Untying the knots of Central Asian carpets

Photo Essay by Victoria Sluka: searching for clues to ancient production techniques among modern weavers

Because they tend to disintegrate entirely in archaeological contexts, textiles are among the rarest media in archaeological analyses. While ceramics, metals, stone and glassware can often be counted, measured, and analysed as large temporal and cultural assemblages, textiles are usually solitary finds without any significant comparative material. This drastically limits the types of conclusions that can be made about wider social customs or cultural traditions, since archaeologists can never be sure if that single data point is representative of the missing wider assemblage, or if the one available textile is unusual or even entirely unique for the period. If the Mona Lisa was the only surviving painting from the Italian Renaissance, how would we know that she was special, even in her own time?

For my doctoral research, I am attempting to address the limitations of the small, temporally and culturally erratic assemblage of a particular type of Central Asian material: pile textiles. Although most commonly associated with carpets, pile textiles are also used for saddlebags, wall hangings, and even clothing items. Carpets popularly known as Oriental or Persian in homes today are pile textiles, as are velvets, corduroys, and most Western towels. Pile refers to the added third dimension in these textiles: they are not just wide and long, but also vertically dimensional and plush. Unlike velvet, corduroy, and towels however, Central Asian pile textiles are generally knotted. In these pieces, the vertical dimension is added by tying knots on the textile substrate: the plush depth of the textile surface is actually the many thousands of loose fibre ends sticking out of the knots, trimmed to a uniform length. By tying the knots with different coloured threads, the weaver generates the patterns and designs of the piece. Although the earliest evidence of knotted pile textiles comes from southern Central Asia in the 2nd millennium BCE (Khlopin 1982), the technique has since been independently invented by and spread through contact to cultures around the world.

Although tools and other paraphernalia associated with textile weaving regularly turn up in the archaeology of Central Asia, complete textiles are exceedingly rare. Even fragments are rare, with only a handful available from the entire region over thousands of years. Therefore, it is currently impossible to execute a robust comparative study of these textiles. I intend to artificially generate a comparative assemblage against which archaeological materials can be compared. To do this, I am collecting carpets and other pile textiles made by traditional handweavers throughout Central Asia. Interviews conducted with each weaver will record the kinds of tools and looms used, structural differences of the textile, idiosyncratic variables like handedness and years of experience, and the origin of the design. From these modern examples, I hope to draw some robust conclusions about how the production methods affect the finished textile. For example, maybe left-handed weavers always tie knots that angle to the right, or those who use a particular tool create more tightly-woven textiles than those who use another tool. If similar patterns are present in the archaeological samples, it would be reasonable to extrapolate a similar set of production circumstances. This tells us not only about the individual weaver, but also about the uniformity of larger technological systems, adherence to standardised guidelines vs. domestic variability, and the extent of trade networks both for raw materials as well as finished textiles.

Because I am primarily interested in technological systems, innovation, and the evolution thereof, my fieldwork tends to focus on documenting the processes involved in pile textile production.

Although we often associate knotted pile textiles with carpets, the structure was (and is!) used for a wide variety of nomadic material culture. Examples include saddlebags (top), decorative yurt bands/hangings (lower left), and cup carriers (lower right). Display at the National Museum of the Republic of Kazakhstan (Nur-Sultan). 2018.

The first step in knotted textile production is acquiring the fibre: wool is a common choice, although cotton and silk are also used in Central Asia. Here, a Turkmen shears a sheep. Ashgabat, Turkmenistan. 2017.

After the wool has been shorn, it must be cleaned to remove dirt, debris, and oils. Although the washing process results in clean wool, it also makes the wool tangled and clumpy. Ashgabat, Turkmenistan. 2017.

Once the wool is clean, it is beaten with long, thin sticks to break up clumps and tangles. The beating makes the wool fluffy and easier to separate with the fingers. Ashgabat, Turkmenistan. 2017.

Next the wool is carded (ladies at the front). Carding involves pulling the wool through a fine-toothed comb. This forces the individual hairs to all lay in the same direction: the same as combing your hair. Notice the lady in the back (orange scarf, blue dress) beating the wool. Ashgabat, Turkmenistan. 2017.

Once the fibres are all lying on the same axis, they can be spun into yarn. Here a Turkmen woman uses a drop spindle. She spins the spindle by hand and feeds the raw wool in. If she feeds too much or too little wool, or if she varies the speed at which she spins the spindle, the yarn will be lumpy and uneven. Ashgabat, Turkmenistan. 2017.

Although the spinning process naturally produces a ball of yarn, it is too dense to be dyed effectively. Instead, the yarn is re-wound into large, loose skeins which allow for more even penetration of dyes. The weaver spins the winder, which unravels the ball and rewinds a skein with uniform tension and prevents tangles. Ashgabat, Turkmenistan. 2017.

The natural skeins are now dyed, traditionally by dipping. Most traditional dyes, both in Central Asia and globally, are made from plants (including madder root, pomegranates, and walnut). Once dry, the wool is ready for weaving. Ashgabat, Turkmenistan. 2017.

The first step in preparing a loom for weaving is adding the warp (vertical structural threads). A low-quality carpet 2 metres wide can have 1,000 or more warps, while higher quality carpets will have several times more. Here a group of weavers work together to add the warp quickly: the woman sitting on top of the loom wraps the yarn around the upper beam and drops the ball of yarn, which her colleagues wrap around the bottom beam and pass back up to her. Samarkand, Uzbekistan. 2019.

Although looms were developed by most cultures worldwide, their precise mechanisms can vary widely. One of the main types of variation is in the shedding system – the way by which the raising and lowering of warps is mechanised. Central Asian carpet weavers use generally similar mechanisms, with a string heddle and a thick shed stick. Because the only weaving pattern used in a pile textile is a plain weave, the shedding system is less complicated than those used for other textile types.

Shymkent, Kazakhstan, 2018;

Samarkand, Uzbekistan, 2019;

Samarkand, Uzbekistan, 2019.

Compare the above shedding mechanisms to these traditional examples from elsewhere in the world:

Dark red yarn woven with a rigid heddle used by the Zuni people of the American Southwest (loom from New Mexico), and;

Indigo yarn woven with a patterned shedding system used by the Thadou-Kuki peoples who inhabit regions in northeast India, Myanmar, and Bangladesh (loom from Manipur). Objects held by the Pitt Rivers Museum, Oxford. 2015.

Once the loom has been warped and the shedding mechanisms added, the weaving can begin. Although it is common for domestic weavers to invent the carpet design as they go along or to draw traditional motifs and designs from memory, professional or commissioned textile designs are usually drawn out in advance. Sometimes called cartoons, these drawings are done on graph paper where each square represents a single knot. The colour of the square indicates the colour of yarn used to tie the knot. Here a Turkmen high school student colours a cartoon. Ashgabat, Turkmenistan. 2017.

The textiles are woven from the bottom-up. Notice the cartoons hanging on both looms.

A professional weaver, the matriarch of a family of artists who own their own gallery, demonstrates her weaving technique in her workshop in Shymkent, Kazakhstan, 2018.

A loom with carpet in progress. Notice the coloured skeins of yarn hanging on the top of the loom to be within easy reach of a seated weaver. Samarkand, Uzbekistan, 2019.

There are many tools unique to pile textiles. The first is the knife: rather than cutting a short length of yarn and then tying it onto the carpet, weavers instead use the yarn still attached to the skein/ball. After the knot is tied, the yarn is cut to sever the knot from the ball. The knives used for this are specially designed to allow for rapid, one-handed cutting. A curved carpet knife is the earliest hard evidence for knotted textile production in Central Asia, uncovered in the Sumbar Valley of Turkmenistan and dating to the 2nd millennium BCE (Khlopin 1982).

Compare the knives: a woman in blue uses a semi-circular knife which allows the weaver to rotate her wrist (as if turning a doorknob) and cut the yarn anywhere along the curved edge. Ashgabat, Turkmenistan, 2017.

Second, the weaver holds a modern, mass-produced knife based off those traditionally used in Afghanistan and South Asia. The cutting blade is shorter and straight, and it includes a hook at the end that allows her to manoeuvre individual threads too small for her fingers. Shymkent, Kazakhstan, 2018.

A close-up of the type of knife used in the middle photograph, made in the 20th century and used in Samarkand, Uzbekistan.

cutting a rug

Ashgabat, Turkeminstan, 2017. Turn up that volume to hear how these artisans keep in the groove!

Another tool needed for pile textiles are shears. Although each individual knot is cut away from the ball as soon as it is woven, the loose ends are not of uniform length. After a whole row of knots is finished, a pair of off-set shears is used to trim the ends to an even length, creating the finished surface.

First, the beginning of a carpet depicting animals, with the top row unshorn – notice how much longer the knot ends are than the finished, trimmed surface. Samarkand, Uzbekistan, 2019.

Close-up of a weaver working left-to-right across a carpet, leaving a line of unshorn knots to her left. Ashgabat, Turkmenistan, 2017.

A pink-centred piece on an unusual horizontal loom, with the top row of knots still unshorn and a pair of shears lying on the surface. Shauldir, Kazakhstan, 2018.

A pair of modern offset shears produced in the 20th century and used in Samarkand, Uzbekistan. Notice that the blades and handle are on different levels, allowing the weaver to hold the shears flat against the carpet surface and create an even finish.

Once the knotting is finished, the textile is cut off the loom and the warp ends are finished by hand. Usually left with loose fringe, the ends can also be braided. Here, women tie knots in the warp to hold the weaving in place but allow the warp ends to remain as fringe. Ashgabat, Turkmenistan, 2017.

A finished carpet of particularly high quality, on display in the Turkmen National Carpet Museum, (Ashgabat). Quality in knotted textiles is measured by the number of knots per square inch. A low-quality carpet may have 10–50 knots/square inch (1.6-7.8 knots/cm2), while a high-quality carpet may have 300 or more per square inch (46.5+ knots/cm2).

One of my research goals is to understand the rate and distribution of design errors in handmade carpets. Standardisation and error studies are used on many archaeological media to understand production circumstances, technological change over time and space, and the relative checks and balances involved in the production process to generate uniform materials.

First, a medallion with two small unrelated errors: the green circle shows a black knot that should be red, and the yellow circle shows a red knot which should be black.

A second medallion has a larger, more obvious group of errors (in the green rectangle): an area of orange and white knots that are inverted suddenly as the weaver worked upwards and realised their mistake.

Both carpets have a traditional Turkmen design (last picture, viewed from the front) and were produced in 2002 in a state-run carpet workshop in Ashgabat. Ashgabat, Turkmenistan, 2017.

As in most parts of the world, weaving and textile work is traditionally associated with women in Central Asia. The immense amount of work required to maintain the clothing, bedding, floor coverings, animal trappings, tenting, upholstery, and other textile items used in a nomadic-pastoral home would have taken up a significant portion of a woman’s day. Unfortunately, the low survival rate of these organic items in the archaeological record means that huge portions of women’s experience, creativity, expertise, and technological ingenuity have been lost. Although they are often archaeologically invisible, I hope that my research will generate novel ways to study their lives through their rare artefacts. It is wonderful to interact with women who work today to maintain traditional technologies and pass their heritage on to their own communities as well as to international interests, like me.

A woman demonstrates her weaving technique for me, after walking me through the set-up and warping of the loom over the previous few days. Ashgabat, Turkmenistan, 2017.

A pair of young women work together on a small carpet – do you see the shadow of another face through the warps? Another pair of young women sit on the other side of the loom, the four working in tandem to create a double-sided carpet. Ashgabat, Turkmenistan, 2017.

A young woman shows me a simple embroidery stitch as she works on a detailed headscarf. Ashgabat, Turkmenistan, 2017.

Me with a weaver, who is the matriarch of a large artisan family who run a private gallery and workshop for textiles, ceramics, leather, and glassworks. I was allowed to weave a row of knots on an in-progress carpet, during which she kindly but firmly corrected my slow technique. Shymkent, Kazakhstan, 2018.

Although traditional textiles are part of the heritage programmes of all five Central Asian states, Turkmenistan puts a particularly dramatic emphasis on them. A purpose-built museum, the Turkmen National Carpet Museum in Ashgabat displays historic and modern carpets from around the region. At the end of May each year, the museum hosts events associated with the National Carpet Holiday, including exhibitions of weaving, embroidery, felting, and other non-textile handcrafts, as well as musical and dance performances and an academic conference. The event is truly a carpet-lover’s dream!

Me, outside the Turkmen National Carpet Museum whose plaza has been covered in traditional red carpets for the event.

The foyer of the Carpet Museum – some of the world’s largest knotted carpets (200 – 300m2) hang on the walls.

A group of women perform a rendition of a traditional song for delegates.

A group of women perform a play relaying the legendary history of carpet weaving in Central Asia – notice a reproduction of the world’s oldest carpet, the Pazyryk carpet (Seager 1964), laying on the ground in the lower right. All Ashgabat, Turkmenistan, 2017.

Opening ceremonies for the Carpet Holiday, Turkmen National Carpet Museum in Ashgabat, 2017. Again, turn up that volume and join in!

Khlopin, I.N. 1982. “The manufacture of pile carpets in bronze age central Asia.” Hali 5(2):116-18.

Seager, G. Wilfrid. 1964. “Oriental Carpets Today.” Journal of the Royal Central Asian Society 51(2):120-126.

Further reading

Alexander, Christopher A. 2010. A Carpet Ride to Khiva: Seven Years on the Silk Road. London: Icon Books Ltd.

Baker, Patricia L. 1997. “Twentieth-Century Myth-Making: Persian Tribal Rugs.” Journal of Design History 10(4):363-374. https://doi.org/10.1093/jdh/10.4.363

Bier, Carol. 1996. “Approaches to understanding Oriental carpets.” Arts of Asia 26(1):66-81. Available at: http://works.bepress.com/carol_bier/49/

Briggs, Amy. 1940. “Timurid Carpets: I. Geometric Carpets.” Ars Islamica 7(1):20-54.

Cammann, Schuyler. 1975. “Review: The Systematic Study of Oriental Rugs: Techniques and Patterns.” Journal of the American Oriental Society 95(2):248-260. DOI: 10.2307/600322

Bohmer, Harald, and Jon Thompson. 1991. “The Pazyryk Carpet: A Technical Discussion.” Notes in the History of Art 10(4):30-36. https://www.journals.uchicago.edu/doi/abs/10.1086/sou.10.4.23203293?mobileUi=0&

Broudy, E. 1979. The Book of Looms: A History of the Handloom from Ancient Times to the Present. Hanover: University of New England Press.

Dalley, Stephanie. 1991. “Ancient Assyrian Textiles and the Origins of Carpet Design.” British Institute of Persian Studies 29:117-135. DOI:10.2307/4299853

Doumani, Paula, N. Dupuy, Robert N. Spengler III, and Michael D. Frachetti. 2017. “Eurasian Textiles: Case Studies in Exchange during the Incipient and later Silk Road Periods.” Quaternary International 468:228-239. http://dx.doi.org/10.1016/j.quaint.2016.09.067

Doumani, Paula N., and Michael D. Frachetti. 2012. “Bronze Age Textile Evidence in Ceramic Impressions: Weaving and Pottery Technology among Mobile Pastoralists of central Eurasia.” Antiquity 86:368-382. DOI: https://doi.org/10.1017/S0003598X00062827

Eiland, Murray L., III. 1993. “The Past Re-Made: The Case of Oriental Carpets.” Antiquity 67:859-863. https://doi.org/10.1017/S0003598X00063857

Fitz-Gibbon, Kate. 2004. “Something Borrowed, Something Red – Textiles in Colonial and Soviet Central Asia.” Textile Society of America Symposium Proceedings 442:124-133. https://digitalcommons.unl.edu/tsaconf/442/

Gasparini, Mariachiara. 2016. “Sino-Iranian Textile Patterns in Trans-Himalayan Areas.” The Silk Road 14:84-96. http://www.silkroadfoundation.org/newsletter/vol14/Gasparini_SR14_2016_84_96.pdf

Geijer, Agnes. 1963. “Some Thoughts on the Problems of Early Oriental Carpets.” Ars Orientalis 5:79-87.

Good, I., J.M. Kenoyer, and R.H. Meadow. 2009. “New Evidence for Early Silk in the Indus Civilization.” Archaeometry 50(3):457-466. https://doi.org/10.1111/j.1475-4754.2008.00454.x

Murphy, Brian. 2005. The Root of the Wild Madder: Chasing the History, Mystery, and Lore of the Persian Carpet. New York: Simon & Schuster.

Seager, G. Wilfrid. 1964. “Oriental Carpets Today.” Journal of the Royal Central Asian Society 51(2):120-126. https://doi.org/10.1080/03068376408731860

Seiler-Baldinger, A. 1994. Textiles: A Classification of Techniques. Bathhurst: Smithsonian Institution Press.

Sheng, Angela. 1998. “Innovations in Textile Techniques on China’s Northwest Frontier, 500-700 AD.” Asia Major 11(2):117-160. https://www.jstor.org/stable/41645543

Spooner, Brian. 1988. “Weavers and Dealers: The Authenticity of an Oriental Carpet.” In The Social Life of Things: Commodities in Cultural Perspective, edited by A. Appadurai, 195-235. Cambridge: Cambridge University Press.

Tehrani, Jamshid, and Mark Collard. 2002. “Investigating Cultural Evolution through Biological Phylogenetic Analyses of Turkmen Textiles.” Journal of Anthropological Archaeology 21:443-463. http://www.ceacb.ucl.ac.uk/ceacb_files/misc/Tehrani_Collard_2002.pdf

Zhao, Feng. 2004. “Evolution of Textiles Along the Silk Road.” In China: Dawn of a Golden Age, 200-750 AD, edited by James C.Y. Watt, 67-77. New York: Metropolitan Museum of Art.

Zhong, H., and M. A. Hann. 1989. “Textile Manufacture in China during the Period of the Warring States (475–221 B.C.).” Journal of The Textile Institute 80(3):403-413. https://doi.org/10.1080/00405008908658295

2 thoughts on “Untying the knots of Central Asian carpets

Leave a Reply