The publication last year of Friedrich Spuhler's catalog of the rugs in the Museum fur Islamische Kunst in Berlin-Dahlem resulted in a number of fairly favorable reviews, and deservedly so.1 The West Berlin rug collection is of sufficient importance as to merit broad and ongoing discussion. The appearance earlier this year of an English language edition, Oriental Rugs in the Museum of Islamic Art (Washington, D.C., Smithsonian Institution Press, 1987), provides an appropriate opportunity to take another look at this milestone catalog. Beyond doubt it is a splendid and essential edifice of rug scholarship, but it is manifestly an edifice with a number of cracks in it. As a tool for further studies it will prove to be indispensable, but it will have to be used with a measure of circumspection.
The English language edition, however, has incorporated several minor improvements which at first suggest that some revision has occurred from the German edition. Rug dimensions which do not occur in the German edition have been added to the plate captions. The paragraphing of the catalog entries has not been indented in the German edition; its appearance in the English version is a small detail which makes the work far easier to use as a reference or research tool. Equally helpful is the substitution of an extensive index of names and rug types for the rather limited introductory glossary of names and technical terms in the German work. That glossary was, perhaps, part of the "focus" problem mentioned later.
Many of the great public museums of Europe were either founded or reorganized during the last century under the influence of the officially supported theory that public collections of great historical or exotic art would improve the aesthetic sensibilities and craft practices of the populace at large; hence, for example, the involvement of William Morris in the acquisition of the so-called Ardebil Carpet by the Victoria and Albert (or South Kensington, as it was then called) Museum. The West Berlin rug collection reflects partly the prevalence of this widely believed conviction and partly the collecting genius of the three early pioneers of rug studies, Julius Lessing, Wilhelm von Bode, and Friedrich Sarre.
This catalog suffers from a certain vagueness of focus. At whom it is directed? Serious students of rug studies, casual lovers of art, or as many people as possible who can be convinced to buy it? Apparently, and doubtless from both the author's and the museum's viewpoint, a scholarly record of one of the world's major rug collections was intended. But why was the German edition fitted out with a glossary of terms like "abrash," "boteh," "field," "gul," and "runner," such as any tiro could pick up in the first popular handbook he stumbled across? (The index, as mentioned, is an improvement in the Washington edition.) Why are there no maps? All rug books should have maps, given the exotic and often obscure places where rugs have been, and continue to be, produced. If, on the other hand, the work is meant to be popular, why is space "wasted" on the bygone losses to the collection, on a concordance of inventory numbers (these are, after all, listed at the close of each entry), or on specialized bibliographies of bibliographies of Islamic art and rug literature? Admirable tools for the specialist, they occupy valuable space which might have been devoted to more basic information for the beginner. There is a curious compression and apparent dogmatism in parts of the text, which seems to have nothing to do with the process of translating from German to English, that hints at pressure to save space. One might suspect that the financial stringencies of modern publishing may have influenced even the venerable art publishing house of Klinkhardt and Biermann, which has published many rug books over the years, to call for compromises in the content of this catalog, aimed more at promoting larger sales than at enhancing the work's academic integrity.
Dr. Spuhler is so preoccupied with the formal, surface organization of pattern on rugs, evidently a specialty of his, that he occasionally neglects, so to speak, to tell about the pattern itself. Consider for example his entry for one of the finest of the postwar Persian acquisitions, a 16th century medallion rug with animal combats (no.78): "The vertical and horizontal central axes divide the field design into four theoretically identical quarters, but in fact the lower half of the field, which was woven first, is more than 5 cm. shorter than the upper. The effect of this seemingly insignificant difference is noticeable in the squashed appearance of the lower half of the medallion and the extended distance between the shield shaped pendant and the palmette blossom above. This irregularity results in the displacement of a little more than 13 cm. within the horizontal central axis between the field and the border..."
From a theoretical viewpoint, this is interesting but overlooks the fact such minor discrepancies could hardly have been noticed by the actual users of a rug four and a quarter meters long. Consider the "palmette blossom," mentioned in the entry as attaching to the upper and lower central edges of the, lamentably, "squashed" medallion. Describing a rather similar rug, though it lacks the animals of the Berlin piece, in his catalog to the Keir Collection, (no. 41)2, Dr. Spuhler writes as follows:
"...Between the medallion and the shield-pendent (in Berlin, it is a pendant), somewhat unusual floral medallions have been inserted in the place of cartouches. They stress the longitudinal aspect of the carpet....
To judge by the relative illustrations involved, these "floral medallions" seem to perform a similar function, under the guise of "palmette blossoms" on the Berlin rug. The quibble here is not with the change of terminology in the intervening decade, but rather with the thought that what was unusual or functional in a pattern element on one rug ceases to be so on another, similar rug barely 10 years later. Surely this is the kind of information that most readers would find interesting and, possibly, even useful.
In some respects, this review is a figurative tale of two cities, Berlin and Washington. Dr. Spuhler himself expressly refers to the succession of directors and writers connected with the Museum fur Islamische Kunst, from Wilhelm von Bode on, as the Berlin school of rug studies.4 However, given the influence that those two eminent rug scholars and theorists, Charles Grant Ellis and May Beattie, have had upon each other and on a younger generation of American rug scholars, and given their varying connections with The Textile Museum over the years, it seems fair to postulate a Washington school of rug studies, based loosely on The Textile Museum. Ellis currently proposes that many of the "Holbein" rugs, possibly the group with kufesque interlace borders, were made in or around Konya, in central Turkey, rather than in the traditionally attributed center Ushak.4 Dr. Spuhler cleaves to the traditional attribution Ushak for the "Small Patterned Holbeins" (nos. 1-3) and the equally traditional location Bergama for his "Wheel" or "Large Patterned Holbeins" (nos. 4-6). Ellis attributes compartment rugs to Damascus and notes that the main body of the group may be divided into hexagonal or octagonal units.5 Dr. Spuhler locates his chess board rugs (nos. 71-73) in Egypt or Turkey, which neatly gets around Ellis's Damascus, and insists that the main body of the group is always divided into hexagonal units. This is unfortunate because the two Berlin "main body" examples (nos. 71 and 72) are in fact divided into octagons!
The West Berlin collection owns three rather atypical, red grounded medallion rugs which look somewhat like late, degenerate Ushaks (nos. 23-25). Dr. Spuhler illustrated two of them (nos. 23 and 24) among others in a fairly recent article, "Uncomfortable Questions about Unknown Turkish Carpets in the Berlin Collection,"6 in which he deplored the restricted and unimaginative attribution of most early Turkish rugs to Ushak. In the present catalog, he cites May Beattie's article, "Some Rugs of the Konya Region,"7 in note 4 of his entry for the Turkish medallion rug (no.23) as a source of further illustrative material but without commenting on her closely reasoned and documented thesis that the cluster of rugs, of which numbers 23, 24, and 25 seem to be a part, were produced in or around Karapinar, a town of the Konya region which was celebrated for rug production in a number of medieval documents. He locates their source in western Turkey. Ellis adopts her view, and even illustrates the same rug as her fig. 8, during his discussion of a kilim-style rug (no. 30) in his recent catalog, Oriental Carpets in the Philadelphia Museum of Art, though he does not cite her either, in his case probably because he and Beattie have discussed and exchanged so much information over the years that they probably regard each other's work as a mutual fund of knowledge.
The list of examples where the latest protagonist of the Berlin school differs roundly from the views of members of the Washington school could be multiplied even from further rugs in this catalog. The multi-niche prayer fragment (no.81) is a case in point: Dr. Spuhler maintains the traditional 16th century Persian identification, whereas Ellis and Louise Mackie are quite certain that it is a 17th century Mughal Indian piece.8 One might almost think that World War II had broken out all over again!
Dr. Spuhler begins his introduction to the section on Spanish rugs with a questionable polarity: "Spanish pile-woven carpets influenced by Oriental designs can be clearly distinguished from Oriental carpets by their distinctive structure."12 The structural difference is well known; the non-Orientalness of the genre seems to have been generally overlooked, except by scholars of the Berlin school. He continues this distinction between Spanish, and by implication European, rugs and Oriental rugs throughout the chapter. He contrasts the unique structure of the so-called Spanish, single warp knots, again by implication European, with the normal Oriental symmetrical and asymmetrical knots, tied around two adjacent warps. He opposes the non-standard border systems of early Spanish rugs to the presumed canonical system of secondary and main stripes seen on "real" Oriental rugs, though here he makes heavy going while explaining the presence of kufesque borders, albeit degenerate kufesque, on rugs with non-Oriental field layouts. He focuses on what he interprets as deviation in the pattern organization of early Spanish rugs from the accepted norms of "classical carpets of the Orient." Simply stated, rugs such as the star and hexagonal lattice fragment (no.137) already referred to, which display large, monumental borders combined with small textile repeats in the field, are intrinsically non-Oriental, at least before the broad dissemination of such design principles during the 19th century.
Strangely, his absorption with surface organization and layout distracts him from making an issue of the native Hispanic, folkloristic humans and beasts that comprise many of the secondary pattern elements on early Spanish rugs. His relative silence on these non-Oriental details may result from the limited scope of the West Berlin Spanish rugs, which lack Western shields of arms, and have lost end panels, such as the detached fragments from the appendix (no.203), on which heraldic bears and boars frequently keep company with lions or engage in combat with the shaggy wild men so popular in European late Gothic art.
In maintaining that early Spanish rugs are non-Oriental in character, Spuhler is paying homage to his predecessors and mentors in the Berlin school. Ernst Kühnel dropped "Hispano-Moresque" rugs from the fourth edition of Bode's Antique Rugs from the Near East,13 even though he went on in 1957 to co-author the first English language monograph on the subject.14 Kurt Erdmann treated them cursorily as "peripheral problems" in his most influential rug books.15 This adherence to the ideas of his Berlin precursors seems to have closed his mind to the recorded connections, discussed in the literature, between the Western caliphate of Cordoba and early Islamic Egypt, where excavated textile fragments exhibit various forms of cut loop technique, some clearly resulting in pile with primitive, single warp knots. It is not necessary to posit a direct link between the few ancient textile fragments with single warp knots found early in this century in Chinese Turkestan and Spain. Erdmann himself devises an acceptable theory to explain the origin of the characteristic single warp knot from Islamic Egyptian sources.17 The hypothesis that the early Spanish border systems do not adhere to classical rug paradigms (he seems to be thinking specifically of Persian rugs) can be refuted by impeccably Oriental examples from this catalogue. The celebrated "Holbein" rug with a row of four large octagon-within-square medallions (no.4) has three borders, each progressively wider towards the perimeter of the rug. The widest outer border contains a complex, open kufesque interlace precisely where kufesque elements normally appear on early Spanish rugs. The vase fragment (no.68) has a single wide border with no inner guard stripe. To judge from a number of other complete examples, there would have been no outer guard either. Finally, about the late dragon rug (no. 103) the author has to say, "the simple vine border without guard stripes is typical of the group."18 His objection to the non- Orientalness of large scale borders combined with small scale repeats in the fields of early Oriental rugs is invalidated by the existence of a whole group of early Seljuk Turkish rugs and fragments.19 Erdmann himself, in alluding to the "Spanish" fragments found at Fostat, in Egypt, comments: "the form world of these first Spanish rugs was similar to that of the contemporary Seljuks of Anatolia."20 A fresh look at number 137 might bear out this formal relationship.
Writers outside of the Berlin school have tended to take a less dogmatic view about the sources of early Spanish rug design, recognizing that there was a community of Islamic cultural influences across Western North Africa, the so-called Maghreb, which included Moorish Spain. The highly cultured Ommayid caliphate soon fell to the Almoravids and Almohads in turn, both Berber dynasties which originated in North Africa and ruled till the first quarter of the 13th century. Louise Mackie especially, who might be considered a member of the "Washington school," has suggested that an amalgam of Islamic and local Hispanic elements, including the survival, both in North Africa and Spain, of Roman mosaic pavements with rug-like geometrical patterns, was operative in the formation of the designs which appear on early Spanish rugs.21 There is a growing body of documentation to suggest that some rugs were woven in North Africa throughout the mediaeval period.22 Charles Grant Ellis, Jon Thompson, and Jenny Housego have all theorized about atypical Mamluk rugs which may have been produced in the Maghreb.23 There is the possibility that some pattern influences may have permeated into Andalusia from North Africa during the centuries of political unity. The border fragments from an "admiral" type rug of the 15th century (no. 138)24 are very similar to those on three early 15th century rugs, now dispersed between The Textile Museum, the Hispanic Society of America, and the Detroit Institute of Arts which display the conjoined coats of arms of Maria of Castille and Alfonso V, the King of Aragon. The central border of each shows a diagonal grid of lozenges strikingly reminiscent both of the pile and flatwoven lozenge repeats on Jaffi Kurdish bags. Dr. Spuhler describes this grid as "a 'textile' design, here bearing a coincidental resemblance to a 19th century Turkish flat weave design." On both sides of this central border are narrower stripes, each with repeated, interlocking horizontal zigzag lines rather like the letter "W." Very similar design elements occur in the 19th century Moroccan hanbels (nos. 207 and 208) which appear on page 320 of this catalog. The "W" configuration of the narrow stripes even appears on a 20th century, flatwoven Zemmour mat in the 1980 Textile Museum catalog, From the Far West: Carpets and Textiles from Morocco.25
An early 19th century Moroccan silk embroidery comprising a 24-point star, nearly two feet in circumference and mounted on a plain panel, was offered for sale at Sotheby's in 1984.26 This star is remarkably similar to the "Crivelli" star medallions which appear on certain 15th century Alcaraz rugs. The rays of this star are broken up into a patchwork like those of the Alcaraz medallions, but the filler patterns are made up of geometrical grids, lozenges and If "W" configurations which almost seem to form a sampler of many of the border designs of early Spanish rugs. It is possible to argue that such similarities, especially when the vast intervals of time are taken into account, are coincidental and technique-conditioned universal patterns, or that both groups of designs go back to common Ottoman Turkish prototypes. However, if further historical evidence is uncovered of rug production in mediaeval North Africa, it may become less incredible to postulate the interaction of rug designs between Moorish North Africa and mediaeval Spain.
Generally, Robert Pinner's translation seems accurate and reads easily. He deserves our gratitude for making this catalog available in English, and the merit due his accomplishment should far outweigh the several specific criticisms which I feel compelled to raise about the English text.
Occasionally too, problems with the English have more to do with the original German text than with his translation. In the section on Persian rugs, especially the large classical pieces, such as the Northwest Persian fragment (no.77), the Herat vine-scroll piece (no.79), and the recently reevaluated, late Indian vine-scroll rug (no.80), Pinner repeatedly uses the phrase "spiral tendril" to designate the curving stems which delineate the principal pattern units of these rugs, and the word "tendril" for stems in general. He is faced with the task of translating the German word Ranke which in its literal meaning describes what English understands as a "tendril," but which through art oriented phrases such as Rankenwerk and Spiralranken has come in German to mean arabesques or decorative strapwork in general. If the word "tendril" has any meaning in English beyond the literal botanical one, that meaning combines features of curling or coiling like a corkscrew with the fineness of a thread, of a filament or a strand of hair. Tendril in English is to a vine or stem as fingers and toes in anatomy are to arms and legs, literally so when one recollects that in the garden real tendrils perform a clinging or grasping function. It is probably no accident that earlier English language rug books have used such phrases as vine scrolls and vine coils (Ellis), spiral vine systems (Ellis translating Erdmann), spiralling scrolls (Wingfield translating Spuhler), and scrolling arabesques (A. U. Pope).27 It is significant that "spiral tendril" occurs only in books translated from German containing the word Ranke, and confusing that tendril is a correct translation - sometimes! Further there is no avoiding the fact translators usually work from the text and plates of a book, and that this may reduce, in their mind's eye, the scale of these often very large rugs. Nor can it be gainsaid that tendril is a pleasanter sounding word, with richer associations, than stem or stalk. In this catalog, nevertheless, the stereotyped translation of Ranke by tendril has led to contradictions in meaning. The primary arabesques on the border fragments of a ruined vase rug (nos. 88-89) are termed tendrils even though they merely arc gently, without coiling; whereas the second, thread-like floral stems are described as "gracefully spiraling vines." This totally reverses the scale of relations inherent in the meanings of these words. This careless use of the word tendril leads finally to such logical absurdities appearing in the entry for a Caucasian blossom rug with the Afshan pattern (no. 110) as "an almost straight tendril" or later as "a stepped tendril pattern." One might hope that, except for genuinely small scale design elements such as the use by Ellis of "tendrils" to describe the background foliate diapers of the Ushak medallion rugs,28 the word would disappear from rug books.
In the entries on the 17th century "Polonaise" rugs, attributed to Isfahan (nos. 84 and 85), the translator describes the main border patterns as "reciprocal pinnacles." In discussing the lower or beginning border of the Caucasian garden rug, dated ca. 1800 (no. 107), he uses the words "reciprocal turret design." These are both correct translations of the German, reziproke Zinne, used in its various grammatical permutations, but they are distracting to English speakers, knowledgeable about rugs, who use "reciprocal trefoil" and to other English speakers who are doubtless perceiving "alternating fleurs de lys." Both phrases are visually more correct in describing such a border in its classical form. I suspect that the author who, as a specialist in Islamic art, is aware that similarly formed decorative crenellations crowned the tops of garden walls in l7th and l8th century Persia, when these forms were consciously used around the perimeters of rugs to provide the user of the rug with symbolic protection, and that he was probably influenced by this awareness in his usage of the word Zinne. In the context (i.e. a row or series of them), however, the German word normally appears in the plural Zinnen, the equivalent of the English "battlements" or "crenellations." Both they and Zinnen imply alternation, making the "reciprocal" redundant except that the latter word is embedded in the usual English descriptions, and these are familiar to most rug students whatever their languages. Hence I surmise "reciprocal pinnacles." But in keeping with the somewhat compressed and dogmatic flavour which at times appears in the text, none of this background is explained.
Another linguistic eccentricity is the translator's choice of "cross bar" to describe the transverse or horizontal bands or panels which appear above the mihrab (and sometimes below as well) in the Ghiordes prayer rugs (nos. 37 and 39) and the end panels which occur in the fields of two of the West Berlin Mamluks (nos. 61 and 64). "Cross bar" is, once more, a correct translation of the German word Querriegel, but surely not the appropriate one in the context of the flat surfaces of rug patterns. To the best of my knowledge, the term appears in no other German rug book29; whereas German authors have frequently chosen the foreign derived Paneel to describe such cross panels, particularly as they appear in Turkish prayer rugs. Querriegel may be suitable in the German text since one of the secondary meanings of the word, the transverse beam, especially over a door in a half timbered construction,30 implies the appearance of flat, surface banding.
In English, however, a bar always implies a three dimensional, elongated object. Even a chocolate bar (the word Riegel is used this way in German) has, or is presumed by the buyer to have, some thickness to it! Therefore "cross bar" seems to be out of place when speaking of rug patterns. Strangely enough, Mr. Pinner was willing to use "horizontal band" to translate Querriegel in the description of the "Transylvanian" prayer rug (no.33). Dr. Spuhler himself used the phrase ein querliegendes Band to describe this feature for the Ghiordes prayer rug (no.37) and the word Querband for this detail, repeated at both ends of the "double-niche" Ghiordes rug (no.38); at this point, however, Mr. Pinner chose to translate Band, which is flat both in English and German, as "bar."
The recumbent, leonine beast within a Renaissance garland of the 16th century Alcaraz fragment (no. 141)31 is portrayed as follows: "Its collar reminds us of a lion in repose; its grinning facial expression suggests a grimace" The beast has an unruly, curly ruff around its neck, which is what the original German Halskrause means. Since the lions on Spanish rugs are often heraldic, and heraldic lions often wear literal collars for a specific meaning, surely "mane" or "shaggy mane" would have suited the context better. The German really seems to be saying: with its "mane," it has the appearance of a lion in repose.
Most users of this catalog will want the essential information without worrying about the niceties of this or that phrasing. Aside from the particular objections raised above, a personal and random sample, it seems that Mr. Pinner has created an admirably competent translation.
