iv. Knotted-Pile Carpets, Designs, Motifs, and Patterns
In this discussion “design” refers to the overall composition of decorative elements on a carpet; the simplest elements in designs are single motifs, which are most frequently combined in more complex units; these units in turn may be arranged in various combinations and sequences to form patterns. The terminology for carpet designs, like that for technique (see iii, above; v, below), is quite controversial, and several sets of nomenclature are currently in use. In academic studies rugs are traditionally classified on the basis of motifs and patterns, with names usually derived from the broader repertory of Islamic art, for example “arabesque” (see below). On the other hand, in trade publications carpets are commonly identified according to a place or period of manufacture, and a variety of terms, generally colloquial, are used to label designs. In the absence of a standard vocabulary, the terminology adopted here is that most widely used in scholarly publications (for an extensive list of colloquial terms, see Varzī, pp. 289-99).
Designs. The most common design format in Persian carpets is a central field (matn; Figure 68a) enclosed by a border (ḥāšīa; Figure 68b-e) of patterned stripes alternating with narrow bands of solid colors. The borders of Persian carpets generally have three or more patterned bands; the widest is called the “main stripe” (Figure 68b) and those flanking it “minor” stripes (Figure 68c). The narrow bands of solid color that usually separate the main and minor stripes are known as “guards” (Figure 68e). Carpet formats may be categorized further as either directional or nondirectional. The former include pictorial, prayer-niche (meḥrāb), and naturalistic vegetal designs intended to be viewed from a single vantage point. Nondirectional designs, on the other hand, are coherent when viewed from any angle; examples include radial designs, as well as abstract repeat patterns associated with tribal weaving (see Āzādī, 1978, no. 40; Edwards, pls. 156-57; Gans-Ruedin, pp. 474, 477, 481).
Some Persian carpet designs may be either directional or nondirectional, depending upon the nature and arrangement of their specific components. The following are particularly common types:
1. Centralized designs are balanced longitudinally and transversely around a central element, usually a medallion (toranj; Figure 68a). In medallion-and-corner (lačak toranj) designs there are a medallion in the center of the field and a quarter-medallion in each corner (e.g., Āzādī, 1978, no. 39; Gans-Ruedin, p. 452). Arguably the best-known examples of Persian carpets with centralized designs are the so-called Ardabīl carpets (Plate CIX in ix, below; see ardabīl carpet).
2. In lattice designs linear motifs intersect to form an overall grid, or lattice. Two, three, or more lattice systems may be superimposed so that their components cross at regular intervals. Although this type of design is most familiar from certain carpets in the so-called “vase technique” attributed to the Safavid period (Beattie, 1976, pp. 40-58, nos. 33-37; see ix, below), a variety of lattice designs are still woven today (e.g., Āzādī, 1978, no. 37; Edwards, pl. 218).
3. In multiple-medallion designs the field is dominated by rows of medallions, either linked or overlapping; they may be arranged either directionally or nondirectionally (e.g., Beattie, 1976, p. 43, no. 7).
Motifs. The following motifs are common in both the fields and borders of Persian carpets.
1. The medallion (toranj) is an ornament, usually curvilinear, that occurs frequently as the central motif(s) in the field design. The most common variants are ogival, circular, shield-shaped, and rhomboid. The edges are generally lobed or stepped.
2. The cartouche is an elongated ogival compartment (Figure 69). When used in a border pattern (as, e.g., in the Ardabīl carpet, see Plate CIX), a single cartouche form may be repeated, or two variant forms may alternate. Inscriptions are generally framed in cartouches.
3. The arabesque (šāḵa-yeeslīmī, lit. “patterned” or “arabesque branch”; Figure 70; see also Survey of Persian Art, figs. 755, 760, 762, 785, 787) is a stylized vegetal scroll with bifurcated leaves. In border patterns arabesques can occur in continuous scrolls, contiguous pairs, or alternately reversed contiguous pairs.
4. The bota (lit. “bush, shrub”; Figure 71a-b) resembles a leaf with a curved tip; it is familiar in the West as the “pear” or “teardrop” motif of paisley shawls (Edwards, pls. 21-22; Varzī, pp. 263-64).
5. The palmette (gol-e eslīmī, lit. “patterned” or “arabesque flower”; Figure 72); see also Survey of Persian Art, figs. 774-80) is a stylized fan-shaped motif somewhat resembling a lotus flower in section. In the most frequent type of palmette border the motif faces alternately inward and outward and is flanked by arabesques.
6. The rosette resembles a fully opened rose seen from above (Survey of Persian Art, fig. 783a-d).
7. The sickle leaf is a long, curved leaf with a serrated concave edge (Survey of Persian Art, fig. 784a-d).
8. The cloud band is an extended ribbon-like motif ultimately derived from Chinese art (Survey of Persian Art, figs. 789a-o, 791a-n).
Patterns. Among the most frequent patterns on Persian carpets are the following.
1 . Herātī (lit. “from Herāt”; Figure 73) is named for the city in which it was formerly thought to have been developed (see ix, below). The defining unit consists of a rosette framed in a rhombus with palmettes at the corners. Four sickle leaves are arranged around the whole. In Persian the pattern is also known as māhī “fish” from the resemblance of the sickle leaves to small fish. The pattern is woven throughout Persia, with local variations (Edwards, pls. 17-19).
2. Mīnāḵānī (literal meaning unknown, possibly derived from a proper name) is based on a two-plane lattice system of scrolling stems, with rosettes at the points of contact. This field pattern is particularly associated with carpets from the Varāmīn area (Edwards, pp. 42-43 and pl. 25).
3. The reciprocal-trefoil (madāḵel lit. “doorways”) pattern, common on borders, consists of two interlocking systems of stylized trefoils (Dimand and Mailey, p. 63, fig. 89; Gans-Ruedin, p. 534).
4. The S-stem (šāḵa-ye mawwāj lit. “wavy branch,” tāk lit. “vine”; Figure 74) pattern, consisting of repeated short stems in an S configuration, is usually a border pattern. The curving stem generally terminates in small vegetal motifs, with larger forms like palmettes or rosettes centered on the diagonals. S stems may be arranged in either single-, double-, or triple-plane systems. The most common single-plane systems are (i) continuous, in which S-stems are sequentially placed (Figure 74b); (ii) contiguous continuous, a sequential arrangement of linked S-stems (Figure 74c); (iii) counterpoised, with pairs of confronting S-stems (Figure 74d); (iv) continuous counterpoised, a sequential arrangement of pairs of confronting S-stems (Figure 74e; for terminology, see Beattie, 1976, pp. 20, 32). In the most common double-plane system contiguous counterpoised S-stems are interlaced with a confronting system. The so-called “crab” (ḵaṛčang) border consists of a double-plane system of interlacing continuous S-stems with stylized flame palmettes, or “crabs,” at their points of contact, and leaves, or “claws,” on their terminals (Edwards, p. 49 and pl. 33). The S-leaf is a variant of the S-stem, in which a leaf covers the entire diagonal of the stem (Varzī, fig. 14).
Šāh-ʿAbbāsī (lit. “of Shah ʿAbbās”), a repeat pattern of various types of blossoms, cloud bands, and palmettes associated with Safavid carpets, e.g., the so-called “Polonaise carpets” (Plate CXI in ix, below; Āzādī, 1978, p. 19; Edwards, p. 43). More recently this pattern has been associated with workshop production in such urban centers as Tabrīz, Kermān, Nāʾīn, and Isfahan (Edwards, pp. 36, 336, and pls. IV, 213; Parhām, pls. 6, 7a-s).
Certain patterns and designs, as well as distinctive renditions of common designs and motifs, have customarily been associated with specific weaving centers: for instance, the “brick” (ḵešt)-patterned rugs of Čahār Mahāl-e Baḵtīārī (Housego, 1978, pl. 89; Edwards, pl. 364; baḵtǰarǰ iii. carpets, p. 562) and the angular medallion-and-corner carpets of Herīs (Edwards, pls. 42, 68-69). Today, especially in weavings for export, such distinctions have been blurred, as pile-carpet manufacturers in Iran, Turkey, Egypt, India, China, eastern Europe, and other areas compete to satisfy the market for particular designs and palettes. For example, the label “Indo-Kirman” is applied to carpets woven in India in a floral design derived from a provincial Persian prototype (see xi, below).
Cartoons (naqša) and samplers (vā-gīra). As noted in iii, above, many carpets are woven without specific models, especially those with traditional designs that have not been produced in manufactories. The weaver may also use a cartoon in which the design (ṭarḥ) has been drawn on graph paper, usually to full scale, and colored; each square (ḵāna lit. “house”) represents one knot. The spacing of the squares serves as a guide to knotting density. Before printed graph paper was available designs were drawn, then painted, and finally divided by superimposed grid lines. Cartoons are prepared by specialized designers (naqqāš; Edwards, p. 206; Parhām, n.p.). Once completed, a cartoon is cut into horizontal sections that can be conveniently placed on the loom in front of the weaver. The knotting sequence on the cartoon may also be called out to the weaver either by a professional design caller (naqšaḵᵛān) or the master (ostād) of the workshop.
Alternatively, the weaver may use a knotted sampler (vā-gīra) in which a section or sections of the field and border designs have been woven according to a cartoon. Although initial preparation of the vā-gīra requires more time and labor than does that of a paper pattern, it provides a more effective representation of a particular choice of colors and design before weaving actually begins, so that costly and time-consuming errors can be avoided. It is also more durable than a paper cartoon. The vā-gīra is noted in descriptions of late 13th/19th- and early 14th/20th-century carpet production, and it has been suggested that it was introduced by Western exporters at that time (Edwards, p. 125). It is quite possible, however, that such carpet samplers were in use earlier.
Today the patterns woven in urban carpet workshops in Persia are usually copied directly from cartoons. The cartoon also plays a vital role in cottage carpet weaving, as it allows an entrepreneur to place orders for a specific pattern with any number of individual loom owners. As a single loom owner may be simultaneously under contract to more than one manufacturing firm, however, it is virtually impossible to copyright designs at the cottage level, and many incidents of design piracy have been recorded in correspondence from foreign manufacturing firms in Iran to their respective consuls (e.g., U.K. Foreign Office 248.1072, 11/21/13).
Designers (naqqāš). The importance of the designer in the hierarchy of artisans involved in carpet production is reflected in the occasional inclusion of designers’ names in custom-woven carpets (Ittig, 1985, no. 5); generally, carpet inscriptions mention only patrons and/or loom owners (ostād), rather than actual craftsmen. Design training has traditionally been conducted according to the apprenticeship system (Parhām, n.p.; Edwards, p. 207); it was not unusual for the skill to be passed down from father to son (Parhām, n.p.; Edwards, p. 207). A designer might work either for a specific firm or on commission, supplying designs to more than one manufacturer (Edwards, pp. 308, 337). During the late 13th/19th and early 14th/20th centuries, a “boom” period for Persian carpets, competition for designers was keen; for example, in Kermān one Āqā Moḥammad, a designer under contract to the Eastern Rug and Trading Company, was lured by higher wages into the employ of the German firm Persische Teppichgesellschaft AG (PETAG; U.K Foreign Office 248.1072, 9/8/13).
Design and provenience. No extant example of a knotted-pile carpet that can be firmly attributed to Persian manufacture before the 10th/16th century is known (see vi-viii, below). Early geographers, historians, and travelers provided only cursory descriptions of carpet designs and manufacture (for an annotated compilation of historical references to trade and production of textiles, including carpets, into the 8th/14th century, see Serjeant). Perhaps the most famous pre-Safavid carpets are the legendary Bahārestān carpet (see bahār-e kesrā), or “Spring of Kosrow,” and the Pazyryk carpet (see vi, below). The former, if in fact it was a woven carpet, was probably of flat-woven, rather than pile, construction (Edwards, p. 2). Recent archeological evidence suggests that the Pazyryk carpet, on the other hand, may not be of Persian origin but a product of Central Asia or Siberia (Kawami, pp. 16-17 and nn. 64-66).
Although the chronology of Anatolian carpets in the 9th-12th/15th-18th centuries is based largely upon comparison with representations in European painting (Mills, pp. 10-33), similar methods are less useful in dating carpets attributed to Persian manufacture (see i, above). Persian carpets, usually of the “Herat” or “Indo-Persian” group, appear in the 11th/17th century, in the paintings of van Dyck, Rubens, and Vermeer, and other European artists, but for earlier periods scholars have traditionally relied for dating on stylistic comparisons with architectural ornament, bookbindings, manuscript illuminations, and depictions of floor coverings and textiles in miniature paintings (see vii, viii, below).
Among the earliest visual evidence for Persian carpets are the depictions of floor coverings in Timurid miniature paintings, where the predominant field designs are small repeat patterns of squares, hexagons, octagons, circles, and crosses; borders often contain repetitious “pseudo-Kufic” designs (Briggs, 1940; for a critique of Briggs’ method, see viii, below). Unfortunately, these paintings give no indication of technique and may represent flat weaves, pile carpets, or even, in some instances, felts. Furthermore, extant Anatolian, Spanish, and Mamluk rugs, as well as carpets depicted in contemporary European paintings, attest that an “octagon” style was in vogue internationally at that time (King, pp. 14-21, pls. 7-12, 14-16, 22-25; Mills, pp. 10-26, pls. 3, 5-7, 11-12). The floor covering that most closely resembles in design those in Timurid paintings is the zīlū, a sturdy cotton flat weave with a double-cloth structure (Wulff, Crafts, pp. 210-11; Beattie, 1981, p. 171, figs. 2, 4, 6-7; see v, below). The date on a fragmentary zīlū with a waqf inscription found in the congregational mosque at Maybod, near Yazd, has been interpreted as 808/1405 (Afšār, pls. 44-45); if that is correct, this fragment would be of Timurid or Turkman manufacture, the oldest carpet firmly attributable to Persian looms.
At present there are only five published pile carpets that can be assigned a Safavid provenience on the basis of inwoven inscriptions and dates (Ittig, 1984, no. 26); the two Ardabīl carpets of 942/1535-36 (see Plate CIX), the Milan hunting carpet (see Plate CVIII) of 929/1522-23 or 949/1542-43, a fragment with a multiple medallion field design in Sarajevo dated 1066/1655-56 (Survey of Persian Art, pl. 1238), and a shaped silk carpet with a directional design of cypresses and flowering plants in Qom dated 1082/1662 (Survey of Persian Art, pl. 1258B). In both their field designs and their predominantly curvilinear motifs, these pieces differ significantly from the repeat patterns of small geometric motifs that scholars have traditionally associated with Timurid carpets, though it is now recognized that “Safavid” features also occurred on floor coverings in Timurid paintings (see viii, below). Interestingly, most of the floor coverings depicted in miniature paintings ascribed to the Safavid period have centralized designs and curvilinear motifs, but none of those illustrated exactly correspond to any extant carpets considered to be of Safavid manufacture.
Carpets attributed to this period are characteristically filled with a variety of stylized floral and foliate motifs, many classifiable as chinoiserie, including lotus and leaf palmettes, split-leaf arabesques, scrolling vines, and blossoms, as well as real and fantastic animals, cloud bands, and cloud knots. Human and animal figures and tableaux from Persian literature also occur occasionally. Despite the wide range of decorative devices employed, these carpets are normally strongly symmetrical on the longitudinal axis and often transversely as well. The terms used to distinguish groups of Safavid carpets, for example, Polonaise, Sanguszko, and Herat, were devised during the early days of carpet studies. Today, a century later, they remain in use as labels of convenience, for no satisfactory classification based on place of manufacture has yet been developed (see ix, below). It is generally thought that the designs for Safavid carpets were produced by court craftsmen, who were also responsible for designs for the arts of the book and for painted architectural decoration. Yet such assumptions must be treated with caution, especially as the development of miniature painting itself is not yet entirely clear; and miniatures formerly thought to have been produced under royal patronage in court workshops have been reattributed to other settings (e.g., Simpson, p. 97). Furthermore, there are no extant carpets known to have been ordered specifically for the Safavid court, though there is contemporary textual evidence that carpets of various other kinds—nomadic, village, and urban, including both commercial weaving and weaving to order—were manufactured in various regions, as they are today (Keyvānī, p. 237; Chardin, VII, pp. 329-34; Du Mans, p. 187; Mańkowski, p. 2433; Thevenot, p. 77). Finally, the assertion that “the golden age of the (classical) class doubtlessly falls within the first half of the sixteenth century” (Bode and Kuhnel, p. 87) is unproven: at present, only three dated pieces (see above) support this statement. The traditional chronology and classifications of Safavid carpets therefore require careful reexamination. More recent studies have stressed structural as well as design features in attributing provenience (e.g., Beattie, 1976, passim); closer attention to primary sources may also provide clearer understanding of the historical context of carpet manufacture.
As few securely dated 12th/18th-century carpets are presently known, only the most general hypotheses about their designs are possible. A preliminary survey suggests that Safavid designs and patterns continued to be woven, albeit with less fluidity. New motifs, like the bota, were introduced later in the century (Housego, 1986, pp. 40-51, figs. 3, 9; see x, below).
The volume of objects and documentary sources relating to carpet production and trade in the 19th century permit a number of general observations on Qajar carpet designs. First, various designs and motifs known from the Safavid and succeeding periods continued in use; in addition, such small repeat patterns as the herātī, mīnāḵānī, and offset bota became established in the repertory, with distinctive regional variations. Furthermore, as part of a contemporary fashion for European imagery in the Persian arts generally (art in iran x. qajar art, p. 631), Western pictorial themes were also included in carpets. As the economic importance of the export market increased in the late 13th/19th and early 14th/20th century, widespread changes in traditional sizes, palettes, and designs were introduced in order to meet the requirements of Western consumers (see xi, below). Finally, it is only from this period that significant quantities of so-called “nomadic,” or “tribal,” carpets survive, with their (usually) bolder geometric patterning (e.g., Housego, 1978, passim), in marked contrast to the types of designs described here. Similarities between certain tribal ornaments, especially the Turkman göl (Pers. gol “flower”), and motifs in floor coverings in Timurid miniature paintings suggest that at least some nomadic carpet patterns represent survivals from the pre-Safavid repertory (Denny, pp. 331-34; Pinner and Franses, p. 88; see viii, below).
Since the Qajar period the designs of Persian carpets have continued to be modified to meet demands of the international market, for example, production of the so-called “American Kirmans” (Edwards, pp. 207-09; see xii, below). As traditional Persian motifs and patterns are increasingly copied in centers outside Iran, design must now be considered more indicative of the targeted market than of the place of origin.
Ī. Afšār, Yādgārhā-ye Yazd I, Tehran, 1348 Š./1969.
S. Āzādī, Fārš-e Īrān/Persian Carpets, Hamburg, 1978.
Idem, “Baluchistan v. Baluch Carpets, Rugs and Other Products,” in EIr. III/6, pp. 645-46.
M. H. Beattie, Carpets of Central Persia with Special Reference to Rugs of Kirman, Sheffield, 1976.
Idem, “A Note on Zilu,” in C. Cootner, ed., Flatwoven Textiles, Washington, 1981, pp. 168-73.
Idem, “Ardabīl Carpet,” in EIr. II/4, pp. 365-68.
W. von Bode and E. Kühnel, Antique Rugs from the Near East, 4th ed., tr. C. G. Ellis, Berlin, 1958.
A. Briggs, “Timurid Carpets,” Ars Islamica 7, 1940, pp. 20-54; 11-12, 1946, pp. 146-58.
I. Chardin, Yoyages du chevalier Chardin en Perse VII, Paris, 1811.
W. B. Denny, “Turkmen Rugs and Early Rug Weaving in the Western Islamic World,” Halı 4/4, 1982, pp. 329-37.
M. S. Dimand and J. Mailey, Oriental Rugs in The Metropolitan Museum of Art, New York, 1973.
R. Du Mans, Estat de la Perse en 1610, ed. Ch. Schefer, Paris, 1890.
A. C. Edwards, The Persian Carpet, 1953.
K. Erdmann, Oriental Carpets, tr. C. G. Ellis, New York, 1960.
E. Gans-Ruedin, Iranian Carpets. Art, Craft, and History, London, 1978.
J. Housego, Tribal Rugs, London, 1978.
Idem, “18th Century Persian Carpets. Continuity and Change,” in Oriental Carpet and Textile Studies 1, 1986, pp. 40-51.
A. Ittig, “A Group of Inscribed Carpets from Persian Kurdistan,” Halı 4/2, 1981, pp. 124-27.
Idem, A Technical and Historical Study of the Qajar Carpet Industry, Ph.D. dissertation, Oxford University, 1983.
Idem, “The Persian Carpet. Previous Studies and Future Directions,” Oriental Rug Review 4/7, 1984, pp. 1-4.
Idem, “The Kirmani Carpet Boom—A Study in Carpet Entrepreneurship,” in Oriental Carpet and Textile Studies 1, 1985, pp. 111-23.
Idem, “Baḵtīārī iii. Carpets,” in EIr. III/6, pp. 561-62.
T. Kawami, “Archaeological Evidence for Textiles in Pre-Islamic Iran,” Iranian Studies 22/2-3, 1989, pp. 1-22.
M. Keyvānī, Artisans and Guild Life in the Later Safavid Period, Berlin, 1982.
D. King, ed., The Eastern Carpet in the Western World from the 15th to the 17th Century, London, 1983.
J. Lessing, Ancient Carpet Patterns after Originals of the 15th and 16th Centuries, London, 1879.
T. Mańkowski, “Some Documents from Polish Sources Relating to Carpet Making in the Time of Shah ʿAbbās I,” in Survey of Persian Art, pp. 2431-36.
F. R. Martin, A History of Oriental Carpets before 1800, Vienna, 1908.
J. Mills, Carpets in Pictures, London, 1975.
S. Parhām, Namāyešgāh-e naqša-ye qālī-e Kermān, Tehran, 1356 Š./1978.
R. Pinner and M. Franses, “Two "Turkoman" Carpets of the 15th Century,” in R. Pinner and M. Franses, eds., Turkoman Studies I, London, 1980, pp. 83-89.
A. U. Pope, “The Art of Carpet-Making. A. History,” in Survey of Persian Art, pp. 2257-2430.
B. W. Robinson, “Art in Iran x. Qajar 2. Painting,” in EIr. II/6, pp. 637-40.
J. Scarce, “Art in Iran x. Qajar 1. General” in EIr. II/6, pp. 627-37.
R. B. Serjeant, Islamic Textiles, Material for a History up to the Mongol Conquest, Beirut, 1972.
M. S. Simpson, “The Production and Patronage of the Haft Aurang by Jami in the Freer Gallery,” Ars Orientalis 13, 1982, pp. 92-119.
F. Spuhler, “Carpets and Textiles,” in Camb. Hist. Iran VI, pp. 698-727.
J. Thévenot, Travels of Monsieur de Thevenot into the Levant, pt. 2, London, 1686.
U.K. Foreign Office Records, Public Records Office, London, no. 248.1072, Sept. 8, 1913, Nov. 21, 1913.
M. Varzī, Honar o ṣaṇʿat-e qālī dar Īrān, Tehran, 1350 Š./1971.
This article is available in print.
Vol. IV, Fasc. 7, pp. 845-850