ENGYCLOPAEDIAOF

BHARATIYADELHI

VIDYA

(rNDrA)

PRAKASHANVARANASI

I]IDITIN WISDOMPROF. SATYA VRAT SHASTRI FELICITATION VOLUME

Edltor

Prof. RAMKARAN SHARMA

[TfF BEST COMPLIMENTT3ROM THE PUBLISHER AIIO

AUTHOA

2.ao{

RatnasrijffanaSheldon Pollock

Measured by the crudest quantitative standards-miles travelled, size of readership,

kinds of language-traditions influenced, numbers of translations and adaptations and

bonowings-Dandin's seventh-century Kivyadaria can safely be adjudged the most important

work on literary theory and practice in Asian history and, in world history a close second

to Aristotle's Poetics. Its impact on the literary histories of southern India, in particular

Kannada, Sinhala, and Tamil in the period 800-1200, is common knowledge among scholars,so too its appropriation into Pali (in the Subodhdlaikan) during the later centuries of thissame epoch, when Theravadin literati undertook a dramatic reordering of their aesthetic

objectives according to Sanskit principles. Less well-known is the influence the KaYydarsa

exircised on Chinese in the formation ofRecent Style Poetry in the high T'ang' and onTibetanafter the remarkable educational reforms initiated by Sakya Pandita (1182-1251).' In view

of these facts, any text pertaining to the history and interpretation of Dandin's work willhold cosiderable scholarly interest. Foremost among such texts, ofcourse, are commentarieson the Karyadania iteself.

It is sobering to realize, however, that the number of published commentaries on theKavyddarla stands in almost directly inverse proportion to the text's importance. only fourhad found their way into print up to 1957 (ofthe two dozen oI more pre-colonial comrnentariesthat we know of). Three of these in fact were published together, in a now rare edition, by

D.T. Tatacharya in 1936, and largely ignored since z The editor believed that nothing wasknown about the authors of these commentaries except their names (and for the third, noteven that). This may not in fact be the case.

The first of the commentaries is printed in Tatacharya's edition in the Srutanupllinlcomposed by oneVadijanghaladeva (or Vadighankala, orVadighar'rghala).' Neither the Printedversion of this commentary nor any of its additional manuscripts tell us anything about theidentity of the author. In 1921, however, a tenth-century Ganga grant was published thatmay have bearing on this question. The record, which is dated Saka 884 (A.D.963), describesthe gift of an agraham, ̂s drutagurudaksi7! or teacher's fee, to one Vadighanghala [sic]Bhatla, and contains a long prasasti describing the scholar. Included in the encomium is th€following list of accomplishments:

ninvadyasahityavidyavyakhyananipupa -.....

sak arajavidyapntipadanapratibuddhabodhapmbodhitava|Iabharaja...hdatv ay at i su gh at a m antr akr amopade dd nu s th an av ai t krt akh i I a di gai gan a...kfsrrarajadeva-

o J /

That is, Vadighanghala was an "expert in the exegesis-of.th€. science of literature" as

well as a "political theorist who sh;pJlit"ttttnliog oitrtt^Y:tl*: king' Ktsnaraja' who'

y"f#lILH"];;;;;;;j; x';iffi "'T"$$T:[,:ii"j:: [:f; T:'.p#:learlier in the Srant ona *no--t'o*ntl "t-iiit t'"L"r king Vadighanghala's acnral patron'

d"i'e" ilrr"tifib". is Krstra lll Reslrakola (r' 939-67)'

It seems to me probable that this scholar is to be identifred with the author of the

s-,,',j,p,ri"i .'i, ;;; * :'.,.:fi,*1,*:,i*i:*",*mi"'T",;:H: ;iT 'Ii i:$s'.t",fir:l,'ffi l:ilf:Tl#'li;:#;;i;;'.. "''*i.n""t uur'iiatit among debaters'-that he acquired bv his mastery;il;;ilGi

i r:1" ftY:.d1t: commentarv itself bv

:"n;;r,;;;;;t lign of thrs retlgious uf itiati;n' and actualluv.beg,;;";y.lrf,;;:[il^i:

;""'r'".*,iunJ;"','; "l "lt.n* and Murari But as the prasast

was a recent convert to tn" 'u"nltiit'g-ffi"t is Jet"tiu"':-"t an orthodox Brahmin-

iust like rnany other tenth-century l"i"i'ti-f"t""t" T:::-":::1" the sreat poer Pampa'

a contemporary ot vadignangnitas (his Vikamarjunavijaya was iomposed around 950)

Such men often retained "gnt o':it"i) Ct*manicai heritaSo 'a Jeature that appears to nave

become institutionalized arnong the Jaina Brahmaqa community of Kamataka' The sumame

sh"n;-i;'ih" g.""t may be taken as evidence of this'

AsforthesecondcommentarypublishedbyTatacharya,thatofTarulavacaspatl,wenow know that this *u, utto" tt?Jin"ti"t'"*n"ttl " ttt noysala court in the mid-thirteenth

century. For, one KeSava shaffi;k;-ii"ntifies himself as the^son of Tartulavacaspati and

lutor to King vira Ramanatha l''-tiSl -l Cl' in the colophon of his own commentary on tne

Kavyadaria.n

ln 1957, Anantalal Thakur and Upendra Jha placed students'of Dandin under a hearry

debt when they edited ri"* " ti;g;";;;;'"'ip[ tnt- '"tuttuble comentary of "Master

Sriiffrna from Sinhala' (acary#'f'"iiiit*titr tiihatajanmanab)'7 The editors did-a

maenificent job in making Jt;';ilf; ;ften imperfect manuscript (though areas tor

improvement rerrain), una r'uue p'Jui-aei un infottea intt-oduction to the questions concemtng

the author. Proceeding "" tht b;;;';;;; r'nui u"tttt or trtt commentary 1"9'h't *t"llll;

ffL#:$il'JJfi ,"ffi:;:'"'J:ffJ ;*:ffi:i:ffi':ii:il{: ffffifi""fff"lxT::fr'f"L1;fr": but they havi nothing more of substance to suggesl'

The work dates itself to trre Z3rd year ofrhe reign 9l l1'J:"t" Thakur and Jha uke

this to be A.D. 93l. tttougt' tlt" ""koning of the successiomf Pila rulers is disputed' and

a truer dare may be g55.. RatnaSri iefers to his patron nol just by name. srlmattuiganarddhipa'

but also as sanabhvunnatanii'"nt"'"t'" rorjre{111^ment of the universarrv

ascendant RasFakutas") tp zaii iftit it hardiy the manner in which to describe a Ra$ga'

kura ramiry resident at cuv' ul'l ;il;t;;;: ;'il :l :::':lli"lf"lX',1ffi1il*il:;i.ii"'"0, ina this is to sav *fiT&:';h,�;:t'l1i;"i"ll irT,i;;i, ",ia'""iLv a rererercecorroborating the existence of

638

to a king of the imperial Rasfiak0tas of Kamataka, and another document, of which the

editors were unaware, makes this identification close to certain.

An inscription found at Bodh Gaya, "palaeographically assignable to circa 9th century,"and published more than a century ago was composed by the very man who wrote the

Katyadaria commentary, "Pandit Ratnasrijiana, the Buddhist mendicant ofthe island Sinhala"

lsiihaladvtpajanmana panditantnairrjanafread: jfianal bhrl1rwa,line l8).r0 The inscriptionrecords the dedication of a repository for incense (? gandhaka$) on the part of the author'spatron-named, again Tunga-in the fifteenth year of his reign.

According to the inscription, this Tunga was second in descent from some one namedNanda (or perhaps Nanna). A number of epithets are given to Nanda in the inscription, oneof which is particularly revealing: "Mattipuradurgadhavala," "Him of Pure White [Fame] ofthe fort of Manipura" (line 4). It is far likelier in terms of phonology that Manipura refersto Manyapura-that is, ManyakheF (Malkhdd), the Ra$FakDla capital in what is today'seastern Karnataka rather than Mainpuri, as suggested by Rajendralal Mitra, the originaleditor of the inscription. This identification becomes more probable when we recall that thefounder of Manyaklita (or perhaps son of the founder), the great king Nrpatuliga Amogha'var$a (814-80), alone among the Raslrak0tas bore the biruda (Ati{ayQ dhavala. Repeatedreferences to him by this title are found in the inscriptions issued during his reign, and inthe remarkabfe work on Kannada poetics produced at his court, the Kaviraiamarga, ilself anadaptation of the Kavyadaria."

Nanda is also called in the grant Mahibhadraka (?) and Gunavaloka, and whereas thefirst sems nowhere attested in connection with Amoghava 6a, -aval6ka is at least an epithet-suffix of the Kannadiga Ra$Fakutas (cf. Kha{gavaloka in the case of Dantidurga, andRanavaloka of Stambha, son of Dhruva).|2 The title "Beholder of (Literary) Excellence"would make good sense in reference to the patron of the Kavirajamarga, whose knowledgeof literature is praised throughout the text. The name Nanda or Nanna itself is not found inreference to Amoghavarsa, either, though Nanna (or Naq4appa) is not uncommon among theearlier Ra$Fakutas.rr What is especially significant, however, for the identification of Nandawith Amoghavarga is tbe fact that RatnaSrl describes Nanda having become a renunciate atthe end of his life:

yai cante tanum utsasarja vidhivad yogtva urtha*ayah (line 4)

(At the end of his life he abandoned his body according to rule, like a yogin, residingin a holy place.)

This fits well with what we know about Amoghavarsa from other works, especially thecelebrated little "catechism" €ntitled the Pninottamntnamafika. Although this work hasbeen variously attributed--most commonly to Sankara, in a highly vedantized version-whatI believe to be the oldest extant manuscript copy of it, a palm-leaf written in an OldKannada script and preserved in the Oriental Research Institute, Mysore, ends with thefollowins verse;

639

vivekat tyakhrajyena raifieyarh ratnamalika /,,racimmoghavaSe4a sudhiyarh sadalanktih //' "

(This Little Garland of Gems, a goodly ornament for the wise,. was composed by King

Amoihauar5a, who on gaining discriminating insight renounced his Kingship )

If the line of reasoning offered so far is conect, then it remains to determine the king

ofthe imperial RA$trakgtasio whom 'Tuiga' refers. No*' among the overloads whose dates

are closeio both the old and revised Pala chronology, "fNpa]tunga" is recorded as a biruda

of both Govinda IV and Krna III; the reign of the first began in 930, that of the second,

as noted earliet ended in 967.15 Govinda IY however, ruled for only six years (he was

succeeded by Amoghavarga III, whose reign lasted only from 935-39), and this would leave

Krsna III as the Tunga who was Ratnsrl's paron. Little additional hard data are available

to corroborate tbis identification; the Bodh Gaya grant gives Tunga the blruda Dharmavaloka,

which I have not yet found used in connection with Knna IlI. There are, however, several

more pieces ofciriumstantial evidence that can be adduced in support. The fact that Ratnafl

calculates Tunga's date according to regnal y€ars conforms with Krsna III's own practice in

his records foimost of his reign lall other Rastsakiitas, like their predecessors the Badami

Calukyes, used rhe Saka era). Moreover, there is an important historical linkage with Sri

Lanka at just this time.

Wehavelittleinformationregardingthecirculationofreligiousprofessionals-andscholars between India and Sri Lanla in the first millennium, but what we do know about

political history suggests a very intense interaction between the two regions precisely in the

mid-tenth ""nt'rry.-krSnu III, iometime around 950 perhaps, invaded the island and was

repulsed (thus according to Sinhala chronicles; the grant to vadijatighala also alludes to the

event). Ai some later point, according to a Cola inscription, a Sri Lankan prince king

i.iuuitutSu Madanaraja, visited the court of'Kannara,' that is, Krsnaraja III.16 Where kings

and armies go, intellectuals generally follow.

One additional, if minor, indication that Ratnadri attended the Rastrakuta court of the

Deccan, and in the tenth century, is offered by the fact that he quotes from the Dam ayanfikatht(Nalacimpa\ of Trivikamabhitta. Trivikrama himself lived in the first quarter of the tenth

""ntr.y, -i worked as an inscriptional poet at the court of Indraraja III, for whom he

composed an important pmiasti in Sa*a glO (AD 915).r/ This is the latest text Ratna$Iquotes in his comrnentary and in fact this appears to be the very first citation of Trivikrama

in alaikara literafire (the next being in the works ofBhojaraja, two generations later).18 No

one outside ofthe Deccan in the tenth century appears to have known the work ofTrivikrama.

As the Bodh Gaya inscription shows, however, Ratnasri had an important conneclion

with the pala world, or at leasf with Buddhist religious and no doubt educational institutions

in the Pala world. This, rather than Tunga's subordination to Pala overlord, is likely lo

account for the dating ofhis commentary according to Pala regnal y€ars Again, if Ratnadrl's

t€xt of the Katyadaia often agr€es with the Tibetan, according to the editors, it often agrees

with that ofVadljanghaladeva as well; indeed, the two sometimes share readings of Da'''din's

mjulasrantha not found elsewhere.le And this would make sense if, as the above logic leads

640

F

,," to conclude, Vadijannghaladeva and Ratnadrl were contemporaries at the RaslrakDta

l.J or rrsnu Ilt (note ihar vadrjanghala is also called Bauddhabudhopama, "like the

guoanu tir."trin rhe mastery of Buddhist doctrine').ru lt is therefore possible to argue.

inrt, tt'..aitort, thal, rather than being dependent on some earliel northem textual tladition

ii'i" Xawaau";u, Rarna6rl 's version of the work comes from the south and was subsequently

;i;;;i into rhe northeast, where it formed the foundation of th€ Tibetan tradition in

""L."ou"n". of the authority of his exegesis and the fact that he was a Buddhist. His close

iJfuri*tftip with centers of leaming in Magadha as evinced in the Bodh Gaya inscription

makes this PerfectlY reasonable.

The fact that all of our other early exegetical works on Daldin (those ofvadijannghala'

Tarunavacaspat i ,andTarupa'ssonKedavaBhaf iaraka)wereproducedinlate-medievalfu:rJudu .ount.y seems to me additional evidence in favour of identifing this as the time

uia pU"" of RainaSri. Dandin spoke with special-force,to Kannadiga (and other southem)

,5"holiust., as he spoke to vernacular scholars such as Srivijaya, avthor of Kavftajamarga'

oill" unonyrnou. uuthor of the twelfth-century Tamil wo:.k, the Tantiyalankanm. This is no

doubt a "onsequ"nce of the acute understanding that Dandin himself, a southemer, brought

o b"u. on the question of how southem poets wrote Sanskrit; his idea 9f vaidatbha marga,

l;;;;;;1., di.". not emerge from thi realm of pure imagination 2r (And it is worth

remarking parenthetically but very explicitly that Dandin's work entirely transcended religious

boundariJs. eside nom adducing in illustration poetry composed by Buddhist authon, Ratnairi

betrays not a sign of anything we could properly call a Buddhisf literary culture Literary

culture is entirely ecumenical: sahitasya sarvapa$adatuat' asBhoJataja put it with epigramma-

tic simplicity.) fhat Daqdin spoke so powerfully at this particular moment, in the mid-tenth

arntury aur"ly has some connection-but whether as cause or effect is unclear-with the

r*pf"ri"" of vemacular writing in the Deccan, of which Pampa's Bhamf4 composed at the

court of the Vemulavd(a Calulyas (vassals and ultimately comPetitors ofthe RasFakutas)'

is the most comPelling examPle.

AninlriguingquestionawaitingfurthelinvestigationconcemsthepossibleidentityofRatnadrijflana"witi a leamed Thera oi Sri Lankan named Ruvan-mt. To Ruvan-mt is ascribed

itr" futupt t""" (sannaya) that accompanies the Siyabaslakara--(Poetics of One'a Own

iunguug"), the enormo,lsiy influential Sinhala adaptatiol' of the Kavyadaria' which itselfis

""rig""'"t"ffy dated to the ninth or tenth century. Nothing further is known about this monk'

whosi name'is usually taken to equate with Ratnamadhu' But it could well be a tadbhava

oi natnamatilpaaa), a name by which RatnaSrl is kuown -in the Sinhala tradition'

n"rr"."tif aa"'*"s iite author ofiwo works on grammat the Sabdarthacina, and a Pafijika

on CunJ.ugoain'. grammar' (Not€ that Ratnairl's knowledge of grammar is everywhere

evident iniis comirentary on Dandin; he is also one of the verl feJv 99.tT9,"3i9*. t"

demonstrate knowledge of a now-lost work on Prakrit grammar by-Hanvfddha )" wheth€r

Ruvan-mi is to be identified with Ratnasrijflana-and indeed, whether he may be the actual

author not only ofthe paraphrase but of thi siyabaslakara itself, which is ascribed tpr.gualivpseudonymouily) to King Sena I (846-66Fseems not beyond the realm ot possrblllty"-

641

2.

) .

6.

't.

8.

9.

.

3.

References & Notes

For a discussion of Kannada Kaviaiamarya, see my "The cosmospolitan vemaculaa (Joumal ofAsran

studies 57,1 0998]: 6-3?).On the sinhala and Tibetan versions. th€ most recent and best accounts are,

respcctively. ihartes Hallisey, "Works and Penons in Sinhala Literary Cultures," and Matthew Kapstein,..Tire Indian Literary ldentity in Tibet," in my forthcoming volume, Lilerary Cultures in Histotyj

Recorrirrucfiors fro; South Asia (Berkely: University of Califomia Press). On the transformation of latepali, se€ steven collions, ..what is Literature is Pali?" in the same volume. The influence of Daudin in

the iamil tradition is considered in Anne E. Monius, In Searr; of "Tamil" Buddhism: Language, Litemry

cultwe, and Religious conmunity in Tafiil-speaking south India (Hdrvard university Dissertation, 1997).The imDact of D;din and related texts in China is discussed in a path-br€aking article by Victor H. Mair

and Tsu-Lin Mei, *The sanskrit origins of Recent style Prosodf (Harya rd toumal of Asiatic stttdies 51.2

[t991], 175-470). tnaccessible to me is the Mongolian trdnslation of Dandin- (Ulanbator: Shinzhlekh

iJkha;nui Akademiin Khelel. 1972. Corpus Scriptorum Mongolorum' Vol' 18, fas€ l)

Tatacharya. D.T. ed., Kavyadada (Tirupati: Shrinivas Press' 1936 [the Upodghata' however, is dated

l94ll). Tatacharya refers (p. i) to earlier editions: of the first chaptet of the first comm€ntary by one

Krishnacharya oithe Govemment Sanskrit College, Madras (for which I can find no bibliographic details):

of complete texa ofthe two other commentaries by M. Rangacharya (Madras: i4 C Narasimhacharya of

the Braimavadin Office, l9l0). I ignore h€re the fou.th commentary, a lat€ (€ighteenth-century?) exegesis

by Krsoakanta (or Knnakinkara) Tartavaglca, a few extracts ofwhich wete publishcd by Thakur and Jha(iee below), 283ff. There are of course likely to be mo.c than those listed in the NCC During ny

ixamination in December, 2000, of the uncatalogued manuscripts on sahiryadistm in the royal library in

Ramnagar Fort, Varanasi, I found two texts of thl Karyadada with commentaries that remain to be

identified.The first spelling is found in the printed colophon to the first panbcieda; the second and third ale

according to Mad-ras Mss. (both tEnscribed fiom a Malayali original; s€ e Tiennial catalogue of Manus9tipts,

Madms R. No. 4347 and R. No. 2746 rcspectively)

Mysore Archaeological Report (MAR) l92l : Plate x, 8 ff ' u' 168'169 (on Kr$Da lll' I I l20 ft) Tatacharya

is fess impressed-with vidi.la;ghaladeva's lcaming than the man himself is (vtfhtiva atipati&amyatipaitmmayati ca, ii\.See MAR l92l: Plate X, l l . 162-63.V. Raghavan, "The Snktiratnahara of tklinngaraya Silrya,' Jouma, of Oiental Researci, MadEs 13

(1939): 105-6.Thakur, A. and Upendra Jha,(eds.) Kavyalaksalam (D"rbhanga: Mithila lnstitute' 1957)'

Thakur and lha, Ka\yalak$anam, 18-20.

Thakur and Jha, Karya laksaDam,2}. For lhe re'vised Pala chronologr, see, for example, Susan L Huntington'

Leaves from the Bodhi T.eet The Aft of Pata lodia (8th-l2th centuries) atd lts Initr,n'ational Legacy(Seattle and London: Dayton Art Inslitute in Association with the University of Washington Press' 1990)'

542.Mitra, Rajendralal, Buddha Caya: The Creat Buddrrisf Tetple (D€lhi: lndological Book House' l9?2

1o;ginattf I rlsl;, .ro+ff.

The inscription merits re-editing. onihe palaeography see EPigraphia Indica(El)

32:l14.For the first, see for example El 6:103; for the second, M.v Seetha Ramiah, ed', Kavirajamagam(Bangalore: Kamataka Sangha, 1968; reprint 1994), 1.24: 1.148; 2.55.

on the use of -aval6ka see El 6: 188-9, where Fleet notes that "Both the birudas ending in turtga and those

cnding in avaldka appear to be, originally, exclusive appellations of the RasFaj0tas of Malkhed, stnce'

as in-the case of th;'bi.dudas ending in vars'a, we cannot tmce the conception of them to any ofier

source." He offers no reason for then ascribing the Bodh Gaya grant to "another bmnch of the Ra$Fa-

klla stock" (p. 189).

t0 .

t

12.

642

{

20.21.22.

,r Mirashi. vV, "The lndragadh Stone Inscription of Ra$Fak0ta Nanoa'" Indian Historical Quarterly 3l' '

, ig55) , qq- toq. Edi ted inEl l : : l l2 :17 '

l1 ::.:T::il,5ii;ir"trltii--*Ilr,Erre:28e(co'ectinsEI 6:rtte).Asmentionedearrier,ruisatt

;:Jd ;; ; ;. in" g.unaton of Nanda according to the Bodh caya grant' though it is not at all clear

i^" "ri",f" we are to-take Ratnadtr's genealo$/ The actual gmndson of Amoghavarsa was Jagattuiga'

illnil;i"";;ii; iu,r,., una n"u.r-** f.ir;t. Indraraja tll, Jagattunga's elder son, ruled between 9l+

;;: i;;";; "^ a "fifteenth" vear in Indn's reig' therefore' it would have been his last' At all events'

i have not found (Nrpa)tunga as a biruda of Indra Ilt'

ro. i"" Nogu.ju, S, "RasFaknF Activities in Ceylon'" in B R Copal' ed ' Ttte Rasfaftotas of Malkhed:'"

iirai", ji 7l"t n istory and Culture (Mysore: Getha Book House on behalf of the Mvthic society' | 994)'

t23-5.

17. See El 9:28. Ratnairl cites from Trivikrama on p 23'

Igi;. ",g,, lnga.p**aia (edited by V Raghavan [Cambridge' MA: Harvard University Press' 19980' 354

lNalacamli l 4), l0l9 (5 14)'

19. Thakur and Jha only remark on this (p 16) but offer no evidence' I have not done a systematic compansoni"

"" i"io.rnJ ."."t of ttre fint par;cc6eaa shows mixed results. on l,6?, both R [aianirt] and v[adijanghalal

i""i'i"rai i"g"riti mrram of e'g', Taruna); on l'76 sarva- (against *'rva.of many mss} on l '95 atyat

i;;;i; ;;;t. of manv mssiand apparentlv Taruna: on l ?l .b€lieve v read kavvam (not *avvq as

n;nre.tl in aneement with R. agarnst iaruna and othen;. Some differences, however, are also found: on

ilii, i"""a."r"rtii,p*, v urr'Jlsiptat; on l'38, R pativate' v bad'tvaa It is impossible to tell from

V ,"'ftatft". f," read in agreement with R on 1 48-49' A iomparison of their interpretations of problematic

passages would be of int€rest.

MAR l92l: Plate X, l. 162.

see "The Cosmopolitan Vemacular" 2l-5'

See his comments on l.3l-5. For Hari\Tddha as Prdkrit poet' see Harivallabh Chunilal Bhayari' lndologial'irai",

lir"^ty "ia p"tf*ning Atts, Pmkit' and Apabhmmsa Studies (Ahmedabad: Prashva Prakashan'

iiqJil ozff. g'sri"m t and iiotogicat studies: Dr' i Raghavan Felicitation uolume' Pp' 61 ff') Ratnadrl's

quoraiion. ofHari't gftlmmar are unfo.tunately unknown to Bhayani'

This Damsraph relies heavily on Martino de Zilva Wickremasilghe' Catalogue of Sinhalese Manuscnpts

in tni g;tisi Museum (London: Bridsh Museum. 1900)' xiii'

643