Home – The Importance of Textile Width”; or read the concluding article (p. 92+) in Archaeological Textiles Review No. 64: “Norwegian double-cloth: warp-weighted loom experiments in a complicated technique” by Katherine Larson and Marta Kløve Juuhl.

Katherine Larson is an Affiliate Assistant Professor of Scandinavian Studies at the University of Washington.