Among the few representations of rugs that survive in 15th century paintings are those of the Venetian master Caroli Crivelli. One such rug appears in a pinnacle of a polyptych, dated A.D. 1482, and was published by Kurt Erdmann.1 In this painting the rug is displayed draped over a sill, directly above the kneeling Virgin of the Annunciation (Illustration 1). It is surmounted by a large jug whose weight anchors the rug in place. Mechanically, a rug so arranged is most stable if its weight is distributed equally on either side of the sill. It follows that the visible portion is half of the rug. We may further assume, given the rule of bilateral symmetry, that the unseen half is a mirror image of the visible portion. Thus reconstructed we can describe a rug pattern consisting of a single large 16-pointed medallion filling a light colored field.
The medallion is constructed of multi-shaped polygons arranged radially about a central square. Two partly detached lozenges appear to fission from the parent medallion. Each polygon exhibits double-outlining, i.e., the outline is itself outlined. This emphasizes the illusion of relief on the surface. Some polygons contain star or bird forms. It was these stylized bird motifs that led Erdmann to categorize this rug as "a fifteenth century Anatolian animal carpet." But the casual inclusion of such motifs is a common feature of Caucasian rugs. Parts of the medallion are adorned with latch-hooks, and the rug ends with a skirt simply furnished with two horizontal stripes. Now, whether this sparse border continues along the sides is a matter for conjecture, since in this painting Crivelli has carefully furled the sides, thus effectively obscuring these details.
Crivelli provides a sill and a jug to guide us in measuring this rug. If we estimate the jug to be 10 inches in diameter, and the sill is three times longer, then the rug, which is furled to fit, must be some 3 feet wide. The ratio of width to visible length is about 2:1. The total length is twice the visible portion plus the depth of the sill. The jug projects slightly over the edge, hence a narrow sill, 6 inches. We compute, therefore, a rug some 3 1/2' feet long by 3 feet wide. An image now emerges of a rug with a square field, an elementary border, and furnished top and bottom with 3-inch plain skirts.
In 1486 (four years later) Crivelli painted another Annunciation2 (Illustration 2). In the foreground, next to a perched peacock, we see the same or a similar jug, now planted. Nearby Crivelli proudly displays his "new" large- patterned Holbein rug, conspicuously scrutinized by the envious peacock. This rug, also only partly displayed, is called "Azerbaijan" by Latif Kerimov (Azerbaijan Rugs, Vol. I, Fig. 7) but "Anatolian" by most others. In the background of Crivelli's lavishly furnished environ, and draped over a parapet, appears another familiar object (Illustration 3). Can this be Crivelli's medallion rug, now four years older?
Displayed in almost the same fashion as before, but with the full width plainly evident, we can detect certain differences among the similarities of these two representations. The bird forms appear more elongated than before and, around the medallion's periphery, a proliferation of latch-hooks, now 10 in number from 4 earlier.
But it is in the elaboration of the border that the major difference exists. A churn or tuning fork (soldat) motif surrounds the field. There follows an outer border stripe, whose vertical sides differ markedly from the horizontal. We observe here the willingness of the weaver to make progressive changes in her designs.
Crivelli's designs are to be found in actual rugs. Published examples span a wide range of places, periods, and weaves. One, in keeping with the single medallion format, was found in a mosque in Savrihisar (Turkey) by Nejat Diyarbekirli (Hali 6/4, p. 459). The 16-point medallion occupies a light red field which it shares with spandrels of mixed design. A main border consisting of Greek crosses is surrounded by a double pearl (88) border. Bird forms but no latch-hooks appear. The rug may well date to 1400 but is so crudely rendered that Crivelli would have disdained it.
A more impressive fragment is preserved in the Budapest Museum for Applied Art.3 While only the right half survives, the design is complete, for the rule of bilateral symmetry also applies here. Two 16-point medallions dominate a yellow field, further filled with octagons and other forms. We are reminded of the horror vacui which progressively packed the fields of 19th century Caucasian rugs. Bird forms occupy their appropriate niches, and a variation of the latch-hook (a "Y" form) also appears. The prominent border of serrate leaves resembles those found on late Western Anatolian products - but, as we shall see, not exclusively.
Two other groups of pile weavings featuring similar medallions are known. Single-warp knot Spanish carpets attributed to the 15th century (but with more than 16 points) are to be found in several museum collections.4 And, as if to balance these early Western rugs, a new group from Kazakhistan were published by M. S. Mukanov in Oriental Carpets and Textiles Studies (0.C.T.S.) I, p. 161, fig. 7. Neither group displays bird forms nor latch-hooks and so should be regarded as derivative.
The largest group of rugs bearing 16-point medallions are to be found in the Caucasus. And here they are found in kilim, soumac, and pile structures.
During the International Conference on Oriental Carpets V in Vienna, Messrs. Robert Chinciner and Emil Salmanov presented photographically an Avar kilim, unfinished upon the loom. Woven in bright chemical colors, with a simplified design, we see this form perpetuated (see 0.C.T.S. III, Part I, p.72).
Perhaps the most interesting rugs in this category are the soumacs of Kuba/Daghestan. Still well represented from the 19th century these non-commercial and infinitely varied weavings are treasured trophies. Almost without exception the 16-point medallion is found enclosed within an octagon. Only a squiggle remains of the bird forms, but latch-hooks abound throughout the weaving. Of particular interest is the border of a soumac published by J. V. McMullan Islamic Carpets, plate 51. It is identical with the serrated leaf pattern of the Budapest fragment, and brings the provenance of that rug into question. Many examples of the 16-point medallion will be found in soumac literature.5
It is more through instinct than logic that Crivelli's rug is considered to be pile knotted. Both Erdmann (Der orientalische Knüpfteppich, Tubingen 1955, p. 17) and Mills (Hali, 1/3, p. 239) express a preference for Western Anatolia as the source of these "animal" rugs. But in 1957 Erdmann published the report of Pierre Belon, traveling in 1555, to the effect that no pile rugs were brought back from Turkey west of Konya.6 Konya was the eastern outpost of the Seljuk empire.
Ulrich Schürmann (Oriental Carpets, London, 1979, p. 77) published a pile rug labeled "Konya" but perhaps more easterly. Latch-hooks adorn the several medallions in a fairly regular manner, but no vestige of birds remains. Of interest here is the anomalous border treatment. After completing one third of the rug, the weaver abandoned the border design in a most unsubtle way. The vine, sporting eagle-head attachments, yields to a series of stars, alternating with stepped polygons. The Inner border changes Similarly. The terminating (horizontal) border shows further modifications. The analogy to Crivelli's depiction of 1486 is inevitable. Aberrant or anomalous border treatment of this kind is not uncommon among Caucasian rugs. But few examples are published since such rugs are regarded as flawed.
Daghestan is the source of a large accumulation of Crivelli- related motifs in the 19th century. Illustration 4 shows a "Crivelli" rug from Tabasaran.7 In common with the soumacs, the bird forms are reduced to (or alternately originated from) squiggles. The fissioning lozenges from Crivelli have fully separated and stand apart as eight-pointed stars on the field. The latch-hooks, forming a radiant halo about the medallions, have become a more assertive aspect of the field design. Adventitious elements fill the remainder of the field, familiar to fourth quarter 19th century Caucasian rugs. A border of intertwined eagle-headed figures, resembling the lower border of Schürmann's example, form a quilloche, an interlace or strapwork pattern, about the field.
We have ample evidence that Daghestan weavers of the 19th century were familiar with the motifs and tendencies of the Crivelli weaver.
To the question posed earlier, did Crivelli own one rug with variant ends, or did he have access to similar rugs, we must conclude with Ettinghausen10 the former. These two halves match each other more closely than any other half discussed. Their costliness would preclude duplication, and border variations are frequent occurrences. We are led to infer that a novice weaver has strung a small loom to produce a modest rug. As her work progressed, the weaver gained in confidence, as shown first by the elaboration of the side borders, later by her burgeoning interest in latch-hooks. Thus, in 1482, Crivelli displayed the initiating end of his rug; in 1486, the terminating end of the identical rug.ll
We are indebted to ibn Khaldun for the significance of the 16-point medallion in his time.12 In his discussion of geomancy "sand writing" he says,
"The sand diviners have given different names to the different combinations and classified them as lucky or unlucky, as is done with the stars. For (the sixteen combinations), they have assumed (the existence of) sixteen "houses." They think that the "houses" are natural and that they correspond to the twelve signs of the zodiac and the four cardines.
Thus the 16-point medallion is regarded as an auspicious symbol. Soon after the discovery of the properties of a magnetic needle, it was employed in the mariners compass. Blending good omens with the four winds, and adding the zodiac for the knowledge of travel by night, the 16-point compass card (now 32 points) became and remains the navigator's mate.
We are further indebted to Ibn Khaldun for preserving a document from Abbasid times, and a map from his own. The document lists the revenues in coin and goods rendered to the treasury at Baghdad from the contributary regions about A.D. 786 (op. cit., Vol. I, pp.361-365). Of the 34 regions listed, three gave partly in rugs as follows:
a) Armenia - Embroidered carpets: 20
b) Ifriqiyah (Tunisia) - Carpets: 120
c) Tabaristan - Tabaristan carpets: 600 pieces.
His map (p. 110) locates Tabaristan north of Armenia and
west of the Caspian Sea. The Arabs frequently incorporated
Christian (Caucasian) Albania within the designation
"Armenia" (see Encyclopedia of Islam, Vol. I, p. 443). The
region south of the Caspian Sea is better known for its
production of silk than for wool. Some 200 years ago a
colony of Kurds were settled at Kalardasht. These continue
to produce carpets on a small scale. No other rugs from this
region are known.
In the 14th century both Genoa and Venice maintained caravansaries in the port of Trebizond. Here valuable merchandise of many kinds, including textiles, were stored. This city, older than Rome, that had offered salvation to Xenophon's beleaguered army, was the eastern outpost of the West. Armenian merchants in the rug trade could have conveyed them to Venice via this port until A.D. 1461. Forin that year this last "Greek Empire" finally succumbed to the armies of Islam. If Crivelli's rug arrived by this route, it was old when he first painted it. Perhaps he acquired his first rug, second hand, in the Ghetto. But to say more is to flirt with Folly.
