The Earliest Known Carpet

A survey of the history of knotted pile carpets requires us to distinguish clearly between carpets woven with a knotted pile and those produced by the much older techniques used for carpets with a looped pile and for flatweaves. The last represent the interweaving of warp and weft threads, while looped pile represents a complicated development of flatweave in which a pile is created by passing the weft over rods laid on the surface. This forms loops, inclined towards the weaver, which are cut open after the completion of one or more rows. In other words, the pile is inserted in the warp-weft system as a loop, not as a knot; it is also referred to as cut loop or cut pile.

We do not know when the rows of loops were superseded by individual knots. In other words, we do not know when knotted pile carpets were first made. In the latter, each knot is individually wrapped round two adjacent warps and the thread cut. This creates the basis for working more subtle designs that could, in theory, change from one knot to the next.

It is generally agreed that nomads were the first people to use the pile knotting technique, a fact which does little to simplify the search for its origins. When scholarly research into Oriental carpets first began in the latter part of the 19th century - references to carpets were collected with encyclopedic discipline from Homer to the Bible. We are now, once again, aware of the great historic value of this work. However, since the early literature covers a mixture of textile wall hangings and floor covers without technical distinction, we are not yet able to exploit such references for the history of the knotted pile carpet. Kurt Erdmann made this clear division a prerequisite for his history of the knotted carpet and at this point I see no reason for abandoning it.

A survey of the history of knotted pile carpets requires us to distinguish clearly between carpets woven with a knotted pile and those produced by the much older techniques used for carpets with a looped pile and for flatweaves. The last represent the interweaving of warp and weft threads, while looped pile represents a complicated development of flatweave in which a pile is created by passing the weft over rods laid on the surface. This forms loops, inclined towards the weaver, which are cut open after the completion of one or more rows. In other words, the pile is inserted in the warp-weft system as a loop, not as a knot; it is also referred to as cut loop or cut pile.

We do not know when the rows of loops were superseded by individual knots. In other words, we do not know when knotted pile carpets were first made. In the latter, each knot is individually wrapped round two adjacent warps and the thread cut. This creates the basis for working more subtle designs that could, in theory, change from one knot to the next.

It is generally agreed that nomads were the first people to use the pile knotting technique, a fact which does little to simplify the search for its origins. When scholarly research into Oriental carpets first began in the latter part of the 19th century - references to carpets were collected with encyclopedic discipline from Homer to the Bible. We are now, once again, aware of the great historic value of this work. However, since the early literature covers a mixture of textile wall hangings and floor covers without technical distinction, we are not yet able to exploit such references for the history of the knotted pile carpet. Kurt Erdmann made this clear division a prerequisite for his history of the knotted carpet and at this point I see no reason for abandoning it.

 known_carpet_2

 

The oldest knotted carpet known to us was discovered by S. J. Rudenko in 1949 in the tomb of a Scythian prince in Pazyryk, in the Altai mountain (it retains its original format of c. 200 x 183 cm and is now kept in the Hermitage Museum in Leningrad). By reference to other objects found in the tomb, it has been dated to the 4th century BC. It has been shown that this carpet is not, as was thought for a long time, a looped-pile textile but a knotted carpet of astonishing technical perfection (3,600 symmetrical knots per square decimeter), which can only have been attained as part of a long tradition. The drawing, especially that of the border frieze with its elks and horses and riders, also contradicts the idea that this carpet might be an example dating from the beginnings of knotted-pile weaving. So far, we only have contradictory opinions as to its place of origin. The style of the drawing corresponds to the art of the Achaemenian period in Persia. Up to now it represents the only complete Oriental knotted-pile carpet known to have survived from the pre-Islamic period, and it is stylistically of a very high standard. It already shows the principles of surface composition for a carpet: a repeat pattern in the inner field and a border composed of a number of stripes. There are no examples of knotted carpets from the next goo years. At the beginning of the 2cth century, Sir Aurel Stein and A. von LeCoq undertook several expeditions to Central Asia, with Turfan as their main objective. Of the tiny carpet fragments they brought back, four are now in the Victoria and Albert Museum, one in the British Museum6 and several more in the National Museum of India in New Delhi. A fragment which was in Berlin, where it was published by F. Sarre, is now missing. This was brought back by LeCoq from Khotcho from the 4th Turfan expedition and Sarre dated it to the 5 th-6th century AD. We have no reason to doubt Sarre's statement that it came from a knotted carpet this is confirmed by T. Falkenberg's technical description and knot diagram. The fragment, which measures 26 x 116 cm, does not reveal a recognizable design. According to the drawing it was woven in the knot which we now call the 'Spanish knot'. Among the fragments of similar size from the area of Lop Nor, which are held in London, there are pieces which are undoubtedly woven in the symmetrical knot (e.g. Victoria and Albert Museum, both from the same carpet). The wool warp is thin. There are S 7 wefts between the rows of knots, which explains the coarse structure. What remains of the design section shows only various horizontal stripes, one of which has lozenge-shaped elements. The range of colors is surprisingly wide and includes up to twelve shades. The suggested date is between the 3rd and 6th century. No attempt has been made to date the pieces more precisely. There is no reason to think that the carpets may have been imported into the area. During the same period, felt carpets were in use, the designs of which were created by filling unspun wool into wool felt. Examples of this technique were found at the same sites. The occurrence at the same time of felt carpets and of the much more labor-intensive knotted pile carpets is surprising. However, a parallel is found in Turkestan carpets of the 19th-20th century which were also made in the same two techniques. This simultaneous use of two techniques makes it more difficult to assess the floor carpets depicted on the Turfan frescoes, which had already been referred to by Sarre. It is impossible to decide whether these paintings show knotted carpets, felts or possibly silk textiles. The designs vary between spiraling vines, some symmetrically and others randomly arranged over the ground, rows of stripes, four-pealed rosettes which cover the ground, and concentric segments of circles in a tile pattern, resembling waves of water. One of the frescoes has a particular bearing on the so-called Chintamani design which was very popular in Turkey in the 16th and 17th centuries. This consists of three balls arranged in a triangle with two parallel curved stripes. The lower part of the fresco shows a textile with this 'tiger and panther skin' design hanging from the roof of a hut and in another place serving as a floor cover for a Brahman. There is no evidence suggesting the use of a knotting technique.

It is a significant fact for the history of carpets that knotted pile carpets were in use in Central Asia in the 3rd to 6th centuries, which in technique were far behind the stage of development of the Pazyryk carpet. The designs, as they appear from the small fragments, were simple, although the wide color spectrum suggests greater experience. The frescoes show that carpets were not exclusively used in tents, either over floor felts or, on special occasions, replacing them, but that they were also employed at audiences or special gatherings where they marked the places at which certain groups of persons stood, and where they offered extra comfort for kneeling or sitting.

There is a double degree of uncertainty about the smallest of the carpet fragments, known as Fustat fragments. They are so tiny that it is virtually impossible to recognize a pattern. But even more important is the fact that they came direct from uncontrolled excavations into the art market where everything that was small and 'mysterious' was misdescribed as from Fustat (Ancient Cairo). According to my experience, these pieces, which are clearly the remains of knotted pile carpets not woven in looped pile - come from Turkish or, in some cases, Spanish fragments of familiar carpet types, of which we know large examples datable to the 13th-17th centuries. The oldest knotted carpet known to us was discovered by S. J. Rudenko in 1949 in the tomb of a Scythian prince in Pazyryk, in the Altai mountain (it retains its original format of c. 200 x 183 cm and is now kept in the Hermitage Museum in Leningrad). By reference to other objects found in the tomb, it has been dated to the 4th century BC. It has been shown that this carpet is not, as was thought for a long time, a looped-pile textile but a knotted carpet of astonishing technical perfection (3,600 symmetrical knots per square decimeter), which can only have been attained as part of a long tradition. The drawing, especially that of the border frieze with its elks and horses and riders, also contradicts the idea that this carpet might be an example dating from the beginnings of knotted-pile weaving.

So far, we only have contradictory opinions as to its place of origin. The style of the drawing corresponds to the art of the Achaemenian period in Persia. Up to now it represents the only complete Oriental knotted-pile carpet known to have survived from the pre-Islamic period, and it is stylistically of a very high standard. It already shows the principles of surface composition for a carpet: a repeat pattern in the inner field and a border composed of a number of stripes.

There are no examples of knotted carpets from the next goo years. At the beginning of the 2cth century, Sir Aurel Stein and A. von LeCoq undertook several expeditions to Central Asia, with Turfan as their main objective. Of the tiny carpet fragments they brought back, four are now in the Victoria and Albert Museum, one in the British Museum6 and several more in the National Museum of India in New Delhi. A fragment which was in Berlin, where it was published by F. Sarre, is now missing.7 This was brought back by LeCoq from Khotcho from the 4th Turfan expedition and Sarre dated it to the 5 th-6th century AD. We have no reason to doubt Sarre's statement that it came from a knotted carpet this is confirmed by T. Falkenberg's technical description and knot diagram. The fragment, which measures 26 x 116 cm, does not reveal a recognizable design. According to the drawing it was woven in the knot which we now call the 'Spanish knot'.

Among the fragments of similar size from the area of Lop Nor, which are held in London, there are pieces which are undoubtedly woven in the symmetrical knot (e.g. Victoria and Albert Museum, both from the same carpet). The wool warp is thin. There are S 7 wefts between the rows of knots, which explains the coarse structure. What remains of the design section shows only various horizontal stripes, one of which has lozenge-shaped elements. The range of colors is surprisingly wide and includes up to twelve shades. The suggested date is between the 3rd and 6th century. No attempt has been made to date the pieces more precisely. There is no reason to think that the carpets may have been imported into the area. During the same period, felt carpets were in use, the designs of which were created by filling unspun wool into wool felt. Examples of this technique were found at the same sites.

The occurrence at the same time of felt carpets and of the much more labor-intensive knotted pile carpets is surprising. However, a parallel is found in Turkestan carpets of the 19th-20th century which were also made in the same two techniques. This simultaneous use of two techniques makes it more difficult to assess the floor carpets depicted on the Turfan frescoes, which had already been referred to by Sarre.

It is impossible to decide whether these paintings show knotted carpets, felts or possibly silk textiles. The designs vary between spiraling vines, some symmetrically and others randomly arranged over the ground, rows of stripes, four-pealed rosettes which cover the ground, and concentric segments of circles in a tile pattern, resembling waves of water. One of the frescoes has a particular bearing on the so-called Chintamani design which was very popular in Turkey in the 16th and 17th centuries. This consists of three balls arranged in a triangle with two parallel curved stripes. The lower part of the fresco shows a textile with this 'tiger and panther skin' design hanging from the roof of a hut and in another place serving as a floor cover for a Brahman. There is no evidence suggesting the use of a knotting technique.

It is a significant fact for the history of carpets that knotted pile carpets were in use in Central Asia in the 3rd to 6th centuries, which in technique were far behind the stage of development of the Pazyryk carpet. The designs, as they appear from the small fragments, were simple, although the wide color spectrum suggests greater experience. The frescoes show that carpets were not exclusively used in tents, either over floor felts or, on special occasions, replacing them, but that they were also employed at audiences or special gatherings where they marked the places at which certain groups of persons stood, and where they offered extra comfort for kneeling or sitting.

There is a double degree of uncertainty about the smallest of the carpet fragments, known as Fustat fragments. They are so tiny that it is virtually impossible to recognize a pattern. But even more important is the fact that they came direct from uncontrolled excavations into the art market where everything that was small and 'mysterious' was misdescribed as from Fustat (Ancient Cairo). According to my experience, these pieces, which are clearly the remains of knotted pile carpets not woven in looped pile - come from Turkish or, in some cases, Spanish fragments of familiar carpet types, of which we know large examples datable to the 13th-17th centuries.

Pazyryk Carpet. New insights.

Christine Klose

The Pazyryk Carpet (fig.1) has frequently been the object of thorough studies. It is the oldest known knotted carpet, preserved in a Scythe tomb (kurgan) from the 4th -3th century B.C. Valuable information is in the book of Sergey I. Rudenko (in English translation: “The Most Ancient Carpets and Textiles of the World”, Moskau 1968), and an article by Ludmilla Barkova (HALI, issue 107,1999, pp. 64-69): “The Pazyryk—Fifty Years on"). 

known_carpet_6

known_carpet_7

As is appears, no one so far has given notice to the fact, that the pattern on the coat of the 24 broad-antlered spotted fallow deers, forming a procession in the second outward wide border, has an anatomical meaning (fig.2).

Lecturing on Oriental carpets in the city of Thun (Switzerland), I was puzzled by the remark of a medical doctor in the audience. He, a passionate hunter, said, that these extra figures are depicting the inwards and the vertebra of the deer, all parts in real positions with nearly clinical precision (fig.3): 1. The heart, just above the, front legs (a yellow framed red sphere, black contoured). 2. The aorta (a long red protuberance on the heart). 3. The maw, on the right hand side of the sphere (a large yellow area with a widening upwards on the end). 4. The intestine, in the rear end (a yellow square surrounded by a light blue and a yellow bow). 5. Possibly the urethra, on the upper part of the right hind leg (a yellow line with a black point), better to see on some others deer on the border. 6. The vertebra, directly below the brown back contour (an alternating black-white chain).

The most common species of a fallow deer in Asia nowadays is known as "Mesopotamian deer" [E. Ueckermann, Das Damwild, (Hamburg, Berlin, 1983) pp.14/15]. Ueckermann also mentions a stone relief (fig.4) from the palace of Dareios in Persepolis (p. 15) showing a feudal servant who offers such a deer for the king.

known_carpet_8

known_carpet_9

One may think that deer were highly valuated; they may have been also connected with ritual procedures. Since Mesopotamia was part of Dareios' empire, one may think of similar rituals there. From Siberia, where Pazyryk is located, we have many representations of deer by Scythian craftsmen. But the character of the Pazyryk carpet is not Scythian. The experts suppose Mesopotamia as its origin. The clue for this supposition is the stony so called threshold carpet from Ninive, now exhibited in the British Museum. It has a very similar pattern as compared with the field of the Pazyryk carpet. Therefore, the name "Mesopotamian deer" and the relief in Persepolis are additional indications for Mesopotamia as origin of the Pazyryk carpet.