Part of the mystery and appeal of the Oriental carpet for collectors and scholars is its ability to incorporate designs from a wide range of sources. However, inquiry into origins of designs should do more than simply satisfy our curiosity. Ideally, it leads us to a better understanding of the economic and artistic contexts in which a rug is woven. This article will examine the use of Ottoman textile designs in Turkish rugs and the impetus for those artistic borrowings. It will also attempt to draw some conclusions as to how we define and use both textiles and rugs and how they relate to our built environment.
The occurrence in Turkish rugs of designs borrowed from textiles has been occasionally noted by scholars. Agnes Geijer proposed that two of the patterns on the so-called Seljuk rugs from Konya and Beyshehir were inspired by silk textiles. Her classic comparison of the field design of one of the rugs to a lotus pattern on a circa 1300 Chinese damask unearthed in Egypt has been reproduced many times (figures 1-2).
An early Turkish rug from Berlin is also quite likely copied from a Chinese textile. There we see a very closely patterned field of cloudbands and alternated stylised lotus flowers. The lotus and cloudbands are foreign elements to the carpet, in that they originate in Chinese art. In a broader sense, the use of such small repetitive patterns is foreign to the Turkic rug design tradition. It shares little in concept with the appearance of compartment and gul rugs as depicted in early miniature painting (figure 37) or with surviving Holbein rugs and their village cousins.
It is remarkable how many of the Konya-Beyshehir rugs have repeating patterned fields, often incongruously combined with wide, monumental borders (figures 3-4). While the borders evoke architectural decorative themes, their fields are like textiles in their use of small repeated motifs. Some also mimic the tonal effect of damasks by using only a few colours of similar value.
The models for the fields of several of these carpets are almost certainly silk textiles woven on a drawloom. Drawlooms are capable of weaving quite complex figures, but designs with short repeats are favoured, as they are easier to program. It is no problem for the a weaver of pile rugs to copy textile patterns. If she chooses to make very curvilinear cloudbands, as in the Berlin rug, it is well within her power, for knotted pile has little limitation of technique to restrict what it can reproduce. The weaver may, however, prefer to choose a stylised drawing of what in the original silk textile is likely a more curvilinear, fluid representation (figure 2). We can see from these early examples that Anatolian rug weavers are quite capable of copying textile designs in varying degrees of stylisation and have an early history of doing so.
Silk textiles and their trade were valued in Anatolia for over a thousand years before the Ottomans consolidated power. However, in the 15th and 16th centuries the importance of silk in the art and economy of the region reached a new level. Anatolia was a nexus in the silk trade linking Europe to the East. The Persians, Turks, and Italians traded, produced, and consumed huge quantities of raw silk and luxury fabrics. The silk trade became a determining factor in state policy and military objective. In Ottoman society, silk fabrics became an important emblem of rank, wealth, and reward. The court encouraged production and it controlled quality. 
With the rise in demand for silk came an explosion of creativity. An artistic vocabulary that was developed in the court ateliers for various media found imaginative application in textiles. Among design layouts, the repeating ogival became the most popular one of all (figure 5, figure 10). Its repetitive nature lent itself readily to composing for the drawloom and to piecing together narrow sections of cloth to make wider panels in continuous patterns. Use of the ogive in textiles saw widespread use, first employed by the Mamluks and then adapted in various forms by the Persians and Italians as well as the Ottomans.
This frenzy of fashion- and design-borrowing is evident in carpets and some kilims as well. Wool was more readily available and considerably less costly than silk. A surviving Ottoman kilim (figure 6) and an Usak pile carpet (figure 7)  are quite direct copies of a popular ogival textile (figure 5). The carpet is remarkable in that it is not just the motifs that have been reproduced: the weaver has also incorporated the concept of the narrow width of a textile woven on a drawloom. Narrow borders cut the ogives at regular intervals across the rug, mimicking the dimensions of silk fabrics. This is a conscious evocation of a luxury textile in a more humble medium. 
Sometimes comparison of the details of a textile and a rug reveal remarkable similarity of detail, and one can be confident the imitation was quite direct . Carnations, tulips, pomegranates, and hyacinths may be disposed in just the same orientations (figure 6, figure 8-9). Other times we cannot find the exact model, but we still know the type of textile design that inspired the rug. The ogival lattice of rumi leaves was a popular layout for silks (figure 10), appearing in various renditions. From there it was copied to wool (figures 11-12), allowing the rug owner to have a durable object in a fashionable style, one that cost less than the original textile.
When one is presented with two rugs that are different versions of the same textile design, it is interesting to try to decide whether both come directly from the original (figures 10-12). A carpet which reproduces great complexity of detail is likely to be quite close to the source (figure 12). Does this mean that a more simplified version is a later copy of that rug rather than of the textile itself (figure 11)?
A look at examples of lineal descendants reveals an increasing confusion as to the original concept (figures 12-14). The single ground colour is lost to a variety of hues, which alters the visual effect by breaking the rug up into compartments.The rumi leaf lattice can then only be recognised if its origin is already known. The weavers of these later rugs in early textile designs were copying from earlier rugs and, unaware of the textile inspiration, had little feeling for the original concept. A case can be made that a rug in a simplified design that retains the graphic clarity and beauty of the original textile is as close to or closer in time than a more complex version (figures 10-12).
The use of a shared design vocabulary for various media of Ottoman art raises the possibility that a design foreign to the rug tradition, even when appearing textile-like, may have a third source. This is particularly true for Ottoman "court" kilims, whose genesis is still an open question. It is clear that the designs featured are foreign to the kilim tradition, but what is their inspiration? A comparison of one such kilim to a textile seems to show obvious design influence (figures 15-16). However, while the overall effect is the same, important details of the drawing differ. In the kilim each saz leaf curls back on itself in the classic way for this style, in contrast to uncurled leaves facing alternate directions in the textile. It is very possible that the silk textile and the wool kilim have taken their inspirations separately from another medium. Ink on paper or tile revetments are possible models for such designs.
In the case of the pile rugs that are the focus of this article, the artistic circumstances of their manufacture seem less ambiguous than those of the kilims. None of these carpets would be characterised as designed "court" products. Even the earliest examples are different from the rugs (carefully drawn from cartoons) that we associate with the output of the Nakkashane or Ottoman imperial workshops. They are neither part of that court carpet production, nor are they copies of them. Among the varied media that could have inspired the designs in these rugs, tile revetments and textiles are likely candidates. They share the same design vocabulary and even some of the same scale of presentation and layouts, notably the ogival lattice.
Textiles, however, are the more likely avenue for the transfer of patterns to rugs. Cloth is very portable, and only a small section is needed for a weaver to see the repeat and make a copy. Textiles enjoyed high status and are obvious agents for the dissemination of fashionable designs.
Textiles inspired a variety of Turkish rugs, some more artistically successful than others, and sometimes the design takes on a whole new life in knotted pile. A rug found at Divrigi is a quite creative and powerful example of converting silk to wool (figures 17-18). Even in examples where the visual effect is very similar in textile and rug, the rug weaver manages to make the design her own. Thus, a popular silk fabric with an Italianate design becomes a Turkish rug by use of the weaver's own visual vocabulary in the details (figures 19-20). It is unclear whether the textile is Italian or was woven in Anatolia in an Italian design. In either case, its genesis and the subsequent transfer of the pattern from silk to wool is evidence of the demand for fashionable designs in the international marketplace.
Another extant carpet design might seem, at first glance, to be an angularized version of the same Italianate textile pattern (figures 20-21). However, the carpet has a reciprocal rather than directional layout and differs in the details of its drawing. The pattern is, in fact, deeply imbedded in the Turanian artistic culture, with widespread use as an architectural motif. Among other places, it appears in pre-Ottoman beylik stonework (figure 22), in Timurid tile decoration, and in Seljuk woodwork. While always drawn as a reciprocal design, it can be used as an overall decoration of a large surface, or replicated as a band to function as a frieze or other architectural divider.
Perhaps it is no surprise to find that this motif's appearance in textiles echoes its architectural uses. It can be seen as overall patterning on tents in miniature painting and as a frieze in Ottoman velvets (figure 23). With such widespread use of this design it remains an open question from exactly where the rug weaver took her inspiration. This is not so much a fashionable textile design as it is a part of the basic vocabulary of the artistic environment.
The theme of architectural conventions used in textiles and rugs is one to pursue when considering the yastik, or cushion cover form. Perhaps the most widely recognised imitation of Ottoman textiles by rugs is the use of velvet yastik designs in wool pile versions. The yastik is a popular format for Ottoman velvets from at least the 16th century onward. Many knotted pile wool yastiks from the 18th and 19th centuries preserve the velvets' designs (figures 24-27).
One curious feature that became a convention in both silk and wool yastiks is the single row of small compartments, or lappets, that frequently appear on each end (figure 28-29). The logic or reason for their existence is not immediately apparent, but a consideration of how velvet yastiks were woven and used suggests some answers.
Weaving a velvet on a drawloom is quite mechanically done, but the setting up of the sequences for the manipulation of the loom to produce the design is a complicated task. The shorter the length of weaving before the design repeats, the easier the set-up. Thus a cloth with repeating carnations has a repeat one carnation long and is easier to program than a medallion design (figure 24 and figure 30). Once the loom has been set up with a repeating design, meter after meter can be woven off. Then, if desired, the cloth, typically only about 65cm. wide, can be sewn side to side to make a continuous broader design.
The simplest way to transform a piece of fabric so it becomes a useable object straight off the loom is to introduce rows of lappets periodically. One thus creates a series of identical yastiks, as many as the length of the warp can accommodate (figures 30-32). In the simplest way, with the least reprogramming of the weaving sequence, a piece of fabric has had its ends "finished." It has been made into furnishing and appropriate to an architectural setting. Perhaps it is not surprising that there are so many examples of wool pile yastiks with textile-like patterns in this same design concept. They have repetitive field patterns, little in the way of side borders, and lappets on each end (figures 33-36). Often the lappets have shrunk from their original proportions on velvet yastiks, but they remain as a reminder of their architectonic defining purpose.
The choice of lappets as a type of end border in defining the form of Ottoman silk yastiks and their wool descendants evolves naturally from the artistic heritage embraced by the Ottomans. Timurid and Jalayrid miniature painting depict yastiks in use, some of which have forms analogous to the end lappets of Ottoman yastiks (figure 38). In those same paintings we see shapes in rows like lappets used generally as dividers or borders in a variety of architectural uses. Examples of the use of rows of lappet shapes to demarcate a space can be seen on tops of walls (figure 37), as decorative friezes such as on the socle or edge of raised platforms for sitting (figure 38), and elsewhere.
There are of course many velvet yastiks, and their knotted pile descendants, which do not use repeating textile patterns. Those with medallion designs we consider to be inspired by the art of the book. There are two other influences to note, however. First, the appearance of lappets at each end has become a defining convention. They usually remain, even when other borders have been added. Second, the yastik with borders all around is an echo of the architectonic design of the oriental carpet. Rugs, composed of borders, guard stripes, and central fields, have always been used to define spaces in both tent and town. So while the Konya-Beyshehir mosque carpets imitate textiles with their closely patterned fields, they establish an architectural grandeur with their wide borders (figures 3-4).
In the nomadic environment rugs are even more important in defining space, where they not only echo architecture, but in a sense are the architecture. What more natural way to integrate a yastik into the built environment than to give it borders so that it looks like a carpet?
The Ottomans wove other textiles in carpet formats, called Nihale. Many velvets survive that were woven with borders along the edges on narrow silk looms in such a way that the product could be pieced together to form carpet-like textiles (figure 17). Brocaded satins were also made in carpet formats, some in medallion designs up to eight meters long.
It is illuminating to see the various instances where rugs and textiles overlap in pattern and conception, for if rugs copy textile patterns, textiles emulate the design of rugs in order to relate to the built environment. An understanding of the ability of rugs and textiles to respond to fashion, absorb foreign influences, and define themselves in the architectural environment can only deepen our appreciation of their versatility and art.
1. This paper was first presented at the Second International Congress on Turkish and Central Asian Carpets in Istanbul on October 16, 1994. The number of illustrations here has been reduced from the 52 colour slides shown then. Back
2. May H. Beattie, "Some Rugs of the Konya Region," Oriental Art 22.1 (Spring 1976): 60-76. See also Jon Thompson's comments in Sotheby's (New York) auction catalogue Turkmen and Antique Carpets from the Collection of Dr. and Mrs. Jon Thompson, December 16, 1993, lot 80. Back
3. Agnes Geijer "Some Thoughts on the Problems of Early Oriental Carpets", Ars Orientalis V, 1963 pp.79-87. Reproduced in Hali 48 ( Dec. 1989) p.45. Back
4. Illustrated in Hali 71 (October/November, 1993), p.87. Back
5. For a discussion of the an early Persian rug fragment also with limited palette and copied from a textile see Louise W. Mackie, "A Piece of the Puzzle", Hali 47 (October 1989), pp.16-23. For other examples of early Turkish rugs with textile-like fields of limited colours see plates 4 and 5 in B. Balpinar and U. Hirsch, Carpets of the Vakiflar Museum Istanbul (Wesel: Uta Hulsey, 1988) and the Sion Usak in Hali 27 (1985) p. 38. Back
6. The other source for small pattern designs on the Konya-Beyshehir rugs is village flatweaves. See Werner Bruggemann, "Ein broschiertes Flachgewebe aus Zentralanatolien", Hali 2,3 (1979) pp.206-211 and Harald Bohmer, "Turkish knotted Carpets and Flatweaves with Similar Designs", Oriental Carpets and Textile Studies 4 (Berkeley, 1993) pp. 57-66. Marla Mallett's precept that a shared design originates in the more restrictive of two given weaving techniques supports their idea that designs migrate from flatweaves to pile weaves rather than the reverse. Deciding which of the remaining Konya-Beyshehir small patterns might come from drawloom textiles and which might come from village flatweaves presents an interesting challange. For example, the flatweave design discussed by Bruggemann and Bohmer also appears in a Qing Zhuang jin silk, reproduced in Gao Hanyu, Chinese Textile Designs (Penguin: London, 1992), plate 119. Back
7. Halil Inalcik, "Harir," Enclopedia of Islam, new ed. 3: 211-218. Back
8. See Louise W. Mackie, "Towards an Understanding of Mamluk Silks", Muqarnas 2 (New Haven, 1984), 127-46, where she also discusses Yuan ogival textiles in Mamluk style. Back
9. It is perhaps significant that this rug was discovered in a mosque in Uskudar, a place whose importance as a weaving center is recognized by the name "Scutari" for a type of late velvet yastik. If textile weaving was a local industry, there may have been a particular local appreciation of textile-like rugs. Back
10. Serare Yetkin, Historical Turkish Carpets (Istanbul, 1981), p. 87. Back
11. Another rendition of the design, perhaps the most confused of them all and better viewed upside down, is the Frauenknecht piece, Hali 42 (Nov/Dec 1988), p 85. Back
12. Jon Thompson has suggested that figure 12 is earlier than figure 11. Loc. cit. see note two above. Some evidence for an early date for figure 11 is found in John Mills' discovery that the depiction in paintings of its border occurs only at very early dates. He cites: 1) Jaime Huguet, Consecration of Saint Augustine, 1463-85. Barcelona, Museum of Catalonian Art . 2) Master of Saint Augustine, Scenes from the Life of Saint Augustine, circa 1500. Dublin, National Gallery of Ireland. Personal communication. Back
13. For examples of that production see Esin Atil, The Age of Suleyman the Magnificant (New York: Harry N. Abrams, 1987), 224-30. Back
14. Nevber Gursu, The Art of Turkish Weaving (Redhouse Press: Istanbul, 1988) pp.67-93. Back
15. Examples of this silk survive in numerous European collections and in the Topkapi Sarai Museum collection as well. Perhaps it was international market demand for salable merchandise that produced another rug in a textile design, a "Damascus" carpet not in the standard "chessboard" design, reproduced in Balpinar and Hirsch, Carpets of the Vakiflar Museum Istanbul (Wesel: Uta Hulsey, 1988) plate 58, inv. no. A-172. Back
16. Other examples are in the Museum of Islamic Art, Berlin and in the Wher collection , the latter published in Turkische Kunst und Kultur aus osmanischer Zeit (Aurel Bongers: Recklinghausen, 1985) II, p. 177. Back
17. On the entrance portal of the Gur-i Amir complex, Samarqand, c. 1400. Reproduced in T.W.Lentz and G.D.Lowry, Timur and the Princely Vision (Smithsonian Institution Press:Washington, 1989) p. 154. Back
18. On the ceiling of the Aslanhan mosque, Ankara. Back
19. As in a Turkman-style painting from circa 1505, reproduced in B.W. Robinson, Persian Paintings in the India Office Library (London, 1976), plate 92. Back
20. Another motif that appears in textiles and rugs permeated the artistic environment and became fashionable as well. The cintamani or three balls and two wavy stripes design preceeded Ottoman art and was then wholeheatedly taken up by it in all media. See Gerard Paquin, "Cintamani", Hali 64 (August 1992) pp. 104-119. Back
21. For archival evidence from 1493 see Gursu, op.cit p. 161. Back
22. Walter Denny, "Ottoman Turkish Textiles" Textile Museum Journal III,3 (December, 1972) pp. 55-66. Back
23. Also frequently in the decorative scheme of wood carvings on Seljuk minbars, another architectonic use. Back
24. Ibid. Atil pp.215-220. For a velvet in carpet design complete with major and minor borders see Fredeich Spuhler, Islamic Carpets and Textiles in the Keir Collection (Faber and Faber: London, 1978) p. 227. Back