MARKET EFFECTS ON THE DESIGN AND CONSTRUCTION OF CARPETS IN THE MILAS REGION OF SOUTHWESTERN TURKEY, 1963-1993


CHARLOTTE A. JIROUSEK

Department of Textiles and Apparel, 208 MVR, Cornell University, Ithaca NY 14853

 

Beginning in 1964 my husband and I were involved in the development of a village carpet weaving cooperative in Southwestern Turkey. We lived with the weavers of Çömlekçi from 1964 to 1966, as part of the first Peace Corps rural community development program in Turkey. Between 1966 and 1969 we continued to work with the cooperative in its efforts to develop markets and quality control standards while working as Peace Corps staff.  Since leaving Turkey in 1969, I have visited Çömlekçi periodically,  most recently  in 1992 and 1994.

 

 The success of the  Çömlekçi cooperative in turn generated carpet cooperatives throughout the province of Mula. This paper describes my observations of changes in the design and construction of  Çömlekçi carpets between 1964 and 1994.

 

ÇÖMLEKÇI CARPETS IN 1964

Carpets produced throughout the region surrounding  Çömlekçi are known as Milas carpets, after the major market town forty miles north of Çömlekçi. In 1964, these carpets were woven primarily for dowries, and sold only during periods of financial hardship. In 1964  a two square meter carpet might sell for $30.  Dealers believed that only the old village rugs had commercial value, and when they did purchase new rugs, they were doctored and passed off as older rugs. European and American export markets at the time favored Persian carpets.

 

As a result, villagers looked upon their carpets as old fashioned and inferior. For those who could afford them, dowries might  include a store bought carpet or two, usually from Isparta. Devaluation of their own carpets also led to the appropriation of new designs such as the so-called "Isparta" design, with its central floral medallion, reminiscent of Persian designs. But whatever the design, carpets were considered an essential part of any dowry, no matter how poor the family might be. Since the local economy was still close to subsistence, the habit of producing the necessities of life from one's own resources had continued. This  motivated village families to continue their weaving.

 

Intensely saturated hues derived from chemical dyes were preferred, such as bright greens, blues, yellow  and pink, arranged to maximize contrast of hue and value. This was in contrast to the classic Milas palette of warm reds, browns, yellows, and perhaps olive green. The chemical dyes are often referred to in the literature as aniline dyes; however, it is probable that the dyes were of various types.  Since these dyes, sold in bulk, came without instructions, they were not used correctly, and therefore tended to bleed. Villagers used the same methods to set the chemical dyes that they were accustomed to using for  natural dyes; that is, the addition of a mordant, usually alum (Ġap), and possibly citric acid in small amounts (limon tozu).

 

The natural dyes were thought of as old fashioned, laborious, and undesirable, although every family had a substantial repertoire of natural dye recipes they continued to use. Certain popular colors such as bright pink, purple, chartreuse, royal blue and wine red were difficult or impossible to achieve with the known natural dye repertoire. Their use was a matter of conspicuous consumption, a mark of luxury. Also, natural dyeing was very time consuming. Because of the seasonal availability of some dye materials, it could take the whole year to dye the full repertoire of colors that might be wanted. Simply going to the market for the colors you had in mind was far easier, if you could afford it. On the other hand, since certain chemical colors such as warm reds and yellows were particularly unstable, more reliable natural dye equivalents were often used.

 

In spite of the popularity of commercial dyes, fifty‑nine different natural dye recipes derived from twenty‑one different plants were identified in Çömlekçi in 1964. The data on the dye recipes was  published in 1982 (Miller).

 

Weaving was done from handspun yarns, utilizing wool from their own sheep. The spinning of the wool for a carpet took from three to six months; weaving could take another three months if the weaver worked alone. In reality, a family would be hard pressed to produce more than one rug a year, since this work was done in between the demands of the agricultural cycles. At the time, tobacco was the major cash crop, and it demanded constant attention from May to October.

 

Thus by 1964 Çömlekçi's weaving traditions were valued neither by their practitioners nor by the marketplace. The carpets were being adulterated by non-traditional designs and colors.Even the weavers themselves preferred the rugs of others. Under these circumstances, it was generally assumed that carpet weaving in Çömlekçi was dying out.

 

THE ÇÖMLEKÇI CARPET COOPERATIVE AND ITS EFFECT ON CARPET DESIGN AND CONSTRUCTION

The successful development of the village's carpet marketing cooperative between 1964 and 1971 significantly altered carpet design and construction. It also altered the attitude of the weavers toward their product. The cooperative was formed in May 1966. Until we left in 1969 the major effort was to establish a quality control system. Factors such as density  of the weave; evenness of surface, selvedges, and plain weave foundation; detail and accuracy of the design and color choices were considered. Because of the unreliability of available chemical dyes, natural dyes were required for Cooperative rugs. 

 

The rugs were at first sold directly to resident foreigners, since dealers were unwilling to believe that the rugs were 100% naturally dyed carpets as we claimed them to be. However, this gradually changed as dealers saw more of the rugs, and began to visit Çömlekçi.

 

After we left Turkey in 1969, the villagers continued to look for markets on their own. In 1970 they obtained a contract from Sumerbank, the national textile monopoly, for what amounted to all the rugs they could produce. The cooperative produced for the Sumerbank contract for about three years, and expanded to include weavers from the surrounding region. The income generated from carpet sales in this period permitted the  Çömlekçi to bring in electricity, and to make the transition from the laborious and unreliable tobacco culture to olives. This was significant evidence of the financial impact of the carpet cooperative,  in that it meant taking fields out of production for the five years it took for the new trees to begin to produce.

 

However, by 1973-74 the dealers had discovered the carpets of the Milas region, and were coming in with better offers. The cooperative, bound to its less profitable Sumerbank contract, essentially collapsed by 1975, though it continued to exist on paper (and does to this day).  When I visited in 1976, I was told that the marketing of rugs in  Çömlekçi had virtually come to a halt.

 

MARKET DEVELOPMENTS IN THE 1980S AND AFTER

The early  Sumerbank efforts to promote carpet production and export were strictly economic in motivation. However, by the early 1980's other Turkish government agencies had begun to realize the economic potential and cultural significance of carpet weaving. Of particular significance was a project undertaken in 1986 by the Ministry of Culture.This ambitious project was intended to document all examples of Turkish rugs found in weaving villages, private collections, and museums world wide, in order to revitalize Turkish village weaving. Çömlekçi  was a significant contributor to this project. The Turkish Ministry of Culture has cited the  Çömlekçi Cooperative as a significant contributor to the effort to revive naturally dyed carpets throughout Turkey. (Turkish Handwoven Carpets, 1987)

 

Carpets were photographed, and scanned in to a computer where the design was converted into a point paper weaving diagram in color. The end product was a series of catalogues originally intended to be an order book for carpet commissions to be woven to strict standards specific to each region. The production aspect of this project has not been realized, beyond ministry-commissioned samples. In spite of this, the project continues. The Ministry of Culture archives now include over 6,000 images of Turkish carpets.

 

The first volume of this catalogue was printed in 1987, followed by three additional volumes published in 1988 and 1990 (Turkish Handwoven Carpets). Copies of the catalogues were distributed to village weaving cooperatives throughout the country. Weaving  diagrams appropriate to particular regions have been distributed free of charge to village weaving cooperatives.  Additional volumes are in press and planned. This ambitious project was  initiated by Güran Erbek,  who unfortunately died in 1989, shortly after the publication of the first two volumes. (Kocakaya, 1994)

 

In  Çömlekçi, meanwhile, the cooperative made a new effort to revive itself. In the 1980's, as the Turkish middle class grew, domestic as well as foreign tourism invaded the region in earnest. Probably inspired by the contacts with the Ministry of Culture carpet project, the  Çömlekçi Carpet Cooperative decided to make a try at an idea we had proposed years before. We had suggested that tourists would enjoy a visit to a real village, and such a visit would also be an opportunity to sell them rugs. The Cooperative contacted tourism agencies in Bodrum in 1986, and the first tours were established.  However, this arrangement quickly fell apart over disagreements that followed the traditional clan lines, and since 1986 two individual families have hosted tours to the village, one from each of the major clans. The rugs they sell include Çömlekçi  rugs, but also stock from neighboring villages, as well as  rugs in designs from Kars, Kula, and Gordes. One of the two households also stocks oversized rugs in traditional Milas designs, but woven in the commercial weaving center of Isparta, where copies of rugs from many regions are being made. On any given summer day, there may be as many as six tour buses coming to  Çömlekçi for the tour. Villagers not directly involved with these two tour opportunities do still weave rugs for sale, and also for their daughter's dowries. They may sell their rugs to one of the two tour families, or to dealers and directly to tourists in Milas or Bodrum at the weekly market.

 

THE EFFECT OF THE ÇÖMLEKÇI CARPET COOPERATIVE ON CARPET DESIGN: EARLY DEVELOPMENTS

Since by 1969 carpets that were naturally dyed in traditional Milas patterns had  higher market value, these became the patterns and colors of choice in Çömlekçi for most people. Few of the Isparta style medallion carpets were woven  even for dowries, as these designs were not accepted by the Cooperative. Initially, when  the rugs were being marketed to foreign residents, the classic Milas prayer rug patterns were in demand, and frequently woven (Figure 1). However, after the Sumerbank  contract was signed in 1970, the rugs were exported, mostly to Germany, where symmetrical designs were in demand. In 1976, no new prayer rugs could be found in  Çömlekçi at all. By far the most common design seemed to be one known as Ada Milas, a classic Çömlekçi pattern. (Figure 2)

 

RECENT DEVELOPMENTS: DESIGNS

After the 1970s, there was an increase in the repertoire of designs in use. Old rugs found in Çömlekçi and neighboring villages have been copied to add to the variety of designs. This occurred as early as 1971, when a pattern was transcribed from a carpet book  by  a teenaged weaver. The Carpet Catalogue produced by the Ministry of Culture in the 1980's has provided a further source of old Milas designs, some of which have been reproduced in Çömlekçi. The asymmetrical prayer rug designs once out of vogue are now in again. Thus the repertoire of Milas patterns is increasing.

 

Today, classic Milas designs are almost universal in Çömlekçi for both dowry rugs and rugs woven for sale. Milas rugs have achieved a certain reputation, and there is a market for the local designs.

 

COLOR AND DYES: EARLY DEVELOPMENTS

By the end of the 1960s in Çömlekçi and many other weaving villages where the cooperatives were established,  natural dyes were now highly valued for both dowry rugs and rugs made for sale. However, a color palette emerged that differed significantly from the historical Milas palette. In addition to customer taste, the supply, cost and labor of certain colors were factors. The most troublesome colors  were the madder reds, due to scarcity of madder near Çömlekçi. In  the Milas region madder (kök boya) is a wild plant, native to rocky mountainsides and forested areas. To reach such areas involved a hike uphill from the village, a direction normal activities did not take most people. Since the roots take more than one season to mature, finding enough madder root for the increasing demand could be difficult. By the early 1970's colors were shifting away from bright clear reds, or deep brick reds, and toward red-browns, soft rose (a weak madder recipe) and other colors that do not require madder at all. Rugs began to look browner and darker.

 

One villager had made a trip to Balikesir in order to learn how to use indigo, not one of the village's traditional colors. The results were impressive, if not typical of Milas rugs. However, indigo was not widely adopted because it was hard to get, and the fermentation vat dyeing procedure was complex and unfamiliar to Çömlekçi weavers.

 

RECENT CHANGES IN  DYES AND COLORS

Most rugs made for sale today are made with commercially dyed yarns. However, because of the emphasis on naturally dyed yarns during the last three decades, the colors in a chemically dyed rug may be indistinguishable from those in a naturally dyed rug. 

 

Many dowry rugs are  woven with commercially dyed yarns, but natural dyes may be still be used. The dowry is a matter of  family prestige and honor which accompanies the daughter to her new family, and cements a new family alliance. Today the naturally dyed rug, woven in traditional Milas designs, is seen as the best, even though it is no longer deemed feasible to produce naturally dyed rugs for the marketplace. This is in marked contrast to the attitudes toward naturally dyed rugs and traditional designs seen in 1964.

 

Some Milas style carpets include fashion colors such as pink and light blue that are not part of  regional color traditions. Carpets that differ radically from the traditional regional palette are more likely to be the product of commercial weaving operations.

 

WASHING RUGS AND THE CURRENT COLOR PALETTE

The washing of carpets is also affecting color choices. In the early days of the cooperative,  Çömlekçi carpets were usually sold unwashed. When dealers  washed rugs, it was usually done to "age" them. To many dealers,  the initial advantage of Çömlekçi's naturally dyed rugs was that, once washed, they had the look of much older rugs, and could be sold to the unwary as such. Although harsh washing drastically changed some of the sensitive natural dyes, at that time, washed chemically  home-dyed rugs tended to bleach out even more,  to beige, grey, and tan.

 

These days virtually all rugs sold are commercially washed, ostensibly to insure color stability in the finished product, though the methods reflect some older deceptive practices. The procedure involves a variety of steps which may be combined or deleted depending on the preferences of the dealer. I observed a batch of rugs being washed in Milas for one of the Çömlekçi dealers. The procedure included singeing the backs, shaving the pile mechanically, and washing with chlorine bleach. Lye, and/or sodium hydrosulfite, or even sulfuric acid may also be used. Rinsing was not as thorough as it might have been. The black knots are hand-clipped shorter than the rest of the pile, clearly in imitation of the way in which the black wool is worn away in antique rugs. However, everyone insisted that these things are done for aesthetic reasons, and I did not observe any dealer claiming such washed rugs were old.  This severe washing and finishing procedure appears to have developed back when dyes were unreliable, and dealers were trying to antique rugs in order to fool their customers. The procedures continue even though dyes are now stable, and rugs are not usually passed off as antique any more- and in spite of the harm these procedures do to the rugs.

 

The result of the procedure I observed is a much softer, somewhat lighter colored rug in which the yellows predominate, as a result of the application of chlorine bleach. Needless to say, this process must weaken the carpet. The Çömlekçi dealer stated he did this because his customers preferred the lighter colors. As a result, rugs currently being woven for sale tend to be lighter in value even as they come from the loom, though still in the classic Milas   hue palette of reds, browns, yellows, and green.  Since the washing lightens the rugs anyway, whether the customer wants it or not, lighter colors are what is available. Meanwhile, villagers see the washed rugs being sold, and then try to weave rugs that duplicate these lighter tones. Their rugs are then washed, and the result is presumably even yet lighter values in the rugs. One must wonder where this will lead. Interestingly enough, rugs woven for dowries are never washed commercially, and everyone seems confident that the  rugs will not bleed when they are eventually washed at home.

 

CHANGES IN CONSTRUCTION: YARNS

Concern with the cost of labor and materials also led to the introduction of commercially spun yarns.and eventually in the 1980's to the widespread use of commercially dyed yarns. The better factory spun yarn such as that sold by Sumerbank was considered to be of good quality, although it differed in several ways from the handspun yarn.

 

The commercial yarns all tended to be somewhat larger in diameter than most handspun yarn. Warps made from hand-plied commercial yarn singles tended to be slightly over 2mm in diameter, whereas the handspun warp yarns were usually finer, with one example measured at 1mm in diameter. The commercial yarn rugs were all between twenty‑six and twenty‑eight knots/square dm, whereas the fully handspun rugs ran as high as thirty knots/ square dm, with one example found at forty‑four knots/ square dm. An advantage of the commercial yarn was its consistent diameter and color.

Today the enormous saving in time, and the availability of acceptable commercially spun yarn, has resulted in the complete abandonment of handspinning. Younger girls may not even know how to spin. Warp is still hand- plied, however. A soft two-ply warp that resembles knitting worsted is available but was not seen in any rug woven in Çömlekçi. It is used by commercial production centers, however, and is evidence of probable manufactory provenance.

 

CHANGES IN CONSTRUCTION: LOOMS

By the 1970s the financial incentives to produce more rugs led to some experiments in loom design. A village carpenter produced a free standing carpet loom with turnbuckle-controlled warp tensioning that was light enough to transport on a donkey with warp in place, for use during the summer migration to the valley to be near fields and livestock. Previously weaving had been largely abandoned during the agricultural season from May to October. A similar loom, but with string heddles (not typical of carpet looms) is used by commercial weavers trained  in Isparta.

 

During the 1980s Milas carpets in general underwent some significant changes because of the proliferation of commercial weaving centers. A few commercially manufactured metal rug looms have been brought into the village by one of the two rug dealer/tour households. In commercial weaving centers such looms are common, and their use results in some differences in finishing. Most  Çömlekçi weavers still use traditional warping and weaving methods.

 

CHANGES IN CONSTRUCTION: FINISHING

Commercially woven carpets are usually produced on long warps, with several carpets being produced on the same warp. The space between each carpet becomes fringe which may or may not be braided. This results in a carpet with fringe at both ends.

 

Village carpets, on the other hand, are warped one at a time. The warp is wound as a continuous loop between two stakes pounded into the ground at the desired interval. There is no cross, but instead the warp threads are chained together at each end as the warp is wound.  The completed warp is attached to the warp and cloth beams of the loom by the insertion of a dowel through the chained loop ends. This dowel fits into a slot in the beam, and is held in place by metal pins. The weaving begins directly on top of the dowel, which results in short loops being left at the starting end of the rug. Sometimes these loops are left unfinished, possibly with the chaining thread still in place. This is very likely in rugs woven for sale. If the chain is removed, a short, twisted fringe results in which the loops can be discerned. The proper traditional method used to finish this end, however, is to twine the loops around one another in clusters of three or two. This results in a neat flat  edge with the look of a braid, and much better warp end protection. (Figure 3) This is almost always done on rugs made for personal use, especially dowry rugs. It is very occasionally done on rugs made for sale.  In any case, a traditional village woven rug can therefore only have a fringe at one end, the end at which the warp is cut from the loom after weaving is completed. 

 

The traditional finish at the fringe end of the  carpet is also different from commercial carpets, and impossible to duplicate on the multi-carpet production warp.  As the single carpet is cut from the loom, the warps are cut in groups of three, and then woven between clusters of three warps picked together, for a distance of six triple warp picks. The picks are then reversed, and the cut ends are brought back through and packed down. Then the next group of three is cut and passed through this same shed, but one triple pick further. The shed is then reversed and the ends returned. The result is again a warp protector edging which resembles a  thick flat braid, but this time with a full fringe appending from the back edge. The last clusters of threads cut are braided to finish off. (Figure 4) Thus even in an unbraided fringe, there will be a single braid at one end of the fringe. The entire fringe may then be braided, though this is becoming less common even on dowry rugs, and is almost never done on rugs woven  for sale. Some rugs being woven for sale in  Çömlekçi are finished off with a final row chained weft following the plain weave weft border and simply cut off.

 

Both traditional end finish techniques can be seen only when a single rug is woven on the warp. Thus rugs with  a single fringe and either loops or a flat twined edging at one end are more likely to be the work of a family loom. They are therefore more likely to be indigenous to the area traditionally associated with the design. If the fringe is braided and the loop end twined, it is even more likely to be a traditional rug woven in the region associated with the design, since commercial producers rarely if ever go to this much trouble for the finishing. If the rug lacks the triple pick warp protector at its fringe end, but has loops at the other end, it was most likely woven for market in a village. On the other hand, rugs with two fringes (braided or not) are very likely to be produced by a hired weaver in a commercial rug weaving operation, and the design is least likely to have a relationship to the place in which it was woven. If the warp is commercially plied, this is a virtual certainty.

 

Sometime in the last fifteen years, the use of a device called the bask2 was abandoned in order to save time. The bask2 is a simple six ply wool cord which is kept wet, and placed in the shed after the insertion of the weft (argeç). It is beaten down on top of the weft and knots, and then removed. Because it is wet, the bask2 dampens the wool of the weft, which allows it to stretch, permitting the weave to pack down more tightly.

 

Mass production of handwoven "village" carpets is now being undertaken throughout Western Turkey. In these manufactories, the designs woven are from a variety of regions. Thus it is no longer possible to determine where a rug was woven by studying its design alone. However, changes in loom design and warping procedures have resulted in some significant technical differences in the construction and finishing of these generic "village" rugs. These finishing methods can provide some clues to the possible provenance of the rugs, though even these features can offer no certainty.

 

CONCLUSION

The marketplace has affected the way Milas carpets are made, the colors that are selected, and the designs used.  This is not a new phenomenon; the marketplace has always affected the design of rugs, as it affects the design of any other object that is bought and sold. The fact that this tradition is evolving is evidence that the tradition is still alive. Even though copies of Milas carpets are being produced in other centers, the core of the tradition is still in place. However, the blurring of regional distinctions that has resulted from this commercial  development  does not bode well for the long term integrity of regional weaving traditions.

 

Every girl raised in Çömlekçi still learns to weave, and the  number of carpets a family tries to provide for her dowry has increased from between three and seven in the 1960s to eight or more in the 1990s. However, weaving carpets is a village activity, and if a bride marries out of the village to a town, as many do, she may never weave again.

 

The commercial value placed on Çömlekçi's rugs by outsiders has fostered pride in the local weaving heritage. The Çömlekçi Cooperative, though  now essentially defunct, did serve a significant purpose in that it brought the weaving and dyeing traditions of Çömlekçi to the attention of the outside world. In 1964 it seemed to be only a matter of time until carpet weaving would be abandoned.  Today it appears that there will still be carpets woven in Çömlekçi in the next century.

 

 

REFERENCES

 

Kocakaya, Bunyamin, Deputy director, Handicrafts Project. Turkish Ministry of Culture, Ankara. Personal communication, June 21, 1994.

 

Miller, Charlotte Jirousek (1982). "Dyes in rugs from the Milas area." Hali (London) v.4, no.3. pp.258-261.

 

Turkish Handwoven Carpets, Catalog No: 1. (1987). G. Erbek, Ô. ànal, D. Y2lmazkaya, eds. Ankara: T.C. kültür Bakanl22.

 

Turkish Handwoven Carpets, Catalog No: 2. (1988). G. Erbek, Ô. ànal, D. Y2lmazkaya, eds. Ankara: T.C. kültür Bakanl22.

 

Turkish Handwoven Carpets, Catalog No: 3. (1990). G. Erbek, Ô. ànal, D. Ylmazkaya, eds. Ankara: T.C. kültür Bakanl22.

 

Turkish Handwoven Carpets, Catalog No: 4. (1990). G. Erbek, Ô. ànal, D. Y2lmazkaya, eds. Ankara: T.C. kültür Bakanl22.