Click for more details.



Size: 23.5 x 13 cm , 27.5 x 10.5 cm, 33 x 9.5 cm, 32 x 8.5 cm

The first three pieces show what Lamm believed to be fragments of border stripes and, as he did not publish the fourth, no reconstruction is available for his opinion. While Lamm recognized the three he published were from several differ carpets, the following discussion will re-consider this and some other points of interest.

Of these examples, Plate Four(10) is in the best state of preservation and clearly shows a fine, small carpet’s complete side and end borders with an additional 2 1/2 inch swatch of rosette decorated blue ground pile as well. Lamm thought this fragment to be from a Holbein Carpet of the 15th century(11). I do not agree. Exactly what field design was present cannot be determined from this small corner but the border treatment is totally inconsistent with a Holbein attribution.

The three other Plates on first glance also appear to be border fragments but not all of them are. In places little of their original design remains and this, together with their small size, complicates knowing anything conclusive. However, it doesn’t prevent reading into what is there and suggesting some tentative ideas. One fact is for sure, they are enigmatic to say the least.

Lamm believed two, Plates Five(12) and Six(13) were from different carpets according to his technical analyses and reconstructions. I don’t agree and will spell out my reasons and also why only Plate Six was originally a border but, regardless, these two were most probably from the same carpet. Although Lamm’s structural analysis implies slight differences those variations are not, in my opinion, at odds with the idea both are from the same carpet - Plate Six a fragment of the border and Plate Five from the field.

Plate Seven’s(14) technical characteristics are, on first glance, also similar and thoughts these three fragments were made at around the same time and in the same place, perhaps even by the same weaver(s), are not without merit. However, Plate Seven was not be part of the same carpet and should be considered as having been part of another one. Taken as a group these three narrow strips of knotted-pile weaving have far more complex designs and appearance than Plate Four. This implies they pre-date it. Furthermore the strong design similarities Plate Six maintains with Plate Four hint it, or one of its contemporaries, was its archetype.

The first clues leading to this assumption are more complex designs and a much more sophisticated articulation of those designs. If Plate Four is interpreted as showing only borders, which is in my opinion the fact, the blue ground area and partially visible designs are from the main border and not part of a field design. If this was the case it is then easy to understand how they are, when combined, a re-interpretation of Plate Six’s rosette and tripartite design. Transliterations like this, when designs like the rosette and tripartite design are divided into two separate border stripes or the corollary when two separate designs like them are combined into one border as Plate Four shows, were not unusual occurrences and many others can be noted when related weavings from different time periods are analyzed.

Fig. 24

Of course this is speculative but let me go even farther out on that imaginary limb. Plate Four’s large rosette main border might well also be a prototype, not for another ancient weaving but for group of much later rugs known as Talish. These carpets, which form an unusually homogenous group in weave, design and age are named after an area of the southern Caucasus near the Caspian Sea where it is believed they were produced. All genuine examples have the same main border of large rosettes alternating with unique pairs of boxes, very special wool quality and brilliant coloration (fig.24). Plate Four displays these same physical characteristics, which may very well signify it was also woven in or around this area.

In addition when each of the four boxes from the Talish border are carefully examined, they can be seen to be divided into eight parts by the intersection of four strokes or solid lines. Unlike the rosette this square divided into eight compartments is a very uncommon design and is rarely seen in any type of weavings made before the middle of the 19th century. Before then it exclusively appears in Talish rugs but after that time it does seem to turn up in other pile and flat-weaves made by a diverse number of weaving groups from other areas. It might appear this design is a flower with eight petals but I cannot buy into that interpretation and must suggest another.

Before doing this brief mention should be made of the cohesive and proscribed character present in all the traditional weaving cultures that produced complex patterned knotted-pile and flat-weaves. Their strict customs were unquestionably the reason why many patterns and individual designs remained viable for such incredibly long periods of time and over such broad geographic distances. A rigid cultural and non-secular tradition bound each weaver within highly defined guidelines requiring exact and faithful adherence. These controls limited personal expression and locked the weaver into reproducing a pattern with all the constituent elements in a proper order and placement.

These weavers wove to express cultural identity and tradition not individuality.

This is extremely important because it enabled these weaving societies to remain resistant to the revolutionary political, religious and economic systems that were introduced throughout the entire Near East beginning in the 12th century. However, it did not make them impervious to these changes. Slowly in some areas and more quickly in others their cohesive character was eventually overwhelmed and, as a direct result, new designs and traditions developed and became popularized.

Exactly how patterns and individual design elements, like those in the Lamm fragments or the Ala al-din/Beysehir Mosque rugs, were affected by or resulted from these changes or how they were then transmitted and carried to other weaving areas is not easily demonstrated. But be assured the relationships posed by these and other similarly constructed parallels present the existence of a potent inter-generational weaving culture that remained viable over long time periods and across great geographic distances.

What were the intermediary steps and cultural relationships connecting them to the subsequent weaving traditions like the Talish or others that can be demonstrated? Mostly they remain unrecognized and even when the most obvious are found no positive evidence is available as proof. This is the situation Plate Six, Plate Four and the Talish rugs present even though they share material and design similarities. But with other less obvious instances where a design has undergone change, unlike in this instance where the rosette and tripartite symbols have remained static and consistent, identifying a relationship like this one can be difficult or impossible. However, sometimes these too can be detected when the process or catalyst responsible for the change is considered.

Three of these - mirror imaging, design doubling and its converse effect the halving of a design - are simple but effective mechanics of design transformation and manipulation. Recognizing their influences can sometimes provide an answer to explain how two designs are related, a newer version generated from a pre-existing one.

For instance, by doubling the tripartite design from Plate Four(fig.25) a
Fig. 25
new design(fig25a) is produced. Notice how very similar to the unusual boxes in the Talish border this new design became. Is this the source of the Talish pairs of boxes design? While being somewhat adventurous, this
Fig. 25a
suggestion is not without merit as Plate Four has the rosette design, similar wool characteristics and the brilliant dye coloration and saturation of the later Talish weaving tradition. These reinforce the possibility Plate Four was woven in the same locale, southeast Caucasus near to the Caspian Sea where the Talish rugs are believed to have originated, and both traditions shared a common design iconography, ie. rosette and tripartite symbols. For the record, these physical details are entirely different from those Plates Five, Six and Seven display but these other weaving do seem to share some iconographic commonality, again the rosette and tripartite symbols, with Plate Four and the Talish rugs. Does this suggest they are the archetype?

Reciprocal patterning or negative imaging, like doubling or halving, can also provide clues to a design’s origin or source. Another long stretch of imagination, yet equally intriguing, proposes a reciprocal of Plate Four’s tripartite element (fig.25) was the source design for a main border design found on an early rare soumak weaving(fig.25b).
Fig. 25b
This is a saddlebag face and it was most probably made in the southern Caucasus, north of Talish, perhaps in the area called Konagend. While not an exact match, the borders of both this Plate and the soumak do display design, color and proportional similarities, especially when the reciprocal of the soumak’s is used for comparison.

This soumak’s large octagon medallion features four large fantastic animals(fig.26). They depict an archaic style of animal
Fig. 26
drawing that appears to have originated in western Anatolia and is unknown in any other weavings from the Caucasus. This furthers conjecture the southern Caucasus and Caspian Sea areas maintained a direct relationship with the weavings traditions present in eastern and central Turkey where these mythical quadrupeds appeared on knotted-pile carpets made from the 13th to the 16th centuries.

By applying the mechanical process of design doubling to a fragmentary design element, like the one that appears in Plate Five, it is possible to suggest how it might have originally appeared. It is unfortunate more of this fragment does not remain to answer the intriguing riddle it raises: Was it from the border or the field? Lamm believed it was from a border while my opinion favors the field. By doubling the design a small complex medallion is produced and, if this suggestion is correct, these medallions would have been placed in rows all-over the field of the carpet
Fig. 27
Plate Five originally came from. Again this might appear speculative, however, other Lamm fragments, like the following two illustrated here, Plates Eight and Nine, as well as several of the carpets from the Ala al-din Mosque (fig.27 and fig.28) have complex medallions of this type in all-over field
Fig. 28

One last mention about these four fragments, Lamm’s structural analyses list a left-handed twist and, as was the case with others, it is actually right-handed or ‘S’ ply. The presumption his description was referring to plying seems most likely but it is also possible he was describing the spinning of each of the two woolen strands that were then plied together to form the warp threads. They, in fact, are spun with a ‘Z’ or left-handed spin and then plied with a ‘S’ or right-handed ply. This would explain Lamm’s apparent error and, as no other information is available, no definite conclusion can be forwarded.

But why is this detail of spin or ply significant? Basically, this difference helps to categorize individual weavings- those from Egypt and the southern tier of the Mediterranean are invariably ‘S’ spun ‘Z’ plied and those from Anatolia and the northern tier ‘Z’ spun and ‘S’ ply. This technicality is not 100 percent positive, however, it does prove useful in generally indicating which of these two broad geographic areas the weaving came from. Presumably Lamm knew this and his left-handed twist was describing spin and not ply.



10. This knotted-pile carpet fragment is illustrated as number 29 in the “Carpet Fragments” catalog.
11. ibid. pg 50
12. This knotted-pile carpet fragment is illustrated as number 7 in the “Carpet Fragments” catalog
13. This knotted-pile carpet fragment is illustrated as number 8 in the “Carpet Fragments” catalog
14. This knotted-pile carpet fragment is illustrated as number 38 in the “Carpet Fragments” catalog


Home   Next Plate

Text 2003 WAMRI - The Weaving Art Museum and Research Institute.
All Rights Reserved. For any problems with this site contact:

click image for larger view