Reconstruction of the tablet-woven textile from Miran

I have admired the tablet-woven band from Fort Miran, ever since I saw the photo of the four-legged animal (lion?) in The Techniques of Tablet Weaving, but it took long time before I mustered courage to weave the reconstruction. If it wasn’t for lockdown, which provided me with plenty of free time, I would be probably still evading it.


The 8th century textile survived in several fragments, which are now kept in various museums (Victoria & Albert Museum, British Museum). Usually, such a fragmentation of an archaeological artefact is a disaster, but this time it helped me, because V&A has a more accurate description on its website, while BM has better photos. 

I charted my first version of the animal pattern based on the photos provided by museums and the one in Collingwood's book – but with all diagonals leading in one direction. The reason why it took me so long to muster courage for this project is that despite my long experience with various tablet-weaving techniques, I still find double-faced 3/1 broken twill confusing. So I charted the twill diagonals in random direction, and only after weaving my first sample I adjusted the pattern so that their direction corresponded to the original band. 

The second sample based on the adjusted pattern

In contrast to the theory of double-faced 3/1 broken twill, now it made perfect sense to me. The change of direction in the twill line was caused by leaving the tablets in the positions, in which they ended up after the colour change, without extra turns aimed at the continuity of the twill pattern in the entire width of the band. Can this be seen as proof that ancient weavers preferred simple techniques to the meticulous neatness favoured by modern people? May be. More likely it indicates the use of two-pack method, in which the directions of twill and colour are changed by flipping the tablets, with no consequences to the turning sequence. (In one-pack method each twill point means a change of the turning rhythm (e.g. two tablets forward - one backward - two forward, instead of continuous two tablets forward - two backward), and therefore it is easier to add those extra turns which allow simpler turning of tablets in following rows.)

Before starting weaving the reconstruction in thin (28/2) wool, as close to the original as possible, one more decision had to be made. Due to my troubles with double-faced twill, I had tried at first to determine the length of the animal motif based on number of turns of the border tablets. I had counted 60. But the animal motif itself had been about 80-90 lines long. Therefore I hesitated whether the museum description was correct. Maybe the whole textile had not been sewn of two but three separate tablet-woven pieces. However, unlike the bottom edge, with the stitches of the brown thread visible even in the low-resolution photograph, in the detail photograph of the upper edge on the BM website, it appears that the weft really passes through. In addition, I have considered the description for V&A based on Lisa Raeder Knudsen's research to be a more credible source of information than if it was created by an anonymous curator.


That's why I decided to weave the rainbow border at the same time as the animal pattern, but not to turn the border tablets each time the pattern tablets were turned. I let the border tablets be idle every fourth row. I have already known from my previous experience that tightly twisted borders make firm beating difficult, even if the threads in the central pattern offer less resistance. Therefore I hoped that by omitting some turns of the border tablets, I would be able to beat more firmly and thus achieve the correct ratio between the warp and weft density. Unfortunately, I was wrong about that. Although the width of my finished band is only 1-2mm wider than the original, the length of one repetition of the pattern is 9cm instead of 6cm.

The first version of the animal band in wool and correct colours

Originally, I planned to weave only the band with the animal pattern, but after having found out that the fabric, to which the band is attached, had been also tablet-woven, I decided to give it a try as well. At least partially - because the original, preserved in a width of 16 cm was woven on more than 180 tablets. (The total width is not known due to the absence of the second selvedge.) As I do not own so many tablets of the same size and the paattern repeats not only vertically but also horizontally, I decided not to weave the full width of the preserved fabric.
At first glance, the geometric pattern seemed easier to draft than the animal pattern. However, when I applied my first assumption that it was a floatwork pattern on the twill background, the lines of the pattern did not meet where they should. Only after weaving a small sample and comparing its reverse side with the photo of the original, I realized that there are also floats hidden in the background and I had to redraw the whole pattern. After solving this obstacle, the weaving itself was quite simple, although - given the width of 60 tablets, mostly rotated individually or in pairs - quite slow. My average weaving speed was about 20 rows per hour. Sometimes I was able to make one row in less 2 minutes, but such a haste often resulted in a mistake, which slowed me down. And there would have been probably much more mistakes and delays if it was not for a striped background. Regularly alternating colors allowed continuous checking whether the tablets are in the right position, and when I forgot what was my last move due to someone disturbing me in the middle of the row, I needed to return only to the beginning of the last colour segment instead of the beginning of the whole row.

The right and wrong sides of the geometrical band

When both parts were ready to be sewn together, I realized a complication in the arcuate shape of the animal band (the result of the border tablets on one side only and/or the uneven weighting of the warp). In addition, I was not satisfied with the proportions of animals, so I decided to weave the animal band once more with lighter weights. I managed to shorten the length of the motif only a little (by 1 cm instead of the desired 3cm), but the small changes I made around the eye of the animal during the second repetition helped me to get straight certain details of the pattern, over which I have been racking my brains. Unfortunately, they also showed me that my pattern is still not an exact copy of the original. I will not weave it for the third time (fourth time, if the test sample is also counted in). There are too many other patterns waitinig to be woven. ðŸ˜€

Finished project


References:

Collingwood, P. (1982): The Techniques of Tablet Weaving, p. 192

Fragment in Victoria & Albert Museum: http://collections.vam.ac.uk/item/O91172/the-stein-collection-fragment-unknown/

Fragment in British Museum: https://www.britishmuseum.org/collection/object/A_MAS-622