Constructing Community: Tamil Merchant Temples in India and China, 850-1281 Risha Lee 

 

Submitted in partial fulfillment of the 

requirements for the degree of 

Doctor of Philosophy 

in the Graduate School of Arts and Sciences 

COLUMBIA UNIVERSITY 

2012

♥ 2012 

Risha Lee 

 

Constructing Community: Tamil Merchant Temples in India and China, 850-1281 Risha Lee 

This dissertation studies premodern temple architecture, freestanding sculpted stones, and Tamil language inscriptions patronized by south Indian merchants in south  India and China. Between the ninth and thirteenth centuries, Indian Ocean trade was at  its apex, connecting populations on European and Asian continents through complex  interlocking networks. Southern India’s Tamil region, in particular, has been described  as the fulcrum of the Indian Ocean circuit; however, knowledge of intra-Asian contact  and exchange from this period has been derived mostly from Arabic and Chinese sources,  which are abundant in comparison with the subcontinent’s dearth of written history. My  project redresses this lacuna by investigating the material culture of Tamil merchants, and  aims to recover their history through visual evidence, authored by individuals who left  few written traces of their voyages across the Indian Ocean. The arguments of my  dissertation are based primarily on unpublished and unstudied monuments and  inscriptions, weaving together threads from multiple disciplines—art history, literature,  epigraphy, and social theory—and from across cultures, the interconnected region of the  eastern Indian Ocean and the South China Seas, spanning the Sanskritic, Tamil, Malay,  and Sinocentric realms. 

My dissertation challenges traditional narratives of Indian art history that have  long attributed the majority of monumental architecture to royal patrons, focusing instead  on the artistic production of cosmopolitan merchants who navigated both elite and non-

elite realms of society. I argue that by constructing monuments throughout the Indian  Ocean trade circuit, merchants with ties to southern India’s Tamil region formulated a  coherent group identity in the absence of a central authority. Similar impulses also are  

visible in merchants’ literary production, illustrated through several newly translated  panegyric texts, which preface mercantile donations appearing on temple walls in the  modern states of Tamil Nadu, Karnataka, and Andhra Pradesh. Moreover, my work  analyzes the complex processes of translation visible in literary and material culture  

commissioned by merchants, resulting from inter-regional and intercultural encounters  among artisans, patrons, and local communities. 

Rather than identifying a monolithic source for merchants’ artistic innovations, in  each chapter I demonstrate the multiple ways in which merchants employed visual codes  from different social realms (courtly, mercantile, and agrarian) to create their built  environments. In Chapter Four, I provide a detailed reconstruction and historical  chronology of a late thirteenth century temple in Quanzhou, coastal Fujian Province,  southeastern China, which both echoes and transforms architectural forms of  contemporaneous temples in India’s Tamil region. Piecing together over 300 carvings  discovered in the region in light of archaeological and art historical evidence, I develop a  chronology of the temple’s history, and propose that Ming forces destroyed the temple  scarcely a century after its creation. In Chapter Three, I interpret stone temples  patronized by the largest south Indian merchant association, the Ainnurruvar, as being integral to their self-fashioning in India and abroad. While the temples do not project a  merchant identity per se, I show that they employ an artistic vocabulary deeply  entrenched in the visual language of the Tamil region. Chapter Two looks at other forms 

through which merchants created a shared mercantile culture, including literary  expressions and freestanding sculptural stones. These texts demonstrate that merchants  engaged in both elite and non-elite artistic production. Chapter One analyzes the  distribution, content, and context of Tamil merchant sponsored inscriptions within the  Indian Ocean circuit, focusing on the modern regions of Tamil Nadu, Karnataka, Kerala,  and Andhra Pradesh. An appendix offers new translations of important Tamil language  mercantile inscriptions discovered throughout south India.

TABLE OF CONTENTS 

List of Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ii List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

iii A Note on Translations and Transliterations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..

x Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 Chapter 1: Tamil Merchants In World History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Chapter 2: Of Symbols and Stones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

Chapter 3: Constructing Community in Stone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

Chapter 4: A Shiva Temple in Medieval Southern China. . . . . . . . . . . . . . . . . . . . . . . 134

Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .187

Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .206

Bibliography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297

Appendix 1: English Translations of Select Merchant Association Inscriptions . . . . 310

Appendix 2: Glossary of Architectural Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 323

LIST OF MAPS 

Map 1 Trade routes connecting India, Southeast Asia, and China, ca. 10-12th centuries (After Hoekveld-Meijer 1980) 

Map 2 Southern India Physical Features and Ports (After Sanjay Subrahmanyam  1990) 

Map 3 Indian Ocean Regional Networks (1300-1500); (After Kenneth Hall 2010) Map 4 Trading Ports and Cities in the Indian Ocean, 618-1500 (After Chaudhuri, 1985) 

Map 5 Tamil Nadu Districts, After Travel Corporation India,  www.tcindia.com/maps/tamilnadu.html.gif 

Map 6 Period 1 (849-985): Overview (Maps 6-17 After Google Maps) Map 7 Period 1: Tamil Nadu detail  

Map 8 Period 2 (985-1070): Overview 

Map 9 Period 2: Tamil Nadu detail 

Map 10 Period 3 (1070-1178): Overview 

Map 11 Period 3: Karnataka detail 

Map 12 Period 3: Tamil Nadu and Sri Lanka detail 

Map 13 Period 4 (1178-1279): Overview 

Map 14 Period 4: Tamil Nadu Detail 

Map 15 Period 4: Karnataka detail 

Map 16 Period 5 (1279-1400): Overview 

Map 17 Period 5: Karnataka and Tamil Nadu detail 

Map 18 South India (Tamil Nadu, Kerala, Karnataka, and Andhra Pradesh)  showing numbers of merchant organization inscriptions by district; color  added by the author (After Karashima 2002) 

Map 19 Merchant stone map (After Google Maps) 

Map 20 Temple locations mentioned in Chapter 3: “Constructing Community in  Stone”  

Map 21 Quanzhou trade routes c.1200 (After Hugh Clark 1995)

ii 

LIST OF FIGURES 

Fig. 0.1 View from southeast of Rajarajesvara Temple, Tanjavur, Tamil Nadu Fig. 0.2 Puspapodigai at the Nakhon si Thammarat Museum, Thailand (Photo courtesy of Narayana Moorthy) 

Fig. 0.3 Kochchenganan Sculptural Panel, Quanzhou Maritime Museum (QMM) Fig. 0.4 Kochchenganan narrative at Tiruvanaikka in Tiruchirappalli District Tamil Nadu (Photo courtesy of Katherine Kasdorf) 

Fig. 1.1 Terracotta sherd from Alagankulam with incised ship, ca. 1st century BCE (After Iravatham Mahadevan 2003) 

Fig. 1.2 Alagarmalai cave entrance, view of ceiling (After Mahadevan 2003) Fig. 2.1 Tiruvalangadu Copper plates of Rajendra Chola, ca. 1012-1044 CE (After The Hindu, last modified December 25, 2009) 

Fig. 2.2 Detail of seal, Tiruvalangadu Copper plates (After The Hindu, last  modified December 25, 2009) 

Fig. 2.3 View of eastern entrance through mahāmaṇḍapa, Shiva Temple in  Pettavaittalai, Kulittalai Taluk, Tiruchirappalli District 

Fig. 2.4 Virakal, Nagainallur, Karur Dt, View of warrior holding sword  Fig. 2.5 Virakal, Nagainallur, View of warrior with sword and shield Fig. 2.6 Merchant stone, Elkamvalasu, Sri Lanka (After Rajagopal 2002) Fig. 2.7 Nagainallur merchant stone, Tiruchirappalli Dt. 

Fig. 2.8 Nagainallur rubbing (After Rajagopal 2002) 

Fig. 2.9 Kaliyampatti merchant stone rubbing (After Rajagopal 2002) Fig. 2.10 Merchant stone at Melnangavaram, Karur District, Tamil Nadu Fig. 2.11 Detail of unidentified object framed by lamps and poles, Melnangavaram Fig. 2.12 Detail of sickle, billhook, and whip, Melnangavaram Fig. 2.13 Detail of sword, Melnangavaram 

Fig. 2.14 View of Korravai, Victorious Durga, Melnangavaram Fig. 2.15 Standing Durga, Tirumandisvaram temple, Gramam, Villupurum Dt. Fig. 2.16 Stone with Elephant Head, Jakarta National Museum (Photo courtesy of  Narayana Moorthy with background modification by the author) 

Fig. 2.17 Boundary Stone at Hampi, Hampi site museum (After Kotriash 1977) Fig. 3.1 “The Vimana,” (After Barrett 1974) 

Fig. 3.2 Western view of Airavatesvara’s vimana at Darasuram, Kumbakonam   Tk., Tanjavur District 

Fig. 3.3 View from East of Adipurisvara temple’s stone superstructure,  Tiruvorriyur, Tiruvallur District (Photo courtesy of the French Institute of  Pondicherry) 

Fig. 3.4 View from southwest of Adipurisvara’s superstructure (Photo courtesy of the French Institute of Pondicherry) 

Fig. 3.5 View from southwest of vimāna and ardhamaṇḍapa with projected  shrines, Manavalesvara temple, Tiruvelvikkudi, Tanjavur District 

Fig. 3.6 Detail of Manavalesvara temple’s adhiṣṭhāna, Tiruvelvikkudi  (Photo courtesy of Katherine Kasdorf) 

Fig. 3.7 Detail of rider on vyāla mount, Manavalesvara temple, Tiruvelvikkudi,  Tanjavur District

iii 

Fig. 3.8 View of southern wall, Manavalesvara temple, Tiruvelvikkudi. Tanjavur  District 

Fig. 3.9 Detail of Ganesha, Manavalesvara temple, Tiruvelvikkudi, Tanjavur  District 

Fig. 3.10 Detail of Nataraja, Manavalesvara temple, Tiruvelvikkudi, Tanjavur  District 

Fig. 3.11 Detail of Agastya, Manavalesvara temple, Tiruvelvikkudi, Tanjavur  District  

Fig. 3.12 Detail of Donor Portrait, Manavalesvara temple, Tiruvelvikkudi, Tanjavur  District  

Fig. 3.13 Detail of Dakshinamurti, Manavalesvara temple, Tiruvelvikkudi, Tanjavur  District  

Fig. 3.14 Detail of Vishnu, Manavalesvara temple, Tiruvelvikkudi, Tanjavur  District  

Fig. 3.15 Detail of Lingodbhava, Manavalesvara temple, Tiruvelvikkudi, Tanjavur  District  

Fig. 3.16 Detail of Ardhanarisvara, Manavalesvara temple, Tiruvelvikkudi,  Tanjavur District  

Fig. 3.17 Detail of Brahma, Manavalesvara temple, Tiruvelvikkudi, Tanjavur  District  

Fig. 3.18 Detail of Shiva and Chandesamurti, Manavalesvara temple,  Tiruvelvikkudi, Tanjavur District  

Fig. 3.19 Detail of Bhikshatana, Manavalesvara temple, Tiruvelvikkudi, Tanjavur  District 

Fig. 3.20 View of southern ardhamaṇḍapa wall, Sakisvara temple,  Tiruppurumubiyam, Tanjavur District 

Fig. 3.21 View of vimāna from west, Vatamulesvara, Kizhpalavur, Ariyalur Dt., TN Fig. 3.22 Detail of pañjara, Vatamulesvara, Kizhpalavur, Ariyalur Dt.  Fig. 3.23 Detail of Nataraja, Vatamulesvara, Kizhpalavur, Ariyalur Dt.  Fig. 3.24 Detail of Kalyanasundarar, Vatamulesvara, Kizhpalavur, Ariyalur Dt.  Fig. 3.25 Detail of Kalyanasundarar image on tōrana over Nataraja niche at  Manavalesvara, Tiruvelvikkudi, Tanjavur District 

Fig. 3.26 Southeast view of Ramasvamy temple at Cheranmadevi, Tirunelveli Dt. Fig. 3.27 Western view of Ramsvamy temple, Cheranmadevi, Tirunelveli Dt. Fig. 3.28 Detail of Ramasvamy temple nāsikās containing dance positions,  Cheranmadevi, Tirunelveli Dt. 

Fig. 3.29 View of Ramasvamy temple’s brick superstructure, Cheranmadevi,  Tirunelveli Dt. 

Fig. 3.30 View of brick superstructure’s central niche containing a seated Vishnu  holding conch and discus at Ramasvamy temple, Cheranmadevi,  

Tirunelveli District 

Fig. 3.31 Northwest view of Devasvamudaiyar temple’s Amman temple,  Cheranmadevi, Tirunelveli District 

Fig. 3.32 View from southeast of Devasvamiyudaiyar’s Shiva temple,  Cheranmadevi, Tirunelveli Dt. 

Fig. 3.33 View from west of Tiruvalisvarar temple, Tiruvalisvaram, Tirunelveli Dt.

iv 

Fig. 3.34 Northwest view of Mudigondesvara vimana, Mudigonda, Kollegal District,  Karnataka 

Fig. 3.35 Detail of Durga, Mudigonda 

Fig. 3.36 Detail of Bhikshatana, Mudigonda 

Fig. 3.37 Detail of Vishnu, Mudigonda 

Fig. 3.38 Detail of Brahma, Mudigonda 

Fig. 3.39 Racigudi, Aihole, Karnataka, (After Sinha 2000) 

Fig. 3.40 Ardhamaṇḍapa of Tiruvalisvarar, Valikandapuram Village, Ariyalur  District 

Fig. 3.41 Devakōṣṭha at Tiruvalisvarar 

Fig. 3.42 Mahamandapa at Tiruvalisvarar, Valikandapuram 

Fig. 3.43 Composite citrakhaṇḍa pillar (with warrior seated astride rearing lion),  Tiruvalisvarar, Valikandapuram 

Fig. 3.44 Warrior astride rearing lion, ca. mid-12th century, detail of eastern gōpura at Somanatharsvami, Palaiyarai, Tanjavur District 

Fig. 3.45 Detail of Hanuman worshipping a linga on citrakhaṇḍa pillar Fig. 3.46 Detail of Rama and Sita on citrakhaṇḍa pillar, Tiruvalisvarar Fig. 3.47 Durga astride a lion, Tiruvalisvarar 

Fig. 3.48 Detail of Aiyanar on citrakhaṇḍa pillar, Tiruvalisvarar Fig. 3.49 Temple tank, Tiruvalisvarar, Valikandapuram 

Fig. 3.50 Detail of tank’s citrakhaṇḍa pillars, Tiruvalisvarar, Valikandapuram Fig. 3.51 View from east of Rajarajacholisvaramudaiyar temple at Koyilpatti,  Pudukkottai Dt. 

Fig. 3.52 Detail of nandis at gateway, Rajarajacholisvaramudaiyar, Koyilpatti,  Pudukkottai District 

Fig. 3.53 View from northwest of Rajarajachoisvaramudaiyar, Koyilpatti,  Pudukkottai District 

Fig. 3.54 View of adhiṣṭhāna on southern side of Rajarajachoisvaramudaiyar,  Koyilpatti, Pudukkottai District 

Fig. 3.55 Detail of kaṇṭha relief showing Karaikkal Ammaiyar, Koyilpatti,  Pudukkottai District 

Fig. 3.56 Detail of kaṇṭha containing Karaikkal Ammaiyar relief, Airavatesvara  temple, Darasuram, Tanjavur District 

Fig. 3.57 View from east of southern shrine at Muvarkoyil, Kodumbalur,  Pudukkottai District 

Fig. 3.58 Wooden temple chariot from Tiruppurumbiyam, Tanjavur District Fig. 3.60 Rearing Horse on ardhamaṇḍapa, Amritagatesvarar temple,  Melakkadambur, Cuddalore District, Tamil Nadu 

Fig. 3.61 Rearing horse on front porch at Somanatesvara, Pazhaiyarai, Tanjavur  District, Tamil Nadu 

Fig. 3.62 View of adhiṣṭhāna and wall from Kampaharesvarar temple at  Tribhuvanam, c. 1212, Tanjavur District, Tamil Nadu 

Fig. 3.63 Citrakhaṇḍa pillars in the mahamandapa of the Airavatesvara temple at  Darasuram, Tamil Nadu 

Fig. 3.64 Citrakhaṇḍa with images of dancing Shiva and Karaikkal Ammaiyar,  Airavatesvara temple at Darasuram

v 

Fig. 4.1 “Tomb of the Foreigner” (Fan ke mu); Quanzhou Maritime Museum Fig. 4.2 Nagapattinam “Pagoda” (After Seshadri 2009) 

Fig. 4.3 View of eastern “pagoda” at Kaiyuan temple, Quanzhou Fig. 4.4 View of front porch showing vyāla bas-reliefs on plinth   Kaiyuan temple, Quanzhou 

Fig. 4.5 Citrakhaṇḍa pillars framing back entrance of main hall Kaiyuan Temple, Quanzhou 

Fig. 4.6 Detail of citrakhaṇḍa pillar showing medallions, Kaiyuan Temple,  Quanzhou 

Fig. 4.7 Indic carvings displayed at Quanzhou Maritime Museum, Quanzhou Fig. 4.8 Indic carvings displayed at QMM, Quanzhou 

Fig. 4.9 View of citrakhaṇḍa pillars at Tianhou Gong temple 

Quanzhou 

Fig. 4.10 Kali sculptural panel (worshipped as Guanyin), Chidian pavilion, Xinjiang  County, Quanzhou 

Fig. 4.11 Bilingual Tamil-Chinese Inscription, c. 1281, Xiamen University Museum Fig. 4.12 Map delineating major architectural features of Old Quanzhou City (After Richard Pearson 2002) 

Fig. 4.13 Diagram showing Drāvida temple’s construction method (After Hoekveld Meijer 1980) 

Fig. 4.14 Detail of column band with scrolling peony and lotus motif, Kaiyuan  Temple 

Fig. 4.15 Detail of Christian gravestone with scrolling peony and lotus motif, QMM Fig. 4.16 Colossal Daoist deity, Qingyuanshan slope, Quanzhou Fig. 4.17 Detail of Daoist deity armrest showing cloud motif 

Fig. 4.18 Indic carving with cloud motif, QMM 

Fig. 4.19 Detail of gravestone with Arabic script and cloud motif, c. 1302, QMM Fig. 4.20 Detail of vyāla reliefs from Kaiyuan Temple’s plinth, Quanzhou Fig. 4.21 Detail of vyālamāla from Tiruvalisvarar at Tiruvalisvaram, Tirunelveli  District, Tamil Nadu 

Fig. 4.22 Detail of vyālamāla in Shiva temple at Agara, Chamarajnagar Dt.,  Karnataka 

Fig. 4.23 Padma jagatī in QMM, Quanzhou 

Fig. 4.24 View of jagatī reused at Kaiyuan Temple, Quanzhou Fig. 4.25 Kumuda at QMM, Quanzhou 

Fig. 4.26 Paṭṭika at QMM, Quanzhou 

Fig. 4.27 Kaṇṭha reused in Quanzhou building (After Wu 2005) Fig. 4.28 Detail of kaṇṭha from Ramasvamy temple, Cheranmadevi, Tirunelveli Dt. Tamil Nadu 

Fig. 4.29 Tirumular panel in Quanzhou (currently at National Palace Museum of  Beijing; After Wu 2005)

vi 

Fig. 4.30 Tirumular panel at Apatsahayesvara temple in Tiruppalanam, Tanjavur Dt.,  Tamil Nadu 

Fig. 4.31 Tirumular panel at Shiva temple in Tiruttandanapuram, Sivaganga Dt. Tamil Nadu 

Fig. 4.32 Kapōta (for dēvakōṣṭha), QMM, Quanzhou 

Fig. 4.33 Guardian figure, ca. 11th century at Anping Bridge, Quanzhou Fig. 4.34 Tōrana with “donor portraits” at QMM, Quanzhou 

Fig. 4.35 Donor portraits at Umamahesvara temple, Konerirajapuram, Tanjavur  District, Tamil Nadu 

Fig. 4.36 Memorial structure on Kaiyuan Temple grounds, Quanzhou Fig. 4.37 Detail of Devasvamudaiyar dēvakōṣṭha, Cheranmadevi, Tirunelveli Fig. 4.38 Ōma at QMM, Quanzhou 

Fig. 4.39 Mālāsthāna at QMM, Quanzhou 

Fig. 4.40 Laśuna at QMM, Quanzhou 

Fig. 4.41 Ghaṭa at QMM, Quanzhou 

Fig. 4.42 Puṣpapōtigai (type 4), QMM, Quanzhou 

Fig. 4.43 Puṣpapōtigai (type 1), QMM, Quanzhou 

Fig. 4.44 Puṣpapōtigai (type 2), QMM, Quanzhou 

Fig. 4.45 Puṣpapōtigai (type 3), QMM, Quanzhou 

Fig. 4.46 Puṣpapōtigai (type 5), QMM, Quanzhou 

Fig. 4.47 Profile of kapōta (Type 1) at Kaiyuan Temple Site Museum, Quanzhou Fig. 4.48 Kapōta (Type 1) at QMM, Quanzhou 

Fig. 4.49 View of iwan with muqarnas at Shengyou si (Ashab) mosque, Quanzhou Fig. 4.50 Detail of muqarnas at Shengyou si, Quanzhou 

Fig. 4.51 Profile of kapōta (Type 2) at QMM, Quanzhou 

Fig. 4.52 “L-shaped face component” comprising parapet at QMM, Quanzhou Fig. 4.53 Stūpikā at QMM, Quanzhou 

Fig. 4.54 Nandi in Quanzhou (Current location unknown; After Wu 2005) Fig. 4.55 Hastihasta fragment at QMM, Quanzhou 

Fig. 4.56 Rearing Lion (Current location unkown; After Wu 2005) Fig. 4.57 Jāla fragment at QMM, Quanzhou 

Fig. 4.58 Door jamb with monkey, Xiamen University Museum, Xiamen Fig. 4.59 Door jamb with devotee, Xiamen Univerity Museum, Xiamen Fig. 4.60 Detail of door jamb with devotee, Xiamen University Museum, Xiamen Fig. 4.61 Detail of Medallion, Narasimha, Kaiyuan Temple, Quanzhou Fig. 4.62 Detail of Medallion, Krishna Felling Arjuna Tree, Kaiyuan Temple,  Quanzhou 

Fig. 4.63 Detail of Medallion, Deer Eating Fungus with Xian in background,  Kaiyuan Temple, Quanzhou 

Fig. 4.64 Detail of Medallion, Chasing Phoenix, Kaiyuan Temple, Quanzhou Fig. 4.65 Detail of Medallion, Krishna Playing Flute, Kaiyuan Temple, Quanzhou Fig. 4.66 Detail of Medallion, Krishna with Gopis, Kaiyuan Temple, Quanzhou Fig. 4.67 Detail of Medallion, Wrestlers, Kaiyuan Temple, Quanzhou Fig. 4.68 Detail of Medallion, Vishnu seated with wives (Lakshmi and Bhudevi),  Kaiyuan Temple, Quanzhou

vii 

Fig. 4.69 Detail of Medallion, Vishnu Mounted on Garuda, Kaiyuan Temple,  Quanzhou 

Fig. 4.70 Detail of Medallion, Gajendramoksha, Kaiyuan Temple, Quanzhou Fig. 4.71 Detail of Medallion, Bhairava, Kaiyuan Temple, Quanzhou Fig. 4.72 Detail of Medallion, Lions Chasing Ball and Ribbon, Kaiyuan Temple,  Quanzhou 

Fig. 4.73 Silver box with chasing Phoenix motif, ca. 1127-1279, Fuzhou Provincial  Museum 

Fig. 4.74 Wrestlers motif on sculptural panel from Darasuram, Tanjavur District Tamil Nadu 

Fig. 4.75 Detail of crocodile in medallion containing Gajendramoksha purana Fig. 4.76 Detail of Narasimha panel from a shrine at the Somanatharsvami temple,  Palaiyarai, Kumbakonam Tk., Tanjavur District 

Fig. 4.77 “Dance mandapa” Shrine at Somanatharsvami temple, Palaiyarai Fig. 4.78 Detail of Garuda, Sivayoganatasvami, Tiruvisanallur Fig. 4.79 Nandi pavilion at Sivayoganatasvami temple, Tiruvisanallur village,   Kumbakonam Tk., Tanjavur District 

Fig. 4.80 Vishnu, ca. 925. Kongu Nadu Bronze. After Dehejia, The Sensuous and  the Sacred, 177. 

Fig. 4.81 Detail of pillar with representation of Shiva and attendant (probably  Karaikkal Ammaiyar) dancing on corpses. Sivakozhutanesvara temple,  Tirusattimutram, Kumbakonam Taluk, Tamil Nadu. 

Fig. 4.82 View of Quanzhou Shiva temple, reconstruction of central niche (bhadra śala). Artist: Mike Lee 

Fig. 4.83 View of Quanzhou Shiva temple from garbagṛha looking to  mukhamaṇḍapa. Artist: Mike Lee 

Fig. 4.84 Detail of medallion containing wrestlers from fragmentary citrakhaṇḍa  pillar in QMM, Quanzhou 

Fig. 4.85 “Stone Bamboo Shoot,” Quanzhou

viii 

A Note on Translation and Transliterations 

I have tried to keep diacriticals to a minimum, with the exception of Sanskritic and Tamil  architectural terminology. Without diacritics, these important terms would make little  sense; I have included an architectural glossary for further clarification, following the  Encylopaedia of Indian Temple Architecture’s. When terms have been transliterated  from Tamil, Sanskrit, or Chinese, I have followed the diacritical systems of Epigraphia  Indica and pinyin. Plurals of these terms employ the English convention of adding s.  Unless otherwise stated, all translations are my own.

ix 

ABBREVIATIONS 

ARE = Annual Reports on Epigraphy 

AMCA = Ancient and Medieval Commercial Activities in the Indian Ocean EITA = Encyclopaedia of Indian Temple Architecture: Lower Dravidadesa IPS = Inscriptions of Pudukkottai State 

SII = South Indian Inscriptions

x 

Acknowledgements 

Studying several countries at once has many benefits, most especially meeting  and learning from extraordinary people with diverse interests. I am indebted to the  following individuals and institutions for their instruction and encouragement.  

I was able to research and write this dissertation with the generous funding of  several institutions: Columbia University’s Department of Art History and Archaeology  funded my first year of research and two additional Mandarin language programs.  Columbia’s Weatherhead East Asian Institute provided me with a Junior Research  Fellowship for a summer of exploratory work as well as the yearlong Julie How  Memorial Fellowship. I also benefited from a Junior Fellowship awarded by the  American Institute of Indian Studies.  

Columbia University provided an ideal environment for researching and writing  this project. Vidya Dehejia offered countless words of wisdom, on topics ranging from  etiquette with Brahmin priests to identifying the sinuous curve of a Chola bronze. I am  indebted to her for many years of intellectual growth. Robert Harrist and his family  accompanied me on an inspiring first trip to Quanzhou, encouraging me to pursue work  in China and planting the initial seeds for love of southern cuisine. Conversations with  Rachel McDermott always steered my thinking in provocative new directions. I owe  Katherine Kasdorf special thanks: trips together around the Karnataka and Tamil regions  were productive and fun. Many photographs in this dissertation attest to her  photographic talent. Laura Weinstein has been a steady source of support and insight; Anna Seastrand sent irretrievable copies of Tamil periodicals from Pondicherry and 

xi 

commented on drafts. Many arguments would have been lost in translation if not for Sam  Sudanandha’s help in reading Tamil epigraphy and Xue Lei and Tsai Chun-yi’s  boundless generosity in translating Chinese texts. In seminars and conferences, Dipti  Khera, Yuthika Sharma, Neeraja Poddar, and Subhashini Kaligotla fostered a productive  atmosphere for rigorous rethinking of material. 

I am grateful to many generous scholars who provided guidance throughout the  research and writing stages. Rick Asher’s enthusiasm for the subject matter and  constructive criticism were major sources of inspiration: I could not have completed this dissertation without his encouragement. Barry Flood’s exciting contributions to the field  and insightful comments compelled me to think about the material in novel ways.  Tansen Sen, Hugh Clark, Padma Kaimal, Phillip Wagoner, Rohit Goel, and Emmanuel  Francis all offered their input at important junctures of its conception.  

In China, I could not have asked for kinder hosts. Ding Yuling and Wang  Lianmao, director and former director of the Quanzhou Maritime Museum, gave full  access to their collection and more, treating me to delicious meals and tours of the  countryside. Daily chats over freshly brewed tea with Wang Lianmao deepened my  understanding of Quanzhou’s rich history, while his assistance in providing local contacts  proved invaluable to my research. The Buddhist monks at the Kaiyuan temple allowed  me the freedom to photograph and measure its architecture. I also thank Lan Dajiu for  helping me to access materials at the Xiamen University Museum.  

My time in India would not have been nearly as productive without the assistance  of the French Institute of Pondicherry. As a visiting scholar there I benefited from its  helpful staff, photo archives, and library on a regular basis. Kannan M. deserves 

xii 

particular note for his generosity and insight. I was truly fortunate to work with Dr. Y.  Subbarayalu, whose astonishing knowledge of epigraphical material was matched only  by his generosity in sharing it with me. The esteemed scholar, Dr. R. Nagaswamy,  devoted hours to teaching me the requisite scripts and structures of medieval Tamil  epigraphy, some lessons occurring at the temples themselves. At the École française  d’Extême-Orient, Valerie Gillet and Dominique Goodall generously shared their photos of temple architecture and sculpture and allowed reproduction of the EFEO’s beautiful  plans and maps. K. Ramachandran of Nagapattinam guided me through his city with  courtesy and enthusiasm. Krishna Devanandan enriched research trips with her quick wit  and sharp eyes. The Adishakti Theater Company and staff, especially Veenapani Chawla  and Arvind Rane, made Pondicherry a second home.  

Thanks are owed to the Department of Architecture and Design at the American  University of Beirut, where I was a research affiliate while writing. 

Most importantly, I owe thanks to my loving family. None of my scholastic  endeavors would have been possible without the support of my parents, Tom and Ida  Jean Lee, who cultivated a love of learning from an early age, and my brother, Mike Lee,  who used his immense artistic talent to make drawings that brought my arguments to life. 

xiii 

1 

INTRODUCTION 

Between the ninth and fourteenth centuries, south Indian trade networks expanded  dramatically, both internally and across the Indian Ocean. Merchants, many of whom  came from the Tamil region in India’s far south, transported merchandise to all corners of  the Asian world via the Indian Ocean. These included luxury items such as spices, horses,  

woven cloth, pearls and gems as well as everyday items such as rice and oil.1 Movement  was not limited to inanimate objects, for trade also disseminated people and ideas across  vast spaces. These exchanges created a world comprised of tightly knit, intercultural  spheres, which many scholars have termed the ‘world system.’ The world system’s apex  was reached in the thirteenth century, when, as Janet Abu-Lughod describes, “there was  an efflorescence of cultural and artistic achievement. Never before had so many parts of  the Old World simultaneously reached cultural maturity.”2 

The coastal port city of Quanzhou, in Fujian province in southern China, attests to  this point. While Chinese records prove that it operated as an international trading  entrepot from as early as the ninth century, it was dwarfed in importance by Guangzhou, a port city located approximately 450 miles to its southwest, until the mid-eleventh century.3 The influx of foreign traders into Quanzhou steadily increased from this time  onwards, with an explosion of activity at the end of the thirteenth century, when a  

 

1For a comprehensive list of trade goods, see Meera Abraham, Two Medieval Merchant Guilds of South  India (New Delhi: Manohar Publications, 1988). 

2 Janet L. Abu-Lughod, Before European Hegemony: The World System A.D. 1250-1350, (New York:  Oxford University Press, 1989), 4. 

3 Guangzhou appears in several Arabic and Persian language travelogues, for instance, the early tenth century Arab geographer, Ibn Khordadbeh’s Treatise of Roads and Provinces (Kitab al-masalik wa’l mamalik) and Abu Zayd’s late ninth century text, An Account of China and India (Ahbar al-sin wa l-hind), trans. Jean Sauvaget, (Paris: Société d’édition les belles letters, 1948). See Chapter Four for a full  discussion of the Chinese historical context.

2 

nomadic tribe of ethnic Mongols foreign to China founded the Yuan dynasty (1271-1368).  Owing to their own foreign status, the Mongols’ policies included granting special  trading privileges to foreigners, which made Quanzhou an even more desirable location  for foreign trade. During this time, it was the most cosmopolitan city in the world and  the crossroads of many non-ethnic Chinese, including Indians, Arabs, Persians, Mongols,  Southeast Asians, Syrians, Armenians, and Italians. Eminent visitors such as Marco Polo  visited the city. According to Marco Polo, 

[Quanzhou is] frequented by all the ships of India, which bring thither  spicery and all other kinds of costly wares . . . hither is imported the most  astonishing quantity of goods and precious stones and pearls . . . And I  assure you that for one shipload of pepper that goes to Alexandria or  elsewhere, destined for Christendom, there come a hundred such, aye and  more too, to this haven of Zayton;4 for it is one of the two greatest havens  in the world for commerce.5  

Later, Polo specifically mentions his encounters with Indian merchants writing, “these  Brahmans are the best merchants in the world, and the most truthful, for they would not  tell a lie for anything on earth.”6 Polo’s high praise resulted from encounters with Indian  

 

4 Zayton, a medieval name for Quanzhou, derives from the Arabic word for olive, zaytun. 

5 Marco Polo, Book of Ser Marco Polo the Venetian: Concerning the Kingdoms and Marvels of the East,  Ed. George B. Parks, (New York: The Macmillan Company, 1927), 241. 

Frances Wood’s popular book, Did Marco Polo Go to China?, (Boulder, Colorado: Westview  Press, 1998) disputes the authenticity of Polo’s accounts of China. Several Mongol period scholars, such as Morris Rossabi, have disagreed with her arguments, faulting the author for her “lack of expertise,” resulting  in “misrepresentations and mistakes.” The current scholarly consensus is that Marco Polo did indeed go to  China. See Rossabi, “Did Marco Polo Really Go to China?,”  

http://afe.easia.columbia.edu/mongols/pop/polo/mp_essay.htm. 

6 Polo, Book of Ser Marco Polo, 285.

3 

merchants during his travels, informing us of the numerous Indian trade diaspora  communities living in India, Southeast Asia, and China at this time. The Indian presence  in Quanzhou is further substantiated by the ruins of a temple devoted to the Hindu god  Shiva now primarily housed in the Quanzhou Maritime Museum and reused as spolia in  the city’s oldest Buddhist temple. The ruins contain several sculptural representations of  Shaivite and Vaishnavite (Shiva and Vishnu) iconography, first noted in 1933 by Ananda  Coomaraswamy.7 Subsequent writings have elaborated on these motifs and convincingly  linked the artistic style to the Tamil region of south India; however, they have not  considered the majority of extant architectural carvings, which, while lacking in  figurative imagery, enable us to reconstruct a substantial portion of the temple. 

On the one hand, this dissertation focuses on the Shiva temple in Quanzhou,  compiling and analyzing all available evidence regarding its construction. I consider the  over 300 extant temple carvings, archaeological evidence, and written texts that reference Indians living in Quanzhou to reconstruct the temple’s history. On the other hand, the  majority of the dissertation seeks to frame the Quanzhou temple within its wider context,  as a product of Tamil merchants’ patronage. In the medieval period, Tamil merchants,  hailing from regions that comprise modern Tamil Nadu, southern Karnataka, southern  Andhra Pradesh, and Kerala (which I shall refer to henceforth as south India), established  permanent communities along the Indian Ocean route that connected India to China,  leaving behind Tamil language stone inscriptions and south Indian style architectural  carvings. Within south India as well, they actively patronized temples and other  institutions, testifying to their frequently powerful positions within society. My  

 

7 Ananda Coomaraswamy, “Hindu Sculptures at Zayton,” Ostasiatsche Zeitschrift 9 (1933), 5-11.

4 

dissertation asks if, how, and why we might see these communities in relation to one  another.  

 This dissertation unites a variety of primary sources, produced by Tamil  merchants and germane to a discussion of their wider role in the premodern Indian Ocean  world. These include stone monuments (temples and sculptural freestanding stones),  primarily in the modern Tamil region and the city of Quanzhou, stone-carved inscriptions  from the same locations, and references to merchants in contemporaneous literary texts.  The arguments of my dissertation are based primarily on unpublished and unstudied  material, weaving together threads from multiple disciplines—art history, literature,  epigraphy, and social theory—and from across cultures, the interconnected region of the  eastern Indian Ocean and the South China Seas (Map 1). Rather than reinforce the notion  that premodern India was static and unchanging, an assumption prevalent in colonial era  studies of the subcontinent, the material culture examined in this dissertation supports the  concept of an interconnected world, where people negotiated local, regional, and global identities on a daily basis. Tamil merchants were but one component of a truly  heterogeneous mercantile cosmopolis that expanded dramatically in the ninth century,  uniting multiple ethnic, political, and religious groups in interlocking trade networks.  These communities comprised a significant portion of the population at port cities throughout the Indian Ocean circuit, as many recent studies have shown.8  

The Quanzhou temple’s existence can be explained only through understanding  the Indian Ocean’s enormous role in shaping south Indian history. The Indian Ocean  

 

8 For example see Elizabeth Lambourn, “India from Aden: Khuṭba and Muslim Urban Networks in Late  13th century India,” in Secondary Cities and Urban Networking in the Indian Ocean Realm, c. 1400-1800,  ed. Kenneth R. Hall, (Lanham: Lexington Books, 2008), 55-97.

5 

covers about 27 percent of the world’s total maritime space and has one of the longest  histories of sea traffic connecting Mediterranean, Arabic, Indic, Malay, and Sinocentric  realms, which some estimate to have lasted around six millennia.9 Individuals living in  port cities had frequent contact with multiple cultural groups. In turn, the flow of goods  and information to port cities affected the development of the hinterland. The ocean’s  inherently intercultural nature and importance as a conduit that connected different  cultural zones is a concept only now receiving attention in studies of premodern South  Asian art history. Studies focusing on art representative of intercultural Islamic exchange  have employed similar theoretical constructs, seeing transregional identity and cross cultural exchange as intrinsic to human experience. Though focused on Tamil merchant  patrons, this study aims to contribute to these intellectual projects.  

Merchant Patrons 

Before continuing, I should mention that no medieval record explicitly states that  Tamil-speaking merchants constructed a Shiva temple in Quanzhou. Rather, the temple’s  history can be inferred through secondary evidence—the absence of a royal name in an  extant bilingual Tamil-Chinese inscription at Quanzhou; traveler’s diaries (such as those  of Marco Polo and Ibn Battuta) recording the presence of Indian merchants at all ports of  trade in the Indian Ocean; and, most importantly, the abundant material evidence  attesting to active pathways of mercantile traffic across South India and Southeast Asia.  

 

9 Michael Pearson writes that “relatively routine and organized trade using the Indian Ocean as a highway”  began with the “rise of early civilizations in the Tigris-Euphrates area, and in northwest India.” Michael  Pearson, The Indian Ocean, Seas in History (New York: Routledge, 2003), 49.

6 

We can study merchants as a distinct group because we know that they frequently  thought of themselves as such. Merchants comprised a distinct class of non-royal  individuals, whose profession required intense mobility and continual negotiation  between local and cosmopolitan identities. Here, I use cosmopolitan as a term that  describes, as Enseng Ho elegantly articulates, “persons who, while embedded in local  relations, also maintain connections with distant places [and] thus articulate a relation  between different geographical scales.” 10 Several studies highlight the merchant’s  integral role in developing “supralocal systems of exchange that cut across (while not  necessarily transcending) ethnic, linguistic, political and religious boundaries.”11 In this  way, many of this study’s actors were transregional—a fact expressed by the period’s  merchant-patronized literary and material culture. 

In hundreds of inscriptions, merchants identify themselves according to  profession, the most common titles being nagarattar, vyabari, vaniyar, and chetti. There  was a huge range of professional differentiation, with merchants specializing in the sale  of specific goods at various locales, including local and regional markets, and periodic  fairs. Some merchant titles refer directly to their specific merchandise, e.g. katriban =  betel leaf merchant, while others were more abstract, designating membership in  supralocal merchant organizations.12 

 

10 Enseng Ho, The Graves of Tarim: Genealogy and Mobility across the Indian Ocean, The California  World History Library 3 (Berkeley: University of California Press, 2006), 31. 

11 Finbarr Flood, Objects of Translation: Material Culture and Medieval Hindu-Muslim Encounter,  (Princeton: Princeton University Press, 2009), 16. 

12 Ancient and Medieval Activities in the Indian Ocean (hereafter AMCA) contains the most complete list of  merchant (individual and organizational) titles, listing over a hundred constituent groups. In N. Karashima  and Y. Subbarayalu’s “Ainnurruvar: A Supra-local Organization of South Indian and Sri Lankan  Merchants,” Ancient and Medieval Commercial Activities in the Indian Ocean: Testimony of Inscriptions 

7 

These professional identities were far from stable. Not only did merchant groups  sometimes distinguish themselves from others in order to gain status, as we see in the  more contemporary example of the Natukkottai Chettis,13 they constantly competed with,  incorporated, and subsumed one another, so that one might list twenty other organizations  and/or professional classes as its subsidiary agents. Mercantile associations were  structurally complex, characterized by constant change according to time and locality.  Indeed, our understanding of these groups is still murky. However, as this dissertation  demonstrates, extraordinary mobility and transregional identity were defining group  characteristics. 

Because the spectrum of mercantile actors in South India is so vast, it is necessary  to focus on a specific group in order to make a convincing comparative study of the  patrons of the Quanzhou Shiva temple. Luckily, one organization presents itself as the  most obvious parallel: the Ainnurruvar,14 roughly translated as ‘The Five Hundred 

Members,’ and known in Kannada texts as the Ayyavole Five Hundred. Although other  merchant organizations proliferated during this period, none were as successful in terms  of geographic penetration and historical longevity. Ainnurruvar inscriptions, dating from  between the ninth and seventeenth centuries, have been discovered throughout south  India and as far away as Sumatra. Many proclaim the organization’s infinite origins with  the following refrain: “we are the Five Hundred of the Thousand Directions of the four  

quadrants of the world (Nānādēsi ticai āyirataiñnūrruvar),” from which it is apparent   

and Ceramic-Sherds: Report of the Taisho University Research Project, 1997-2002. (Tokyo: Taisho  University, 2002), 76-82. 

13 See David West Rudner, Caste and Capitalism in Colonial India : The Nattukottai Chettiars (Berkeley:  University of California Press, 1994). 

14 Pronounced I-noo-troo-var.

8 

that their transregionalism (both real and imagined) was integral to group identity and  cultural worldview.15 Although we will never know if its members participated in  constructing the Quanzhou temple, the organization’s literal and imagined geographical  and ideological expansiveness strongly suggests that it would have impacted the policies  and behavior of contemporaneous merchant groups.  

The Ainnurruvar’s architectural commissions and associated inscriptions form the  skeleton of the current study. The association has been studied in several works, most  notably Meera Abraham’s Two Medieval Merchant Guilds of South India, Kenneth Hall’s  Trade and Statecraft in the Age of the Colas, and most recently, the edited volume  Ancient and Medieval Commercial Activities in the Indian Ocean. The current study is  indebted to these groundbreaking works, which have collected, translated, and analyzed a  huge amount of empirical data pertaining to the Ainnurruvar, which I have used as a  literal and virtual roadmap. 

The Ainnurruvar’s composition and associated regions fluctuated greatly, with the  greatest concentration of inscriptions commissioned between the eleventh and thirteenth centuries. Inscriptions naming the Ainnurruvar are found all over south India, but mostly  in the modern states of Tamil Nadu and Karnataka. Invariably, as the Ainnurruvar rose to  power, they incorporated smaller local organizations into the fray, attested to by multiple  inscriptions that name the Ainnurruvar organization as the main patron, often followed by  a long list of other merchant organizations. Since the inscriptional content varies much  

 

15 Many contemporaneous South Indian organizations (including merchant, royal, agrarian, and other  groups) used numbers into their titles. For example, the “Bellagunda 300,” or the ‘Kemgal Five Hundred,’  which operated in Nolambavadi territory. See Andrew Cohen, Temple Architecture and Sculpture of the  Nolambas: (Ninth-Tenth Centuries), (New Delhi, India: Manohar, 1998), 26. Also see Chapter 2 on the  subject of incorporating sacred numerology into the Ainnurruvar eulogy.

9 

according to region and time period, scholars have struggled to define the organization,  often using the word “guild” as a descriptor.  

Describing the Ainnurruvar’s power structure is even more difficult. Several  studies have attempted this feat, but crucial structural elements, e.g. central governing  bodies, legalistic codes, hierarchy, either remain unidentified or highly debated. For  example, Kenneth Hall proposes that the Ainnurruvar operated through the Chola state  mediated institution of the nagaram, or commercial city center, which regulated local and  long distance trade. In this vision, there is a sharp distinction between local and itinerant  merchants, who converge upon the nagaram to exchange wares within a highly  systemized forum. Hall characterizes the Ainnurruvar as permanently itinerant carriers of  exclusively exotic and foreign merchandise who are barred from conducting trade at the  local level of the village.16 While Hall acknowledges that the systems are highly complex,  varying throughout time and by region, he assumes the village to be a relatively static and  

self-contained entity. 

 

16 Kenneth Hall proposes the nagaram as the main unit governing mercantile exchange, drawing much  inspiration from Quentin Skinner’s “high-order and low-order market” model, best articulated in Skinner’s  “Marketing and Social Structure in Rural China.” In Hall’s own words,  

Skinnner mapped communities of exchange, market towns, and their dependent  

territories. He then overlaid contemporary units of government administration to  

establish relationships between marketing structures and government. (P. 124)  

He then writes,  

The distribution of nagaram in the Cola domain, as well as supporting epigraphic  

evidence, indicates that the South Indian nagaram functioned similarly to  

Skinner’s market centers, serving a series of local villages by connecting them to  

the upper levels of a marketing system. (P. 125) 

Perhaps in his eagerness to comply with Skinner’s model, Hall oversimplifies many of the “state’s”  organizational features. Scholars such as L. Thiyagarajan, for example, dispute Hall’s central claim that  there was only one nagaram per nadu (territorial unit in Chola period). The most problematic aspect of his  evaluation is the strict division between “local” and “itinerant” merchants. As subsequent research  suggests, the borders between these communities are far from neatly drawn. Hall summarizes Skinner in  Trade and Statecraft in the Age of Colas, 1st ed. (New Delhi: Abhinav Publications, 1980), pp. 124-5. 

10 

Meera Abraham argues that little evidence directly connects the Ainnurruvar to  the institution of the nagaram, noting that records contain many instances of the  merchants acting independently. Moreover, she rejects Hall’s binary division between communities of local and itinerant traders, acknowledging the high degree of interaction  between the two. Her self-described “micro-study” of two merchant organizations, the  Ainnurruvar and the Manigramam, results in an unprecedented compilation and analysis  of related inscriptions. One of its greatest insights is of the organizations’ internal  diversity—Abraham concludes that the Ainnurruvar managed a myriad of other  professionals: weavers, basket makers, potters, leather workers, market gardeners, and  peasant farmers, to name a few.17 Later studies expand upon this notion, defining the  Ainnurruvar as a “merchant organization which overarches all the substantial merchant  organizations formed in some particular area, locality or town.”18 

While these studies have advanced our knowledge of the Ainnurruvar, their  definition remains vague and fails to emphasize the organization’s strongly transregional  dimensions. This dissertation’s inclusion of associated material culture—in particular, its  study of inter-communal architectural collaborations and associated artistic forms— 

expands the available textual evidence dramatically, while its interpretive thrust  advocates a shift away from understanding art within the totalizing paradigms of South  Indian kingship. By expanding the textual and interpretive foci, this study attempts to  gain insight into the social history of premodern merchants.  

 

17 Abraham, Two Medieval Merchant Guilds, 5. 

18 Karshima and Subbarayalu, “Ainnurruvar: A Supra-local Organization of South Indian and Sri Lankan  Merchants,” in AMCA, p. 87.

11 

Beyond Royal History 

Popular South Indian history has always been a story of kings. In Tamil Nadu,  India’s most southeastern region, children learn the state’s royal roots from grammar  school onwards. An edict from the third century BCE, attributed to the Mauryan emperor,  Ashoka (ca. 304-232 BCE), records three great south Indian kingships: the Pandya, Chola,  

and Chera, which once divided the geographical region corresponding to modern Kerala  and Tamil Nadu. By the fourth century, mysterious invaders, identified as the Kalabrahas,  are said to have seized the region. Little is known about this time period. The sixth through eighth centuries open a brighter chapter, initiated by the Pallava kings, who  reigned over the majority of northern Tamil Nadu, and at times, the southernmost  portions of Karnataka and Andhra Pradesh. The nation’s golden age is synonymous with  the Chola monarchs (849-1279), who conquered great swaths of land as far north as  Andhra Pradesh’s Godavari River and outside India’s modern boundaries to Sri Lanka,  the Maldive Islands, and beyond. Though the dynastic chronology continues, popular  scholarship recognizes the Cholas as its definitive climax, a fact which is illustrated by  semi-annual festivals still held at the Rajarajesvara, or Brihadisvara, temple in Tanjavur (Fig. 0.1), to commemorate its patron, the Chola king Rajaraja I (r. 995-1010), and his  unparalleled cultural contributions to Tamil civilization. National newspapers report that  these festivals “not only [pay] tribute to the King and the temple he built but [serve as] a  reminder to [the] young generation about [their] past glory.”19 

These sentiments echo those by K.A. Nilakanta Sastri, one of the early twentieth 

century’s most prolific Indologists and grandfather of Chola studies, who wrote, “Tamil  

 

19 G. Srinivasan, “Tribute to Temple’s Grandeur,” last modified September 24, 2010,  http://www.hindu.com/2010/09/24/stories/2010092458140200.htm

12 

civilization may be said to have attained its high water mark under the Cola empire of the  tenth to the thirteenth century.”20 In his oft-cited text, The Cōḷas, Sastri augments the  Chola’s impact on Tamil society even further, arguing that the royals enjoyed continuous  control of the Tamil region from the second century BCE onwards.21 Emphasizing the  Chola empire’s pervasiveness and historical continuity makes the study’s purpose self 

explanatory: the author finds the glory of Tamil civilization within Chola history.  Similar attitudes color studies of Tamil art and architecture, which have focused  almost exclusively on royal commissions. In fact, until only a few decades ago, scholars  assumed that the Chola monarchs commissioned all architectural monuments, (quite a  proposition since the period’s temples number well in the hundreds).22 Several tomes  have been written on Rajaraja I’s Tanjavur temple alone. 23 In these studies, the  Rajarajesvara’s architectural perfection becomes analogous to perfected modalities of  kingship. Percy Brown provides the consummate example of this logic in his celebrated  tome, Indian Architecture (Buddhist and Hindu Periods). For Brown, Rajaraja I’s temple  demonstrates that “the Chola dynasty had been made aware of its vast power and had had  its character revealed to itself . . . The first ruler to become conscious of this sense of  

 

20 K. A. Nilakanta Sastri, The Colas, (Madras: University of Madras, 1935), 2. 

21 Sastri states that “the history of the Colas falls naturally into four divisions: the age of the literature of the  Sangam, the interval between the close of the Sangam age and the rise of the Vijayalaya line, the  Vijayalaya line which came to prominence in the ninth century A.D. and lastly, the Calukya-Cola line of  Kulottunga I and his successors from the third quarter of the eleventh century to about the middle of the  thirteenth.” Ibid. 

22 For example R. Champakalakshmi writes, “almost all Chola temples have foundation inscriptions which  identify the patrons under whom the temples were built . . . invariably the consruction of the vimana is by  a royal personage or a chieftan; there is not a single instance in the entire Cola period of any vimana being  constructed by a vellala or a group from the local populace.” Quoted in Leslie Orr, “Cholas, Pandyas, and  ‘Imperial Temple Culture’ in Medieval Tamilnadu,” In The Temple in South Asia, edited by Adam Hardy,  

(London: British Association for South Asian Studies, The British Academy, 2007), 237. 23 See B. Venkataraman, Rajarajesvaram, the Pinnacle of Chola Art, (Madras, India: Mudgala Trust, 1985).

13 

their own might was Rajaraja the Great (985-1018), which he proceeded to inaugurate by  a superb architectural monument. It must have been a profoundly spiritual impulse which  moved this ruler to commemorate the material achievements of his line . . .”24 

These analyses beg reconsideration for several reasons. First, it is clear that many  non-royal actors played important roles within ‘premodern’ or ‘medieval’ Indian society,  and excluding them paints an inaccurate historical picture. Second, the reason why  scholars have emphasized India’s dynastic history is theoretically charged. Over the past  century, the premodern, non-western world has been used as a foil against which to  define the modern nation state. Benedict Anderson’s classic, Imagined Communities,  exemplifies this world-view. In the premodern era, he argues, religious community and  dynastic realm were “taken-for-granted frames of reference” and that “all the great  classical communities conceived of themselves as cosmically central, through the  medium of a sacred language linked to a superterrestrial order of power.”25 This line of  thinking leads to theories of a rupture between premodern (or medieval/ancient) and  modern time periods predicated upon the premodern’s presumed difference from the modern. In other words, Anderson presupposes, rather than demonstrates, a difference  between the two periods.26  

 

24 Percy Brown, Indian Architecture (Buddhist and Hindu Periods), (Bombay: D B. Taraporevala Sons and  Co. Private LTD, 1956), 103. 

25 Benedict Anderson, Imagined Communities Reflections on the Origin and Spread of Nationalism, (Rev.  ed. London ; New York: Verso, 2006), 13. 

26 Benedict Anderson succinctly describes this rupture/binary: “in Western Europe the eighteenth century  marks not only the dawn of the age of nationalism but the dusk of religious modes of thought . . .  nationalism has to be understood by aligning it, not with self-consciously held political ideologies, but with  the large cultural systems that preceded it, out of which-as well as against which-it came into being.” Ibid,  11-12.

14 

These theories have affected the last century’s scholarship of the non-western  world. In its more egregious form, in justifying Europe’s claims to political hegemony  over its colonies, colonial era scholars described non-western societies as diametrically  opposed to notions of modernity. India, in particular, represented the modern’s  anathema—a society driven by irrational impulses and ruled by despotic kings. As  scholars have shown elsewhere, Indian nationalist historians, in direct response to these  orientalist accounts, produced studies extolling India’s glorious, indigenous past. 27  Sastri’s work exemplifies this phenomenon. Nationalist, decolonizing motivations  produced a rhetoric that perpetuated the monarchic fixation, albeit with a positive valence.  In recent years, scholars have criticized the exclusion of non-royal actors from studies of  South Indian temple architecture, noting that the emphasis on royalty relies upon a  colonial logic separating center from periphery, with the latter being subject to the  former’s domination.28 As scholars produce more studies of premodern cultures, the  sharp division between ancient world and modern nation-state blurs. In direct  contradiction to Anderson’s equation of premodernity with a ‘cosmically central’  worldview, these studies chart premodern actors negotiating complex socio-cultural  identities on a daily basis, engaging in activities and pursuits unrelated to religious piety.  

 

27 See Partha Mitter, Much Maligned Monsters: A History of European Reactions to Indian Art, Chicago:  University of Chicago Press, 1992, and Tapati Guha-Thakurta, Monuments, Objects, Histories: Institutions  of Art in Colonial and Postcolonial India, (New York: Columbia University Press), 2004. 

28 See Padma Kaimal “Early Cola Kings and “Early Cola Temples”: Art and the Evolution of Kingship.”  Artibus Asiae 56, no. 1/2 (1996), 5, and Andrew Cohen’s Temple Architecture and Sculpture of the  Nolambas: (Ninth-Tenth Centuries), (New Delhi, India: Manohar, 1998). 

Studying inscriptions from early Buddhist stupas, Vidya Dehejia has demonstrated that many  individual non-royal donors frequently sponsored their construction. See Vidya Dehejia, “Collective and  Popular Bases of Early Buddhist Patronage: Sacred Monuments, 100 B.C. – A. D. 250,” in The Powers of  Art, ed. Barbara Stoler Miller, (New York: Oxford University Press, 1992), 35-45.

15 

In identifying some of the past century’s scholarly motivations for studying  kingship, I do not mean to suggest that monarchies were unimportant; rather, my aim is  to complicate our understanding of the dynamic that existed between social classes. In  many instances, the merchants in this study recognized and subordinated themselves to  the sovereignty of the ruling monarchs. This is apparent in inscriptional dating practices,  which, in the opening salvo, date the donation according to the reigning king’s regnal  year (e.g. “this was written in the third regnal year of the Chola king, Rajaraja I). In  many cases, there are striking parallels between mercantile and monarchical cultural  practices, especially literary conventions and artistic style, reflecting a relationship  characterized by interdependency. In other instances, however, (increasingly during the  late Chola period), merchants acted without royal approval, cooperating with non-elite  populations. I explore these issues in Chapters Two and Three. 

The Tamil Trade Diaspora 

I identify the dissertation’s subject population as “Tamil” for a number of reasons.  On a literal level, most of the primary textual sources I consulted are written in Tamil  script and language, save for a few in Sanskrit and Kannada. On an analytic level, this  descriptive is misleading, since a cursory investigation of merchant groups testifies to the  prevalence of transregional activities and identities in the premodern period, a fact which  blurs the boundaries between linguistic communities. Sources show that the regions  corresponding to modern day Andhra Pradesh, Karnataka, Kerala, and Tamil Nadu were  connected in interlocking political, economic, and cultural networks from early times. In  Chapter One, for example, I examine information contained in inscriptions (both content 

16 

and density) to demonstrate the high degree of interaction between merchant groups in  Tamil Nadu and Southern Karnataka. Linguistic, artistic, and religious exchange was  ubiquitous: for example, many temples in southern Karnataka are inscribed with Tamil  text, but written in vernacular Kannada script. Likewise, Kannada text appears in  inscriptions on temples in Kannanur, Tamil Nadu where the Karnataka-based Hoysala  rulers founded their Tamil capital in the thirteenth century.29 Additionally, as analyzed in  Chapter Two, an Ainnurruvar-commissioned praśasti (eulogy) appears in inscriptions in  Tamil and Kannada languages. Narrating the origins of the Ainnurruvar in poetic prose,  little variation exists between the Tamil and Kannada texts; indeed, vocabulary and  phraseology are nearly identical.  

As such, I justify employing the term “Tamil merchants” as a group descriptive based on evidence collected outside India, where merchants elected to use distinctly Tamil linguistic and artistic conventions. Archaeologists and epigraphists have identified  at least nine Tamil script/language inscriptions at different locations in Southeast Asia.30  These are not the only indicators of a Tamil presence. Although little work has been  done on the subject, conversations with scholars and curators have revealed several sites  in Southeast Asia housing Dravida (south Indian-style) carvings, clearly from religious  monuments. Photos from south Thailand’s Nakhon si Thammarat Museum reveal a  distinctive form of bracket capitals known as puṣpapōtigai in Tamil architectural  terminology (Fig. 0.2). While this material demands closer inspection, its sheer presence  

 

29 Dorasamudra (Halebid) in Karnataka remained their primary capital. 

30 See N. Karashima, “Tamil Inscriptions in Southeast Asia and China,” in AMCA, pp. 10-18, and Jan  Wisseman Christie, “The Medieval Tamil-Language Inscriptions in Southeast Asia and China,” Journal of Southeast Asian Studies 29, no. 2 1998: 239-68.

17 

suggests that in the premodern period Tamil-speaking diaspora communities participated  in trade networks that connected South India, Sri Lanka, Southeast Asia, and Southern  China into a single circulatory sphere. Similarly, in choosing to record donations using  Tamil script (and sometimes Sanskrit), rather than Kannada, Telugu, or Malayalam, it is  likely that merchants self-consciously maintained a Tamil identity abroad, even while  continually engaging with individuals hailing from a wide variety of ethnic, political, and  religious affiliations, connected by the world system. Though we still do not fully  understand how these affiliations worked, we know they varied greatly according to each  community. Elizabeth Lambourn, for example, writing on connections among Muslim  communities across the Indian Ocean, proposes that “khuṭba networks,” in which names  of particular rulers were mentioned in the Friday sermon (khuṭba), allowed individual  communities to formalize their relations with particular geographical centers and  religious leaders.31 

This dissertation also argues that studying settlements of Tamil merchants living  outside of India is crucial to understanding how merchants operated in South India. Even  though these communities left fewer written records than the Muslim polities just  mentioned, I believe that it is possible to recover something of their connective filament,  largely from visual evidence. Drawing on the work of Indian Ocean scholars such as  Philip Curtin and K. N. Chaudhuri, I investigate the artistic production of South Indian,  Tamil-speaking merchants in India as well as that of trade diasporas. In Curtin’s  

 

31 Elizabeth Lambourn, “India from Aden: Khuṭba and Muslim Urban Networks in Late Thirteenth-Century  India,” in Secondary Cities and Urban Networking in the Indian Ocean Realm, c. 1400-1800, ed. Kenneth  Hall, (Lanham: Lexington Books, 2008), 55.

18 

formulation, “trade diaspora,” describes mercantile communities with diverse systems of  governance. He writes,  

Sometimes scattered settlements of the same culture had no formal  ties of any kind. They were united only by the solidarity that could  be built on the sentimental ties of common religion, language, or  distant kinship. At the other extreme were trade diasporas that were  founded as political entities with each node under central control.32 

To date, no study has analyzed these Tamil merchant communities’ connected histories in  the Indian Ocean circuit. This lacuna has resulted, in part, from the fact that empirical  data on individual communities of the time is sparse. Moreover, an academic tendency to  study particular areas or regions individually as opposed to cross-culturally has prevented  comparative work, particularly in the discipline of art history. Visual and literary cultural  studies, produced in and around the Indian ocean littoral, have fortunately begun to buck  this trend, exposing the anachronism of scholars who apply modern epistemological  understandings of space and nation to people of the premodern past who were much more  mobile than typically assumed.33  

 

32 Philip D. Curtin, Cross-Cultural Trade in World History, Studies in Comparative World History  (Cambridge Cambridgeshire ; New York: Cambridge University Press, 1984), 7. See also K. N. Chaudhuri,  Trade and Civilization in the Indian Ocean : An Economic History from the Rise of Islam to 1750 (Cambridge Cambridgeshire ; New York: Cambridge University Press, 1985); Kenneth R. Hall, Secondary  Cities and Urban Networking in the Indian Ocean Realm, C. 1400-1800 (Lanham: Lexington Books, 2008). 

33 Many scholars have debunked the equation of mobility with modernity. For instance, Enseng Ho  emphasizes that before modernism, “experiences of mobility involved complex and subtle interplays  between absence and presence in many dimensions: tactile, visual, auditory, affective, aesthetic, textual,  and mystical.” Engseng Ho, The Graves of Tarim: Genealogy and Mobility across the Indian Ocean, The  California World History Library 3 (Berkeley: University of California Press, 2006).

19 

While existing studies of Indian Ocean merchant communities do attend to  inscriptions etched into stone surfaces across South India and Southeast Asia, they  generally treat this information as purely textual, and neglect analyzing the material  objects themselves. In contrast, my work brings material culture to the analytical fore by  highlighting the narrative potential of visual sources, treating monuments, art objects, and  artifacts produced by or linked to South Indian merchants as important historical “texts.”  In many cases, these items are the sole remnants of once vibrant mercantile communities.  

I have conceived of this project as a first in a series of works on the artistic  production of Tamil trade diaspora communities within the Indian Ocean circuit. In my  fieldwork, I visited sites in southern India (coastal and inland Tamil Nadu and southern  Karnataka) as well as the city of Quanzhou in China. At each of the 149 sites I visited,  epigraphical and architectural structures/remains attest to the historical presence of  Tamil-speaking merchant communities. In Chapter One, by analyzing the information  contained within the epigraphical record, in addition to merchant-produced material  culture, I argue that a transregional, perhaps even transcultural merchant identity was  forged. Merchants were less connected by commercial partnerships than by shared sign systems of solidarity, which art and literature played essential roles in signaling. As a  result of this transregional self-imagination, I contend, merchants eventually gained  power and independence from central authority, becoming important local elites  themselves. This non-royal perspective in fact fits nicely with manuals describing royal  codes of conduct, such as the Arthaśāstra. This text is regarded as premodern India’s  most important administrative manual for kings, and expresses “a vision of political  authority that is multicentered, necessarily shifting, and automatically encompassing a 

20 

wide range of semi-autonomous forms or intermediate authorities within the ambit of the  realm.”34  

My arguments figure into current scholarship on the nature of political power in  South Asia between 1000-1500. Identifying and analyzing the phenomenon of  vernacularization in southern Asia from roughly the second millennium onwards—in  which vernacular literary cultures languages displaced the more homogenized and  cosmopolitan Sanskrit literary culture—Pollock argues that these changes “not only  correlate with transformations in social identity but appear at times to converge with a  shift in the perceived scope of political power.”35 Whereas before, Sanskrit literary texts  had circulated within a vast geographic expanse, from Central Asia to Sri Lanka and from  Afghanistan to Annam, now individuals who chose to enunciate their identities through  more localized vernacular literature elected to break not just with a language, but also  with cultural communication and self-understanding. 36 At the heart of Pollock’s  argument is the belief that choosing to write in a language itself communicated a very  particular vision of the world.37 My dissertation makes similar contentions about the  significance of the architectural patron’s choice of form and content—the deployment of  

 

34 James Heitzman, Gifts of Power: Lordship in an Early Indian State, (Delhi ; New York: Oxford  University Press, 1997), 15. 

35 Sheldon Pollock, “India in the Vernacular Millennium: Literary Culture and Polity, 1000-1500, Daedalus 125, no. 3 Early Modernities (Summer 1998), 42.  

36 Ibid., 45-46. 

37 Pollock sees “the rise and spread of Sanskrit inscriptions [as] a synecdoche for a range of literary-cultural  (and political-cultural) practices;” similarly, I wonder whether we can interpret the spread of Tamil  language inscriptions and architecture across the Indian Ocean as signifying discrete political-cultural  practices. Pollock, “The Cosmopolitan Vernacular,” Journal of Asian Studies 57, no. 1 (February 1998), 6. 

21 

Dravida style carvings—to the forging of a cosmopolitan Tamil merchant identity in the  Indian Ocean circuit between 850 and 1281.  

Kochchenganan in China  

These concerns are illustrated better, perhaps, by examining a sculptural relief  from the Quanzhou Shiva temple (Fig. 0.3). Carved on a granite slab measuring roughly  27” (length) x 8” (depth) x 18” (height), it depicts a caparisoned elephant approaching a  linga (aniconic symbol of Shiva) underneath the branches of a flowering tree. The  elephant lifts its trunk mid-step and delicately places a large lotus flower atop the linga.  The relief illustrates a well-known tale of the early Chola king, Kochchenganan. In one  of his past lives, Kochchenganan was a spider that wove a daily web to protect a Shiva linga located in a forest. An equally devout elephant also paid daily homage to the linga 

by lustrating it with water from its trunk, thereby removing the spider’s web. The  fortunes of both animals changed one day, when the spider, infuriated by the elephant’s  continued destruction of his work, attacked it and bit its trunk. The elephant smashed his  trunk against the ground, killing the spider, but then died from the venomous bite.38  Ultimately, Shiva rewards both creatures for their devotion with honorable reincarnations. 

Kochchenganan iconography rarely appears outside of the Tamil region. The  Periyapuranam, the twelfth century Tamil literary work in which the story is recounted,  specifies that the event occurred on the banks of the Kaveri River, one of the major  centers of cultural production within the Tamil region. An iconographic parallel to the  

Quanzhou panel appears in the maṇḍapa of the Jambukesvarar Shiva temple in  

 

38 Sekkizhar, trans. TN Ramachandran, St. Sekkizhar’s Periyapuranam, (Tanjavur: Tamil University, 1990),  4197-4202.

22 

Tiruvanaikka, Tiruchirappalli District, (Fig. 0.4), stylistically assignable to the Nayaka  period (ca. sixteenth-seventeenth centuries).39 The most noticeable difference between  the two Kochchenganan reliefs is that the Tiruvanaikka version contains several  additional figures. These are the goddess, Lakshmi, who extends a hand in benediction  on behalf of Shiva, the sage, Agastya, who sits in a devotional posture underneath the tree,  

and a spider, Kochchenganan, which seems to hover in dead space over the linga. One  could generalize that the Tiruvannaikka relief contains more of the story’s narrative  elements; however, the Quanzhou version contains enough iconic elements to identify the  tale: an elephant with upturned trunk, lotus, linga, and flowering tree with individuated  leaves. It also maintains the same composition as Tiruvanaikka’s, albeit reversed, with  the elephant approaching the linga from the viewer’s right, with the tree on our far left.  Yet there are also differences. Strongly linear in design, the Quanzhou elephant and linga,  while well proportioned, are executed in a flat relief, suggesting a hand less accustomed  to portraying sculptural volume. The artist has depicted the elephant ears with rigid  triangular folds, a stylistic treatment unseen in India, but more akin to Chinese  ornamental patterns such as the dense and angular cloud motif that appears at the linga’s  base in the Quanzhou relief.  

The above example encourages us to rethink several concepts that have been  central to scholarship on south Indian history and society. First, its unexpected location  in China demonstrates the artistic patronage of Tamil merchants, who voyaged through  the Indian Ocean to arrive there, revising the traditional view of Indian art as an  

exclusively royal enterprise. Second, it expands Indic visual culture’s geographical reach,  

 

39 Local legend states that Kochchengannan constructed the original temple in Tiruvanaikka.  Balasubrahmanyam, Middle Chola Temples, 392.

23 

since the narrative depiction’s existence outside India attests to a larger circulatory  network within the Indian Ocean. Finally, it points up the cross-cultural transmission of  artistic style by expressing a melding of Sino-Indic concepts and techniques, encouraging  a reading that emphasizes the artist’s and patron’s active engagement with aesthetic form,  and the transformation of preexisting representational modes.  

. . . 

I have organized this dissertation typologically and thematically, so that each  chapter focuses on different primary sources and questions. Chapter One, “Tamil  Merchants in World History,” provides a historical framework, emphasizing the  multidirectional networks that connected the Tamil region to the Indian Ocean world,  moving to a more focused discussion of the Ainnurruvar merchant association. Chapter  Two, “Of Symbols and Stones,” studies the Ainnurruvar eulogy as blending of text and  object, analyzing its literary precursors and sculptural expressions to show that both elite  and non-elite merchant cultures are represented. Chapter Three, “Constructing  Community in Stone,” examines temples commissioned by the Ainnurruvar, and argues that their Drāviḍa-style constituted a transregional architectural phenomeon that signaled  a common identity among its patrons. Finally, Chapter Four, “A Shiva Temple in  Medieval Quanzhou,” works to recover the ruined Shiva temple’s visual logic and its  patrons’ place in Quanzhou society.  

The dissertation’s title, “Constructing Communities,” has a dual meaning. Firstly,  it refers to the dynamic process through which mobile, mercantile groups related to each 

24 

other and other populations. Merchants forged ties in south India and beyond through the  use of architectural and sculptural mediums; by creating their lived environment in stone  and prose, merchants transported memories and culture to new locales, transforming their  identities through replication and reinterpretation of familiar spaces. Architecture and  sculpture expressed their ambitions and ideology to a wider audience, but also provided  tangible guidelines for internal self-fashioning. Secondly, the title addresses the  historian’s fraught task of reconstructing the past through diverse and always fragmentary  material. The scaffolding of this constructive process is perhaps more visible in the  following study, as it raises even more questions than it answers, but I hope that at the  very least it sheds light on the rich possibilities of cross-cultural research. 

25 

CHAPTER 1: TAMIL MERCHANTS IN WORLD HISTORY 

This chapter situates Tamil merchants in their wider geographic and historical  contexts. First, I outline the geographic and seasonal factors affecting the Tamil and  Indian Ocean regions.1 Second, I focus on merchants from the Tamil region, narrowing  onto the Ainnurruvar, a transregional South Indian merchant organization active between  the ninth and fourteenth centuries. While the organization has been the subject of several  studies, it has not been situated within larger conversations about the structure and  organization of premodern South Indian polities. Building on earlier conceptions of  “ritual polity,” I propose the category of “transregional polity” to describe the  Ainnurruvar’s social and political practices.2  

Geography and Seasons 

Sandwiched between the Gulf, Red Sea, and Bay of Bengal, south India was an  important transit point for maritime traffic circulating between the Middle East and  China; several studies consequently have dubbed it the “hinge” or “fulcrum” of the  Indian Ocean.3 The South Indian peninsula is defined by a huge triangular plateau,  beginning in the Deccan’s Vindhya Mountains and extending to the subcontinent’s  

 

1 This section owes much to Michael Pearson’s, The Indian Ocean, Seas in History (New York: Routledge,  2003), 13-26. 

2 As James Heitzman writes, “the paradigm of the ‘ritual polity’ suggests that cultural meaning may explain  the formation of the early state, the legitimation of its authority, and the spatial configurations of its  political units.” In “Ritual Polity and Economy: The Transactional Network of an Imperial Temple in  Medieval South India,” Journal of the Economic and Social History of the Orient 34, no. 1/2 (1991), 23. 

3 Abu Lughod adds that the Indian Ocean constituted a “great ‘highway’ for the migration of peoples, for  cultural diffusion, and for economic exchange.” Abu-Lughod, p. 261, while Pearson terms it the “fulcrum,”  Indian Ocean, p. 53

26 

southernmost tip (Map 2). The Eastern and Western Ghats stretch from north to south,  dividing the region, but are nowhere impassable.4 The mountain ranges separate low  lying coastal plains from the slightly elevated plateau, producing two distinct coastlines:  the Malabar Coast in modern Kerala, and the Coromandel Coast in modern Andhra  Pradesh and Tamil Nadu. 5 Several rainfed rivers flow eastward from the Western Ghats  towards the Bay of Bengal, with the Krishna and Godavari rivers in the north and the  Kaveri, Palar, and Vaigai in the south.  

The Coromandel coastline has no natural harbors and constant high surf, making  docking ships there particularly treacherous. Still, many ports are known to have existed  in early south India, including Arikamedu, Kaveripattinam, Periyapattinam, Alakankulam,  Korakai, and Kayalpattinam.6 Most harbors were located at the mouths of rivers that  eventually connected to the Bay of Bengal, providing protection from the open sea to  

 

4 Michael Pearson, Indian Ocean, 16. 

5 The name, Coromandel, derives from a Portuguese corruption of the Tamil term, “Cholamandalam,”  literally, “land (circle) of the Cholas.” 

6 These are confirmed by archaeological remains and textual references. For an overview of port locations  in ancient southern India, see Himanshu Prabha Ray, and Jean-François Salles, eds. Tradition and  Archaeology: Early Maritime Contacts in the Indian Ocean. New Delhi: Manohar Publishers and  Distributors, 1996. 

Chinese, Arabic, and European traveler diaries name many ports along the southern peninsula, and  scholars have devoted much work to identifying these locations. For example, see Noboru Karashima’s  identification of “Dabadan” in Chinese sources with modern Periyapattinam in southern Tamil Nadu.  “Trade Relations between South India and China During 13th-14th Century A.D,” Journal of East-West  Maritime Relations 1 (1989): 59-81.  

Much scholarship exists on individual ports, although the majority awaits systematic  archaeological excavation. For a geographical perspective on the selection of ports, see Jean Deloche,  “Geographical Considerations in the Localisation of Ancient Sea-Ports in India,” Indian Economic and  

Social History Review 20, no. 4 (1983): 439-48 and “Études Sur La Circulation En Inde. Iv: Notes Sur Les  Sites De Quelques Ports Anciens Du Pays Tamoul.” Bulletin de l’Ecole française d’Extrême-Orient 74  (1985): 153-57. For studies of individual ports, see Roderich Ptak’s “Yuan and Early Ming Notices on the  Kayal Area in South India,” in China’s Seaborne Trade with South and Southeast Asia (1200-1750), edited  by Roderich Ptak, Aldershot: Ashgate, 1999, 137-56, as well as Vimala Begeley’s numerous studies of  Arikamedu, especially, The Ancient Port of Arikamedu: New Excavations and Researches 1989-1992, Vol.  22, Mémoires Archéologiques. (Pondicherry: Centre d’histoire et d’archéologie, École française d’Extrême Orient, 2004).

27 

moored ships. Attesting to this fact, Marco Polo mentions ‘Palayakayal,’ which  translates as ‘Old Lagoon,’ as a major port.7 

The monsoon’s importance to Indian Ocean travel cannot be overstated. Every  year, winds push southwest from early June to mid-September, and northeast from  November to March, producing two rainy seasons. Travelers had used the monsoons as a  marker for navigating the Indian Ocean for millennia, demonstrated by numerous  premodern travel accounts offering logistical and proverbial wisdom on navigating  monsoons. For example, one Arab author writes that “he who leaves India on the 100th day [2 March] is a sound man, he who leaves on the 110th will be all right. However, he  who leaves on the 120th is stretching the bounds of possibility, and he who leaves on the  130th is inexperienced and an ignorant gambler.”8 

The monsoon impacted agricultural practices in the Tamil region as well;  intensive periods of rainfall promoted particular methods for collecting water to cultivate  wet crops such as rice paddy. Farmers developed three primary irrigation methods: man made embankments constructed near natural depressions to contain rainwater; channeling  water from rivers into man-made canals or tanks; and digging wells to underground water  supplies.9 The topographies of individual regions determined irrigation method and  farming layout. For instance, eastern districts, receiving more rainfall than western, to  this day utilize interdependent water systems in their farmlands, while western farmlands  

 

7 Y. Subbarayalu posits that Palaya-Kayal is “10 km to the north of Tiruchendur and about 30 km south of  Tuitcorin. . . [and] 3 km interior from the coast.” See “Chinese Ceramics of Tamilnadu and Kerala Coasts,”  in Tradition and Archaeology, 109-114.  

8 In Michael Pearson, Indian Ocean, 21 (footnote 25) 

9 James Heitzman, Gifts of Power: Lordship in an Early Indian State, (Delhi; New York: Oxford  University Press, 1997), 24.

28 

relied on individual tanks that divided the land into discrete units. 10 Indeed, it appears  that irrigation systems impacted the formation of the Tamil region’s oldest political  structures; Y. Subbarayalu suggests that before royal administration, groups of villages  with common water networks partnered together.11  

In the Tamil country’s littoral regions, waves constituted another deep structural  element. High waves along the Coromandel coastline made it difficult for sailors who  were unfamiliar with the terrain to navigate its waters. We might imagine the difficulty of sailing through an account written in 1183 by the religious pilgrim, Muhammed ibn  Ahmad Ibn Jubair. When arriving at the important Gulf port of Jiddah, Jubair marvels at  the captain’s navigational skills writing, 

The entry into [Jiddah] is difficult to achieve because of the many reefs  and the windings. We observed the art of these captains and the mariners  in the handling of their ships through the reefs. It was truly marvelous.  They would enter the narrow channels and manage their way through  them as a cavalier manages a horse that is light on the bridle and tractable.  They came through in a wonderful manner that cannot be described . . .12 

While deep structures affected the development of the Indian Ocean’s  interconnected regional networks, they were not wholly responsible. Advances in  

 

10 K.C. Alexander, “Some Characteristics of the Agrarian Social Structure of Tamil Nadu.” Economic and  Political Weekly 10, no. 16 (1975), 666. 

11 These regions become the territorial unit known as the nadu in Tamil inscriptions. Y. Subbarayalu, The  Political Geography of the Chola Country. Madras: Tamil Nadu State Department of Archaeology, 1972,  22. 

12 Muhammed ibn Ahmad Ibn Jubair, The Travels of Ibn Jubayr (1183-1185 AC), trans. R.J.C. Broadhurst,  London, Jonathan Cape, 1952, pp. 69 et seq. Quoted in Michael Pearson, Indian Ocean, 19.

29 

maritime technology between 1000-1500 CE segmented the previous east-west maritime  route into port centered regional networks, producing distinct navigational circuits  corresponding to the Bay of Bengal, Melaka Straits, Java Sea, South China Sea, and Sulu  Sea (Maps 3 and 4). In this network, “there was no hierarchical trade structure  corresponding to markets with a single clearing house or a single core with peripheries  with which it traded on terms of unequal exchange. Rather, the Bay of Bengal was a  poly-centric networked realm,” in which north-south movement of tropical goods was  commonplace.13 

The Indian Ocean’s segmentation resulted in shorter routes for individual ships.  Whereas before 1000 CE, a single boat might have made the entire trip from the  Arabian Sea to China, now it stopped in south India to transfer goods onto a new boat  with a different crew, which completed the journey to southeast Asia and China.  Explaining the change, Pearson suggests that traders realized the inefficiency of using  one ship per voyage, since it required long respites at port waiting out the monsoon. He  also acknowledges that the segmentation might have resulted from the increased power of  South Indian merchant organizations, which now took charge of the goods from Arabia.14  

Early Merchant History in Tamilakam (ca. 300 BCE-700 CE) 

Early Tamil literature refers to the region between the Vengadem Hills (Tirupati  in modern Andhra Pradesh) and Venadu (in southern Kerala) as Tamilakam, or “the  

 

13 Kenneth Hall, “Ports-of-Trade, Maritime Diasporas, and Networks of Trade and Cultural Integration in  the Bay of Bengal Region of the Indian Ocean: C. 1300-1500,” Journal of the Economic and Social History  of the Orient 53 (2010), 113. 

14 Pearson, Indian Ocean, 87-88.

30 

abode of the Tamils.15” Tamilakam denotes what Burton Stein has termed a cognitive  region, imparting a shared cultural identity among its inhabitants, expressed in common  mythic and symbolic beliefs.16 Other scholars have observed that, “classical Tamil  literature is explicitly conscious about the close relationships among language,  geographical territory, and culture.”17 Less studied is its converse implication, which is  that a distinctly Tamil cognitive region necessitates that its inhabitants were attuned to a  world much larger than their immediate surroundings. The Tamil region’s centrality to  Indian Ocean trade must be held partially accountable for this cognizance.  

In the third century BCE, Ashoka, the Mauryan emperor who controlled most of  northern India, authored several inscriptions referring to southernmost peninsular India  that mention three kingships lying outside his realm: the Chola, Pandya, and Chera. By  the sixth century, Tamilakam consisted of four regions inhabited by the Chola, Pandya,  Chera, and the newly formed Pallava dynasty. With the exception of the Chera, based in  modern western Kerala, each of these territories developed around rivers: the Pallavas  around the Palar, the Cholas around the Kaveri, and the Pandyas around the Vaigai.  Expanding agricultural activities and the need for constant irrigation likely provided the  impetus to settle in these locations. As mentioned earlier, rivers also were important to  

 

15 This geographic description occurs in the first extant Tamil grammar, the Tolkappiyam, defining “the  good world where Tamil is spoken (stretching from) northern Venkatam to Kumari in the South.” Martha  Selby and Indira Peterson, introduction to Tamil Geographies: Cultural Constructions of Space and Place  in South India, eds. Indira Viswanathan Peterson and Martha Ann Selby. (Albany: State University of New  York, 2008), 4. 

16 Burton Stein, “Circulation and the Historical Geography of Tamil Country,” The Journal of Asian Studies  37, no. 1 (1977), 7. 

17 Selby and Peterson, Tamil Geographies, 4.

31 

maritime trade, providing boats shelter from the open sea and uninterrupted passage  upriver to the Tamil hinterland.  

Archaeologists have discovered many Roman goods at these locations, which  point to trade networks connecting India, Southeast Asia and the Roman west.18 For  example, a pottery sherd from a storage jar, ca. first century BCE, inscribed with Tamil Brahmi script, was discovered in Quseir-al-Qadim, an ancient Roman settlement on  Egypt’s Red Sea Coast.19 A terracotta sherd from Alagankulam dates from the same  period, and carries a rare illustration of sea vessels frequenting the region at that time. It  bears an incised outline of a Roman ship with riggings and mast, identifiable as a three master Roman sailing ship, one of the largest ships of the time, carrying Greco-Roman  goods between India and Egypt (Fig. 1.1).20 By the first century, Tamil country was an  important hub for overland traffic from northwest India, western maritime traffic with  Arabs and Romans, and eastern maritime traffic with Southeast Asia. Roman coin hoards  discovered in Karur and Coimbatore districts suggest that overland trade routes  connected southern India’s east and west coasts. This is confirmed by Greek and Latin  sources, like The Periplus of the Erythrean Sea, composed in the first century CE by an  Egyptian Greek merchant, who describes the trade routes he used to travel to Africa and  India, including those connecting the Kerala coast to Tamil Nadu’s hinterland. A half  century later, the Roman geographer, Ptolemy, writing of India’s southern region, names  

 

18 H.P. Ray, “Maritime Archaeology: An Overview” Tradition and Archaeology, 4. 

19 “Tamil Brahmi Script in Egypt,” last modified Nov 21, 2007,  

http://www.hinduonnet.com/2007/11/21/stories/2007112158412400.htm. New evidence continually points  to earlier dates for international maritime commerce. 

20 Iravatham Mahdevan, Early Tamil Epigraphy from the Earliest Times to the Sixth Century A.D,  (Cambridge, MA: Harvard University Press, 2003), 6.

32 

two ports (which scholars identify with modern Nagapattinam and Kaveripattinam), and  refers to the Cholas as the “Soras.”21 

Many literary texts refer to the prosperity and cosmopolitanism of Kaveripattinam  (Puhar). The Manimekhalai, one of most important epics from the Tamil “Sangam”  school of literature (ca. 300 BCE-300 CE), is set in Kaveripattinam; intriguingly, its  author, Shattan, was both a renowned poet and noble merchant. Shattan likens the city to  a beautiful woman whose many adornments reflect the city’s prosperous maritime trade: 

The moats filled with clear water, embellished with innumerable flowers,  sounding with the song of a thousand kinds of bird, form a ring around her  ankle. The surrounding walls, commanded by towers, are her diamond studded girdle. The gates, surmounted by staffs with flags flying are her  shoulders laden with many necklaces . . . The vast palace, thousands of  years old, of matchless splendor, commanding the city, the residence of  the Chola king who wears a necklace of orchid tree leaves . . .22  

Other passages in the Manimekhalai explicitly mention merchants working in bustling  bazaars, conducting business as they “murmured their prayers, making offerings of  flowers.”23 Another Sangam period work, the Pattinappalai (ca. 100 BCE-100 CE) lists  the exotic Indian Ocean goods available in Kaveripattinam. The city “had an abundance  of horses brought over the water, sacks of black pepper brought [overland] in carts,  

 

21 Sastri, The Coḷas, (Madras: University of Madras, 1935), 22. See also K. A. Nilakanta Sastri, Foreign  Notices of South India; From Megasthenes to Ma Huan, 1st ed. (Madras: University of Madras, 1972). 

22 Cattanar, Alain Daniélou, and T.V. Kopalayyar. Manimekhalaï: The Dancer with the Magic Bowl, (New  Delhi, India: Penguin Books, 1993), 22. 

23 Ibid., 34.

33 

gemstones and gold from the northern mountains, and sandalwood and eaglewood from  the western hills, pears from the southern seas and coral from the eastern seas, grains  from the regions of the Ganga and Kaveri, foodgrains from Ceylon, and the products of  Burma and other rare and great commodities.24  

The Pattinapallai’s reference to Sri Lanka (Ceylon) opens discussion onto the  Tamil region’s relationships with Sri Lanka and Southeast Asia. Sri Lanka, although not  conceptually part of Tamilakam, shared many commonalities with it well before the  Chola king, Rajendra I, seized control of the region in the eleventh century. Notices in  Greek, Roman, Persian, Arabic, and Chinese language sources, archaeological  excavations, and epigraphy—written in early Brahmi, Sinhalese, and Tamil scripts— provide information about the many trade emporia along its coasts and hinterland. Indian  Ocean trade was crucial to Sinhalese economy, demonstrated by the variety of foreign  coin hoards discovered there dating from the first century onwards.25 Additionally, two  Sri Lankan Buddhist narratives, the Dipavamsa (ca. fifth century CE) and the  Mahavamsa (ca. sixth century CE) describe merchandise and personnel circulating  between Sri Lanka and Myanmar.26  

Further east in southern Thailand, the Andaman Coast, which spans the Thai Malay peninsula, was a stopping point for ships sailing the Indian Ocean route and  became a lively trading center. Scholars have discovered several smaller objects dating  

from the early centuries of the first millennium and inscribed with Tamil-Brahmi script,  

 

24 Pattinappalai, lines 185-192, quoted in Kanakalatha Mukund, The Trading World of the Tamil  Merchant: Evolution of Merchant Capitalism in the Coromandel, (Chennai: Orient Longman, 1999), 15. 

25 See Osmund Bopearachchi’s “Seafaring in the Indian Ocean: Archaeological Evidence from Sri Lanka,”  in Tradition and Archaeology, 59-77. 

26 Malini Dias, “Historical Relations of Ancient Sri Lanka and Myanmar,” AMCA, 172.

34 

from which the existence of Tamil speaking settlements has been interpreted. The  earliest find is a pottery sherd that has been dated to about the second century CE, which is inscribed with fragmentary letters that are possibly part of the Tamil word for monk.27  Another small rectangular stone bears a short inscription in third century CE Tamil Brahmi script, identifying it as “the great goldsmith’s stone,” (perumpatan kal), implying  that Tamil speaking artisans had settled in Thailand by this time.28 Tamil merchant  communities also existed in Lobu Tua (Barus), Sumatra as indicated by a late eleventh century Tamil inscription, which relates the local Ainnurruvar organization’s decision to  tax merchant ships docked at port.29 

Missionary activity and religious pilgrimage also impacted the formation of  Indian Ocean networks. Buddhist monks from the subcontinent flooded Sri Lanka and  Southeast Asia at the beginning of the first millennium, traveling as far as China by the  first century CE.30 Most famously the Chinese Buddhist pilgrims Faxian and Xuanzang  authored travel diaries describing their voyages to the subcontinent in the fourth and  seventh centuries respectively, emphasizing India’s cultural sophistication.31 Faxian’s is  

 

27 The three letters are ‘Tu Ra O . . .’ , which might be part of the word ‘turavon.’ Boonyarit Chaisuwan,  “Early Contacts between India and the Andaman Coast,” in Early Interactions between South and  Southeast Asia: Reflections on Cross-Cultural Exchange, eds. Pierre Yves Manguin, A. Mani, and Geoff  Wade, (Singapore: Institute of Southeast Asian Studies; Manohar), 89. 

28 In 1992, Noboru Karashima et al. translated the greenish rectangular stone from Khuan Luk Pat, which is  now located in the Wat Khlong Thom Museum, Krabi province, southern Thailand. “Tamil Inscriptions in  Southeast Asia and China,” AMCA, 10 and 156. 

  

29 Y. Subbarayalu, “The Tamil Merchant-Guild Inscription at Barus, Indonesia: A Rediscovery,” AMCA,  20-21. 

30 This was not the first contact between India and China; merchants had navigated these paths as early as  the second century BCE, when Han Chinese official recorded their plans in the Shiji (Records of the Grand  Historian) to take over the trade route connecting southwestern China to India. Sen, Buddism, Diplomacy,  and Trade, 3. See Chapter 5 for a lengthier account of contact between India and China. 

31 See Faxian, Trans. Henri Giles, The Travels of Fa-Hsien (399-414 A.D.) or, Record of the Buddhistic 

35 

the sole extended account of Indian Ocean travel from this time (415 CE), and contains a  description of a boat ride between Sri Lanka and Srivijaya, in which two hundred Brahmanical merchants accompanied him.32  

While overseas commerce was essential to the Tamil country, it was not the only  factor in its socioeconomic development. R. Champakalakshmi has argued that “external  stimuli” provided the primary motivation for internal development within Tamil country.  The surge of coastal and overseas trade along the Coromandel Coast between 300 BCE  and 300 CE, she writes, prompted an “urban revolution” and the creation of “trade  enclaves.”33 In this narrative, urban centers in the Tamil hinterland developed in order to  accommodate increased traffic towards the coast. The assumption of unidirectional  movement from village to town overlooks complex interactions among multiple centers,  however.34 Sanjay Subrahmanyam has proposed a tripartite model for analyzing trade,  differentiating among coastal, overland, and overseas networks; all of these pathways  exhibit “inter-dependence” and “independence,” an observation which overturns the  assumption that inland villages were less significant than ports in India’s commerce.35  

  

Kingdoms, (London: Routledge & Paul, 1956); Sally Hovey Wriggins, The Silk Road Journey with Xuanzang, (Boulder, Colorado: Westview Press, 2004). 

32 Michel Pearson, 55. 

33 R. Champakalakshmi’s suggestion that the “urban revolution” was followed by either urban ‘devolution’  or “breakdown of earlier tribal forms” between the third and sixth centuries CE is also problematic. She  maintains that the Kalabrahas (i.e. a dynasty of evil kings) invaded Tamil country, destroying social and  political institutions, when no written records exist for this period. See Trade, Ideology and Urbanization:  South India 300 BC to AD 1300, (Oxford: Oxford University Press, 1999), 37. 

34 The conception of limited movement of goods, i.e. from village to town (small to big), simplifies the  complex reality of trade in premodern south India. This model appears in, ‘Inland trade’ in Habib and  Tapan Raychauhuri eds., The Cambridge Economic History of India, Volume I, 325-59.  

35 Sanjay Subrahmanyam, The Political Economy of Commerce: Southern India, 1500-1650, (Cambridge  England; New York: Cambridge University Press, 1990), 46-90.

36 

The unidirectional trade model also neglects important evidence from the Tamil  country’s interior that contradicts direct funneling between coast and hinterland.36 These  occur in the first written references to South Indian merchants, appearing on the walls of  rudimentary stone shelters and cave dwellings in the former Pandyan region of Tamil  Nadu, near to the modern city of Madurai. Iravatham Mahadevan’s translations of these  Tamil-Brahmi script inscriptions inform us that Tamil speaking merchants patronized  many religious and cultural institutions, indexing complex inland trade networks not  connected to ports.37 The first known example comes from Mangulam village near  Madurai, where a Tamil-Brahmi script inscription (ca. second century BCE), which  records that a merchant association (nigama) sponsored the construction of the cave’s  stone beds. The northern origin of the term nigama and the fact that most of the Tamil Brahmi inscriptions have Buddhist or Jain associations “may indicate a northern origin  for at least a section of the traders.”38 A more significant example exists at Alagarmalai,  

located about twenty kilometers northeast of Madurai, in a cave reached after a steep  

 

36 Generally speaking, commercial and social networks do not function within such neatly defined  geographic zones and temporalities, and much information suggests that the Tamilakam’s agriculture and  economy developed unevenly. Moreover, Champakalakshmi bases her argument on flawed interpretations  of early Tamil Sangam period literature, especially the concept of tinai. Chamapakalakshmi defines tinai as different “eco-systems” of Tamilakam, extracting data about the different types of landscapes and people  that comprised early Tamil Nadu. As scholars like Martha Ann Selby have noted, tinai is a more nebulous  concept. Selby prefers to define tinai as “context,” where emotion is the only fixed quality. As she writes,  “this context is sweeping, and includes geographical space, time, and everything that grows, develops, and  lives within that space and time, including emotion.” Martha Ann Selby, “Dialogues of Space, Desire, and  Gender in Tamil Cankam Poetry,” in Tamil Geographies, 25. 

37 Meera Abraham reads the inscriptions’ Jain-related content as evidence of early trading connections  between Karnataka and Tamil Nadu. She reaches this conclusion on the assumption that Jain  establishments “were often established in areas of commercial importance” and that the disciples of  Bhadrabahu brought Jainism to the Tamil country by the third century BCE. Abraham, 49. 

38 Kesavan Veluthat, The Early Medieval in South India, 25. On the term nigama, see Romila Thapar, who  records multiple donations made to Buddhist chaityas in the western Deccan from the Maurya period  onwards. In these instances the donors are described as being from the nigama, which has been interpreted  as a market center. “Patronage in Community,” in The Powers of Art: Patronage in Indian Culture, ed,  Barbara Stoler Miller, (Delhi; New York: Oxford University Press, 1992), 26.

37 

ascent through rocky hillsides (Fig. 1.2). An inscription records that around the first century BCE, a group of merchants spent one night or more sleeping in the cave’s stone  beds. The cave’s high ceilings provided shelter from the dry climate, while a deep,  freshwater pool quenched their thirst. Multiple weatherworn Tamil-Brahmi script  inscriptions are scrawled across the ceiling. They record that merchants, each identified  by their wares (sugar, salt, iron implements, and textiles), donated money to the nearby  Jain monastery, which shows that merchants specializing in the sale of particular goods  traveled together within a single organization. 39 Thus, the earliest textual references to  merchants in the Tamil region record them acting as architectural and cultural patrons,  endowing monuments, institutions, and objects in order to establish and elevate their  burgeoning sociopolitical identities.  

The historical record essentially vanishes during the “Kalabhra” interregnum  (300-600).40 Reliable records reappear after 600 CE, most of which are dated according  to Pallava dynastic years.41 The Pallava worldview incorporated aspects of north Indian  political order, which viewed kings as upholders of dharma (morality) and ritual purity.  Champakalakshmi observes that no vaisya (merchant) caste existed in the region at this  time. She proposes that this absence resulted from the growth of Brahman-centric power  

 

39 Mahadevan, 15. 

40 Hetizmann suggests that kings in the Deccan temporarily displaced the South Indian kings’ seat of power  (Heitzman, Gifts of Power, 2). Veluthat proposes an alternative reading, in which Brahmanical ideology  solidified its grip on South Indian society, resulting in major social changes, including a turn to  agriculturalism. He dismisses the common interpretations (e.g. Sastri) of northern invaders or Jain  hegemony. Veluthat, The Early Medieval in South India, 45-7. 

41 The Pallavas migrated from neighboring Andhra Pradesh, eventually settling in northern Tamil Nadu,  making Kanchipuram their capital.

38 

networks between the seventh and ninth centuries.42 By the eighth and ninth centuries,  south Indian commercial networks were deeply entrenched in Southeast Asia. Jan  Wisseman Christie’s analysis of Javanese records, for example, shows that by the late  ninth and early tenth centuries, relations between India and Java were so entrenched that  inscriptions classified Indian merchants according to regional affiliation. Merchants  came from “the east coast districts of what are now Tamil Nadu, Andhra Pradesh, and  southern Orissa . . . These three regions remained a constant in the Javanese lists from the  ninth to the fourteenth century.”43 

The ninth through thirteenth centuries saw a dramatic expansion of trading  networks in India and abroad. Many inscriptions from this period contain detailed lists  describing accurate pricing of commercial goods, duties, and tax laws, created and  ratified by associations of merchants, suggesting their sophistication and autonomy. The  largest of these was the Ainnurruvar (referred to as the Ayyavole Five Hundred in  Kannada), which Y. Subbarayalu and N. Karashima have defined broadly as a “supra 

local organization of South Indian and Sri Lankan merchants.”44  

 

42 Champakalakshmi concludes that that the “growth of the brahmanical agrahara in an agricultural setting,  where trade was not a very important factor in the economy of these regions at this stage and the number of  communities dependent on trade must have been fairly small to have been practically ignored in the records  of this period.” Trade, Ideology and Urbanization, 42. 

43 An 883 inscription from Kalinguran, Kedu in central Java lists foreigners residing at Javanese ports.  These are: “mainland Southeast Asians from Campa (Champa), Remman (Ramanyadesa in Mon Lower  Burma), and Kmira (Cambodia); the listed South Asians came from Kling (Kalinga), Aryya  (Aryapura/Ayyavole), Pandikira (in Karnataka), and perhaps Singhala (Sri Lanka).” Jan Wisseman  Christie, “The Medieval Tamil-Language Inscriptions in Southeast Asia and China,” Journal of Southeast  Asian Studies 29, no. 2 (1998), 244. Christie notes that many Javanese inscriptions emphasize the  popularity of South Indian block printed textiles, which presents another fruitful avenue for future research  of Tamil merchant material culture. See Christie, “Texts and Textiles in ‘Medieval’ Java,” Bulletin de  l’Ecole française d’Extrême-Orient 80 (1993): 181-211. 

44 N. Karashima and Y. Subbarayalu, “Ainnurruvar: A Supra-local Organization,” AMCA, 72.

39 

The Ainnurruvar (‘The Five Hundred Members’) 

The following section analyzes the Ainnurruvar’s organization and structure  through inscriptions. The earliest references to the organization date from 800 CE and  were discovered on the Lad Khan and Gaudara Gudi temples in Aihole, Karnataka. Both  inscriptions are somewhat opaque because the donors’ title, ‘the Five Hundred,’  designates a group of Brahmin priests, not merchants.45 Connection with the merchant  organization is undeniable, however, as later Ainnurruvar inscriptions frequently mention  the group’s origin in Aihole, referring to the merchants as the “Five Hundred svāmins  (priests) of Aihole”46 and as the “sons of Durga, the goddess of Ayyapolil (Ayyavole or  Aihole).” By the tenth century, the Ainnurruvar had become a transregional mercantile  corporate body, their activities spanning south India. Most inscriptions appear on  temples, marking important donations by individuals or groups of members, but some  also appear on freestanding planted stones (studied in the next chapter). Membership  crossed sectarian lines; while the majority of Ainnurruvar donations describe patronage  of Shaivite or Vaisnavite institutions, several inscriptions describe donations made to  mosques, Buddhist, and Jain temples. 

Inscriptions mention the Ainnurruvar in various ways. Many contain partial or  entire versions of the Ainnurruvar praśasti (eulogy), an elaborate and surprisingly  consistent eulogy describing the fame and merit of the organization. At other times,  Ainnurruvar members appear as signatories within a larger list. The first inscription  featuring the Ainnurruvar as a merchant organization appears in 927 at Munisandai in  

 

45 The inscriptions record donations made by the Five Hundred Chaturvēdins or mahājanas, meaning  Brahman priest. AMCA, 227-8. 

46 For example, see Sirasangi inscription from Belgaum District, Karnataka (1186 CE) in AMCA, 262-263. 

40 

Pudukkottai District, Tamil Nadu, where they donated funds for a temple tank. 47  Individuals also identified themselves as members by incorporating the organization’s  name into their (its) own personal title or by naming places and things after it. For  instance, in the town of Virachchilai, in Pudukkottai District of Tamil Nadu, an  “Ainnurruvar Peraiyar” and “Ainnurruva Devan” sold a piece of land named “Ainnurruva  Mangalam” in 1282.48 The respective suffixes, peraiyar and devan, connote two males  using Ainnurruvar as a distinguished title.  

Distribution of Inscriptions 

The following section maps the geographical and temporal distribution of South  Indian merchant organizations.49 I divide the ninth through fourteenth centuries into four  sub-periods, “which correspond to the political fortunes of the [Chola] dynasty and to the  broad outlines of development in political economy.”50 I have added a fifth sub-period  (1279-1400), as merchant inscriptions continue to be found in large numbers until  approximately 1400. I illustrate each of these sub-periods with a map showing  inscriptional sites (also see Map 5 for reference). 

 

47 P. Shanmugam and Y. Subbarayalu, “Texts of Select Inscriptions of the Merchant Guilds,” AMCA, 228. 48 IPS 393 and 421. 

49 I have used Subbarayalu and Karashima’s list of inscriptions from AMCA, but excluded those postdating  1400, which appear sporadically.  

50 Heitzman elaborates: “Sub-period one (849-985) was a preliminary stage when the Cholas originated,  weathered major political storms, and consolidated their control over the central part of the Tamil country.  Sub-period two (985-1070) saw the rapid expansion of Chola power and the zenith of its influence under  the greatest of its kings. Sub-period three (1070-1178) was initiated by the joining of the Chola and the  eastern Chalukya thrones, and thus the general consolidation of the Cholas’ largest power base. Generally,  however, this was a time of some retrenchment and a growing, if subdued, political dissolution within the  empire. Sub-period four (1178-1279) began with Kulottunga III trying to hold on under intensified  pressure from outside and within, leading at the end of his reign of the rapid collapse of the Chola political  system.” Heitzman, Gifts of Power, 21.

41 

Period 1: 849-985 (Maps 6-7) 

The vast majority of inscriptions are concentrated in the Chola heartland on both  banks of the Kaveri River between Tanjavur and Kumbakonam. Two isolated  inscriptions are found in coastal cities in Kerala, and there is a single inscription from  Takuapa, southern Thailand. 

Period 2: 985-1070 (Maps 8-9)  

There is a leap in the number of inscriptions, the majority of which are found in  Tamil Nadu. Inscriptions still cluster around the Kaveri, but the most noticeable increase  is further to the west around Tiruchirappalli and Karur Districts. Pockets of inscriptions  appear around Pudukkottai and Tirunelveli districts. 

 The first Ainnurruvar inscriptions are found in Karnataka, but are restricted to  southern Mysore and southern Dharwad districts. 

Period 3: 1070-1178 (Maps 10-12) 

Inscriptions in Tamil Nadu decrease visibly, while Karnataka experiences a  sudden profusion of them, most of which are concentrated in the north. In contrast to the last two periods, there is virtually no activity around the Kaveri  in Tamil Nadu. Inscriptions appear in regions south of Pudukkottai near to Madurai.  Additionally, a cluster of inscriptions appears around inland Erode District. The first  Ainnurruvar inscriptions appear in Sri Lanka’s central region and west coast. A single  inscription appears in Barus, Sumatra. 

42 

Inscriptions continue to concentrate in southern Mysore, although mostly between  Kollegal and Chamarajnagar districts. Meanwhile, there appears to have been continuous  expansion to northern Karnataka, with a huge number of inscriptions appearing between  Davanagere and Bijapur. The majority of these are concentrated in the west, especially  near Kolhapur.  

Period 4: 1178-1279 (Maps 13-15) 

By the late twelfth century, merchant inscriptions reappear in Tamil Nadu around  the Kaveri, as in periods one and two, but are closer to Tanjavur district. They have also  expanded into the southern coast, especially the Madurai region. Inscriptions also are  recovered from northern Sri Lanka. 

Inscription locations in the Karnataka region remain constant with those from  period three, with nodes in southern Mysore, Bangalore, Davanagere, and Bijapur,  although sites have shifted slightly eastward.  

New for this period is a number of inscriptions discovered southeast of Hyderabad  in Andhra Pradesh, the most numerous of which come from Vishakhapatnam, a well known port during this time. 

Period 5: 1279-1400 (Maps 16-17)  

In the final period, we see continued development of Tamil region sites in the  Kaveri delta region and Pudukkottai district. There is also an inscriptional node east of  Coimbatore. Scattered inscriptions occur along the coastal circuit between modern  Karaikkal and Chennai. 

43 

In Karnataka, inscriptions decrease sharply, though not in the strongholds of  Mysore and Bangalore. The previous centers, Davanagere and Bijapur, are emptied and  we find the only pocket of inscriptions near Hospet.  

Inscriptions continue to be found in Andhra Pradesh, though further southwest  than in period four. 

. . . 

To summarize, epigraphical distribution varies according to time and region.  Select areas receive high amounts of mercantile patronage within different sub-periods;  dense clusters of inscriptions in these areas suggest centralized nodes i.e. “activity centers”  of mercantile activity. These include Tamil Nadu’s Kaveri and Pudukkottai regions,  southern Karnataka’s Bangalore and Mysore districts, and northern Karnataka’s  Davanagere and Bijapur districts (Map 18).  

While the reasons behind frequent mercantile migration are unclear, it is likely  that they initially “depended on states for the expansion of their operations.” Daud Ali  has suggested that merchants first migrated south from Aihole into the Tamil region  during the Rastrakuta’s tenth century incursions.51 The merchant inscriptions’ distribution  pattern then mirrors the Cholas’ rise to power, appearing first in the Chola heartland of  the Kaveri Delta Region, and expanding north into southern Karnataka at the beginning  of the eleventh century, when Rajaraja I raided Talakad, the Ganga capital in southern  

 

51 Daud Ali, “Between Market and Court: The Careers of Two Courtier Merchants in the Twelfth-Century  Deccan,” Journal of the Economic and Social History of the Orient 53 (2010): 189.

44 

Karnataka.52 Multiple eleventh century Tamil inscriptions on newly constructed Dravida  style temples attest that the southeastern Karnataka region was incorporated culturally  into Chola territory during this period. The initial eleventh century appearance of  inscriptions in Sri Lanka and southeast Asia probably resulted from Chola military  prowess: beginning in 993, Rajaraja I launched campaigns into Sri Lanka, which his son,  Rajendra I, continued during his own reign. Rajendra I expanded further east than his  father, raiding Malaysian peninsula ports in 1017 and 1025. These raids demonstrate  Rajendra I’s commitment to commercial expansion in the Indian Ocean, for by seizing  control of Malaysian ports, Rajendra I wrested power away from the Srivijayan empire (the Southeast Asian kingdom based in modern southern Malaysia), who were the Cholas’  

chief adversaries in conducting trade with China. Attesting to the fierce competition  between the two entities, Chinese records reveal the Srivijayans’ direct efforts to block  south Indian merchants from accessing Song China markets.53 

Meanwhile, merchant associations operating in Karnataka appear to have veered  away from Chola operations by the mid-eleventh century, expanding to the north.  Between the eleventh and twelfth centuries, there is a surprising dearth of inscriptions in  the Tamil region, paired with a sudden explosion of them in Karnataka. Merchants  continue to expand into Karnataka’s northern regions until the thirteenth century. Ali  suggests that many merchants were closely allied with the Hoysalas of Dorasamudra,  especially during the twelfth and thirteenth centuries, and several inscriptions attest to  

 

52 Abraham, 54. 

53 Tansen Sen, “The Military Campaigns of Rajendra Chola and the Chola-Srivijaya-China Triangle,” in  Nagapattinam to Suvarnadwipa: Reflections on the Chola Naval Expeditions to Southeast Asia, eds.  Hermann Kulke, K. Kesavapany, and Vijay Sakhuja, Singapore: ISEAS, 2009, pp. 61-75. 

45 

Karnataka based merchants building careers for themselves within royal courts. Between  the twelfth and thirteenth centuries in Tamil country, merchant inscriptions continue to  increase even as the Chola monarchy declined. Kenneth Hall posits that during this time,  the power of regional centers and local “lords” including “merchants who were able to  establish an early association with rising local magnates,” increased dramatically and  further destabilized the Chola monarchy.54 

Challenges in Defining the Ainnurruvar 

Defining the Ainnurruvar as a distinct group has advantages and shortcomings.  The main advantage is that by studying the corpus of inscriptions as a discrete unit, we  acknowledge that premodern merchant actors self-identified as part of a larger  organization, one that they imagined as continuous with local identities. In other words,  by identifying as members of the Ainnurruvar, merchants conceived of themselves as  belonging to a cognitive region that was much larger than their immediate surroundings.  The major shortcoming, however, is the implication that the organization functioned  consistently and within discrete spheres. Under scrutiny, the organization was not  centrally organized, making it difficult to describe the Ainnurruvar’s precise structure and  organization. I outline the larger issues that complicate definition of the Ainnurruvar in  point form below.  

Size. In comparison to other premodern South Indian polities, the Ainnuruvar’s  inscriptional realm is enormous and unwieldy, covering an area of around 612 miles  measuring from north to south. The northernmost inscription appears at Warangal,  

 

54 Hall, Trade and Statecraft, 4.

46 

Andhra Pradesh, and the southernmost in Tirunelveli, Tamil Nadu, with several  inscriptions appearing in Sri Lanka and Southeast Asia. Their geographic reach is  astonishing, encompassing territories associated with the Rastrakuta, Western Chalukya  of Kalyan, Eastern Chalukya, Hoysala, Ganga, Yadava, Kakatiya, Chola, Chera, and  Pandya dynasties.  

Networks. Scholars have written extensively on transactional networks in  premodern south India, identified through analyzing temple inscriptions. James  Heitzman, for example, maps the transactional network of Rajaraja I’s temple in Tanjavur  by recording the type and place of origin of numerous endowments. He finds that the  majority of donations to the temple came from within the Kaveri Delta Region, “a rough  isosceles triangle with an area of approximately 7,000 square kilometers.”55 In the  Ainnurruvar’s case, scanty epigraphical texts preclude a similar study. Isolated  inscriptions, such as at Piranmalai village in Ramanathapuram District, Tamil Nadu,  contain long lists of individual donors who identify their home towns (often with distant  places), but these few sources are insufficient for reconstructing extensive transactional  networks.56 

Authority. No Ainnurruvar inscription references a central authority (group or  individual) as responsible for regulating political and economic activities.57 Moreover,  

 

55 Donations included foodstuffs, livestock and money as well as humans, such as temple dancers, service  workers, and priests. See Heitzman, “Ritual Polity and Economy,” 51. 

56 For Piranmalai translation see P. Shanmugam and Y. Subbarayalu, “Texts of Select Inscriptions,” AMCA,  275-283. 

57 Meera Abraham provides an excellent overview of arguments for and against centralized administration.  She records that A. Appadorai, T.V. Mahalingam, and LD. Barnett, interpreted the title ‘Five Hundred,’ as  denoting 500 individuals of an “apex group” that controlled its members. K.N. Sastri merely described  them as “the most celebrated” guild. K. Indrapala interpreted them as a “loosely organized body,” without 

47 

no charter enumerates the criteria for membership in the association. 58 Numerous  inscriptions attest that Ainnurruvar members partnered with other merchant and village  organizations, recording decisions made in local villages, and, less frequently, inter regional meetings. For instance, a ca. 1200 inscription from Vahalkada in Anuradhapura  District in Sri Lanka, records that the Ainnurruvar, in partnership with several other  merchant organizations as well as a local branch of warriors, convened to honor the  latter’s heroism. The warriors had protected the townspeople from a greedy chieftan’s  continued monetary extortion, and multiple groups gathered to commemorate their  bravery by planting a stone in their honor.59 

Partnerships. Although the Ainnurruvar was the most frequently mentioned  merchant organization, it was not the only one of its kind. For centuries, Tamil merchants  had operated within diverse trading associations, several of which have been the focus of  individual studies. In particular, the Anjuvannam and the Manigramam merchant  organizations are mentioned independently in a number of inscriptions. 60 Quite  frequently, especially from the twelfth century onwards, the Ainnurruvar name would be  

  

central administration, as did G. S. Dikshit, who believed that the association “formed small and workable  federations extending over a district or two.” Abraham, 74. 

58 Two written texts, currently located in Kerala, describe or allude to merchant organization charters, but  do not mention the Ainnurruvar. The Syrian Christian church at Kottayam possesses an undated copper  plate (which interestingly references an otherwise unspecified group as the “Ārunūrruvar” i.e. the 600) that  records the Anjuvanam and Manigramam’s ‘72 rights and privileges,’ elaborating upon the organization’s  rules and regulations. At the bottom of the plate, we find individual signatures in Pahlavi, Arabic, and  Hebrew scripts. Another copper plate from Cochin in Kerala, dated 1000 CE and known as the “Cochin  Jewish copper plate of Bhaskara Ravi,” grants “the Anjuvannam’s 72 rights” to an individual named  ‘Issuppu Irappan’ i.e. Joseph Raban. In Y. Subbarayalu, “Anjuvannam: A Maritime Trade Guild of  Medieval Times,” Kaveri: Studies in Epigraphy, Archaeology and History : Professor Y. Subbarayalu  Felicitation Volume, (Chennai: Panpattu Veliyiittakam, 2001), 145-146. 

59 See Appendix 1 (28) for full translation. 

60 See Subbarayalu for an analysis of the Anjuvannam and its possible ties to the Arabian Ocean in  “Anjuvannam,” For an account of the Manigramam see Meera Abraham. 

48 

accompanied by several other merchant organizations, individuals, and institutions. In  some cases, the Ainnurruvar was clearly the predominant organization, with other  individuals and organizations listed as subsidiaries, leading N. Karashima and Y.  Subbarayalu to suggest that the Ainnurruvar was an umbrella organization that  incorporated smaller organizations. While this doubtlessly occurred, it was not always  the case, as the Ainnurruvar also appeared as an equal or lesser signatory party. Thus,  while it is certain that the Ainnurruvar sometimes incorporated other groups, it remains  unclear if these organizations were subsumed permanently, or if their listing as subsidiary  groups reflected isolated instances of inter-mercantile partnerships.61 That being said,  because of the Ainnurruvar’s extensive geographic reach and frequent interaction with  other mercantile organizations, it seems likely that their institutional practices impacted  other merchant groups. In other words, by focusing on the Ainnurruvar, I argue that we  can infer larger patterns about how merchant organizations functioned in general.  

Medieval South Indian Polity  

In Gifts of Power: Lordship in an Early Indian State, James Heitzman provides  the best summary of the century-long debate over the structure of ‘Chola polity’ (849- 1279).62 Though ostensibly concerned with the political structure of the Chola monarchy,  these debates opened conversations about many other social groups’ involvement in  regional politics, including merchants. Heitzman emerges with a description of the  

 

61 Subbarayalu and Karashima regard the Nanadesi, Padienen-vishayam, and Padinen-bhumi organizations  as the same as the Ainnurruvar. The Manigramam and Anjuvannam guilds are also included in the study,  but their inscriptions are much fewer in number. N. Karashima, “South Indian and Sri Lankan Inscriptions  Relating to the Merchant Guilds,” AMCA, 4. 

62 See Heitzman’s first chapter, “The Chola Empire: History Theory, Method,” in Gifts of Power, especially  pp. 11-20.

49 

‘Chola Empire’ that weaves together many of these arguments. His description of  medieval south Indian political structure focuses on the importance of local ‘lords,’  whose ritual practices supported the dynastic ambitions of kings. The concept of  ‘lordship,’ Heitzman writes,  

suggests that attributes of birth and merit, reinforced by ritual  demonstrations, supported landed and commercial elites who projected  themselves as a separate and superior group . . . Their authority, resting on  land and labour, received legitimation and support through ritualized  action ranging from religious donations to military campaigns. . . This  assemblage of lords, emerging primarily from the agrarian economy,  supported dynastic ambitions of kings. As the leaders of assembled lords,  kings could not easily rework institutions into centralized bureaucratic  entities.63  

In emphasizing that landed and commercial elites commanded much power, Heitzman  restores agency to a host of individuals involved in medieval South Indian politics. As  Heitzman demonstrates, the Chola kings’ political strategies varied according to time and  space, including increased centralization around 1000, when kings “intervened directly in  the extraction of agrarian produce from the central area of their empire,” thus changing  the previous relations of production.64 In response, local lords tightened control over  their land to prevent future royal taxation, eventually creating “a new level of production  and property relations which undermined [the Chola] polity.”65 Merchants, in this vision,  

 

63 Heitzman, Gifts of Power, 19. 

64 Ibid., 218. 

65 Ibid.

50 

played crucial roles in determining the fate of the Cholas by reaching into their own  networks to topple the status quo.  

 Heitzman’s model of lordship builds on Burton Stein’s previous work on the  region. Stein defines Tamilakam as a cognitive region: a territory with more or less fixed  geographic borders whose inhabitants formulate similar mythic and symbolic  associations with the land.66 Stein develops this concept in contrast to a functional 

region: the actual rather than imagined movement of individuals through the region.67 Stein argues the primary functional region of administrative and political action was not  the royal court, but the nadu, a territorial unit comprised of a group of villages.  According to Stein, each nadu handled local affairs individually, retaining a “tribe like or  segmentary character,” thus forming the basis of Stein’s theoretical model, the  segmentary state, in which “each locality was to a high degree separate and  autonomous . . . [but] the circulation of particular ritual specialists provided an  overarching ideological framework which tied hundreds of these localities into a single  loosely integrated entity.”68  

  

66 Stein in turn builds on Bernard Cohn’s concept of “historical regions.” Historical regions are  demonstrated by textual terminology such as “French pays, the German Landschaft, and the Indian desh.  Stein, “Circulation and the Historical Geography of Tamil Country.” The Journal of Asian Studies 37, no. 1  (1977), 9. See also Bernard S. Cohn, “Regions Subjective and Objective: Their Relation to the Study of  Modern Indian History and Society,” in Regions and Regionalism in South Asian Studies: An Exploratory  Study, ed. Robert I. Crane (Durham: Duke University, no. 5, 1967), 6-7. 

67 Stein’s circulatory agents are the Tamil Shaivite saints who authored devotional hymns between the sixth  and tenth centuries that describe sacred sites primarily in the Kaveri delta region of Tamil Nadu. These are  compiled in the Tevaram, in which 63 hymnists/saints (nayanmars) refer to a total of 537 shrines. 274 of  these are the hymns’ primary subjects, and 300 temples are located in the Kaveri Delta region. By mapping  the holy site locations visited by the itinerant saints, Stein argues for a functional region centered on the  Kaveri. Stein, “Circulation,” 13. 

68 Stein, “Circulation,” 16.

51 

Though scholars adamantly criticized Stein for his lack of empirical rigor, few  disagreed with the basic premise of his second proposition, that nadus were cognitively  connected through a “ritual polity.” Ritual polity derives from early pan-Indian religious  texts that describe the king’s primary obligation as upholder of dharma, or morality,  requiring gift giving to righteous persons and institutions. Through gift exchange, Stein  proposed, king and subject depended on reciprocal gift giving to ensure the dharmic 

equilibrium. Evidence for a ritual polity can be found on the most basic level of  inscriptional practice, in which donors date contributions by using one or more  combinations of the king’s name, title, and regnal year. Thus, all of the region’s  donations automatically legitimate the king’s authority by invoking his name. Ritual gift  giving between subject and king mirrored the ultimate gift exchange, which occurred  between god and devotee in the temple. In religious ceremonies, the deity receives ritual  service and offerings, which the mediating priest then returns to devotees as prasadam,  proof of god’s grace.  

Stein’s distinction between cognitive and functional networks is instructive for  studying the Ainnurruvar, as its members seem to have been cognitively and functionally  connected within a “transregional polity.” Although the Ainnurruvar had no ritual  figurehead like the Chola king, there was much emphasis on belonging to a cognitive  network that transcended regional boundaries. This is particularly apparent when  studying the Ainnurruvar eulogy (see next chapter), which describes a set of boundaries  delimiting the members’ cognitive network. These are much less straightforward than  Tamilakam’s territorial boundaries; rather than naming actual locations demarcating the  Ainnurruvar’s cognitive network, the Ainnurruvar eulogy (praśasti) claims that its 

52 

members’ geographic affiliations were infinite, coming from “the four corners of the  earth and a thousand directions.” As noted above, the Ainnurruvar’s functional networks  remain unclear due to lack of data, but several inscriptions attest that long-distance,  transregional networks sometimes connected the organization. A fourteenth century  inscription from Piranmalai in Tamil Nadu’s southern Ramanathapuram District records  many Ainnurruvar representatives gathering from all over the Tamil region and southern  Karnataka to make decisions regarding taxation rates on merchandise, for example.69 

In many other instances, however, local Ainnurruvar branches operated  independently.70 Indeed, Ainnurruvar activities were not limited to trade, and varied according to village and region. For example, several inscriptions from Jambai in Tamil  Nadu’s Southern Arcot district describe Ainnurruvar members serving as the village’s  judicial heads. A ca. 1054-5 record reports that a man was penalized a sum of 32 kasu as  a result of having precipitated a woman’s suicide. The man is identified as an  

overzealous tax collector, whose continued threats to the tax-evading woman caused her  

 

69 Y. Subbarayalu provides a full transcription in AMCA, pp.275-78, followed by a slightly modified  version of K.V. Subrahmanya Aiyer’s English translation, which appeared in ‘Largest Provincial  Organisations in Ancient India’, Quarterly Journal of the Mythic Society (new series) 45, pp. 29-47, 70-98,  270-86 and vol. 46, pp. 8-22. 

70 It seems that other professional organizations functioned similarly, seen in the example of the  Vishwakarma kula, a transregional organization of craftsmen in medieval peninsular India. Named after  Vishvakarma, the “lord of the arts” and “carpenter of the gods,” the organization first appears in twelfth  century inscriptions, comprising five socially and economically differentiated jatis, while today, the  Vishvakarma comprise a distinct caste who are “dominant in the elite and prestigious domains of practice  where the discourses of sastra are most refined.” Vijaya Ramaswamy notes that in medieval times,  “community solidarity in certain situations may have taken precedence over jati solidarity.” Affiliation  with a larger organization thus enabled Vishwakarma kula members increased mobility and power while  operating within individual regions. 

The full reference from the Mahabharata is as follows: “Viswakarma, Lord of the arts, master of a  thousand crafts, carpenter of the gods and builders of their palaces divine, fashioner of every jewel, first of  craftsmen by whose art men live, and whom, a great and deathless God, they continuously worship.”  Mahabharata, Adi Parva verses: 1.60.27, 1.60.28 and 1.60.29. Quoted in Vijaya Ramaswamy,  “Vishwakarma Craftsmen in Early Medieval Peninsular India,” Journal of the Economic and Social  History of the Orient 47, no. 4 (2004), 548. Also see Samuel Parker, “Text and Practice in South Asian  Art: An Ethnographic Perspective,” Artibus Asiae 63, no. 1 (2003), 11.

53 

to commit suicide by drinking poison.71 Another inscription, dated over two decades  later (ca. 1065-6), records that the Ainnurruvar ordered a shepherd to make a monetary  contribution to the temple, a slap on the wrist for having murdered his own wife during a  quarrel.72 In Tiruvidaimarudur, Tanjavur district in Tamil Nadu, the local merchants 

were administrators of temple donations, working alongside Brahmins (sabha), temple  priests, and clerks (devakanmis).73 Here, the Ainnurruvar demonstrated their authority by  having a maṇḍapa named in their honor.74 Such autonomy might have served as a Chola  state strategy to encourage of the merchants’ already successful operations. Although  writing at a much later time than our period, the sixteenth century Portuguese traveler  Duarte Barbosa notes that the Hindu rulers along the Malabar coast appreciated the  customs revenues of wealthy traders and thus left them alone.75 

Evaluating the inscriptional data suggests that the Ainnurruvar structure  governance emulated the larger model of lordship characterizing the Chola state. Local  Ainnurruvar branches, for the most part, operated within discrete functional networks but  were affiliated with one another through common expressions of identity and ideology.  As the next chapters show, these were not only communicated through inscriptions, but  

 

71 SII 22, no. 80. 

72 SII 22, no. 91. 

73 See inscriptions in SII 3 nos. 92 and 347. 

74 The maṇḍapa was destroyed in an early 20th century renovation. It was named “Ticaiyayirattainurruvar  maṇḍapa,” and commissioned by the kaikkolar, who probably were the Ainnurruvar’s private army. ARE 253 of 1907. See also SII 19, no. 4. Tiruvidaimarudur’s main temple, the Mahalingasvami, was destroyed  in the 20th century so I cannot evaluate the maṇḍapa’s form. For more information about mercantile  involvement in medieval Tiruvidaimarur, see Kenneth Hall, “Peasant State and Society in Chola Times: A  View from the Tiruvidaimarudur Urban Complex,” Indian Economic and Social History Review 18, no. 3  and 4 (1981): 393-400. 

75 Duarte Barbosa, A Description of the Coasts of East Africa and Malabar in the Beginning of the 16th Century, trans. Henry E.J. Stanley, (London: Hakluyt Society, 1866), 148. 

54 

through literary and artistic language. In fact, the expansiveness of the Ainnurruvar’s  imagined community (or cognitive network) was often a source of legitimacy and power.  In contrast to the Chola monarch’s ritual polity, formulated through acknowledging the  overlordship of a single ruler, the Ainnurruvar stressed their transregional roots and  cosmopolitan origins, turning the conventional order on its head. Intriguingly, their  rejection of a definite center and assertion of the organization’s porous borders allowed  them to flourish in widespread places, grafting their overarching ideology onto whichever  local base suited them. 

55 

CHAPTER 2: OF SYMBOLS AND STONES 

The previous chapter analyzed the structure of the Ainnurruvar mercantile  organization through inscriptional distribution and content. I contend that although  transregional partnerships among different Ainnurruvar branches were frequent, there  was no central regulating body, and individual branches had the autonomy to make 

decisions about business transactions and corporate partnerships. Despite the many  functionally disconnected Ainnurruvar branches, however, inscriptions emphatically  insist on the organization’s unity, describing its members as “the Five Hundred of the  Thousand Directions of the Four Quadrants of the world.”1 One of the more interesting  expressions of a unified Ainnurruvar community is a panegyric prefacing many of its  donations. Rich in literary allusion, the panegyric (praśasti) connected its members  within a cognitive region by creating a cosmopolitan identity that built on preexisting  visual and textual language from multiple registers of society.  

The Ainnurruvar praśasti appears in inscriptions discovered in the modern Tamil,  Karnataka, and Andhra regions, and is written in Tamil, Kannada, or Telugu, usually  depending on its location. It might be more accurate, in fact, to call them “praśastis,”  because they follow no identifiable ur text, and individual inscriptions present many  variants of the form. They exhibit common themes, language, tropes, and format, and  seem to have had a limited production between 1000-1300 CE. 

In crafting their praśastis, the Ainnurruvar adopted a pan-Indic literary form,  

overtly aligning themselves with royal courts, where poets had used the genre in stone  

 

1 This is the Ainnnurruvar’s full title: Nānādēsi ticai āyirataiñnūrruvar.

56 

and copper plate inscriptions as early as the first century CE to enumerate the king’s  genealogy and heroic feats. Daud Ali, for example, in studying the Ainnurruvar praśasti, has emphasized its direct borrowings from royal praśasti..2 Its close associations with  royalty are underscored by the fact that the Ainnurruvar often included the current king’s  praśasti or regnal year before inscribing their own.3 On the one hand, these borrowings  attest to mercantile efforts to elevate their status through appropriation of elite imagery  and syntax, and on the other, it gestures to real relationships connecting the two realms,  

primarily due to the royal court’s dependence on mercantile groups for luxury goods.  While the desire to emulate a royal model certainly figured into the Ainnurruvar  

praśasti’s creation, it was not the only impetus. Seen from a different angle, I suggest  

 

2 Praśastis come under the umbrella of Sanskrit kāvya (literature), and are written in Prakrit and Sanskrit.  They appear on caves, tank foundations, and pillars. Daud Ali has made several informative studies of  praśasti. On the Ainnurruvar praśasti (emphasizing connections to the Hoysala court in the medieval  Deccan), see “Between Market and Court: The Careers of Two Courtier Merchants in the Twelfth-Century  Deccan,” Journal of the Economic and Social History of the Orient 53 (2010): 185-211. On Chola praśasti see “Royal Eulogy as World History: Rethinking Copper-Plate Inscriptions in Cōla India.” In Querying the  Medieval: Texts and the History of Practices in South Asia, eds. Ronald Inden, Jonathan Walters and Daud  Ali, (New York: Oxford University Press, 2000), 169. 

3 I have opted to call the Ainnurruvar eulogy praśasti as opposed to meykkīrtti, a parallel tradition of  eulogizing kings in Tamil literature, after reading Manu Francis and Charlotte Schmid’s recent study of  Chola period meykkīrtti. Francis and Schmidt reject Sheldon Pollock’s argument that meykkīrtti “appropriated a Sanskrit aesthetic and a range of its literary models into their languages for both political  and imaginative expression,” arguing that the two genres remained distinct from one another in literary  history. Specifically, meykkīrtti “never contain genealogy and focus on only one king” and “deal almost  exclusively with conquests, while praśasti are more diverse in theme.” However, the distinctions are not  entirely transparent. Francis and Schmid also claim that praśasti “are unique pieces of poetry composed  afresh, whereas meykkīrtti are mostly found at the beginning of stone inscriptions and are available in  dozens or hundreds of copies.” Challenging this distinction is the fact that identical portions of  Ainnurruvar praśasti appear throughout the Tamil, Karnataka, and Andhra regions, adhering more to the  prerequisites of the meykkīrtti genre in this respect. There are many other overlaps, as Tamil region  Ainnurruvar frequently lifted phrases and literary devices from Chola meykkīrtti. Moreover, Francis and  Schmid specify that “praśasti are always part of a royal inscriptions whereas meykkīrtti generally are not.”  The Ainnurruvar praśasti contradicts this claim since it always introduces a non-royal donation.  

Thus there are numerous problems with the terminological choice of ‘praśasti’ for defining the  Ainnurruvar eulogy. I have chosen to call them as such because numerous Tamil texts specify that  meykkīrtti are written in āciriyappā meter (providing a link with Puram poetry from the Tamil Sangam  period), which the Ainnurruvar eulogy does not use. For more information on meykkīrtti, see Schmid and  Francis’s “Preface,” in Pondicherry Inscriptions, vol. 2 edited by Bahour S. Kuppusamy and  Vijayavenugopal, (Pondicherry: Institut français de Pondichéry; École français d’Extreme-Orient, 2006), v xxi.

57 

reading the Ainnuruvar praśasti, in its formulation and expression of a collective  mercantile identity, as actively forging connections with non-elite realms as well. As I  show in this chapter, the Ainnurruvar complemented the praśasti’s written textual content  with a material object: freestanding stones (hereafter merchant stones) which were  planted in the raw earth outside of the temple’s domain, and often carved with the  praśasti’s key phrases and associated emblems. These stones were usually placed in non 

elite locations, near tanks or canals located on the outskirts of villages, or in the middle of  agricultural fields. Their high degree of visibility is noteworthy, suggesting that they  served as visual proclamations of Ainnurruvar identity and authority to all members of  the community, many of whom would not have been allowed access to the temple. We  must remember that entrance and access level depended on criteria such as jati and birth  rite, which predetermined the worshipper’s physical relationship with the temple space by  restricting the areas s/he was allowed to enter.4 As I explore below, many of the  merchant stones bear sculpted emblems that non-elite and frequently illiterate populations would have recognized. These include emblems associated with farming, such as  ploughs and scythes, and emblems associated with warfare, such as swords, axes, and  bows. These emblems indicate that the viewing audience consisted of agricultural  laborers and soldiers. Seen from this perspective, the stones suggest that the Ainnurruvar  were intermediaries between elite and non-elite populations.  

 

4 Various agamas enumerate different sets of criteria governing temple access, some specifying the areas  that worshippers were allowed to pray from, like the ardhamandapa, mukhamandapa etc. See Richard  Davis, Ritual in an Oscillating Universe: Worshiping Siva in Medieval India, (Princeton, N.J.: Princeton  University Press, 1991), 70 and Annapurna Garimella, “Early Vijayanagar Architecture, Style and State  Formation,” Unpublished Dissertation, Columbia University, 2002, 45-6.

58 

To add another layer of complexity to this study, it is important to note that not all  merchant stones were Ainnurruvar commissions. Several stones assert in writing to have  been erected by other merchant groups, and yet, the emblems adorning them are nearly  identical to those displayed on stones explicitly commissioned by the Ainnurruvar. This  suggests that the Ainnurruvar, rather than inventing symbolic icons anew, employed  emblems that were already in use by mercantile and agrarian populations at large.  Incorporating pre-existing iconography into their group identity would have enabled the  Ainnurruvar to communicate effectively across widespread audiences, and also would  have encouraged less prominent mercantile associations to conduct business with them.  

Following Daud Ali’s direction in his interpretation of the Chola monarch  Rajendra I’s praśasti, I suggest reading the Ainnurruvar praśasti and its visual  expressions as indexing the organization’s conscious intervention and manipulation of  political and social codes. The texts should be interpreted as “the dialogical utterances of  [merchants] actively speaking to and positioning one another, rather than as static  monological documents passively expressive of some political (or social) reality.”5 In  other words, instead of reading these texts as “documentary sources,” (e.g. earlier studies  typically used praśasti to generate dynastic chronology), I analyze them as vehicles  through which Ainnurruvar members actively articulated group identity. As I show,  independent Ainnurruvar branches authored praśasti simultaneously in different places,  producing variants that innovated upon common themes, while translating them to fit  local contexts. In crafting praśasti in text and icon, the Ainnurruvar articulated beliefs  and practices its members deemed crucial, and created a powerful imagined identity.  

 

5 Ali, “Royal Eulogy as World History,” 166.

59 

Using royal rhetoric as well as abstruse symbolism, the Ainnurruvar forged a group  identity that transcended its functional borders.  

Ainnurruvar Praśasti 

The Ainnurruvar praśasti prefaces many of the organization’s donations in Tamil  Nadu, Karnataka, and Andhra Pradesh, tying the organization’s members into a single  cognitive network. The prose emphasizes the merchants’ cosmopolitan status in no  uncertain terms, and must have been instrumental in the Ainnurruvar’s self-fashioning. It  begins with an invocation for good fortune, followed by praise for the organization’s  mythic/divine origins, heroic deeds, moral qualities, learning and artistic skills, as is  typical of the genre.6 The praśasti is followed by the donation’s specificities, including  detailed information about the donation’s terms, local context, and members. They share  a common linguistic format, having a shorter Sanskrit opening followed by a single  vernacular language, Tamil, Kannada, and less frequently Telugu, depending on the  inscription’s regional location. Some diverge from the above format, adding phrases in  either the vernacular language or Sanskrit.7 Whatever the vernacular language, the  praśasti versions display a high degree of overlap, employing similar imagery and  

 

6 Richard Salomon provides an overview of the praśasti format in Indian Epigraphy: A Guide to the Study  of Inscriptions in Sanskrit, Prakrit, and the Other Indo-Aryan Languages, (New York: Oxford University  Press, 1998), 112. 

7 I add the caveat that many scholars of Tamil literature might find issue with labeling Tamil a vernacular language. Using many examples from Tamil literature, dating as early as the Sangam period, they have  argued against Sheldon Pollock’s assertion that in premodern India Sanskrit writing was associated with the preeminent form of political power, and the primary language of political elites. Identifying Sanskrit as the  hegemonic language denies the “dynamic multilingualism that has defined South India since the earliest  times.” For more information about these arguments, see Jennifer Clare’s introduction in Passages:  Relationships between Tamil and Sanskrit, Kannan, M., and Jennifer Clare, eds., (Pondicherry: French  Institute of Pondicherry; Tamil Chair, Department of South and Southeast Asian Studies, University of  California at Berkeley, 2009), xxiv. 

60 

identical phrases. Thus the length of the individual praśasti could vary widely, ranging  from a few select phrases to a lengthy paean. The following section outlines the  Ainnurruvar praśasti’s most common themes and phrases, using new translations of the  texts.8 

As an introduction to the Ainnurruvar praśasti, I analyze the first portion of an  inscription on a freestanding stone from Samuttirapatti in Madurai District, Tamil Nadu.9 The inscription dates to the mid-eleventh century and is written in Tamil and Sanskrit; I  quote it at length because it contains many elements recurring in other Ainnurruvar  praśasti versions.  

Let it be auspicious! In the 26th year of King Vikrama Cola Pandya Tevar, we who protect the whole earth and possess the 500 charters, whose chest  is adorned by Lakshmi and are the bravest on earth and descended from  the deities Sri Vasudeva, Kandali, Mulabhadra and were the children of  the goddess Sri Durga of Aiyapolil (Aihole), who became friends of the  ‘old goddess.’10 

Happily transacting in the 18 pattinam, 32 velapuram, 64 covered markets,  the chettis, the merchant chettis, the Valanjiyar (kavarai), Chamunda  Svamigals, the group of 300 warriors of sorts carrying victorious parasols,  

 

8 Unless otherwise noted, I have translated the Ainnurruvar praśasti versions with the generous assistance  of Y. Subbarayalu. Original language transcriptions can be found in Y. Subbarayalu and P. Shanmugam’s  “Appendix 2: Texts of Select inscriptions of the Trade Guilds,” AMCA, 227-283. I have used their number  

systems to identify inscriptions, e.g. AMCA 1 corresponds to inscription [1] in Subbarayalu and  Shanmugam’s appendix.  

9 The stone is currently located in the TN Mahal Museum in Madurai. I was unable to photograph it. 

10 Svasti śrī! Samasta bhuvanāśraya pañcaśata vīrśāsana laksaṇe lakshmi vakshastala puvana parkārama  śrī vāsudēva kaṇṭali mūlapattirōrpava śrī ayyapolilpura paramēcurvarikku makkaḷāki śrī palapatṭārakikkut  tuṇaivarāki. The inscription’s initial phrases are entirely Sanskrit with the exception of the Tamil dative  case ending (–ukku) after ‘Paramesvari’ and the Tamil word for ‘sons’ (makkal). AMCA 8, 232-3.

61 

the 700 swordbearers, warriors, the shop keepers on the big street, the  merchants with pasumpai (literally the sack containing merchandise), and  writers, all of whom made Kali grow emaciated and weak and eradicated  all enemies on earth in every direction, for whom charity grows, whose  fame is known everywhere, whose righteous scepter was present  everywhere, and who happily carry on the samaya dharma (merchant code  of conduct).  

The Ainnuruvar of 1000 directions, the four corners of the earth, and of  the 18 countries, the Ainnurruvar from Paniya nadu, and our sons, the  merchant warriors, and our sons the servants of patinenbhumi, and [the  warriors] met together in Paniyanadu, Rajendra Chola Valanadu, in  Rajarajapandi Nadu . . . the big assembly named for the beautiful Pandya,  under the Ainurruvan Tamarind tree. The Tisai Ayirattainurruvars’ good  deed is arranged as follows: 

The preceding sentences describe an event in which warriors in the service of the  Ainnurruvar killed several of their enemies. The conflict is described in detail,  concluding with a decision to honor the warriors by transforming their town into an  erivīra-paṭṭinam (literally, “warrior port”), a designation that would have given the city’s  warrior inhabitants special rights and a monetary reward.11 The Samutirapatti praśasti 

consists of distinct sections (delineated here by paragraph) that unfold in fluid  progression. The prose moves from the group’s universal to local associations,  

telescoping into a precise temporal moment and geographic location, where the details of  

 

11 There have been many interpretations of erivīra-paṭṭinam, such a ‘mercantile town’ (T. N Subramaniam); ‘fortified mart’ (Venkatrama Ayyar); and ‘market-towns protected by warriors (K. Indrapala).  Subbarayalu and Karashima recently formulated a new definition for erivīra-paṭṭinam in AMCA, specifying  that it “is the name of the town conferred on it by the merchants of Ayyavole (Ainnurruvar) in appreciation  of the brave deed done for them by the virakodiyar, who are stated in many inscriptions to be our  (Ainnurruvar merchants’) sons.” Karashima, “South Indian and Sri Lankan Inscriptions,” AMCA, 8. 

62 

the decision and the events it commemorates are elaborated. The first portion, written in  Sanskrit, is the most commonly repeated refrain within the Ainnurruvar praśasti corpus.  It describes the association’s divine origins and unparalleled heroism: we infer that  because they protect the whole earth, their heroism is equivalent to that of the gods. By  listing their divine ancestors (Vasudeva, Kandali, and Mulabadra), and naming the great  goddess Durga as their mother, the Ainnurruvar formulate a genealogy akin to meykkīrttis  commissioned by contemporaneous Chola kings, which claimed the dynasty’s descent  from the sun.12 The statement that Lakshmi “adorned their chests” alludes to a sexual  relationship with the goddess, a claim made by nearly all royal houses that commissioned  eulogies.13 In using the metaphor, the Ainnurruvar insert themselves into preexisting  modes of Puranic discourse, a “transcendent otherworldly set of myths employed by  kings to legitimate their authority.”14 Puranic world order depended on a paramount  overlord i.e. king, who was assisted by Vishnu, embodying the god on earth. Vishnu  favored only a single king at a time, hence to claim ownership of Vishnu’s bride,  Lakshmi, was tantamount to proclaiming overlordship of the entire universe.  

From this perspective, the Ainnurruvar’s assertion that they possess the goddess  seems to challenge the existing hierarchy. In the mid-eleventh century, the time of the  Ainnurruvar’s inscription, the Chola monarchy was at the height of its power, making the  Ainnurruvar’s challenge to its hegemony all the more intrepid. In fact, the borrowed  

metaphor was probably not intended as provocation; rather, in reproducing royal rhetoric  

 

12 For example, see H. Krishna Sastri, ed. and tr. “The Tiruvalangadu Copper Plates of the Sixth Year of  Rajendra-Chola I,” SII 3, 383-439. 

13 Francis and Schmid state that the practice of mentioning goddesses as spouses derives from Sanskrit  praśasti, even though it is common to most meykkirttis. “Preface,” Pondicherry Inscriptions, xvi.  

14 Ali, “Royal Eulogy as World History,” 177.