Pramāṇa · Search · About · API

Early Indian Architecture: I. Cities and City-Gates, etc.; II. Bodhi-gharas c9: Introduction: The Language and Process of Early Indian Architecture

Early Indian Architecture: I. Cities and City-Gates, etc.; II. Bodhi-gharas c9

Language
English

Contents

Introduction: The Language and Process of Early Indian Architecture

c9.2

Michael W. Meister

c9.3

Architecture has meaning through the process of its making, as well as from the cultural system within which it is made. Early Indian architecture had a strong pragmatic basis, functional in its rhetoric, rooted in carpentry and other building crafts. The visual language of early Indian architecture was, in fact, rooted in this pragmatism. Wattle-and-daub huts with thatched domed roofs were used by forest dwellers; vaulted rectangular wooden shelters, also thatched or shingled, acted as village assemblies. Elevated urban palaces had upper storeys with cantilevered balconies supported by struts; lattice screens; and outer terraces with balustrades and small open pavilions. A distinctive type of wooden roofing using barrel-vaults and dormers gave to these Indian cities — in the images that have come down to us — an original appearance unique to the Indian subcontinent.

c9.4

Ananda Kentish Coomaraswamy set out to document this urban language of form at its utilitarian and practical level in his “Early Indian Architecture” series of articles for *Eastern Art* — particularly Part I: “Cities and City-Gates” — much as he had sought out evidence for the surviving traditional architectural practices in Sri Lanka in his earlier monograph, *Mediæval Sinhalese Art* (1908).¹ He did so primarily by searching texts for descriptions and sculpted reliefs and murals for representations of ancient architecture.

c9.5

¹ Mediæval Sinhalese Art, a Monograph on Mediæval Sinhalese Art and Crafts, Mainly as Surviving in the Eighteenth Century with an Account of the Structure of Society and the Status of the Craftsman, Broad Campden, Gloucestershire: Essex House Press, 1908; Second Edition, New York: Pantheon Books, 1956.

c9.6

Yet India’s early thinkers also had long expressed a deep belief in the “iconicity” of the perceived world: its capacity to act, in Plato’s terms, as the “shadow” of “real” forms. This medieval “realism” was also attractive to Coomaraswamy and to a large degree transformed his knowledge of the built environment. India’s earliest architecture took not only the pragmatic forms of town and village but also those of a variety of highly charged symbolic monuments, such as free-standing pillars, mounds, and the tree-enclosures that Coomaraswamy discussed in his “Early Indian Architecture, Part II: Bodhi-gharas,” that served iconic ends.

“Aśokan” Pillars and “So-called Bell” Capitals

c9.s1.1

In this regard, Coomaraswamy tucked away one of his major contributions to our understanding of early Indian architecture in the two brief notes he wrote for *The Indian Historical Quarterly* on the “Origin of the Lotus-capital” — that typical “so-called bell-” shaped capital of Maurya-period pillars.² Significantly, he began the first of these notes by distinguishing that "the problem... may be considered from two points of view: (1) that of morphology, and (2) that of significance."³

c9.s1.2

² “Origin of the Lotus- (so called Bell-) Capital,” *The Indian Historical Quarterly* 6 (1930): 373-375; “Origin of the Lotus-Capital,” *The Indian Historical Quarterly* 7 (1931): 747-750

c9.s1.3

³ In the "Origin of the Lotus Capital," however, Coomaraswamy admitted that "decorative variations on the simpler themes of the Plant style are fast developing in early Indian art, and such subordination of meaning to ornament is a part of a normal development that takes place in any art" (p. 748). Coomaraswamy's first article was in response to Achyuta Kumar Mitra, "A Bell-Capital from Bhuvaneśvara," *The Indian Historical Quarterly* 5 (1929): 693-699. His second was in answer to Mitra's reply: "Origin of the Bell-Capital," *The Indian Historical Quarterly* 7 (1931): 213-244.

c9.s1.4

While other scholars had emphasized the formal decorative link between these capitals and Persepolitan prototypes, Coomaraswamy saw this borrowed form — as had the pillars' Mauryan craftsmen — in terms of the ripe lotus (fig. 1), its ripe seedpod acting, within the Indian system, as an appropriate pedestal to carry emblems of high power.⁴ This significance Coomaraswamy traced to "the oldest Indian cosmology, that of water," where, he explained, "we meet at once with the idea that water is the source and support of all things, particularly the source of life, and the support of the earth."⁵

c9.s1.5

⁴ Coomaraswamy's discussion of the lotus-pedestal appeared first in "Early Indian Iconography, II: Śrī Lakṣmī" *Eastern Art* 1 (1929): 174-189. As pointed out by A.K. Mitra, "the resemblance of the bell-capital to the calyx of a flower reversed" had been pointed out as early as 1875 by Rajendralal Mitra ("Origin of the Bell-Capital," p. 213). Coomaraswamy's contribution was morphological: he made the connection between the lotus-capital and a succession of standards bearing a variety of emblems. He pointed out that "my theory... regards the lotus capital as simply the termination of a shaft, and not as a *cihna*" ("Origin of the Lotus-Capital," p. 748).

c9.s1.6

⁵ "Origin of the Lotus- (so-called Bell-) Capital," p. 374. Coomaraswamy goes on to quote from the *Śatapatha Brāmaṇa* ("the lotus means the waters; thou art the back of the waters; this earth lies spread on the waters") and to state that: "Here the original significance of the lotus as representing the waters... is very clearly stated, and there is no need to invoke the later mystical ideas about a world lotus and *maṇḍalas*." See also John Irwin's more recent discussion of the topic: "'Aśokan' Pillars: A Reassessment of the Evidence," *The Burlington Magazine* 115 (1973): 706-720; "Part II: Structure," *The Burlington Magazine* 116 (1974): 712-727; "Part III: Capitals," *The Burlington Magazine* 117 (1975): 631-643; "Part IV: Symbolism," *The Burlington Magazine* 118 (1976): 734-753.

c9.s1.7

By separating the technology of carving from the cultural context within which this adapted form had been applied, Coomaraswamy was making a major statement about the iconic validity of these pillars as architecture. He did not, however, in this brief overview, place the pillars with any particularity into the historical context within which they were made. Thus the "significance" of the morphology was allowed to stand for the significance of the individual object.⁶

c9.s1.8

⁶ Irwin, "'Aśokan Pillars," also tends to make the cosmological frame equivalent to each pillar's particular meaning.

Morphology and the Symbolism of Architecture

c9.s2.1

Dan Sperber, in his study entitled *Rethinking Symbolism*, wrote that "by asserting that symbolism is a cognitive mechanism, I mean that it is an autonomous mechanism that, along side the perceptual and conceptual mechanisms, participates in the construction of knowledge and in the functioning of the memory."⁷ This statement seems apt in discussing the transformation wrought by the translation of early urban architectural forms into stone architecture in India.

c9.s2.2

⁷ *Rethinking Symbolism* (Cambridge Studies in Social Anthropology), trans. Alice L. Morton, Cambridge: Cambridge University Press, 1975, pp. xi-xii.

c9.s2.3

Much of what we know of the early phase of this architecture comes from Buddhist cave-excavations of the early centuries B.C. and A.D. These recreate free-standing assembly halls, at times with wooden rafters hung inside and wooden screens but with other details carved in stone (see Plates, pp. 39, 40, 42, below). Coomaraswamy used these excavations largely as if they were transparent sources for the wooden palaces from which their architectural language seemed derived. He wrote of them in his *History of Indian and Indonesian Art* (1927) that:⁸

c9.s2.4

⁸ *History of Indian and Indonesian Art*, New York: E. Weyhe, p. 28. Percy Brown, *Indian Architecture (Buddhist and Hindu Periods)*, Bombay: Taraporevala Sons & Co., 1942 (4th ed., 1959), takes the same tact.

c9.s2.5

These *caitya*-halls are excavated copies of wooden structural buildings as clearly appears in the literal imitation of timbered construction; occasionally wood was combined with stone, forming a screen of concentric ribs within the arch of the entrance, or applied to the stone ceiling to represent rafters, and in one or two cases part of the original woodwork has survived.

c9.s2.6

Such an estimate, however, might best fit the façade to the Mauryan-period Lomas Ṛṣi cave in Bihar, which replicates the construction of a small village thatched structure with great exactitude.

c9.s2.7

By the time of the Buddhist caves in the Western Ghats, on the other hand, the process of excavation had begun a development already that Coomaraswamy had characterized as “from the functional to the decorative” in reference to the elaboration of arched windows and their integration into a decorative web that was cast over the vertical bands of north-Indian temple-towers in later periods (below, p. 55; and Plates, pp. 52, 56-66).

c9.s2.8

I have instead written of this cave-temple period that:⁹

c9.s2.9

⁹ “Architecture: India, Introduction,” *Grove Dictionary of Art*, London: Macmillan Publishers Limited (in press).

c9.s2.10

Architects within this rock-cut tradition soon began to rearrange wooden forms, creating stone façades for excavations in ways that transformed carpentry sources into a viable decorative surface. Through clever manipulation of scale, they could make a façade suggest an almost unlimited range of constructed storeys. If such complex carved façades at first seem intended as substitutes for the labourious construction of a wooden original, they also came to represent a process of formal abbreviation and compaction that quickly became a source for architectural creativity in its own right. By the time stone temples began to be built in the fifth century A.D., such compaction of architectural form took on an elaborate and self-conscious symbolic vitality.

c9.s2.11

If the wood rafters hung from the interiors of some of these excavations represent almost the only carpentry surviving from ancient India, the way they have been integrated into the stone excavation had little to do with timber structuring. A variety of constructional expedients¹⁰ were used to rest this heavy vaulting on stone ledges and to bind its mighty beams to their lithic frame. The shape of this inner vault and that of the façade screen did not need to correspond. And those who designed the surrounding stone façade, as with the façade of a Roman stage-set, could play a variety of games with scale to create an appropriate effect. The fluidity of the sculpted stone may have encouraged such play; it in any case provided the opportunity.

c9.s2.12

¹⁰ James Fergusson had written instead of “constructive necessities,” *History of Indian and Eastern Architecture* (1876), revised by James Burgess, 2 vols., London: J. Murray, 1910, vol. I, p. 325.

c9.s2.13

Stella Kramrisch has suggested a set of the conceptual mechanisms used by artisans to create the surface of a temple. These she has defined in the following way:¹¹

c9.s2.14

¹¹ “The Temple as Puruṣa,” in *Studies in Indian Temple Architecture*, ed. Pramod Chandra, New Delhi: American Institute of Indian Studies, 1975, p. 43.

c9.s2.15

All these shapes originally were functional parts of a structure. Embodied in the temple and on its walls they retain some of their functional meaning even though not fulfilling their original function.... The rules regulating the use of these and other motifs as parts of the... integument of the temple wall are: (1) diminution of the original shape in proportion to the size of the monument and further proportionally graded reductions...; (2) repetition or identical shapes either in the vertical or in the horizontal...; (3) splitting of one entire motif into parts; (4) super-imposition in the third dimension of one shape upon the other; (5) inscribing one motif or theme into a different kind of theme or motif; and (6) contraction of several themes juxtaposed and or superimposed, according to the above rules, into one complex new entity.

c9.s2.16

A grasp of the perceptual and conceptual aspects of this process, however, makes it even more important that we understand the cognitive aspects of this "complex new entity's" symbolic basis. "Symbolic interpretation," according to Sperber, "is not a matter of decoding, but an improvisation that rests on an implicit knowledge and obeys unconscious rules."¹² Thus the conception of the temple as shelter — whether as palace or hut — precedes the architectural form.

c9.s2.17

¹² *Rethinking Symbolism*, p. xi.

c9.s2.18

In addition to the utilitarian aspects of Coomaraswamy's "Early Indian Architecture," his essays provide an incomparable foundation for understanding the symbolic underpinnings of India's religious and secular architecture. Coomaraswamy himself found that carrying this dialogue between "morphology" and "significance" further was a difficult task.

c9.s2.19

His essay on "The Symbolism of the Dome" in 1938, for example, had very little to say on morphology at all.¹³ In reviewing Heinrich Zimmer's *Kunstform und Yoga*, published in 1926, Coomaraswamy had written that:¹⁴

c9.s2.20

¹³ *Indian Historical Quarterly* 14 (1938): 1-56. Reprinted in Coomaraswamy, 1: *Selected Papers, Traditional Art and Symbolism*, ed. Roger Lipsey, Princeton: Princeton University Press, 1977, p. 415-464. He writes, p. 415: "The origin of any structural form can be considered either from an archaeological and technical or from a logical and aesthetic, or rather cognitive, point of view; in other words, either as fulfilling a function or as expressing a meaning. We hasten to add that these are logical, not real distinctions; function and significance coincide in the form of the work; however, we may ignore the one or the other in making use of the work as a thing essential to the active life of the body or dispositive to the contemplative life of the spirit."

c9.s2.21

¹⁴ *Artibus Asiae* 4 (1930-32): 78-79. Cited in Heinrich Zimmer, *Artistic Form and Yoga in the Sacred Images of India*, trans. by Gerald Chapple and James B. Lawson, Princeton: Princeton University Press, 1984, pp. xviii-xix.

c9.s2.22

No more valuable book for understanding of Indian art, the answering of the fundamental problem "Why is it what [it] is?", has yet been published.... No history of art, or of aesthetic theory, which does not take into account the Hindu point of view here so clearly expounded, can be regarded as at all complete.

c9.s2.23

It was in his 1930 note on "*Pāli kaṇṇikā:* Circular Roof-Plate"¹⁵ written in the same period as the "Early Indian Architecture" essays that Coomaraswamy could take pleasure in declaring that "the present discovery of the roof-plate as a typical architectural device... is of considerable interest for the history of the dome in India." He felt obliged, of course, to continue that, "like other wooden methods of construction, it would naturally have been copied in stone."¹⁶

c9.s2.24

¹⁵ *Journal of the American Oriental Society* 50 (1930): 238-243; reprinted in Coomaraswamy 1: *Selected Papers*, pp. 459-464.

c9.s2.25

¹⁶ Ibid, p. 462.

c9.s2.26

In his 1938 essay, however, Coomaraswamy would write instead that "we are here mainly concerned with significance.... We propose to ask rather *why* than *how*.'"¹⁷ Coomaraswamy's personal rush. towards metaphysics was apparent in the change between these two essays, as it had also been in the haste by which he "cleared the decks" early in the 1930s of the vast accumulation of data he had collected on specific morphologies.

c9.s2.27

¹⁷ Ibid., p. 415.

c9.s2.28

Fig. 1. Coomaraswamy's drawing of the lotus-pedestal.

c9.s2.29

*Coomaraswamy and Temple-Form*

c9.s2.30

The final problem with which Coomaraswamy concerned himself in "Early Indian Architecture: IV. Huts and Related Temple-Types," however, had been a morphological one: that of the formal sources for temple architecture. From the hermit's hut, Coomaraswamy felt that a "dome and cornice shrine" had evolved (below, p. 119) that was "still easily recognizable" in the development of Dravidian architecture.¹⁸ Of temple architecture in northern India he wrote that "a slightly different treatment of the reduplicated dome and cornice, with far greater compression of parts, was at the same time giving rise... to the simpler forms of the Nāgara *śikhara*" (p. 119, below).

c9.s2.31

¹⁸ He states that, following the temples at Māmallapuram, "it becomes unnecessary to emphasize the stylistic continuity in the southern development from this time onward."

c9.s2.32

In his *History of Indian and Indonesian Art* (1927) he wrote that:¹⁹

c9.s2.33

¹⁹ *History of Indian and Indonesian Art*, p. 83.

c9.s2.34

The original view propounded by Fergusson I believe to be the correct one. That is that the Nāgara spire, however elaborately developed, really represents a piling up of many superimposed storeys or roofs, much compressed. The key to this origin is the *āmalaka*; properly the crowning element of a tower, its appearance at the angles of successive courses shows that each of these corresponds in nature to a roof. Thus the Nāgara and Drāviḍa towers both originate in the same way.

c9.s2.35

In "Huts and Related Temple-Types," however, he concluded that "it is to be inferred that the whole scheme [of the Nāgara temple's *śikhara*] is developed from an original type of domed shrine surmounted by a single *āmalaka*" and went on to say that "we cannot pretend to offer here a complete solution to the *āmalaka* problem, that is, as to the actual origin of the form itself" (below, pp. 111-112).

c9.s2.37

Fig. 2. Robert DeJager, Watercolor, "The Temple Stripped Bare" (© Meister).

"Early Indian Architecture" and the Question of Nāgara Origins

c9.s3.1

The "origin" of the Nāgara temple, however, goes back to Sperber's issue of the triadic relationship between conception, perception, and symbolism. From Vedic times to now, the lobed myrobalan (*āmala*) has been recognized as a significant and purifying medicinal fruit.²⁰ It could, like the lotus, cognatively "represent" the water cosmology Coomaraswamy had so ably restored to our present understanding in his other work (fig. 1). As capital of a pillar, as with the lotus-"bell" in the Mauryan period, this powerful seed-form stood "as" — not "for" — cosmic potentiality. Along the shaft of this world-axis, placed at the center of a temple (fig. 2), were strung "*bhūmi*" levels — altars as well as worlds — like beads on a rosary.²¹

c9.s3.2

²⁰ The *āmalaka* is the fruit of the Emblic Myrobalan; *amala* can mean "spotless, stainless, clean, pure, shining." Sir Monier Monier-Williams, *A Sanskrit-English Dictionary*, Oxford: Oxford University Press, 1899.

c9.s3.3

²¹ *Bhūmi* = earth, storey. For Śiva's conquering of the "three worlds" see Michael W. Meister, "Śiva's Forts in Central India: Temples in Dakṣiṇa Kosala and Their 'Dæmonic' Plans," in *Discourses on Śiva*, ed. Michael W. Meister, Philadelphia: University of Pennsylvania Press, 1984, pp. 119-143.

c9.s3.4

This is one of the "realities" that lies behind the form. To it could be added the sheltering concept of the monastic hut (the "*gandhakuṭī*"). Over it, architects eventually would cast the "shadow" forms of a palatial architecture appropriate for the temple's function as shelter for a manifesting divinity (figs. 3-6).²²

c9.s3.5

²² Michael W. Meister, "On the Morphology of a Symbolic Architecture: India," *Res* 15 (1986): 33-50.

c9.s3.6

Coomaraswamy's analysis of the formal language of early Indian architecture has made it much easier for us to read the palatial elements combined in a structure such as the seventh-century brick temple at Rajim in Madhya Pradesh shown in the Frontispiece (p. ii).²³ The barrel-vaults, hip-roofs, domed pavilions, and pillared walls compacted into this structure are familiar forms (figs. 3, 4). The square corner aediculae, like the *kūṭa* pavilions of a Dravidian shrine, signal the appropriateness of Coomaraswamy's conclusion that "Nāgara and Drāviḍa towers both originate in the same way."²⁴

c9.s3.7

²³ See also Michael W. Meister, "Prāsāda as Palace: Kūṭina Origins of the Nāgara Temple," *Artibus Asiae* 49 (1989): 254-280.

c9.s3.8

²⁴ *History of Indian and Indonesian Art*, p. 83.

c9.s3.9

The significance of the morphology, however, is not sufficient to explain the symbolism of the object. Coomaraswamy's separation of morphology and significance does not go far enough. To end this discussion, then, I might accept Sperber's "possible hypothesis" that "the basic principles of the symbolic mechanism are not induced from experience but are, on the contrary, part of the innate mental equipment that makes experience possible."²⁵

c9.s3.10

²⁵ *Rethinking Symbolism*, p. xii.

c9.s3.11

The Nāgara temple replaces palatial forms with what I have called "symbolic substitutes," making each monument, to steal the phrase, "a thousand points of [sacred] light." In place of the *kūṭas* of palace architecture, each aedicula reproduces the temple and reduces it again to its symbolism of altar and pillar (fig. 5).²⁶ The final form of the Nāgara temple thus is as mysterious as it is powerful (fig. 6), weaving together (as had cave-temple façades) what is conceived, perceived, and symbolically understood.

c9.s3.12

²⁶ Michael W. Meister, "De- and Re-constructing the Indian Temple," *Art Journal* 49 (1990): 395-400.

c9.s3.14

Fig. 3. Rājīm, Rājīvalocana temple, separation of *kūṭas* (© Meister).

c9.s3.15

Fig. 4. Rājīm, Rājīvalocana temple. Palatial structuring of the façade (© Meister).

c9.s3.16

Fig. 5. Ālampur, Viśva Brahmā temple. Kūṭina implications of a Nāgara temple’s façade (© Meister).

c9.s3.17

Fig. 6. Robert DeJager, watercolor of a Nāgara śikhara (© Meister).

c9.s3.18

*Reference Bibliography: M. W. Meister*

c9.s3.19

Altars and Shelters in India. *aarp* (*Art and Archaeology Research Papers*) 16 (1979): 39.

c9.s3.20

Analysis of Temple Plans: Indor. *Artibus Asiae* 43 (1982): 302-320.

c9.s3.21

Construction and Conception: Maṇḍapīka Shrines of Central India. *East and West*, new series 26 (1978): 409-418.

c9.s3.22

Dārra and the Early Gupta Tradition. In *Chhavi II, Rai Krishna Dasa Felicitation Volume*, edited by Anand Krishna, pp. 192-205. Banaras: Bharat Kala Bhavan, 1981.

c9.s3.23

De-and Re-constructing the Indian Temple. *Art Journal* 49 (1990): 395-400.

c9.s3.24

*Encyclopædia of Indian Temple Architecture*, edited by M. W. Meister, M. A. Dhaky, and Krishna Deva. Delhi: Oxford University Press, 1983, 1986, 1988, 1991.

c9.s3.25

Essay in Indian Architecture, An. *Roopa Lekha* 41 (1973): 35-47.

c9.s3.26

Forest and Cave: Temples at Candrabhāgā and Kansuāñ. *Archives of Asian Art* 34 (1981): 56-73.

c9.s3.27

Geometry and Measure in Indian Temple Plans: Rectangular Temples. *Artibus Asiae* 44 (1983): 266-296.

c9.s3.28

Hindu Temple: Axis and Access, The. In *Concepts of Space, Ancient and Modern*, Edited by Kapila Vatsyayan, pp. 269-80. New Delhi: Indira Gandhi National Centre for the Arts & Abhinav Publications, 1991.

c9.s3.29

*Making Things in South Asia: The Role of Artist and Craftsman*, edited by M. W. Meister. Philadelphia: South Asia Regional Studies Department, University of Pennsylvania, 1988.

c9.s3.30

Maṇḍala and Practice in Nāgara Architecture in North India. *Journal of the American Oriental Society* 99 (1979): 204-219.

c9.s3.31

Measurement and Proportion in Hindu Temple Architecture. *Interdisciplinary Science Reviews* 10 (1985): 248-258.

c9.s3.32

Muṇḍeśvarī: Ambiguity and Certainty in the Analysis of a Temple Plan. In *Kalādarśana, American Studies in the Art of India*, edited by Joanna G. Williams, pp. 77-90. New Delhi: Oxford & IBH Publishing Co., 1981.

c9.s3.33

On the Development of a Morphology for a Symbolic Architecture: India. *Res, Anthropology and Aesthetics* 12 (1986): 33-50.

c9.s3.34

Prāsāda as Palace: Kūṭina Origins of the Nāgara Temple. *Artibus Asiae* 49 (1989): 254-280.

c9.s3.35

Reading Monuments and Seeing Texts. In *Śāstric Traditions in Indian Arts*, edited by A. Dallapiccola, 2 vols., I, pp. 167-173; II. pp. 94-108 & pls. 81-88. Stuttgart: Franz Steiner Verlag, 1989.

c9.s3.36

Śiva's Forts in Central India: Temples in Dakṣiṇa Kosala and Their ‘Dæmonic’ Plans. In *Discourses on Śiva*, ed. M. W. Meister, pp. 119-143. Philadelphia: University of Pennsylvania Press; 1984.

c9.s3.37

Sub-Urban Planning and Rock-Cut Architecture in India. In *Madhū, Recent Researches in Indian Archaeology and Art History*, edited by M. S. Nagaraja Rao, pp. 157-164. Delhi: Agam Kala Prakashan, 1981.

c9.s3.38

Symbol and Surface: Masonic and Pillared Wall-Structures in North India. *Artibus Asiae* 46 (1985): 129-148.

c9.s3.39

Temple Building in South Asia: Science as Technology's Constraint. In *Science and Technology in India*, edited by Peter Gaeffke and David A. Utz, pp. 31-36. Philadelphia: University of Pennsylvania, Department of South Asia Regional Studies, 1985.

c9.s3.40

Udayeśvara Temple Plan, The. In *Śrīnidhih, Perspectives in Indian Archaeology, Art and Culture*, pp. 85-93. Madras: New Era Publications, 1983.

c9.s3.41

Early Indian Architecture

c9.s3.42

Parts 1 and 2

c9.s3.43

FIG. I. UPPER STOREY (*uttamāgāra*)

c9.s3.44

FIG. 2. SECOND STOREY (*ardhatala*)

c9.s3.45

FIG. 3. GROUND FLOOR (*āditala*) WITH BRIDGE AND MOAT

c9.s3.46

*Plans of a city gate (based on Kauṭilīya Arthaśāstra and early reliefs)*

c9.s3.47

EARLY INDIAN ARCHITECTURE¹

c9.s3.48

¹ Mainly as referred to in the Pali literature and as represented in the reliefs of Bharhut, Sāñcī, and Amarāvatī. Terms are cited in Pali, Sanskrit, or Prakrit, according to their source. Where no particular text is cited, it is to be understood that the term is of very common occurrence, or that the sense is well known. Abbreviations: Cull., *Cullavagga*; DhA., *Dhammapada Aṭṭhakathā*; DHA, Acharya, *Dictionary of Hindu architecture*; Div., *Divyāvadāna*; DN., *Dīgha Nikāya*; IAT., my *Indian architectural terms, JAOS*, vol. 48, 1928; J., *Jātaka* (Fausböll); KA., *Kauṭilīya Arthaśāstra* (Bk II, Ch. 21 unless otherwise stated); Mhv., *Mahāvaṁsa*; Mv., *Mahāvagga*; Mil., *Milindapañha*; MN., *Majjhima Nikāya*; MSA., my *Mediaeval Sinhalese art*; PvA., *Pethavatthu Aṭṭhakathā*; SBE., *Sacred Books of the East* The *vaḍḍhakī, vaḍḍhakī* (Skt. *vardhaki*), architect, builder, carpenter (also shipwright, J. VI, 432), is by far the most important and oftenest mentioned of the craftsmen. Cf. Miln. 345. "When men see a pleasant city, they know by inference how great the *vaḍḍhakī* was." In J. II, 208, we have a Brāhmaṇa*vaḍḍhakī*. Tools used by *vaḍḍhakis* include adzes (*vāsī*, J. II, 18 and IV, 344); chisels (*nikhādana*, J. IV, 344), hammers (*muggara*, ib.), measuring line (*sutta, kāla sutta*, J. IV, 344 and VI, 332, pegged out with wooden pegs, *khāṇuka*, J. VI, 332); they work *sippānurūpena* "according to *silpa-plans*," perhaps a reference to *śilpa-śāstras* (cf. *sippa, sippavant* in Pali Dictionaries). DHA rendering of *vardhaki* as "painter" is absurd. In Mhv. LXXXVIII, 106, 107, *iṭṭhikā-, cuṇṇa-, dāru-* and *silā-vaḍḍhakīs* are mentioned, i.e., bricklayers, plasterers, timberers (carpenters) and masons. While *vaḍḍhakī*, alone, is the carpenter as architect, we also get *tacchaka* as woodworker, Mil. 413.

c9.s3.49

By ANANDA K. COOMARASWAMY

I. Cities and City-Gates, etc.

c9.s4.1

The old Indian city is by no means to be regarded as a typically Brahmanical institution. Apart from the fact that we have remains of well-developed pre-Aryan cities in the Indus Valley sites,² and from the fact that the Vedas make occasional reference to the "cities of the Dasyus," it is to be observed that in the Brahmanical law books, which are very nearly, if not quite contemporary with the architectural period to be discussed below, cities are despised, and there are no ceremonies for urban life; the *Baudhāyana Dharma Sūtra*, II, 3, 6, 33, says "It is impossible for one to obtain salvation, who lives in a town covered with dust."

c9.s4.2

² "The true creators of the city state were the Sumerians . . . the higher city civilisations of the Old World from the Mediterranean to India do form a great unity, and it seems possible to affirm their fundamental community alike in type and origin," Dawson, *The age of the gods*, pp. 117, 118. In other words, the city in Indian culture belongs to the "Early Asiatic" inheritance, and certainly cannot be regarded either as characteristically Aryan, or as a recent development in Maurya times.

c9.s4.3

Flourishing cities nevertheless certainly existed in centuries preceding the Christian era. Megasthenes writing, about 300 B. C., described Pāṭaliputra as over nine miles in length. References to and descriptions of cities, and city organization in Buddhist literature, in KA., and in the Epics, and the representations in early Buddhist reliefs all suggest that in the Maurya period we have already to do, not with any new developments, but with the continuation of a long-established and familiar type of urban organization.

c9.s4.4

Some of the most detailed descriptions of cities are to be found in Mil. 1, 34, and 330 ff. The last of these, despite its length, may be quoted in full. "Just as the architect of a city, when he wants to build one, would first search out a pleasant spot of ground, with which no fault can be found, even, with no hills or gullies in it, free from rough ground, and rocks, not open to danger of attack. And then when he has made plain any rough places there may still be on it, he would clear it thoroughly of all stumps and stakes, and would proceed to build there a city fine and regular, measured out into quarters, with excavated moats and ramparts about it, with stout gate-houses and towers, with market-places, cross-roads, street-corners, and public squares, with cleanly and even main roads, with regular lines of open shops, well-provided with parks, gardens, lakes, lotus-ponds, and wells, adorned with many kinds of temples of the gods, free from every fault. And then, when the city stood there in all its glory, he would go away to some other land." There follows a long list of those who will inhabit such a city, nobles, soldiers, craftsmen, and of the foreigners who will resort to it: "all these coming to take up their residence there, and finding the new city to be regular, faultless, perfect, and pleasant, would know 'Able indeed must that architect have been by whom this city was built!' Just so . . . the Blessed One's City of Righteousness has righteousness for its rampart, and fear of sin for its moat, knowledge for its city-gate, zeal for the *City gates, from Sāñcī* gate-tower, faith for the pillars at its base, mindfulness for the guard, and wisdom for the palace, the Suttantas for the market place, the Abhidhamma for the public square, the Vinaya for its judgment hall, and constant self-possession for its street.'³

c9.s4.5

³ Based on the translation in SBE XXXVI, with verbal changes.

c9.s4.6

FIG. 4. CITY OF KAPILAVATTHU

c9.s4.7

FIG. 5. CITY OF JETUTTARA (Vessantara Jātaka)

c9.s4.8

FIG. 6. CITY OF KUSINĀRĀ (WAR OF THE RELICS)

c9.s4.9

Let us now assemble the available data in a more systematic way. Cities (*nagara*, *pura*) are built by city or master architects (*nagara-vaḍḍhakī*, *mahā-*, Mil. 1, 2, 330; J. VI, 332), assisted by other carpenters (*vaḍḍhakī*, J. II, 18; IV, 153, 344; VI, 4, 27, etc.) and bricklayers (*iṭṭhaka-vaḍḍhakī*, Mhv. XXX, 5; J. VI, 333). All these live in villages, and are summoned when work is to be done. Of the carpenters in particular we learn that they lived in villages of their own, going up river to cut and fetch timber for town houses (J. II, 18; VI, 427). Eighteen guilds of craftsmen are employed in building a new city, and amongst these guilds (*seṇi*) are cited the aforesaid carpenters, the superior smiths (*kammāra*), workmen (*cammakāra*),⁴ decorators and painters (*cittakāra*), "and others skilled in various crafts (*sippa*)" (J. VI, 427, cf. list of craftsmen in Mil. 331), who work according to the traditions of their craft (*sippānurūpena*, J. VI, 332).

c9.s4.10

⁴ *Cammakāra*, "leather-workers," but often apparently equivalent to *rathakāra*, wright or artisan generally, e.g., J. V, 174; *PvA.* 175. If *cammakāra*, as is not improbable, is equivalent to blacksmith or ironworker, this would be with reference to the use of leather bellows, cf. MSA., p. 194. In J. VI, 432, sappers using leather sacks (*cammāluka*) dig earth and build ramparts; but the *cammakāra* are evidently craftsmen, not coolies.

c9.s4.11

The most conspicuous and necessary parts of a city are the moat (*pārikhā*) and rampart (*pākāra*), gates (*dvāra*, *gopura*), more specifically gate-houses (*dvāra-koṭṭhaka*), with their defense towers (*dvāra-aṭṭālaka*, *gopura-°*), other defense towers within and near, but not upon the rampart (*antaraṭṭāla*, J. II, 400), and the king's palace (*pāsāda*, *harmya*, *rāja-nivesana*, *vimāna*, etc.). Then there are other mansions (*nivesana*, *kūṭāgāra*) and houses (*geha*), temples (*devaṭṭhāna*, Mil. 92, 330; *koṣṭhaka*, a temple granary, KA., Ch. 25), granaries (*koṭṭha*, *koṭṭhaka*), resthouses or sarais (*sālā*, *puṇyasālā*, J. I, 200 = *vissamana-sālā*, DhA. I, 269/270), halls and arenas for sport (*kīla-sālā*, *kīla-maṇḍala*, J. VI, 332/333 monasteries (*paṇṇa-sālā*), almonries (*dāna-sālā*, at city and palace gates, J. I, 262, IV, 402, VI, 97, 484, 487; *dānagga* "in hundreds," Mil. 2) elephant stables (*hatthi-sālā* J. VI, 432, etc.), shops (*āpaṇa*), bazaars (*antarāpaṇa*, Mil. 2), saloons, cook-shops, taverns, slaughterhouses (*pānāgara*, *odaniyā-ghara*, *soṇḍā*, *sunā*, J. VI, 276, etc.).

c9.s4.12

There are parks (*uyyāna*, Mil. 34; J. III, 238, VI, 333), gardens (*ārāma*, Mil. 34) and flower-gardens (*pupphārāma*, DhA. I, 270), lotus ponds and bathing tanks (*pokkhariṇī*, *nahāna-pokkhariṇī*, *nāhanodaka*, *udapāna*, *taḷāka*); sacred trees (*rukkhacetiya*, often at the gates). There are main streets (*rāja-magga*, *mahā-patha*, *toranamagga*, J. I, 199/200, III, 217, etc.), ordinary streets (*vīthi*, J. I, 89, and *antaravīthi*, J. I, 340), alleys and blind lanes (*patatthi*, *sandhi-bbūha*, J. III, 217), a main public square (*singhāṭaka*, Mil. 62; J. VI, 276), a market place (*caccara*, Mil. 332, meaning rather dubious), other squares and street crossings (*cattuka*, Mil. 34; J. I, 326 and 340; *sandhi*, Mil. 330). That some streets at least were exclusively inhabited by members of particular trades or castes, as in modern Indian cities, is indicated by J. VI, 485, *vessānaṁ vīthiyā*, *vessa-vīthiyā*, "street of the Vaiśyas or merchants." Evidently the four main streets (*catumahāpatha*, J. I, 199/200) led direct from the gates to the central square (*singhāṭaka*) on which the palace abutted; for a city elder from the *singhāṭaka* can see a man coming from East, South, West or North (Mil. 62); a man proceeds directly from the city gate to the palace (J. VI, 412); and the *catumahāpathe*, as the place where a sarai is built (J. I, 200/201) is clearly equivalent to *singhāṭake*. A drain or sewer (*niddhamana*, *niddhamana-magga*, J. I, 425, 490) leads out of the city, and it is possible for a man to make his escape by it. Outside the city are suburbs (*nigama*, J. VI, 330), and rural villages (*yava-majjhaka-gāma*, J. VI, 330).

c9.s4.13

The city is laid out in quarters, *bhāgaso mitaṁ*, Mil. 1, J. VI, 46, etc. The plan is rectangular, usually square, with four gates, one in the middle of each wall, facing [Figure: FIG. 8. CITY OF KUSINĀRĀ (WAR OF THE RELICS) .]

c9.s4.14

FIG. 7. FROM AN ARCHITRAVE, MATHURĀ (CAT. I, 38)

c9.s4.15

[Figure: FIG. 9. CITY OF JETUTTARA (*Vessantara Jātaka*)]

c9.s4.16

FIG. 10. CITY OF KOSALA

c9.s4.17

FIG. II. CITY OF RĀJAGAHA

c9.s4.18

*City gates, from Mathurā (Fig. 7) and Sāñcī (Figs. 8–11)*

c9.s4.19

the four quarters (J. I, 262; II, 171, 194; III, 129; IV, 83, 425; VI, 330, 347),⁵ and as we have seen, four main streets led from these gates to the centre of the city. The gates are closed at night or in time of war (J. II, 412; VI, 406), and opened in the morning in time of peace. In the most general way the gates are called *dvāra*, or less often *gopura* (J. VI, 276, has the gloss *dvāra=gopura*); but more specifically, the term gate-house (*dvāra-koṭṭhaka*) is constantly employed. As the city is approached from without, the gate-house (also called *daḷha-gopura*, *daḷham-aṭṭāla-koṭṭha*, *daḷha-gopura-aṭṭāla-koṭṭha*) came into view, with its defense towers (*dvāra-aṭṭālaka*, *gopura-aṭṭālaka*)⁶ on the city side of the moat.

c9.s4.20

⁵ But one Sāñcī relief (Fig. 6) shows two gates (one of unique form), not far apart in a single line of wall. Obviously, large cities must have had more than four gates; Pāṭaliputra, according to Megasthenes, had sixty-four. In Mhv. LXXXVIII, 116, the wall of Pulatthipura is circular: cf. representations of Ujjain in Jaina MSS., e.g. *Cat. Indian Collections, Boston*, Pt. IV, Pl. X (folio 75).

c9.s4.21

⁶ In the *Uttarādhyayana Sūtra*, *ṭīkā*, IX, 18 (Charpentier, p. 314), gloss on *gopuraṭṭāla*, we have "the towers are military structures (*āyodhasthānāni*) connected with the rampart-house (*prākārakoṣṭha*, presumably = *dvārakoṭṭhaka*).

c9.s4.22

The city is always surrounded by a moat or moats (*parikhā*, Skt. *paligha*) or at least a dry ditch. In the reliefs, the moat is indicated by growing lotuses, and in many cases the actual water is also shown; occasionally we see women from the city going down to the moat to fetch water, and in one case there is access on the city side through a little gate, so that the moat on the city side becomes effectively a *ghāṭ* (Fig. 4). In the texts, the water is usually taken for granted, but in J. VI, 432, a water-moat, *udaka-parikhā*, is specified. In Mil. I, the moat is "deep," *gambhīra*. In Mhv. XXV, 48, we find reference to a great triple moat, *timahāparikhā*; and the gloss to *utkiṇṇantara-parikhaṃ*, "deep excavated inner moat" of J. IV, 106, defines three successive external moats, *udaka-*, *kaddama-* and *sukha-parikhā*, one within the other, respectively a water moat, a mud moat, and a pleasure moat. J. VI, 276, has *parikhāyo*, moats, in the plural.

c9.s4.23

On the city side of the moat rises the wall (*pākāra*, Skt. *prākāra*), from a foundation or plinth (*vapra*),⁷ which in the case of a dry ditch should perhaps be called a glacis. It is generally indicated in the reliefs that city walls are made of brick, but in one Sāñcī relief (Figs. A and 4), we find what is evidently a wooden wall of palisade construction,⁸ and in this case the wall is continued along the edge of the bridge leading to the gateway; the *Uttarādhyayana Sūtra*, 30, 16 is authority for a *paṁśuprākāra*, earthen (laterite) wall, of a city. Mil. I has *paṇḍara-pākāra*, white wall.

c9.s4.24

⁷ *Vapra* may be sometimes synonymous with *pākāra* (DHA, 534), but in KA XXIV we have *vaprasyopari prākāram*.

c9.s4.25

⁸ Doubtless heavy planks laid horizontally and fastened by solid uprights. This type of construction is well seen in the conduit (*niddhamana*?) walls recently excavated at Pāṭaliputra (*Illustrated London News*, March 24, 1928, p. 477).

c9.s4.26

Sometimes the wall is shown with reëntrant angles. It is finished off at the top either by a coping or more usually by battlements. In no case (single exception, Fig. 16), are towers on the walls represented, nor have I found any reference in the texts to the existence of such towers; defense towers are either the regular *dvāraṭṭāla* of the gate-houses, or free-standing towers (*antaraṭṭāla*) near the walls but within them (Figs. B, 6, 8, 11, etc.). The walls could be manned, having, no doubt, an inner gallery or pathway a little below the top, and approached by stairs as suggested in the plan (Fig. 3). The top (*matthaka*) of the wall was of some width: in J. VI, 275, Puṇṇaka's horse gallops along the *pākāra matthaka*.

c9.s4.27

The gate-house was approached by a bridge (*saṅkama*, Mil. 92, *saṁkrama*, KA) in the case of a water-moat, or earthen causeway (*pūrdvārī mṛtkūṭa*, *Śiśupālavadha*, III, 68, gloss) when the city was defended only by a dry ditch. Such a bridge or causeway seems to have been designated by the curious name of "elephant's nail" (*hastinakha*).⁹ Such bridges, like thoroughfares, are public places, unsuited for serious talk (Mil. 92).

c9.s4.28

⁹ "Elephant's nail," originally a kind of capital composed of addorsed elephants, their forefeet projecting beyond the abacus so as to make the toe-nails conspicuous when seen from below; secondly, a balcony or bridge supported by such a capital; finally, even a causeway substituted for such a bridge. See IAT, pp. 258, 259. (It was by an oversight [thinking of *kesara-nakha*] that I adduced *kesa-nakha* (which means hair and nail-paring relics) as a parallel here).

c9.s4.29

The traveller, crossing the bridge, and entering the city, passes between two high towers, the *dvāra-aṭṭālaka*, *gopura-aṭṭālaka*; these are forwardly projecting members of the gate-house, towers of defense, to be manned in case of necessity (J. II, 217, and VI, 409, and as represented in the reliefs). In J. II, 244/5 we read of a king who ascends (*abhirūhitvā*) and again comes down from (*otaritvā*) a gate-tower; in all probability these towers contained the stairways which led to the upper floors (KA, Ch. 21, discussed below).

c9.s4.30

FIG. 12. CITY OF RĀJAGAHA(?)

c9.s4.31

FIG. 13. CITY OF BENARES (*Chaddanta Jātaka*)

c9.s4.32

FIG. 14

c9.s4.33

FIG. 15. CITY OF KUSINĀRĀ (PARTITION OF THE RELICS)

c9.s4.34

FIG. 16

c9.s4.35

FIG. 17

c9.s4.36

The main part of the gate-house is even or continuous (*prākāra-sama*, KA, and as seen in the reliefs), with the rampart on either side, and in the centre bridges the great hall (*sālā*) of the gate-house, and connects the two towers already mentioned; the two ends contiguous with the rampart extending to right and left of the forwardly projecting towers are similar to these in appearance, and form, in fact, two other towers of defence. The space between the projecting towers and immediately in front of the great gate-hall is the "mouth" (*mukha*,¹⁰ KA, Ch. 21).

c9.s4.37

¹⁰ *Mukha* = entrance, not as structure, but as space: *cf*. the *bahimukha*, entrance from without, of apartments (*ṭhāna*) in a *sarai* (*sālā*), J. VI, 333.

c9.s4.38

It is often clearly indicated in the reliefs that the whole basement of the gate-house is of brick. In any case, the basement wall is unbroken up to the level of the second storey, except that high up on the front face of each of the forward towers there is a small horizontal slit window; or more rarely a group of small square apertures; these loopholes probably served the double purpose of lighting the stairway inside, and of a post for archers. The ends of the supporting beams of the second or mezzanine floor (*ardha-tala*, KA) are generally clearly shown, projecting at the top of the basement wall. Probably with reference to its pavilion-like construction (*sthūṇāvabandha*, KA, Ch. 21) this second floor is designated *harmya* (*ib.*); a railing (*vedikā*), pillars (*thambha*) and cornice are always clearly shown. Above this is a top floor, supporting an attic-house or roof apartment (*uttamāgāra*, KA, Ch. 21), with brick walls, four gable window ends (*mahāvātayāna*) and thatched, barrel-vaulted roof (*chadana*), having its ridge (*kūṭa*), surmounted by finials (*ghaṭa* or *kalasa*, DhA, I, 414) on each gable and at intervals along the roof ridge. No doubt the barrel-vaulted roof was constructed as usual of curved rafters (*gopānasiyo*) resting against an internal roof-ridge (*kūṭa*, Mil. 38; DhA, I, 414), or if with apsidal ends, against circular or semi-circular roof plates (*kaṇṇikā*).¹¹ In J. VI, 125, the gateway of the Tāvatiṃsa heaven is called *citta-kūṭa-dvāra-koṭṭhaka* "gate-house of the painted (or decorated) roof-ridge." The same gate-house is set about with statues of Indra "as though guarded by tigers;" evidence, perhaps, that statues of a reigning king might be set up at a city gate.

c9.s4.39

¹¹ See my Pali *kaṇṇikā* = *circular roof-plate*, to appear in J.A.O.S., vol. 50.

c9.s4.40

The actual gateway, opened or closed as occasion required, consisted of a pair of heavy wooden panels (*kavāṭa*, Skt. *kapāṭa*), sometimes iron-bound or studded (*ayo-kammata-dvāra*, Mhv. XXV, 28) and turning on tenons above and below. The upper parts of these leaves closed against the top of the archway, (*toraṇasīraḥ*, KA), the lower part against the heavy *indakhīla* (Skt. *indrakīla*) embedded in the ground between the pillars of the *toraṇa*, and forming a low threshold.¹² The gate leaves are not often visible in the reliefs, but can be very clearly seen in Figs. A and 4 where a half-opened leaf partly conceals a soldier standing in the entrance; and in Figs. C and 6 we see the besiegers battering on the leaves of the door (*cf*. *kavāṭe ṭhāpetvā*, "beat against the door panels", in this case, however, of a palace, J. IV, 182). Having regard to the necessity of providing passage for elephants with riders, we may suppose the dimensions of the gate aperture to have been something like twenty by eighteen feet, and many existing mediæval city or fort gateways are quite as large as this; and from this, some idea of the size of the whole gate-house can be deduced.

c9.s4.41

¹² *Indakhīla*, in DhA, II, 180, 181, a symbol of stability. Generally, "threshold," *e.g.*, Div. 250, 20; 365, 1: *Mahāvastu*, I, 195, 16; I, 235, 12; I, 308, 7: *Madhyamakasūtravṛtti*, 199, 13. In J. I, 89, *indakhīle ṭhito* "standing at the gate" (of Kapilavatthu). In DN. II, 254, we have *chetvā khilaṃ* (v. l. *khīlaṃ*) *chetvā palighaṃ inda-khīlaṃ ūhacca-m-anijā* "all bars and bolts are hewn in twain, the threshold-stone dug up." Professor Clark informs me that *indakhīla* is rendered in Tibetan (*e.g.* *Mahāvyutpatti*, ed. Sakaki, 5582) by words meaning "threshold" or "steps at the threshold." *Khīla* alone = pillar (DN. II, 254), and *kīla* has this sense when *indrakīla* denotes "Indra's flagstaff" as in *Mahābhārata*, VI, 59, 122.

c9.s4.42

FIG. 18. JAIPUR, CITY GATE (AFTER REUTHER)

c9.s4.43

FIG. 19. BĪJĀPUR, FORTRESS GATE (AFTER REUTHER)

c9.s4.44

*Later city and fortress gates.*

c9.s4.45

FIG. A. DIAGRAMMATIC SKETCH OF A GATE-HOUSE, MOAT, AND BRIDGE (SEE FIG. 4)

c9.s4.46

The gate-leaves are framed at the sides by heavy jambs, usually called *esikā* (Mil. 332; J. II, 94/95 and VI, 276, glossed as *thambhas* set firmly in the ground) but the word *esikā* may include all the gateway pillars, that is the bolt post, jamb, and *toraṇa*-upright at each side (cf. Fig. 20). In J. VI, 276, *toraṇāni* is glossed *piṭṭha-samghāta* (the usual terms for jamb in the case of smaller doorways) but this must be regarded as a rather casual use of terms, not strictly correct: for the use of *toraṇāni* in the plural cf. Fig. 20. When closed, the gate-leaves are secured by one, or perhaps usually two (KA) heavy cross-bars (Pali *paligha*, Skt. *parigha*)¹³ resting in slots in the bolt-posts immediately behind the jambs, for example *Buddhacarita*, V, 82, *guru-parigha-kapāṭa saṃvṛttā*, "the heavy bar and gate-leaves being closed." We also find mention of *aggaḷāni*, bolts, which are well known in connection with smaller doors; in J. VI, 276 and 483 the Commentator makes *aggaḷāni*=*dvāra-kavāṭāni*, but this again must be regarded as casual. In KA the fastenings of the *aṇidvāra* are called *hasti-parigha*. J. VI, 483 gives us *citraggala*, decorated bolt, and J. V, 169, *rajataggala*, silver bolt. When the gate is attacked by elephants, stress is laid on breaking down the *toraṇa* or *toraṇas*, breaking the *paligha*, and uprooting the *esikāni* (J. II, 94/95; DN. III; 14, 254).

c9.s4.47

¹³ Pali *paligha* (=Skt. *parigha*) or *palikha* is always a cross-bar closing a gate, *parigha* (Skt. *paligha*) or *parikha* a moat. In J. II, 400, *parikhaṁ bhinditvā*, the sense requires *palikhaṁ*, and perhaps an emendation should be made.

c9.s4.48

FIG. B. GATE-HOUSE, AND BUILDINGS WITHIN THE CITY-WALL. THE STRUCTURE IN FRONT TO RIGHT IS A MOVABLE VEHICLE, A KIND OF PALANKEEN. AMARĀVATĪ, MADRAS MUSEUM

c9.s4.49

Passing between the gate-posts (*esikāni*) under the arch (*toraṇa*) of the door, and between the two swinging door panels (*kavāṭa*) or when the great gates are closed, through the *cūḷa*- or *aṇidvāra* which opens through one of the large panels (J. VI, 391, 399, 401, and KA),¹⁴ the traveller enters *Later city and fortress gates.*

c9.s4.50

¹⁴ A smaller door cut in one of the large gate-leaves is a constant feature of mediæval gateways, and I have seen many examples. In J. VI, 406, the *cūḷa-dvāra* and *dvāra* (here=*kavāṭa*) of a city gate are contrasted; but *cūḷa-dvāra* is not exclusively a technical term for a door in a great gate, for in J. VI, 432, 460, *cūḷa* and *mahā-dvāras* are simply small and large doors of interior chambers. In the *Gobhila Gṛhya Sūtra*, IV, 7, 19–20 *aṇidvāra* is the back door of a house.

c9.s4.51

FIG. 20. CITY GATE DABHOI, GUJARĀT

c9.s4.52

FIG. 21. FORTRESS GATE, GWĀLIAR

c9.s4.53

a large hall (*sālā*, KA, Ch. 21) roofed by the second storey (*ardha-tala*) of the actual gate-house. On either side of him are stone platforms (*pratimañcau*).

c9.s4.54

An integral part of the gate-house is the *toraṇa* (*varaparava gopura-toraṇa*, "splendid gate-arch," Mil. 1), or arch against which the gate-leaves close from within; the upper part of this arch being designated *toraṇa-sirahs* (KA) and the threshold *inda-khīla*, Skt. *indrakīla* (see above). Such arches are easily recognizable in mediæval fortress gates, e.g. at Dabhoi, Fig. 20, and Gwālior, Fig. 21, but they cannot be made out at Sāñcī or Amarāvatī, where we can hardly ever see far enough into the *mukha* for the actual gateway to be visible. In one case, at Sāñcī (Figs. A and 4) there is evidently no such arch, but there is a kind of semi-circular door-stop at the top. At Bharhut, no representation of a city gate is preserved, but we have several illustrations of *dvāra-koṭṭhakas* of palace or temple enclosures (Cunningham, *Bharhut*, Pls. XVIII, XIX); these are of a simpler type, but large enough to admit an elephant and rider, and here a pointed arch of the kind elsewhere so familiar in the early art, is a conspicuous feature of the structure.

c9.s4.55

I cannot satisfactorily explain the "high *patthaṇḍila*" with which a city *gopura* is provided, Mhv., LX, 3; the upper stories (*cf. tala* in KA) of the gate-house may be meant. *Cf.* Geiger, *Culavamsa*, trans. 1, 214; and MN, II, 155.

c9.s4.56

FIG. C. A SECOND TYPE OF CITY-GATE, REPRESENTED ONLY IN ONE RELIEF (SEE FIG. 6). SĀÑCĪ

c9.s4.57

Quite distinct from the *toraṇa* of the actual gateway are the freestanding *toraṇas* which are frequently represented as situated at the bridge-end remote from the city; through these arches one must pass when entering or leaving the city, but it is obvious that they served an ornamental and honorific purpose, and had no value for defense. In function they differ in no way from the *toraṇas* elsewhere represented, or extant as at Sāñcī, as set up at the entrance to any sacred or honorable area or enclosure; and no doubt because they are taken for granted in this sense, and because they have no defensive value, we do not find them referred to in the literature.

c9.s4.58

KA describes the gate-house of a fort (*durga*), substantially identical with that of a fortified city (for *durga-nagara cf. Āghāṭa-durgge* in the colophon of a Jāina manuscript described on p. 237); the nomenclature has already been made use of above. The passage is full of difficulties; but as these can be, at least in part, resolved by comparison with the passages cited already, it will be desirable to offer here a new rendering of the greater part of it.

First 155 of 1,041 passages. Read on

Cites (68)