THE THEOLOGIZATION OF THE RAMAYANA IN SOUTH INDIA, 1250-1600

  • 11 357 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up

THE THEOLOGIZATION OF THE RAMAYANA IN SOUTH INDIA, 1250-1600

THE UNIVERSITY OF CHICAGO A DISSERTATION SUBMITTED TO THE FACULTY OF THE DIVINITY SCHOOL IN CANDIDACY FOR THE DEGREE

1,608 376 2MB

Pages 235 Page size 616.321 x 794.16 pts Year 2007

Report DMCA / Copyright

DOWNLOAD FILE

Recommend Papers

File loading please wait...
Citation preview

THE UNIVERSITY OF CHICAGO

THE THEOLOGIZATION OF THE RAMAYANA IN SOUTH INDIA, 1250-1600

A DISSERTATION SUBMITTED TO THE FACULTY OF THE DIVINITY SCHOOL IN CANDIDACY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

BY AJAY K. RAO

CHICAGO, ILLINOIS JUNE 2006

UMI Number: 3219570 Copyright 2006 by Rao, Ajay K.

All rights reserved.

INFORMATION TO USERS

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

UMI UMI Microform 3219570 Copyright 2006 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code.

ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, MI 481 06-1346

TABLE OF CONTENTS

...

LIST OF TABLES.................................................................................................... ill

INTRODUCTION..................................................................................................... 1 CHAPTER 1:R&4A AS KING, R&4A AS GOD: VXMIKI'S EPIC IN COURTLY AND TEMPLE SPHERES...............................................16 CHAPTER 2: FROM MANIPRAVXA TO SANSKRIT: TRANSLATING VIBHISANAS SURRENDER..............................................-61 CHAPTER 3: MODES OF FIGURAL READING IN R m A Y A N A COMMENTARIES............................................................................................... .I19 CHAPTER 4: THE ALLEGORY AND TOPOGRAPHY OF A R & I A MESSENGER POEM.............................................................................165 CONCLUSION..................................................................................................... .208 BIBLIOGRAPHY................................................................................................. .222

LIST OF TABLES

TABLE 1: CONNOTATIVE READING VS. SLESA ....................................153

TABLE 2: TYPOLOGICAL READING VS. ALLEGORICAL READING...............................................................................158

TABLE 3: &X$A-LIKE READING................................................................... 160

TABLE 4: FIRST SEVEN VERSES OF MEGHADUTA AND MA

SANDESA ..................................................................................... .175

TABLE 5: STRUCTURAL COMPARISON OF MEGHADUTA AND M

A SANDESA .......................................................... 177

...

Ill

INTRODUCTION

The Riimiiyana, arguably the most renowned work of Sanskrit literature, is today regarded by many as a religious text. This conception of the Riimiiyana is the result of a historical process that began in the ~rivaisnavacommunity of South India, 1250 to 1600. The ~rivaisnavas,the first religious order to venerate the hero of the epic, Riima, from the ninth century, engaged in a project I call the theologization of the Riimiiyana. The dominant reception of the Riimiiyana, dating back to the Ramiiyana's own self-presentation, was as a work of literary culture ( k w a ) . By theologization, I refer to the efforts of ~rivaisnavaintellectuals to transform this receptive history, in effect rendering the Riimiiyana into a soteriological work. The primary concern of this dissertation is to show how this project was achieved through specific interpretive techniques, both those imported from vernacular traditions and those modified within Sanskrit aesthetics.

0.1 Materials

The primary sources for the dissertation are evidence for the ~rivaisnava hermeneutic project: full-length commentaries, poetic retellings, and praise poems in Sanskrit. Among the vast corpus of ~rivaisniivaliterature dealing with the Riimayana, the work of two intellectuals, Vedanta Desika (b. 1268) and Govindaraja (mid-sixteenth century), stands as the most serious effort to engage the receptive history of the Riimiiyana as kdvya. Since my objective is to situate the work of these authors historically, I present here a somewhat detailed summary of ~rivaisnavacommentarial texts and secondary literature devoted to the Riimiiyana. A comprehensive outline of these works is presented nowhere else in the dissertation. The chronology is primarily my own. The ~rivaisnavas,a religious order devoted to the paramount overlordship of the god Visnu, are active today throughout South India, especially Tamil Nadu, Karnataka, and Andhra Pradesh, and are primary constituted of Tamilspeaking groups. ~rivaisnavaswere socially organized from the tenth century. The ideological orientation of ~rivaisnavascoalesces around the authority of both Tamil and Sanskrit sources: the devotional poetry of the Tamil aviirs, the liturgical Piificaritra Agamas, and the philosophy of Visistiidvaita. ~rivaisnava metaphysics is encapsulated in the body-embodied relationship ( s a r i r a - M bhdva): just as the body is related to the soul, so also the individual soul is related to the divine, with Visnu present in all beings as the inner controller

(antaryimi). The ~rivaisnavaconcept most important for the interpretation of the Rtimiiyana is surrender, orprapatti, which, unlike bhakti (participation in the divine), is available to everyone regardless of class or gender. The first phase of ~rivaisnavaengagement with the Ramayana was in Manipraviila (mixed Tamil and Sanskrit) commentaries on the Tamil Divyaprabandham poetry and in esoteric (rahaysa) works. All the important Divyaprabandham commentators wrote in the twelfth and thirteenth centuries: Tirukkurukai Piran Pillan (twelfth century), Nafijiyar (b. 1182), Nampillai (beginning of the thirteenth century), Vatakkutiruviti Pillai (b. 1217), and Periyavaccan Pillai (b. 1228). The authors of the rahasya works, whose ideas formed the basis for the rival Vadagalai (Northern) and Tengalai (Southern) schools, lived during the fourteenth and fifteenth centuries: Vedanta Desika and Pillai LokSciirya (b. 1205) (along with his influential commentator, Manavala Mamuni, b. 1370). Two rough contemporaries wrote rahasya works devoted entirely to the Ramiiyana: Periyavacctin Pillai's Abhayapradiinasiira and Tanislokam and Vedanta Desika's Abhayapradiinasiira. These Manipraviila works do not constitute the primary objects of inquiry. These are tradition-internal texts which were not meant to circulate widely; they evidence no concern with the status of the Ramiiyana as kdvya. I therefore discuss the Manipraviila literature only insofar as it relates to the Sanskrit commentaries.

The full-length, verse-by-verse Sanskrit commentaries produced primarily in the fifteenth- and sixteenth-centuries represent the final phase of the ~rivaisnavaengagement with the Riimiiyana. It is these sources, rather than the Manipraviila works, that provide the primary evidence for theologization. ~rivaisnavaswere the primary players in Sanskrit commentarial activity on the Riimiiyana, which may to some degree be measured by the fact that four of the six commentaries used in the preparation of the Varodara critical edition were srivaisnava. My primary focus on Govindariija's Bhaana is due to a number of factors. This commentary was a virtual compendium of previous ~rivaisnava interpretations, incorporating the comments of earlier writers verbatim directly into the body of the text. This propensity means that, in a sense, a careful study of the B h w a involves a study of all preceding ~rivaisnavaapproaches to the epic. In addition, Govindariija's BhQana is the most renowned Sanskrit commentary, both within and without the ~rivaisnavacommunity. The earliest extant Sanskrit commentary on the Riimiiyana is the Vivekatilaka by Udiili ~ a r a d a r i i j a .This commentary is wholly concerned with text-critical issues (i.e. proper readings of variants, settling the literal meaning of difficult passages, etc) and displays almost no trace of the religious affiliation of its

I have consulted two grantha manuscripts of the Vivekatilaka, R 3409 at the Madras Oriental Research Library and 9386 at the Sarasvati Mahal Library in Tanjavur.

author. In this sense it represents a line of interpretation distinct from either the early Manipraviila rahasya works or later Sanskrit commentaries. The difference between this early commentary and those that followed may indicate that more ambitious exegesis required a more or less fixed recension with editorial issues relegated to the background. In a series of brief articles, V. Raghavan traced references in the &u commentary of Vadakkutiruviti Pillai to establish the terminus ad quem of the Vivekatilaka as 1250 C E . The Vivekatilaka itself makes mention of the fact that the author was a "Cola Pandita," belonging to a group of Brahmins who served under the Cola regents as military generals (with the title, Brahma-Maharaja). It also refers to earlier commentaries, although these have not survived. At least five other Sanskrit commentaries probably antedate Govindariija. All of these commentaries were produced under the same set of socio-cultural circumstances, during the Vijayanagara period in the fifteenth and sixteenth centuries. Almost the entirety of the cryptically titled, Rdmdnujiya, a short work expounding esoteric interpretations for select verses in the mode of the rahaysa works, perhaps datable to the early fifteenth century, is incorporated into Govindariija's ~ h ~ a n Four a . ~other unpublished commentaries were not

V. Raghavan, "Uddi's Commentary on the Rtimgyana: The Date and Identification of the Author and the Discovery of His Commentary" in Annals of Oriental Research 1940-42,l-8.

available to me. These are those of Venkatakrsnadhviirin, a student of Adi Van satagopa of the Ahobila Matha (c. 1450-1500), Vaidyaniitha Diksita, also the author of the law book Smrtimukhamuktdphala (c. 1500), and Isvara Diksita's

Brhadvivarana and Laghuvivarana (1518). The only commentary rivaling Govindariija's in terms of circulation is the

Tirthiya of Mahesvaratirtha (c. 1500-1550 CE). Again, not much is known about Mahesvaratirtha except that he calls himself the son of ~iirii~anatirtha? Although there is a common misconception among traditional scholars that Mahesvaratirtha lived after Govindariija, this is an impossibility, since Govindariija himself explicitly refutes the reading of Mahesvaratirtha on 2.101.1. Mahesvaratirtha's commentary is a lengthy one and Govindariija incorporates it more fully than any other, to the extent that the Tirthiya itself constitutes a bulk of the B h m a . The only post-Govindariija commentary I have consulted is Atreya Ahobila's

Vdmikihrdayam. Based on the fact that his teacher is Pariinkusa, the sixth While there remains no doubt about the possibility of the historical Rfimfinuja authoring the text, virtually nothing is known about the actual author and the text's provenance. P. P. S. Shastri (Annals of the Bhandarkar Oriental Research Institute, 1942,413) cites 1400 as a possible date, which is corroborated by a number of the other critical editors of the Baroda edition of the Rfimfiyana, but not much evidence for this claim is provided. This same Shastri also identifies this Rfimfinuja as the Kandfidai Rfimfinuja Aiyangfir who was the raja-guru for the Vijayanagara king Sfiluva Narasimha, but Kandfidai Rfimfinuja Aiyangfir lived in the mid-sixteenth century. A more plausible explanation is that the Rluminujiya is an eponymous work titled for the school of the Rdmanujiyas, i.e. the ~rivaisnavas. M. Krishnamachariar, History of the Classical Sanskrit Literature (Madras: Vaijayanti Press, 1906) 24.

pontiff of the Ahobila Matha, the text may be assigned to the first quarter of the seventeenth century. The composition of so many of these commentaries within the Ahobila Matha indicates that this single institution, which flourished through Vijayanagara patronage, was the epicenter of the ~rivaisnava hermeneutic project. We can date Govindaraja's own monumental commentary similarly on the basis of the identity of his teacher and his citation of other contemporary figures. Govindariija credits the inspiration of Vedanta satagopa Jiyar, the fifth head of this same Ahobila Matha, who was council to the chiefs of the Niindyala family according to an inscription dated December 1548; this satagopa Jiyar probably presided at Ahobila from 1548 to 1557. Govindariija also credits the encouragement of Bhavaniiciirya, who lived during the reigns of the first Tuluva Vijayanagara kings in the early sixteenth century. Finally, Govindaraja obliquely refutes the opinions of Appaya Diksita, whose dates are uncertain but roughly coterminous with those offered here, mid-sixteenth century. Combining these various references, 1550-1575 CE appears to be the probable range for the activity of Govindariija. According to the colophons, Govindariija belonged to

Sadhu Subrahmanya Sastry, Tirumalai Tirupati Devasthanams Inscriptions: Report on the Inscriptions of the Devasthanam Collection with Illustrations (Tirupati: TI'D Religious Pubications Series, 1930) 278. K. V. Rangaswami Aiyangar, "Govindaraja,"Annals of the Bhandarkar Oriental Research Institute 1942,40-41.

the Kausika gotra and was the son of Varadariija; he also wrote a commentary on the Taittiriya Upanisad. In the invocatory verses of the Bhaana, Govindaraja says that while visiting Tirupati he dreamt that crowds of devotees instructed him to compose his commentary. Govindaraja's sectarian affiliation is unclear, which may in part reflect the liminal status of the Ahobila Matha, which became an exclusively Vadagalai institution only with the seventh head, Ahobila Jiyar, according to Vtidulaviraraghava's Triniiatpra&ottaram. Additional evidence for the Ahobila Matha's liminal status is the fact that it is the only Vadagalai institution that accepts the proto-Tengalai @u commentary. In addition to these Sanskrit commentaries, Sanskrit poems by Vedanta Desika provide further evidence for the theologization of the Riimayana. Vedanta Desika, who lived in Kiifici during the century preceding the Vijayanagara empire, was one of the seminal intellectual figures of the latemedieval ~rivaisnavacommunity whose sectarian positions provided the foundation for the Vadagalai school,. His voluminous and diverse oeuvre includes philosophical works in Sanskrit, rahasya works in Manipraviila, and poetry in Sanskrit and Tamil; altogether he is credited with composing over onehundred thirty works. Among these are a number of allegorical poems and praise-poems in Sanskrit; I examine three that deal directly with the Ramayana: the Hamsa Sandesa, Pddukdsahasra, and Mahdvlravaibhava. These works in

many ways provide a parallel to the dynamics evident in Ramayana commentaries, producing in poetic form the image of the Ramayana envisioned by the commentators. None of these sources can be understood in isolation, and so I also draw from a wide body of archeological, epigraphic, and intertextual evidence. Despite the historical significance of this material, it has received almost no previous scholarly a t t e n t i ~ n . ~

0.2 Methodology

Given that ~rivaisnavaintellectuals do not present any theoretical analysis of the principles governing their interpretation of the Rfimayana, identifying and analyzing these principles requires a methodologically grounded perspective from the outside. In an introductory essay to her translation of the Kiskindha K m a , Rosalind Lefeber exemplifies a standard philological approach to Sanskrit commentarial

'

Articles which discuss the Sanskrit RCimSyana commentaries include Krishnamachariar (1906), Raghavan (1940-42), Shastri (1942), Aiyangar (1942), G. H. Bhatt, "Ramayana Commentaries," Journal of the Oriental Institute, Baroda 14: 350-361, and Rosalind Lefeber, The Rdmdyana of Vdlmiki:An Epic ofAncient India, Volume IV:Kiskindhdkanda, ed. Robert P. Goldman (Princeton: Princeton University Press, 1994). Those treating the interpretation of the Ramayana in Tamil and Manipravda are Vasudha Narayanan, "The Riimayana in the Theology and Experience of the ~rivaisnavaCommunity" in Journal of Vaisnava Studies 2 , 4 (Fall 1994) 55-89, and Patricia Mumme, "Riimiiyana Exegesis in Tengalai ~rivaisnavism,"in Many Rdmdyanas: The Diversity of a Narrative Tradition in South India (Berkeley: University of California Press, 1991,203-215.

sources: "A close consideration of the commentator's point of view forces us to perform a kind of triangulation, to measure the intellectual and emotional distance between ourselves and the commentators, and thereby perhaps perceive as well the distance between ourselves and the text" (Lefeber, 1994). But while of heuristic value for translation, delimiting commentaries to the status of a reflection of the original fails to account for the productive work done by commentary, with the result that projects such as the one explored in this dissertation would fail even to come into view. Instead we can conceive of supplementary works such as commentaries, doxographies, anthologies, and poetic retellings as characterized to a greater or lesser degree by distance from their originals. A reader such as Govindariija operates not only as an individual but also as a member of an interpretive community. This is all the more evident with the BhQana, which, as I have mentioned, is not an autonomous work but rather an amalgamation of Srivaisnava views on the Riimiiyana. The ~rivaisnava community was constituted not only through a shared metaphysics but also as a sociological collectivity requiring initiation and display of insignia. We can conceive of the constitutive features of Govindariija's interpretive practice as the On interpretive communities, see Umberto Eco, The Limits of Interpretation (Bloomington: Indiana University Press, 1990), Umberto Eco, The Role of the Reader: Explorations in the Semiotics of Texts (Bloomington: Indiana University Press, 1984), and Hans Robert Jauss, Towards an Aesthetics of Reception (Minneapolis: University of Minnesota Press, 1972).

Ramayana as a text, the earlier receptive history of the Ramayana, the ~rivaisnavacommunity, and Govindaraja himself. Some consideration of the general characteristics of ~rivaisnavahermeneutic practice may shed light on what is distinctive about works like Govindar&jals commentary. For ~rivaisnavas,as for other Sanskrit intellectual communities, metaphysical, epistemological, and interpretive principles were brought to bear on the reading of any authoritative text, so that, even as resources from the past provided a source of innovation in the present, these principles shaped the understanding of the source. For such intellectuals, understanding and explanation were one and the same act. One of the special problems posed by Govindar&jalscommentary is the application of such an approach to the reading of narrative. Not only is the composition of a supplementary work a productive act, it also represents a concrete intervention in the world.' The methodological approach of this dissertation is a historical one, which pays special attention to the historical agency of readers. It is an excavation of the particular project in which these works were engaged-the

theologization of the RbGiyana.

This formulation is drawn from Ronald Inden's Querying the Medieval: Texts and the History of Practices in South Asia (Oxford: Oxford University Press, 2000). Here Inden develops the concept of a scale of texts from R. G. Collingwood's scale of forms; a "scale of texts" refers to the overlapping relationships, both dialectical and eristical, which allow a given text to present itself atop a hierarchical network of other texts and traditions. By composing a supplementary work atop a scale of texts, an agent (or collectivity) is able to re-orient the resources of the past toward historically concrete interventions in the world.

0.3 Chapter Outline

This dissertation is divided into four chapters. Each chapter treats aspects of the central concern of the dissertation: the specific interpretative techniques employed in the theologization of the Riimiiyana in the ~rivaisnavacommunity, 1250-1600. In the first chapter, I argue that the shift in the reception of the Riimiiyana from kdvya to theology also involved an institutional shift from court to temple. Srivaisnavas played a crucial role in the process, especially with the establishment of a royal Riima cult in the Vijayanagara empire. The primary evidence for the early reception of the Riimiiyana as kdvya are kdvya retellings and the treatment as the primeval poem ( d d i - k m a ) . The institutional locus for k&ya was the court. As an illustration of how ~rivaisnavaintellectuals sought to transform this receptive history, I compare the conception of Riima as a divine king in the epic with ~rivaisnavaphilosophical conceptions of divinity and incarnation, and I discuss Govindariija's characterization of the Rimiiyana as a work of tradition

(smrti). The assimilation of the Riimiiyana into the institutional locus of the temple begins with the imaginative affiliation of the figure of Riima with temple icons in Tamil devotional poetry and the use of Riimiiyana stories in ~rivaisnava temple oratory; it culminates in the construction of ~rivaisnavaRiima temples at

Vijayanagara in the fifteenth and sixteenth centuries. The shift from court to temple can help explain the two dimensions of the hermeneutic project of Riimiiyana commentary: the incorporation of vernacular modes of interpretation oriented within the temple (discussed in the second chapter) and the modification of the aesthetic categories of Sanskrit (discussed in the third chapter). In the second chapter, I explore the first dimension--the translation of Manipraviila (mixed Tamil and Sanskrit) into Sanskrit scholastic commentary. Manipraviila works may be characterized as the reification of temple oratory, in fact the "voice" of ~rivaisnavismas vernacular theology. The translation from Manipraviila to Sanskrit represents a transmission from a circumscribed, esoteric discourse to one that was in some sense pan-Indic. Within the rahaysa works, ~rivaisnavasapply Manipraviila techniques to the Riimiiyana, including what I call performative substitution and the use of Riimayana stories as parabolic exemplars. Govindariija and others redeploy these techniques through the filter of Sanskrit scholastic categories in new ways, as is evident in Govindariija's practice of connotative reading. As examples of the translation from Manipraviila, I focus on two sections of Govindariija's commentary: first, the eighteen traditional meanings of the Riimiiyana passed down from Sri ~ a i l a Piirna, and then Vibhisana's surrender, which ~rivaisnavasconsider to be the most significant event in the epic. As a parallel case of the incorporation of

vernacular forms into Sanskrit, I also consider the relationship of Riima praise poems in Sanskrit to temple liturgy and Tamil devotional genres. In the third chapter, I examine the second dimension of the hermeneutic project, involving the novel application of categories of Sanskrit poetics to the Riimiiyana. The specific focus is on Govindariija's readings, which closely approximate but ultimately do not conform to the normative characteristics of the category of double entendre (slesa). First I explain what slesa is and then I examine the differences between Govindariija's slesa-like readings and the definitions of slesa by Sanskrit aestheticians. Among the verses that Govindariija treats in this fashion is what is traditionally considered to be the inaugural verse of Sanskrit poetry (ma nisada). A comment on this verse is tantamount to a comment on the entire history of receiving the Riimiiyana as kdvya, and Govindariija suggestively splits the verse into separate literal and theological levels of meanings in such a way as to make it closely resemble slesa. By comparing the technique of slesa-like reading with others including connotative reading and allegorical reading, I provide a comparative grammar of Govindariija's figural reading techniques. Finally I suggest that Govindariija's decision to read something resembling a slesa into Riimiiyana verses is related to his conception of the Riimiiyana as a special kind of kdvya, one that articulates ~rivaisniivasoteriological principles. This dimension of the hermeneutic project was primarily oriented back out into the world of kavya.

In the fourth chapter I show how the various aspects of the theologization of the Ramayana examined in the first three chapters-the

intuitional shift to the

temple, the incorporation of vernacular interpretive techniques into Sanskrit, and the modification of categories of Sanskrit aesthetics--enhance our ability to perform a close reading of a different kind of text, a poetic composition: the Ha&a Sandesa of Vedanta Desika. This poem, which describes how Rama asks a swan to convey a message to Sita, is based on the model of Sanskrit messenger poems and the Sundara Kfinda narrative of Hanuman's dialogue with Sit% In the H a m a Sandesa, this swan stands allegorically for the teacher, dciirya, whom the lord (Rama) sends to recover the imprisoned soul (Sita); the path traversed covers several of the divine places (divya desas), the institutional centers of the ~rivaisnavacommunity. My analysis of the Hamsa Sandesa develops directly from the analyses of the first three chapters: 1)the topographical shift from epic space to the space of ~rivaisnavadivine places represents symbolically the shift in institutional locus from court to temple; 2) the allegorical form, almost unprecedented in Sanskrit literature, reflects the incorporation of a Manipravfila interpretive techniques into Sanskrit; and 3) as a messenger poem, the Haws Sandesa invokes and seeks to transform the generic norms of a class of Sanskrit kdvya. In a short conclusion, I present some reflections on the broader significance of this dissertation.

CHAPTER 1 R & I A AS KING, R&IAAS GOD:

VALMIKI'SEPIC IN COURTLY AND TEMPLE SPHERES

Viilmiki's Riimiiyana refers to its own composition as the first instance of poetic imagination, and Sanskrit poets and literary critics long considered it to be the primeval poem (idi-kdvya). It is this conception of the Riimiiyana as kdvya that ~rivaisnavaintellectuals sought to transform through commentaries

and poems produced from the thirteenth through sixteenth centuries. The new image of the Riimiiyana for ~rivaisnavasis best exemplified by the classification of the Riimiiyana as a work of tradition (smfli) chronicling the life of its hero, Riima, not as mere divine king but as paramount deity. In this chapter, I argue that this shift in reception also entailed a shift in institutional locus from court to temple. While the site of Sanskrit literary culture was the court, that of ~rivaisnavaintellectual activity was the temple. The interpretive strategies examined in this dissertation make sense in light of specific practices of the

temple. ~rivaisnavaswere the first religious order to assimilate the Riima story into temples, first in generic Visnu temples and later in specific temples dedicated to Riima. As the earliest example of the identification of Riima and Visnu icons, I view the ninth century Tamil poetry of Kulacekariilviir as emblematic of this process. Generic temples to Visnu were the paradigmatic site for the adaptation of the Riima story by ~rivaisnavasthroughout the period under consideration. ~rivaisnavasalso played a crucial role in the establishment of a royal Riima cult in the Vijayanagara empire, established in 1336. Although there is some evidence for the existence of earlier Riima temples from the tenth century in the Tamil region and the mid-twelfth century elsewhere, Riima worship became culturally significant only during the fifteenth and sixteenth centuries at Vijayanagara. Since the ~rivaisnavaaffiliation of nearly all the Riima temples in Vijayanagara has been largely overlooked by recent scholarship, I devote close attention to this development as a striking parallel to the hermeneutic project of Riimiiyana commentaries which took place at almost the exact time and place.

1.1 Rimayana as kavya

One key element of the narrative of the Riimiiyana itself and of all its early kdvya retellings, which we may use as a point for comparison with the later

~rivaisnavainterpretations, is the divinity of Riima. A generation of early Riimiiyana scholars argued that the very representation of Riima as divine in the epic reflects subsequent accretion. Many factors strongly vitiate against this view, including the internal logic of the narrative, where Riivana's boon requires him to be killed by a human/ divine agent, and references to Riima's identity with Visnu at key junctures in the narrative. I draw these points from the persuasive argument of Sheldon ~ o l l o c k The . ~ peculiar image of divinity depicted in the epic bears morphological affinities with other examples of the boon motif in Sanskrit literature, such as the stories of Sunda and Upasunda, Tiirakii and the birth of Skanda, and the myth of Hiranyakasipu, whereby a man-god is required to counteract the dangerous potency of an ascetic. One reason this conception of the divine has engendered such confusion is its contrast with Jewish and Christian or even later Hindu theism: the hero here is bothfilly human and also in some sense divine. Complicating matters is the fact that Riima at times himself displays ignorance of his divine nature. This image of divinity mirrored another where the agent was both a man and a god: the ideal of divine kingship. The earliest expository discussion of this ideal occurs in the Riijadharma section of the Mahiibhiirata, but the association For the most fully developed presentation of this view see Hermann Jacobi, Das Rarnayana: Geschichte und Inhalt (Bonn, F. Cohen, 1893). Sheldon Pollock, "The Divine King of the Rarnayana" in The Rarnayana of Vdrniki: An Epic ofAncient India, Ed. Robert Goldman, Vol. 3 (Princeton: Princeton University Press, 1991) 1555.

of Visnu with kingship is also a common theme in the Puranas. Following the Central Asian ~ a k a and s Kusiinas, who on the precedent of the Indo-Greeks used the title "king of kings" and "son of God," Gupta rulers implemented this ideal in the world, as evident in the reference to Samudragupta's participation in Visnu in the Allahabad Pillar Inscription of 379 CE. The conceptualization of royal power was as a series of concentric circles of overlordship emanating from the ruler and, ultimately, the divine itself; therefore any portent or disaster impacting the king, either identified with Visnu or viewed as the paradigmatic devotee of h a , also affected the kingdom.3 Rama's identification with Visnu is therefore closely related to the ideal nature of Rama's rule, rdma-rdjya. The Riimiiyana's poetic idealization of kingship provides a possible explanation for why it, rather than any other early narrative, attained the status of primeval poem (ddi-kdya), the thematic and formal model for the entire corpus of epic poetry and classical narrative representation. This conception of the Riimayana dates all the way back to Aivaghosa in the second century and is repeated later by Kiilidiisa, Bhavabhiiti, and Riijasekhara (early tenth ~ e n t u r y ) . ~ Its basis is the Riimiiyana's own self-presentation as the foundational moment

1draw this characterization of ancient Indian conceptions of divine kingship from Ronald Inden, "Kings and Omens" in Purity and Auspiciousness in Indian Society, Ed. J. B. Carman and Frederique Apffel Marglin (Leiden: E. J. Brill, 1985) 30-40. Sheldon Pollock, "Sanskrit Literary Culture From the Inside Out," in Literary Cultures in History: Reconstructions From South Asia, ed. Sheldon Pollock (Berkeley: University of California Press, 2003) 80-81.

for poetic composition, evident in Viilmiki's first verse (ma ni+ida) and in verses identifying the Riimiiyana as a kdvya (1.4.7-1.4.9). Asvaghosa's choice of the Riimiiyana as precedent in composing the oldest extant Sanskrit epic poem (mahdkwa), the Buddhacarita, is highly significant.Scholars have long noted the modeling of the Buddhacarita on Viilmiki's epic, evident in the parallels between Riima's departure from Ayodhyii and the description of Siddhiirtha's exodus from the kingdom and battle with ~ i i r a . ~ Along with the Rdmopdkhydna of the Mahiibhiirata (Mbh 3.257-76), the Buddhacarita represents the earliest evidence for the reception of the epic. The dominance of the receptive history of the Riimiiyana as kdvya is evident in the perfunctory treatment it receives in the Purhnas. The Rdmopdkhy&za itself presents a summary of the main events through near direct quotation from the North-Eastern recension, excepting the Biila and Uttara Kiindas with which

Asvaghosa's Buddhacarita is only one of many Buddhist and Jaina works which evidence familiarity with Viilmiki's account of the Rama story. Others include the Pali Dasaratha Jdtaka, which is well-known in South Asia but later (at least fifth century CE) than two Chinese translations of lost Sanskrit works (dated 251 CE and 472 CE); whereas in the Dasaratha Jdtaka, Rama, Sits, and Laksmana are siblings (children of Dasaratha), in both of the Chinese translations Rama is a Bodhisattva. The Dasaratha Jdtaka includes some verses resembling those in Valmiki's epic, but a much closer relationship is evident in the narrative of the Vessantara Jdtaka. Examples of early Jaina Rama stories include the third century Prakrit Vasudevahiq& of Sanghadasa, the third century Prakrit Paumacaiiya of Vimalasuri, and the ninth-century Sanskrit Uttarapur&zaof Gunabhadra. See G. S. Altekar, Studies on Vdlmiki's Rdmdyana (Pune: Bhandarkar Oriental Research Institute, 1987) and Margaret Cone and Richard F. Gombrich, tr. The Perfect Generosity of Prince Vessantara: A Buddhist Epic (Oxford: Clarendon, 1977). See for instance, The Buddhacarita; or, Acts of the Buddha, Ed. E. H. Johnston (Calcutta: Baptist Mission Press, 1935-1936).

it demonstrates no familiarity. But in most of the Puriinas, such as the Viiyu, Visnu, and Bhiigavata Puriinas, the Riima story is simply incorporated amid the genealogical lists. One interesting exception is the Piitala Khanda of the Padma Puriina, which recounts Riima's return from Lank2 and the celebration of the horse sacrifice, an intervening account of the life of Krsna, and a concluding discussion between ~ i v and a Riima on topics including instruction on the method of lingo, worship. This portion of the Padma Puriina, however, is likely quite late.7 The plethora of poetic retellings of the epic, almost a class of literature unto itself, provides the strongest evidence of the reception of the Riimiiyana as

kdvya. Some of the most famous include Bhiisa's Pratimdndtaka and Abhi~ekandtaka(c. 3rdcentury CE), Kiilidiisa's Raghuvamsa (c. 4"' or 5th century), the Bhattikdvya (c. 6 or 7 century), Bhavabhiiti's Mahdviracarita and

Uttarardmacarita (early gthcentury), and Muriiri's Anarghardghava (ninth or

'

I draw this conclusion from details of the textual transmission of the Padma Puriina; while Hazra lists the earliest possible date for the composition of the Piitala Khanda as 800 CE, it remains more likely that the references to ~ i v were a attached to the text at a very late date, perhaps even the early modern period. There are two recensions of the Padma Puriina, the Western (including six Khandas) and the Bengali (including five Khandas). While all published versions are based on the Western, scholars identity the Bengali as older. The picture is further complicated by the fact that the Padma Puriina comprises a number of different strata and incorporates independent mahatmyas (e.g., the Bhiigavatamiihiitmya) and praise poems. The outline of the three parts of the Patala Khanda is based on the more well-known Western recension; in the Bengali recension only chapters 29-96 and 100-122 correspond to some degree with the first and third parts of the Western recension. The intervening sections are far different, with emphasis on the story of Riivana, the burning of Tripura, and the kings of the solar dynasty. See R. C. Hazra, Studies in the Puranic Records on Hindu Rites and Customs (Dacca: University of Dacca, 1940).

tenth century). None of these radically transforms Viilmiki's narrative structure, except occasionally to elide the events of the Biila and Uttara Kiindas (as is the case with Bhiisa, Bhatti, and Muriiri), to revise ethically questionable scenes (as in Bhavabhiiti's Uttarardmacarita or Muriiri's casting of Viili as the aggressor in his fight with Riima), or to foreground later events (such as Riima and Sit3 toying with asceticism in Bhiisa's Pratimdndtaka or Mantharii as ~ u r ~ a n a k hini i disguise in Bhavabhuti and ~ u r i i r i ) Pollock . makes a quick survey of some retellings whose association with historically identifiable kings is most clear: A large number of dramas and other forms of narrative based on the Riima theme in Sanskrit, Prakrit, and regional languages were commissioned by, performed before, or indeed composed by kings over a thousand-year period; from the court of the Viitaka (less likely Kashmiri) king Pravarasena in the fifth century (Setubandha), to that of Yasovarman of Kanauj in the seventh (Rdmdbhyudaya), Bhimata of Kalaiijara in the eighth (Svapnadasdnana), Bhoja of D3rii in the eleventh (Campurdmdyana), to that of ~iviijiin the seventeenth (the Rdmdyana of ~iimdas).'

The institutional locus for Sanskrit kdvya was the court, and despite the treatment of mythic themes, the central concern of kdvya, as courtly poetry, remained eulogizing the activities of the sovereign: conquest, political intrigue, and erotic love.1Â

J. L. Brockington, Righteous Rdrna: The Evolution of an Epic (New York: Oxford University Press, 1984). Sheldon Pollock, "Ramayana and Political Imagination in India" in Journal ofAsian Studies 5 2 , 2 (May 1993): 262.

And this conception of the Riimiiyana as k w a received explicit theorization within Sanskrit aesthetics, most famously by handavardhana (ninth century) in Dhvanydloka 1.5. Here, handavardhana defines the essence of kdvya itself as poetic sentiment (rasa), or more specifically the suggestion of poetic sentiment (rasa-dhvani), by citing Viilmiki's first verse (mi nisdda, 1.2.15; I discuss the various figural readings of this verse in detail in chapter three). In his Locana commentary, Abhinavagupta teases out the implications of this key kdrika : ''That very meaning (i.e. the suggestion of poetic sentiment, rasa-dhvani) is the essence of poetry. For example, long ago the grief (soka) produced from the separation of the two herons became the poetry (slokatvd) of the first poet."11 The "grief cannot refer to Viilmiki himself but only to the heron, for aesthetic response--the experience of rasa-is rapture; Abhinavagupta further clarifies this 'grief as the stable emotion (sthdyibhdva) for the sentiment of pity (karunarasa) which is dominant in the Riimiiyana. Hence the rasa in question pertains to the entire epic and not only to this particular verse. Later we will have occasion to return to the question of the predominant rasa of the Riimiiyanalo On kavya as court poetry see David Smith, Ratndkara's Haravijaya: An Introduction to the Sanskrit Court Epic (New York: Oxford University Press, 1985) and, more recently, Daud Ah, Courtly Culture and Political Life in Early Medieval India (Cambridge: Cambridge University Press, 2004.

l1 kavyawtma sa evdrthas tathd cddikaveh purdl krauncadvandaviyogotthah sok& slokatvam @atah// Dhvanydoka of handavardhana, with the Locana commentary of Abhinavagupta and the Bdapriya subcommentary of Rgmasaraka, ed. Pt. Pattabhirgma hstri, Kashmir Sanskrit Series 135 (Varanasi: Chowkamba Sanskrit Office, 1940) 1.5. All translations in this dissertation are mine unless otherwise noted.

the most important ~rivaisnavacommentator, Govindariija, identifies this as erotic passion (smgdra) instead of pity with important implications--but what is relevant at the outset is that this identification of the Riimiiyana with kdvya is simply presumed and not in need of clarification.

1.2 Ramayana as Smrti

This dissertation examines the direct engagement of ~ r i v a i s ~ a vintellectuals a with this receptive history of the Riimiiyana as kdvya. In this section I discuss the characterization of the Riimiiyana by ~rivaisnavasas a work of tradition (smrti) and the ways in which the ~rivaisnavaconception of Rfima's divine nature as Visnu departs from the ideal of a divine king, a man-god, described above. My focus is on a few key passages in Govindariija's BhMana, the most important ~rivaisnavacommentary on the Rfimiiyana. The composition of the ~rivaisnavaRfimfiyana commentaries and other secondary literature represents a later stage of a phenomenon that begins with the treatment of the Riima story in vernacular and religious literature throughout South (and Southeast) Asia from the beginning of the second millennium. In the South, other relevant works include the putatively "lost" Miilaramdyana cited by Madhva (Miidhava, fourteenth century), Appaya Diksita's Rdmayanatdtparyanimaya (~aiva,sixteenth century), and the Niimasiddhiinta cult at Tanjavor to which Tyiigarfija later belonged (~aiva,

nineteenth century). Better known are works from the North including the Yogavdsiflha (twelfth or thirteenth century Kashmir) and the Adhydtmaramdyana (fifteenth century), which influenced Tulsi's Rdmacaritmdnas (sixteenth century) as well as later works like the Rdmatdpaniyopanisad and Rdmarahasyopanisad (sixteenth or seventeenth century) and the full-length commentary on the Riimiiyana by the great grammarian Niigesa Bhatta (eighteenth century). These latter works share an idiosyncratic Advaita philosophy, the combination of Vaisnava and Saiva elements, and the introduction of innovations in the basic plot such as the "shadow" Sitii abducted by Riivana in place of the original sitii.12 My focus on the ~rivais~avas is due in part to the precedence and sustained nature of their use of Riima stories. More specifically, it is because the ~rivaisnavas,more than any other community, engaged in the hermeneutic project I call the theologization of the Riimiiyana, which sought to transform the receptive history of the Riimiiyana as k w a . The passages I would like to look at are Govindariija's general introduction and his gloss on 1.4.7. and 1.5.1. In the first two, Govindariija makes what I view to be an implicit argument about the text's religious character, asserting that the Riimiiyana qualifies as both a k w a and a work of tradition (smrti) explicating ---

l2 For a treatment of the "shadow"Sit2 motif, see Wendy Doniger, Splitting the Difference: Gender and Myth in Ancient Greece and India (Chicago: University of Chicago Press, 1999).

the meaning of the revealed Veda ( ~ e d o ~ a b r h a n a r nThis ) . ~ ~statement of the dual status of the epic-as

both poetry and theology-serves

as a precis for the

overall modality of Govindariijalscommentary; while Govindariija affirms that the Riimiiyana is a k w a (since it "teaches like a lover" as per Mammata's taxonomy of didactic forms) he provocatively extends its status to an authoritative comment on revelation. The particular view of divinity that Govindariija and others mapped onto the figure of Riima is closely related to the central concept of surrender (prapatti). Explicit theorization of this concept begins in Manipraviila works only from the twelfth century. As mentioned in the introduction, surrender is valorized as accessible to all, unlike the performance of bhakti, which, according to ~rivaisnavas,is comprised of the observance of rituals for which only males of the highest three classes are qualified. The fourteenth-century philosophers Sudarsana Stiri and Vediinta Desika cast canonical sets of divine characteristics into the broader categories of transcendence (paratva), and accessibility (saulabhya), both of which are of significance for the hierarchical relationship between agents involved in surrender; for surrender to be efficacious, the one who gives refuge (saranya) must be both capable and willing to provide refuge. l3 +am r m y a n a m M s n a m sitdydicaritam mahatipaulastyavadham ityeva cakara caritavratahli All citations of the Rsmiiyana are from the reprint of the Venkatesvara Steam Press edition: Valmikiyaramdyana, ed. Rajendra Nath Sharma (New Delhi: Nag Publishers, 1990). (Hereafter cited in the text with the prefix, "R.) Note that Kumiirila anticipates this characterization of the Riimiiyana as traditional smrti in his seventh-century Tantravarttika.

John Carman has shown that although Ramanuja never used the terms paratva and saulabhya in this technical manner, several aspects of his presentation prefigure the later theorization.14 An examination of the lists of Visnu's auspicious qualities (kaly~nagunas)that divide into two broad categories in the introduction to the Bhagavad Gita Bhdsya serves as an epitome of the ~rivaisnavaconception of the godhead. The six qualities (sadgunas) conform neatly with the notion of the lord's transcendent nature orparatva: knowledge (jndna), strength (bala), lordship (aisvarya), energy (virya), power (sakti), and splendor (tejas). Later, another quite different set of qualities conforms to the notion of accessibility or saulabhya: compassion (karunya), gracious condescension (sausllya), protective affection (vatsalya), and generosity (auddrya); further along in the passage, Ramiinuja also lists the additional qualities of friendship (sauharda) and love ( a n ~ r d ~ aWhile ) . ~ ~ the Visistiidvaita philosophical conception of divinity rests on the metaphysics of Visnu as the ensouler of the universe (contingently identified with the world on the analogy of co-ordinate predication statements), the more basic concepts ofparatva and saulabhya are more germane to the developing theory of surrender.

l4

John Carman, The Theology of Rdmdnuja (New Haven: Yale University Press, 1974).

l5 Bhagavad Gitd: With the Bhaya of Rdmdnuja and the Tdtparyacandrika of Veddnta Desika, Ed. M. Ragachariar (Srirangam, Vani Vilas Press, 1907). Hereafter cited in the text with the prefix, "BG."

Also relevant to the ~rivaisnavainterpretation of the Riimiiyana is a distinctive philosophical understanding of avatdra. Post-Riimiinuja philosophers grouped Visnu's descents into a fourfold taxonomy: 1)vibhava, the lord's manifestation in incarnations; 2) area, the lord's presence in temple icons; 3) @ha, the fourfold Piificariitra emanation; and 4) antayiimi, the lord's presence as the inner controller within all beings. To demonstrate the relationship between these philosophical ideas and the ~rivaisnavaapproach to the Riimiiyana, we can compare Riimiinuja's commentary on Bhagavad Gitii 4.8 and Govindariija's long opening prose passage. Riimiinuja here discusses at length the purpose of Visnu's incarnation:

1) [For the protection of the good] 'Siidhus," the "good,"are the best among Vaisnavas practicing the described dharma who seek my protection. If they do not see me even for a moment they consider it to be an eternity, they find no solace in continuing to live, and their limbs become weak--but the nature of my name and acts transcends language and cognition. I protect them by allowing them to see me and my acts and to speak with me. 2) [For the destruction of the wicked] 3) [For the establishment of dharma] For reinstating the Vedic dharma, which consists of venerating me, I make myself an object of veneration. 4) [I am born in every era].16

16 sadhava uktalabanadharmasild vaisnavdgresard matsam5rayanepravfltd manndmakarmasvarUpdndm avdnmanasagocaratayd maddarsandd ,ae svdtmadhdranaposa~disukhamalabhamdnd anumdtrakalam api kalpasahasram manvdndh prasithilasarvagdtrd bhaveyuh iti matsvarUpacestitdvalokandldpddeddnena te$im pantrdpiya tadvipaiitdndm viniikiya ca ksmasya vaidikadhannasya madMdhanasvarUpasya drddhyasvarUpapradarsanenatasya sthdpandya ca devamanysyddirUpena yuge yuge sambhavdmil (BG 4.8)

What is of particular interest in this key passage is that Riimiinuja casts all the

avatdras of Visnu as efforts to implement Vedic practice, now identified with devotional worship and facilitated by the divine presence in the world. Now notice the similar reasoning that Govindariija gives for the birth of Riima: The lord of ~ r iall , his desires fulfilled, endowed with all auspicious qualities, the overlord, sat on his throne with his wives in the divine world of Vaikuntha. Those who were eternally released (nityamuktas) always served his lotus feet, but he saw those ignorant beings who also deserved to serve his feet but did not attain him, stuck to primeval matter at the time of dissolution like drops of gold stuck to honey. His mind filled with compassion, he gave them senses and bodies so that they could reach him. But as if diverted by the current of the river into the ocean with rafts meant for crossing the ocean, they became attached through their bodies to other objects. And even when he promulgated his own command in the form of the Veda ('!idsandt sdstram") so that they could discriminate between the real and the unreal, they did not respect it because of their incomprehension, false understanding, and misinterpretation. Like a king wanting to approach and discipline his subjects transgressing his command, the lord wanted to descend in the fourfold form of Riima, etc to teach living beings through his own conduct. l7 l7 Vdlmikiyardmdyana With the Commentaries of Govindardja, Rdmdnuja, Mahesvaratirtha, and the Commentary Known as Tanisloki, ed. Gagavishnu Srikrishnadasa, reprint of the Venkatesvara Steam Press Edition (New Delhi: Nag Publishers, 1994). Hereafter cited in the text with the prefix, "R." sriyahpatir avdptasamastakdmah samastakalydnagundtmakah sarvesvarah "vaikunthe tu pare loke sriyd sdrddham jagatpatih/ dste visnur acintyatma bhaktair bhdgavataih saha//" ity uktarityd irivaiku&ikhye divyaloke irimahdmanima&ape sribhiiminildbhih saha ratnasimhdsanam adhydsino nityair muktais ca nirantaraparicaryam&acarananallino pi tadvad eva svacaranayugalaparicarandrhdn api taddhindn pralayeprak@ivilindn madhficchis@magnahemakanasad~n ksmajndndn jivdn avalokya "evam samsflicakrasthe bhrdmyamdne svakarmabhih/ jive duhkhakule vimoh krpa kdpy upajdyate//" ity uktaritya dayamdnaman$a "vicitrd dehasampattir isvardya n i v e d i t u m / p ~ a m eva krtd brahman hastapddddi saniyutd//" ity uktaprakiirena mahadddisamsrstikramena te@m svacaranakamalasamdsrayanocitdni k a r a m l e v a r & t i dattvd naditarandya dattaih plavair nadiraydnusdrena sdgaram avagdhamdnesv iva tesu tair visaydntarapravanesu te@m sadasadvivecandya "sdsandc chdstram" ity uktarityii svasdsanariipam veddkhyam sdstram

The overlap between these two accounts is symptomatic of the normative dimension of the ~rivaisnavahermeneutic project, which is oriented to a systematic metaphysics and epistemology. Govindariija's etiology of Riima's status as divine emerges not from his close reading of the text, but instead from a prior Visistiidvaita conception of Visnu's incarnation as facilitating sensory experience of the divine presence. Govindariija lays out this direct superimposition in his commentary on 1.5.1, where he explicates the eighteen traditional meanings of the Riimiiyana putatively passed down from Riimiinuja's maternal uncle, Sri ~ a i l Purna. a The first of the eighteen meanings extends the six signs of what a work is about (tdtparya lingas) in Mimiimsii textual analysis to prove that the central purport of the Riimiiyana is the supremacy of Visnu. With the first of these signs, introduction (upakrama), Govindariija identifies Riima with Visnu as brahman, the universal spirit described in the Upiinisiidic quote, "Know that from which these beings are born, through which, being born, they live, and that into which they enter again. That is brahman.''18 Here we have an excellent example of the distance that must be traveled between the Visistiidvaita conception of Visnu as pravartydpi tasminn apratipattiviprattipatyanyathdpratipattibhis tair andd,rte svasdsamdtila~hinam janapadam svayam eva sddhayitum abhiyiydsur iva vasudhddhipatih svdcdramukhena tan siksayitum rdmddiriipena caturddhdvatitirsur antardmaraganaih sadruhinair abhyarthitah svfirddhakasya dasarathasya manoratham api purayitum caturdhdvatatdra. l8yato vd imam bhutani jdyante yena jatani jivanti yat prayanty abhisamvisanti tad v@inasva tad brahmal

monotheistic, universal overlord and the man-god, divine king represented in the Riimiiyana. For the third determiner, repetition (abhydsa), Govindariija seeks to show how the plot of each book of the Riimiiyana unfolds a separate aspect of Visnu's supremacy: 1) in the Biila Kiinda, Visnu creates the world; 2) in the Ayodhyii Kiinda, Visnu's causes of the sustenance of the world; 3) in the Aanya Kiinda, Visnu grants moksa (as with Jatiiyu); 4) in the Sundara Kiinda, Visnu possesses all virtues; 5) in the Kiskindhii Kiinda, Visnu's destroys all; 6) in the Yuddha Kiinda, Visnu is the object of knowledge in the Vedanta (with the praise of Mandodari); and in the Uttara Kiinda, Visnu's is the source of the creator himself. (R. 1.5.1) Govindariija effectively reduces the heterogeneity of the narrative to a set of propositions about Visnu's divine nature, shifting the frame from the text into a received conceptual schema. ~rivaisnavasaccomplished similar transformations through a variety of sophisticated strategies, including connotative reading, allegorical reading, hyperanalysis, performative substitution, and slesa-like reading, which I examine in the next two chapters.

1.3. Ramayana in Temples

As I will try to demonstrate throughout this dissertation, the novelty of these interpretive strategies is directly related to the institutional shift in the reception of the Riimiiyana from the court (the sphere of kdvya) to the temple (the sphere

of ~rivaisniivatheology).19 In chapter two especially I study a number of Sanskrit works in relation to temple liturgy, didactic oratory (upanydsa), recitation, and performance. Although a detailed analysis of the characteristics of the temple as an institution is beyond the scope of this dissertation, there are a number of studies on the topic, including Stella Kramrisch's The Hindu Temple (1980), George Michell's The Hindu Temple:An Introduction to its Meaning and Forms (1988), and the 1977 volume of the Indian Economic and Social History Review devoted to South Indian Temples, edited by Burton stein2' In the remainder of this chapter, I chart the assimilation of the Riimiiyana into temples in a series of stages: 1) the imaginative affiliation of elements of the epic narrative with existing icons of Visnu; 2) the first Riima temples; 3) Riima liturgies; and 4) the royal Riima cult at Vijayanagara. Given the relationship between the temple as an institutional locus and the interpretive practices used in Riimiiyana commentary, I discuss these developments in some detail, paying special attention to the role of ~rivaisnavasat Vijayanagara. This

19

I find Bourdieu's theory of fields of cultural production to be useful for conceiving the shift from court to temple in the reception of the Rfimfiyana. Specific fields might include religion, law, the literary or aesthetic, each containing any number of subdivisions and subfields; these fields do not stand as reified wholes but are rather permeable with overlapping boundaries. See Pierre Bourdieu, The Field of Cultural Production: Essays on Art and Literature, ed. Randall Johnson (New York: Columbia University Press, 1993). 20 Special Number on South Indian Temples, Ed. Burton Stein, Indian Economic and Social History Review 1 4 , l (Jan-March 1977). Stella Kramrisch, The Hindu Temple (Calcutta, University of Calcutta, 1946). George Michell, The Hindu Temple:An Introduction to Its Meanings and Forms (New York, Harpar and Row, 1977).

last area may provide a positive contribution to existing scholarship, since the connections between ~rivaisnavasectarian agents, Vijayanagara kings, and the construction of the first monumental Riima temples have not been adequately emphasized in previous scholarship.

1.3.1 Rama in Tamil Devotional Poetry

The first stage in drawing the Riimiiyana into the institution of the temple was to identify the figure of Riima with the worship of Visnu in existing temples. The earliest evidence for this practice is in the poetry of the aviirs, especially Tirumankayiilviir, Nammiilviir, Periyiilviir, h t i i l , and Kulacekariilviir, in both individual verses and entire songs (stanzas or decads). The poetry of the mviirs is the first known bhakti treatment of Riima stories in any language or literature that we know of, and this primacy attests to the Riimiiyana's special significance for Tamil Vaisnavas from their earliest period. Notable features of the dviir poems on Riima include: 1) evidence for a separate Tamil oral tradition of Riima stories; 2) the voicing of paternalistic and erotic love through specific characters from the epic; and, most significantly, 3) the superimposition of scenes from the Riimiiyana onto Vaisnava pilgrimage sites. Though I have drawn from Vasudha Narayanan's excellent summary of this material, I develop my discussion of Kulacekariilviir through my own close

reading of the relevant sections of the Perumdl Tirumoli with the thirteenthcentury commentary of Periyaviicciin ~ i l l a i . ~ ' Narayanan identifies four stories referred to in the a v i i r poetry which find no precedent in Viilmiki's Riimiiyana: 1)Visnu lies as a baby on Brahmii's lap and warns him not to grant Riivana's boon (cited in Poykayiilviir, Peyiilviir, and Tirumiilisai aviir, the earliest aviirs); 2) Rama touches a squirrel who assists in building the bridge to Lanka, imprinting the three marks visible on all squirrels today; 3) Sitii binds Riima inside the house with a garland of jasmine flowers; and 4) Riima shares a meal with Hanumiin. Note that all of these incidents involve Visnu's proximity to and intervention in the world (his antipathy towards Riivana, affection for the squirrel, intimacy with Sitii, and accessibility to Hanumiin). The fact that these stories are so apposite to a bhakti sensibility raises the possibility of an early Tamil Vaisnava substratum of Riima legends.

I would like to focus in more detail on a few verses from Kulacekariilviir's Perurndl Tirumoli whose novelty is difficult to overestimate. Kulacekara, a ninth century Cera king who claimed (hyperbolically) suzerainty over the Piindya and Cola regions, for the first time connected the Riimiiyana with a contemporary temple site, the Govindariija shrine at the ~ a i v center a of Cidambaram (reconfigured as Citrakiitam, the mountain where Riima and Sit&spent their

21

Narayanan, "The Riimiiyana in the Theology and Experience of the ~rivaisnava

Community," 1994.

forest exile).22 In the first half of each verse of the tenth decad of the tenth section, Kulacekara recounts events from the epic only to assert in the subsequent half that the very Riima who performed these acts stands before him at Cidambaram. As examples, we can look at two sets of verses, 10.1 and 10.3-4.23In the first half of 10.1, Kulacekara identifies the fulfillment of the gods' request (recast by Periyaviicciin Pillai into an act of surrender) as the purpose of Visnu's incarnation. In the second half, he compares Riima with the image of Govindariija: for Riima bears the distinctive lotus-eyes of Visnu (sinkan edurikarumukilai; Sktpundarikdqa) and is dark like a big black cloud that has drunk the water of the ocean leaving behind only sand. Then Kulacekara connects the presence of Riima with a soteriological objective: "he appears in Cidambaraml Citrakiita so that those who missed experiencing him in his previous avatdra can see him now." Here we have an institutional parallel to the mapping of the philosophical conception of avatdra onto Riima, since for Kulacekara the sensory experience of Riima facilitated by the Govindariija image fulfils Visnu's incarnation. And as with all the verses, the lingering last image is a statement of the aviir's own devotional experience: "When will I see

22 K. A. Nilakanta Sastri,A History of South India From Prehistoric Times to the Fall of Vijayanagar (Madras, Oxford University Press, 1976). 23 Perurndl

Tirumoli of Kulacekariifir (Kanci: Kadaikkumidam Granthamala Office, 1966).

him to the full satisfaction of my eyes?" (enro kolu kan kulirakkiinu nale). Kulacekara neatly constructs an imaginative act of worshipping Riima in a sequence of images, first through the identification of Riima with the Govindariija icon and then with allusions to the veneration of that icon by himself and other devotees. A similar structuring of half-verses can be seen in 10.3 and 10.4. In the first half of 10.3, Kulacekara ties together Riima's heroic acts of breaking &vats bow (at Sitii's svayanivara) and winning Visnu's bow (from Parasuriima), referring to him as "viran tannai'l-the

great hero. In the second half, he describes the high

walls of Cidambaram and his desire to "worship at the feet of those who worship at his feet." The first half of 10.4 lists pivotal scenes from the Ayodhyii Kiinda such as Kaikeyi's exile of Riima and Bharata's receipt of Riima's sandals; it ends with Riima approaching the Citrakiita mountain, which provides Kulacekara with a transition to the second half about Citrakiital Cidambaram. Finally, Kulacekara says suggestively that those residing in this pilgrimage site, satiating their eyes with a vision of the lord, surpass those eternally in the divine realm (the nityasaris), since the latter stray from his presence through their engagement in works. This last statement anticipates the strong Visistiidvaita view of iconography, whereby the image is neither a symbolic representation nor merely infused with divine energy through consecration but rather a full embodiment of the divine on earth as iconic incarnation or arcavatira.

The relatively insignificant status of Kampan's monumental Tamil Rgmgyana, the Iramavataram, for ~rivaisnavasalso merits attention. While an oral tradition identifies Kampan as a ~rivaisnavaand today one can find a special hall constructed for Kampan in the ~ r i r a n ~ acomplex m (a Hoysala construction of the thirteenth century), it is significant that the Iriimavatdram spawned no ~rivaisnavacommentarial tradition; moreover, the Iramavataram is rarely, if ever, cited in the rahasya literature despite the bhakti elements of Kampan's retelling, evident for example in the depiction of the Ahalya episode. A possible explanation for the absence of theological commentaries on the Irdmavatdrarn is its perceived proximity to the emerging regional court of the Colas. Some of the a v a r poets also refer to what must count as the most powerful association between the Riimayana and ~rivaisnavaiconic mythology: identifying the image of Ranganatha at ~ r i r a n ~ a m the, central pilgrimage site for ~rivaisnavas,as Riima's gift of a family heirloom (kula-dhana) to Vibhisana as described in the last chapter of the Yuddhakiinda (6.131.88). Here Rama presents gifts to Hanuman, Sugriva, Laksmana, and Vibhisana upon ascending the throne at Ayodhya. According to the oral history surrounding ~rirangam, this gift was none other than the reclining image of Ranganatha (the implication being, as stated in the Srirarigamahdtmya, that the icon remained in its current location during Vibhisana's return trip to Lanka). Periyalvar and Tontariitipoti make explicit mention of this connection. In the Periyalvdr Tirumoli, Periyalviir

says that in Srirangam the reclining Visnu "directs his flower-like eyes towards the fortified city of Lankii, for the sake of ~ i b h i s a n a . (Narayanan, "~~ 1994, 65). Similarly, in the Tirupalliyelucci, Tontariitipoti refers to the lord at Srirangam as 'the king of Ayodhyii" who "with his bow, ruined the whole clan of people at Lankii." This foundational association demonstrates just how central the Riimfiyana was to the identity of the community. Kulacekara's approach provides an early precedent for the veneration of Riima in generic Vaisnava temples. We must remember that the construction of temples dedicated specifically to Riima represents the culmination of developments in discourse and practice over several hundred years and that the radical element of the early reception of the Rarniyana in the ~rivaisnava community was precisely the identification, visually and, for ~rivaisnavas, ontologically, of Riima with the universal iconic imagery of Visnu. Throughout the period under consideration, the oral Manipraviila discourse and other vernacular forms associated with the Riimayana, as discussed in the next chapter, occurred in these generic Vaisnava temples and not in Rama temples. The identification of Riima with Visnu was therefore a crucial step in the institutional shift from court to temple.

24

Narayanan, "The Riimiiyana in the Theology and Experience of the ~rivaisnavaCommunity," 1994, 65.

1.3.2 The First Temples and Liturgies

I turn now to the Riima temples and Riima liturgies in the South prior to the monumental ~rivaisnavaRiima temples at Vijayanagara. I view this material as primarily negative evidence-the

absence of a significant royal Riima cult until

Vijayanagara. The existence of these precedents however, establishes the conceptual and practical resources from which ~rivaisnavasectarian agents and Vijayanagara regents drew. The early Cola regents were the first ever to consecrate Riima temples. As Pollock has shown, whereas scenes from the Riimiiyana appear in temple wall friezes from at least the fifth century CE, the figure of Riima was not the object of veneration-the

actual installed temple icon.25 The majority of these scenes,

especially as depicted in a number of eighth century temples in the Tamil region, are in fact ~ a i v aexamples ; include the images of Riivana shaking Kailiisa (in the Olakkannesvara temple at Mahsbalipuram) and Riivana disturbing Viili worshipping a liriga (in the Kailiisaniitha temple of ~ i i i i c i ) . ~ ~ The early construction coincides with the practice of Cola kings identifying themselves with the persona of Riima. Aditya I took the title, "Riima bearing a bow," kodada-rdma, and his son Parfintaka I referred to himself as "Riima in

25

Pollock, "Riimiiyana and Political Imagination in India," 1993,265.

26 R. Nagaswamy, " ~ rRiimiiyana i in Tamilnadu in Art, Thought and Literature" in The Rcimciyana Tradition in Asia, Ed. V. Achaean (New Delhi: Sahitya Academy, 1980) 409-429.

battle," sa@ma-riima,

during his conquest of ~ a n k a The . ~ ~very first temples

are primarily restricted to two areas: 1) villages surrounding Tanjavor, containing bronze images of kodada-rima, Rama holding a bow in his left arm and an arrow in his right; and 2) the area around Kafici in what is now Chengalpattu talak. Most are dateable to the reigns of Parantaka I (907-955) and Rajaraja '(985-1016) (as well as the intervening period of political instability). While archeologists and art-historians have devoted considerable attention to the Tanjavor bronzes,28almost nothing is known about the temples in which they were located, since these are neither attached to existing pilgrimage centers nor accompanied by adequate epigraphic evidence. Three sites closely grouped together comprise this group: Paruttiyur (c. 950 CE), Tiruccerai (end of the tenth century), and Vadakkuppanaiyur (end of the tenth century). The actual bronzes, which P. Sundaram (1963), Douglas Barrett (1965), and R. Nagaswamy (1983) have studied, are almost identical in appearance, with k o d a d a Rama accompanied to the right and left by Laksmana and Sits, wearing an ornament worn by emperors, a ritual thread, and a jeweled crown. The identification

'

Although Aditya took this name we have no temples or images from his reign. On Pariintaka as Sangriimariighava see South Indian Inscriptions, ed. E. Hultzsch and others (Chennai: Government Press, 1890-) 11, no. 76. Hereafter cited in the text with the prefix, "SII." 28 The best preserved of the bronzes, from Vadakkuppanaiyur, is now in the Madras Museum; dates are usually determined through changes in style.

between king and god emblemized in these first images appears to be a logical extension of Riima's status as the ideal divine king. As R. Champakalakshmi points out: [Riima's] exploits were still an important subject of comparison with the achievements of the kings and their subordinates. For instance the Tiruvalangadu copper plate record of the sixth year of Riijendra says that while Riighavendra (Riima) constructed a bridge across the ocean with the assistance of the monkeys and killed with great difficulty Ravana, the king of Lankii, the general of Rajaraja I crossed the ocean by ships and destroyed the lord of Ceylon and hence Riima was surpassed by this Cola general.29 The temples in Chengalpattu, on the other hand, were probably closely related to the new bhakti treatment of the epic in the h i i r poetry. These include the earliest dateable Riima temple-the

Kodandariima temple at

Maduriintakam, dedicated by Pariintaka I in 914 CE.~' Today this remains a living temple housing the image of Ayodhyii-perumal (also called "Eri kathii Riima"), and a popular local legend recounts that a colonial collector had a vision of Rama while inspecting the Maduriintakam tank during a flood.31

29 R. Champakalaksmi, Elements of Vaisnava Iconography in the Tamil Country (New Delhi: Orient Longman, 1981) 122.

' SII V, no. 126. 31 However, other than names of the deities in the Cengalpattu temples--several of which contain the affix, "perumd," the Tamil name for Visnu common in the fivfir poetry--it is impossible to assign an explicit ~rivaisnavaaffiliation in the absence of any textual evidence. The only explicit reference to an early ~rivaisnavaRfima temple is in the chronicle of ~rirangam,the Koil O m , which states that Vikrama Cola (1118-1135) had a shrine for Rfima built there. Other sites in Cengalpattu include: 1) Uttiramerur, where the wife of Aditya I1 (Karikfila) (955-969), Villfivfin Mahfidevi, set up an image of Rfima in the Tiruvfiyottiyai temple;

I point to the Cola bronzes as negative evidence because they are confined to a few sites and there is no evidence of widespread worship of Riima. Riima never attained the status as patron deity for the Cola empire, and the overall significance of the sites does not extend beyond their probable status as an 'inaugural moment" in the construction of Riima temples. The earliest sources for Riima liturgies and iconography are in the

Caturvargacint-i

liturgical manual of Hemiidri and a number of Vaikhiinasa

Agamas. The descriptions in the Vaikhiinasa sources conform closely to the

k o h d a icons in the Cola region and elsewhere.32 The liturgical manual of

2) the Raghavadeva temple at Ukkal, bearing a 988 inscription of Rajaraja; and 3) Vadamadurai, containing the Adikesava temple in which Rajaraja gifted an areca grove on the occasion of the marriage of Nabirattiyar (Sita) with Tiruvayoddicakravarti (Rama) in 1037. Also located in this area is the ancient Viraraghava temple at Tiruvallur, a ~rivaisnavadivine place (diva desa) about which Tirumangaiyiilvftr sang. Interestingly, while the deity is named llVirarfighava"the icon here is a reclining Visnu; similarly in Uttiramerur the primary icon is Sundaravarada and in Vadamadurai it is Adi Kesava. All of which suggests that as with Kulacekara's poetry, the earliest temple innovation was the identification of the figure of Rama with generic Vaisnava images. The conqueror Rajaraja also had R2ma images installed in other sections of the Tamil region, including Raghavacakravartigal in the Narasimha temple at Ennayiram (South Arcot district), the Rarnaswtimi temple at Sermadevi (Tirunelveli district), and a shrine for Hanuman in the Visnu temple at Timmalpuram near Kafici. One reason these Cengalpattu sites, located within thirty kilometers of each other, warrant special attention is the fact that Uttiramerur, established by Nandivarman Pallavamalla, 731-796, was an important settlement of Vaikhanasa Brahmins (who practiced a Vedicized form of Vaisnava liturgy). On the Cengalpattu temples, see SII 111, no. 32, SII 111, no. 262, SII 111, no. 341, SII V, no. 180, SII V, no. 335, SII 111, no. 341 and Census of India 2001: Temples of Tamil Nadu: Kancheepuram District (Delhi, The Controller of Publications, 2003). On Uttimamerur, see Fracois Gros and R. Nagaswamy, Uttaramerur: legendes, histoire, monuments, (Pondichery: Institut francais d'indologie, 1970). 32 T. A. Gopinatha Rao included in his Elements of Hindu Iconography a two-page excerpt from an unspecified Vaikhanasa source, the Bhr@ Marici Samhita contains a brief description in the section on the ten avatdras, and similar instructions occur in the Atri ~ a m h i t a Rama .~~ should appear dark, two-armed in the kodanda posture, and bearing Visnu's distinctive ~rivatsa mark and kaustubha jewel. T. A. Gopinatha Rao, Elements of Hindu Iconography (Chennai:

Hemiidri, however, warrants some attention insofar as it provides a precursor for the most important Vijayanagara royal festival, the Mahiinavami (today called Dasera). Hemiidri, the royal chaplain of the Yiidava king Riimacandra (1271-1312

CE) and his predecessor, Mahiideva (1261-1271), Vaisnavas who chose Garuda as their ensign, oversaw the construction of a number of distinctive dry masonry temples including a Riima temple at Riimtek. Riimtek is located about fortyfive kilometers from Niigpur in Maharashtra, and there are strong mythic associations for this site: it is believed to be both the place where Riima killed the h d r a ascetic ~ambtikaand also the Riimagiri where the semi-divine yaksa in the Meghaduta is exiled.33 Besides the Riima temple, the site includes the surrounding Ghatesvara, Suddhesvara, Kediira, and Anjaneya (Hanumin) temples.34 Riimacandra dedicated these temples in a stone inscription datable to the last quarter of the thirteenth century.35

Law Printing House, 1914-1916). An almost identical description is found in Atri Samhita, Samurtarcanadhikarana, Ed. P. Raghunatha Chakravarti Bhattacharya and M. Ramakrishna Kavi (Sri Venkatesvara Oriental Series, No. 6) 60: 17-25. 33 Henry Cousins, List ofAntiquarian Remains in the Central Provinces and Berm (Calcutta: Archaeological Survey of India, 1897).

34 See Omkar Prasad Verma, A Survey of Hemadpanti Temples in Maharashtra (Nagpur, Nagpur University, 1973).

35 Epigraphica Indica, Ed. V. Mirashi and L. R. Kulkarni (Calcutta: Archaeological Survey of India, 1888-) Stone Inscription of Ramagiri, pg. 7.

Hemiidrils Caturvargacintdmani quotes from a Piificariitra source, the Agastya Samhita, which describes three separate festivals-the

Riimanavami,

Riighavadviidasivrata, and ~ a h i i n a v a m i Hemiidri .~~ was a pivotal figure in the inauguration of Riima worship, especially as it seems (despite wide misconception among scholars) that this Agasfya Samhita is likely apocryphal. The Agasfya Samhita describing these rituals is not among the Piificaratra manuscripts H. Daniel Smith collected for his descriptive catalogue nearly thirty years ago. Instead, he includes an altogether different text bearing the same title; as his editor Agehananda Bharati puts it: This work on manuscript is to be clearly distinguished from another work of the same name, also accorded status in the Piificariitra literature, and which has gone through several published editions; while the homonymous work in print has attracted some attention as a Piificariitra piece because of its publication-despite the fact that it is hard-to-find, hence "rare," and not originally a part of Smith's Agama collection-yet 36 Ramanavami, the earliest of the three in the calendar during the month of Caitra, marks the birth of Rama. Occurring during the passive period of the ritual cycle in the spring, it could not have been a large-scale royal ceremony; Hem2dri1sdescription indicates instead a domestic service. The worshipper is enjoined to fast and sleep on the ground the previous night thinking of Rama, worship Rama in the morning, and then dispense gifts to Brahmins. The description of the ideal of donating a Rama image demonstrates the overall emotional tenor of Hemadri's instructions: Someone who gives an image of R2ma (srirdma-pratimd-pradah) is ensured liberation. With mantras to Rama as the only means, humbly and with bhakti one should call out to Rama and pray, "I make a donation of an image of Rama. Be pleased with my devotion, you are srirama himself." Then bathing that image with oil, offer clothes, etc, and think of Rama in the heart. Similarly, in his description of the Ramadvsdasivrata, a fast occurring on the twelfth day of the lunar fortnight of the month of Jyestha, Hemadri refers to R2ma as the "highest god" (paramam devam), enjoins worshipers to call out, "OMnamo rdghavdya," and describes the image as bearing Visnu's distinctive conch and discus. Caturvargacintdmani of Hem& (Varanasi: Kasi Samskrta Granthamala 235,1985).

that work, almost exclusively concerned with devotion to Rama, is clearly ''apocryphal" while the manuscript at hand, fragments of which had been dispersed and brought together here, is judged to be " a ~ t h e n t i c . " ~ ~ This rival version of the Agastya Samhita was probably compiled subsequent to Hemadri's citation. We can therefore say that, in effect, Hemiidri invented the Riima liturgies. The liturgies of Hemiidri are the only evidence for the impact of a non~rivaisnavatradition on the nascent Riima cult in Vijayanagara. By comparing Hemiidri's instructions for the Mahiinavami, foreign accounts of the Mahiinavami festival at Vijayanagara, and an important post-Vijayanagara liturgical manual, I will later briefly consider how these liturgies were expanded and developed at Vijayanagara.

1.4 ~rivaisnavaRama Temples at Vijayanagara

It was only during the Vijayanagara empire, established in 1336, that the Riimiiyana assumed a hegemonic position as the mythological frame for an imperial order. The monumental Riima temple at the heart of the royal center of the capital, the Riimacandra temple, and the identification between Vijayanagara kings and the figure of Riima have received a great deal of

37 The Smith Agama Collection: Sanskrit Books and Manuscripts Relating to P&icardtra Studies, Ed. Agehananda Bharati (Syracuse, NY: Maxwell School of Citizenship and Public Affairs, Syracuse University, 1978).

attention from archeologists of ~ i j a ~ a n a ~ a rYet a . ~almost ' all of this scholarship is focused on the structure of the Riimacandra temple and its significance for Vijayanagara kingship, without any consideration of the sectarian character of Riima worship at Vijayanagara. In what follows, my primary interest is to demonstrate the significant role played by ~rivaisnavasin this process. The project of promulgating a Riima cult at Vijayanagara provides a striking parallel to the hermeneutic project of Sanskrit Riimiiyana commentaries at exactly the same time and place. At least seven of the ~rivaisnava commentaries were produced in the fifteenth and sixteenth centuries, three, including the Bh-a

of Govindariija,within the Ahobila Matha, an early

Vadagalai institution heavily patronized by Vijayanagara regents. Both these projects represent late developments, ambitious extensions of ~rivaisnava temple practices dating back to at least the thirteenth century. The Vijayanagara Riima temples mark, in some sense, the culmination of the institutional shift in the reception of the Riimiiyana from court to temple. Close attention to the role of the ~rivaisnavasat Vijayanagara also provides a broader

38 See for instance A. L. Dallapiccola, John Fritz, George Michell, and S. Rajasekhara, The Ramachandra Temple at Vijayanagara (New Delhi: Manohar Publications and American Institute of Indian Studies, 1991) and John Fritz, George Michell, and M. S. Nagaraja Rao, Where Kings and Gods Meet: The Royal Center at Vijayanagara, India (Tuscon: University of Arizona Press, 1984),

historical context for situating the work of Govindariija and the other Riimiiyana commentators. The approach I employ to establish the role of the ~rivaisnavasin the development of a royal Riima cult at Vijayanagara is a diachronic perspective: this cult culminated in the fifteenth and sixteenth centuries. Although Vijayanagara was founded in 1336, much of the major construction of the Riimacandra temple itself and the entirety of the other Riima temples can only be dated to the period of the Tuluvas, when ~rivaisnavasincreasingly wielded influence in the empire. All post-Riimacandra temples bear the ~rivaisnava insignia. As further evidence of the involvement of ~rivaisnavasin the royal Riima cult, I also focus on the axial systems and circumambulatory routes which oriented the city and its environs towards the Riimacandra temple. Vijayanagara was a Saiva (likely Kiilamukha) kingdom during the first dynasty of the Sangamas, and the ensign of the Vijayanagara regents remained the local form of h a , Viriipiiksa, until the Aravidu king Venkata I1 replaced it with Venkatesvara, long after Siiluva Narasimha, Krsnadevariiya, and Acyutariiya had replaced ~aivismwith Vaisnavism--more specifically ~rivaisnavism-as the state religion. The geo-mythic associations of Vijayanagara are all from the Kiskindhii Kiinda, including the lake Pampii (also the name for the consort of Virupiiksa), the Miilyavanta mountain, the Rsyamiika hill, and the Aiijaniidri mountain, several of which later became the

locations for important Riima temples. While a few pre-Vijayanagara inscriptions refer to these associations from the eleventh century, there is no evidence of pre-existing Riima worship, barring an isolated twelfth or thirteenth century Hoysala kodada temple.39 The construction of the Riimacandra temple at the heart of the city, probably during the reign of Devariiya I (1406-1422), marks the first and most significant step in the reconceptualization of Vijayanagara kingship around the figure of Riima. All evidence points to Devariiya and the other Sangamas retaining the ~ a i v form a of Virupiiska as the titular deity for the empire, despite this new association with Riima. The famous inscription of Devariiya reads, "Just as Viini was gracious to Bhoja Riima, Trupurambii to Vatsa Raja, and Kiili to Vikramiiditya, so is Pampii now gracious to king ~ e v a r i i ~ aWhile . " ~ ~the association of the goddess with Riima is also present in the Mahiinavami festival, what is striking in this inscription is the absence of any explicit reference to Riima in a temple dedicated to his worship. This central place for Riima in a ~ a i v akingdom may be congruent with a peculiarly Saiva division of labor: while the ~ a i v aking orients himself soteriologically towards Siva as ideal devotee, he represents in his own person Riima the ideal king on earth. The division is

39 This Hoysala temple is located in Chikmanglur talak, site number 91. Op. Cit. S. V Padigar, The Cult of Vishnu in Kamataka (Ph.D. diss) 380.

' SII IX, pt. 11,no. 573.

exemplified by the respective locations of the Riimacandra and Virupiiksa temples: whereas the Virupiiksa temple was a public building, according to Fritz, Michell, and Rao, "the limited space within the principal shrine suggests a restricted use (for the kings, his priests, and perhaps, also his ministers and high

official^)."^^ In addition to the Riimacandra temple, there are records of grants by Devariiya I1 (1423-1446) to a few other Riima temples, including the Riima temple in the Advaita Raghuttama Matha at ~ o k a r n a . ~ ~ The most important festival at Vijayanagara, the Mahanavami, came to be associated with the figure of Rama. But what is unclear is when exactly this occured. This festival occurred in the month of ~riivanaduring the crucial transition from passive to active periods in the ritual cycle. Our knowledge of the martial and celebratory character of this festival, the most popular in Vijayanagara, derives almost wholly from the accounts of four foreign visitors: Nicolo Conti, Abdul Razzaq, Domingo Paes, and Fernao ~ u n e z The . ~ ~first two, those of the Italian Conti and Persian Abdur Razzak, are incomplete; in

41 Fritz, Michell, Rao, Where Kings and Gods Meet: The Royal Centre at Vijayanagara, India, 1984,149.

42 An inscription of 1433 of Devariiya I1 records a gift of a village to a temple of Riimacandra; Annual Report of South Indian Epigraphy (Chennail Caclutta: Government Press, 1887-), B. K. no. 119. Both Devarsya I1 and Mallikiirjuna patronized the Raghuttama matha at Gokarna and the temple of Rams in it. 43 Translations of the relevant portions of the first two accounts are contained in R. H. Major, Ed, India in the Fifteenth Century (Frankfurt: Institute for the History of Arabic-Islamic Science at the Johann Wolfgang Goethe University, 1994). For the latter two accounts see Robert Sewell, A Forgotten Empire; Vijayanagara (London: George Allen and Unwin, 1924) 253-264,357-360.

fact, Abdul Razzaq describes a festival spread over only three days, occuring at a different time during the year. The Portuguese visitors who provided more complete accounts visited Vijayanagara only in the sixteenth century, at the height of ~rivaisnavainfluence in Vijayanagara. It therefore remains entirely possible that consequential changes occured in the performance of the Mahiinavami from the time of the Kiiliimukha Sangama dynasty to that of the ~rivaisnavaSiiluva, Tuluva, and Aravidu dyanasties. We can note the divergences even in pre-Vijayanagara liturgical descriptions of the Mahiinavami, from ~ a i v sources a such as the Devi-bhiigavata Puriina to Vaisnava sources such as the Visnudharmottarapuriina and Hemiidri's Caturvargacintiimani, which elide the bali-diina which marks the culmination of the festival for ~aivas. The most significant innovation in the performance fo the Mahiinavami at Vijayanagara is the addition of a tenth day, Vijaya Diisami, the day of victory. After the lustration of weaponry on the ninth day, during this tenth day, on a central platform in front of the Riimacandra temple, the king identified himself with Riima's triumphant return from Ayodhyii as described in the end of the Yuddhakiinda, granted honors, and reviewed the army in an ostensive exercise of military and political power. No description of a tenth day occurs in preVijayanagara liturgical accounts, and we may even infer that the "vijaya" in the title is a not-so-suble reference to the empire of its origin. Given the paucity of evidence corroborating the performance of this festival before the sixteenth

century, the Vijaya Dasami itself could well have been a very late developement. We know that the Mahiinavami Dibba, the platform just described, was constructed in three phases, the last in the sixteenth century (again, the period of ~rivaisnavainfluence at ~ i j a ~ a n a ~ a r a ) . ~ ~ With the end of the Sangama dynasty, a discernable ideological shift occurred favoring ~rivaisnavasand the pilgrimage center of Tirupati, first with Siiluva Narasimha (1486-1493) and later with the Tuluva kings Krsnadevariiya (1509-1530) and Acyutariiya (1530-1542). The close relationship between Vijayanagara royal agents and ~rivaisnavasectarian agents is evident in events like the sacking of ~rirangamin the fourteenth century, after which the Vijayanagara general Gopaniirya triumphantly returned and reconsecrated the movable icon protected in Tirupati. As the work of V. Rangachari (1916-1918),

T. Viraraghavacharya (1953), and Arjun Appadurai has shown, Vijayanagara royal agents played a pivotal role in institutionalizing the Tengalai-Vadagalai divide from this period.45 While the emergence of Tengalai institutionalism can

44 A d a Verghese, Religious Traditions at Vijayanagara, as Revealed Through its Monuments (New Delhi: American Institute of Indian Studies, 1995) 106. 45 V. Rangachari, "The Successors of Riimiinujii and the Growth of Sectarianism among the ~ri-~aishnavas," Journal of the Bombay Branch of the Royal Asiatic Society 24 (1914-15): 277-312. V. Rangachari, "The History of ~ rVaishnavism: i From the Death of Sri Vedanta Desika to the Present Day," Quarterly Journal of the Mythic Society 7: 2 (January 1917): 106-18 and 7:3 (April 1917): 197-209. V. Rangachari, "Historical Evolution of Sri Vaishnavism in South India" in The Cultural Heritage of India, ed. H. Bhattacharya, 2nded (Calcutta: Ramakrishna Mission Institute of Culture, 1953-58,vol. 4 (1956): 163-85. T. Viraraghavacharya, History of Timpati: The Thimvengadam Temple (Tirupati: Tirumala-Tirupati Devasthanams, 1977-). Arjun Appadurai,

be traced to the formation of the ~riran~aniirii~ana Jiyar Atinam in the fourteenth century at ~ r i r a n ~ athrough m the partnership between the Vijayanagara generals Gopaniirya and Siiluva Kunta and Periya Krsnariiya Uttamaniimbi, Vadagalai institutionalism began after 1500 with Kandiidai Riimiinuja Aiyangiir. During the Siiluva and Tuluva dynasties, the same period in which the monumental post-Riimacandra Riima temples were built, ~rivaisnavasattained unprecedented influence at Vijayanagara. Almost all of the new temples were ~rivaisnava,often bearing the Tengalai or Vadagalai insignia, and several were dedicated to Srivaisnava deities: Ranganiitha, Venkatesvara, and Varadariija, along with images of the a v i i r ~ The . ~ ~most significant development was the rise of Tirupati, beginning with the change in dynasty with Siiluva Narasimha, who had been a patron of Tirupati even before becoming king with endowments through his chosen mediary, the enormously influential Kandiidai Riimiinuja Aiyangiir. Through his affiliation with the Vijayanagara regent, this Kandiidai Riimiinuja Aiyangiir was able to take charge of a Riimiinuja Kiita established for the benefit of non-Brahmin ~rivaisnava.He also became the guardian of the

Worship and Conflict Under Colonial Rule: A South Indian Case (Cambridge: Cambridge University Press, 1981). 46 Verghese,Religious Traditions at Vijayanagara, as Revealed Through its Monuments, 1995, 69-84.

gold treasury at Tirupati and even vied with the Uttamaniimbis for influence at ~riran~am.~~ The status of Tirupati continued to rise with the Tuluva emperors Krsndevariiya and Acytadevariiya. Anila Verghese notes that "out of the 1250odd epigraphs published by the Devasthiinam there are fewer than 150 records of the pre-Vijayanagara period, while 59 records are of the Sangama period, 168 of the Siiluva period, 229 of the reign of Krsnadevaraya, 251 of Acyutadiivariiya's period, 176 of Sadiisiva's reign and 192 of Venkata 11's period."48(Verghese 1995,70-71). Krsnadevariiya made Venkatesvara his patron deity, visited Tirupati seven times (as recorded in Devasthiinam inscriptions), and composed the Amuktamdyadd in Telugu narrating the life of the Tamil female saint h d i i l . Acyutariiya's regard for the pilgrimage site was so great that he had himself crowned emperor first in the presence of Venkatesvara at Tirupati and then again at Kiilahasti and at Vijayanagara. With the Aravidus, ~rivaisnavainfluence reached its apogee, epitomized by the aforementioned replacement of the ensign of Viriipiiksa with that of Venkatesvara.

47 Tirupati-Tirurnalai Epigraphic Series (Chennai: Tirupati Sri Mahant's Press and TirumalaiTirupati Devasthanams Press, 1931-38)VI, pt. 11, no. 224. 48 Verghese, Religious Traditions at Vijayanagara as Revelealed Through Its Monuments, 1995, 70-7 1.

The sudden emergence of Riima temples at this time and the simultaneous shift in the sectarian affiliation of Vijayanagara kings were not mere coincident events. We can correlate the rise of ~rivaisnavainfluence at Vijayanagara and the construction of Riima temples with the identity of the royal chaplains of these regents in the fifteenth and sixteenth century. Several of these belonged to the Tiitiiciirya family, said to be descended from ~risailaPtirna, Riimiinuja's maternal uncle who instructed him in the special meanings of the Riimiiyana (according to hagiographic accounts such as the Guruparampardprabhava). These Tiitiiciiryas remained expert redactors and exponents of the Riimiiyana. The Prapanndmfla, composed by Anantiiciirya (also a Tiitiiciirya), recounts the story of two members of this family from Ettur, Narasimhiiciirya and Rangiiciirya, who converting the last Sangama king, Viriipfiksa I1 (14651486), to Vaisnavism through their recitation of the Riimiiyana. The only attention scholars have paid to this remarkable account is to refute its historical veracity given the absence of any corroborating epigraphic e~idence.~' But the status of Prapanndmrta as a historical document may be more indirect, and in fact, deeper: this account, whatever its accuracy, provides valuable insight into ~rivaisnavaconceptualization of the relationship between Vijayanagara kingship, ~rivaisnavasectarian figures, and the Riimiiyana as medium of

49 See for instance B. A. Saletore, "Vaisnavism in Vijayanagara" in D. R. Bhandarkar Volume, ed. B. C. Law (Caclutta: Indian Research Institute, 1940) 193-195.

exchange. It demonstrates a presumed identification between the Riimiiyana narrative and ~rivaisnavatheology. Influential Tiitiiciiryas at Vijayanagara included Krsnadevariiya's guru Venkata Tiitiiciirya and Riimariiya's (1615-1633) guru Paiicamatabhafijanarn ~iitiiciirya.~~ Among the Riima temples of fifteenth and sixteenth century Vijayanagara, four in particular merit consideration: 1) the aforementioned Riimacandra temple (early fifteenth century), 2) the Miilyavanta Raghuniitha temple (sixteenth century), 3) the Pattiibhiriima temple (sixteenth century, during the reign of Acutariiya), and 4) the Kodandariima temple (early seventeenth century). What is interesting, in the first place, about these temples is their great structural homology. While the fifteenth-century Riimacandra temple does not bear the ~rivaisnavainsignia, it is the first temple ornately built in the distinctive Tamil s t y l e . In addition, in the mid-fifteenth century temple added to the Riimacandra complex during the reign of Mallikiirjuna (1447-1465), images of the aviirs and Riimiinuja appear in the reliefs of the pillars. Araviti Vengalariiju made this affiliation within the primary temple itself by installing

Epigraphia Camatica (Bangalore: Mysore Government Central Press, 1886-) XIV, Md. 115; Epigraphia Indica (Calcutta/ New Delhi: Archaeological Survey of India, 1892-) XVIII, 110111.

*'Verghese, Religious Traditions at Vijayanagara as Revealed Through Its Monuments, 1995, 47.

icons of the mviirs, as recorded on an undated inscription on the west wall of the north gateway.52 What are we to make of this development? Such explicit sectarian markers indicate that at some point in time Srivaisnavas likely took control of the Riimacandra temple, which may well have earlier been a Saiva temple given the centrality of Virupiiksa and the association of the goddess with the Mahiinavami during the Sangama period. This fact in and of itself may have far-reaching implications for our understanding of the sectarian character of Riima worship at Vijayanagara, even in what became the core temple of the empire. In contrast to the Riimacandra temple, other Riima temples at Vijayanagara have not received significant attention. Further study of the characteristics of these temples may tell us a great deal about the Riimacandra temple itself and about the nature of Riima worship at Vijayanagara. All of these temples bear either the Tengalai or Vadagalai insignia. Many of them are built on sites bearing mythic associations with events from the Riimiiyana. The Miilyavanta Raghuniitha temple, believed to be situated on the mountain where Riima stayed before the campaign for Lankii (also the setting for Vedanta Desika's Hamsa Sandesd), contains a core built around a large boulder which dates to the Sangama period, but the temple structure itself is from the sixteenth century. In

52 Vijayanagar: Progress of Research, ed. D. V . Devaraj and C. S. Patil (Mysore: Directorate of Archaeology and Museums, 1983-84), no. 67.

this large complex, images of the aviirs, Riimiinuja, and the Tengalai insignia appear in several places on the pillars. The Pattiibhiriima temple, actually the largest extant Riima temple in the city, was dedicated by Acyutariiya according

~ these two temples, the Kodandariima to an inscription dated 1 5 3 9 . ~Unlike temple bears the Vadagalai insignia, and although here as well the images were carved from a boulder, the temple itself probably dates to the very last period before the collapse of the empire; it is said to be the site for Sugriva's coronation. In all there are eight extant Riima temples at Vijayanagara. We can also note that there is evidence of the endowment of these temples by a variety of groups, including royal agents, subordinate rulers, private citizens, and merchant guilds, indicating the possibility that the cult of Riima had a life of its own in addition to its role in the ideology of kingship. One feature of these newly built temples-their

location-may

provide the

most telling sign of the relationship between the ~rivaisnavasand Riima worship at Vijayanagara. As Fritz, Michell, and Rao have shown, the arrangement of temples was organized into axial systems and circumambulatory routes establishing "the importance of the Riimacandra Temple as the nucleus of the royal center." Given the status of the Riimacandra temple as the private shrine of the king, these axial systems transformed the geography of the city itself into

' SII IX, pt. 11, no. 595.

an emblem of the identification between king and god. The groundbreaking work of Fritz, Michell, and Rao in tracing these axial systems has rightly received a great deal of attention and shaped scholarly conceptions of the structure of Vijayanagara as a city. But what is also important is that Fritz, Michell, and Rao identify the mythological associations of surrounding points such as the Tungabhadra River, Matanga Hill, the Mdyavanta Hill, noting that ''significantly,these prominent natural features are associated with episodes from the Ramayana epic; that is, the story of the god to whom the temple is dedicated" (Fritz, Michell, Rao 1984, 149). While some of these associations antedate Vijayanagara, the layout of axial systems could only have been a product of a historical process in which Srivaisnavas played a crucial role. Several of these sites, such as Mdyavanta, came to prominence only as ~rivaisnavatemples in the fifteenth and sixteenth centuries. ~rivaisnavas,therefore, were involved in the construction of the landscape of the Vijayanagara capital as a virtual theophany. The mapping of the identification of the figure of Rama and the Vijayanagara king with the layout of the city was not, therefore, a mere synchronic fact of the Vijayanagara world but rather the result of a project on the part of both regents and sectarian agents. Much more empirical work must be done to uncover how ~rivaisnavaswere involved in the promulgation of Rama worship in Vijayanagara. Another

avenue of future research is the exact relationship between these twin ~rivaisnavaprojects of the Vijayanagara period, the establishment of monumental Rama temples and the composition of monumental Ramayana commentaries. We may note that there is clear evidence for the close connection between the Ahobila Matha, in which Govindarfija's commentary and others were composed, and Vijayanagara from the time of Siiluva Narasimha (1486-91), who was a devotee of the Narasimha image at Ahobila. Further examination of patterns of Vijayanagara patronage may shed light on the convergence between the building of temples and writing of commentaries in the fifteenth and sixteenth centuries.

1.5. Conclusion

In this chapter I have tried to argue a number of points. First, although the conception of Rama as divine is a crucial aspect of the idealization of divine kingship in the epic narrative, the dominant early reception of the Riimiiyana was as kivya. It was only in the late-medieval period that religious ordersespecially the ~rivaisnavas--conceived of the Riimiiyana as a different sort of text, a work of tradition (smfli). This shift in reception also involved a shift in institutional locus from court to temple. The new use of the Rarnayana by ~rivaisnavaswas predominantly in generic Vaisnava temples, as exemplified in Kulacekariilviir's identification of Riima with the Govindariija image at

Cidambaram in the ninth century, and was marked by the application to the Rama story of practices of the temple, such as oral pedagogy, liturgy, and vernacular devotional genres. In the next two chapters, I examine the ways the integration of these practices of the temple are related to the specific interpretive strategies used in Sanskrit Riimiiyana commentaries, in the effort to transform the receptive history of the Riimiiyana as kdvya.

CHAPTER 2 FROM M A N I P R A V ~ ATO SANSKRIT: TRANSLATING VIBHISANA'S SURRENDER

The epic representation of the surrender of Riivana's brother, Vibhisana, to Riima is in the language of calumny, political intrigue, and strategic policy. Whereas the structuring of this incident provides a showcase for Riima's ideal embodiment of benevolent royal virtues, for ~rivaisnavasit takes on far greater significance as the exemplar for soteriological conceptions of surrender. No other event, including the banishment of Riima, the abduction of Sitii, or the killing of Riivana, merits such rigorous attention. Vibhisana's surrender also marks the place where Govindariija borrows most heavily from the Manipraviila works of Periyaviicciin Pillai and Vedanta Dedika. This chapter uses Vibhisana's surrender as a focal point for examining the larger issues of the relationship between Manipraviila and Sanskrit commentary, the situatedness of Manipraviila discourse within the institutional locus of the temple, and the specific transformations requisite for the integration of the

modes of vernacular interpretation into the modes of Sanskrit. I begin by characterizing the Manipraviila linguistic register itself as the emblematic ''voice" of ~rivaisnavavernacular theology and proceed to close readings of the exegesis of Vibhisana's surrender, the eighteen traditional meanings of the Ramfiyana from ~rifiaila~urna, and the parallel integration of vernacular forms evident in Vedanta Desika's Rama praise-poems. The institutional shift into the temple, discussed in the previous chapter, explains to a large degree the vernacular forms that ~rivaisnavasadapted to the Riimiyana. Texts were used in specific ways in temples, as sources for oral pedagogy, liturgy, recitation, and performance. In this chapter, the sections dealing with the eighteen meanings and Vibhisana's surrender illustrate one type of integration: the translation and redeployment of forms of oral discourse into scholastic prose. The section dealing with Rama praise-poems illustrates a different type of integration: the stylization in Sanskrit kdvya of the discourse of liturgy and genres of Tamil devotional poetry. This integration of vernacular forms represents one dimension of the hermeneutic project of theologization, the other being the novel use of the categories of Sanskrit aesthetics, which I examine in the next chapter. The translation of Manipravala into Sanskrit captures the historical and logical trajectory of ~rivaisnavacommentarial writing on the Ramayana. The earliest theological treatments of the Riimayana within the ~rivaisnava

community are Manipraviila texts: the Divyaprabandham commentaries and the esoteric rahasya works of Pillai Lokiiciirya, Periyaviicciin Pillai, Vedanta Desika, and Manaviila Miimuni. All of the important Divyaprabandham commentators wrote in the twelfth and thirteenth centuries, and the rahasya authors lived during the fourteenth and fifteenth centuries. The composition of full-length commentaries incorporating these theological readings into Sanskrit represents the final phase of ~rivaisnavahermeneutic activity directed towards the Riimiiyana. Nearly all of these commentaries (six of which are extant) date to the sixteenth century, the period which witnessed the construction of all four ~rivaisnavaRiima temples in Vijayanagara, the rise of ~rivaisnavainfluence at Vijayanagara, and the patronage of ~rivaisnavamonastic institutions such as Ahobila where most of the commentarial writing took place. In this sense, the translation from Manipraviila marks a more profound historical shift. This shift entailed a widening of audience from a restricted one to one that was in some sense universal, and with it an effort to transform received conceptions of the Riimiiyana. Attention to this incorporation of the modes of temple oratory into Sanskrit may help elucidate the dominant theological reading practice employed by ~rivaisnavas,whereby esoteric interpretations are inserted directly within the literal level of meaning; this practice I call connotative reading over and against the overt slesa-like reading examined in the next chapter.

2.1. Manipravala as the 'Voice"of ~rivaisnavaVernacular Theology

2.1.1. What is Manipravala?

The term "Manipraviila"refers to the combination of Sanskrit with South Indian languages (Tamil, Malayalam, Kannada, and Telugu) on the metaphor of coral (pravda) inlaid with pearls (mani). Within these regional literatures, various patterns of combination receive metalinguistic theorization as distinct language phenomena, though elsewhere forms of discourse bearing the marks of such linguistic hybridization are only retrospectively categorized as Manipravala; the ~rivaisnavaManipraviila literature belongs to the latter class. The commentators on the Divyaprabandham and rahasya authors never use the term "Manipraviila,"despite the fact that their works represent the most extensive examples of such literature in ~ami1.lThe differences between the Manipraviila of the ~rivaisnavasand other more explicitly theorized forms of Manipraviila may facilitate a more nuanced understanding of how the ~rivaisnavasmade these linguistic patterns their own. Although the earliest examples of Manipravala-style discourse in Tamil appear in Pallava inscriptions dating to the fifth century, no reference to mixed K. K. A. Venkatachari, "Observations," Journal of Oriental Research, Madras 1994-1997: 9497) 95.

language occurs in the early classification of linguistic borrowing in the classical Tamil treatise on grammar and poetics, the Tolk@piyam, produced at almost the same time. Rather, the Tolkdppiyam divides words in Tamil into a fourfold taxonomy: 1)Tamil words of everyday usage (iyar col); 2) Tamil words found only in literary works (tiri col); 3) words borrowed from neighboring regions (ticaic col); and 4) words from Sanskrit (vatu col). This categorization precludes Manipraviila, since in the Tolkdppiyam Sanskrit vocabulary is normatively adapted to the phonemic features of Tamil (absence of aspiration, consonant clusters, etc). In contrast, later Manipraviila is characterized by the preservation of the phonemic qualities of Sanskrit words. The first reference to the term, "Manipraviila,"in Sanskrit is found in Abhinavagupta's tenth-century Bhdrati commentary on the Nfitya ~iistra.Even though Abhinavagupta's Kashmir milieu may seem a distant site for awareness of southern language phenomena, Abhinavagupta's integration of known forms of linguistic hybridity within the technical categories of Sanskrit grammar provides an important emic account of what Manipraviila is. Abhinavagupta defines Manipraviila as the mixture of Sanskrit nominal stems (prdtipadikas) and regional language forms (deiibhdsyd). This reference occurs in the context of the preferable language for the introductory stanzas in dramas (dhruva songs), identified by Abhinavagupta as either "half-Sanskrit" (ardha-samskflam) or a

Priikrit like ~ a u r a s e n i Here . ~ we get a restriction of the use of the term Manipraviila specifically to the South: whereas in Abhinavagupta's native Kashmir half-Sanskrit is called htakulam, this same entity is called is called Manipraviila when the combination is with southern languages. Abhinavagupta's brief definition antedates the earliest ~rivaisnavaManipraviila literature, though it conforms well to the joining of Sanskrit verbal and nominal stems and compounds with Tamil endings found therein. The reference may be to Malayalam Manipraviila (which includes specific varieties used by temple drummers) or to other early forms of Tamil Manipraviila.

2.1.2. Manipravala as a Cosmopolitan Vernacular

One way to think of the cultural novelty of Manipraviila discourse is as a special case of what Pollock has called the cosmopolitan vernacular. Pollock charts the remarkable rise of local languages over against Sanskrit in the sphere of literary culture from the end of the first millennium to 1500, first in the 'workly," literary portions of inscriptions and then in the production of monumental works of poetry and elaborate theorization of vernacular literary forms. What is especially interesting in the challenge posed to the cosmopolitan

Nafyascistra with the Commentary ofAbhinavagupta, ed. Manavalli Ramakrishna Kavi (Baroda: Central Library, 1926-64).

discourse of Sanskrit by this process of vernacularization is, as Pollock says, that it "typically takes the form of domesticating the literary apparatus (themes, genres, metrics, lexicon) of the superposed cultural formation that set the rules of the literary game."3 The appeal of such a literary apparatus was at least in part because it supplied the most powerful conceptual equipment available for the political and cultural world of the emerging regional courts. This picture is complicated significantly in the Tamil region by the fact that Tamil, unlike any of the other vernacular languages, possessed a classical literature--the Sangam corpus--and sophisticated theorization of grammar and literature beginning with the fourth century Tolk~ppiyam,rendering the incorporation of aesthetical and grammatical categories from Sanskrit unnecessary and redundant. And yet, this is exactly what takes place in the aforementioned eleventh-century Viracoliyam, modeled closely on Dandin's K&yadar,fa in its adaptation of the figures of sense (arthdamkaras) to specific examples of Tamil literature. In characterizing Manipraviila as a cosmopolitan vernacular, we may turn to a striking aspect of the first definition of Manipraviila in Tamil, which occurs in the Viracoliyam, namely that it appears to refer to courtly poetry: unlike the

Sheldon Pollock, "The Cosmopolitan Vernacular," Journal ofAsian Studies 57,l (February, 1988): 8.

mere adoption of Sanskrit phonemes called viraviyal, Manipriiviila verse is defined as a mixture of Tamil and Sanskrit words without the presence of the etukai second syllable rhyme of the lines of couplets.4 Yet we have no extant Tamil Manipraviila poetry from this period. The absence of an enduring Tamil Manipraviila literary tradition contrasts with the Manipraviila literature emerging in Kerala from the eighth or ninth century, which is elaborately theorized in the fourteenth-century (Sanskrit) Lildtilakam. Here Manipravda is not a marginal form of literary composition, but rather a way of distinguishing the amalgamation of Sanskrit into the Western dialects eventually termed Malayalam from the standardized Tamil of the East. This Malayalam Manipraviila may provide the best material for retrospectively reconstructing the nascent Tamil Manipraviila poetry referred to in the Viracoliyam. Given the presence of this earlier literature, the use of Manipraviila by ~rivaisnavasand Jainas may be seen as a derivative practice, the self-conscious adoption of a courtly cosmopolitan vernacular by specific religious communities. We may go further and point out that this religious dimension appears to be an aspect present even in the eleventh-century Viracoliyam itself. Anne Monius has persuasively argued, through a close reading of Puttamittiran's autocommentary on the Viracoliyam, that the Buddhist affiliation of its author, its K. K. Raja, "Malayalam and Manipraviilam" Journal of Oriental Research, Madras 1994-1997: 64-67, 101.

almost exclusive citation of Buddhist literature, and claims for Tamil's status as the language of Avalokitesvara indicate that the text's primary function was to provide a literary "technology"for a cosmopolitan Buddhist community of readers, which allows for a standard form of analysis of post-classical Tamil Buddhist literature (like the Manimekhalai and Cilappatikdram) and Sanskrit sources. Hence what is at stake in this particular Tamil theorization of the cosmopolitan vernacular is the forging of a distinct literary identity for a religious community. But Monius herself acknowledges that the Viracoliyam at the same time conforms to Pollock's attention to the medieval vernacular court, as it is named for the Cola patron Virarajendra Cola (1063-1069), standardizes the grammar of what appears to be an inscriptional language, and plays out regional cultural politics in the mode of Kampan's ~ramdvatdrarn.~ The Manipravala of the ~rivaisnavasrepresents a highly specified usage proper to the ~rivaisnavatemple, a usage which, even more so than the anthology in the Viracdiyam commentary does for Buddhists, provides a distinctive "voice" for the ~rivaisnavacommunity. The inscription of this new voice, first by Tirukkurukair Piran Pillai in the eleventh century and Parasara Bhatta in the twelfth, marks the earliest use of Manipravda in theological Anne Monius, "The Many Lives of Dandin: The Kdvyddarsa in Sanskrit and Tamil," International Journal of Hindu Studies 4:1 , April 2000, 1-37. Monius, "The Many Lives of Dandin: The Kdvyddarsa in Sanskrit and Tamil," 2000,29-30.

literature. Conceiving of Manipraviila as merely the reification of ~rivaisnavism as the end of the two Vedas, Tamil and Sanskrit (ubhayaveddnta, itself a fairly late appellation), obscures the sociocultural dynamics behind this development. For instance, the very vernacularity of Manipraviila was at the same time cosmopolitan, not only through the presence of Sanskrit but also through its use from the fifteenth century by native speakers of Kannada and Telugu. This emphasis on institutional locus should not underplay the significance of the obvious linguistic hybridity of Manipraviila. The combining of distinct languages is a novel factor apparent even in the orthography of Manipraviila works, as K. K. A. Venkatachari has noted: Manipraviila was generally written in Tamil script with grantha characters for those sounds not in Tamil, or else it was completely written in Telugu script with the addition of the Tamil characters for the two sounds peculiar to Tamil (underline 1 and underline r); when Tamil script was used, often the direct Sanskrit quotations were completely written in Grantha characters; or if a Sanskrit technical term or unfamiliar word was commonly used in Tamil, we find that the Sanskrit sounds may appear in Grantha characters, though often the word itself may have been in such common use that it was given a Tamil form.' We may productively think of ~rivaisnavaManipraviila as a linguistic register, a language variety distinguished by situational context or purpose. Socioloinguistic studies of register, distinguished from dialect as referring to situational variety instead of varieties associated with speech groups, include

' Venkatachari, "Observations," 1994-1997, 95-96.

several features associated with Manipravala, such as shared vocabulary, intonation, syntax and phonology, formulaic sequences, and special terminology. Although registers are normally associated with vocational domains or rhetorical modes of discourse, the formalized language of the Manipravala falls more easily within the category of register than within mere speech genre or style. And the presence of idiomatic vocabulary recognizable only to ~rivaisnavasrenders Manipravala the ultimate "insider" register, one which would be opaque to those not participating in the ~rivaisnavatemple.

2.1.3. Manipravala and Temple Oratory

~rivaisnavasexplicitly refer to the Manipravda works as the record of temple oratory. We may go further and characterize the entire ~rivaisnava Manipraviila corpus as the sedimentation in writing of the modes of oral discourse. As a form of writing, Manipravala developed different compositional styles as well as differences in the relative presence of Tamil or Sanskrit lexical items, as summarized by Venkatachari: Pillan, the first ~rivaisnavato use the Manipravala prose used long compound sentences; when he uses Sanskrit words, he provides Tamil endings. Next, Nafijiyar avoided long compound sentences in favor of

From Charles A. Fergusen, "Dialect, Register, and Genre: Working Assumptions About Conventionalization"in Sociolinguistic Perspectives on Register, ed. Douglas Biber and Edward Finegan (New York: Oxford University Press, 1994) 20.

simple sentences; when he uses Sanskrit words, he provides Tamil endings. Vatakkutiruviti Pillai and Periyaviicciin Pillai use more Tamil vocabulary (the ratio was about 2:1), while Pillai Lokiiciirya used still more Tamil vocabulary (3:l) and introduced the sutra style into Manipraviila prose. Vediinta Desika reverted to a Sanskrit-dominated Tamil prose (3:2). . . Alakiyamanaviilaperum21n2yaniirtsManipraviila is largely Tamil (4:l) and difficult prose, whereas Manaviilamiimunikal who came after him used the simplest prose style of all the A~iiryas.~ Yet despite this diversity, almost all examples of ~rivaisnavaManipraviila retained a close affinity with the temple lecture (upanydsa). Some references to the oral sources of Manipraviila writing occur in the Divyaprabandham commentaries but these are more frequent in the later hagiographical literature, especially Pinpalakiya Perumiil Jiyar's fourteenthcentury Guruparampardprabhdva. Two anecdotes associated with the thirteenth-century teacher Nampillai are illustrative. Guruparampardprabhdva 373 recounts a regent witnessing crowds emerging from Nampillai's lectures which rival the crowds at the worship chamber in Srirangam.10 This is a clear reference to public discourse; hagiographies, however, are more concerned with the circumscribed teacher-student transmission that the &A terms "oran ~ a l i . ' ' ~ ~ The most famous example of the latter concerns the competing records of Venkatachari, 1994-1997,95. lo Agyirappati Guruparampdprabhdva of Pinpamya Perumdl, ed. S. Krishnaswami Ayangar (Timcci: Puttur Agraharam, 1975) 373. Hereafter cited in the text with the prefix, "G."

l 1 Vasudha Ntinarayanan, The Vernacular Veda: Revelation, Recitation, and Ritual (Columbia, South Carolina: University of South Carolina Press, 1994) 105.

Nampillai's lectures by Periyaviicciin Pillai and Vatakkutiruviti Pillai. Unlike Periyaviiccfin Pillai, who received his teacher's permission to commit a commentary to writing, Vatakkutiruviti Pillai surreptitiously wrote his own by night. When Nampillai came to know of this unauthorized commentary, he at first denounced it and only came to accept it at a much later point (G 391). But even if the Manipravda commentaries require authoritative approbation from within the tradition, the continual production of new commentary signals an unusual open-endedness to exegetical practice summarized in the single phrase ~rivaisnavasuse to describe performative commentary: "text of enjoyment" (anubhava-grantha). The objective of such commentarial activity is not, as with Riimgnuja's ~ rBhlisya, i to provide an authoritative adjudication of determinative meaning but rather a creative participation in the original, which may explain the large number of extant commentaries for both the Divyaprabandham and the Riimiiyana. As in a temple lecture, the Manipravda commentaries weave narratives around a set of conceptual propositions, employing intricate world play, improvisational techniques, repetition, and novel forms of citation. Narayanan has demonstrated that such practices in the Tiruviiymoli Vdcakamdlai, for example, apply "a performing art technique to a verbal commentary," paradigmatically through glossing a single word or phrase in multiple ways on the model of the elaboration of a melodic theme (rdgd) or

dance

movement^.'^

Another feature evocative of performance is the

compatibility of mutually contradictory interpretations, albeit in conformity with the set conceptual framework; what is fixed is not the meaning of the verse itself but rather the preconception. The application of such Manipravala strategies to the Ramayana may be best observed in the contrasting evaluations of Bharata's relationship to Riima in Vatakkutiruviti Pillai's (b.1217) @u, Pillai Lokiicfirya's (b. 1205) ~rivacanabhz@ana,and Vedanta Desika's (b.1268) Abhayapraddnasdra. In one case Bharata's experience is a model for measuring an Aviir's progressive spiritual development, in another it is disparaged in favor of conformity with divine will, and in the third it is contrasted to more active performances of surrender. In the third section of the general introduction (mahiipravesa). Vatakkutiruviti Piiiai draws an extended comparison between Nammiilviir and Bharata in terms of developing stages of servitude to the lord (kainkarya), the state after liberation for ~rivaisnavas.Just as Bharata experiences intense grief upon returning from his maternal uncle's house and hearing Kaikeyi address him as king ("riijan"), similarly Nammiilviir gives voice to the pain of separation from the lord in the Tiruviruttam. When Bharata travels to the Citrakiita forest

l2

Narayanan, The Vernacular Veda, 1994,109.

with his ministers and army he bows his head to Riima in an act of propitiation, a display of longing matched by the longing of Nammiilviir's Tiruvasiyam. Residing in Nandigriima, Bharata's desire intensifies, as does Namm8lv%-'sin the Tiruvantati. Finally, after Riima returns and is consecrated king, Bharata is reunited with him; so, too, Nammdviir realizes the lord with the ~iruviiymo.Zi.~~ Compare this favorable attitude towards Bharata with the way that Pillai Lokacarya and his commentator Manaviila Mamuni develop the contrast between Rama's intimacy with Guha and rebuff of Bharata into a classic Tengalai argument in hvacanabh@ana 142-144. Bharata fails to convince Rzima to return to Ayodhya even though he is accompanied by his ministers and army (this request, though unsuccessful, constitutes his act of surrender); on the other hand, Riima himself approaches Guha and accepts him despite Guha's status as a tribal hunter. The message of the story, according to Manavala Miimuni, is that the lord's unmitigated independence (nirinkuia-svatantrya) entails that only his will determines liberation irrespective of acts of surrender. For Tengalais (unlike the Vadagalais) surrender is not a means (updyd). Accordingly Bharata's putatively meritorious act of seeking surrender is

l3 Tiruvfiymoli of Nammfilvfir with the Commentary 'IIdu'Iof Vatakkutiruviti Pillai, ed. B. R. Purushothama Naidu (Chennai: University of Madras Press, 1957-62).

transformed into a fault insofar as it conflicts with Riima's own intention to kill ~iivana.'~ Vedanta Desika's view of Bharata's request in the Abhayapraddnasdra is diametrically opposite: here Bharata's actions are not viewed as impertinence but rather are valorized over Laksmana's insistent request to serve, proving the superiority of passive dependence (p6ratantrya) over active servitude

(kairikarya). Ultimately it is Vedanta Desika's reading that Govindariija adopts. What is evident is that this same event, Bharata's request, is lifted out of its immediate narrative context into a didactic arena for divergent polemical purposes; this is the dominant Manipravala exegetical approach, which I refer to as connotative or parabolic reading to distinguish it from the more overt slesalike and allegorical readings examined in the next chapter. When Srivaisnavas use Vibhisana's surrender as a parabolic exemplar, this is not because it stands as an allegory or type, but rather because there is a deeper significance to the narrative. The term most commonly employed to describe this deeper significance is "esoteric" meaning (rahaysdrtha), indicating that the hidden resonances of these narratives would be apparent only to those initiated into the ~rivaisnavaconceptual world. How such meanings may be communicated in a l4 $ r i v a c a n a b h ~ ~ aof~ Pillai a Lokciclirya with Commentary of M a p a v + a m h p i (Puri: Sriraghunandanmudralaya, 1926). Hereafter cited in the text with the prefix, "SVB." For a discussion of this issue, see Mumme, "RCimCiyana Exegesis in Tengalai ~rivaisnavism,"1991,205206.

manner accessible to all is one of the central hermeneutic challenges for Govindariija.

2.2. Eighteen Traditional Meanings

2.2.1. Global Approach to the Epic Narrative

The most significant transformation wrought by integrating Manipravda oral discourse into Sanskrit is the redirection of the unsystematic extracting of isolated verses to verse-by-verse exegesis in context, in effect turning the provocative treatment of isolated incidents into a comment on the Riimhyana itself as verbal icon. Govindariija accomplishes this task in an especially insidious fashion, given that his gloss of the vast majority of verses operates through straightforward grammatical analysis and makes no mention of esoteric meanings. This grammatical dimension in particular gives Govindariija possibilities for detailed hyperanalysis only hinted at in the Manipraviila works, such that the esoteric meanings are traded in as the original meaning itself. In his comment on 1.5.1., Govindariija collects all the stories involving acts of surrender under the rubric of the eighteen traditional meanings of the Riimiiyana passed down from Ritmanuja's maternal uncle ~ r~ia i l Piirna. a This section provides an excellent place to examine the architectonics of Govindariija's translation from Manipraviila: all the stories are adapted directly

from the works of two rough contemporary authors, Vedanta Desika's Abhayapraddnasdra and Periyaviiccan Pillai's Tanislokam and Abhayapraddnasdra, and all are imbedded back into the text in their respective contexts. These three works differ vastly in structure and style. Periyavaccan Pillai's Tanislokam is a free-flowing, stream of consciousness meditation on individual verses organized in their order within the epic. Periyavacciin Pillai provocatively picks apart verses syllable by syllable without regard to narrative sequence but, rather, through associations with a wide array of sources, colloquial expressions, and rhetorical substitutions. Examples range from framing the first verse examined, 1.19.19 (Visvamitra identification of Rama's special nature), with the Puiuia Sfikta to the dozens of mutually contrasting interpretations of 5.21.3 (where Siti places a blade of grass between herself and Ravana). The Abhayapraddnasdras of both Periyavaccan Pillai and Vedanta Desika revolve around Rama's granting of protection to Vibhisana. Although these two texts are structured around this specific incident, they retain all the features of the oral Manipravala style. Tracing the relationship between this schematic outline and readings of specific verses allows us to isolate the procedures by which Govindariija integrates the Manipraviila performative style into the style of Sanskrit scholia. In what follows I present for each of the most important of the eighteen

meanings of the Riimiiyana the concept elaborated, the stories used as exemplars, and then the specific new information these stories communicate. After applying the six Mimiimsg signs of what a work is about (tdparya lirigas) to establish that the first meaning is Visnu's supremacy, Govindariija

identifies surrender (prapatti) as the second meaning. As discussed in the introduction, surrender is the central soteriological concept of the ~rivaisnava community, and the identification of this concept with the Riimiiyana allows Govindariija and others to refer to the epic itself as the "systematic discourse on surrender" (prapatti-Gstra). The stories cited here are used for almost all of the eighteen meanings. These stories include the following acts of surrender: the gods to Riima, Trisiinku and ~ u n a h s e ~ htoaVisviimitra, Laksmana to Rama, Bharata to Riima, the sages of Dandakiiranya to Riima, the demonic crow (kdkdsura) to Riima, Sugriva to Riima, Vibhisana to Riima, Riima to the ocean, and Trijatii to Sitii. Each exemplifies a different aspect of surrender. Some of the stories are major turning points in the narrative; others are minor incidents especially apposite to the dynamics of surrender. Laksmana's request to accompany Riima to the forest demonstrates that one should perform surrender in the presence of a mediator (purusakdra) such as Sitii. Similar is the treatment of Bharata's acceptance of Riima's sandals; according to Govindariija, the fact that Riima did not return with Bharata does not imply that he failed to

honor Bharata's request, but instead results from the priority of the god's prior request that Riivana be killed. The sandals, in this sense, are emblems not only of Riima's reign but also of the eventual fulfillment of Bharata's surrender. This revisionary presentation in effect transforms the narratological logic, with the series of acts of surrender replacing the dramatic flow of the epic emplotment of events. Other stories appear far more tangential. These include Riima protecting the sages in the Dandakiiranya, the punishment of the demonic crow ( k ~ k a u r a ) , Sugriva's negligence, Vibhisana's surrender, and Trijatii's exhortation to the female demons. The first three of these demonstrate respectively that acts such as residing in the lord's domain, falling before the lord, and folding one's hands in supplication qualify as surrender. The last example best illustrates how minor incidents may provide definitional resources for interpreters, demonstrating the dialectic dimension of the hermeneutic process-with

reflections on the Riimayana informing the

conceptualization of surrender as well as the other way around. The transference of Trijatii's surrender to the female demons (and Vibhisana's to his four companions) conduces to the popular Tengalai notion that Riimiinuja's performance of surrender extends to all ~rivaisnavas.And Vedanta Desika cites this passage in his discussion of surrender to the goddess in both the Rahasya Traya Sara and the ~ a r a ~ ~ a t i ~ a d ~ a b h d s y a .

Another incident whose significance for ~rivaisnavasis incommensurate with its position in the epic sequence is the perceived failure of Riima's entreaty to the ocean on the grounds of an absence in the requisite power disparity between suppliant and protector. The issue of restrictions on the object of surrender (visaya-niyama) has its source in the twenty-third siitra of Pillai Lokiiciirya's ~rivacanabh&ana,which Govindariija translates: "Even though there is no restriction regarding the place, time, qualified aspirant, and purpose of surrender, still the object of surrender is restricted" (SVB 23). Although one may perform surrender to other beings, the paradigmatic object of surrender is Visnu, and the criterion for identifying an appropriate object--one who is both capable (samartha) and compassionate (k-ika)-parallels

the two aspects of

Visnu's divine qualities, supremacy (paratva) and accessibility (saulabhya). The next meaning, servitude (kainkaryd), marks a crucial step in identifying these acts of surrender with the technical conceptualization ofprapatti. The terminology used for surrender is significant: while in the rahasya works the termprapatti is generally used, here Govindariija adopts the near synonym 'sarandgati," a more literal translation for which might be "seeking protection." What is of note is that sarar@gati, and notprapatti, occurs within the language of the epic. Adducing episodes such as Vibhisana's sarandgati as examples of surrender may appear to follow in a straightforward manner, but Govindariija takes a crucial step further: in contradistinction to mundane acts of surrender,

the immediate results sought in these examples are ancillary (dnusangika) to the ultimate goal of servitude. With this third meaning, Govindariija identifies all the various performances of surrender as oriented towards soteriological ends. Although the gods explicitly requested Riima to kill Riivana, their birth in the wombs of the celestial apsarases shows that their real goal was to serve him on earth (R 17.5.24). Govindariija offers similar explanations for each story, concluding that surrender is the chief subject (pradhdna-visaya) of the Riimiiyana, that the Riimiiyana itself constitutes an extended act of surrender (inrdmdyanam dirghasarandgati), and that all other meanings are subordinate to this unifying thematic focus.15

Of the remaining meanings, twelve through eighteen cover subsidiary topics such as the "sixty-four arts;" only four through eleven are explicated in context: 4) the mediator (purusakdra); 5,6, and 7) the threefold taxonomy of qualified aspirants (adhikdrasvadpa); 8) the five secret forms of knowledge (arthapaicakajcna); 9) the state of helplessness (dkiicanya, ananyagatitva); 10) the dcdrya; and 11) release from transmigration. The section on qualified aspirants (meanings five through seven) is one case where the contrast between the pedagogical imperative of generating exemplary

l5

The phrase used is: tadupayuktatayetaresam arthdndm pratipcidanam

models and the interpretation of the narrative plot is most apparent. Govindaraja reiterates Visnu's fourfold avatdra outlined in the opening prose passage: as Rama, Laksmaga (the embodiment of servitude), Bharata (the embodiment of dependence), and ~ a t r u ~ h (the n a embodiment of dependence on the lord's devotees, bhdgavata-pdratantrya). What is interesting in this taxonomy is the progressive hierarchical relationships between the various terms in the rahaysa works, with dependence valorized over servitude and dependence on the lord's devotees valorized in turn over mere dependence (on the criterion of absolute subservience). But extolling ~atrughnaover Bharata seems counterintuitive, given that the former accompanies the latter as subordinate at every stage in the epic. This is an example of the distance to be traveled from performative parabolic exegesis to contextual readings predicated on fidelity to the epic narrative. With economy of presentation, Govindariija's eighteen traditional meanings constitute a precis of the stories of surrender in the rahasya works. Each of these stories operates as a prism through which readers may recognize the meanings in context.

2.2.2. Exegesis in Context

Translating from Manipraviila into Sanskrit in the context of the relevant verses requires considerable hermeneutic effort on the part of Govindariija.

Although in some places Govindariija is able to reproduce the repetition and word play from Manipraviila nearly verbatim, at others he harmonizes these into the running commentary or modifies them through greater attention to grammatical detail. Translation from the rahasya works accounts for most of Govindariija's connotative reading, but this he accomplishes in a relatively restrained manner primarily in episodes where such a treatment seems most apposite, such as Laksmana's servitude, Sit& forbearance, Jatiiyu's liberation, and Vibhisana's surrender. The first example I wish to examine is Laksmana's request for servitude (nitya kairikarya) described in 2.31. For ~rivaisnavas,surrender results in a state of active service to Visnu called kairikarya. This term, defined as "a pleasuregenerating activity" (pritijanakavydpdravi@a) is a secondary derivation from kirikara, servant, which is in turn etymologically derived from the phrase a servant may utter, "What may I do, I' kim karomi). The concept of servitude is grounded in Visistiidvaita metaphysics, where the relationship between human and divine operates on the grammatical analogy of the "property-owner" relationship (sva-svdmi bhavd). After learning of Riima's exile, Laksmana persistently entreats Riima to allow him to accompany Rama to the forest. For Govindariija, this entire dialogue charts the stages of servitude, beginning with the performance of surrender and proceeding to the formal request for servitude. In 2.31.2, Laksmana first

responds to the news that Riima has been exiled: "That famed joy of the Raghus firmly grasped his brother's feet and addressed Sits and Rsghava of the great vow."16 Govindariija links Laksmana's address to both Riima and Sit%with the necessity of the presence of the mediator, but his stronger claim is that the very grasping of Rama's feet represents an act of surrender. Other elements of the verse correspond to the six components (angas) of surrender. This list of components, derived from Piificariitra Agamic sources, represents an analysis of the progressive performance of surrender and constitutes a point of controversy between Vadagalais and Tengalais regarding whether surrender is a unified or composite act. Grasping "firmly" (gadham) signals great trust ( m a h a v i i v ~ a )while , Laksmana's being "extremely famous" (atiyaia) indicates his intention to do good ( ~ n u k ~ l y a s a n k a l pand a ) avoidance of evil (pratikulyavarjana). These examples demonstrate the skillful way a set conceptual pattern is transposed onto the narrative through the literal words of the text. Govindariija pursues the analogy in more detail in a few key verses. Verses three and four constitute the formal request (kainkarya-prarthanam); Laksmana asks, "If your mind is set on going to the forest filled with deer and elephants,

16

sa bhratu6carapu gzdham nipidya raghunandanahl sitam uvacatiya6a raghavam ca mahavratam/l ( R 2.31.2)

then I will accompany you to the forest in the front, bearing a bow. With me by your side you will wander many forests, with birds and herds of animals resounding on all

sides.^'^'

Govindargja views verses five and six as a statement

of the absence of an ulterior motive (ananyaprayojanatva), another key aspect of servitude. In verse five, Laksmana issues a series of denials of otherworldly enjoyments: ''Without you I would not desire conquest over the gods, immortality, or even lordship over the ~ o r l d s . ' Govindarsja '~~ recasts each of these as an alternative conception of liberation. The fact that Laksmana does not desire conquest over the gods indicates that he rejects liberation without servitude. Laksmana's rejection of immortality refers to the state of isolation (kaivalya) advocated by Advaitins and in Samkhya. Finally, Laksmana rejects lordship over all the worlds, identified by Govindargja as becoming the god Brahma himself. We can imagine the different ways such explanations would have played out before ~rivaisnavacongregations and for generic Sanskrit readers. The verse receiving the most attention is 2.31.22: "Make me your servant, there is no impropriety in this. I will accomplish my goal, and your needs will

17yadigantum krtd budhir vanam mrgagajdyutaml aham tvdnugamisydmi vanam agre dhanurdharahll mayii sametoranyiini bahuni vicarisyasi1pak;ibhir m.rgayiithaiS ca sa%hus&ini samantatah// (R 2.31.3-4) nu devalokiikramanam niimaratvam aham v ~ eai~aryam l niipi lokaniim kamaye nu tvaya vina// (R 2.31.5)

also be taken care o f . " Here, the loading of esoteric meanings is especially intense, with multiple readings offered for the verse as a whole. That Laksmana is a "follower" (anucara) is a clear analogue to the servitude of kainkarya, and Govindariija supplies the qualifier "who am a dependant" (iesabhutam) in his gloss of the word "me1'(in technical discussions, sesa is used interchangeably with kairikarya). The level of minutiae subject to this comparison is evident in the morphology of the verb itself: the middle voice of "to make" ( k u y w a ) indicates that Laksmana's act of surrender will also be beneficial for you, the lord. Govindariija then reads the absence of impropriety two ways, either as an indication of the fact that the relationship of servant to lord (sevya-sevaka dharma) will not be violated or that the untoward events warned by Riima will not occur. The phrase, "your needs" indicates that this is an absolute case of servitude, performed for the sake of the other and not for one's self. Govindariija then offers an alternative reading for the entire verse. A wordplay develops from Laksmana's traditional association with the serpent-vehicle of Visnu, &a, the same word as the synonym for kairikarya. So Govindaraja renders the request as a call to restore the intrinsic nature of Laksmana as servant1 serpent ( s e d sew), and the absence of impropriety is really only the

l9 k u y a mam anucaram vaidhannyam neha vidyatel k.rt&tho ham bhavigiimi tava carthah prakalpatell (R 2.31.22)

lack of any contradiction with Laksmana'sl ~ e s a ' eternal s state of dependence (sesatva). The phrase, "I will accomplish my goal" refers to the possible objection, "If you don't perform servitude, what do you lose?" According to Govindariija, what Laksmana stands to relinquish is his very essence. As elsewhere Govindariija develops the basic outline for these esoteric

readings from a Manipravda work: the Tapiilokam, which includes a commentary on two of the verses considered here, 2.31.2. and 2.31.25. But the influence of Periyavaccan Pillai in this case extends only to the precedent of utilizing Laksmana's request as a model for servitude, not to direct translation. What does this example of Laksmana's surrender tell us about the relationship between Govindariija's interpretive practice and that of the rahasya authors? Again, this pattern of presenting contrasting, sequential interpretations stems directly from the oral substratum, where elaborating variations on a set theme mark the virtuosity of the exegete. The difference for Govindariija is that this spontaneity is achieved through careful grammatical analysis taken to extreme levels of detail, a kind of hyperanalysis. Another similar example of translation from Manipraviila is Govindariija's gloss of the verse known among Srivaisnavas as either Sitii's "last" verse (caramailoka, on the model of the "last"verse of the Bhagavad Gita, one of the three ~rivaisnavarahasya mantras) or Sita's protection-granting verse (abhayapraddna-doka, parallel to Rama's own protective statement to

Vibhisana). This verse (6.116.44) is embedded in a long passage near the end of the Yuddha Kiinda. When Hanumiin is about to attack the female demons who threatened Sits during her imprisonment in the Moka grove, Sitii prevents him and instead advocates forgiveness. This "last verse" is memorized and recited by ~rivaisnavasand represents for them the classic example of the goddess's role as mediator. Again, the Manipraviila source is Periyavaccan Pillai's Tanislokam. Several aspects resemble Riima's own granting of protection. The identity of those surrendering is in both cases demons, enemies whose character appears dubious to those advocating punishment. In each case the statement is framed by a legendary account, for Rama with the vignette of the dove, and for Sit3 with the words of the bear. Both protection-granting verses are presented as universal maxims. Govindariija provides an account of the parable of the bear. A hunter, chased by a tiger in the forest, climbs up a tree where a bear rests. The tiger says to the bear, "This hunter is the enemy of animals in the forest, why don't you push him down?" But the bear reasons that such an act would result in a loss of his dharma and instead peacefully falls asleep. The tiger takes advantage of the situation and says to the hunter, "Look, I'll protect you, just make this sleeping bear fall." And the hunter proceeds to do just that, but the bear grabs onto another branch and remains in the tree. So the tiger takes a different tack and again encourages the bear to push the hunter, pointing out that after all the

hunter just double-crossed him. The bear, however, remains firm and says, "I won't hand him over to you, even though he wronged me." The words of the bear cited by Sits are: "A wise person does not take notice of the wrongdoing someone does to someone else. Rather, one should guard one's own conduct, for good conduct ornaments good people."20 Next comes Sitii's protection-granting verse, an open-ended statement easily lifted out of the immediate context: "Whether dealing with someone wicked or virtuous, or even one deserving to be killed, oh monkey, a noble person should be compassionate. No one is free from

fault^.'^

It is the last part of the verse

(nu kaicin niiparadhyati, which we might better translate as, "nobody's perfect"), which receives the most attention from Srivaisnavas almost as a sort of pithy aphorism. While the contours of this passage are obviously conducive to ~rivaisnava connotative reading, Govindarfija performs a number of suggestive moves in his gloss on the protection-granting verse predicated on an apriori assumption: Sitfitsidentification with the goddess, Sri. And yet this simple identification

20 na parah papam adatte paresam papakannanaml samayo raksitavyas tu santaS caritrabhusanahll (R 1.116.44)

21papanamva Subhanam va vadh~irhanamplavangamal karyam k a r ~ a maryena na k d c i n naparddhyatill (R 1.116.45).

turns the passage into a distinctively ~rivaisnavaconception of the mediator, a conception developed over a millennium after the composition of the epic. Govindariija's first significant move is to supply the entire story of the bear while commenting on verse 1.16.43. This account of the interaction between the tiger, the hunter, and the bear is Govindariija's own; the Riimiiyana itself only refers to the words uttered by a bear, and Periyaviicciin Pillai seems unaware of it. Not only is the story an apparent invention, but the length of the commentary at this juncture supplies inordinate weight to what is a mere passing reference in the original. On the other hand, the remaining verses in the chapter elicit only a one- or two-line gloss from Govindariija. The portion drawn from the Tanislokam takes the form of an imaginary, ironic dialogue between Sit3 and an interlocutor (presumably Hanuman). As Govindariija provides the word-by-word gloss, each new phrase becomes a further development in an argument by Sitii whose ultimate objective is to convince Hanuman to forgive the female demons. Govindariija begins by undermining the opening pair of opposites, wicked and virtuous, with Sitii declaring, "either wicked, which is your meaning, or virtuous, which is my meaning." But this stance is immediately undermined when Govindariija's Sits refers to the familiar ~rivaisnavanotion of the lord's preference for those who are imperfect, like dirt requiring washing. The rhetorical shift here reverts to the idea that the person in question is, in fact, sinful (even though Sits's first

statement asserted her virtue). Govindariija reinscribes the aporia in the verse ("whether sinful or virtuous") into pseudo-scholastic alternative positions. The second statement renders the first into an ironic remark, whose ultimate meaning would be as follows: I, Sitii, consider them to be virtuous because their sin is, for me, a virtue. Govindariija's Sitii also dispenses with two objections voiced from the perspective of normative social practice (dharma iiistra). Sitii's extreme position seems to vitiate the basic rule of dhiirmic justice, whereby only those deserving punishment are punished and not the innocent (dandy0 da&antyah nadandyah). Govindariija's Sitii offers the last rhetorical option, "even those who deserve to be killed," as a direct rebuttal of this position to demonstrate that surrender transcends all other ethical norms. A second objection rests on the possibility of an unwarranted overextension (atiprasariga) of the principle of compassion, since the guilty would not be punished. Sitii's response is the laconic last phrase, 'nobody's perfect" (nu kaicin n~paradhyati). What should be clear from this example is that Govindariija's subtle method allows him to transform an ancillary passage into a major theological statement while at the same time maintaining a high deal of fidelity to the literal meaning of the text. The relative weight Govindariija grants to sections such as Sitii's 'last" verse enhances their significance and reorients the reader's perspective on the epic as a whole. And each reading is grounded on implicit interpretive

principles, although these are nowhere explicitly theorized; for instance, for Sitii's "last" verse, the intertextual (and intratextual) principle through which the analogue of other "last"verses may be brought to bear, or the more general principle of improvisation. In contradistinction to this subtle incorporation of Manipravda readings, occasions for more overtly sectarian interpretations also arise in Govindar~ja's commentary. A specific example is the liberation of the vulture Jacayu. The description of Jatayu's death occurs towards the end of the Aranya Ksnda (3.673.68) after Riima and Laksmana have discovered Sits's absence.22 The precedent for the treatment of these chapters appears to be the life of Parsiara in the Guruparampariiprabhava. The commentators focus their attention on a few specific verses: 3.67.9, 3.67.22 and 3.68.29-30. In 3.67.9, Govindariija explains the use of the adjective "illustrious"(mah~bhiigam)as a reference to the glory of Jatayu's having abandoned his body, indeed his very life, for the lord. In 3.67.22, Mahegvaratirtha views the description of Rama's face being covered with tears (biispap@amukhah) as the lord's distress on seeing a calamity befall his devotee. These brief remarks prepare the reader for the more detailed

22 Other notable examples of this more direct style include the praise of Mandodari (referred to as the Mandodari-catu&loki, R 6.114.14-17) and the praise of Tara (R 5.24.31) and Rama's granting a second chance to Ravana on the battlefield (R 6.41.66).

interpretation of Rama's bestowal of liberation in 3.68.29-30: " 0 king of vultures, I give you permission to go by the paths of those devoted to sacrifice, those who keep the fires, those who do not return again, those who give up the earth. Go to the highest worlds, o noble soul, go consecrated by me."23 Govindaraja offers four separate readings, each of which places special weight on the contrast between the ultimate liberation of Ja$iyu and ordinary life after death. First, Govindargja views each of the paths as a world corresponding to the earthly life-stations (dramas, in the sequence of the verse): householders, forest-dwellers, renunciates, and perpetual students. Or, on the second reading, the paths may refer respectively to sacrifice (yajfia), charity (dana), penance (tapas) and renunciation (sanyaa); in either case, Japiyu may proceed not only to these worlds, but beyond them to the heavenly world of Visnu. The framework of the last two readings is parallel. In the third reading, Rgma's consecration and permission are grouped together, with emphasis placed on the last phrase, now reinscribed as "come to me" (miim vraja). In the last reading, the permission after consecration is linked to specific mantras which qualify one for the respective worlds.

23

ya- gatir yajnasilanam dhitdgnei ca yd gatihl aparavartincim ya ca yE ca bhumipradayinamll (R

3.68.29-30)

These four interpretations develop a progressively sectarian Vaisnava understanding of liberation, from the identification of the worlds of Vienu to the construal of a devotional relationship and the positing of esoteric mantras. Through such a sequential approach, Govindariija is able to constrict the reference of a fairly open-ended statement to meanings conducive to the ~rivaisnavaemphasis on divine intervention in the achievement of liberation. The Jatayu incident plays a pivotal role in the unfolding of the events that follow. And within the logic of the epic narrative itself, this type of interaction marks an aspect of Rama's status as divine king. As Pollock has shown, Riima's salvific association with Vienu often appears as the beneficent relationship of a royal agent with his subjects; other similar cases include Rama's release of ~ a b a r iViriidha, , and Ahal~ii.'~ The commentary here provides an excellent example of the way ~rivaisnavasprovocatively tap into and foreground features of the text's basic structure. Given the centrality of Rama's association with Visnu for ~rivaisnavas,much is at stake in the interpretation of such incidents. Regarding our overall discussion, Jatiiyu's liberation is also significant in that Govindariija here engages in patterns of performative substitution as he does elsewhere, but this time without the precedent of a Manipraviila source.

24

Pollock, "The Divine King of the R%miiyana,"1991.

2.3. Vibhisana's Surrender

Vibhisana's surrender is the single most important event in the Riimiiyana for ~rivaisnavas.The two rahasya Abhayapraddnasdras of Vedanta Desika and Periyaviicciin Pillai that treat Vibhisana's surrender culminate in Riima's protection-granting verse (abhayapraddna sloka), for which they are named, after lengthy discussions of nearly every verse of the preceding two chapters (6.17-18). Here I focus especially on the ways Govindariija incorporates idiomatic, colloquial Manipraviila expressions and his use of framing devices such as categories drawn from dharma sastra.

2.3.1. The Monkeys' Strategic Deliberation and Rama's Rejoinder At the end of chapter 6.16, Vibhisana abandons Riivana after repeatedly advising him to return Sitii. Vibhisana appears in the sky opposite the place where the monkeys and Riima have gathered to prepare for war. Announcing that he is the estranged brother of the wicked Riivana, Vibhisana declares his desire to be brought immediately before Riima, the protector of all (sarvalokasaranydya). A lengthy discussion between the monkeys ensues, with each stating reasons why Vibhisana should not be trusted: some believe him to be a spy, while others disdain him as a demon (moreover, a demon who has just abandoned his own brother); some want to send a counter-spy to have

Vibhisana pass a test of faithfulness, or just to kill him (with a sharp stick, ''dudenu tivred). Finally, Hanumiin expresses a dissenting opinion. Strategically, Hanuman sees faults in each of the previous views; for instance, Vibhisana could not be tested without being entrusted with a great task, and this would lead to either his ascertaining the monkeys' intentions and deceiving them or their losing his friendship if he is sincere. Furthermore, Vibhisana's appearance and demeanor engender trust, and so Hanumiin advocates accepting him. Hanumiin's words provide an opening for Riima, whose response represents one of his more gracious acts in the epic. The structuring of this narrative, framed by the cynical perspectives of the monkeys couched in the technical language of political strategy (nitiiiistra), augments Riima's embodiment of ideal kingship as marked by benevolent rule. But as with the representation of Riima's encounters with Ahalyii and Jatiiyu, the language and analogies used by Riima bespeak an absolute, divine kingship highly conducive to ~rivaisnava conceptions of divinity. Riima says that he too has an opinion about Vibhisana: one should never abandon someone who approaches in friendship, even if he may have faults. Moreover, Riima points out, rivalries are quite common in royal families and it should not surprise the monkeys if Vibhisana in fact wants to overthrow Riivana. Riima also feels that Vibhisana's intentions are immaterial, for Riima

himself is capable of killing anyone with just his little finger (awlye@. Finally, Riima presents his clinching arguments: 1) the story of the dove; 2) the words of Kandu, son of Kanva; and 3) the protection-granting verse (R. 6.18.35). Riima asks Sugriva to bring forward whoever is asking for protection, whether he be Vibhisana or Riivana himself (yadi vd r d v a d svayam), and Sugriva at last accepts (R. 6.18.36). The entire dialogue rings with the immediate political context: spying, usurping kingship, treason, and asylum. But language used to describe the dynamics between Vibhisana and Riima, including related words for protection (saramm), seeking protection (saraqdgati), and protector (saranya), provide ample openings for ~rivaisnavainterpreters to insert sustained meditations on the nature of surrender itself.

2.3.2. Connotative Reading Just as with the other episodes, Govindariija frames the narrative with a conceptual schema, the six components (angas) of surrender. These components include: 1) intention for good (dnukulyasya sarikalpah); 2) avoidance of evil (prdtikulyasya varjanam); 3) great trust (mahdvisvdsa); 4) the request for protection (g~pt~tvavaranam); 5) entrusting one's self (dtmaniksepa, also considered the principal, a@, of the other components); and 6)

helplessness (kdrpanya). Together these components provide an analytic description of surrender. According to Govindaraja, Vibhisana's possession of the first two, ethical components, intention for good and avoidance of evil, is already evident in the preceding chapters through his continued counsel to Ravana. In addition, Govindaraja identifies his great trust at the end of the sixteenth chapter, where Vibhisana leaves Ravana in full faith that Rama will accept him. Vedanta Desika provides more details in his Abhayapraddnasdra. For Vedanta Desika the request comes in verse fourteen, where Vibhisana declares himself to be seeking Rama's protection (rdghavam saranam gatah), and the actual performance of surrender, Vibhisana's entrusting of himself (dtmaniksepa), occurs in the next verse where Vibhisana asks the monkeys, ''Immediately introduce me to the great Rama, the protector of all, as Vibhisana who is close at hand.''25 With the more detailed analysis of a rahasya commentary, Vedanta Desika identifies this request as a complete submission of the burden of protection (rakM-bhara-samarpana) outlined in the ~rivaisnava double mantra (dvaya). He goes on to identify Vibhisana's helplessness in the

25 sarvalokasaranyaya raghavaya mah~itmanel nivedayata mam h i p r a m vibhisanam upasthitam (R, 6.17.15)

very first words he utters to the monkeys (6.17.10): "The one named Riivana is the wicked demon ruler. I am his younger brother, known as ~ i b h i s a n a . " ~ ~ Mahesvaratirtha, however, presents a slightly different analysis, with both the avoidance of evil and the actual entrusting of self occurring in a single verse, 6.17.14: "And so being ill-treated and insulted like a servant, I left my sons and wives and come to Riima for protection ( ~ a r a n a m ) . "These ~ ~ contrasting applications of the identical six-component rubric again indicate the desideratum for proficiency in applying schema to the narrative, rather than adjudicating meanings once and for all. As an exemplar for surrender the scene requires the presence of a mediator,

and Govindariija identifies this role with Laksmana and the monkeys. Although the mediator is paradigmatically the goddess or an acdrya, any distinguished devotee (bhagavata) may also qualify. Govindaraja repeatedly explains the reluctance of Laksmana and the monkeys to accept Vibhisana as an expression of their overwhelming emotion and concern for Riima's safety. The first verse of chapter seventeen describes Vibhisana appearing in the vicinity of Riima with Laksmana (salaksmanah); Govindariija explains that this reference to Laksmana

ravano ndma d u y t t o r*aso rdksasesvarah/ tasydham anujo bhrata v i b h i w a iti srutah// (R. 6.17.10) 26

27 so h a m parwitas tena ddsavac cdvamdnitah/ tyatkvd putrdms ca darams ca rdghavam saranam gatah// (R. 6.17.14)

may only be a mere identifier (yydvartaka) of Rama or it may indicate the presence of a mediator (purusakarasannidhyam). Similarly, when Vibhisana's speech is introduced in 6.17.9, the specific form of his great trust is his temerity in approaching those opposed to him as mediators, knowing full well that as bh@avatas they would ultimately accept him. The strongest potential obstacle to constructing this episode as an act of prapatti is the possibility that Vibhisana's surrender might be a mere stratagem for usurping kingship, a refrain voiced not only by the monkeys, but also by Rama himself in 6.18.9 when he raises the more subtle ( s ~ m a t a r a m point ) about familial rivalries. Vedanta Desika addresses this issue at length, and Govindaraja embeds his arguments in his commentary on the relevant verses. The crucial component is the particularly ~rivaisnavaview of the boon Vibhisana receives from Brahma as recounted in Uttara Kanda 7.10.31-34. Here Vibhisana asks only to always have his mind intent on dharma--identified by ~rivaisnavasas the dharma of surrender itself. While this anecdote refurbishes Vibhisana's propensity for surrender, an additional, more direct argument is based on a simple psychological application of a distinction imported from grammatical theory, that between what is internal (antarangd) and external (bahiranga). Vibhisana's own words carry more weight in ascertaining his intentions than those of others. Finally, Vedanta Desika considers the question of Vibhisana's eventual acceptance of

kingship: if he only desired to perform servitude, why was he even offered to rule Lanka? Vedanta Desika's rejoinder is through two citations: 1)as with the other acts of surrender, the worldly result is merely ancillary (anusarigika), for which Vedanta Desika supplies a corroborating verse from the Visnudharmottara ~ u r ~ n2)aRama ; in effect orders Vibhisana to accept kingship when he grants him the Iksvaku family heirloom (the ~ r i r a n ~ aicon) m in Uttara Kanda 7.108.27-29.

2.3.3. Idiomatic Expressions and Hyperanalysis

Govindaraja directly translates a number of idiomatic expressions from Manipravala and incorporates the rahasya readings to a greater degree than anywhere else, with his commentary running to dozens of lines of text for each verse. Examples of such colloquial expressions include several in the commentary on 6.17.1, where Vibhisana is said to approach Rama within a moment (muhurtena). This dexterity is compared to "someone walking over hot coals" (angdranikarapariksiptevartmani padanyasavat) and, even more poignantly, to a freed calf who rushes forward refusing to see anything except his mother's udder (vigalitabandharajjor vatsasya mdturiidhah sparjam antarena

. sarirdrogyam arttkims ca bhogdrhs caivdnusangikanl daddti dhyiiyiniim nityam apavargaprado bar$.. Visqudhamottara PurMa (New Delhi: Nag Publishers, 1985). 28

madhyadesddarkqzavat), a metaphor drawn directly from Periyavaccan Pillai's Tanislokam and based on Tiruvaymoli 1.30. Another examples occur in the commentary on the next verse, 6.17.2. Here we are told that Vibhisana comes as if from a forest fire into a cool lake.29 And Vibhisana's abandoning Ravana for his sworn enemy amounts to an increase in the eminence of Rama at the expense of that of Ravana on the analogy of the possession of a horn (smga), a Tamil expression akin to "a feather in his cap." Later, in 6.17.9 Govindaraja refers to the superfluousness of Vibhisana verbally announcing his surrender after approaching as being similar to "salt on top of salt" (ksate ksdravat). These idiomatic metaphors represent concrete evidence of the oral sources for Manipravala readings. While I have been able to trace many of these as direct translations from the two Manipavala AbhayapradLinasdras, others can only be inferred to have their origins in an oral substratum. These usages are not carefully constructed poetic metaphors but rather the sort of concrete imagery one would imagine as part of common parlance. The best example for the way Govindaraja employs detailed grammatical analysis without resorting to overt figural reading is his comment on 6.18.3,

29

davdgner nirgatya sitalahrade patitum ivdgacchantarn

where Riima says, "I cannot in any way abandon someone whom I meet in friendship, even if he has faults. And good people would not censure this."30 Even as Govindariija draws heavily from Periyaviicciin Pillai and Vedanta Desika, his reliance on the technical analysis of Sanskrit grammatical categories is his own innovation. The forms in question are the two optative tenses in the verse for the verbs abandon (root tyaj) and to be (root as). The use of the optative (vidhi Un) covers an extremely wide semantic range in classical Sanskrit, which Govindariija exploits to generate multiple, seemingly faithful construals of the literal sense. For the first optative, Govindaraja offers two additional readings to the one I have provided in my translation. First, the optative is glossed as an imperative (vidhi): I will not abandon him. Next, this same optative expresses possibility (sambhGvanGydm Un) generating thereby a rhetorical statement: how could I possibly abandon him?

An even more counterintuitive pair of alternatives is offered for the second optative. First, again, Govindariija glosses the optative as an imperative (vidhi): he should be full of faults. And later this same optative becomes even more pointed as an entreaty (prGrthanGyGm Un): may he have faults. (Here the

30 mitrabhdvena samprdptam na tyajeyam kathamcanal doso yady api tasya sydt satdm etad agarhitam// ( R 5.18.3)

implication, as with Siti's granting protection, is that the faults of the suppliant only adds to the glory of the protector). In the next chapter, I analyze other examples where alternative glosses construct doubled utterances sometimeseven involving phonemic resegmentation similar to double entendre, slesa. Again, I view these patterns of repetition and substitution as emblematic of oral performance rather than of scholastic commentary. Bauman in his work on verbal art includes parallelism among the linguistic features keying performance genres, defined as "the repetition, with systematic variation, of phonic, grammatical, semantic, or prosodic structures." 31 Here as with the rahasya works the parallelism involves improvisational techniques oriented towards the notion of commentary as a ''text of enjoyment" (anubhava-grantha). But what is unusual about the practice of Govindarsja is that he carries out such substitution patterns through Sanskrit grammatical analysis.

2.3.4. Rama's Protection-Granting Verse

The turning point in Rama's argument occurs in verses 6.18.24-28 through references to the vignette of the pigeon and the words of sage Kandu. The story of the pigeon resembles in many ways that of the bear alluded to in the context 31 Richard Bauman, VerbalArt as Performance (Rowley: Newbury House Publishers, 1977) 18-19.

of Sitii's own protection-granting verse, with Govindariija again supplying the details. This time, the imagery is more graphic. A hunter enters a forest, shoots, kills, and eats a female pigeon but remains unable to quench his hunger. Falling at the foot of a tree, he calls out to the forest spirit (vanadevatd) for protection (saranam) with his hands folded in supplication. By chance, this tree happens to be the same one in which the pigeon's male partner resides. This pigeon, hearing the hunter's call for protection, offers himself into fire so the hunter may eat him, despite the fact that this same hunter killed his own partner. Though a version of this tale occurs in the Visnu Purina, it more closely resembles Buddhist gifts of the flesh and presents a striking contrast to the parallel incident of a hunter killing one among a pair of lovemaking herons (the inaugural md niNda verse, 1.2.15). In the latter case, Viilmiki's curse of the hunter symbolically prefigures the epic emplotment of the separation of the two protagonists, Riima and Siti. Here as well Govindariija draws this connection, the implication being that if a mere pigeon treated the hunter in this manner, how much more so should they accept Vibhisana (or for that matter, even Rivana). The verses attributed to Kandu (son of Kanva) are as follows: "Scorcher of foes, out of compassion one should not kill a helpless suppliant who seeks protection (sarawatam) with hands folded. One who is self-possessed should

protect the life of an enemy seeking protection (sara@gata) from others, whether this person be proud or in

distress.^'^^ I have not been able to

determine the source for these verses (the ka&-gdthd),

but Govindariija seems

to locate them in a dharmasdstra work, given the classificatory schema through which he organizes Riima's arguments. This he draws from the Yajiiavalkyasmyti: "The source of dharma is declared to be fivefold: 1)revelation (sytih); 2) tradition (smcih); 3) right conduct (saddcdra); 4) one's own benefit (svasya ca priyam dtmanah) and 5) desire born of purposeful intention (sarikalpajah kdmo) ."33 The manner with which Govindariija foregrounds this schema in relation to the decisive moment in the acceptance of Vibhisana-Riima's granting verse-epitomizes

protection-

his domestication of Manipraviila readings within

tight Sanskrit scholastic categories. According to classic dharmasdstra theory the evaluation of normative conduct is through the precedent of Vedic tradition and the conduct of practitioners of this tradition.

32 badhfijaliputam dinam ydcantam sara@gataml nu hanydd dnrsamsyartham api satmm parantapall drto vd yadi vd d.ptopare@m sarandgatah/ ar$zpr&zdnparityajya rabitavyah krtdtmana// (R.5.18.27-28)

srutih s m . ~ i hsaddcdrah svasya ca priyam dtmanah/ samyak sarikaIpajah kdmo dharmamiilam idam sm,rtamll 33

Govindariija identifies all the five sources in Riima's argument. Although Vediinta Desika also makes reference to the Yajiiavalkyasmrti verse, Govindariija uses it as a framing device for analyzing Riima's entire speech in this chapter. First, he supplies a mysterious Upiinisiidic fragment which begins, ''For this reason they do not hand over someone to be killed who has sought protection

..

The next logical step is corroboration through tradition

(smrti) with the dharmasiistra verses of Kandu and the precedent of established practice (siglicdra) rendered through the story of the pigeon. According to Govindariija, Riima demonstrates the benefit to himself through his pragmatic political argumentation with the monkeys. And he expresses his desire in the protection-granting verse with the reference to a previous vow; the verse itself reads as follows: "I give protection to anyone who surrenders to me even once asking, 'I am yours.' This is my vow."35 The two Abhayapraddnasdras place special significance on the gloss of the word, "once" (sakrd) in the protection-granting verse, with Periyavacciin Pillai charting an oral tradition of interpretations of this single word dating back to Riimiinuja's original core disciples. The first cited is that of Kuresa, Riimiinuja's scribe. According to Kuresa, the word sakrd functions as the substitute (ddesa} 34

tasmdd api vadhyam prapannam nu pratipayacchanti

35 sakrdeva prapanndya tavdsmiti ca ydcate/ abhayam sarvabhOtebhyo daddmy etad vratam mama// (R. 5.18.35)

for "immediately,"sahasd; hence the sense is that protection ought to be granted to one who approaches straight off like Vibhisana. Next comes the views of Embar, Riirninuja's cousin, a significant figure in the line of succession but the author of no known works. For Embar, sakrd means only once. Before performing surrender one is forever immersed in transmigratory existence and afterwards enjoys eternal happiness--therefore a single act of surrender is sufficient. And according to Parasara Bhatta, son of Kuresa and student of Embar, the restrictive pronoun eva which follows sakrd means that repeated performances are in fact prohibited (with any semblance of repetition involving only the pleasure of remembering the original event). Vedanta Desika, interestingly enough, adopts this last position, which is especially notable given the standard Vadagalai view concerning the status of secondary acts of surrender. Whereas Tengalais adhere firmly to the principle that surrender may be performed only once, Vadagalais also consider surrender to function as a repeatable form of penance (payascitta). Several aspects of this tradition of analyzing the word sakrd are striking. First, Periyavaccan Pillai's attribution of alternative views represents an unusually explicit presentation of the transmission of esoteric readings through successive generations of teachers and students. These readings also touch most directly upon subsequent sectarian conceptions of surrender, even though there is little substantive disagreement between them. As an additional mark of

Govindariija's effort to shroud the esoteric interpretations in a universal idiom, Govindariija elides these rival interpretations of sakrd entirely.

2.4. Rama Praise Poems

I would like to now briefly consider a genre of ~rivaisnavacomposition where we find similar dynamics as the translation from Manipraviila to Sanskrit: Sanskrit praise-poetry (stotra-kdvya) to Riima. I focus specifically on Vedanta Desika1sMahaviragadya, a short prose poem, and portions of his Pddukdsahasra, a much longer poem. Just like the translation from Manipraviila, these poems also involve the incorporation of vernacular forms and forms proper to the temple, namely Tamil devotional poetry and the language of liturgy. Yet this incorporation does not entail direct translation, but rather a stylized adaptation; neither do these Sanskrit poems conform to the generic norms of Tamil poetry, nor were they likely used in liturgy.

I do not focus on more popular literature in Sanskrit, including the ~rivaisnavadaily recitation (prayana) of verses invoking the Riimiiyana or the 'Giiyatri Riimiiyana" which matches each of the twenty-four syllables of the Vedic Giiyatri hymn with verses from the Riimiiyana spaced into roughly onethousand verse intervals. Although these practices are germane to an examination of the integration of the Riimiiyana in temple liturgy, they are of uncertain provenance. Moreover, since they are in no way engaged with the

Ill

status of the Riimiiyana as k w a , they are tangential to my overall study of theologization.

2.4.1. Generic Relationship of Sanskrit and Tamil Praise Poems

The relationship between Riima praise-poetry and vernacular forms may be made more clear by considering the broader historical context of transformations in Sanskrit stotra-k&ya in the Tamil milieu from the tenth century. Sustained composition of lyrical stotras marks the influence of Tamil devotional genres, as evident in the works of ~rivaisnavaslike Yiimuna, Kuresa, and Pariisara as well as in the Bhiigavata Puriina; as Friedhelm Hardy puts it regarding the direction of influence between the Bhiigavata Puriina and a v i i r poetry, "the BhP uses the fiviirs and not vice versa.36 This transformation of the stotra from lists of philosophical epithets to poetic descriptions of icons dovetails with the revaluation of stotra as a form of liturgical utterance or mantra in Piificariitra Agamic 1iteratu1-e.37Yet the most ornate examples of this

poetry such as the stotras of Vedanta Desika, replete with long compounds,

36 Friedhelm Hardy, Yiraha Bhakti: The Early History of Krsna Devotion in South India (New Delhi: Oxford University Press, 1983. 37 Nancy Ann Nayar, Poetry As Theology: The hivaiwava Stotra in the Age of Ramanuja (Weisbaden: Harrassowitz, 1992) 22-25.

complex prosody, and rhetorical figures, represents a highly stylized form which was probably never recited in temples. The relationships between the classical Tamil Sangam corpus and Vaisnava and ~ a i v devotional a genres is a vast topic, but I will limit my discussion to the generic form that bears most directly on the Mahdvlragadya and Tamil Riima bhakti: pdtan tinai, the Tamil panegyric. Although akam poetry, (poetry of love) more strongly informs Tamil representations of the relationship between soul and god, it is the heroic tone ofpuram poetry (poetry of war) that is relevant for understanding Riima bhakti. Pdtan tinai, which is included among the seven subdivisions ofpuram poetry in the Tolkiippiyam, is the basis for Vaisnava devotional poetry from the earliest examples in the Paripiital to the poetry of the Alviirs, including portions of the Tiruviiymoli. A. K. Ramanujan has carefully traced the correspondences between classical Tamil and bhakti panegyrics, noting that "themes such as praise of a hero's (god's) fame, praise of a victorious hero (god), and praise of a king (god) for providing shelter and security fall into this category.''38

38 A. K. Ramanujan (with Norman Cutler), "From Classicism to Bhakti" in The Collected Essays ofA. K. Ramanujan, ed. Vinay Dharwadker (Oxford: Oxford University Press, 1999) 236.

2.4.2. Mahaviragadya and Padukasahasra as Heroic Panegyric

The hyperbolic praise ofpdtan tinai appears most apposite to Rama bhakti, where the overall mood is that of respectful honoring rather than of erotic love as with Krsna bhakti. I turn now directly to Vedanta Desika's Mahdviragadya and Pddukdsahasra, paying particular attention to the elements most germane for a comparison with the Ramayana commentaries: 1)the Tamil panegyric style; and 2) the elements of Sanskrit stotra-kdvya. Vedanta Desika constructs the Mahavlragadya, a prose poem, as a set of descriptive epithets to Riima, each recounting a separate episode from the epic; the overall effect is to render every stage of the Ramiiyana narrative into a chronicle of Riima's heroic exploits. The structure of the poem is as a single, extended sentence comprised of vocatives culminating in a final indeclinable ("I bow down to you. again I bow down to you" narnas tepunas te n ~ r n a h ) . ~ ~ The choice of events represented and the length of treatment conduces to the overall heroic mood. Notice for instance the paired descriptions of the breaking of &vatsbow ("your intoxicated club-arms were powerful enough to break the great bow of h a " ) and acquisition of Visnu's bow ("you showed your supremacy in taking back your own bow").40Even more typical is the string of

'

Vedanta Desika Stotramda, ed. U. V. Ramadesikacarya (Chennai: Lifco, 1966). Hereafter cited in the text with the prefix, "MV."

short epithets describing Riima's killing of demons, including Viriidha ("tiger to the deer of Viriidha"), Trisiras ("the sun casting away the darkness of the three heads of Trisiras"), Khara ("the terrible wind which felled the wicked tree called, 'Khara"), and the demon army ("the great elephant who destroyed the grassy forests of the 14,000 demons).41 These violent images are matched by phrases such as "one of unimpeded furious deeds" (unapdya-sdhasa, MV 36), "hero" (vira, MV 50), "one whose enemies are as non-existent as a flower in the sky" (kha-puspita-ripu-paksa,MV 72), "lion among the Riighavas" (rdghava-simha,

MV 76), etc, and also by the lengthy treatment of the war from stanza fifty-five to sixty-nine. While this account of heroic acts evokes the panegyric style ofpdtan tinai, the poem remains fully an ornate Sanskrit kdvya, as evidenced by the elaborate compounding, rhetorical effects, and poetic imagery. The long compounds in particular make it appear unlikely that the gadya was ever actually recited in temples. I point to two specific examples. The first is the description of Riima's liberation of Ahalya:

40 kha&aparasu-koda&a-praka&a-kha&ana-sau~a-bhujadanda nija-dhanur-fikarsana-prakasita-pfiramesthya (MV 19) 41 druhina-hara-valamathana-durfiraba-saralaba (MV 27) trisirah-sirastritaya-timara-nirfisa-vfisarakara(MV 31) kharatara-khara-tam-khandana-cads-pavana (MV 33) dvisapta-ra~sahasra-nalavana-vilolana-mah~kalabha (MV 34)

(MV 15)

''You were praised by the beautiful wife of the great sage. The dust of your lotus feet, which released her form as a stone, were resplendent from the ambrosia-like Gangii river which flows down in a jagged path to ~ i v a 'matted s hair bearing the crescent moon."

jadakirana-sakaladhara-jatila-natapati-muk~a-tata-natanapa.&vibudhasarid- atibahulamadhugalana-lalitapada-nalinaraja-upamrditanijavrjina-jahatupala-tanumcira-paramunivara-yuvati-nfia(MV 12) Another extremely long compound recounts Riima's encounter with Parasuriima: ''Your iron-rod arrows diverted the merit of Parasuriima. Parasuriima had already propitiated his father with a vast lake full of the bloodsprings of families of kings, the foremost among whom was Kiirtavirya. This Kartavirya could defeat Riivana, jagged from bearing the teeth of the elephant of Indra, the tormenter of the world eager to play ball with Mount Kailiisa."

kmtaharasikari-kantuka-vihrtyunmukha-jagadamntuda-jitahari-dantibharita-pfihutara-tdtaka-tarpita-pitmka-bhrgupati-sugati-vihatikara-natapamdisu-parigha (MV 20) Of the Sanskrit poetic tropes in the Mahavlragadya, the most prominent is the sound figure of phonemic alliteration (anuprdsa). An excellent example of such anuprdsa is the seventh stanza describing Riima's birth: ''Since the whole universe comes into existence depending on you, you are the ultimate cause. Yet hiding this fact from everyone, you were born as a son to Kausalya just like any other child." kosalasutd-kumdrabhdva-kaficukita-kdramra (MV 7) The repetition of initial "kt'sounds is a straightforward, easily identifiable form of alliteration, but more complex forms occur in other stanzas, especially

fourteen and forty-four. Note the repetition of whole syllables in each of these verses; (locana in fourteen and daksa in forty-four): ''You were like a moon to the partridge-eyes of the beautiful-eyed women of Maithila." maithila-nagara-sulocand-locana-cakora-candra (MV 14)

''He reassured the heart of his friend by moving his little left two slightly, which was capable of casting away the pile of bones of Dundubhi. Those bones were very hard and as massive as Kailasa."

These various elements indicate that poems like the Mahdvlravaibhava were carefully constructed works of art, which were modeled on popular hymns used in temples yet lacked the mantra-like liturgical quality of devotional poetry. Another poem containing similar features is Vedanta Desika's Pddukdsahasra. A legendary account tells of Vedanta Desika composing this poem of a thousand-verses to the sandals of Ranganatha, the icon at ~ r i r a n ~ a m , overnight in rejoinder to a challenge from a literary rival. The relevance of the Pddukdsahasra to our study involves the nearly a hundred verses which associate these sandals of Ranganatha with Rama's sandals granted to Bharata. As with the Mahdviragadya, the mood of these verses is heroic praise, rendered hyperbolic in the veneration of this simple possession of Rama, the intimacy with which engenders a greater distance from Rama himself. Typical is verse

T h e lion Rama entered the forest as he pleased, his heart adamant in the desire to tear asunder the elephant of Ravana. But as if embracing your cub Bharata with motherly affection, you did not abandon the lair of Kosala to calamity." dasagrivastamberamadalanadurddntahrdaye vihdrasvdcchandydd visati raghusimhe vanabhuvam/ svavdtsalyakrodik~abharatasdvevabhavati nirdbddhdm pdddvani nu vijahau kosa~a~uhdm//4~ The maternal compassion of the sandal contrasts directly with the image of violent conqueror in a far more dramatic manner than the standard contrast between Visnu and the goddess. This role of the sandal is the central theme of the retelling of the Rama-Bharata encounter in the Pddukdsahasra. And as with the Mahdvlravaibhava, Vedanta Desika recreates this style of panegyric in a form replete with figures (alamkaras) of Sanskrit poetics. The most common figure is poetic supposition (utpreksa), which Vedanta Desika uses repeatedly to convey the embodiment of the sandal. As with the other examples of translation and redeployment examined in this chapter, Vedanta Desika's Sanskrit praise-poem appears to construct a new form of personification based on utpreksa which parallels the new forms of allegorical reading and allegorical composition discussed in the next two chapters.

42~~dukasahasra of Vedanta Desika, ed. Pandit Kedaranatha and Wasudeva Laksman Sastri Panashikar (Varanasi: Chaukhambha, 1984) 5.18.

2.5. Conclusion

I have tried to show the ways the shift in institutional locus from court to temple outlined in the first chapter entailed the application of interpretive techniques specific to the use of texts in vernacular oral lectures and liturgy. The trajectory of influence may be best captured in the translation of provocative Manipravda readings into Sanskrit, given the status of Manipravala as the "voice" of ~rivaisnavismas vernacular theology, especially in the section on Vibhisana's surrender and Riima's protection-granting verse. While Govindaraja's commentary incorporates readings from Vedanta Desika's and Periyavaccan Pillai's respective Abhayapraddnasdras, he does so through Sanskrit technical grammatical analysis and dharma sdstra categories in a manner far different than the modes of temple oratory. The parallel example of RAma praise-poems includes a similar incorporation of Tamil devotional and liturgical discourse within the idiom of Sanskrit kdvya. The incorporation of the modes of vernacular temple discourse into Sanskrit marks a counterpart to the modification of categories of Sanskrit aesthetics so as to accommodate ~rivaisnavacategories and concepts. These two dimensions come together in some of the readings I examine in the next chapter, including the linking of performative substitution and allegorical reading with readings resembling an important category of Sanskrit aesthetics, double entendre (slesa).

CHAPTER 3 MODES OF FIGURAL READING IN ~ Y A N COMMENTARIES A

In his commentary on the verse that the Riimiiyana itself characterizes as the inaugural instance of Sanskrit poetry, and elsewhere, Govindariija produces figural readings that bifurcate the text into separate literal and theological levels of meaning. While "figural," as well as the related terms "figuration" and 'tropology," may refer generally to any language that departs from the standard usage of words or syntax so as to produce poetic effects, in this chapter I am concerned with a particular trope identified in Sanskrit poetics, double entendre

(slesa). By a practice of reading, I mean the process through which Govindariija is able to identify in the text multiple levels of meaning resembling slesa. What should emerge is an assessment of how Govindariija's reading of multiple levels of meaning resembles and yet differs from both the theorization of slesa in Sanskrit aesthetics and the practice of slesa in Sanskrit poetry. This chapter is divided into two parts. In the first, I examine three separate types of

In the second, I compare and contrast these readings with others in Govindariija's commentary including connotative reading and allegorical reading, so as to provide a comparative grammar of figural reading techniques. Finally I raise the question of why Govindariija would chose to adapt and modify categories of Sanskrit aesthetics and what this choice tells us about his conception of the Riimiiyana as a kdvya. Such an engagement with Sanskrit aesthetic categories represents a second dimension of the hermeneutic project of Riimiiyana commentary, the counterpart to the integration of vernacular forms into Sanskrit outlined in the last chapter.

3.1 What is slesa?

Double-entendre, or slesa, is a rhetorical trope in which a single phonetic sequence yields numerous meanings. As Appaya Diksita (16th century) defines it in his classic textbook on poetic figures, the Kuvalaydnanda, slesa is "the stringing together of multiple meanings" (n&zarthasamsrayahsleso).' Depending on the type of slesa and depending one's language ontology, slesa may be said to involve homonyms or homophonemic utterances--in other words a complex form of paronomasia. A classic English example occurs when Mercutio in Shakespeare's Romeo and Juliet is stabbed, knows he is to die, and

Kuvalaydnanda ofAppaya Dikita, Ed. Vasudeva Sharma. (Mumbai: Tukaram Javaji, 1903) 63. Hereafter cited in the text with the prefix, "K."

says: "Ask for me to-morrow and you shall find me a grave man;" here the pun rests on the two meanings of the single word "grave," a serious person or a corpse in a grave.2 ~ l e s ahowever, , includes a much wider range of phenomena than such forms of paronomasia, not only multiple meanings of single words, but also alternative splitting of compounds, construing of syntax, and even renderings of morphological and phonemic elements. ~ l e s aemerges as a major category of analysis in Sanskrit poetics and a distinctive mode of composition, with the development of an entire technical apparatus (including associative lexicons) and the production of "double-stream" poems, dvisandhdna-kiivya,from the beginning of the second millennium. Yet despite this diversity, there are certain standard features of slesa against which we can measure Govindariija's slesa-like readings. As an example, we can turn to a verse from Dandin (c. 725),which Appaya Diksita later cites: ''Rising in fame, handsome, the circle of neighboring rulers devoted to him, this king captures the hearts of people with his light taxes." (Contextual Meaning) 'Rising atop the eastern mountain, resplendent, a red-hued orb, the moon captures hearts with its gentle rays." (Non-contextual Meaning)

Romeo and Juliet of William Shakespeare (New York: Washington Square Press, 1992) 3.1.97-98.

asau udayam *huh kdntimdn raktama&alah/ rdjd harati lokasya hrdayam mrdulaih karaiw (K 64) Here the slesa rests on the double meanings of several words: udayam (fame, the eastern mountain); raktam (devoted, red color); ma&alam (circle of neighboring rulers, orb); raja (moon, king); mrdulaih (light, gentle); and karaih (taxes, rays). Even in this relatively straightforward example, the sophistication necessary for both composing and identifying slesas should be evident. With this example in the background, I turn now to a detailed analysis of the three forms of slesa-like readings in Govindariija's commentary.

3.1.1. The Inaugural Verse Commenting on the inaugural "md ni@dat'verse (1.2.15) is in effect tantamount to commenting on the entire receptive history of the Riimiiyana as kdvya, given the foregrounding of the verse in the epic's own presentation of its aesthetic qualities and the attention it receives in earlier Sanskrit aesthetic theory. The setting is as follows. The poet, Viilmiki, has just visited the divine sage Niirada, who recounted in brief the story of the acts of Riima, best among men. Thinking about what he has heard while returning to his hermitage with his student, he beholds a disturbing sight: a hunter kills a heron in the act of love before his very eyes. Viilmiki utters a curse: 'Hunter, may you never have progeny, since you killed one among this pair of herons in the thrall of sexual desire."

md n i s d a pratistiim tvam agamah sdsvatih sam$z/

yat krauficamithundd ekam avadhih kdmamohitam// (R 1.2.15) Much to Vamiki's amazement, this utterance issues forth in a special form, with four feet each containing an even number of syllables and the melody of a vind set to musical time, as Valmiki himself remarks to his student three verses later (R 1.2.18). It is in this context that the phrase that handavardhana notes in his assessment that pity ( k a r w ) is the central rasa of the RSmayana occurs: ''Let this which has been uttered by me in sorrow (soka) be known as verse (sloka),not in any other way" (slokdrttasya pravflto m e sloko bhavatu ndnyathd)

(R.2.40). In the end, the god Brahma appears before Valmiki to explain that the md nisada verse is the result of his own gift of composition; in fact, it marks the very first instance of poetry in the Sanskrit language. Brahmii then grants Valmiki the power of omniscient narration so he may recount all the events of the Riimayana: "Whatever happened to that intelligent one, whether in secret or out in the open. . . all will become known to you" (rahasyam ca p r a k d h m ca yadvrttam tasya dhimatah. . . sarvam viditam te bhavisyati) (R 1.2.33-34). Since the verse is set apart and foregrounded in this way at the very commencement of the narrative, it demands close attention from interpreters. Govindaraja first provides a literal gloss of the verse, which contains a grammatical irregularity, focusing attention onto the elements that become the

axis of the slesa-like reading. The issue is that the aorist (lun) retains the augment "a," even though this should normatively be elided to form the injunctive. The prescriptive application of Paninian grammar to epic usage is a common feature of epic commentary, but Govindariija explores this particular irregularity in extreme detail. First, he provides a fairly straightforward resolution by citing the seventh-century Kiiiikd position that through a different morphological analysis of the negative particle, (md vs. man), the augment would no longer be prohibited.3 But, as though this solution were unsatisfactory, Govindariija also refers to another, more unusual explanation, that of the now lost twelfth-century Durghatavpti by the Buddhist grammarian Maitreya ~ a k s i t a Maitreya . Raksita resolves the problem of the unwanted augment by providing an alternate construal of the phonemic sequence, anticipating Govindaraja's slesa-like reading in taking the md to mean sri (or laksmi), actually an attested lexical meaning. The solution is to avoid the augment altogether and read the phrase

The Kasika explanation is as follows: asarhsaydm bhzitavac ca iti luwtidesdt lun, lrditvdd an, ndyam marig api tu mdsabda iti kdsikdkdrah, tendddgamepi na virodha. Harold Coward and K. Kunjunni Raja include this entry on Maitreya Raksita: "This Buddhist grammarian in eastern India lived between 1092 and 1122, according to Yuddhisthira Mimamsaka, who thinks he may have been a Bengali. In addition to works on Buddhist Grammar, including Dhdtupradipa, Dhurghcti, and a Tantrapradipa on Jinendrabuddhi's Kdsikanyasa (a fragmentary manuscript, which is listed as residing at the Asiatic Society Library in Calcutta), he appears to have written a tikd on the Mahdbhdsya, which has been lost." Harold G. Coward and K. Kunjunni Raja, Encyclopedia of Indian Philosophers, Volume V: The Philosophy of the Grammarians, Ed. Karl Potter (Delhi: Motilal Banarsidas, 1990) 207.

as, "o hunter, unfortunate one, may you never have progeny" (md n i s d a prati@n tu ama gamah). What Maitreya Raksita has done is replace "tvam agamav with "tv ama gumah," "ama" being a bahuvrihi compound meaning 'unfortunate" (i.e., without laksmi, alaksmika). Although Maitreya Raksita' construal is unnatural in terms of word order, it provides Govindariija with a precedent for taking the word m i to mean hi?. Here, unlike the connotative readings of Laksmana's servitude or Vibhisana's surrender, Govindariija separates literal and figurative levels of meaning. The gloss of the individual elements of the verse stands on its own without the intrusion of sectarian concepts and terminology, providing a coherent explanation of the verse within its immediate narrative context.

3.1.1. ~lesa-likeReading One

Govindariija introduces the actual slesa-like reading as reflective of an earlier ~rivaisnavatradition that treats the verse as a benediction (mangaldcarana). Govindariija explicitly locates the opening for such a reading through a revaluation of Brahmii's gift of composition--noting the impropriety in making the first verse an imprecation--but he does not develop the implications of this Brahmii's involvement provides evidence connection as do other ~rivaisnavas.~

The only evidence of such a tradition is a brief slesa-like reading by Mahesvaratirtha along the same lines.

for ~rivaisniivasof the status of the Riimiiyana as a work of tradition (smfii or explication of the Veda, Vedopabrhanam); interestingly, the late-sixteenthcentury commentator Atreya Ahobila goes so far as to locate the intentionality for the slesa-like level of meaning with Brahmii and not Viilmiki, the implication being that this level would be unknown to Viilmiki himself. This assertion is linked to deeper epistemological questions that ask how readers go about identifying slesas, and it opens suggestive possibilities for demarcating the literal and figurative levels clearly. But Govindariija does not go down this path. Instead, he simply states that the real point of the verse is a benediction (mangaldcaranaparo yam slokah) and launches into the slesa-like reading. Govindariija identifies the sort of bitextuality we normally associate with slesa, both the kind that occurs in slesas such as the king-moon example cited earlier, and the kind that involves phonemic resegmentation. An example of the latter, involving the same element upon which Govindariija's own slesa-like reading rests--the word, m i , as meaning ~ri--isfound again in the Kuvalaydnanda. I cite the half-verse twice to represent orthographically the two different ways it may be read: ''May the all-giver, the lord of ~ rwho i held up the mountain and the earth, protect you." 1draw the terms, "bitextuality"and "resegmentation" from Yigal Bronner's unpublished dissertation on slesa, Poetry at Its Extreme: The Theory and Practice of Bitextual Poetry ( ~ l e s ain ) South Asia, University of Chicago, 1999.

(1)sarvado mddhavah pdydt sa yo gum gdm adidharat1 'May he who is always the lord of Umii and who held up the Gang3 river protect you." 2) sarvadomddhavah payat sa yo gangdm adhidharat1 (K 63) Here the doubling of meaning cannot be confined to a play on the polysemy of single words. Rather, the same sequence of sounds is divided, on the one hand, " and, on into the words "all-giver" (sarvada) and "husband" (dhava) of " ~ r i (mi), the other, into the words "always" (sarvadd), and "husband" (dhava) of Umii (umd). This double reading is made possible by the rules for euphonic combination (sandhi) in Sanskrit. While other examples involving resegmentation similarly rest on features proper to the Sanskrit language such as the splitting of compounds, such resegmentable utterances, or oronyms, occur in almost all languages. In English, we find this more frequently in speech than in writing, given the rigidity of writing conventions; we can think of combinations such as "bean ice1'/"be nice." With Udbhata's (c. 800) definition, this kind of slesa came to be known as a sabda-$esa, i.e. a slesa in which the phonemic dimension itself rather than meaning is predominant. Govindariija's reading rests on both standard lexical meanings and more fanciful etymologies and derivations. Replacing the negative particle, md with md as the goddess ~ r(as i it is taken in both Maitreya Raksita's grammatical explanation and in the Kuvalaydnanda), Govindariija combines it with the next

word, niMda. Based on the derivation of this word from the first class root sad plus the prefix, ni, he takes niMda to mean "place of residence" (nivdso). The meaning for the newly formed compound is: "one in whom the goddess ~ r i resides," i.e. Visnu. The entire verse is thereby transformed into the desired benediction. We should note that this reading, though grammatically plausible, is based on no attested sense of the word niMda, which is actually derived from a different verb-the

fourth class verb sad meaning to scatter or finish rather than the first

class sad. But it is upon this foundation that the entire slesa-like reading rests. Govindartija further glosses the compound as meaning ~rinivtisa,a common ~rivaisnavaepithet for Visnu and the name for the icon at Tirupati. Tirupati, as we saw in the first chapter, was the primary center for ~rivaisnavasectarian activity during the Vijayanagara period and the place where Govindartija says he was inspired in a dream to compose his commentary. With this vocative in place-ma-niMda

as ~rinivtisa-~ovindarfijafills out

the other details of the slesa-like reading. First, the verb, now an aorist and not an injunctive, is understood to function semantically like an imperative. The object of this verb,pratis@m, on the secondary level means greatness (mahatmya). Govindartija then reads "one among a pair of herons" (krauficamithunad ekam) as referring to Riivana (paired with his demon wife, Mandodari), ostensibly because of the etymological derivation of kraufica from

the root, kruiic, meaning "to make curved or crooked," perhaps referring to a sort of moral crookedness; in the Kum&asambhava, the mountain on which the demon Taraka resides is also called krauiica. "In the thrall of sexual desire" (kdmamohitam) is now rendered as "filled with desire," i.e. R2vanats lusting for Sitli. Finally, Govindariija gives the slesa-like reading in its entirety: "May you, o &inivdsa, who killed the one among the pair of demons filled with lust, be forever victorious." Govindarlija concludes the slesa-like reading by glossing a rhetorical figure which he does not identify on the literal level-poetic

reason (kdvyalingam). In

kdvyalingam, an incredible or unbelievable statement in a verse is explained by the reason given; this reason may be provided either for the meaning of a sentence (v&y&?hahetukam) or for an individual word (paddrthahetukam). Presumably, it is Visnu's eternal glory which must be explained by his killing of Rlivana. Govindaraja merely mentions the presence of this figure, but this reference exemplifies the degree to which his analysis of the md ni$ida verse is suffused with the categories of Sanskrit poetics. And this dimension is most evident as Govindariija theologizes the verse through a reading closely resembling slesa, one of the distinctive technical features of ornate Sanskrit k w a. The ways in which Govindariija's reading of md niNda resembles slesa-with the construal of bitextuality through both resegmentation (sabda-slesa) and

double meanings (artha-s1esa)-should be clear. I would now like to look at the two primary ways this reading differs from slesa: 1) the absence of a tropic relationship between the two levels of meaning; and 2) the status of the secondary level of meaning as the product of reading.

3.1.2 Why This is Not a slW

Although the relationship between Sanskrit poetics and Sanskrit poetry is not a straightforward one, both in poetics and in poetry one dimension of slesa remains constant: the copresence of slesa with other figures. For a given bitextual utterance to be characterized as a slesa, the two levels of meaning must convey a relationship inherent in another figure (alamk6ra). This second figure may be a simile (upama), metaphor (riipaka), or another of the sense-based figures (arthalamkaras); Udbhata (c. 800) characterizes this relationship as the appearance (pratibha) of the second figure. The exact nature of this relationship constitutes one of the major points of debate in the poetic tradition, and the central issue is a taxonomical one: if slesa is copresent with other figures, what determines its autonomy viz. these figures? Answers to this question range from subsuming all other figures involving slesa under slesa

itself, as Udbhata does, to taking slesa itself to be a mere mode of these other figures, as Dandin does.7 To illustrate how such a tropic relationship functions in a slesa, we can return to the king-moon example cited earlier. Here, the relationship between the king and the moon is a metaphor (riipaka), which we might rephrase as: the king is the moon. This metaphor also governs the double meanings of the other elements of the verse such as fame/ the eastern mountain (udayam) and taxes/ rays (kara). Such tropic relationships are easily identifiable wherever ilesas occur in Sanskrit poetry and are in fact often marked by the presence of explicit indicators (iva for upamii, eva for riipaka, api for virodha, etc). One notable exception to this rule, which would seem to provide a possible opening for Govindaraja's reading to be classified as a slesa, is handavardhana's distinction between slesa and suggestion based on the denotative capacity of sound (sabdasaktimzdadhvani). In demarcating the domain for this category, handavardhana must differentiate it from slesa which it closely resembles. According to handavardhana, most previous examples of ilesa may now be subsumed under the new rubric of sabdasaktimiiladhvani. If the figure related to the bitextual expression is merely implied (&@a), it is a case of suggestion; if this figure is directly expressed

'

Bronner, Poetry at Its Extreme: The Theory and Practice of Bitextual Poetry (~lesa) in South Asia, 1999,259-260,

(through the presence of the words such as api for contradiction, or virodha, and adhika for distinction, or vyatireka) it is a case of slesa. But even more relevant to Govindaraja's reading is handavardhana's identification of a form of slesa in which no separate figure is present at all: the co-narrating of multiple plot elements (vastus). This category would seem to correspond better to the md niMda reading than a slesa requiring the presence of another figure, not only because it is difficult to conceive of a tropic relationship between the two levels as Govindariija lays them out, but also because it is not clear what could provide the focus of comparison as the word raja does in the king/ moon example. Govindariija's reading could in theory be classified as a slesa if slesa were characterized only as a form of bitextual utterance. The actual example handavardhana provides, however, is not so straightforward. The verse simultaneously describes the exploits of Visnu and ~ i v using a the exact same phrase from the earlier example of sabda-slesa ("the all-giver, lord of Sri," sarvado mddhavah, and "he who is always the lord of

May the all-giver, lord of Sri, protect you. He is the one who destroyed the cart, who once made his very body that conquered the demon Bali into a female form, who killed the raised serpent. That one who resides in sound upheld the mountain and the earth. The immortals praise him as, "destroyer of the head of R2hu.I' He himself gave a home to the Andhakas.

yena dhvastamanobhavena balijitkdyah purd stikfio yas codvrttabhujarigahd ravalayo agam gdm ca yo dhdrayatl yasydh* Mimacchirohara iti stutyam ca ndmdmar* paydt sa svayam andhakaksayakaras tvdm sarvado madhavah// May he who is always the lord of Umii protect you. He is the one who destroyed the god born of mind, who once made the conqueror of the demon Bali into his own weapon. His necklace and bracelets are serpents and he bore up the Gangs river. The immortals praise him as, 'Hara," the bearer of the moon on his head. He himself destroyed the Andhakas.

yena dhvastamanobhavena balijitkdyah purdstrtrikfio yascodvrttabhujarigahdravalayo gangdm ca yo dhdrayatl yasydhiih iaiimaacchiro hara iti stutyam ca ndmdmarduih pdydt sa svayam adhakaksayakaras tvdm sarvadomddhavahlI (D. 2.21) It should be apparent that this example is not substantively different from the king/ moon verse and that a similar metaphorical relationship between Visnu and ~ i v could a easily be posited.8 It seems that this category primarily allows handavardhana to preserve some scope for desa, now considerably marginalized by suggestion. So with the exception of this somewhat complex example of co-narrated plot elements, Govindariija's reading would fail to qualify as a slesa due to the absence of another figure. The fact that Govindariija's secondary, slesa-like level of meaning is wholly generated through a process of reading marks another departure from the normative hermeneutic protocols of Sanskrit. The question of whether the The actual king/ moon verse, however, is as Appaya Diksita tells us cited by handavardhana as a case of suggestion, because the two levels of meaning are respectively contextual and non-contextual.

esoteric meanings identified are proper to the text itself or the projection of its interpretive community is an issue for all of Govindariija's figural readings, but it is especially relevant in the case of slesa-like meanings that have the radical potential to alter the text itself. In Sanskrit, such hermeneutic questions are invariably adjudicated through appeals to authorial intention (vivaksa), which explains Atreya Ahobila's attribution of the slesa-like level of meaning to Brahmii's authorship and not Viilmiki's. The criterion of intentionality is somewhat different from that of the presence of the tropic relationship. While it would seem that reading an utterance as bitextual would lack the presence of another figure, it is theoretically possible to read double meanings which do in fact bear this kind of relationship. Govindariija's slesa-like reading is not a slesa, on both grounds: neither is the tropic relationship present, nor does this reading appear to represent the intended meaning, resting as it does wholly on grammatically plausible but unattested meanings for words like niMda and krauiica. The identification of a slesa is normally not a subject of controversy for Sanskrit readers, in part because of the presence of several standard features. Authors often explicitly identify a given verse as a slesa. Bronner also points out genre-based, implicit conventions such as "the unmistakable distinction between the bitextual and alliterative blocks," in Subandhu's Vdsavadatta, which serve "as an implicit yet clear sign for the location of slesa." Moreover, the use

of certain vocabulary and words indicating the presence of another figure (iva, eva, api) enables readers to identify a slesa. (Bronner, 1999,401-403). Govindarfija's technique, however, is hardly unprecedented in Sanskrit literary criticism. Other far more ambitious examples that turn slesa into a reading practice include Ravicandra's construal of the erotic collection, the Amamiataka, into a treatise on liberation (believed to be authored out-of-body by ~ankara)and Krsnadatta's transformation of every epithet in the Gitagovinda into praise for ~ i v a .But these examples, too, probably would not conform to the implicit and explicit criteria for identifying slesa and represent a similar second-order application of slesa compositional practice.

3.1.3. Govindaraja's Mudralamkara Reading But this single &a-like reading does not exhaust the fecundity of meaning at play for Govindargja in the md niMda verse. Rather, it is only the first of a series of interpretations that produce multiple meanings for the entire ten-verse passage preceding the md ni,pida verse. Through this intricate procedure, the single word krauiica in the md nisdda verse is made to yield no less than six different meanings. These subsequent readings of md ni@da are all applications of another figure from Sanskrit poetics: the figure of signing

(mudrdlamkdra). Although it is quite clear that this is what Govindariija has in mind, other scholars have not taken note of it.' The Kuvalaydnanda defines mudrdlarhkdra as the "indication of indictable meanings" (siicydrthasiicanam). Mudrdlamkdra occurs most commonly in benedictory verses (niindis) at the beginning of plays but also in other modes of composition, both within kdvya and without. The mechanics of this figure is clearly distinct from slesa in that the double meanings (usually names or titles) are merely indicated by the words of the verse and form no secondary level with a coherent syntax. Govindariija himself cites the injunction regarding the use of mudrdlamkdra in benedictory verses: "Whether through meaning or sound, there should be some indication of [the contents] of the poem" (arthatah sabdato vdpi mandk kdvydrthasiicanam).

An apposite example of how mudrdlamkdra differs from slesa is the first verse of Bhiisa's Pratimdndtaka, an important Sanskrit retelling of the Riimiiyana:

Others who have treated this famous verse include Bronner, Poetry at Its Extreme: The Theory and Practice of Bitextual Poetry (~lesa)in South Asia, 1999, and Robert Goldman in "Translating Texts Translating Texts: Issues in the Translation of Popular Texts with Multiple Commentaries" in Translation East and West, Ed. Cornelia N. Moore, Vol. 5 (Honolulu: EastWest Center and University of Hawaii, 1992) 93-106. Another provocative reading of mfi nisada built on the notion of the verse as the kernel of the poem (kavya bija) is that of Atreya Ahobila, who identifies various forms of suggestion therein, including the suggestion of a plot element (vastu-dhvani), suggestion of a figure (alamkdra-dhvani), suggestion of rasa (rasa-dhvani), suggestion based on sound (sabda-sakti-udbhava-dhvani),suggestion based on meaning (arthasakti-udbhava-dhvani),etc.

May the god of the furrow protect us in life after life. He is pleased with good verse, is beautiful on account of his good neck, and bears auspicious marks. That awe-inspiring lord is the unmatched enemy of the one who caused the goddess to cry. sitdbhavah pdtu sumantratusa(i?sugrivardmah sahalaksmanas cat yo rdvandryapratimas ca devyd vibhisandtmd bharto anusargamll lo Bhiisa's verse includes an indication of the names of various characters from the Riimiiyana including Sitii, Sumantra, Sugriva, Laksmana, Riivana, Vibhisana, and Bharata, as well as the title of the play, Pratimd. But since these meanings do not form an independent syntax, the verse is even less similar to a slesa than the slesa-like reading Govindariija provides for md niNda. Just as the Pratimdn@aka introduces narrative elements through the words (sabdatah) themselves, other plays and poems do so through meaning (arthatah). Thus, in Kiilidiisa'sAbhijfidnasakuntald, the description of creation in the introductory verse presages Visviimitra's creating heavens for Trisiinku. Other, non-kdvya, examples of the use of mudrdlamkdra include the Vfttamanimdld, which indicates the names of meters in each verse of a story as a mnemonic. Govindariija's first mudrdlamkdra reading involves an indication of each of the six books of the Riimiiyana (through meaning rather than sound, since the titles of the books are not directly stated). "Md-nisdda,''as ~riniviisa,refers to Riima's marriage to Sitii and the Biila Kiinda. "Pratistdm tvam agamah," now lo Bhasandtakacakram, Plays Ascribed to Bhdsa, ed. C. R. Devadhar (New Delhi: Motilal Banarsidas, 1962) Pratimdt@aka 1.

rendered as "keeping one's vow," refers to Riima respecting the word of his father and the Ayodhyii Kiinda. "~dsvatihsam@' refers to Riima continuing to fulfill promises to the sages in the Dandakiiranya and the &anya Kiinda. With the mysterious word, "krauiica,"the remaining four books are indicated, again on the basis of different etymological meanings from the root kruiic: going crooked and being small (kmiica gatikautilyalpibhdvayoh). The killing of the krauiica recalls Riima's killing of Vali and the Kiskindhii Kiinda, since Viili and Tiira as animals move laterally (tiryac). But krauiica as "small" also conveys Sitii's emaciation from separation and the Sundara Kiinda. As per the earlier slesa-like reading, kraunca also indicates the killing of Riivana and the Yuddha Kiinda. Finally, the Uttara Kiinda is also included with Sitii's emaciation. This first mudrdlamkdra reading construes the principle events of the Riimiiyana in a manner far more in the mode of kdvya retellings than Govindariija's summary of the books in the section on the eighteen traditional meanings, where Visnu's creation of the world is identified as the purport of the Biila Kiinda, Visnu's sustenance of the world that of the Ayodhyii Kiinda, and so on. The second and third mudrdlarhkdra readings are far more ambitious. Govindariija situates both in response to an objection regarding the propriety of Viilmiki representing his own experiences in the first fifteen verses of the chapter. After Niirada narrates to him in brief the story of the life of Riima, Viilmiki returns to his hermitage with his student Bharadviija; he prepares to

bathe in the river, only to witness the killing of the heron. The objector believes that this is a digression akin to describing the path to the Mandara mountain in heaven to one traveling to the Malaya mountain, since all that is relevant is Viilmiki's possession of the special qualifications for narrating the epic. Hence the chapter ought to begin with Brahmii's boon. This objection provides Govindariija with a convenient opening: he responds that the entire passage from verse three to md ni@da is yet again an indication of the meaning of the poem, this time of the two central themes of the epic described in 1.4.7: the killing of Riivana (paulastyavadha) and the great acts of Sitii (sitdyds caritam mahat). The verse itself reads: "He who had made a vow [Vdmiki] composed the entire Riimiiyana poem ( k w a ) known as "The Killing of Riivana" and "The Great Acts of Sitii" (kdvyam rdmdyanam knsnam sitdyG caritam mahat) (R. 1.4.7). Govindariija renders this explicit reference to Ramiiyana as kdvya into a summary of its theological content by applying the two themes as an extended mudrdlamkdra generating new slesa-like readings. Both readings rest on the presumption that Viilmiki's omniscient vision allows him to perceive distant events as if before his very eyes; thereby his comments to his student and his witnessing the killing of the heron convey multiple meanings related to the events of the epic. The first reading for the theme in question describes Viilmiki viewing Riima killing Riivana. The passage begins with Viilmiki pointing out a bathing spot to his student:

'Bharadviija, look at this bathing spot free from dirt, beautiful, with clear water like the mind of a good person." akardamam idam tirtham bharadvdja niiiimaya/ ramaniyam prasanndmbu sanmanusyamano yatha// (R. 1.2.3). In the "Killing of Riivana," Viilmiki actually views with his mind's eye the location of the building of the bridge over the ocean as described in the Yuddha Kiinda, perhaps considered a sacred bathing spot (tirtha) because of the pilgrimage site of Rilmesvaram. On the literal level Viilmiki asks his student to see (nisdmaya) the bathing spot, but on the figurative level he asks him to hear (nidmaya) the story. The next verse subjected to this mudrdlamkdra reading is 1.2.9: 'But the sage saw a pair of inseparable, sweetly chattering herons walking nearby." tasydbhyiiie tu mithunam carantam anapdyinam/ dadarsa bhagavdns tatra krauiicayor cdrunisvanam// (R. I .2.9) Since the word krauiica also means "demon" for Govindariija, the reference is to the pair Riivana and Mandodari. Several of the words bear similar double meanings: walking (carantam) means experiencing pleasures (bhogam bhuiijdnam); anapdyinam refers to Brahmii's boon to Riivana ensuring his near immortality; and sweetly chattering (c&-unisvanam) means enjoying the music of vinds. While some of these meanings are secondary lexical senses (as in the case of carantam), others simply reflect different referents (as in chnisvanam). On this reading, the "hunter" is none other than Visnu, who kills Hiranyakasipu,

~ i s u ~ i i land a , Riivana in successive lives. Other transformations include 'bearing a jeweled crown" for "tawny-headed" (tdmrasirsena, referring to the heron's being bloodied) and "possessing a vehicle" for "winged" (patrha, since patra can also mean vehicle). Viilmiki's pity is now turned to disgust (according to the Amarakosa this is another meaning for the word karund, jugupsd karzqzii g h r w ) , his comment "this is unethical" (adharmo yam) refers to Riivana's misdeeds and not the "hunter's," and the md niMda verse is, as in the slesa-like reading, a benediction to Visnu. 'The Great Acts of Sitii" reading proceeds along almost identical lines in describing Riivana's abduction of Sitii. Again, Viilmiki sees this event in his mind and describes it to his student. The bathing spot refers to the Godiivari river, and its environs to the Paficiivati hut where Riima and Sitii reside. On this reading, the two krauiicas are Riima and Sitii, thin on account of their asceticism, and the hunter is Riivana. Several of the adjectives describing Riima, the male krauiica separated from his partner, bear sectarian associations: ''inseparable," unapiyinam, means without beginning or end (ddyantaiz,lnyam); and "sweetly chattering," cfirunisvanam, means the promulgator of the Vedanta (sarvaveddntapravartakam). Finally, Viilmiki's pity (kdrunya) is directed towards

Sitii and is the reason for his composing the epic. The md nipida verse remains a curse directed at Riivana.

This phrase, "The Great Acts of Sitii," bears great significance in the Manipraviila rahasya works. It is quoted by Pariisara Bhatta and Pillai Lokiiciirya, and is alluded to in Vediinta Desika's reference to the "three separations" (vislesa-traya) from Riima that Sitii is forced to undergo. The significance of this phrase is due to the idea that the Riimiiyana, as the technical treatise on surrender (prapatti-sdstra), promulgates the soteriological role of the goddess as mediator. These last two examples are different from other mudrdlarhkara "indications of contents" in two significant ways. First, the interpretation of the preceding ten verses borders on complete slesa-like readings rather than a mere indication of names. Govindariija posits two separate levels of meaning, first providing a straightforward grammatical gloss of the verses in their own context. Compare the transformations of 1.2.3 and 1.2.9, whose secondary meanings form a parallel syntax, with the Pratimdndtaka and Abhijndnasakuntala examples. Second, the very fact that these meanings are wholly the result of reading is again distinct from mudrdlamkdra, where such meanings should normatively be an intrinsic feature of the text itself. Other Sanskrit figures like mudrdamkara, even more than slesa, conform to highly predictable patterns that allow Sanskrit aestheticians to characterize them as operating on the literal level (abhidha) rather than the figurative (lakwd) or suggested (dhvani).

3.1.4. ~lesa-likeReadings Two and Three: Verses Related to Allegorical Reading and Performative Substitution

One of Govindariija's most innovative interpretive practices is his allegorical reading of the entire Sundara Kanda narrative. I will treat this topic later in this chapter and also in the next. At this stage I wish only to introduce the basic contours of the allegory in order to contextualize slesa-like readings for 5.27.51 and 5.1.1. My interest, as with the md ni@ida verse, is to explore the degree to which these readings approximate the technical features of slesa. The allegory centers on Hanuman's journey to Lanka and the message he delivers to Siti. The Divyaprabandham commentators compare this message to that of a sectarian teacher (dcdrya) to an individual soul (jiva). Many of the characters and plot elements of the Sundara Kfinda narrative are treated allegorically: Rama is Visnu; Hanuman the dcdrya; Lanka the body; Ravana and Kumbhakarna are respectively ego (ahamkdra) and selfishness (mamakdrd); Indrajit and the other demons are desire (kdma), anger (krodha), greed (lobha), intoxication (mada), jealousy (mdtsarya), ostentation (dambha); the asoka grove is the mind or heart; and Sit%is the soul imprisoned in transmigration. The most concise synopsis of these correspondences is given by Vedanta Desika in a verse from the Sankalpasuryodaya (1.71). Govindaraja introduces these associations in slesa-like readings of two verses in the Sundara Kanda. The allegorical background is more complete for

5.27.51; the verse itself is mysterious and, like md nipida, seems to invite symbolic interpretations. Sitii is sitting forlorn in the Asoka grove: Trijatii sees an ominous dream and exhorts the other female demons to propitiate Sitii, when a bird calls out encouragingly from a nearby tree: 'And a happy bird who made that branch his home, always chattering sweetly, speaking words of welcome, seemed to encourage her with pleasure.'' paksi ca sdkhdnilaya prahmah punah punas cottamasdntvavddil susvdgatam vdcam udiraydw punah punas codayativa hptah// (R 5.27.51) Govindariija first provides a literal gloss, with no substantive modifications except for the identification of the verse as an augur (sakuna-nimittam) and the observation that some identify this bird as none other than Garuda, Visnu's consort. The allegory here involves the bird (instead of Hanumiin) as the dcdrya and Sitii as the soul. The structure directly parallels that of Hanumiin's subsequent conversation with Sitii in the Asoka grove, and a direct precedent occurs in the Divyaprabandham commentaries, where the names of dcdryas are associated with specific birds. According to Govindariija, the verse outlines the distinguishing characteristics of an dcdrya. The word bird ("winged,"paksi) on the secondary level refers to the dcdrya because the wings refer to movement (gamana) which may be understood as movement along the spiritual path. Govindariija quotes the following verse as support: "Just as birds travel in the

sky by means of both wings, so too the enjoined highest path is through knowledge and action" (ubhdbhydm eva paksdbhydm yathd khe paksipim gatihl tathaiva jndnakarmdbhydm niyatd paramd gatiw). Other words and phrases bearing double meanings include the "branch" (sdkhd) on which the bird is perched as the Vedic branches, the "words of welcome" (susvdgatam vdcam) as mantras passed down by tradition, and ''encouraging" (codayati) as the acdrya's command.ll The significance of other phrases is also impacted by this identification of the bird with the dcdrya: for example, the repeated (punah punas ca) quality of the bird's sweet chatter indicates the dcdrya's persistence in conciliating the soul who struggles with the lord and contrasts with the lord himself, who refused to teach Arjuna again when Arjuna asked him to do so during the Asvamedha sacrifice. This comparison allows Govindariija to assert hyperbolically the supremacy of the dcdrya to Visnu (bhagavoto pi atisaya) Govindariija's gloss for the verse which begins the Sundara Kiinda proceeds along similar lines, this time with Hanumiin identified with the dcdrya; the verse itself describes Hanumiin's path through the sky: T h e n the destroyer of enemies (Hanumiin) went by the path of the celestial singers in search of the place where Sitii had been taken by

l1 Govindargja develops the

alternate meaning of codaydti from the standard meaning of the root cud, to impel or enjoin, and the susvdgatam reading from two separate successive glosses of su: sustu sampraddydvicchedo yathd tathdcdryaparamparayd svasmai dgatam.

tat0 rivaganitdydh sitiy* satrukarsanahl iyesa padam anvesmm cira@carite pathill (R 5.1.1) The allegorical meaning is that the iciiya goes off in search of the student who is to be instructed. Govindariija concretizes the allegory through double meanings for almost every word. The "destroyer of enemies" is none other than the icdrya, since the synonym t'guru"is derived from the syllables gu (meaning darkness) and ru (meaning destroying). Similarly, "the path of celestial singers" may be understood as the path of previous iciryas due to the etymology of the word icirya from the prefix i plus the root car (hence "cirana" means dcdrya and "carite" means practiced). As per the set of correspondences outlined above, Riivana is non-discrimination, aviveka (in that he causes false talk, rivayati), and Sitii is the soul (similarly unborn and of the feminine gender by virtue of its natural state of dependence). Govindariija's reading of these two verses makes allegorical meanings concrete in a form very close to slew The readings are both similar to and different from slesa in much the same ways as in the m i ni@da reading. As with the m i nisida reading, these readings construct bitextual utterances conveying primary and secondary levels of meaning similar to a slesa. Here as well, the absence of a clear tropic relationship between levels and the fact that the

secondary meanings are produced completely as a result of reading is at odds with slesa. Some aspects of Govindariija's reading of these two verses, however, do not so closely resemble the ma ni,pida reading. First, the bitextual effect is more akin to a slesa based on meaning (artha-slesa) than a slesa based on sound (sabda-slesa). The artha-slesa-like dimension is evident in the use of multiple meanings for words such as sakhii (tree branch, Vedic branch), codayati (encourages, enjoins), andpadam (place, determination). Second, it is far more conceivable in this case to construe a metaphoric relationship between levels, as in the Visnul Siva example of handavardhana, the focus of comparison being the bird or Hanumiin, on the one hand, and the de&ya, on the other. This dimension, however, is not at all explored or elaborated upon by Govindariija. A further difference is that there remains even less possibility that the meanings could have been intended by Viilmiki, given that they are developed from a ~rivaisnavaallegorical tradition and not merely evocative of the Riimiiyana's plot as a kiivya (khya-blja). The double meanings also often rest on symbolic associations of words, such as bird (paksi) and Sitii, rather than lexical meanings. The last slesa-like reading practice is related to the technique I labeled in the last chapter performative substitution. Specifically, performative substitution refers to the tendency to amplify a given phrase or verse numerous, mutually contrasting ways in the mode of the repetition and substitution patterns of

verbal performance art. For ~rivaisnavas,this practice reflects the conception of the source as a "text of enjoyment" (anubhava grantha). Govindariija's commentary on 2.2.13 is an example of performative substitution couched as a slesa-like reading. Govindariija introduces Dasaratha's praise of Riima prior to his being named crown-prince by asserting that the secret meanings to be discussed stem from Viilmiki's phrasing and not the real-life Dasaratha's intention. In each of the preceding types of slesa-like readings similar caveats occur: for the md niMda verse as a response to possible objections or through the boon of Brahmii, and in the readings related to allegory through the explanation for the allegory itself. The verse reads: ''Indeed the illustrious older brother of Laksmana is a suitable leader, through whose command all the three worlds would be well-ruled." anuriipah sa vai ndtho laksmivdn laksmandgrajah/ trailokyam api ndthena yena sydn ndthavattaram// (R. 2.2.13) The primary focus of Govindariija's reading is the word I have translated as ''suitable," anuriipah. This is an example of aprddi-samdsa compound in Sanskrit, a tatpurwa whose first member is a preposition. In such a compound, the preposition itself is a truncated form of a past participle to be supplied in the grammatical analysis. Since these forms are highly predictable and of a fairly limited set, identifying them does not normally require elaborate interpretation. Examples include "an eminent teacher" (pragatah dcdryd

prdcdryah), "excessive wind" (prakso vdtahpravdtah), "exceeding a garland" (atikldntam mdldm atimdlah), e t c . Govindariija uses this opening to provide a number of successive readings for anuriipah in the performative substitution mode, with each new reading bearing sectarian resonances. So, anuriipah may mean having a pleasing form (anukdam riipam yasya asau), referring to Riima/ Visnu being an object of meditation for all the gods; or an attended form (anugatam riipam yasya asau), referring to Visnu's being the embodied soul for all beings; or a form with a continued sequence (anusylitam riipam yasya asau), referring to Visnu's avatdras. Finally, Govindariija reads anuriipah in a manner resembling a most elaborate form of slesa, the resegmentation of individual phonemes bearing symbolic meanings as listed in specific slesa lexica. Since V indicates the god Rudra, an-u-riipa becomes a Bahuvrihi compound meaning ''having a form other than Rudra." Govindariija provides two similar syllabic slesa-like readings for the phrase 'sa vai ndtho," in each case recombining the words in the manner of a sabdaslesa into a compound, "savaindtho." In the first reading, the compound refers to Visnu because vih means bird. Vaindtha is a secondary nominal formation meaning "lord of the bird," Brahma's vehicle is the swan, and so savaindtho thereby means "one connected with the lord of the bird," and Visnu bears l2 These examples are drawn from M. R. Kale, A Higher Sanskrit Grammar (Delhi: Motilal Banarsidas, 1972) 142.

Brahma on his navel. As fanciful as this reading may appear, it rests entirely on the common lexical meaning for vi as bird and is grammatically sound. The second reading is closely related and differs only in taking vainatha, as lord of the birds, i.e. Garuda. Praising Visnu as one whose vehicle is Garuda is one of three stipulated epithets; the other two are one who is accompanied by the goddess ~ rand i one who is served by devotees, as in Yiimuna's Catu@loki. This reading, just like the other slesa-like readings, is both similar to and different from a slesa for much the same reasons: despite the construction of a bitextual utterance, it lacks any semblance of a tropic relationship between levels and is clearly the result of reading rather than the intended meaning. Yet Govindariija's gloss of this verse is interesting in part because it employs the most technical aspects of slesa practice (resegmentation, symbolic syllables) in the service of theological exegesis directly in the mode of oral performance.

3.2. Other Reading Techniques

In this section, I compare the slesa-like readings with other reading techniques, connotative reading and allegorical reading. While the slesa-like readings approximate a category proper to Sanskrit poetics, the other forms of figural reading are completely foreign to Sanskrit, being either imported directly from Tamil Manipraviila or produced through the translation of the modes of Manipravda into Govindarfija's Sanskrit commentary. My use of the terms

'connotation" and "allegory"reflects an attempt to characterize reading techniques which are nowhere theorized by the ~rivaisnavacommentators themselves, and by discussing the similarities and differences between these and the slesa-like readings I hope to provide a comparative grammar of the figurative reading techniques employed by Govindariija. This comparison may help to illustrate the significance of the desa-like readings and also provide an analytical framework for concepts discussed elsewhere in the dissertation (i.e. connotative reading in chapter two and allegory in chapter four).

3.2.1. Connotative Reading

Govindariija's parabolic elaboration of the stories summarized in the last chapter, including Vibhisana's surrender, are examples of connotative reading. Connotation may be distinguished from denotation as surplus meaning, the element added to the content of an expression by an addressee." I think this may be a useful way of conceiving how Govindariija adds the concept ofprapatti to the narrative content of Vibhisana abandoning Rgvana for Riima. With Vibhisana's surrender refigured asprapatti, the entire narrative leading up to

l3 This is a stylistic distinction between denotation and connotation rather than one that is semiotic or logical. See Goren Sonesson,Pictoral Concepts:Inquiries into the Semiotic Heritage and Its Relevance for the Analysis of the Visual World (Lund: Lund University Press, 1989).

this event conduces to ~rivaisnavatechnical accounts of the six components (arigas) ofprapatti, the role of the mediator, etc. Of course, this characterization ofprapatti as supplementary to the narrative is my own; Govindariija's commentary collapses the sectarian concept with the narrative itself in a manner unlike Manipraviila oral and written discourse, where the didactic context is explicit. It interesting to note that Govindariija never uses the concept of suggestion (dhvani) to describe his readings of Vibhisana's surrender, Laksmana's servitude, or Jatiiyu's liberation. Suggestion is a third power of language beyond the primary (abhidhd) and secondary (laksani) levels of meaning, which involves resonances available to sensitive readers. One variety of suggestion in particular, suggestion in which the literal meaning itself conveys another meaning (vivaksitanyaparavdcya-dhvani),would seem to parallel the identification of meanings like prapatti. But for Govindariija,prapatti is not the suggested sense of Vibhisana's surrender, but rather its literal meaning, or else an esoteric meaning (rahasydrtha) available only to members of the community but nonetheless inseparable from the narrative content. The basic difference between connotative reading and slesa-like reading is that in the former, unlike the latter, there is no split between primary and secondary levels of meaning. The following diagrams illustrate this point:

Table 1 Connotative Reading vs. ~ l e s a

Connotative Reading

+

sign

+

meaning1

meaning 2

~lesa-likeReading

+ sign

meaning1

I 1

(normative tropic

1

relationship)

I

I +

meaning 2

In these diagrams and the others in this section, I use the term, "sign" as a neutral descriptor, a heuristic for comparing reading practices irrespective of the details of differing language ontologies. This concept of sign is a category encompassing different chunks of texts, from words to phrases to sentences and entire narratives. For the slesa-like reading, this entails a different standard of measure than that used in Sanskrit poetics, where the analysis is based on individual words and their meanings. So, for instance, we can think of the phonemic string, "mani-,yi-da" as a single sign divided into two, I'mi nisida" (do not, o hunter) and 'mi-ni@da" (lord of ~ r i )with , two separate meanings. The point I wish to make is that in the slesa-like readings, the meaning of a sign does not lead to another meaning, but the sign itself is construed as conveying multiple, separate meanings. And the separateness of the primary and secondary meanings is even more pronounced in the slesa-like readings than in slesa itself, given the absence of a tropic relationship between them. On the other hand, the " + ' I

I have used

to capture the surplus quality of connotation would be replaced by an " = , I t

an

absolute identification of primary and secondary meanings, for Govindariija himself, since Vibhisana's surrender for Govindariija isprapatti.

3.2.2. Allegorical Reading Govindariija himself refers to the functioning of the allegory of the Sundara Kiinda in his comment on 5.27.51 and 5.1.1: these allegorical meanings, which provide the background for the secondary slesa-like level of meaning, are applicable to the entire narrative of the book, so that a one-to-one correspondence between Hanumiin and icarya, Sits and the soul, and so forth, pertains throughout-even

though there is no indication of this in context (R.

5.27.51, R 5.1.1.). Yet slesa-like reading and allegorical reading represent two different interpretive codes, with a qualitatively different relationship between levels of meaning, even though Govindariija brings these together in the two verses examined above. There is a virtual absence of allegorical composition in Sanskrit literature. A very small class of rLipaka orpratika dramas resembles allegory, with narrative and philosophical levels of meaning and characters standing for abstract concepts. These include Jayantabhatta's tenth or eleventh century Agamadambara and Krsnamisra's Prabodhacandrodaya. There are also a number of Jaina allegorical legends, including Siddharsi Sari's tenth century

Upamiti-Bhavaprapafica-Katha.But aside from these few examples, allegory is not a class of Sanskrit literature. Allegory is also unknown to the Sanskrit aesthetic tradition. Its stream of two stable sets of meaning would seem similar to the simultaneous appearance

(rather than superimposition as per the trope of metaphor, riipaka alamkdra) of non-contextual meaning in collapsed expression (samdsokti). But the temporal duration of allegory in narrative differs sharply from the momentary quality of samdsokti. Just as with connotation, another possible classificatory option would be suggestion in which the literal meaning conveys another meaning

(vivaksitanyaparavdcya-dhvani),since this is exactly what occurs in allegory. Interestingly, Govindaraja himself uses the term "suggestion on the scale of the entire text" (prabandha-gata-dhvani,as opposed to suggestion on the level of the word, pada-gata-dhvani, or suggestion on the level of the sentence, vdkya-gatadhvani) in his commentary on 5.27.51 to describe the allegory. While this connection seems a thought-provoking way to capture the dynamics of allegory, Govindariija does not elaborate. Andprabandha-gata-dhvani is a term used in Sanskrit aesthetics exclusively for poetic sentiment, rasa. The absence of Sanskrit allegory contrasts with the presence of allegorical composition and reading in Tamil and other South Asian literary traditions. Not only do the Divyaprabandham commentaries and rahasya works like the Acdrya Hrdayam contain allegorical readings similar to that of Govindaraja, but allegorical composition is an element in South Asian vernacular languages and in both bhakti and sufi traditions. Yet none of these contains a sustained theory of allegorical practice.

What Govindaraja's allegorical reading most resembles is a form of Biblical figural interpretation. In descriptively applying the term, "allegory,"it is important to bear in mind that this is not a homogeneous category within Western poetics itself, but rather refers to widely divergent forms from the classical trope of allegoria to Biblical figural interpretation, hermetic traditions, and personification poetry. Furthermore, we must distinguish between allegorical reading in the mode of Govindariija, which projects allegorical meanings onto an antecedent narrative, and the supplementary interpretation that merely identifies the hidden content of an allegorical composition. To clarify the most relevant form of Biblical figural interpretation, I turn to the distinction between typology and allegory outlined in Augustine's On

Christian Doctrine. This is a basic distinction between interpretive codes or levels rather than the later fourfold classification designating different categories of meaning (history, anagogy, typology, allegory). In typology, the referents or events themselves signify real events in the future, while in allegory the meanings of words signify other concepts. So, for instance, the crossing of the Jordan river is read typologically as indicating something occurring in the future, while meanings like "lamb" or "winds" allegorically indicate other meanings like "Christ"or "souls." The following diagrams illustrate this difference, the crucial point of departure being the ontological status of the secondary level of meaning:

Table 2 Typological Reading vs. Allegorical Reading

Tv~olosicalReading

sign

+

meaning1

+

meaning 2 (ontological, future)

Allegorical Reading

sign

+

meaningi

+

meaning (not ontological)

Martin Irvine summarizes the difference between typology and allegory: "On the semantic level, allegory and typology are distinct: allegory functions on the vertical axis of signification and the correspondences among general terms, whereas typology operates on the horizontal axis of reference and the contiguity

of distinct agents and events in sacred history considered as proper names or unique referents."14 Govindariija's allegorical reading closely resembles Augustine's allegory and not typology. Not only do the allegorical meanings not signify events, they are not ontological, but rather are embedded in the immediate plot structure, so that ~rivaisnavasnever view Hanumiin as an iiciirya in other contexts; similarly, Sitii meaning merely the individual soul,jiva, stands in contrast to her unquestioned ontological status among ~rivaisnavasas the incarnation of the goddess ~ r i . These diagrams should demonstrate why Govindariija's allegorical reading is distinct from both his connotative reading and slesa-like reading. As in the very derivation of the word, in allegory the secondary level of meaning is other (allos) than the primary, while in connotative reading it is added to (or, on the view of the ~rivaisnavas,identical with) the primary. Furthermore, in allegorical reading, the secondary level is not viewed as being real in the way it is in connotative reading. Returning to the diagram for slesa-like reading may help us to clarify the difference between Govindariija's allegorical reading and his slesa-like reading:

l4 Martin Irvine, The Making of Textual Culture: Grammatica and Literary Theory, 350-1100 (Cambridge: Cambridge University Press, 1994) 264. I have also drawn the examples for typology and allegory from Irvine.

Table 3 ~lesa-likeReading

~lesa-likeReading

+ sign

meaning1

I 1

(normative tropic

1

relationship)

I

I

+

meaning 2

In ilesa-like reading, the secondary level of meaning comes from the sign itself rather than from the primary level of meaning, as it does in allegorical reading. So, whereas the sign "ha-nu-mGn"yields the meaning the monkey Hanuman, which in turn produces the allegorical meaning dcdrya, "md-ni-,ycZdat'is itself split into primary and secondary meanings. This difference in the relationship between levels of meaning is even more pronounced in the slesa-like reading than in slesa itself, since slesa-like reading lacks the normative relationship

between an instrument and object of comparison; in fact there is no obvious relationship between the primary and secondary levels of meaning in slesa-like reading.

3.3. Conclusion

We are now ready to return to the question posed at the beginning of this chapter: why choose to read verses in a manner so closely resembling and yet ultimately distinct from a category of Sanskrit poetics like slesa? My answer is that this approach is emblematic of the larger hermeneutic project of transforming the antecedent receptive history of the Riimiiyana as kdvya, what I call the theologization of the Riimiiyana. The slesa-like readings examined in this chapter represent different aspects of this project: 1) slesa-like reading one is voiced entirely within the idiom of kdvya, yet the secondary meanings are sectarian; and 2) slesa-like readings two and three domesticate vernacular interpretive techniques to the idiom of kdvya. ~lesa-likereading one parallels to Govindariija's characterization of the Riimiiyana as kivya. Govindariija explicitly classifies the Riimiiyana as a kdvya, first in the introductory passage where he cites Mammata's taxonomy of didactic forms (with kdvya teaching like a lover), but also elsewhere as in his comment on the bird-aciirya verse, 5.27.51, where he argues that the Riimiiyana is an

example of highest kdvya (uttama k*a,

as opposed to madhyama kdvya and

adhama kdvya), in which suggestion is predominant. This acceptance of the Riimiiyana's status as kdvya is evident in Govindariija's commentary on three verses in succession: 1) 1.4.7, which states that the Riimiiyana is a k m a (kivyam rdmayaqam kcsnam); 2) 1.4.8, which states that the Riimiiyana is to be recited and sung (pathye geye ca madhuram); and 3) 1.4.8, which states that all nine rasas are included in the Riimiiyana (R 1.4.7-9). Yet, ultimately, Govindariija's Riimiiyana is a theologized kdvya. For instance, in 1.4.9, Govindariija argues contra handavardhana that passion-inseparation (vipraZambhas.mgdra) and not pity (karzqzd) is the predominant rasa of the epic. The rasa of passion conduces especially well to the devotional relationship between human and divine that undergirds the allegorical reading of the Sundara Kiinda. Just as in these passages, so too in Govindaraja's slesalike and mudrdlamkdra readings of md niNda, the Riimiiyana is rendered as a theologized kdvya. While Govindariija identifies the verse as the kernel of the poem (kwa-bija), indicating the contents as other kdvya benedictory verses do, the secondary meanings themselves are sectarian (e.g., ~riniviisaas a name for Visnu). ~lesa-likereadings two and three are distinguished by their relationship with Manipravda forms of interpretation; desa-like reading two with allegorical reading and gesa-like reading three with performative substitution. As

discussed in the last chapter, the composition of Sanskrit commentary was a phase subsequent to the interpretation of the Ramiiyana in Manipraviila esoteric works. The relationship between slesa-like readings two and three and allegory and performative substitution may reflect an effort to domesticate these techniques in a form more familiar to Sanskrit (albeit not one conforming to technical definitions of slesa). It is difficult to p a g e the success of the ~rivaisnavahermeneutic project. Nevertheless, there is some evidence for the reception of Govindariija's reading of md nisda. Within kdvya itself, Ramacandra Buddhendra (Andhra Pradesh, date uncertain) incorporates the slesa-like reading almost word for word in his commentary on Bhoja's ~ a m ~ u r d m d ~ aFor n a .Ramacandra ~ Buddhendra, the theologized inaugural verse itself becomes the standard against which a kdvya retelling of the Riimiiyana is measured. And within the rival intellectual community of Advaita Vedanta, Niigesa Bhatta in his eighteenth-century Tilaka commentary does the same, an all the more remarkable fact given that he draws from a different recension of the Ramayana (Northern) and primarily adheres to ideas developed in the Adhyatma Ritmayana and Tulsi's Riimcaritmfinas such

l5 Campurdmdyana of Bhoja with the Commentary of Rdmacandrabudhendra, ed. Narayan Rama Acarya (Mumbai: Nirnayasagar Press, 1956).

as the shadow-sitii.16 These examples may indicate that the theologized Riimiiyana lost its explicit marking, so that Govindariija's slesa-like reading could be accepted in worlds where ~rivaisnavametaphysics was not accepted. In the next chapter I examine a poem, the Hamsa Sandesa, which embodies in its form the image of the Riimiiyana envisioned in the Ramiiyana commentaries. Its allegory turns the technique of allegorical reading into a mode of composition, and its sectarian content and topography renders it a theologized Rams kdvya.

l6 Ramdyana of Valmiki With the Commentaries Tilaka of Rama, Ramiiyanasiromani of ~ivasaha~a, and Bhmana of Govindaraja,ed. Satkari Mukhopadhyaya, 8 vols. (New Delhi: Parimal Publications, 1990) 1.2.15.

CHAPTER 4 THE ALLEGORY AND TOPOGRAPHY OF A R&MA MESSENGER POEM

The HaWa Sandesa of Vedanta Desika (1268-1369 CE), which recounts the imaginary vignette of Riima conveying a message to Sits via a swan, evokes the plot of the Sundara K m a through a generic form familiar from Sanskrit (and Tamil) literature, the messenger poem (sandesa-kavya). And yet, even as it alludes repeatedly to the Riimiiyana and betrays its close structural modeling on the Meghadfita, it is also a work of remarkable newness, transforming a classic depiction of longing and passion into a meditation on the significance of the teacher (dcdrya) for devotional practice. I explore here this central paradox: how the innovation of the Hamsa Sandesa is rendered through an engagement with antecedent sources. While the Meghaduta itself belongs to the literary series beginning with the Sundara Kanda, the Haws Sandesa seeks to transform this series in two key ways. First, its allegorical form brings the narrative of the Sundara Kanda into

conformity with specifically ~rivaisnavasoteriological conceptions; second, the construction of the route collapses the ancient action of the epic in time and space with the institutions of the late-medieval ~rivaisnavacommunity. I conclude with a discussion of how, as a special case of the theologization of the Ramayana, aspects of the Hamsa Sandesa correspond to each of the aspects examined in the previous three chapters, the shift from court to temple, the incorporation of vernacular forms into Sanskrit, and the modification of elements of Sanskrit kdvya.

4.1 Literary Antecedents

The highlight of Hanumiin's acts in the Sundara K a d a is the meeting he has with Sitii. Seated in a tall tree in the asoka grove, Hanuman sings the praises of Riima to gain Sits's trust, speaks to her in sweet and clear Sanskrit, displays Rama's signet ring, and receives in return an intimate anecdote (the story of the kdkdsura) to be conveyed to Rama. Commentators on the Meghadiita explicitly refer to the Sundara K w a as a source, a highly plausible claim given the primacy of the epic for Sanskrit poetic imagination and the close affinities between the separation of the lovers, the non-human messenger, the message of reassurance, and the aerial route. Of the more than sixty commentaries on the Meghadiita produced after the tenth century, Daksinavartaniitha's Pradlpa (most likely produced in

Tiruvalaiiculi, a small village near Kumbhakonam in the Cola region) makes the relationship with the Riimiiyana its centerpiece. Even as Daksinavartaniitha explicates the obvious references to the epic-the

setting in the RZimagiri

hermitage, the favorite ~ a i v image a of Riivana shaking Kailiisa (1.58), and the comparison with Hanuman's message towards the end of the second half (2.37)-he

also foregrounds deeper analogues. This influence is concretized

with imagery drawn directly from the Ramiiyana, especially in the descriptions of the yaksa's wife in the second half: her longing for his embraces even in a dream (2.28) recalls 5.34.21 (svapnepi yady aham viram); the braid of unkempt hair to be untied upon return by theyaksa (2.29) is an exact replica of the image of Sit2 in 5.20.8 (ekavem dhard sayya); the auspicious throbbing of the left thigh is just as it was with Sit%before Hanumiin's arrival, recounted by Trijatii in 5.27.50; and the yaksa's clasping of snowy breezes in the hope that they have touched his wife is drawn from 6.5.6, where Riima makes a statement to the same effect.' The literary series of messenger poems based on the Meghaduta includes dozens of extant works in Sanskrit and works in Pali, various Prakrits, and emerging vernacular literary cultures like Tamil, Telugu, Kannada, Malayalam, Sinhala, and Tibetan. The Sanskrit poems follow a standard structure: 1)

Meghadfita of Kdiddsa, critically edited by Sushi1 Kumar De (New Delhi: Sahitya Academy, 1957) (Hereafter cited in the text with the prefix, "ME.")

request to the messenger; 2) description of the route to be traveled; and 3) message. Many eulogize the love affairs and conquests of the royal patrons, spatially mapping out the boundaries of the realm in the manner of a royal victory tour (dig-vijaya). Examples of such courtly poems include the collection of twenty-three sound-based paronomasia verses (yamaka) of Ghatakarpara (one of the nine jewels of Vikramiiditya VI, 1076-1127 CE), the Pavanaduta of Dhoyi (patronized by the third Sena regent, King Laksmanasena, 1178-1206 CE), and the Kokilasandeh of Uddanda (patronized by the Zamorin king Miinavikrama, 1466-74 C E ) . ~The specificity of the route as a conceptualization of politically meaningful space is evident in the last two of these poems. The wind in the Pavanaduta passes from the Malaya mountains to the Piindya country, the Tiimraparni river, Kiifici, the Kiiveri river, Godiivari, Kalinga, the Narmadii river, and finally the Sena capital, Vijayanagara, in effect tracing in reverse Laksmanasena's southern conquest. So too the bird in the Kokilasandesa travels from Kifici to the Kampii river, the Cola region, the Viinmayi river, Koyyiiyam, the Ranakhala region, and the capital at Jayantamangala, significantly crossing through the north-eastern corner of the Hoysala empire.

For an exhaustive list of Sanskrit messenger poems, see Satyavat Sastri, Essays on Indology, (New Delhi: Meharchand Lachhrnandass, 1963).

Messenger poems were especially popular in vernacular traditions where regional space was significant, involving what Steven Hopkins calls the poetic crafting of a religio-geographic irn~zginaire.~ The aerial description of a concrete topos provided the perfect vehicle for charting this imaginaire in a manner analogous to the first vernacular versions of the epics. An even closer parallel to the Hahsa Sandeia is provided by didactic poems

that adapt the messenger motif to philosophical and religious pedagogy, a number of which were composed by Jaina intellectuals. The Piirsvdbhyudaya of Jinasena produced during the reign of the Riistrakiita king Arnoghavarsa I (814874) depicts the route to Alaka as paved by Jaina temples. Similarly, the Jaina Meghadiita of Merutunga (born 1346, a generation after Vediinta Desika), recounts a message to the mendicant Neminiitha from his bereaved, separated family. Finally, the Divyaprabandham corpus itself contains a number of poems where birds serve as messengers between the female devotee (the heroine, or pariiwa-niiyaki, as per classical Tamil akam poetry) and the lord. In Manipraviila commentaries on the Tiruviiymoli, which devotes four whole decads to this theme, the romantic relationship is viewed as an imaginative vehicle for expressing the &var poet's own devotional experience. In each of

Steven Hopkins, "Lovers, Messengers and Beloved Landscapes: Sandesa Kdvya in Comparative Perspective" (forthcoming) 1.

these four decads (1.4,6.1, 6.8, and 9.7), respective forms of the lord are invoked: 1) the Paficariitra emanations (vyiiha); 2) the incarnations (vibhava); 3) the lord as transcendent and immanent (paratva, antawmi); and 4) iconic forms (area). The messages are delivered by a variety of birds-storks,

koel birds,

swans, cakravdka birds, herons, parrots, and tinypuvai birds-identified

by the

commentators as dcdryas, their "wings" being knowledge and right conduct4

4.2 Comparing Messengers The Hamsa Sandesa contains a high degree of what A. K. Ramanujan has called self-reflexive reflexivity, the meta level on which a work makes explicit its linkages with other works.5 The primary reflexive aspect--the similarity between Hanuman and the swan as messengers from R&mato Sit&-is generated through ample hints to the reader, with the entire prospective action of the Hamsa Sandesa framed with the Riimiiyana narrative in relief.

I have based my understanding of the Hamsa Sandesa in large part on the published modern commentaries of Krishna Brahmatantra Parakala Swami (Parakala) and Uttamur Viraraghavacharya (Uttamur), although I also have in my possession several manuscripts of earlier commentaries summarized by 4

Tiruvilymoli of Nammillvilr, ed. B. R. Purushottama Naidu (Chennai: Madras University Tamil Department Series, No. 18, 1957-1962). A. K. Ramanujan, "Where Mirrors are Windows: Towards an Anthology of Reflections," History of Religions 28.3 (1989) 187-216.

Parakala and ~ t t a m u r The . ~ commentators betray slightly different approaches to the status of the H a M a Sandesa as poetic re-creation. In keeping with the rahasya tradition, Uttamur describes the Hamsa Sandesa as a "direct experience" (s&ad anubhava) of the Riimiiyana, in contrast with the more or less indirect relationship between the epic and the Meghadiita. This is an interesting turn of phrase, given the use of the very similar phase, "text of experience," anubhavagrantha, to refer to ~rivaisnavaoral commentary (U 1.1). Parkala on the other hand, makes use of a category of poetic virtues and faults: the Hamsa Sandesa possesses the virtue of abridgement, sanksepa, in summarizing the Riimiiyana narrative (P 1.1). Both Parakala and Uttamur identify one essential distinction between the Hamsa Sandesa and the Riimiiyana narrative: while the Riimiiyana represents fact, the Hamsa Sandesa is a work of fiction (kalpita) (U 1.1, P 1.1). This reflexivity breaks into the reader's awareness in the series of images in the second section, where descriptions of Sitii's forlorn state repeatedly evoke scenes from the epic. I present here translations of some of these parallel verses

Hamsa Sandesa of Vedanta Desika, with the commentary of Uttamur Viraraghavacharya (Chennai: Ubhaya Vedanta Granthamala, 1973); hereafter cited in the text with the prefix, "HS." Hamsa Sandesa of Vedanta Desika, with the commentary of Krishna Brahamatantra Parakala Swami (Mysore: Mysore Oriental Research Institute, 1917). In addition to his own commentary, that of Krishna Brahmatantra Parakala Swami, and the supplementary gloss (tippani) on P by Kasturi Rangacharya, Uttamur Viraraghavachar lists the following commentaries: the Sudhti, quoted by Parakala Swami and the commentaries of Vtidhula ~rinivtistirya,author of a commentary on Tattvamukttikaltipa, ~vettiran~anarfi~ana Stistri, and Melaipalayam Nadtidur Srirangactirya. I have obtained the following manuscripts from the Mysore Oriental Library: 1) E 26786, P.7211, grantha, paper; 2) 26787, P.9556110, grantha, palm leaf; 3) E26790, P.9556111, grantha, palm leaf (with commentary of ~ribhtisyasrinivtistiryatanuja).

to convey a sense of the closeness of the imagery, as well as of the way the images are lyrically developed and embellished. The second verse of the second section refers to divine women captured by R3vana waiting for Sita to liberate them: ItR3vana keeps the wives of the world-protectors (lokapdas) at a distance for fear of a terrible death from their curses; while they, who are as pure as your white color, could burn fire itself, they anticipate their own protection by the fiery power (tejas) of Sit3 and bear their imprisonment."7 Here, three separate verses from the Rfimayana are listed by Uttamur as sources: 1)R 3.37.18: "That Sit3 possesses an immeasurable fiery power (tejas); you are not capable of killing her in the forest whose refuge is R%matsbow"8;2) R 5.22.20: "Because R%mahas not ordered it, and so that I may preserve my penance, I do not reduce you to ashes, o ten-headed one, though you deserve to be ashes'"; and 3) R 5.55.29: "Through her penance, true speech, and fidelity to her husband, she herself could burn fire; fire cannot burn her.'''' All these verses occurring in widely disparate contexts revolve around the symbolic imagery of fire, penance

'

HS 2.2. ugraih sdpair upahatibhiyd rabasd duramuktah dagdhum yogyLi hutavaham api tvatpriydvamuddh$z/ utpasyanto janakatanaydtejasaiva svara&m rodham yasydm anuvidadhate lokapdldvarodh~ihll R 3.37.18. aprameyam hi tat tejo yasya sd janakatmajd/ nu tvam samarthas t h hantum rdmaciipasraydm vane11 R 5.22.20. asandesdt tu rdmasya tapasas cdnupdlandt/ nu tvam kunni dasagnva bhasma bhasmdrha tejasdll lo R 5.55.29. tapasd satyavakyena ananyatvdc ca bhartaril api sd nirdahet agnim nu tdm angnih pradhakshyatll

(tapas, also meaning heat), and tejas, which can mean luster, brilliance, or prowess, a fiery power derived from Sits's absolute fidelity to her husband. As Rama describes Sit%and her locale to the swan, more direct resemblances

follow in sequence. For instance, these words of the gods describing Rsvana in the Balakanda are obviously the source of Vedanta Desika's depiction of the asoka grove: "The sun does not afflict him, the wind does not blow at his side, even the ocean with its tremulous waves does not quiver upon seeing him."ll Compare H.S. 2.7.: "You will reach an arbor which was previously crushed by Hanumiin's trampling, where the wind is afraid of just the slight anger [of Ravana], where the sun appears like the moon, where all the seasons are pleasant, which is lit by asoka trees which are like the fire of Sits's grief."12 Other resemblances revolve around the description of Sits's longing. The ''mixture of sorrow and pleasure" which the swan will experience when seeing the beauty of Lanka's palaces and the pitiful state of captured women therein (2.6) is similar to Sit& words to Hanuman: "What you have said, oh monkey, is like a mixture of poison and the nectar of immortality, for Rama thinks only of

"R 1.15.10. nainam saryahpratapatipdrsve vdti nu m h t a h l calorrnimdi tam ddtva samudro pi nu kampatell l2 HS 2.7. isatkopdt cakitapavandm indusandigdhasGryim nityoddrdm ,Hubhirakhilair niskute v&aviiflm/ sitdsokajvalanasahajaih tatra diptdm asokaih dpadyeth* prathamalulitdm iiiijaneyapracdrai(d1

me but is overcome with grief."13 Later, the following, poignant verse develops Valmiki's repeated accounts of Sita's "crying, heavy breathing" (R.5.33.11): "Her eyes are empty, she breathes heavily, her lotus-face shrivels, the tears from her eyes flow like a mountain waterfall, she laments again and again; my Sit2 is from some unavoidable fate reduced to this pitiable state; this very slender-bodied one despairs within" (2.23).14 Even the passing reference to unavoidable fate ("vidhind durnivdred) seems to flow from words uttered by Sita, "Fate is certainly unavoidable for living creatures" ("vidhir nibmm sanhdryah prdnindm").15 What results is a tapestry of correspondences available for readers to identify which repeatedly focuses attention back towards the familiar original, reprocessing the narrative and developing it in a new poetic idiom. Vedanta Desika's referencing of the generic norms of messenger poetrywhich in a sense represents a receptive history of the Sundara Kdda--is structured in an even more self-conscious manner. Some of the correspondences are obvious: the two sections of about fifty verses (in the Meghadiita called the Piirva-megha and Uttara-megha, in the Hamsa Sandesa

l3 R 5.37.2. amflam visasams#am tvayd vdnara bhdsitamlyac ca ndnyamand rdmo yac ca sokapardya&ll l4 R. 5.33.11. "rodandd atini@v&vcid bhiimisamsparsandd api" HS 2.23. siinyd drstih svasitam adhikam militam vakrapadmam dhdrakdram nayanasalilam sdnubandho vildpah/ ittham dainyam kim api vidhind durnivdrena nitd sd m e sitd tanutaratanus tapyate niinam antahll

l5

R 5.37.4. "vidhir niinam asamhdryahprdnindm"

called the Prathamdsvdsa and Dvitiydsvdsa), manddkrdnta meter, the mirroring paths (Riimagiri north to Kailiisa in the Meghaduta; Rusyaktita south to Lanka in the Hamsasandesa), and matching seasons (the rainy season in the Meghaduta; the subsequent sarad season in the Hamsasandesa). But the

modeling is actually much stronger. Take for instance the first seven verses of both poems:

Table 4 First Seven Verses of Meghaduta and Hamsa Sandeia

MEGHADUTA

1. Setting, curse ofyaksa, separation from lover

1. Setting, Hanuman has returned from visiting Sits; R h a ,

separated from her, spends a sleepless night

2. Appearance of cloud, state of yaka

2. Appearance of swan while Rfima prepares army

3. Reaction of yaksa,

symbolism/ association of cloud

3. Resemblance to Sitfi, Rfima's dumbfoundedness

4. Welcome fromyaha; message

4. Riima's resolution to send

message; honoring of swan

5. Anomaly of insentient cloud as messenger, confusion of one in state of love

5. Reason for sending message with swan vs. cloud, mountain, etc.; more honor than Hanumiin

6. Praise of cloud, servant of Indra, statement of helplessness on part of yaka

7. You protect all from heat and must protect me now; go to Alakii

6. Praise of swan, transport of

Brahmii; qualities of Sarasvati

7. Divinity of swan, best of birds; appropriateness of condescending to send message

We can see that each verse in the Hamsa Sandesa corresponds closely to its counterpart in the Meghaduta. So the statement of the strangeness of an insentient cloud serving as a messenger is transformed into a positive assertion of the merits of a swan, even over Hanuman; similarly, the seventh verse of the Meghadiita, which puns on the cloud's ability to protect from heat (tapa, which also means suffering in Sanskrit), is matched by a hyperbole (atisayokti) in the

Hamsa Sandesa on the special status of swans as divine. The chart below shows how this patterning, though not exact, continues for the entire poem:

Table 5 Structural Comparison of Meghaduta and H a h a Sandeia

MEGHADUTA

HAMSA SANDESA

1.1-11introduction

1.1-14 introduction

1.12 take leave of Tunga mountain

1.15 taking leave of lotus

1.13 now listen to path

1.16 fly up, looking down on tribal women

1.14 siddha women remarking on your taking off 1.15-62 path until Alaka

1.17-1.60path until Lanka 1.21-1.22 description of Tirupati 1.23-28 description of Kaiici

1.27-37long description of Ujjain, Avanti

1.38-46description of ~rirangam

2. 1-11description of Alaka

2.1-2.6 description of asoka grove, location of Sit2

2.12-18 description ofyaksa's house 2.19

description of yaksa's wife

2.20-28 description of possible states of wife

2.10

physical description of Sit2

2.11-23 description of states Sita may be experiencing

2.29-2.35 further descriptions, comments

2.24-2.27 instructions to swan

2.36 preamble to message

2.28-2.32 preamble to message

2.37 explicit comparison to Sit2 and Hanuman 2.38-49 message

2.33-2.46 message

2.48-52 conclusion

2.47-2.50 conclusion

The result of this modeling is that the generic form of the messenger poem is continually made explicit for readers through signals and allusions, only to be transformed through the Hamsa Sandesa's implicit allegory. And given that the Meghadiita itself seems to be structured on the Sundarakwa narrative as shown above, transforming this generic form amounts to challenging a particular receptive history of the epic itself-as k&ya--reorienting the latter within the frame of a distinctively ~rivaisnavaconception of the soteriological role of the acarya.

4.3 The Allegorical Form of the Hamsa Sandeia

4.3.1 Allegory, Passion, and Bhakti The most novel element of the Hamsa Sandesa is its allegorical form, which corresponds with the allegorical reading of the Sundara Kiinda in the Divyaprabandham commentaries and Govindariija's Riimiiyana commentary, as

discussed in the last chapter. Here, according to the model adopted by Vedanta Desika, Rama stands for the lord, Hanuman is the acdrya, the ocean is the sea of transmigration (bhava-sagara), Lank2 is the body, Riivana is the mind, his ten heads are the ten senses, the asoka grove is the inner organ in which the lord resides, and Sit2 is the individual s o u l . The popularity of this emblematic paradigm is marked by the attested existence of the Srivaisnava maxim: "For yogis, he [Rama] neutralizes the terrible mind-demon of ten-heads with a ray of arrows that is discrirninati~n."~~ Again, this allegory is entirely contextual, embedded in the immediate epic plot structure, so that ~rivaisnavasnever view Hanumiin as the acarya or Sita as the individual soul, jiva, elsewhere. This allegorization of the emotional relationship between devotee and deity is facilitated by the overt erotic nature of the messenger poems, as highlighted in Parakala's characterization of the central rasa of the H a h a Sandeia as passion-in-separation (vipralambhasmgdra). This identification of vipralambha$pigiira as the predominant rasa of course corresponds to Govindarfija's later argument (contra handavardhana) that vipralambha$mgdra is predominant in the Ramiiyana itself. In the Divyaprabandham poetry, the longing of a young woman (viewed by commentators as a stand-in for the Alvar l6 As mentioned in the last chapter, this set of allegorical correspondances is summarized in Vedanta Desika's Sankalpasuryodaya 1.71. Sankalpasuryodaya of Vedanta Desika, ed. T. K. T. Viraraghavacharya (Chennai: Ubhaya Vedanta Granthamala, 1971) 1.71. l7 dasendriyiinanam ghoram yo

manorajanicaraml vivekasarajdena samam nayati yogindm11

himself) for the lord is the situation of choice for depicting the painful and arduous nature of separation. Quoting from a number of digests on rasa in his comment on the first verse, Parakala states that the H a d a Sandeia evokes passion-in-separation due to travel (pravaa-vipralambhsmgdra),i.e. incidental rather than intentional, We can closely observe this erotic dimension in the number of verses involving the trope of collapsed expression (samaokti), where erotic symbolism is joined to natural imagery. In 1.11,the non-contextual meaning of a lover drinking liquor and enjoying the company of his paramour is grafted onto the pleasant breeze accompanying the swan on the path: "The perfumed breeze that drinks the fresh juice of red lotuses moves at will among the lotus groves blooming at night and becomes moist with the ichor of elephants will enjoin the fragrances to compete with each other to please you."18 Here the lotuses (padmini) are identified with a class of women (padmini), drinking their fresh juice (navamadhu) is akin to a kiss, and so on. Similarly romantic liaisons are construed between the swan and a lotus in 1.15, Kannada and Telugu women in 1.20, the Svarnamukhari river in 1.23-24, and the Tiimraparni river in 1.52. While these images do not relate directly with the romantic relationship

l8 H S 1.11. araktanam navamadhu Sanair apibanpadmininam kalonnidre kuvalayavane dhiirpzmanas salilaml svinno danair vipinakarinam saumya sevisyate tvdm amoddndm aharnahamikam W a n gandhaviih$tlI

between Rama and Sits, they do conduce to the overall mood of passion as the commentators repeatedly point out. In one verse towards the middle of the second section (2.22), the merging of the rasa of passion and the practice of bhakti becomes direct: "She restricts her mind from all other objects, all states are held in check, and as per the scripture of erotic love (kdma) she places her mind in me alone. Due to the excellence of her meditation with her heart soft from inner melting, she experiences concentration free from distinctions (nir~ika~lasamddhi)."~~ Here Vedanta Degika plays ironically on the erotic intensity of Sit& longing for Rama. Uttamur connects the verse with a similar verse from the Ramayana: "She does not look at either the female demons or these flowers, fruits, and trees. With her heart steadfast she sees only Rama;" here the first quarter of the H a d a Sandeia verse parallels the first half of the Sundara K w a verse, while the remainder parallels the second half. 20 Uttamur proceeds with a technical discussion of the yoga references which articulate a specifically ~rivaisnavaconception of meditation. Again this is accomplished through a quarter-verse by quarter-verse gloss. After a statement

l9 H S 2.22. cetoydtim iamayati bahis sarvabhaume nirodhe m a y ekusmin prapihitadhiyam miinmatheniigamena/abhyasyantim anitarajuso bhiivanaydh prakarpit sviintendntarvilayam~dunii nirvikalpam samadhim//

20 naiM pasyati rdksasyo nemin puspaphaladrumani ekasthahrdayi niinam rimamevdnupasyatilI

of the goal of yoga as the restriction of mental activity, he says that the states

(sarvabhaume) refer to the first three of the five yogic stages associated with various degrees of activity and inactivity; the latter two stages, one-pointedness

(ekdgratd) and restriction (niruddhatd), are actually valorized in the second and then in the third and fourth quarters respectively. The last phrase of the verse, "concentration free from distinctions"

(nirvikalpakam samddhi) draws these yoga references toward a topic of great significance for Vi6is;gdvaita philosophy: attributive knowledge

(dharmabhlltajfidna). As Riimiinuja elaborated it in the great thesis in Brahmasutrabhasya 1.1.,Visistiidvaitins view knowledge as both qualified by the object perceived and self-luminating (arthaprakdsika and svayarhprakdsika). Consciousness as a quality of the self (atman) is here compared to the flame of a lamp or candle, which illuminates objects with its light but is not dependent on objects for its luminous power. This view of consciousness is related to the critique of Advaita, where consciousness is not a mere quality but the self itself, and also to the development of a distinctive theory of bhakti as outlined in RiimSnujatscommentary on Bhagavad Gita 9.34. According to this view, bhakti is characterized by remembrance, a constant meditation on the lord similar to an unbroken stream of oil poured from one vessel to another, wherein the normal distinction between knower and known is effaced.

And yet, as closely as these images would seem to adhere to standard Visistsdvaita concepts, one phrase stands out as highly incongruous: Sits is said to meditate on Rama according to the "scripture of erotic love" (mdnmathendgamena). The reader is left wondering whether the romantic context is in the service of the philosophical allegory or the yogic context itself serves as a metaphor for Sitg's longing.

4.3.2 Esoteric Readings

As discussed in the last chapter, allegory is not commensurate with any category in Sanskrit aesthetics. What we have is literary practice that breaks through the established categories of Sanskrit poetics and aesthetic theory. Of the forms of discourse categorized in Western poetics as allegory, perhaps the form which most closely resembles the esoteric aspect of the H a d a Sandesa is hermetic teaching, where the precise figurative import remains shrouded in mystery. To some degree, this ambiguity is endemic to allegory, where the secondary sense must necessarily be supplied through interpretation. Uttamur brings out the esoteric meanings in an unusual manner by reading each verse in a manner similar to a slesa, a move which allows him to fuse the supplementary discourse of commentary with the original. These ilesa-like readings are highly idiosyncratic, even more so than those of Govindargja examined in the last chapter. In his introduction, Uttamur is cognizant of the ways his commentary

marks a departure from previous practice, but he emphasizes this departure is based on a central principle: "Whatever I have written that is special is out of consideration of the fact that in this poem, as in the Riimiiyana, spiritual (ddhydtmika) meanings are suggested. Bearing this in mind, knowing that the connoisseurs will refine what I have illustrated for the first time for public disclosure with respect to both individual verses and the poem as a whole, bowing down to the lotus-like feet of the dcdya, Riima, and Sitii, I am ~ o n f i d e n t . "The ~ ~ sense of what is unprecedented in Uttamur's commentary is his making explicit what was always implicit: the esoteric rahasya meanings. While we must be careful not to conflate Uttamur's virtuoso interpretations with the text of the Haws Sandesa itself, they do provide a useful insider's view into a dimension of the poem which remains otherwise obscure. What is especially complex about Uttamur's approach is that he lays out two separate allegorical levels, at least for the first several verses of the first section:

1) the message sent from the individual soul to the lord; 2) the message sent from the lord to the individual soul. Hence the literal triad of Riima-swan-Sits gives rise to the allegorical triads of both lord-iiciiya-soul and soul-iiciiryalord. The particular trajectory of these two allegorical readings follows a

tatra tatra yad viseppzoktam asti, srimadrdmdyanasyevdsya kdvyasydpy ddhydtmikdrthavyaiijakatvamastiti manasikrtya dikpradarsanamdtrabuddhyd tatra tatra pfihak ca bahuslokdrthasamdhdrena ca yad utprek$tam asti, tat sarvam pardnuktam sahrday* sambhavya rava~ pariskarisyantiti ~ m a d ~ c ~ a s r i s i t ~ m a c a r a ~ r a v i n d a ppratyemil

distinct logic: mirrored messages from the soul to the lord and from the lord to the soul echo the reciprocity of the devotional relationship. Uttamur's collapsing of these two allegorical levels is clearest in his commentary on the very first verse of the poem: "Born in the dynasty of the sun, respecting his humanness, the god, the husband of ~ r ialert , in the search for Sit& endured a night that was like an eon (kalpa);he had just been reassured [as to Sitii's well-being] by ~ a n u m i i n . "On ~ ~the first esoteric reading, Uttamur reads the adjectives describing Rama from the perspective of an aspirant seeking liberation: the word god (devah) is restricted to its etymological meaning of luminosity, "alert in the search for Sitii,"means lost in the wheel of transmigration, i.e. engaged in seeking protection for one's parents and children, and so on.23 While much of the meaning of the verse remains stable here--Ri+ima is simply interchanged for the soul and adjectives describing him are rendered applicable to the plight of the soul--at other places this first allegorical level is possible only if the given verse is taken apart and put back together in a radically new way. But the true complexity of Uttamur's approach is the way in which the first and

22 HS 1.1. vaniie jdtah savituranadhe mdnayan mdnusatvam devah srimi.in janakatanaydnvesane jdgarUkah1 pratydydte pavanatanaye niscitarthah sa kiimi ka lpdkardm katham api nisdm dvibhdtam visehell

23 Uttamur plays on the epithet used to descibe Sita, "daughter of Janaka" (janaka-tanayd, the word janaka, the name for Sit& father, meaning "parent" or "source"), as engaged in seeking protection for ones' parents and children, i.e. lost in the wheel of transmigration, and so on.

second allegorical levels are combined with the very same string of syllables giving rise to directly opposed meanings. The second esoteric reading follows much more easily from the literal meaning and the only phrase amplified through slesa-like reading is the "seeking the daughter of Janaka" compound, now reinterpreted as "seeking souls who realize him as their progenitor (janaka) and hence are his offspring (tanayii)." The rest of the commentary on this verse is devoted to spelling out the allegory, with Hanuman identified as an acdrya, his return associated with a reminder to Visnu to protect transmigrating beings, and 'morning" (vibhiitam) viewed as the maturation of attributive knowledge. Uttamur also draws the two allegorical levels together for the third verse describing Rima's first vision of the swan: "The hero's eyes were riveted on this swan, whose gait resembled Sit&, whose form is the symbol on her silk garment, who made the same sound as her anklets. Caught up in her he lost his mind for a moment; how sharp is the punishment of the god of desire at any given moment!"24 Here, the description of the swan conduces to the standard definitions of an acdrya. "Whose gait resembled Sitii's"is interpreted as one whose compassion towards souls rivals ~ r i ' s"whose ; form is the symbol on her silk garment" (taddukd&ikamurtau) means he receives the garlands, clothes,

24 H S 1.3. tasmin sitdgatim anugate taddUkuldnkamUrtau tanmaiijirapratimaninade nyastanispandadmihl viras cetovilayam agamat tanmaydtmd muhiirte sanke tivram bhavati samaye sdsanam minaketoh//

and food offered to the divine couple, and "who made the same sound as SitG1s anklets" refers to the sound of the dcdrya's sweet teachings. The second esoteric reading is somewhat more straightforward than the first. The thirteenth verse of the first section contains a beautiful metaphor of the clouds during the autumn season forming umbrellas for the swan: "The autumn clouds will serve as umbrellas in the sky for you, with the subtle rays of the sun forming supporting rods, and, being multicolored at the ends, the refracted ~~ identifies a series of rainbow will appear as a multicolored ~ 1 0 t h . "Parakala metaphors (riipakas) between the rays of the sun and rods, etc., which facilitates the simile (upamd) between the clouds and umbrellas. For Uttamur, on the esoteric level, students of the dcdrya carry the umbrellas for him, an image of service to royalty and divinity. This rahaysa reading resembles a resegmentation slesa, with "rainbow" split into "bow of Indra" and the appellation for Indra, one of a hundred sacrifices, 'satamakha," lifted from its compound and transformed into an epithet for the dcarya. Next, the adjective "of the sarad season" (sdradd) neatly works as a name for the goddess Sarasvati, since the word used for clouds (vdrivdz$z, literally, "bearing water") can also mean "casting off obstacles" (van read like

25 HS 1.13. siik.yndkarair dinakaraih kalpitdntas s a l w sdropdnt* satamakhadhanus sesacitrd*ukena/ iidhdh pascdd ucitagatind vdyuna rajahamsa chatrayeran nabhasi bhavatah sdradah vdrivdhLih//

virika, opposing, obstructing, and the upapada from the root vah having the extended sense of "taking away"). Another resegmentation slesa-like reading is formed with the two separate words collapsed into a compound (sdraddvdriv&$z)meaning the students "who cast off obstacles to the dcdryats eloquent speech." Uttamur's esoteric reading at several places appears entirely symbolic, avoiding overt slesa-like reading and more closely resembling straightforward allegorical reading. For instance, in 1.54, the neat symmetry of the simile (upama) in the verse seemingly lends itself to an esoteric analysis: "Waiting a moment there, like an arrow from the bow-shaped beach quickly cross over and make Lanka your target; the sound of your bow-string will become manifest in the resounding streams that flow from the mountains on both sides." Here the speech of Hanumiin is recalled: "Just as the arrow released by Rama goes breaking through the wind, so I will go to Lankii which is protected by ~ i i v a ~ aWhile . " ~ ~the swan as a whole is compared to the arrow, its wings also resemble the feathers of the arrow, the shore of the ocean is like a long bow, the two mountain peaks are the two front portions of the bow, and the rivers flowing down from them are the bow-strings; just as these rivers produce a resonant sound, the bow-strings are also sonorous.

26 yathd

rdghvaninnuktah Sarah Svasanavikramah/gacchet tadvat gamisydmi lankdm

rdvanapdlitdml

Uttamur develops his esoteric reading on the basis of the following verse: T h e bow is the OM, the arrow is the soul; one should ceaselessly know the universal soul (brahman) which is said to be its target and be like the arrow, fixed on it."27 Here the symbolically rich set of images of bow, arrow, and target are appropriated to the context of soteriological knowledge. What is perhaps most interesting about this particular reading is that it is built up from the metaphorical level, the metaphor of the verse that provides the bow, arrow, and target imagery; we have seen the same approach above with the umbrella verse. In this example the only element linking the verse with the overall poetic framework is the identification of the swan with the iiciirya. We can see a nearly opposite approach in the first verse of the second section, where a blend of symbolic and ilesa-like readings is built on a detailed series of imagistic correspondences between the swan and the women of Lanka. The verse is an obvious parallel to the first verse of the second half of the Meghadiita, where the palace in Mount Kailasa is depicted as closely resembling the cloud. Here, nearly every linguistic element of the detailed series of matching adjectives linking the swan with the women of Lanka is tapped for figurative associations. While most of these literal correspondences are ontological, some are based on word-play or metaphor.

27pranavodhanuh saro hydtmd brahma tallaksam ucyatel apramattena veddhavyam saravat tanmayo bhavetl

For example, the swan's beautiful sound (carunada) is like the women's' anklets in producing pleasant noise (sasinj*); similarly, the swan's whiteness (gauram) is compared to the women's white limbs (apandurangyah). Elsewhere, the parallels are merely linguistic. An interesting set is the pair of adjectives "bhalliik~am"and '!smara$arad@o." The former is another word for swan, based on the similarity between a swan's eyes and the crescent-arrow, or bhalla; the latter is a bahuvrihi compound containing a simile (similar to "one with a face like the moon," candramukhi): their eyes are like the arrows of the god of love. The comparison with arrows is quite different in each case, its shape in the former and an alluring capacity in the latter.28 On the esoteric level, Uttamur identifies the women that the swan encounters with good people that the a c w a might meet in the mundane world, who have the capacity to become like him.2g

manasi jndl jndnam y m m tah. Even more contingent on a linguistic rather than ontological similarity is the pair, "miinasdrham"and "manoj&&" Both clearly involve elements of the nominal root (pratipadika) "manas,"or mind. This meaning is clearly present in the second, where the compound may interpreted two ways depending on whether the bahuvrihi is glossed with an objective or subjective genative: 1) "alluring,"in that they are the object of the heart's knowledge; 2) "captivating,"in that they know the hearts of men. In the first it only remains as the etymological root for the lake MZnasa in Kailasa.

''

The de~a-likereading occurs for the adjective "with pleasant gaits," (lalitagaman@) which is trasfomed to mean that their conduct is beautiful on account of being approved by the iastras, Here 'kamana" is used in the extended metaphorical sense of "conduct," far removed from its standard lexical meaning of "going." On the other hand, "with white limbs" (apd&urarigryah) is simply glossed as being endowed with the quality of goodness (sattva), presumably on account of the association of the color white with purity, etc. Here, no attempt is made to transform the compound literally.

4.4 Mapping "Sectarian"Space

4.4.1 What Kind of Space?

As with the other messenger poems, the description of the route in the Hamsa Sandesa is not incidental but rather crafted in a purposeful manner, so that the Riimiiyana mythic action becomes mapped in time and space onto the institutions of the late-medieval ~rivaisnavacommunity. Before moving to an examination of the specific sites described, it will be helpful to pause to consider exactly what kind of space Vedanta Deiika is fashioning and what this reveals about the overall constitution of ~rivaisnavacommunity. ~rivaisnavasself-consciously eulogized specific pilgrimage centers as divine places (divya desas), demarcating of the boundaries of the ~rivaisnavasocial community in a way that both overlaps with and transcends other forms of regional space in late-medieval South India. I compare this imagined space of the ~rivaisnavacommunity with two other spatial constructions: agrarian space and vernacular space. These may be said to be constructed in the sense that such spaces may be seen as always involving a mediation between geography or distribution, on the one hand, and socio-cultural imagination, on the other.

In the political world of late-medieval South India, spaces of region and locality were produced through shifting patterns of agricultural exchange. David Ludden discusses a diachronic series of agrarian zones (beginning from the ninth century with the nddu as a territory of ethnic identity), all of which are subsets of the 'ancient specters' of region described in the Sangam corpus, the four macro-regions (mandalams): Tondaimandalam (land of the Pallavas), Colamandalam, Piindyamandalam, and ~eramandalam.~' As we will see below, Vedanta Deiika invokes these macro-regions, which represent an important vector against which he develops the space of the ~rivaisnavacommunity. As R. Champakalakshmi has shown, the temple centers associated with ~ a i v and a Vaisnava communities in fact represent one process of the centralization of regional space from the period of the The larger space of vernacular literary culture oriented to the emerging regional courts represents a conception of the local perhaps even more germane for ~rivaisnavas.The characterization of Visistiidvaita as Ubhayavediinta (the end of both Vedas, Sanskrit and Tamil), with its concomitant claim of Tamil as a language of revelation, meant that language choice was a highly consequential

30 David Ludden, "Specters of Agrarian Territory in Southern India," The Indian Economic and Social History Review 39,2 & 3 (2002) 233-257.

See R. Champakalakshmi "The Study of Settlement Patterns in the Chola Period: Some Perspectives" in Man and Environment 14,14 (1989) 91-102.

issue for ~rivaisnavas.Some of the relevant issues with respect to the linguistic register of Manipravda, in particular, were discussed in the second chapter. But even as the space of the ~rivaisnavacommunity incorporates these social and linguistic boundaries, it is qualified by something more, namely the fact that the ~rivaisnavaswere a religious community. Again, demarcating the ~rivai~navas as a distinct institutional field does not qualify the divya deias as 'sacred" spaces in the mode of Eliade, whose account of an "inherent capacity" realized through theophany, hierophany, or signs obscures the way societies construct spatial meaning, even in the case of sites believed to be autochthonous. In late-medieval South India, the composition of mihitmyas or sthalapuranas associated with particular temples demonstrates the carefully constructed significance of such sites. These texts describe the special powers associated with a site, its mythic history, and the adventures of the worshipped deity. The divine places are normatively catalogued as a list of 108 pilgrimage centers.32 ~rivai~navas understand these to be the temples described by the aviirs, all sites in which Visnu resides in an iconic form (area) that is viewed not just as a representation but as an incarnation (arcivatira); the emphasis,

32 For a list and description of the divine places see L. V. Gopalan, ~rivaisnavaDivya Desams (Chennai: Sri Visistadvaita Pracharini Sabha, 1972) and Katherine Young, Beloved Places, unpublished diss., McGill University, 1980.

therefore, is on the direct presence of the lord on earth. Various Sanskrit, Manipraviila, and Tamil terms are used synonymously to refer to these sites: ukantarulinanilarikal (places that are beloved and blessed), ukant-urellam (all beloved places), tayappatikal (unique place), vilakpqzasthdna (unique place), ukantarulinadesankal (places graciously loved), and, simply, astottarasatashtdndni (108 places). These terms date to the time of Riimiinuja's immediate disciples, but a a g i a Manaviila Diisa compiled the standard list accepted today. The four most significant sites are those said to have been frequented by Riimgnuja: ~ r i r a n ~ a Tirupati, m, Kiiiici, and Melkote. While 96 of the 108 sites are in South India (primarily in the Tamil country), two are not of the earth: the milk ocean upon which Visnu sleeps and his heavenly realm of Vaikuntha. What we have, then, is a demarcation of space predicated on the relationship between human and divine worlds, a spatial mapping of the local as coterminous with the divine.

4.4.2 Divine Places along the Route In the route depicted in the Hanisa Sandeia, we can see how this constructed space of the ~rivaisnavacommunity, rendered coterminous with the action of the Riimiiyana itself, operates on a number of different vectors oriented toward the divine places, the ultimate marker of the boundaries of this space. As Steven Hopkins has pointed out, the aerial view recalls the world conquest (dig-

vijaya) described in the Sankalpaswodaya, which again marks Southern, especially Tamil space, as the boundary of the ~rivaisnavacommunity.33 But the fact that the swan is made to pass through no less than seven of the divine places, including the central triad of Tirupati, Kafici, and ~ r i r a n ~ a m , demonstrates that regional space is subordinated to the space of the religious community. Before Rama tells the swan the details of the route, in 1.18 he describes two separate paths to the South: "Hanuman described to me two paths: a (shorter) western path close to the Sahya hills near Coorg, which is always rainy and unpleasant, and an eastern path filled with wondrous sites and kingdoms; take the latter path, but keep your eyes from these distractions for my sake."34 On other occasions as well, Vedanta De6ika invokes the special qualities of the southern region indicating not just a preference for familiar locales but also an assertion of the favor of the divine towards this region. The pragmatic instruction to avoid the western path on account of rains facilitates a contortion of the route of the swan from the western edge of the Deccan (the Mdyavanta mountain, presumed in medieval India to be near the city of Vijayanagara)

33 Hopkins, "Lovers, Messengers, and Beloved Landscapes: Sandesa Kdvya in Comparative Perspective," Forthcoming, 4-6.

34 mdrgau samyag gama hanumatd vamitau dvau tayos te sahydsanno py anagha subhagah pascimo nityavarsahf prdcinesu pratijanapadam samhatdv adbhutdndm magnd d r s a katham api sakhe mat k,@ete nivdrydl ( H S 1.18)

towards the southeastern edge of the Bay of Bengal-the

exact area in which

Tamil is spoken. On the pretext of prescribing an easier route, Vedanta Degika has Rama construct a pathway through the primary centers of ~rivai~nava intellectual and cultural activity-Kaiici, Tirupati, and ~ r i r a n ~ a m And . the admonition to avoid wondrous sites ironically draws attention to these very sites, the divine places which are present "in every kingdom." As elsewhere, a specific image from the epic provides the frame, Hanuman's own journey South and account of the two paths. As I chart the route described, I will pay special attention to the specific divine places Vedanta Dedika refers to and to the ways these overlap with other conceptions of regional space. Note that the 108 d i v a deias are normally classified according to the n@u agricultural zones-Cola,

Pandya, Malai, Nadu

(middle), Tondai, and Vada (northern).

A few verses earlier (1.9), Rama's advocacy of the southern region is even more specific: "Go to the motherly southern regions, adorned with divine places, beautiful on account of sandal forests, pervaded with the wind of the Malaya mountain, and full of pearls, in order to revive my beloved, Sita. But endure this minor fault: it is the land of the demons."35

35 HS 1.9. sthanair divyair upacitagupc?m candanaranyaramyam m u k t m t i m malayamarutiim mataram dak$na£amasmatpri~aijanakutanaya jivitdrtham ca gacchan ekam raksam iti sakhe do;aleSam saheth$zll

The first specific regional reference is to Kannada and Telugu speaking women in the rural border area in verse 1.20: "You will take pleasure in the joy mixed with love of the girls guarding the paddy fields, who sleep in beds made of foliage amidst the sugarcane, please each other with chatter, and sing songs proper to the Karnataka and Andhra regions."36As elsewhere, Vedanta Degika retrospectively projects contemporary notions of space (here linguistic usage) into the hoary past, such that the speaking communities of Kannada, Telugu, and Tamil are conceived of as reified transhistorically. Interestingly, Vedanta Deiika mentions no specific sites for the entire Kannada region, including Tirunarayanapuram (Melkote), whose significance rivals Tirupati and Kaiici by the Vijayanagara period. It is a particularly Tamil regional space which Vedanta Degika chooses to foreground. As we will see, not only the entire Kannada region, but vast swaths of Cola and Pandya land are also cast in the background in favor of Tirupati, Kafici, ~rirangam,and other pilgrimage centers. While Vedanta Deiika does not explicitly clarify the criteria for the choice of itinerary, the commentaries, especially Uttamur, are quite clear that the objective is to provide opportunities for the direct experience of the lord as present in his iconic forms.

36 icchucchdye kisalayamayam talpam dtasthusi@m salldpais tair muditamanasdm sdisamraksikdnami kam&indhravyatikaravasdt karbure gitibhede muhyantindm madanakalusam maugdham dsvddayethdh/f ( H S 1.20)

The route proper begins at Tirupati. The Tirumalai hill was widely considered to be the outer boundary for the Tamil region, as attested first in the classical Tamil grammar, the Tolkappiyam (fourth-sixth century CE). As S. Krishnaswami Aiyangar points out, "These hills form a feature of the frontier half a degree to the north of Madras, extending the whole length from the Mysore plateau and stretching eastwards to almost near the coast at Ponneri, and thus constitute a prominent feature of the northern extremity of the Tamil land.l13' Aside from the geographic prominence of the hills, in even the earliest references there is an awareness of a shift in language use in this area, from Tamil to the "vaduku" language, a somewhat amorphous category meaning simply, "Northern speech.'' Tirupati is the most important divine place in the northern region, or Vadanadu. Vedanta Deiika's description of the site covers four verses, 1.211.24, and includes descriptions of the hill, the main icon in the Tirupati temple, and the Kanakamukhara river where ~ i v is a believed to have worshipped V i g . The first verse is typical of how Vedanta Degika introduces these pilgrimage sites, with poetic imagery weaving mythic associations with the physical landscape: "Just ahead the Anjanadri mountain (Tirupati) will please the eyes; it is correctly deemed by people to be the serpent adi-iesa himself, because it is

37 S. Krishnaswami Ayangar, A History of Tirupati, Vol. 1 (Madras: C. Sambaiya Panthulu, 1940-41) 2.

where Visnu resides, it bears the earth, has jewels inlaid in its peaks/ hoods (iiro), and is joined with small clouds that appear like skin just cast 0 f f . 1 ' ~This ~ description plays upon the idea that the places of Visnu's residence are his dependents (iesas), even as Tirupati more specifically appears like the divine serpent &:a on which he rests.3g As with all the other divine places, the overall description of the site is followed by a description of the icon located therein in 1.22: "You may also worship the destroyer of Madhu there, just as crowds of people worship him, both gods descended on earth and mortals climbing up, their mutual inequality erased by the common preponderance of purity (sattva)." The image of effacing mutual differences references a specific point in ~rivaisnavasoteriology regarding the effaced hierarchy (t~ratamya)between beings during servitude, regardless of differences in caste or gender subject to karma. Parakala sees this reference as a case of collapsed expression (samdsokti), since servitude is not the contextual meaning. So in this single verse, Vedanta De6ika ties associations of the divine place Tirupati with the serpent ~ e s to a allusions to a specifically ~rivaisnavaconception of liberation.

38 HS 1.21. v i ~ o vr m d avanivahanad baddharatnais sirobhih sesas s & & Zayam d iti janais samyag unniyamdnahl abjair yukto laghubhir acirdn muktanirmokakalpaih agrebhavi tad anu nayane ra fijayan afijanadrihll 39 The verse itself is patterned through the trope of anumanalaihkara,where reasons are provided for a fanciful description, which is in turn supplemented by an u t p r e k ~with the clouds as cast-off skin.

The description of Kafici is much more detailed than that of Tirupati, covering eleven whole verses (1.25-1.35); Vedanta Degika places special importance on his own birthplace near Kafici (at Viiakkoli Koil). Kafici contains several separate divine places, three of which Vedanta Desika specifically refers to: Tirukacci, Tiruttanka (or Toopal), and Tiruvega. While Tirukacci refers to Kaficipuram proper and Tiruvega to the Veghavati river, Vedanta Desika devotes the most space to a description of the Hastisaila shrine at Tiruttanka. Here as well, a close parallel to the Meghadtita is involved given Kiilidiisa's eleven verses on his own hometown, Avanti. These verses dedicated to Kafici contain the only references to Tondainadu. As with Tirupati, Vedanta DeSika first introduces Kafici with a verse full of poetic embellishment, an elaborate desa playing on the name "Kaficittas a word for a particular kind of ornament: "You will see the waist ornament of the earth, inlaid with many jewels and forever making beautiful musical sounds. The lord of Hastisaila, Satyakiima, being beyond the worldly opposites always casts his eyes there.'t40While this comparison with this belt around the hips replete with bells and jewels presents an alluring image, the suggestion of an amorous relationship between the incarnation of Visnu, Satyakama, and the earth is a form of apparent opposition (virodh~bhasd),given the fact that the lord "is 40

ndndratnair upacitagundm nityasangltanddam bhiimair draksyasy ucitavibhavam b h a a n a m tatra kd ticiml yasydm nityam Yahitanayano hastisailddhivdsi dvandvdtitas s khalu p u q o drsyate satyakdmahll ( H S 1.26)

beyond worldly opposites" and not subject to desire. But this construal is a highly suggestive way of characterizing the nexus between the divine and mundane believed to be at work in the divine places. Next, Vedanta De6ika outlines a description of the icon and the mythic history associated with Kaiici. As with Tirupati, in 1.27 Rama enjoins the swan 'to bow down with head bent in

devotion.^'^^ A legendary account augments the

illustriousness of the Hastigiri shrine in particular. In 1.25 we are told that the lord who is "the bridge for all worlds" (i.e. in crossing the ocean of transmigration) once took the form of a bridge to assuage the anger of the goddess Sarasvati towards her husband Brahmii, who had performed a sacrifice there with another woman, ~ a t ~ a v r a tA a .few ~ ~verses later in 1.33Vedanta De6ika includes a very specific image of the icon at Hastigiri, described as an "emerald-colored cloud" "streaked with the lightning of ~ a k ~ m iThroughout ."~~ this section and others, the commentators corroborate their readings with sections of the relevant miihiitmyas (such as the Hastigirimiihitmya). Although most of the miihiitmyas date back only a few centuries, the shared myths and geographic references attest to the careful manner in which Vedanta De6ika constructs the route.

41

HS 1.27. pranamya nagarim bhaktinamrena mUrdhnd

42

HS 1.25. sakalajagatam ekasetus

43

H S 1.33. marakatasilamecakamviksya megham, laksmividyul lalitavapusam

The intervening Cola region and Kaveri river before ~ r i r a n ~ ado m not warrant significant attention from Vedanta DeSika, and the swan is urged to 'cross over the intervening areas.''44While Vedanta Deiika describes the Kaveri river poetically, no other site is deemed worthy of attention between Kafici and ~rirangam,a region consisting of vast portions of the Colaniidu including Tafijavor, Kumbhakonam, and other important centers. The verses describing srirangam are especially significant for a number of reasons. Aside from its centrality to the srirangam community, the relationship between the Ranganatha icon and the Ramayana is the strongest and oldest material link between a ~rivaisnavainstitution and the epic narrative. The verses themselves, 1.38-1.46, cover two distinct divine places--Tiruveliarai and ~rirangam-as well as an extended description of the Kiiveri river as it passes through the area. The description begins with three full verses (1.38-1.41) dedicated to the relatively minor Tiruvellarai shrine, including a comparison of a from the nether regions (1.38) (as in the ~vetadrihill with the serpent ~ e s risen the section on Tirupati) and a desa confating the luster of the divine gaze with a lamp in the temple (1.39). The emphasis on this site may indicate a pragmatic consideration, since it is situated on a mountain like Tirupati and later Tirumaliruficoiai (1.49, on the Vrsabha mountain), which would presumably be

44

H S 1.36. janapadam atho madhyamam larighayitva

suitable for the swan to touch down upon. The environs depicted in 1.38 (the western part of the Cola country called Varsa), 1.40-1.41 (the forest and areca grove nearby), and 1.42-1.43 (describing in vivid detail the Kiiveri river and the crocodiles and pearls contain therein) correspond closely with descriptions in the irirangamah~tmya.These correspondences are so specific that Parakala even lines up specific phrases from the mah~tmyawith quarters of verses in the H a d a Sandeka (especially in 1.45). The institutional relationship referred to above concerns the famous family heirloom (kula-dhana) of the Ranganatha icon that Riima gave to Vibhisana as recompense for his assistance in ousting Riivana. The source for this connection is a verse in the last sarga of the Yuddha Kanda (6.131.88), where Riima presents gifts to Hanuman, Sugriva, Laksmana, and Vibhisana upon ascending the throne at Ayodhyii. Parakala's commentary deals with this issue at length, with particular emphasis on the divergence between the comments of Maheivaratirtha and Govindariija. While Govindargja states simply, "tradition holds that family heirloom means the family heirloom of the Iksviiku kings, i.e. the icon at ~ r i r a n ~ a mMaheivaratirtha ," insists that this family heirloom refers only to the kingship of ~ a n k i i Parakala .~~ uses several citations to dismiss Mahesvaratirtha's view and argues on grounds of logic that since Riima had

45

kuladhanam iksvdkukuladhanam firangavimdnam iti sampradayah

already granted kingship over Lank2 to reiterate this as a gift would be redundant. More than anywhere else, here in the section on ~ r i r a n ~ athe m possibility of anachronism between the contemporary pilgrimage sites and the action of the Rgmayana is directly addressed. Other than Tiruvellarai, the only material object described is the ancient Puskarini tank associated with the removal of the disease of the moon. And Vedanta Desika is careful to have Rama describe the ~ r i r a n ~ aicon m prospectively: "From thinking of that place arisen from Vaikuntha, containing the image of the lord within it like an emerald within a golden box, my mind races to the primordial lord of long eye-lashes, the support of the life of Laksmi, reclining on adi-sesa with his hand as a pillow."46 This is a vivid description of the famous reclining Visnu icon in ~rirangamas it will appear after installation. Here we get a glimpse of the final argument against anachronism: unlike the rest of the South, the divine places are deemed to be eternal sites of the lord's presence on earth.

46 H S 1.45 tire tasya viracitapadam sadhubhis sevyamanam Sraddhayogad vinamitatanus sesapitham bhajethW yasmin asmat kuladhanatayd saumya saketabhdjah sthdnam bhdvyam munibhir uditam Snmato rarigadh&nnahll HS 1.46 sattve divye svayam udayatas tasya dhdmnah prasangdt maiijusaydm marakatam iva bhrdjamdnam tadantah/ ceto dhdvaty upahitabhujam Sesabhoge Sayanam dirghapdrigam jaladhitanaya jivitam devam adyarnll

4.5. Conclusion

I want to conclude by placing the Hamsa Sandesa within the broader context of the theologization of the Ramayana, first by showing how the various aspects of theologization examined in the previous three chapters enhance our understanding of this poem. As discussed in chapter one, the shift in the reception of the Ramiiyana from k w a to theology also involved a shift from court to temple. In the Haws Sandesa, this institutional shift is mirrored by the constricting of epic space to the space of the ~rivaisnavadivine places. Such a mapping of space represents imaginatively the assimilation of the Ramayana to the sphere of the ~rivaisnavatemple. As discussed in chapter two, this institutional shift facilitated the incorporation of vernacular forms of temple discourse into Sanskrit. The allegory of the Hamsa Sandesa involves just such a process, as Vedanta Desika draws on the precedent of Manipravala allegorical reading to fashion a distinctive style of Sanskrit allegorical composition. And this transfer from an explicit form of theological discourse to an insidious form of Sanskrit composition (evident in the Hamsa Sandesa's esoterism) mirrors other practices such as connotative reading. As discussed in chapter three, when ~rivaisnavaintellectuals interpreted the Riimiiyana they also adapted and modified the categories of Sanskrit kdvya. The Harhsa Sandesa, as a Sanskrit messenger poem, possesses many of the distinctive features of high Sanskrit kdvya, including poetic figures, etc, as identified in Parakala's commentary. And

yet its allegorical form and explicit sectarian content make it a highly unusual kdvya . The Hamsa Sandesa represents a special case of the hermeneutic project of Ramayana commentary, the specific interpretive techniques by which ~rivaisnavaintellectuals sought to transform the conception of the Ramayana as k w a . We can view the poetic retelling of the Hamsa Sandesa itself as a form of interpretive practice. A. K. Ramanujan has written eloquently about the 'reflections' in South Asian folk practices of retellings and embedded narratives, where familiar stories are performed and modified so that each version involves subtle variations, and a masterwork like the Ramiyiina can be represented poetically hundreds of times without exhausting its capacity to generate new meaning4' Ramanujan speaks of the vertical relationship between works through a series of semiotic typologies, grouped under the rubrics of reflexivity (responsive, reflexive, and self-reflexive) and translation (iconic, indexical, and symbolic). These heuristic paradigms illustrate how a new work can invoke, transform, and comment upon previous works. In the first triad, whereas responsive texts redefine both text A and B, reflexive texts react inversely to text

A, similarly, in the second triad, even as iconic texts geometrically resemble their predecessors, indexical texts are focused on concrete historical contexts,

47 A. K. Ramanujan, "Where Mirrors are Windows: Towards an Anthology of Reflections," 1989.

and symbolic texts produce a counter-text, interpreting the antecedent. It is in this sense that a work based on a prior model can embody significant newness. The Hamsa Sandesa's evocation of the receptive history of the Riimiiyana-the Sundara Kiinda narrative and Sanskrit messenger poems based on this narrative --parallels the hermeneutic project of the Sanskrit commentaries on the Riimiiyana. What is distinctive about the Hamsa Sandesa is that it embodies in its form the image of the Riimiiyana that the commenators envision: a theologized Riima kdvya.

CONCLUSION

I have explored the specific interpretive techniques employed in the theologization of the Riimiiyana in the ~rivaisnavacommunity (1250-1600) in a variety of ways throughout this dissertation. In this brief conclusion, I summarize how each chapter addresses this issue and then move on to some remarks about the broader significance of the dissertation and avenues for future research.

5.1 Dissertation Summary

In the first chapter, I showed how the shift in the reception of the Rfimayana from k w a to smyti also entailed a shift in institutional locus from court to temple. Demarcating this shift was crucial to the analyses of the second and third chapters, for two reasons: 1) theologization itself, as I define it, involves the efforts of ~rivaisnavaintellectuals to transform this very receptive history of the Riimiiyana as k m a ; and 2 ) the specific interpretive techniques used in the theologization of the Riimayana involve the integration of practices of the

temple into Sanskrit. As evidence of the reception of the Riimiiyana as kdvya, I discussed the Riimiiyana's self-presentation as a k w a , references to the Riimiiyana as the primeval poem (ddi-kdvya) by Asvaghosa (second century), Kiilidiisa (fifth century), Bhavabhuti (eighth century), and Riijasekhara (tenth century), Sanskrit poetic retellings, and the centrality of the epic for Sanskrit aesthetics. As evidence for the changed conception of the Riimiiyana as srnfli, I cited passages from the most important ~rivaisnavaRiimiiyana commentary, Govindariija's BhQana . The Riimiiyana was treated as an authoritative text from the beginning of the second millennium in a variety of places in South (and Southeast) Asia, but nowhere in as sustained a manner as in the ~rivaisnavacommunity. As an example of importance of the institutional locus of the temple to this development, I examined Kulacekariilviir'simaginative affiliation of elements of the Riimiiyana narrative with the Govindariija icon at Cidambaram. Throughout this entire period, ~rivaisnavastreated stories from the Riimiiyana in generic Vaisnava temples, and it was in just such temples that the interpretive techniques used in Riimiiyana commentaries were developed. Only in the Vijayanagara period were the first monumental temples dedicated to Riima built. I argued that the role of ~rivaisnavasin the establishment of a royal Riima cult in fifteenth and sixteenth century Vijayanagara marked a project that

was parallel to the composition of Sanskrit Riimayana commentaries at the exact same time in a closely related place. In the second chapter, I examined one side of the ~rivaisnavahermeneutic project, the incorporation of vernacular temple forms into Sanskrit commentary. Each example involved some aspect of translation or redeployment. The first case, examined at some length, was the translation of Manipraviila rahasya treatments of Riimiiyana stories into Govindariija's Sanskrit commentary. Characterizing Maniparaviila as the "voice" of ~rivaisnavismas vernacular theology, I examined the relationship between Manipraviila writing and Manipraviila temple oratory, as well as specific Manipraviila practices such as the use of stories as parabolic exemplars and the techniques of performative substitution and improvisation. This dynamic of translation captures the historical trajectory of Riimayana commentary, with Manipraviila works produced from the twelfth through fourteenth centuries and Sanskrit works produced during the fifteenth and sixteenth centuries. As examples, I looked closely at Govindariija's commentary on the eighteen traditional meanings and Vibhisana's surrender, emphasizing transformations such as framing through Sanskrit scholastic categories, the heavy use of Sanskrit grammatical analysis, and the engendering of new techniques such as connotative reading. As a second case, I looked at the way praise-poems to

RZma stylize liturgical utterance (mantras) and genres of Tamil devotional poetry as Sanskrit kavya, In the third chapter, I examined the other side of the ~rivaisnava hermeneutic project, the application of modified Sanskrit aesthetic categories to the RZmZyana. These two sides of GovindarZjatscommentarial practice are closely related: just as vernacular forms from the temple were brought into Sanskrit, so too the categories of Sanskrit were made to resemble these vernacular forms. The Sanskrit aesthetic category examined in chapter three was double entendre (slesa), with three types of readings containing elements resembling slesa yet significantly different from slesa as defined in Sanskrit poetics, where the levels of meaning must almost invariably bear a tropic relationship. The first type involved readings of the verse traditionally considered to be the inaugural instance of Sanskrit poetry (ma nisada); commenting on this verse is tantamount to a commentary on the entire tradition that took the RamZyana to be a kiivya. The second and third types were related respectively to allegorical reading and performative substitution. fiesa-like reading differs in significant ways from other reading techniques in GovindarZjatscommentary, such as connotative reading and allegorical reading. Unlike connotative reading, slesa-like reading splits the text into literal and figurative levels. Whereas connotative reading conflates concepts such as prapatti with the source text, slesa-like reading keeps these separate. aesa-like

reading also differs from allegorical reading in that in slesa the secondary level of meaning develops from the sign itself, while in allegory its source is the first level of meaning. Finally, 1 considered why Govindaraja would choose to read in this fashion, in an idiom so close to a category of Sanskrit poetics and yet ultimately distinct from it. The answer I proposed was twofold. Govindaraja's practice of slesa-like reading is emblematic of his conception of the Riimiiyanii as a special kind of kdvya, a theologized k w a . In addition, the slesa-like readings allow Govindaraja to domesticate vernacular interpretive techniques such as allegorical reading and performative substitution into the idiom of Sanskrit aesthetics. In the fourth chapter, my close reading of Vedanta Desika's H a M a Sandesa was based on the analyses of the previous three chapters. Each feature identified in the poem corresponded to an aspect discussed in an earlier chapter. We can better understand the charting of the path of the swan through the ~rivaisnavadivine places, including Tirupati, Kafici, and ~ r i r a n ~ a m in ,light of the institutional shift from court to temple. The allegory of the Hamsa Sandesa, where the swan represents the dcdrya conveying a message from the lord to the soul, is an example of the integration of vernacular forms into Sanskrit. And the fact that Hamsa Sandesa represents a peculiar kind of kdvya, which closely conforms to generic norms of the Sanskrit messenger poem yet

stands apart in its allegorical form, is yet another case of the modification of Sanskrit categories.

5.2 Significance of Research

As outlined in the introduction, my approach to the ~rivaisnavaRfimfiyana corpus is based on an understanding of the act of commentarial writing as a historically concrete intervention in the world, and I would like to use this methodological perspective as a way to outline two categories of possible contributions of this dissertation: historical and hermeneutic. While I have characterized this intervention as the effort to transform the receptive history of the Riimiiyana as kivya, this material also touches upon a number of other related historical processes at work in late-medieval South Asia. First, the theologization of the Rilmiiyana is an example of the more general absorption of kdvya and Sanskrit aesthetics within religious orders from around the beginning of the second millennium. From the tenth century, we see a remarkably similar set of developments in the Kashmir Saiva world, including the formulation of a Saiva aesthetics (with Abhinavagupta in the tenth century) and ~ a i v mahdkdvya (with Kalhana and Mankha in the twelfth century). For a the Kashmir Saivas, aesthetics and metaphysics were mutually constitutive, with the drama functioning as a speculative paradigm for conceptualizing the

c o ~ m o s .Other ~ such examples include kavya-esque bhakti literature such as the Bhiigavata Puriina (eleventh century, Tamil region) and the Gitagovinda of Jayadeva (twelfth century, Bangla region). And the Gaudiya Vaisnavas also developed concepts such as bhakti-rasa in the aesthetics of Riipa Goswiimin

( 1 6 century). Further studies of such examples of theologization may enable us to historicize the relationship between religion and aesthetics in South Asia. Just as I have focused here on the shift from court to temple, it may be useful to consider the institutional locus of these other examples. While the Kashmir ~ a i v a provide s a striking parallel in terms of the disintegration of the Kashmir empire and the drying up of patronage for Sanskrit literati, Jayadeva's Gitagovinda is a possible counter-example since its author, Jayadeva, was the court poet of the Sena king, Laksmanasena. Comparative study of these three cases-the

hermeneutic project of Riimiiyana commentary in the South, ~ a i v a

aesthetics in Kashmir, and the Gitagovinda as court poetry sung in temples in the East-may

indicate the degree to which these reflect related socio-cultural

processes.

1

Guy Leavitt is currently completing a dissertation on this topic, The Poetics of ~ a i v i s min Kashmir, 840-1200, University of Chicago. On the use of the phrase, "speculative paradigm" with respect to Abhinavagupta, see Edwin Gerow, "Abhinavagupta's Aesthetics as a Speculative Paridigm," Journal of the American Oriental Society (April, 1994) 114: 2, 186-209.

Second, the Rarnayana commentaries allow us to examine more localized historical transformations within the ~rivaisnavacommunity from the thirteenth through sixteenth centuries, a period that witnessed the use of a new linguistic medium, Manipravala, the formulation of the key soteriological concept for ~rivaisnavas,~ra~atti, and the split between the Tengalai and Vadagalai schools. This dissertation may help reconfigure our understanding of this last development in particular. While the lateness of the institutionalization of the Tengalail Vadagalai split is well known, the absence of rival positions even in Ramayana commentaries produced in the fifteenth and sixteenth centuries indicates that doctrinal positions were not reified at this point either. This appears all the more surprising given thatprapatti, the focal point of Tengalail Vadagalai debates, provides the rubric for interpreting the epic narrative. The foundational figures for the Tengalai and Vadagalai schools are, respectively, Pillai Lokacarya (b. 1217) and Vedanta Desika (b. 1268), but no Tengalai or Vadagalai institutions date to their lifetimes. As Arjun Appadurai has carefully shown, such institutions emerged as a result of the gifting of temples by warrior leaders associated with the Vijayanagara empire though sectarian mediaries from 1350 to 1700.~The key moments in the institutionalization of the Tengalai school were the ~riranganarayanaJiyar Appadurai, Worship and Conflict Under Colonial Rule, 198 1.

Atinam at ~ r i r a n ~ ainmthe early fourteenth century, the collaboration between the generals Kopanna and Siiluva Kunta and members of the Uttamaniimbi family, and the activities of Manavila Mimuni in the fourteenth century. The institutions of the Vadagalai school date somewhat later, from 1500, with the patronage of Tirupati by the Vijayanagara king Siiluva Narasimha through his mediary Kandiidai Riimiinuja Aiyangiir, the Ahobila Matha in Kurnool District, Andhra Pradesh, the Brahmatantra Parakiila Sviimi Matha in Mysore, and the influence of the Tsticharya raja-gurus at Vijayanagara. From the eighteenth century, the Tengalail Vadagalai split became more rigid through British legislative intervention, resulting in competition for the control of temples. The picture that emerges from this genealogy of institutional development is that the Tengalais and Vadagalais arose first as competing doctrinal schools and by the early sixteenth century evolved into distinct social communities. The schools are demarcated through conflicting positions on the status of prapatti.3 Vadagalais viewprapatti as a means (updya), but Tengalais do not; Vadagalais permit repeated performance ofprapatti as a form of penance (prayaicitti), but for Tengalaisprapatti must be performed only once; Vadagalais argue that the twice-born must continue to adhere to the norms proper to class and life-station (varqairama-dharma), but Tengalais believe thatprapatti For a summary of these debates, see Patricia Mumme, The ~ r i v a i ~ a vTheological a Dispute: Vedanta Desika and Manavdlamhuni (Chennai: New Era Publications, 1988).

supersedes all other ethical norms. The Riimiiyiina commentators never touch upon such issues even though they adduce detailed technical analysis ofprapatti to the narrative. Typical is Govindariija's application of the six components (angas) ofprapatti to Vibhisana's surrender without any mention of the controversy surrounding the status of these components, with Tengalais viewing prapatti as a unified act, and Vadagalais as a composite act. Govindariija also quotes equally from proto-Tengalai and proto-Vadagalai sources. What are we to make of this apparent anomaly? A possible explanations is the liminal institutional location of the Sanskrit Riimiiyana commentaries. As mentioned in the introduction, the heads of the Ahobila Matha, where the commentary of Govindariija and several others were composed, espoused only Vadagalai positions from the time of the seventh head, Ahobila Jiyar (at least several decades after the period of Govindaraja's own teacher) according to the seventeenth-century Trimsatprasnottaram. Further evidence of the liminal status of the Ahobila Matha is its acceptance of the @u commentary of Vadakkutiruviti Pillai, which is rejected by other Vadagalais on account of its proto-Tengalai positions. But institutional location, in and of itself, is an insufficient explanation for the absence of explicit Vadagalai or Tengalai positions. Not all the Sanskrit Riimiiyana commentaries were produced at Ahobila, and none articulate distinctive Tengalai or Vadagalai positions. We must therefore conclude either that there were specific contexts where such

positions were ignored or that the ideological, as well as institutional, split was much later than earlier imagined. The historical dimension discussed directly in this dissertation is perhaps of more timely significance than the transregional process of theologization and the Tengalai-Vadagalai split: the possible relationship between the ~rivaisnava Riimiiyana commentaries and the establishment of Riima temples at Vijayanagara. The archeology of Rama temples is of course of contemporary political relevance, given that the Riimjanmabhiimi (birthplace of Riima) has been the most divisive symbolic issue for the mobilization of support for Hindu nationalists since the late 1980s. One avenue for future research is a comparison of these two projects of fifteenth and sixteenth century Vijayanagara. Both represent ambitious expansions of ~rivaisnavapractice: just as the commentaries mark the translation from Manipravda to Sanskrit, so also the Riima temples turn the treatment of the Riimiiyana in Vaisnava temples into veneration of the figure of Rama. The Riima-cult at Vijayanagara itself merits special attention. Vijayanagara provided a crucial template for early modern Rama traditions throughout South Asia. It was only at Vijayanagara that Riima worship became culturally significant; never before had Riima been the patron deity for an empire. Examination of epigraphic, archeological, and textual evidence may shed more

light on the nature of the participation between ~rivaisnavasectarian agents and Vijayanagara regents in the establishment of a royal Riima cult. Specifically,we can ask, what was the relationship was between the ~rivaisnavaconception of the Riimayana asprapatti-sastra and the identification between Vijayanagara kings and Riima? In addition to turning our attention to the relationship between the ~rivaisnavaRiimiiyana corpus and broader historical processes, the idea of commentary as an intervention in the world can highlight the ways this dissertation may also contribute more broadly to our understanding of South Asian hermeneutics. The study of Sanskrit commentary can point us to the differences between the way we read and interpret texts today and the way intellectuals did so in the past. In part because in Sanskrit intellectual history the predominant mechanism for innovation was (paradoxically) through engagement with antecedent sources, a vast archive of scholia, doxographies, and citations survives as concrete evidence of precolonial reading practices. Yet Indologists typically treat such a secondary source as wholly derivative or at best the record of an originary audience and thereby a reflection of the original, rather than itself the object of inquiry. The hermeneutic engagement with the past examined in this dissertation includes various forms of figural reading. It may be of interest to compare these forms of figural reading with those found in other cultures and traditions, as

well as with other Sanskrit hermeneutic traditions such as Mimiimsii. Although the ad hoc quality of the Riimiiyana commentaries differs from the explicit interpretive principles of Mimiimsii philosophical hermeneutics, there are many similarities between these commentaries and the Mimiimsii approach to the Vedic corpus; both involve the practice of reading parts through an apriori understanding of the whole and the projection of metaphysics onto an antecedent source through the application of implicit or explicit interpretive principles. Comparison of these and other such examples may allow us to lay out some general characteristics of South Asian hermeneutic practice. This dissertation includes for the first time technical analysis of a number of interpretive techniques (such as slesa-like reading and connotative reading). Since most of these have no obvious counterparts in Western rhetoric, additional study of such practices may nuance contemporary theories of allegory and figural language. This analysis may also be brought to bear on similar examples of reading practices not theorized in Sanskrit aesthetics, such as Nilakantha's seventeenth-century mantrarahasya works.4 One measure of the success of the ~rivaisnavahermeneutic project is the treatment of incidents such as Vibhisana's surrender in other religious orders See Christopher Minkowski, "Nilakantha Caturdhara's Mantrakdsikhaw Journal of the American Oriental Society 122:2 (2002) 329-344; Christopher Minkowski, "Meanings Numerous and Numerical: Nilakantha and Magic Squares in the Rgveda, FestchriftElizvenkova, forthcoming; and Christopher Minkowshi, "Nilakantha Carurdhara and the Genre of Mantrarahasyapariksfi" in Proceedings of the Second International Vedic Workship, ed. Y. Ikari, Kyoto, forthcoming.

familiar with the ~rivaisnavacommentaries, and I intend to look at a number of such sources as I continue to explore this topic. I referred briefly in the third chapter to the incorporation of Govindariija's slesa-like reading by Riimacandra Buddhendra and Nagesa Bhatta; similar comparisons could be made for a number of other passages. Although by far the predominant site for the theologization of the Riimiiyana was the ~rivaisnavacommunity, there were also other examples in the South including Appaya Diksita's Ramdyanatdtpaiyanimya, which posits a ~ a i v areading of the Riimiiyana based on suggestion, and Madhva's citation of a text referred to by no one else, the Mzilardmiyana. The theologization of the Riimiiyana is a vast and important topic which merits much more investigation, and these are only a few of the directions in which I hope to take this project.

BIBLIOGRAPHY

Inscriptions

Annual Report of South Indian Epigraphy. Chennai/ Caclutta: Government Press, 1887-. Epigraphia Camatica. Bangalore: Mysore Government Central Press, 1886-. Epigraphia Indica. Calcutta/ New Delhi: Archaeological Survey of India, 1892-. Vijayanagar:Progress of Research. Ed. D. V. Devaraj and C. S. Patil. Mysore: Directorate of Archaeology and Museums, 1983-84. South Indian Inscriptions. Ed. E. Hultzsch and others. Chennai: Government Press, 1890-. Tirupati-TirumalaiEpigraphic Series. Chennai: Tirupati Sri Mahant's Press and Tirumalai-Tirupati Devasthanams Press, 1931-38. Primary Sources

Aflyirappati Guruparampdprabhdva of Pinpalakiya Perumdl. Ed. S. Krishnaswami Ayangar. Tirucci: Puttur Agraharam, 1975. Atri Samhita. Ed. P. Raghunatha Chakravarti Bhattacharya and M. Ramakrishna Kavi. Sri Venkatesvara Oriental Series, No. 6. Bhagavad Gitd: With the B h w a of Rdmdnuja and the Tdtparyacandrikd of Veddnta Desika, Ed. M. Ragachariar. Srirangam: Vani Vilas Press, 1907. Bhdsan@akacakrarn,Plays Ascribed to Bhdsa. Ed. C. R. Devadhar. New Delhi: Motilal Banarsidas, 1962.

Bhdsandtakacakram, Plays Ascribed to Bhdsa. Ed. C. R. Devadhar. New Delhi: Motilal Banarsidas, 1962. Campurdmdyana of Bhoja with the Commentary of Rdmacandrabudhendra. Ed. Narayan Rama Acarya. Mumbai: Nirnayasagar Press, 1956. Caturvargacintdmani of Hemddri. Varanasi: Kashi Sanskrit Granthamala 235, 1985. Dhvanydloka of Anandavardhana. With the Locana commentary of Abhinavagupta and the Bdlapriya subcommentary of Riimasiiraka. Ed. Pt. Pattiibhiriima ~iistri.Kashmir Sanskrit Series 135. Varanasi: Chowkamba Sanskrit Office, 1940. Hamsa Sandesa of Veddnta Desika. Unpublished Mss, Mysore Oriental Library, E 26786. Hamsa Sandesa of Veddnta Desika. Unpublished Mss, Mysore Oriental Library, 26787. Hamsa Sandesa of Veddnta Desika. With the commentary of Krishna Brahmatantra Parakala Swami. Mysore: Mysore Oriental Research Institute, 1917. Hamsa Sandesa of Veddnta Desika. With the commentary of Uttamur Viraraghavacharya. Chennai: Ubhaya Vedanta Granthamala, 1973. Kuvalaydnanda ofAppaya Diksita. Ed. Vasudeva Sharma. Mumbai: Tukaram Javaji, 1903. Meghadiita of Kdliddsa. Critically edited by Sushi1 Kumar De. New Delhi: Sahitya Academy, 1957. Ndtyakistra wih the Commentary ofAbhinavagupta. Ed. Manavalli Ramakrishna Kavi. Baroda: Central Library, 1926-64. Pddukdsahasra of Veddnta Desika. Ed. Pandit Kedaranatha and Wasudeva Laksman Sastri Panashikar. Varanasi: Chaukhambha, 1984. Perumdl Tirumoli of Kulacekardlvdr. Kanci: Kadaikkumidam Granthamala Office, 1966.

Altekar, G. S. Studies on VdlmikilsRdmdyana. Pune: Bhandarkar Oriental Research Institute, 1987. Appadurai, Arjun. Worship and Conflict Under Colonial Rule: A South Indian Case. Cambridge, Cambridge University Press, 1981. Bauman, Richard. VerbalArt as Performance. Rowley: Newbury House Publishers, 1977. Bhatt, G. H. "Riirnayana Commentaries," Journal of the Oriental Institute, Baroda 14: 350-361. Bourdieu, Pierre. The Field of Cultural Production: Essays on Art and Literature (New York: Columbia University Press, 1993). Brockington, J. L. Righteous Rdma: The Evolution of an Epic. New York: Oxford University Press, 1984. Bronner, Yigal. Poetry at Its Extreme: The Theory and Practice of Bitextual Poetry lev) in South Asia. Unpublished Dissertation, University of Chicago, 1999. Carman, John. The Theology of Rdmanuja (New Haven: Yale University Press, 1974).

Census of India 2001: Temples of Tamil Nadu: Kancheepuram District. Delhi, The Controller of Publications, 2003. Champakalaksmi, R. Elements of Vaimava Iconography in the Tamil Country. New Delhi: Orient Longman, 1981.

. "The Study of Settlement Patterns in the Chola Period: Some Perspectives." In Man and Environment 14,14 (1989) 91-102. Cone, Margaret and Gombrich, Richard F, Tr. The Perfect Generosity of Prince Vessantara:A Buddhist Epic. Oxford: Clarendon, 1977. Cousins, Henry. List ofAntiquarian Remains in the Central Provinces and Berdr Calcutta: Archaeological Survey of India, 1897. Coward, Harld G. and Raja, K. Kunjunni. Encyclopedia of Indian Philosophers,

Volume V: The Philosophy of the Grammarians. Ed. Karl Potter. Delhi: Motilal Banarsidas, 1990. Dallipiccola, A. L., Fritz, John, Michell, George, and Rajasekhara, S. The Ramachandra Temple at Vijayanagara. New Delhi: Manohar Publications and American Institute of Indian Studies, 1991. Doniger, Wendy. Splitting the Difference: Gender and Myth in Ancient Greece and India. Chicago: University of Chicago Press, 1999. Eco, Umberto. The Limits of Interpretation. Bloomington: Indiana University Press, 1990.

.

The Role of the Reader: Explorations in the Semiotics of Texts. Bloomington: Indiana University Press, 1984. Fergusen, Charles A. "Dialect, Register, and Genre: Working Assumptions About Conventionalization." In Sociolinguistic Perspectives on Register. Ed. Douglas Biber and Edward Finegan. New York: Oxford University Press, 1994, 15-30. Fritz, John M., Michel, George, and Rao, M. S. Nagaraja. Where Kings and Gods Meet: The Royal Centre at Vijayanagara,India. Tuscan, Ariz., University of Arizona Press, 1984. Gerow, Edwin. "Abhinavagupta'sAesthetics as a Speculative Paridigm." Journal of the American Oriental Society (April, 1994) 114: 2, 186-209. Goldman, Robert, "Translating Texts Translating Texts: Issues in the Translation of Popular Texts with Multiple Commentaries." In Translation East and West, Ed. Cornelia N. Moore, Vol. 5. Honolulu: East-West Center and University of Hawaii, 1992. Gopalan, L. V. fivaisnava Divya Desams. Chennai: Sri Visistiidvaita Praciirini Sabhii, 1972. Gros, Fracois and Nagaswamy, R. Uttaramerur:legendes, histoire, monuments, Pondichery: Institut francais d'indologie, 1970. Hardy, Friedhelm. Viraha Bhakti: The Early History of K w a Devotion in South India. New Delhi: Oxford University Press, 1983.

Hazra, R. C. Studies in the Puranic Records on Hindu Rites and Customs. (Dacca: University of Dacca, 1940). Hopkins, Steven. "Lovers, Messengers and Beloved Landscapes: Sandesa Kdvya in Comparative Perspective." Forthcoming. Inden, Ronald. "Kings and Omens" in Purity and Auspiciousness in Indian Society. Ed. J. B. Carman and Frederique Apffel Marglin. Leiden: E. J. Brill, 1985. 30-40.

. Querying the Medieval: Texts and the History of Practices in South Asia. Oxford: Oxford University Press, 2000. Irvine, Martin. The Making of Textual Culture: 'Grammatica' and Literary Theory, 350-1100. Cambridge: Cambridge University Press, 1994. Jacobi, Hermann. Das Ramayana: Geschichte und Inhalt. Bonn, F. Cohen, 1893. Jauss, Hans Robert. Towards an Aesthetics of Reception. Minneapolis: University of Minnesota Press, 1972. Johnston, E. H., Ed. The Buddhacarita; or, Acts of the Buddha. Calcutta: Baptist Mission Press, 1935-1936. Kale, M. R. A Higher Sanskrit Grammar. Delhi: Motilal Banarsidas, 1972. Kramrisch, Stella. The Hindu Temple. Calcutta, University of Calcutta, 1946. Krishnamachariar, M. History of the Classical Sanskrit Literature. Madras: Vaijayanti Press, 1906. Lefebre, Rosalind. The Riimiiyana of Viilmiki:An Epic ofAncient India, Volume IV:Kiskindhiikwa. Ed. Robert P. Goldman. Princeton: Princeton University Press, 1994. Ludden, David. "Specters of Agrarian Territory in Southern India. The Indian Economic and Social History Review 39,2 & 3 (2002) 233-257. Major, R. H. India in the Fifteenth Century: A Collection of Narrative Voyages to India. London: Hakluyt Society, No. 22, 1857.

Michell, George. The Hindu Temple:An Introduction to Its Meanings and Forms. New York, Harpar and Row, 1977. Minkowski, Christopher. "Meanings Numerous and Numerical: Nilakantha and Magic Squares in the Rgveda. Festchrift Elizvenkova. Forthcoming. . "Nilakantha Carurdhara and the Genre of Mantrarahasyapariksii." In Proceedings of the Second International Vedic Workship. Ed. Y. Ikari, Kyoto. Forthcoming.

. "Nilakantha Caturdhara's Mantrakasikha&a." Journal of the American Oriental Society 122:2 (2002) 329-344. Monius, Anne. "The Many Lives of Dandin: The Kdvyddarsa in Sanskrit and Tamil." International Journal of Hindu Studies 4:1, April 2000, 1-37. Mumme, Patricia. "Riimiiyana Exegesis in Tengalai ~rivaisnavism."In Many Ramdyanas: The Diversity of a Narrative Tradition in South India. Berkeley: University of California Press, 1991, 203-215.

.

The Sfivaisnava Theological Dispute: Vedanta Desika and Manavalamdmuni. Chennai: New Era Publications, 1988. Nagaswamy, R. "Sri Riimiiyana in Tamilnadu in Art, Thought and Literature." In The Rdmayana Tradition in Asia, Ed. V. Achaean. New Delhi: Sahitya Academy. Narayanan, Vasudha. "The Riimiiyana in the Theology and Experience of the ~rivaisnavaCommunity" Journal of V a i ~ a v aStudies 2,4 (Fall 1994) 55-89.

.

The Vernacular Veda: Revelation, Recetation, and Ritual. Columbia, South Carolina: University of South Carolina Press, 1994. Nayar, Nancy Ann. Poetry As Theology: The &vaisnava Stotra in the Age of Ramdnuja. Weisbaden: Harrassowitz, 1992. Padigar, S. V. The Cult of Vishnu in Karnataka (Unpublished Ph.D. diss). Pollock, Sheldon. "The Cosmopolitan Vernacular." Journal ofAsian Studies 57, 1 (February, 1988): 7-37.

. "The Divine King of the Riimiiyana" in The Rdmdyana of Vdlmiki:An Epic of Ancient India, Ed. Robert Goldman, Vol. 3. Princeton: Princeton University Press, 1991. 15-55. . "Riimiiyana and Political Imagination in India." Journal of Asian Studies 52, 2 (May 1993): 261-297. . "Sanskrit Literary Culture From the Inside Out." In Literary Cultures in History: Reconstructions From South Asia. Ed. Sheldon Pollock. Berkeley: University of California Press, 2003) 39-130. Raghavan, V. "Udiili's Commentary on the Riimiiyana: The Date and Identification of the Author and the Discovery of His Commentary." In Annals of Oriental Research 1940-42: 1-8. Raja, K. K. "Malayalam and Manipraviilam."Journal of Oriental Research, Madras 1994-1997: 64-67,99-109. Ramanujan, A. K. "From Classicism to Bhakti." In The Collected Essays ofA. K Ramanuan. Ed. Vinay Dharwadker. Oxford: Oxford University, 1999.

.

"Where Mirrors are Windows: Towards an Anthology of Reflections," History of Religions 28.3 (1989) 187-216. Rangachari, V. "Historical Evolution of Sri Vaishnavism in South India." In The Cultural Heritage of India. Ed. H . Bhattacharya. 2nded. Calcutta: ~amakrishnaMission Institute of Culture, 1953-58, vol. 4 (1956): 163-85.

, "The History of ~ rVaishnavism: i From the Death of Sri Vedanta Desika to the Present Day," Quarterly Journal of the Mythic Society 7: 2 (January 1917): 106-18 and 7:3 (April 1917): 197-209.

. "The Successors of Riimiinujii and the Growth of Sectarianism among the ~ri-~aishnavas." Journal of the Bombay Branch of the Royal Asiatic Society 24 (1914-15): 277-312. Rao, T. A. Gopinatha. Elements of Hindu Iconography. Chennai: Law Printing House, 1914-1916. Saletore, B. A. Social and Political Life in the Vijayanagara Empire. Madras: B. C. Paul and Co., 1934.

. "Vaisnavism in Vijayanagara." In D. R. Bhandarkar Volume. Ed. B. C. Law. Caclutta: Indian Research Institute, 1940. 193-195. Sastri, K. A. Nilakanta. A History of South India From Prehistoric Times to the Fall of Vijayanagar. Madras, Oxford University Press, 1976. Sastri, Sayavat. Essays on Indology. New Delhi: Meharchand Lachhmandass, 1963. Sewell, Robert. A Forgotten Empire: Vijayanagara. London: George Allen and Unwin, 1924. Shastri, P. P. S. "Riimiiyana Commentaries," Annals of the Bhandarkar Oriental Research Institute 1942: 410-417. Shastry, Sadhu Subrahmanya. Tirumalai Tirupati DevasthZnams Inscriptions: Report on the Inscriptions of the Devasthdnam Collection with Illustrations. Tirupati: TTD Religious Pubications Series, 1930.

The Smith Agama Collection: Sanskrit Books and Manuscripts Relating to Pdncardtra Studies. Ed. Agehananda Bharati. Syracuse, NY: Maxwell School of Citizenship and Public Affairs, Syracuse University, 1978. Smith, David. Ratndkara's Haravijaya: An Introduction to the Sanskrit Court Epic. New York: Oxford University Press, 1985. Sonesson, Goren. Pictoral Concepts: Inquiries into the Semiotic Heritage and Its Relevance for the Analysis of the Visual World. Lund: Lund University Press, 1989. Stein, Burton, Ed. Special Number on South Indian Temples, Ed. Burton Stein, Indian Economic and Social History Review 14, 1 (Jan-March 1977). Venkatachari, K. K. A. "Observations," Journal of Oriental Research, Madras 1994-1997: 94-97,95. Verghese, Anila. Religious Traditions at Vijayanagara,as Revealed Through its Monuments. New Delhi: American Institute of Indian Studies, 1995. Viraraghvacharya, T. History of Tirupati: The Thiruvengadam Temple. Tirupati: Tirumala-Tirupati Devasthanams, 1977-.

Young, Katherine Young. Beloved Places. Unpublished diss. McGill University, 1980.