10. CHAPTER SEVEN MUKULABHATTA AND PRATIHARENDURAJA

CHAPTER SEVEN MUKULABHATTA AND PRATIHARENDURAJA

We have seen that a central element in Anandavardhana's

defense of his approach to literary analysis is his contention

that the theories devised by previous analysts of both

literary language and language in general cannot provide an

adequate account of the ways in which poetry functions. All

previous efforts at literary analysis, Anandavardhana argues,

have overlooked the existence of dhvani, most critically rasa-

dhvani which, in his view, provides the basis for aesthetic

coherence in literary texts. He takes great pains to

demonstrate that what he calls dhvani has not been dealt with

by earlier Ālamkarikas under some other name; Anandavardhana

views his dhvani theory not as a re-analysis or an improved

treatment of aesthetic phenomena dealt with by his

predecessors, but rather as an examination of phenomena which,

despite their crucial role in the creation and appreciation of

aesthetically effective poetry, lay entirely outside the scope

of Alamkarasastra as it had so far been practiced. He further

argues that the process by which this dhvani is understood by

the readers or hearers of a poem cannot be accounted for in

terms of the two modes of signification (direct and figurative

expression) generally accepted by philosophers of language,

and makes it necessary to postulate a third, specifically

poetic, mode of signification, namely suggestion.

Anandavardhana's treatment of dhvani thus presents a

direct challenge to established theories of both literary

304

305

language and language in general. This attack, naturally

enough, provokes a response among the adherents of these

theories. The two authors whose works we will examine in this

chapter are concerned to uphold the adequacy of existing

linguistic and literary theory in the face of Anandavardhana's

attack. Their works are the earliest critical responses to

the Dhvanyaloka which are available to us, and they reveal the

extent and rapidity of the transformation in Alamkārasāstra

produced by Anandavardhana's work.

Anandavardhana's radically new approach to literary

analysis raised new questions and new expectations which works

on Alamkarasastra were expected to meet. Although the works

of Mukulabhatta and his pupil Pratihārendurāja represent a

conservative attempt to defend established linguistic and

literary theory against Anandavardhana's charges of

inadequacy, their defense of these theories takes the form of

an effort to "update" them -- to elaborate or extend them so

that they can be seen to account for the phenomena

Anandavardhana believes they need to account for but cannot.

Their quarrel is not with Anandavardhana's claim that the

indirect, non-literal communication of narrative elements

(vastus), figures of speech (alamkaras), and especially rasas

plays a pivotal role in aesthetically effective poetry, but

only with his contention that this non-literal communication

cannot be explained within the framework of existing literary

and linguistic theory and demands the acceptance of a

306

previously unrecognized poetic phenomenon (dhvani) and,

concomitantly, a previously unrecognized semantic mode

(vyañjana). Their defense of existing theory is thus carried

out within the context of a general acceptance of

Anandavardhana's model of literary aesthetics.

Mukulabhatta

Mukulabhatta identifies himself as the son of Bhatta

Kallata.1 Kallata is known as the author of a commentary on

the Spandakarika (seminal work of the Spanda ["Vibration"]

school of monistic Saivism) and, according to some, of the

Spandakārika itself. According to Kalhana,2 Kallata lived

during the reign of Avantivarman (r. 855-883), and was thus

more or less contemporary with Anandavardhana. Thus

Mukulabhatta's work should date from the late ninth or early

tenth century.

The only extant work of Mukulabhatta is his

Abhidhāvrttamātrka ("Fundamentals of the Denotative Function")

a brief examination of the modes in which words convey

meaning. Mukulabhatta accepts only two such modes, direct and

figurative expression; Anandavardhana's "suggestion", in his

view, falls entirely within the scope of the latter.

Mukulabhatta's work and, in particular, his critique of

Anandavardhana, while they were quite influential in tenth and

1Mukulabhatta, Abhidhāvrttamātrka, ed. Revaprasad Dvivedi, Vidyabhavana Samskrta Granthamala 165, Varanasi, 1973, p. 73.

2RajataranginI, V. 66 (Stein's ed., Vol. 3, p. 74).

307

eleventh century Kashmir, have been almost entirely neglected

by modern scholars in the field of Alamkarasastra. Kane and

De both recognize that Mukulabhatta is aware of and refers to

the Dhvanyaloka and note, in passing, that he would include

dhvani within the scope of laksana (figurative expression) ,

but neither takes the time for even a cursory survey of his

arguments against Anandavardhana.3 Krishnamoorthy, in his

otherwise thorough survey of critical response to the

Dhvanyaloka,4 takes no notice whatsoever of Mukulabhatta's

opposition to Anandavardhana's theory of suggestion. Gerow is

not only unaware of Mukulabhatta's response to Anandavardhana

but suggests, anachronistically, that Anandavardhana was

influenced by Mukulabhatta's treatment of the literal and

figurative modes of signification.5 Kunjunni Raja, in his

Indian Theories of Meaning, makes serious use of

Mukulabhatta's work in his own treatment of laksana. He

recognizes that Mukulabhatta defines laksana so as to include

everything Anandavardhana would call suggestion and labels him

3Cf. P.V. Kane, History of Sanskrit Poetics, op. cit., pp. 192, 218; S.K. De, History of Sanskrit Poetics, op. cit., Vol. I, pp.73-4, and Vol. II, pp. 141-2, 179-80.

4K. Krishnamoorthy, The Dhvanyaloka and its Critics, Kavyalaya Publishers, Mysore, 1968.

5Edwin Gerow, Indian Poetics, Vol. 5, Fasc. 3 in A History of Indian Literature, ed. Jan Gonda, Otto Harrassowitz, Wiesbaden, 1977, p. 258.

308

as an opponent of the dhvani theory;6 but here again there is

no concerted effort to examine the arguments Mukulabhatta

offers in support of this view.

The only serious discussion of Mukulabhatta's views I

have found in the secondary literature on Alamkārasāstra is in

Mukund Madhav Sharma's The Dhvani Theory in Sanskrit Poetics.7

Sharma is a thoroughgoing proponent of the dhvani theory and

does all he can to find fault with Mukulabhatta's criticism of

Anandavardhana, reiterating those arguments advanced against

him by Abhinavagupta in his commentary on the Dhvanyaloka and

by Mammata in his Sabdavyaparavicara; nevertheless, he is

virtually alone among modern authors in recognizing

Mukulabhatta's importance as an early opponent of the

Ānandavardhana's views. He regards resistance to the theory

of suggestion (rightly, I believe) as Mukulabhatta's primary

object in writing the Abhidhavrttamatrka, rather than as an

incidental matter dealt with by him only in passing.

The general neglect of Mukulabhatta's work by modern

scholars of Alamkārasastra is largely attributable to its

somewhat problematic genre-character. The Abhidhavrttamatrka

does not present itself as a work of Alamkārasāstra. Its

stated purpose is, by means of an analysis of the primary

6K. Kunjunni Raja, Indian Theories of Meaning, Adyar Library and Research Centre, Madras, 1969 [second ed.], p. 295.

7Mukund Madhav Sharma, The Dhvani Theory in Sanskrit Poetics, Chowkambha Sanskrit Studies, Vol. LXIII, Varanasi, 1968, pp. 224-234.

309

(mukhya) and figurative (lāksanika) modes of verbal

signification, to provide a clear understanding of language --

necessary, it is argued, for the definite ascertainment

(arrived at through the medium of language) of the nature of

things in the world and, thereby, for the attainment of all

human ends.8 Although in concluding his work Mukulabhatta

refers directly to the usefulness of a correct understanding

of these modes in the composition of literature, he does not

explicitly present this literary application as the primary

goal of his analysis, but refers once again to the importance

of such an understanding in all human endeavor.9

If the Abhidhavrttamātrka does not clearly fall within

the scope of Alamkarasastra, however, neither does it belong

to any existing genre of Sanskrit philosophical literature.

Mukulabhatta does not present his work as belonging to any

recognized school of philosophical thought. His general views

on language correspond fairly closely to those of the Bhätta

Mimamsakas and he draws on Mimamsa sources in his treatment of

figurative language, but he tries to harmonize his semantic

analyses with the views of the Grammarians and Prabhākara

Mīmāmsakas as well. More importantly, in providing

illustrations for the various sub-varieties of figurative

expression he discusses he employs examples drawn from poetry

as well as from everyday (and occasionally Vedic) language.

@Mukulabhatta, op. cit., p. 1.

9Ibid., pp. 72-73.

310

This concerted interest in poetic usage has no parallel in

Sanskrit linguistic philosophy. Poetic references of any kind

are extremely rare in Sanskrit philosophical works, and there

are certainly no such works which display anything like the

degree of attention Mukulabhatta devotes to the investigation

of poetic expression.

Apart from his clear interest in matters of poetic usage

which ordinarily fall solely within the province of

Alamkārasāstra, the response generated by Mukulabhatta's work

in tenth and eleventh century Kashmir indicates that he was

viewed, at least primarily, as an Alamkarika. As already

noted, both Abhinavagupta (c. 1000) and Mammata (c. 1100),

Anandavardhana's most prominent followers during this period,

are aware of Mukulabhatta's work and respond to his criticism

of the theory of suggestion. Mammata's Sabdavyaparavicara is

both an imitation and a detailed critique of the

Abhidhavrttamatrka, incorporating as much of its analysis of

literal and figurative usage as is consistent with the dhvani

theory (often copying Mukulabhatta's text virtually word for

word) while attacking the e arguments it advances for the

inclusion of dhvani within the scope of figurative expression.

In Vallabhadeva's (c. 925) commentary on Sisupalavadha 3.32

there is an apparent reference to Mukulabhatta in which he is

referred to as "one who knows alamkāras [and/or

Alamkārasāstra] " [alamkāravid] .10 In contrast to the careful

10Cf. below, p. 335.

311

attention it receives from later Alamkārikas, Mukulabhatta's

work is entirely ignored by linguistic philosophers outside

the field of Alamkārasāstra.

What we know of Mukulabhatta's personal connections

likewise suggests that he should be viewed primarily as an

Ālamkārika. His two known pupils both worked in the area of

Alamkārasāstra: Pratīhārendurāja, who wrote the earliest known

commentary on Udbhata's Kāvyālamkārasārasamgraha, and

Sahadeva, who wrote the earliest known commentary on Vamana's

Kavyalamkarasutravrtti. Both state that they learnt the works

they comment on from Mukulabhatta, and Sahadeva states further

that it was Mukulabhatta who revived the study of Vamana's

work, which had apparently by his time fallen into

obscurity.11 All that we know of his intellectual life would

thus seem to indicate that Alamkarasastra was his primary

field of study. 12

The Abhidhavrttamatrka should thus be viewed as a

11Cf. Kavyalamkarasarasamgraha of Udbhata Laghuvrtti of Pratiharenduraja, ed. N.D. Banhatti, pp. 1, 92. with the

Sahadeva, Kavyalamkarasutravrttitippana, p. 1 (Sahadeva's commentary has not been printed, but I was able to obtain a copy of a modern transcript of the work from P.L. Shaji, director of the Oriental Manuscript Library of the University of Kerala [Manuscript T.316]). Sahadeva informs the reader that Vamana's work, the transmission of which had been interrupted [bhrastāmnāyam], was restored [samuddhrtam] by Mukulabhatta, who had obtained a manuscript from some source.

12Mukulabhatta's son Harsata, while not an Ālamkārika, was active in the related discipline of metrics. He is the author of a commentary on the Jayadevachandas (ed. by H.D. Velankar in Jayadaman: A Collection of Ancient Texts on Sanskrit Prosody [Haritosamāla 1, Bombay, 1949], pp. 1-40) .

312

contribution to the field of Alamkärasastra, but one which

departs in important ways from the mode of treatment

conventional to such works. It represents, in fact, the first

example of what was to become a new sub-genre of

Alamkarasāstra. Short works dealing with the various modes of

verbal expression very much after the manner of Mukulabhatta

are not uncommon in later Alamkārasāstra; the best known (but

by no means the only) such works are Mammata's

Sabdavyāpāravicāra and Appayadīksita's Vrttivārttikā. 13

Mammata's Kavyaprakasa and many of the later systematic works

on Alamkarasāstra devote a chapter to the same topic.

The primary objective of such works, including

Mukulabhatta's, seems to be to provide an account of semantics

in general, more or less in the terms of Sanskrit linguistic

philosophy, which can serve as a theoretical basis for the

treatment of more specifically poetic topics (Anandavardhana's

dhvani, as well as those matters dealt with by earlier

Ālamkārikas -- the analysis of alamkāras, gunas, etc.). Until

Mukulabhatta's time, no one in the field of Alamkārasāstra

seems to have felt the need to provide such a systematic

philosophical basis for their analysis of the distinctive

features of poetic language. It is only in the wake of

13Because these works generally accept Anandavardhana's vyañjana as a third semantic mode and are often, like the two examples mentioned above, composed by authors who have written other works on poetics, they have generally been seen, unlike Mukulabhatta's work, as belonging unproblematically to the field of Aiamkārasāstra.

313

Anandavardhana's work that the relationship between poetics

and philosophical language analysis becomes a pressing

problem.

It has already been noted (cf. above, Chapter Five, pp.

194ff) that, up to Anandavardhana's time, the Alamkarikas had

developed their analysis of literary language for the most

part independently of the analyses of language in general

formulated by linguistic philosophers, and these two groups

had, in fact, developed two largely distinct conceptual

vocabularies in terms of which language could be discussed.

By introducing philosophical techniques of language analysis,

in particular the Mimamsakas' teleological approach to text

interpretation, into the study of poetic language,

Anandavardhana began to bridge the conceptual and

terminological gap which had previously separated literary

from non-literary language theory. Furthermore,

Anandavardhana's insistence that the philosophers' analysis of

semantics is insufficient to fully account for poetic usage

and that an additional semantic mode, vyañjana, must be

recognized if such usage is to be adequately explained places

his theory in explicit conflict with all established schools

of philosophical thought. Subsequent Ālamkarikas, whether

they accepted or rejected Anandavardhana's vyañjana, could

hardly ignore this conflict. The need to take a stand, one

way or the other, on the existence of vyañjana raised the

question of the linguistic/philosophical basis of the concepts

314

and categories employed in Alamkārasāstra to a prominence it

had not so far held.

So, Mukulabhatta's apparent departures, in the

Abhidhavrttamatrka, from the topics and the mode of treatment

typical of earlier works on Alamkarasastra can be attributed

in large part to his attempt to deal with this newly pressing

question of the philosophical basis of literary analysis.

Mukulabhatta's quarrel with Anandavardhana and his defense of

the adequacy of earlier philosophical theories of language

even in the area of literary analysis also help explain these

departures. The genre is, in this case, part of the message;

given Mukulabhatta's contention that poetry can be

satisfactorily explained and analyzed in terms of established

general-linguistic theory, without resorting to a special

theory of poetic language or postulating a specifically poetic

mode of expression, it is quite understandable that his work

should present itself as an analysis of language in general,

rather than of poetic language in particular. The

Abhidhavrttamatrka treats poetic, Vedic, and everyday language

in terms of a single, unified theory of signification.

Direct and Figurative Expression

For Mukulabhatta, all meanings can be divided into two

broad classes: those which are understood directly through

their association (whether natural or conventional) with the

individual words of an utterance, and those which are arrived

at indirectly through reflection on these individual word

315

meanings.14 He calls meanings of the first type "primary"

(mukhya) ;15 all meanings of the latter variety he designates

as "figurative" (laksanika). Thus, like the Bhatta MImamsakas

(whom he follows in most respects), Mukulabhatta allows a very

wide scope to the concept of figurative usage, such that

virtually any expression will have to be regarded as conveying

some figurative sense.

Consider, for example, the Vedic injunction "A cow should

be sacrificed" (gaur anubandhyah). Here the word "gauh" by

itself, according to the standard Mimamsa view, 16 cannot

refer directly to any particular, individual cow, but only to

the eternal, universal class-property "cow-ness". If the word

is understood only in its primary, literal meaning, however,

the injunction will make no sense; one cannot sacrifice "cow-

ness", but only some particular, individual cow. In order to

derive a coherent meaning from the sentence, one must

understand the word "gauh" to refer, figuratively, to some

individual possessing the class-property in question.17 As

14Mukulabhatta's claim that the understanding of any meanings not directly expressed invariably derives from the directly expressed meanings of the words employed, rather than the words themselves closely parallels the view o such meanings later advanced by Mahimabhatta in his Vyaktiviveka (cf. below, Chapter Ten, pp. 516-518).

15Equivalent to what Anandavardhana and most others call vacya (" [directly] expressed").

16Cf. above, Chapter Three, pp. 70ff.

17Abhidhavrttamātrkā, pp. 2-3 In support of his contention that words cannot refer directly to individuals, he cites, without attribution, the following half-verse:

316

nearly all sentences must similarly be understood to describe,

or to enjoin action with regard to, individual entities rather

than abstract, universal class-properties, figurative

expression can be seen to play a vital and pervasive role in

language of all sorts.18

It is this expansively defined and pervasive figurative

mode of expression which absorbs the bulk of Mukulabhatta's

attention. His treatment of the primary or direct mode of

expression is quite brief and proves to be relatively

uncontroversial. There are, for Mukulabhatta, four sorts of

attribute with which words may be directly associated and,

thereby, four sub-varieties of primary expression. There are

words which express universal class-properties (jāti-sabdas --

"Denotation does not reach the thing to be specified; its power is exhausted in specification" (visesyam nābhidhā gacchet ksinasaktir visesane) -- denotation, i.e. direct expression, cannot refer to particular entities, but can only indicate them indirectly by referring to those attributes which characterize them -- nct only class-properties like cow- ness, but other repeatable properties such as color or motion. The same words, likewise unattributed, are quoted by Abhinavagupta (Locana, p. 55), Mammata (Kavyaprakasa [Mysore: Oriental Research Institute Series 120, 1974], Vol 1., pp. 51- 52, Sabdavyaparavicara [Delhi: Indu Prakāsan, 1977], p. 6), and several later Ālamkarikas. Editors and contemporary scholars of Alamkārasāstra have failed to trace the source -- Kamalākarabhatta identifies the words as Kumarilabhatta's (Mīmāmsākutūhalam, ed. Pattābhirama Sastri, SaravatIbhavana Granthamala 123, Varanasi, 1987, pp. 80-81) . They cannot be found in any of his extant works; presumably they are from his lost Brhattīka.

18Again, Mukulabhatta is echoing the view of the Bhätta Mīmamsakas -- cf. above, Chapter Three, pp. 115-117.

317

e.g. "cow"), words which express qualities19 (guna-sabdas --

e.g. "white"), words which express actions (kriya-sabdas --

e.g. " [he] cooks"), and words which, through an act of will on

the part of a speaker or group of speakers, are made to refer

to particular individuals (yadrccha-sabdas -- i.e. proper

names) . This fourfold classification of words is derived

ultimately from Patañjali's Mahabhasya, which Mukulabhatta

quotes in support of his view.20

This classification of primary signification has no

relevance to the question of vyañjana; Mukulabhatta deals with

the topic quickly, and it seems little more than a necessary

prelude to his far more elaborate discussion of figurative

usage. Anandavardhana's followers, and later Alamkārikas in

general, have no serious quarrel with Mukulabhatta's treatment

of the matter, and it is incorporated virtually without

alteration into both of Mammata's works.21

It is to the analysis of figurative signification

(laksana) that Mukulabhatta applies most of his attention, and

it is here that he begins to break new ground. Mukulabhatta

is, in fact, the first to seriously attempt a complete and

19Distinguished from class-properties in that they are incidental to and not constitutive of the entities which possess them.

20Abhidhāvrttamātrkā, pp. 4-5. Cf. Patañjali, Vyakarana- mahabhasya, Vrajajīvan PrācyabharatI Granthamala 23, Chowkhamba Sanskrit Pratishthan, Varanasi, (reprint) 1992, Vol. I, p. 101.

21Kāvyaprakāa, op. cit., pp. 66-79; Sabdavyaparavicāra, op. cit., pp. 2-4.

318

systematic account of figurative usage in all its varieties.

Questions of metaphorical, metonymical, or other types of

figurative usage are frequently dealt with in earlier

philosophical literature, but usually only as they crop up

incidentally in connection with other matters.22

Mukulabhatta divides laksana into two broad categories

and, within them, a total of six subdivisions. A given

instance of laksana may be either upacāra-misra ("combined

with superimposition"23) or suddha ("pure" -- i.e. free of

such superimposition) .24 An example of the suddha type is

the expression gangayam ghosah ("A village on the Ganges").

Here, the word gangayam ("on the Ganges") cannot be understood

literally, as it is impossible for a village to rest on the

actual waters of the river; we understand it to refer,

metonymically, to the bank of the Ganges. We mentally

substitute for the contextually impossible literal referent of

the term a closely connected referent which fits the context.

This indirect mode of reference does not, however, depend upon

the ascription of any qualities of the first referent to the

22For the more important discussions of figurative expression in the earlier philosophical literature, cf. Mīmāmsāsutra 1.4.23 and 3.2.1 (with Śabara's Bhāsya, Kumarila's Tantravarttika, and Prabhakara's Brhati thereon), Nyayasutra 2.2.62 (with Vatsyayana's Bhasya and Uddyotakara's Vārttika), Patañjali's Mahabhasya on Pānini 4.1.48, Bhartrhari's VakyapadIya 2.263ff. and 'Dharmakirti's Pramanavārttika 2.35-37.

23That is, the superimposition of the literal referent upon the actual (figurative) referent.

24Abhidhāvrttamātrkā, p. 11.

319

second. One referent has simply been replaced by another,

without affecting our understanding of this second

referent.25

In upacara-misra laksana, on the other hand, such

transference of qualities does take place. In the expression

"The Punjabi is a cow" (gaur vahikah), the word gauh ("cow")

cannot be understood literally; its juxtaposition with the

word vähikah requires that it be understood as referring to a

Punjabi, a human being. Such figurative usage requires some

connection between the literal and actual referents of the

term in question and, in this instance, we understand the

connection to be the possession of similar qualities.

Here the word "cow" has its primary meaning blocked, because of the impossibility of its co-referentiality26 with the word "Punjabi"; via the figurative indication of qualities (slow-wittedness, laziness, etc.) belonging to the Punjabi which are similar to the qualities (slow- witted-ness, laziness, etc.) belonging to a cow, [the word "cow"] is superimposed [upacarita] on the Punjabi, who possesses qualities (slow-wittedness, laziness, etc.) similar to the qualities (slow-wittedness, laziness, etc.) belonging to a cow.27

25But cf. below, pp. 326-327 and 351-354.

26Samanadhikaranya -- lit. "the property of having the same substratum"; two or more words in a sentence which are syntactically parallel and are understood to refer to a single entity are said to "have the same substratum".

27Ibid., p. 11. This rather cumbersome description of the process by which metaphor is understood is clearly modelled on Kumarilabhatta's treatment of the topic -- cf. Tantravärttika on 1.4.23 (Ānandasrama Sanskrit Series 97 [second ed.], Vol. 2, p. 313), and above, Chapter Three, pp. 109-110. Kumarila emphasizes the distinction between metaphorical and non- metaphorical modes of figurative expression. In the latter, the first, literal referent figuratively indicates a second directly related to it (e.g. the Ganges and its bank). In the

320

Such metaphorical usage does not merely cause a hearer to

mentally substitute one referent (the Punjabi) for another (a

cow); if it did, the expression "The Punjabi is a cow" would

be a mere tautology. The word "cow" refers to the Punjabi

and, in doing so, alters our understanding of the Punjabi by

emphasizing his cow-like qualities. It is this alteration of

the hearer's understanding of the figurative referent that

distinguishes upacāra-miśrā from suddhā laksanā.

Suddhā laksana is further divided into two types: laksana

("indication") and upadana ("inclusion"). The first of these

is seen in examples such as "A village on the Ganges"; here

the literal referent (the river Ganges) is replaced by a

second (the bank) related to but nonetheless distinct from it.

In upadana-laksana the second, figuratively understood,

referent incorporates the first, literal, referent within

itself. Mukulabhatta's example is the Vedic injunction "A cow

should be sacrificed". Here, as already noted, the word "cow"

is thought to refer literally only to the universal class-

property "cow-ness"; it is only through figurative expression

that it can refer to a particular, sacrificable cow. Because

the figuratively understood referent (the particular cow)

includes within itself the literal referent (the class-

property "cow-ness", which it possesses), this is considered

former, the relation between the literal and figurative referents is indirect; qualities belonging to the first are identified with similar qualities belonging to the second and, on the basis of this identification, the entities possessing these qualities are identified.

321

as a case of upādāna-laksanā. 28

Mukulabhatta takes this category to include examples of

what the Bhatta Mimamsakas call srutarthapatti ("necessary

presumption [arthapatti] based on what is heard") -- cf. above,

Chapter Three, p. 91. The stock example of such arthapatti is

this: if we are told that "Fat Devadatta doesn't eat by day, "

we understand, without being told, "He eats at night. "29 The

literal meaning of the expression makes sense only if we

understand this additional unstated meaning, and this unstated

meaning is understood in addition to the literal meaning of

the expression rather than as a replacement for it; thus this

and other such examples are, according to Mukulabhatta, to be

included within the scope of upadāna-laksanā.

The category of upādāna-laksanā is of particular

importance for Mukulabhatta's case against Anandavardhana. We

have seen (above, Chapter Five, pp. 226-230, 233-237) that a

crucial feature of Anandavardhana's argument for the existence

of suggestion as distinct from figurative expression is his

contention that in figurative expression the figurative

28Tbid., p. 13. Given the persistent controversy in Sanskrit philosophical thought regarding the existence and status of universals, the example may seem problematic. A clearer case would be "Have the spears enter" (kuntān pravesaya). Here, as spears cannot enter on their own, the word "spears" (kuntän) is understood to refer to men carrying spears; the literal referent is clearly included as part of the actual, ultimately, figurative referent -- the from Mahabhasya on 4.1.48 example is drawn,

Śabdavyapāravicāra, op. cit., p. 6). (cf. also Mammata,

29Abhidhavrttamatrka, p. 13. Cf. also Ślokavārttika, Arthapatti 51ff.

322

meaning invariably supplants, rather than supplements, the

literal sense of the word or phrase used figuratively.

Because many examples of dhvani, all those classified as

vivaksita-vacya ("in which the expressed meaning is

intended"), involve the appearance of a second meaning without

the effacement of the first, they cannot, Anandavardhana

argues, be explained as instances of figurative expression.

Mukulabhatta's upādāna-laksanā, however, allows for just such

supplementation of a first meaning by a second within the

domain of figurative expression, and thus goes a long way

toward undermining Anandavardhana's case for the distinctness

of suggestion from figurative expression. We shall see that

for Mukulabhatta all cases of vivaksita-vācya dhvani,

including the all-important category of rasa-dhvani, are to be

explained as instances of this type of laksanā.

Upacāra-miśrā laksanā is divisible into four types: the

superimposition (upacara) involved in such figurative usage

may be gauna (based on qualitative similarity) or suddha

(lacking a basis in similarity) and, in each case, this

superimposition may take the form of aropa (superimposition

proper) or adhyavasana (absorption). The example of upacāra

already given, "The Punjabi is a cow", is gauna, since it

depends on the qualitative similarity between the Punjabi and

a cow. An example of suddha upacara is the expression "Ghee

is long-life" (ayur ghrtam).30 Here the figurative use of

30Abhidhavrttamatrka, op. cit., p. 16.

323

the word "long-life" to refer to ghee is based not on any

similarity between these two, but rather on a relationship of

cause and effect; the ingestion of ghee is thought to cause

long-life, and is thus designated as long-life itself. This

figurative mode of reference significantly colors the hearer's

understanding of the ghee, emphasizing its life-giving

properties, and should thus be considered a case of upacāra.

Both these examples of upacara take the form of aropa,

defined by Mukulabhatta as that "in which one thing is

superimposed upon another without suppressing the difference

between the thing imposed and the object of imposition. "31

In each case, the term employed figuratively is juxtaposed

with another which is applied literally to the referent in

question; these figurative expressions are not phrased in such

a way as to efface the distinction between "the Punjabi" and

"a cow", or "ghee" and "life", respectively. In cases of

adhyavasana, this distinction is effaced.

Where, because the object of imposition is meant to be expressed as absorbed in the thing imposed, a denial of the nature of the object of imposition is made, there there is adhyavasāna.32

As an example of such absorption in a case of gauna

laksana, Mukulabhatta gives the use of the term rajan ("king")

to describe anyone who wields political power. The term, it

is argued, can refer literally only to members of the ksatriya

31Ibid., p. 18.

32Ibid., p. 18.

324

varnas, but is used metaphorically to refer to members of

other varnas when they, by exercising political authority, act

as ksatriyas typically act.33 Such king-like persons are

figuratively referred to as "rajan", without the accompaniment

of any literally applicable term; thus the distinction between

such a person and one who might literally be designated by the

term "rajan" (i.e. a ksatriya) is suppressed. In fact, the

suppression of this distinction is so thorough and so much

ingrained in conventional usage that the figurative character

of such expressions may not be immediately apparent; it is

only through analysis that the literal and figurative

applications of the term "rājan" can be accurately

differentiated.34

The example given for adhyavasana in a case of suddha

upacara is use of the expression "the Pañcalas" to refer to

the country in which the Pañcalas dwell. The word "Pancala"

is, literally, the proper name of an ancient king. It is used

figuratively (in the plural) to designate his descendants and

their subjects and, by a further figurative extension, is used

(still in the plural) to refer to the geographic area

inhabited by these people. This mode of reference suppresses

the distinction between the country and its inhabitants, and

is thus a case of adhyavasana; the superimposition of the two

33For the argument that the word rajan can literally refer only to ksatriyas, cf. Sabarabhasya and Tantravarttika on Mīmāmsāsūtra 2.3.3 (op. cit., Vol. 3, pp. 151-162).

34Ibid., p. 18.

325

depends on a relationship of inhabitant and inhabited, rather

than qualitative similarity, and is thus a case of suddha

upacara. As in the case of "rajan", such usage is so much a

part of conventional usage that its figurative character

generally goes unnoticed.35

Mukulabhatta groups these six varieties of laksanā in

pairs (two suddha, two based on aropa and two on adhyavasana)

on the basis of the degree of identification between the

literal and figurative referents in expressions of each type:

Pure [suddhā] laksana operates figurative referent] which is independent [tatastha], with respect to [a

aropa with respect to one which is not far removed [adūraga], and adhyavasana with respect to one which is swallowed up [nigīrna] due either to linguistic convention or to extreme closeness [asannataratva] .36

In a case of suddha ("pure") laksana such as "A village on the

Ganges", one understands that the village in question must

stand on the bank of the Ganges rather than on the river

itself. The fact that the bank is referred to metonymically

by the term "Ganges" does not, however, cause us to identify

the river and the bank; we continue to understand the two as

entirely distinct from one another. "Since it [the Ganges] is

used merely to point out [the bank], we do not understand any

coloring [uparaga] of the bank by this particular river,

because the bank is understood to be independent

35Ibid., p. 18.

36Ibid., p. 20.

326

[tatastha] ."37 Our understanding of the bank is not

"colored" (i.e. effected or transformed) by its association

with the Ganges.

By contrast, in a case of äropa ("superimposition") such

as "The Punjabi is a cow", our understanding of the figurative

referent of the word "cow" (i.e. the Punjabi) is definitely

colored by its literal referent. The word "cow" is here used

not simply to refer to the Punjabi, but precisely to emphasize

his cow-like qualities; the figurative usage thus results in

a partial identification between the literal and figurative

referents of the word " Cow". Due to such partial

identification, the figurative referent is said to be "not far

removed" from the literal in such cases. In cases of

adhyavasana ("absorption") , the distinction between the

literal and figurative referents is suppressed; their

identification is thus total, and the figurative referent

(e.g., in the example of rajan, a non-ksatriya ruler) is thus

said to be "swallowed up" by the literal one (e.g. a king) .

This "swallowing up" of one referent by another may derive

from conformity to conventional usage (as in the case of

rajan), or may be used deliberately to indicate the "extreme

closeness" or virtual identity of the literal and figurative

referents.

The assignment of a particular example of laksana to one

of these three categories is not always a clear cut matter.

37Tbid, p. 20.

327

It is the degree of identification between the literal and

figurative senses of a term that determines this

classification, and this degree of identification depends, at

least in part, on the intention of the speaker using the

figurative expression. The expression "A village on the

Ganges" causes us to understand that the village actually

rests on the bank of the Ganges rather than on the river

itself, typically without causing us identify the Ganges with

the bank in any way; it is therefore classified as suddha

laksanā. However, when this expression is used with the

intent of communicating such identification, it will have to

be classified differently:

When the bank, without its own form being suppressed, is meant to be understood as colored by this particular river because of its not being far removed from the particular river denoted by the word "Ganges" then, even in the previous example ["A village on the Ganges"], there will be superimposition [adhyaropa38] ... this expression is used to communicate the extreme And when

closeness of the particular river to the village, suppressing the form of the bank so that it is swallowed up by this particular river ... then there is absorption [adhyavasāna] .39

So one and the same figurative expression, depending on

whether or not it is meant to convey either qualitative or

total identification between the literal figurative referents,

may belong to any of these three categories. Our ability to

determine this intention in any particular case will of course

depend on context.

38The terms aropa and adhyaropa are used synonymously.

39Ibid., pp. 20-21.

328

Mukulabhatta's classification allows for a total of six

sub-varieties of laksana (two suddha, two which depend on

aropa, and two which depend on adhyavasana) . Within this

taxonomy he strives to include all possible cases of non-

literal expression (that is to say, any expression in which

the individual words convey anything beyond the particular

lexical meanings with which they are invariably associated) .

If he can do so successfully, Anandavardhana's claims for

vyañjana as a third, specially poetic, mode of expression will

collapse.

The chief difficulty associated with the analysis of

laksana is its variable, conditional character. A given word

may, in different contexts, convey any of a wide variety of

figurative senses. What are the factors that lead one to

understand a particular figurative sense in a given context,

and how exactly does one learn to properly use and comprehend

figurative expressions? The acquisition of the capacity to

use and understand words in their literal senses is a

relatively straightforward process. As a child one learns, by

observing the linguistic behavior of one's elders, that each

word is invariably associated with a particular meaning.40

4°Even when words are used figuratively, they still convey their literal, invariably associated meanings; it is only through its connection with the literal sense of a word that any figurative sense can be understood. As Sabara says, "How can one word be used in the sense of another ?... We say: via the expression of its own (Sabarabhāsya on Mimāmsasūtra 1.4.22; Ānandasrama Sanskrit Series 97, second meaning."

ed., Vol. 2, p. 315 -- cited by Mukulabhatta, ibid., p. 25.

329

Whenever the word is heard subsequently, this meaning will be

understood.

Obviously, the process by which one learns to use and

understand words figuratively cannot be quite so simple.

First of all, figurative meanings cannot, by Mukulabhatta's

own definition, be associated directly with the words that

convey them; a figurative meaning is one which is understood

indirectly, through its connection with the literal meaning of

a word. Such a connection with the literal meaning of a word

cannot, however, be regarded as a sufficient condition for

one's understanding of a figurative meaning in any given

instance; if this were the case, the same figurative meaning

would be understood whenever the word was used.41 Indeed,

every word would convey all possible figurative meanings --

i.e. all meanings connected in any way with its literal

meaning -- all the time. There must be further restrictions

governing the figurative use of words.

The proper use and understanding of figurative language,

Mukulabhatta argues, must, like all other aspects of

linguistic usage, be learnt through the observation of the

linguistic behavior of one's elders (vrddha-vyavahāra) .42

Just as we observe, in the linguistic practice of our elders,

that individual words are each invariably associated with a

single, primary, literal meaning, we observe, similarly, that

4lAbhidhāvrttamātrkā, p. 24.

42Ibid., p. 25.

330

these words, occasionally and under specific circumstances,

may be used to convey other meanings related to their primary

senses. We can correctly use or understand a word in a

particular figurative sense only if we have either seen that

word itself so used or seen other, similar, words used to

convey similar figurative senses under similar circumstances.

Where no precedent of either kind is found, figurative usage

will not be possible and, if attempted, will fail to

effectively convey the desired sense.

This analysis of the role of precedent in justifying

particular figurative usages produces another threefold

division laksana, one which Mukulabhatta justifies by citing

a verse of Kumārilabhatta:

Because of their capacity for implication, some kinds of laksana are conventionalized, functioning just like direct expression; others may be created others cannot be, because of their incapacity.43 now; still

Figurative meanings which are common and thoroughly

conventional can be learnt in direct association with the

words that convey them, just as literal meanings are learnt;

an example is the use of the word rajan to refer to non-

ksatriya rulers.44

43Ibid., p. 25 -- cf. Tantravarttika on Mīmāmsāsūtra 3.1.12 (Ānandasrama Sanskrit Series 97 [second ed.], Vol. 4, p. 46).

44That such usage is conventional and learnt in the same

figurative. manner as literal usage does not make it any the less Mīmāmsakas (and presumably for Mukulabhatta) the association between words and their literal For the

meanings is natural and not the product of arbitrary convention (either human or divine). The word rajan has such a natural connection only with the meaning "ksatriya"; if one

331

To exemplify figurative usage which does not depend in

this way on convention, but may be newly created by a given

speaker, he quotes the following verse:

The clouds, with cranes wheeling against them, smear the sky with dark, wet splendor; the winds contain droplets of water; the joyful cries of the friends of the clouds [i.e. the peacocks] are sweet. All right, let them be so! I am Rama, whose heart is very hard; I can endure all! But how will Sīta survive? Alas, my queen, be strong! 45

This is the first literary example of laksana cited by

Mukulabhatta, and he has pointedly chosen a verse already

mentioned by as Anandavardhana as an example of dhvani.

Mukulabhatta notes and discusses three terms in the verse

which are used figuratively: "smear", "friends", and "Rama".

The first two cases are relatively uncontroversial, and

neither is discussed by Anandavardhana. The word "smear"

cannot here be understood literally; the dark color of the

clouds partially conceals the sky in much the same way as a

viscous, smearable substance partially conceals the surface to

which it is applied and, on the basis of this qualitative

similarity, the clouds are said, metaphorically, to "smear"

the sky. Friendship is possible only between sentient beings,

so the description of the peacocks as "friends" of the clouds

must likewise be understood metaphorically; peacocks wait

anxiously for the rain-clouds to come and rejoice when they

uses it in any other sense, one is speaking figuratively,. whether one realizes this or not.

45Abhidhāvrttamātrka, p. 27. Cf. above, Chapter Five, p. 200.

332

arrive and their behavior thus resembles that of friends. Our

capacity to understand such metaphorical expressions is not

dependent on convention; we need never have actually seen the

words "smear" and "friend" used in this way to understand the

desired meanings in this verse. It is enough if we have, in

the linguistic practice of our elders, seen other words used

to designate things qualitatively similar to their ordinary

referents, when their literal meanings are for some reason

inapplicable.

Mukulabhatta argues that, like "smear" and "friends", the

word "Rama" is here used figuratively. Here he directly

challenges Anandavardhana, who gives the use of the word

"Rama" in this verse as an example of arthantarasamkramita-

vacya dhvani. 46 Anandavardhana argues that the use of the

word "Rama" in this context conveys its own meaning (the

person designated by the proper name Rama) transformed

(parinata) into a set of suggested properties. The name is

used not to identify the speaker (which is already known from

context) but to suggest the great sufferings Rama has endured

in the past (the loss of his kingdom, exile in the forest,

etc.).

Mukulabhatta is in agreement regarding the meaning

ultimately conveyed here by the word "Rama"; his objection is

only to the characterization of the process by which we

understand this meaning as suggestion. This process, he

46Cf. Dhvanyaloka, Kashi ed., pp. 167-9.

333

argues, is essentially the same as that by which we understand

the figurative meanings of "smear" and "friends". The word

"Rama" cannot here be understood to convey its ordinary

meaning, not because the person referred to is not Rama, but

because his identity is already known. If the word were

understood literally, it could serve no communicative function

whatsoever; the basis for the inapplicability of the literal

sense in this case differs from that in the first two

examples, but it is nevertheless inapplicable, and this

inapplicability remains an essential prerequisite for the

conveying of any non-literal sense. This inapplicability of

the literal sense makes us look for a related figurative sense

which would be contextually applicable; because Rama proclaims

his own name in support of his contention that he can endure

the suffering of separation from Sita, the use of the name

"Rama" brings to mind those properties of Rama which

demonstrate his capacity to endure suffering. Thus our

understanding of the so-called "suggested" meaning in this

instance can be explained as simply another case of figurative

expression. As in the two previous cases, we need not

actually have ever seen this specific word used in this

figurative sense to understand its use here. It is enough if

we have seen other words used similarly. It is only by

extrapolating from cases of figurative usage we have actually

observed that we can, as speakers, create new figurative

expressions and, as hearers, properly comprehend such newly

334

coined expressions.

The requirement that newly created figurative expressions

must be based in this way on extrapolation from known cases

limits the scope of figurative usage, and accounts for the

third class of figurative expressions mentioned by Kumārila:

those which, because they cannot properly convey their

intended meaning, cannot be newly created by speakers.

Mukulabhatta gives the following example:

[The city of Dvärakal which, reddening the quarters of the sky in the middle of the ocean, shining like a land of gold, divided and illuminated the water like the flame of the oblation-bearer which is the face of a stallion's beloved.47

The coastal city of Dvaraka [Krsna's home] illuminates the

surrounding oceans with its golden sheen and, in doing so, is

compared with the legendary mare's-head fire

(vadavāmukhāgni) .48 The expression by which the poet refers

to this fire, "the oblation-bearer which is the face of a

stallion's beloved", is regarded by Mukulabhatta as an attempt

at figurative usage which goes beyond what linguistic

precedent will allow for and thus fails to effectively convey

its intended meaning. The fire used in the Vedic sacrifice is

thought to convey burnt offerings to the gods and is thus

47Abhidhāvrttamātrkā, p. 29. madhyesamudram kakubhah pisangīr yā kurvatī kāncanabhūmibhāsā/ turangakāntānanahavyavahajvāleva bhittva jalam ullalāsa// (Māgha, Sisupālavadha 3.32)

48A tremendous, undying flame, in the shape of a mare's head, which is currently submerged beneath the sea but will one day emerge to destroy the world -- for the story of its origin, cf. Mahabharata 1.169-171 (Poona ed.) .

335

called "oblation-bearer", a term which is commonly applied by

extension to fire in general. This particular fire has a

shape like that of the head of a female horse, an animal which

would typically be loved by a male horse. Thus we can see,

through a series of several stages, that the poet means by

this expression to refer to the mare's-head fire, although it

does not bear oblations and is not the actual head of a mare

beloved by an actual stallion. But this mode of designation

is simply too indirect and, according to Mukulabhatta, such

usage is not sanctioned in established linguistic practice

(such judgements, in this and other specific cases, will

necessarily be somewhat subjective, but the general principle

is clear enough) .49

So, although figurative meanings need not, like literal

ones, be learnt in direct association with the words that

convey them and it is possible to use and understand

figurative expressions not seen previously, the requirement

that such new expressions be formed on the analogy of

figurative usage as we have observed it in the linguistic

practice of our elders (vrddha-vyavahāra) places some limits

Vallabhadeva, in his commentary on this verse, notes that the 49Apparently referring to Mukulabhatta's criticism,

use of the term "oblation-bearer which is the face of a stallion's beloved" to refer to the mare's-head fire is "beyond what is desired by those who know alamkaras [or Alamkāraśāstra]" (cf. Śisupalavadha with the commentary of Vallabhadeva, ed. Ram Chandra Kak and Harabhatta Shastri, Bharatiya Book Corporation, Delhi, 1990, p. 106 -- I emend the [unintelligible] printed text "alamkaravidaticchaya" to read "alamkāravidatīcchayā") .

336

on the use of figurative language, and averts the threat of

linguistic chaos mentioned earlier whereby every word would

invariably, along with its fixed literal meaning, figuratively

convey all possible related senses. By observation of vrddha-

vyavahara we learn not only the fixed literal meanings of

words, but the particular contextual conditions under which

these words may be used to figuratively convey meanings

related to their literal meanings in particular ways.

Conditions of Laksana

Having established generally that our ability to

successfully use and understand figurative expressions depends

on, and is limited by, our observation of the use of such

expressions in vrddha-vyavahara, it remains to specify the

necessary conditions for figurative expression which are to be

deduced from such observation. Like Anandavardhana, and like

the Mimamsakas before him, Mukulabhatta believes that there

are three conditions which are essential for any figurative

expression: the impossibility of taking the word or phrase in

question in its literal sense, some relation between the

literal and the intended figurative sense, and some motive on

the part of the speaker for adopting an indirect mode of

expression.50

The first of these conditions, the impossibility of the

literal sense, is understood very broadly by Mukulabhatta, and

this expansive understanding of the factors which make it

50Abhidhāvrttamātrkā, p. 50.

337

impossible to understand an expression literally plays a

crucial role in his critique of Anandavardhana. In many of

the standard examples of figurative usage, the inability to

interpret an expression literally results from some logical or

physical impossibility -- one cannot simultaneously be a human

being and a cow, so both of the terms in the expression "The

Punjabi is a cow" cannot be understood literally; a village

must rest on some solid surface, so the word "Ganges" in "A

village on the Ganges" cannot be taken to designate its

literal referent, a particular river. But we have already

seen that such logical or physical impossibility is not

essential. In the example discussed above, "I am Rama, whose

heart is very hard; I can endure all!", the word Rama cannot

be understood literally, not because of any logical difficulty

in doing so (the speaker is, in fact, properly designated by

the name "Rama"), but because the speaker's intent in using

the term will not be comprehensible if it is taken literally.

The act of speech presupposes some communicative intention on

the part of the speaker; here the word "Rama", taken

literally, can fulfill no such communicative intention (as the

speaker's identity is already known), and it is this

intentional impossibility which leads us to look for some

meaning beyond the literal.

Thus, the inability to understand an expression literally

need not arise from a logical impossibility in the content of

the expression; any aspect of the communicative process which

338

cannot be properly understood if the expression is taken

literally can force us to interpret an expression

figuratively. For Mukulabhatta, the need to interpret an

expression figuratively may arise from factors related to the

speaker (vaktr), the sentence (vakya -- i.e. the words which

make up the expression), and the directly expressed meaning

(vācya) .

The first of these three is exemplified as follows:

Hey, neighbor-woman! while. Keep an eye on our house too for a The father of this child generally won't drink tasteless well-water, so I am going alone from here to the river where it is thick with tamala trees; let the dense clusters of reeds with their hard, broken ends scratch my body. 51

The unstated meaning we are to understand from this verse is

that the speaker is going to the river to meet secretly with

her lover; the mention of the need to fetch river-water is

meant to conceal her actual intention to meet her lover, and

the mention of the scratching of her body by reeds is meant to

conceal the actual source of the scratches which will be made

during her love-making. In each case, the false statement

causes us to understand, figuratively, the truth it is meant

to conceal. "This concealment," says Mukulabhatta, "is

understood through reflection on the fact that the speaker is

s1Ibid., p. 31. drstim he prativesini ksanam ihāpy asmadgrhe dāsyasi prāyo naiva sisoh pitā 'sya virasāh kaupīr apah pāsyati/ ekākiny api yāmi tad varam itah srotas tamālākulam nīrandhra vapur alikhantu jarathaccheda nalagranthayah// (Subhasitaratnakośa 807 -- ascribed to Vidyå)

339

a woman of bad character. "52 Reflection on the character of

the speaker leads us to conclude that her statements are meant

to conceal her real intentions and thus causes us to imagine

some state of affairs these statements could plausibly meant

to conceal. It is only our awareness of the bad character of

the speaker (itself derived from familiarity with the

conventions of Sanskrit verse) which leads us to look beyond

the surface meaning for some hidden truth. There is no

difficulty in understanding either the construal of the words

of the sentence (väkya) or in the meaning they express (vācya)

which would, in the absence of such knowledge regarding the

speaker, lead us to look for any sense beyond the literal one.

The second type of impossibility which leads us to

interpret an expression figuratively, that which relates to

vakya, is illustrated by the following verse:

"Having already obtained Śri, why would he again produce in me the pain of churning? I cannot imagine for him,

52Ibid., p. 32. How exactly do we know the speaker to be a woman of bad character? This seemingly crucial question is not dealt with at all by Mukulabhatta (nor, as far as I can determine, by other Ālamkarikas in their treatment of similar examples). The verse in question is not part of any larger work, so there is nothing outside of the verse itself to give us information regarding the character of the speaker. Our knowledge of the character of the speaker in such cases would seem to depend on an awareness of genre-convention. Verses describing the efforts of deceitful women (almost invariably married village-women) to arrange secret meetings with their lovers represent a substantial genre of Sanskrit and Prakrit muktakas (single verses). In this verse, we recognize a scenario familiar from other verses of this genre -- a married village-woman going alone to a secluded spot -- and it is our recognition of the verse as belonging to this genre that lets us understand that the speaker is a woman of bad character and leads us to look for a hidden meaning in her words.

340

his mind free of lassitude, his former sleep. Why should he build a bridge again, being attended by the lords of all the islands?" The ocean, when you approach, appears to tremble, as if entertaining such doubts.53

The same verse was given by Anandavardhana as an example of

alamkara-dhvani; Mukulabhatta's concern is, of course, to show

that the process by which we understand alamkaras which are

not directly stated is one of laksana, rather than any kind of

"suggestion".

In the verse quoted he finds several unstated alamkāras

beyond the unstated rūpaka (the metaphorical identification of

the king addressed in the verse with Visnu) dealt with by

Ānandavardhana. The movement of the ocean is identified with

"trembling" (actually a property of animate creatures only),

and this "trembling", because doubt is mentioned as its cause,

is further identified with the shaking of the head seen in

people beset by doubts. The identification of things which

are not truly identical is recognized as one of the sub-types

of atiśayokti (hyperbole), and thus both of these

identifications should be seen as instances of this alamkāra.

Both should be seen as instances of laksana as well. In each

case there is the superimposition of one thing on another, on

the basis of qualitative similarity between the two, and in

both cases the distinction between the thing superimposed and

the object of superimposition is suppressed; both thus fall

within the category of upacāra-misrā laksana which is gauna

s3Ibid., p. 33. Cf. above, Chapter Four, p. 129.

341

and takes the form of adhyavasana (like the case of rājan

mentioned earlier).

The verse contains a directly expressed alamkāra, namely

utpreksa -- the attribution to one entity of qualities or

actions properly belonging to another: in this case the

statement that the ocean looks "as if [it were] entertaining

such doubts". This utpreksa depends on yet another unstated

hyperbole, the false identification of the ocean with a

sentient being capable of entertaining doubts.54 Again, the

distinction between the two is suppressed and their

identification rooted in similarity, so this unstated

atisayokti, like the first two, may be classified as a case of

upacāra-misra laksana which is gauna and takes the form of

adhyavasāna.

The fourth and final unstated alamkāra Mukulabhatta finds

in this verse is the metaphorical identification (rūpaka) of

the king being addressed with Visnu (offered by Anandavardhana

as an example of rüpaka-dhvani). Each of the three "doubts"

which the ocean is fancifully said to entertain involves the

consideration and dismissal of the possibility that the king

will "again" perform some action formerly performed by Visnu --

Visnu obtained his consort Śri [the personification of royal

54In support of his view that the figure utpreksā, here and elsewhere, requires some element of hyperbole, Mukulabhatta (p. 34) quotes Udbhata's definition of utpreksā

describes (Kavyalamkarasarasamgraha, 3.3 [Banhatti's ed., p. 46]), which the (atisayānvitā) . figure as "joined with hyperbole"

342

prosperityl by churning the ocean, he sleeps on the ocean

during the interval between each cosmic dissolution and re-

creation, and, incarnated as Rama, he built a bridge across

the ocean to the island of Lanka, where his wife Sita was

being held prisoner by the rāksasa-king Rāvana. These

"doubts" could properly be entertained only with reference to

Visnu, but are here entertained with respect to a mortal king;

this can be accounted for only if we understand the king to be

identified (at least in the "mind" of the ocean) with Visnu.

Mukulabhatta agrees with Anandavardhana in viewing this

implied identification as an unstated rūpaka; as usual, it is

only with regard to the process by which we understand this

figure that he differs. The doubts which the ocean is said to

entertain are properly applicable only to Visnu, but Visnu is

nowhere mentioned in the verse and there is no way to connect

these doubts with him; this apparent inconsistency in the

expressed meaning of the verse is explicable only if we

understand the king to be identified with Visnu. Because our

understanding of this unstated rūpaka arises from an

inconsistency in the expressed meaning of the verse, it may be

explained as a case of laksana. Like the three unstated

alamkāras discussed earlier, this rūpaka is conveyed by a

process of upacāra-mirā laksanā which is gauna (because the

identification of the king with Visnu is based on their

similarity) and takes the form of adhyavasāna.

Thus all of the unstated alamkäras found in this verse

343

can be explained as cases of laksanā. In each case, our

understanding of the figuratively expressed alamkāra arises

from the impossibility of intelligibly construing the words of

the sentence if they are taken literally.55 The mention of

trembling and doubting in connection with the non-sentient

ocean, and of the entertaining of Visnu-specific doubts with

respect to one who is not Visnu, cannot be properly understood

unless we interpret these expressions figuratively in the

manner already described. Thus, the literal impossibilities

which lead us to find figurative meanings in this verse are

connected with the sentence (vākya).

The third and final type of inconsistency which may lead

us to seek for some meaning beyond that which is directly

expressed is that which depends on the directly expressed

meaning (vacya) itself. Mukulabhatta gives the following

example:

The arrows of the love-god cannot be deflected; the spring-season spreads out in every direction; the moon's rays produce madness in my heart; the cuckoos are enchanting; this youthful stage of life is borne with difficulty, due to the weight of these swelling breasts. O friend! How can I now endure these five unbearable fires?56

Here, the unbearability of a variety of factors generally

associated with the production and intensification of love

55Ibid., p. 34.

56Ibid., p. 37. durvāra madanesavo diśi diśi vyājrmbhate mādhavo hrdy unmādakarāh śaśānkarucayaś cetoharāh kokilāh/ uttungastanabhāradurdharam idam pratyagram anyad vayah sodhavyah sakhi sāmpratam katham amī pañcāgnayo duhsahāh//

344

(emphasized by the metaphorical identification of these

factors with fires) would be inexplicable if we did not

conclude that the speaker was suffering from separation from

her beloved; thus, the need to account for an apparent

implausibility in the expressed meaning of the verse produces

an awareness of the rasa vipralambha-srngara (love in

separation). Because this apparent implausibility leads us to

understand a second, unstated meaning which does not displace

(but, on the contrary, justifies) the directly expressed

meaning, this example falls within the category of upadāna

laksanā (cf. above, pp. 320-322) .

While Mukulabhatta differs with Ānandavardhana in

explaining what he would call rasa-dhvani as a kind of

laksana, his remarks on this verse indicate that he generally

shares Anandavardhana's view of the aesthetic centrality of

rasa. "Although this vipralambha-srngara is only indicated

figuratively, it is predominant with respect to the directly

expressed meaning, by reason of its being delightful to the

hearts of people of taste [sahrdayas]."57 This is the only

occasion on which Mukulabhatta discusses the relative

predominance of an expression's literal and non-literal

meanings. The question of predominance has no real bearing on

his primary task: to show that non-literal meanings, whether

they are predominant or not, can be understood only through

the process of laksanā. Nevertheless, this brief comment

57Ibid., p. 38.

345

strongly suggests that he accepts, or at least has no quarrel

with, Anandavardhana's claims for the primary importance of

rasa in poetry.

At least one of these three types of impossibility or

surface inconsistency (relating to vaktr, vākya, or vācya) in

the literal meaning of a verse or sentence must be present if

we are to understand any additional, figurative meaning from

it. In illustrating laksana based on these three types of

surface inconsistency Mukulabhatta has deliberately chosen

examples which would on Anandavardhana's theory be regarded as

cases of vastu-dhvani, alamkara-dhvani, and rasa-dhvani,

respectively. It is not clear that Mukulabhatta wishes to

claim that all vastu-laksana depends on inconsistency related

to the speaker, all alamkara-laksana on that related to the

sentence, and all rasa-laksana on that related to the

expressed meaning. Nevertheless, he has shown that examples

of all three types of dhvani discussed by Anandavardhana can

be explained as cases of laksana. This is important, in that

a major element of Anandavardhana's defense of the distinction

between dhvani and laksana was his contention that no one

could reasonably argue that alamkāras or rasas could ever be

conveyed by laksana (cf. above, Chapter Five, pp. 236-237).

The second of the three conditions necessary in order for

any figurative expression to be properly understood is that

there must be some relation between the literal meaning of the

expression and the figurative sense it is meant to convey. If

346

any relation at all, however tenuous, were to be accepted as

sufficient for figurative expression to operate, we would

again be faced with a situation in which any given term could

figuratively convey virtually any sense and it would be

practically impossible to understand the particular figurative

sense intended by any expression. It is therefore necessary

to limit and specify the sorts of relation between literal and

figurative senses which will allow laksana to function

properly. Several lists of such possible types of relation

are found in earlier philosophical treatments of laksana. 58

Mukulabhatta accepts five possible types of relation, and

quotes, as authority for this view, the following verse of the

Mīmāmsaka Bhartrmitra: 59

Laksana is considered to be of five kinds: arising from connection [sambandha ] with the denoted meaning, from similarity [sādrsya], from inherence [samavāya], from contradiction [vaiparitya], and from connection with an

58Cf. Mimāmsasūtra 1.4.23, Nyayasūtra 2.2.62, Mahābhāsya on Pānini 4.1.48.

59This is one of only two known direct quotations from Bhartrmitra's works (the other is in Cakradhara's commentary on Jayantabhatta's Nyayamamjari [Sivakumarasāstri Granthamala 5, Vol. 1, p. 298]). He is known as the author of a work called Tattvasuddhi (cf. ibid., and also Umbeka's Tātparyatīka on Kumarila's Ślokavarttika [University of Madras, second ed., 1971, p. 3.]). Kumārila and Jayantabhatta both criticize his views on the nature of sound (cf. Slokavarttika, Sabdanityatā 130-132 [ed. Svāmi Dvārikādāsa Sāstrī, Varanasi, 1978, p. 540] and Nyayamañjari, op. cit., pp. 298 and 317), and Kumārila accuses him of having made Mīmāmsa into a materialist philosophy (cf. Slokavarttika, Pratijñasutra 10 [Univ. of Madras ed., p. 3]).

60Including, but not limited to, physical contact.

347

action [kriyāyoga] .61

Laksana based on relations of contact and similarity can be

illustrated, respectively, by the familiar examples "A village

on the Ganges" and "The Punjabi is a cow". In the first case,

the word Ganges refers not to the river itself but to the bank

adjacent to it; in the second, the word "cow" refers not to a

cow but to a person thought to be qualitatively similar to a

cow.

of laksanā base As an example based on a relation of

inherence, Mukulabhatta gives the expression "The umbrella-

carriers are going. "62 Here the word "umbrella-carriers"

(chatrinah) is used to refer to the entire retinue of a king,

even though only one of them carries the umbrella (a symbol of

royal power, to be held over the king). Because there can be

only one umbrella-carrier, the word, here used in the plural,

cannot be understood literally; we therefore understand it to

refer to the group of royal followers to which the umbrella-

carrier belongs, even though the rest are umbrella-less. In

laksana based on contradiction, a word is used ironically to

indicate the opposite of its literal meaning; for example, use

61Abhidhāvrttamātrkā, p. 50. This verse, like many of Mukulabhatta's citations from Mimamsa works, becomes a regular fixture of discussions of laksana in later Alamkārasāstra texts. It is quoted by Abhinavagupta (Locana, Kashi ed., p. 28), Mahimabhatta (Vyaktiviveka, Kashi Sanskrit Series 121, p. 119), and Sahadeva (Kavyalamkarasutratippana, op. cit., p. 68), among others.

62Abhidhavrttamatrka, p. 54. The example is drawn from

p. 326) . Śabara -- cf. Mimamāsūtrabhāsya on 1.4.23 (op. cit., Vol. II,

348

of the term "agreeable" (bhadramukha) to refer to a person

known to be disagreeable.63

In laksana based on "connection with an action"

(kriyayoga), as Mukulabhatta explains it, the figuratively

expressed meaning is not connected directly with the literal

meaning, but rather with the word or words which convey this

meaning. He gives the following example:

You are Prthu by reason of your good qualities, Rama by reason of your fame; you are Nala, you are Bharata/you are not addicted to greed; in a great battle, you are Satrughna/you are a killer of enemies; in the world, you are Janaka because of your stability/you are the producer of stability; since, by your great deeds, you bear the names of ancient kings, why are you not Mandhatr/why will you not support me, even though you lord,

have conquered the three worlds?44

Here, in the expression mahati samare satrughnas tvam ("In a

great battle, you are Satrughna/you are a killer of enemies")

the king being eulogized is called by the proper name

"Satrughna", not because of any similarity to the person

ordinarily designated by this name, 65 but because of his

connection with the etymologically recoverable sense of this

63Ibid., p. 56.

64Ibid., p. 56. prthur asi gunaih kīrtya rāmo nalo bharato bhavān mahati samare śatrughnas tvam ksitau janakah sthiteh/ iti sucaritaih khyatim bibhrac cirantanabhūbhrtām katham asi na māmdhātā deva trilokavijayy api// The verse is a variant of Subhasitaratnakosa 1417, where it is attributed to one Viryamitra (known only from this and other poetic anthologies). Udbhata quotes the verse (with the same variant reading) as an example of the figure slesa (cf. Udbhata's Commentary on the Kavyalamkara of Bhamaha, p. 37) .

65Satrughna was one of Rama's younger brothers.

349

name. This king kills his enemies in battle, and it is on the

basis of this action that he is identified with Satrughna,

whose name means "enemy-killer". Mukulabhatta's discussion of

the verse is confined to the figurative use of the name

Satrughna, but he would presumably explain its punning use of

the names of other kings in the same way.

Indeed, it would seem that Mukulabhatta wishes to explain

all double-meaning expressions (both the alamkāra slesa and

what Anandavardhana would call sabdasaktimula-dhvani -- cf.

above, Chapter Four, pp. 166-172) as instances of this type of

laksana. "In laksana based on connection with an action the

figuratively understood meaning is rooted in the power of word

[sabdasaktimula], since it follows from the expressive power

of the parts of the word or words [used figuratively] . #66 In

any expression conveying a double meaning, the parts of the

polysemic word or phrase must convey meanings apart from the

literal meaning of the word or phrase as a whole. It seems

odd to describe the basis of such figurative usage as

"connection with an action", but if the concept of "action"

(kriya) is understood broadly enough to include being and/or

becoming, this will pose no problem, as the second meaning

understood from the parts of the polysemic word or phrase can

always be considered an action -- obviously, Mukulabhatta's

choice of the term kriyayoga is dictated by his acceptance of

66Abhidhavrttamatrka, p. 65.

350

Bhartrmitra's typology. 67

In addition to inconsistency in the literal meaning of an

expression (related to vaktr, vakya, or vacya) and connection

between this literal meaning and the intended figurative

meaning (of one of the five types just mentioned), there is a

third condition necessary for any expression to be understood

figuratively: that there should be some motive (prayojana) for

the adoption of a figurative mode of expression rather than a

literal one.8 This motive, according to Mukulabhatta, may

be of two kinds.69 The first type of motive is mere

conformity to conventional usage (rūdhi), as in the case of

the word rajan ("king"). The word rajan is very frequently

applied to non-ksatriya rulers; even if one realizes that it

can be literally applied only to ksatriyas, one may, in

conformity with established convention, use the word

figuratively to refer to non-ksatriyas. In such cases,

conformity to convention is itself sufficient motive for

resorting to figurative usage.

The second possible motive for resorting to figurative

67How Bhartrmitra himself understood the category of laksana based on kriyayoga, of course, we have no way of knowing.

68Like Anandavardhana (and unlike the Mīmamsakas), Mukulabhatta understands the prayojana required in cases of laksana as a purpose on the part of the speaker for choosing an indirect, non-literal mode of expression, rather than as a contextually useful application for the figuratively understood meaning (cf. above, Chapter Three, pp. 111-112 and Chapter Five, pp. 227-230) .

69Ibid., p. 50.

351

usage is the intention to communicate some special aspect of

the thing figuratively described, which could not be

effectively conveyed literally. In the example "I am Rama,

whose heart is very hard; I can endure all, " the word "Rama"

figuratively connotes not merely Rama's general capacity to

endure suffering, but the many particular sufferings he has so

far endured: the loss of his kingdom, his exile in the forest,

the abduction of Sita, and so on. To list explicitly and

exhaustively all the sufferings Rama has endured throughout

his life, even if it were possible, would be tedious and

aesthetically ineffective. By figurative use of the word

"Rama", the poet is able to allude, in a compact and

aesthetically effective way, to the full range of sufferings

Rama has been able to endure. It is the need to convey in

this succinct manner a great many aspects of Rama's experience

and character which serves as the motive for figurative usage

in this case.

Mukulabhatta finds a similar motive in the familiar

example of laksana, "A village on the Ganges" (gangayam

ghosah) . "Here, the motive for the laksanā is to make known

that the bank possesses properties such as purity,

delightfulness, etc. which are not commonly known to belong to

it, and which inhere in the same locus as Ganges-ness."70

Here Mukulabhatta appears to contradict his earlier analysis

of this expression. "A village on the Ganges" was given as an

70Ibid., p. 53.

352

example of suddha ("pure") laksana (cf. above, pp. 318-319);

the transference of properties from the literal to the

figurative referent of the term employed figuratively is

precisely what such "pure" laksana is meant to be pure of. If

the use of the term "Ganges" to refer to the bank of the

Ganges causes us to understand that certain well-known

properties of the river belong to its bank as well, the

expression "A village on the Ganges" cannot legitimately be

regarded as a case of suddha laksana; it involves the partial

superimposition of the nature of the Ganges upon its bank and

should thus be classified as a case of upacara-misra laksana.

We have seen (above, pp. 326-327) that, for Mukulabhatta,

expressions such as "A village on the Ganges", while typically

to be explained as cases of suddhā laksana, may also (if our

understanding of the figurative referent is meant to be

"colored" by the literal referent of the term employed

figuratively) be considered as cases of upacāra-miśrā laksanā

(based on either aropa ["superimposition"] or adhyavasāna

["absorption"] ). In order save Mukulabhatta from apparent

self-contradiction we may suppose that his explanation of the

motive underlying the figurative usage in the expression "A

village on the Ganges" should apply only when it is meant as

an example of upacāra-misrā laksanā. But this explanation

leaves an important question unanswered: what exactly is the

motive for figurative expressions such as "A village on the

Ganges" when they are meant as examples of suddhā laksana?

353

Ānandavardhana discusses two separate types of figurative

usage which, due to their conventional character, serve no

special aesthetic function: those recognizably figurative

expressions which are used in conformity with established

usage (such as the use of the word "tells" in "This bed of

lotus petals tells of the affliction of this emaciated

woman") ,71 and those which are so thoroughly conventionalized

that they are not generally recognized as figurative at all

(such as the use of the word lavanya [literally, "saltiness"]

in the sense of "beauty") .72 Does Mukulabhatta mean to

include both of these types under the heading of figurative

expressions motivated by convention (rudhi)? All the examples

he gives of conventionalized laksanā (rājan, pañcalah, etc.)

are of the second type (where a particular word is

conventionally associated with a particular non-literal sense,

and where the figurative character of the usage is not

immediately apparent). If expressions of the first type are

also accepted as cases of laksana motivated by convention, "A

village on the Ganges", when meant as an example of suddhā

laksana, might be regarded as falling into this category;

saying "on the Ganges" when one means "on the bank of the

Ganges" could be seen as merely a common manner of speaking

requiring no special motive apart from conformity to accepted

usage.

71Cf. Dhvanyaloka, p. 143 and above, Chapter Five, p. 224.

72Ibid., p. 146, and above, Chapter Five, p. 224-225.

354

Thus, the apparent self-contradiction in Mukulabhatta's

analysis of the motive element in the expression "A village on

the Ganges" could perhaps be explained. Mukulabhatta,

however, makes no attempt to provide such an explanation and,

indeed, shows no awareness of any apparent inconsistency.

Here, as elsewhere, his discussion is extremely terse; one is

sometimes left in the dark as to his specific understanding of

particular examples of figurative usage and even certain

elements of his own theory of figurative expression. His

analysis of the nature of prayojana and its role in the

process of figurative signification is certainly not as

thorough or clear as one might wish.

In any case, these three factors -- the impossibility of

the literal sense of an expression, a connection between its

literal and intended figurative senses, and some motive for

the speaker's resort to a non-literal mode of expression --

constitute the necessary preconditions for the operation of

laksana. We have seen that Mukulabhatta, in his discussion of

these conditions and his selection of examples to illustrate

their various aspects, has tried to lay the groundwork for the

inclusion of Anandavardhana's dhvani within the scope of

laksanā. He has shown how even poets' communication of

alamkaras and rasas without their being explicitly stated can

be explained within the framework of laksana, and has analyzed

several of Anandavardhana's own examples of dhvani as cases of

laksanā. He has yet, however, to argue directly for the

355

inclusion of dhvani in laksana or even to mention dhvani

explicitly. It is to the question of the relation between

figurative expression and Anandavardhana's dhvani that he

turns next.

The Inclusion of Dhvani in Laksana

Most of the key points in Mukulabhatta's attack on

Anandavardhana have already been established in his discussion

of the taxonomy of laksana and of its necessary conditions, so

when he finally turns to directly consider the question of

dhvani he does so very succinctly. In his discussion of

upadana-laksana he has established, contra Anandavardhana,

that figuratively understood meanings need not entirely

supplant, and may rather supplement, the literal meanings

which give rise to them. He has shown that alamkaras and

rasas, as well as narrative elements (vastus), can be conveyed

figuratively. He has thus already substantially undermined

Anandavardhana's arguments for the existence of suggestion as

a mode of signification distinct from figurative expression.

Mukulabhatta's primary concern, in his treatment of

Anandavardhana's dhvani, is to show how the subdivisions of

dhvani correspond to his own subdivisions of laksanā, in

particular his fivefold classification of laksana on the basis

of the type of connection between the directly expressed and

figurative senses of an expression.

Given this connection, based on five types of relation, which we have previously described, the directly expressed displacement sense is sometimes subjected to total [atitiraskāra], sometimes meant to be

356

expressed [vivaksita], and sometimes not meant to be expressed [avivaksital; [this discussion is undertaken] in order to show the respective scope of these three categories which have previously been shown by the Sahrdayas.73

"Sahrdayas" is the term by which Mukulabhatta (and

Pratiharendu-raja) refer to Anandavardhana and his followers.

We have seen (above, Chapter Five, pp. 198ff) that

Anandavardhana divides dhvani into two broad classes: that in

which the directly expressed meaning is not literally intended

(avivaksita-vacya), and that in which it is literally intended

but subordinated to some further suggested sense

(vivaksitānyapara-vācya) . The first type has two sub-

divisions: that in which the expressed meaning is totally

displaced (atyantatiraskrta) and that in which it is

transformed into another meaning (arthantarasamkramita -- what

Mukulabhatta here calls avivaksita-vācya) .74 It is these

three classes that Mukulabhatta wishes to consider in relation

to his own fivefold typology of laksana.

In laksana based on a relation of similarity (sadrsya) or

contradiction (vaiparitya), the directly expressed meaning

will invariably be totally displaced by the figurative sense

to which it gives rise. In the verse "The clouds, with cranes

wheeling against them, smear the sky with dark, wet

73Ibid., p. 58.

74 Anandavardhana' s sub-division of vivaksitanyapara-vācya dhvani (on the basis of the presence or absence of noticeable sequence between the expressed and suggested meanings) is of no interest for Mukulabhatta's purposes, and he ignores it here.

357

splendor ... " (above, p. 200), the figurative use of the words

"smear" and "friends" (in the expression "friends of the

clouds", referring to peacocks) depends on a relation of

similarity -- the partial obscuring of the sky by the clouds

resembles smearing, and the joyful behavior of the peacocks at

the coming of the clouds resembles that of friends. In each

case, the directly expressed meaning of the word used

figuratively is entirely displaced; there is no actual

smearing and no actual friendship involved in the situations

described. "Here, because the words 'smear' and 'friend'

refer to the things with which their own meanings are

compared, their own meanings are entirely unconnected with the

event [described in the verse] . "75 Likewise in laksana based

on contradiction: when the term "agreeable" (bhadramukha) is

ironically applied to a person known to be disagreeable, its

directly expressed meaning is totally displaced -- the

figuratively understood sense is precisely the opposite of the

literal. 76

In laksana based on relations of connection (sambandha)

or inherence (sama (samavaya) the situation is somewhat more

complicated. Laksana based on either of these relations may

be of two types, taking the form of either upādāna

("inclusion"), in which the figurative sense is understood in

addition to the literal one without replacing it, or laksana

75Ibid., p. 58.

76Ibid., pp. 58-59

358

("indication"), in which the figurative sense replaces, rather

than supplements, the literal one (cf. above, pp. 320-322).

Among these, in laksana which has the form of upadana, since, given the inclusion [of the figurative sense along with the literall, there is intent to express the directly expressed sense, it is said to be vivaksita; in such cases the Sahrdayas have shown, in the realm of poetry, that [the expressed meaning] is meant to be expressed while subordinated to another meaning [vivaksitānyapara]. But in cases of laksana the directly expressed meaning is not meant to be expressed [avivaksita], because it is transformed into another meaning [arthāntarasamkramita] .77

As an example of the upadana type he gives the expression

"Fat Devadatta doesn't eat by day." From this statement we

understand, without being told, that Devadatta eats at night;

this understanding (which is, for Mukulabhatta, a figurative

meaning) supplements but does not replace the literal meaning

we understand from the sentence. "Here an effect, namely

fatness -- qualified by the absence of eating during the day --

is itself meant to be expressed [vivaksita]; in order to

justify itself it indicates its cause, namely eating at night,

in a case of laksana based on connection [sambandha] .178

His example of the laksana type is "I am Rāma, whose

77Abhidhavrttamatrka (ed. Brahmamitra Avasthi, Delhi: Indu Prakāśana, 1977), p. 52. In his edition (p. 60) Dvivedi omits the words here translated as: "But in cases of laksana the directly expressed meaning is not meant to be expressed." This reading of the text is clearly mistaken -- it makes Mukulabhatta give "transformation into another meaning" (arthantarasamkramitatva) as a reason for regarding expressed meanings as vivaksita when, by his own reasoning as well as that of Anandavardhana, it would, on the contrary, be a reason for regarding it as avivaksita.

78Abhidhāvrttamātrka (Dvivedi's ed.), p. 60.

359

heart is very hard." In this expression the directly

expressed meaning of the word Rama is not totally displaced --

the speaker is, in fact, Rama -- but it is not what is meant to

be conveyed by the word in this context. "Here the directly

expressed meaning of the word 'Rama', namely the son of

Dasaratha, is not taken up for its own sake [svaparatvena],

since it is transformed into other suggested79 properties; it

is therefore not meant to be expressed [avivaksita], but is

not totally displaced [atyantam tiraskrtam], since it can be

in some way connected with the sentence-meaning via these

suggested properties. "80 The point of the speaker's

declaration that he is Räma is not to identify himself, but to

allude to the many sufferings he has so far had to endure in

life and thereby indicate his capacity to endure his

separation from SIta. In this and similar cases the directly

expressed meanings of the words used figuratively are neither

totally displaced nor are they what is meant to be expressed

by these words.

The example already given of laksana based on a relation

of inherence (samavaya), "The umbrella-carriers are going, "

79Mukulabhatta's use of the term "suggested" (vyangya) here is odd and seemingly rather careless, since he has already explained, in discussing this very example, that these properties are conveyed figuratively (laksita), and since he is currently arguing the total inclusion of Anandavardhana's dhvani for the scope of laksana, eliminating the need to accept "suggestion" as a distinct mode within

of verbal expression.

80Ibid., p. 63.

360

can, according to Mukulabhatta, be explained either as a case

of upadana-laksana or as one of laksana-laksana, and may thus

be understood as either vivaksita-vācya or avivaksita-vācya,

respectively. Its classification as one or the other depends

on whether we take the word "umbrella-carriers" (chatrinah) to

refer figuratively to the non-umbrella-carrying members of the

king's retinue or to the retinue as a whole. "For here, when

the word "umbrella-carrier" [chatrin], because of its

plurality, indicates even those [members of the retinue] who

lack umbrellas in order to justify its syntactic connection

with this plurality, then, since inclusion [upādana] of the

umbrella-less is being made on the basis of a relation of

inherence81 [samavaya], the directly expressed meaning, the

umbrella-carrier, is vivaksita. "82 The word chatrinah, on

this understanding, refers literally to the one person who

actually does carry an umbrella and also -- because the plural

suffix attached to it could not otherwise be understood --

refers figuratively to the many umbrella-less persons who make

up the rest of the king's retinue.

However, "when umbrella-carrier-ness, because its

connection with plurality would not otherwise make sense, is

taken as applying to the group [samudāya] then, because it is

the group that is meant to be expressed [vivaksital, the

81Since both the umbrella-carrier and his umbrella-less associates are parts of a greater whole, i.e. the retinue, their relation is said to be one of inherence.

82Ibid., p. 60.

361

directly expressed meaning [i.e. the single umbrella carrier]

is not meant to be expressed [avivaksita]. Even so, because

he is included within the group, the umbrella carrier too can

easily, via the group, be connected with the action [described

by the sentencel, and thus the expressed meaning is not

totally displaced.83 Thus the single example, " The

umbrella-carriers are going," can be taken as an illustration

of both upādāna- and laksana-laksana based on a relation of

inherence. If it is understood as upādana, the directly

expressed meaning of the word "umbrella-carrier" will be

vivaksita; if understood as laksana, this directly expressed

meaning will be avivaksita.

The fifth and final type of relation on which figurative

usage may be based is "connection with an action" (kriyayoga) .

As we have already seen, Mukulabhatta wishes to explain

expressions relying on punning or double entendre as cases of

figurative usage based on such a relation. In such punning

usage the directly expressed meanings of the word or words

used figuratively will sometimes be totally displaced and

sometimes meant to be expressed. In the example already given

of laksana based on connection with an action, "In a great

battle, you are Satrughna/ you are a killer of enemies," the

directly expressed meaning of the word satrughna is meant to

be expressed. "For here, the word satrughna, while causing us

to understand [the king's] agency in the act of killing

83Ibid., p. 63-64.

362

enemies by means of laksana based on connection with an

action, communicates its own meaning, i.e. the son of

Daśaratha, as a standard of comparison [upamāna] . #84 The

expression identifies the king being addressed with Satrughna,

and this identification can be understood only if the word

satrughna is taken to refer to both Satrughna and the enemy-

killing king. The literal referent is not replaced by the

figurative one, but is rather placed in a relationship of

comparison with it.

In this type of laksana, total displacement of the

literal sense will be possible only when the double-meaning

word or phrase need not be understood to convey both its

figurative and literal senses at once (as in the case of

satrughna) ; in practice this will happen only when the double-

meaning word or phrase is repeated. Mukulabhatta gives as an

example the expression "purusah purusah. "e5 The

illustration, taken out of context, is cryptic. It appears to

be a citation from the following verse (quoted in full by

Hemacandra) :

A hero who moves in battle, at the forefront, killing a great many enemies: that person [purusa] is like a purusa ["one who goes in front", "one who protects"] .86

84Ibid., p. 66.

85Ibid., p. 65.

%nighnann abhimukhah śūro 'nekaso bahuśah parān/ samgrāme vicaraty esa purusah puruso yathā// Cf. Hemacandra, Kavyanusasana, ed. Pandit Sivadatta, Kāvyamala 71, Nirnayasagara Press, Bobmay, 1901, p. 239. Since the words he quotes are unintelligible out of context,

363

Here, in order for the simile to be intelligible and non-

tautologous, the second word purusa must be understood to

convey not its literal meaning, "a human being", but a second,

figurative meaning -- one based on its etymologically

recoverable sense. The Unadisutra (IV.76-77) allows for two

possible derivations of the word purusa: the agentive suffix

kusan applied to the verbal roots vpr (meaning "to fill" or

"to protect") or Vpur (meaning "to go in front") will produce

the desired form.87 The meanings "one who goes in front" and

"one who protects" are both potentially applicable to a

military hero such as is described in the verse; it is not

clear from Mukulabhatta's remarks which he had in mind.88 In

any case, the first word purusa is meant to convey only its

literal sense and the second to convey only its etymologically

based figurative sense; the simile expressed in the verse

depends on the comparison of these two separate meanings. The

Mukulabhatta apparently assumes his audience will be familiar with the verse; it was very likely quoted in some earlier Alamkara work known both to him and to Hemacandra (probably Udbhata's Bhamahavivarana -- since the passage in which Hemacandra quotes the verse paraphrases a portion of Udbhata's work; cf. Gnoli, Udbhata's Commentary on the Kavyalamkara of Bhamaha, op. cit., p. 21) .

87Cf. Unadisutra with the commentary of Svetavanavasin, (Vol. I of The Unadisutras in Various Rescensions, ed. T.R. Chintamani, Madras University Sanskrit Series No. 7, Madras, 1933 [reprint Navrang, New Delhi, 1992]), pp. 155-156.

88Hemacandra is similarly unclear as to his understanding of the etymological sense required by the verse. He says only that (op. cit., p. 239) "The second word purusa, because it is meant to convey a suggested sense rooted in the power [of word], refers to a man of heroic acts [avadātakarman]."

364

verse involves a play on the meaning of the word purusa, but

each instance of this word ultimately conveys only one

meaning; the literal meaning of the second word purusa is thus

totally displaced by its figurative meaning. So laksana based

on "connection with an action", as Mukulabhatta understands

it, falls partly within the scope of Anandavardhana's

vivaksita-vacya category, and partly within that of

atyantatiraskrta-vācya.

In his analysis of the nature and conditions of laksana,

Mukulabhatta has already established that examples of all the

types of dhvani recognized by Anandavardhana can be explained

as cases of figurative usage, and thus that there is

significant overlap between Anandavardhana's dhvani and

laksanā; this examination of the correspondence between

Anandavardhana's categories of dhvani and his own fivefold

division of laksana aims to go beyond this by showing that

Anandavardhana's dhvani can be entirely subsumed within the

scope of laksanā. Mukulabhatta's treatment of the dhvani

question here is brief and schematic, as he himself

acknowledges:

This matter is not fully investigated here, as a great deal would have to be said [to investigate it fully]; but it has been mentioned here in order to show, in a general way, that dhvani, which has been described by the Sahrdayas as if it were something new, falls within the scope of laksana. Learned people should carefully examine this [argument], and not summarily dismiss it.89

Given that the express purpose of the Abhidhavrttamātrka is to

89Ibid., p. 66.

365

provide a concise analysis of the literal and figurative

semantic modes, a thorough and full-blown attack on the dhvani

theory (of the sort later made by Mahimabhatta) would be

beyond the scope of the work. What is presented here is more

the outline of a possible response to the theory of suggestion

than a detailed critique of it.

Mukulabhatta's plea that readers should not casually

dismiss his argument for the inclusion of dhvani within

laksana would seem to indicate that Anandavardhana's theory of

suggestion had already obtained a wide degree of acceptance

(at least among Alamkarikas) by this time; he apparently feels

that his attack on the existence of suggestion as distinct

from figurative expression may be rejected without receiving

the attention it deserves.

Although Mukulabhatta's work presents an attack on

Dhvanyaloka, this attack is very narrowly focussed and is in

fact confined to a single point, namely Anandavardhana's

insistence that dhvani cannot be identified with figurative

usage. He has no quarrel with other aspects of

Anandavardhana's theory: the general importance of non-literal

meanings in poetry, the threefold division of such meanings

(into vastu, alamkāra, and rasa), and the paramount importance

of rasa in aesthetically effective poetry. He explicitly

acknowledges the predominant role of rasa in producing

aesthetic pleasure for the audience of a poem (cf. above, pp.

344-345), and shows, through his analysis of various example

366

verses, that all three types of non-literal meaning dealt with

by Anandavardhana can be conveyed by laksana.

Mukulabhatta's argument with Anandavardhana is thus

carried on within a broad framework of agreement regarding the

nature of poetry and the role of non-literal meaning in the

production of aesthetic pleasure. His work defends the

adequacy of existing philosophical theories of non-literal

expression against Anandavardhana's attack, but this "defense"

takes the form of a demonstration that Anandavardhana's

teleological, rasa-centered model of aesthetics can be

accepted and accounted for within the framework provided by

these theories; that this model of aesthetics ought to be

accepted and accounted for, and that any analysis of non-

literal usage, in order to be accepted as valid, must be able

to account for it, seems to be taken for granted.

Mukulabhatta quarrels with certain aspects of Anandavardhana's

theory of poetic semantics, but apparently accepts without

reservation his understanding of the tests such a theory

should be required to meet and the range of phenomena it

should be expected to explain.

Pratīhārendurāja

Mukulabhatta's pupil, PratIharendurāja, has a similar

quarrel with Anandavardhana. As with Mukulabhatta,

PratIharenduraja's primary object is to defend earlier

language theory -- in this case literary rather than

philosophical -- against Anandavardhana's charges of

367

inadequacy, rather than to attack Ānandavardhana's

teleological, rasa-centered model of literary aesthetics.

Since he names Mukulabhatta (c. 875-925) as his teacher,

PratIharenduraja's work should date from the early- to mid-

tenth century. His only known work is a commentary, entitled

Laghuvrtti, on Udbhata's Kavyalamkarasarasamqraha. Some

scholars have argued that he should be identified with

Bhattenduraja, whom Abhinavagupta names as one of his teachers

in the field of Alamkārasāstra.90 The terms Bhatta and

PratIhara are both honorific titles, here prefixed to the same

personal name, Induraja. Abhinavagupta's works date from the

late tenth and early eleventh century, so his teacher's dates

would at least roughly correspond with those of

PratIharenduraja, raising the possibilty that these two

Indurajas are one and the same person. Kane, De, and

Krishnamoorthy reject the identification (rightly, I think) ,

partly on the basis of discrepancies between the views of

Bhattendurāja (as reported by Abhinavagupta) and

PratIharenduraja and partly because the honorific Bhatta is

applied consistently to Abhinavagupta's teacher and not to

90Cf. K.C. Philosophical Study, Pandey, Abhinavaqupta: An Historical and

[second Chowkhamba Sanskrit Studies, Vol. : ed.1, Kāvyalamkarasarasamgraha of Udbhata, 1963, pp. 213-14; N.D. Banhatti ed., Bombay Sanskrit and Prakrit Series, No. LXXIX [second ed.], 1982, pp. xxiv-xxviii; Mukunda Madhava Sharma, The Dhvani Theory in Sanskrit Poetics, Chowkhamba Sanskrit Studies, Vol. LXIII, 1968, pp. 14-15; Anita Karan, PratIharenduraja's Contribution to Sanskrit Poetics, Delhi, 1988, pp. 12-15.

368

Udbhata's commentator.91

Because PratTharenduraja's work, unlike Mukulabhatta's,

falls clearly and unambiguously within the field of

Alamkarasastra, his response to Anandavardhana's dhvani theory

has not suffered the same degree of neglect at the hands of

modern scholars of Alamkārasāstra. His treatment of

Anandavardhana's views has, however, provoked a certain amount

confusion and even hostility among these scholars. De

(History of Sanskrit Poetics, op. cit., p. 76) and

Krishnamoorthy (The Dhvanyaloka and its Critics, op. cit., p.

249. label him an opponent of the dhvani theory, while Gerow

claims that he "fully accepts" it (Indian Poetics, op. cit.,

p. 262). Sharma finds in him a "disguised loyalty for

Anandavardhana", and argues that this, combined with his

nominal defence of Udbhata's views, makes it "more precise to

recognize him as a bundle of confusions" (The Dhvani Theory in

Sanskrit Poetics, op. cit., pp. 12-13). It is true that

PratIharenduraja is neither a wholehearted proponent nor an

uncompromising critic of Anandavardhana; however, this is

hardly sufficient ground for a charge of incoherence. Like

Mukulabhatta, he presents a narrowly focussed critique of a

particular feature Of Anandavardhana's argument while

apparently concurring with his general understanding of the

91Cf. P.V. Kane, History of Sanskrit Poetics, Motilal Banarsidass [fourth ed.], 1971, pp. 205-7; S.K. De, History of Sanskrit Poetics, Firma KLM Private Ltd. [second ed.], 1960, Vol. I, pp. 74-76; K. Krishnamoorthy, The Dhvanyaloka and its Critics, Kavyalaya, Mysore, 1968, p. 249.

369

nature and function of poetry.

Pratiharenduraja's examination of the dhvani question is

presented as an epilogue to his commentary on Udbhata's work.

The problem which gives rise to the discussion is this:

Udbhata's work deals only with alamkāras, but Anandavardhana

and his followers have argued that the most aesthetically

significant aspect of poetry is not alamkara but dhvani; how

can Udbhata's neglect of this supposedly crucial element of

poetry be explained or justified? PratIhrenduraja's answer is

a simple one: Udbhata cannot be faulted for omitting the

consideration of dhvani, because what Anandavardhana calls

dhvani is already included in the alamkāras as Udbhata has

defined them. 92

Anandavardhana has divided instances of dhvani, on the

basis their suggested content, into three classes: vastu-

dhvani, alamkara-dhvani, and rasa-dhvani. PratIhārendurāja

cites examples of all three types and tries to show that each

can be included within one or another of the alamkāras dealt

with by Udbhata. As an example of vastu-dhvani he gives the

following verse:

.. [Visnu], who, by a forceful command in the form of a blow from his discus, made the love-play of Rahu's wives devoid of the unfettered play of embracing, so that only kissing remained [possible] .93

92Laghuvrtti on Kāvyalamkarasarasamgraha [Banhatti's ed.], p. 85. Udbhata's

93Ibid., p. 86; cf. above, Chapter Four, pp. 178-179. The choice of example seems odd; Anandavardhana (Kashi ed., p. 225) quotes the verse as an example of the alamkāra

370

Here the statement of an effect, the inability of Rahu's wives

to embrace their husband, causes us to understand the unstated

cause of this effect, the severing of Rahu's head by Visnu.

The unstated meaning we understand from this verse is neither

a rasa nor a figure of speech, so the verse is taken to be a

case of vastu-dhvani. The process by which we understand this

unstated meaning fits the definition Udbhata has given for the

alamkāra paryāyokta ("oblique statement") :

Paryāyokta is that [figure] in which something is communicated in a manner which is devoid of the functioning of directly expressed meanings and directly expressive words and which consists of [implicit] understanding.94

The severing of Rahu's head is not explicitly mentioned in the

verse, but is nevertheless understood; thus the verse is an

example of the alamkara paryayokta as defined by Udbhata.

As we have seen (above, Chapter Four, pp. 157ff.) ,

Anandavardhana himself rejects the identification of dhvani

with alamkäras which contain a suggested element. For him,

such suggestive alamkāras and true cases of dhvani can be

clearly differentiated on the basis of the relative

paryāyokta, not vastu-dhvani. however, Anandavardhana himself concedes that some cases of Elsewhere (pp. 118-19)

paryayokta may be included within the scope of dhvani, provided that the unstated meaning conveyed by such verses is their primary meaning (cf. above, Chapter Four, pp. 163-164). Here the unstated meaning, Visnu's decapitation of Rahu, is the main purport of the sentence, and thus the verse ought, arguably, to be considered a case of vastu-dhvani by Anandavardhana's own standards. Nonetheless, it is surprising that PratIharenduraja did not choose a more unambiguous example of vastu-dhvani for his demonstration.

94Kāvyālamkārasārasamgraha 4.6 (ibid., p. 55).

371

predominance of their suggested and directly expressed

meanings. Only if the suggested meaning of a verse is

subordinate to its expressed meaning can this suggestion be

called an alamkāra; if the suggested meaning is predominant,

it will be taken as an instance of dhvani. It is true that

Udbhata's definition of paryayokta, and definitions of such

alamkäras in general, make no reference to the subordination

of the unstated meanings they convey. Nevertheless,

Anandavardhana and his followers argue that the designation of

these as alamkāras (lit. "ornaments") necessarily implies

subordination. Ornaments, by definition, beautify something

other than themselves; this, it is argued, necessarily implies

their other-directed and subordinate character.

PratIhārenduraja considers and rejects this argument for

differentiating dhvani from Udbhata's alamkāras:

"Here [in the verse ' ... (Visnu), who, by a forceful command ... '], the understood vastu, because it is predominant, should appropriately be called a thing to be ornamented [alamkarya] rather than a cause of ornamentation; how then can this be called an alamkāra?" To this objection we reply: even a predominant thing, if it is a cause of beauty for the things subordinate to it, can become a means Of ornamentation; in everyday experience we see statements such as "Servants are ornamented by their masters." understood meaning Here too, even though the is predominant, there contradiction in calling it an alamkāra, since it is a is no

cause of beauty for the expressed meaning subordinated to it.95

The use of the term alamkara itself does not necessitate

subordination of the suggested to the expressed meaning, and

95Ibid., p. 86.

372

there is no other basis on which to conclude that Udbhata's

definitions of paryayokta and other suggestion-based alamkāras

are meant to apply only to cases of subordinate suggestion.

This and similar examples of vastu-dhvani meet Udbhata's

definition of paryayokta, and ought thus to be regarded as

examples of that alamkara.

In the verse " [Visnu], who, by a forceful command ... ",

the expressed meaning is meant to be expressed (vivaksita);

cases of vastu-dhvani in which the expressed meaning is not

meant to be expressed (avivaksita) may similarly be included

in the alamkaras defined by Udbhata. PratIhārendurāja

considers the (by now familiar) example, "I am Rama, whose

heart is very hard; I can endure all! "96 This example, like

the previous one, conforms to the stated definition of

paryayokta. The word "Rama" is used not in order to convey

its own expressed meaning (as the speaker's identity is

already known), but in order to bring to mind the many

sufferings Rama has endured (loss of his kingdom, exile in the

forest, etc.). These sufferings are understood without being

explicitly mentioned, and this use of the word "Rama" can thus

be taken as a case of paryāyokta. The same general

principle -- understanding of what is not explicitly stated --

will apply in all cases of vastu-dhvani: "Thus, in other cases

as well, when a vastu is understood, it will be called

paryayokta; therefore, when a vastu is understood, there is no

6Ibid., pp. 86-87; cf. above, pp. 200 and 331.

373

property of poetry called 'dhvani' which causes this

understanding, apart from this [paryāyokta] . "97

What Anandavardhana calls alamkära-dhvani can likewise be

accounted for within the framework of Udbhata's alamkāras.

Again Pratīharenduraja considers both vivaksita-vācya and

avivaksita-vācya cases. For the first type he gives the

following example:

O you with the long, trembling eyes; since now, when your smiling face which fills the quarters of the sky with the splendor of your beauty is here, he does not become even slightly agitated, therefore I think this ocean is just a mass of water/ an inert mass.98

The ocean ordinarily becomes turbulent in the presence of the

full moon, and the speaker's expression of surprise and

disparagement at its failure to do so in the presence of a

beautiful woman's face is taken to implicitly indicate the

metaphorical identification (rüpaka) of this face with the

moon. Anandavardhana gives this verse as an example of

rüpaka-dhvani. The unstated alamkara is understood in

addition to the directly expressed meaning of the verse, and

does not supplant or displace it; it is therefore vivaksita-

vācya. There are two ways, Pratiharenduraja argues, to

97Ibid., p. 87.

98Ibid. p. 87. lāvanyakāntiparipūritadinmukhe 'smin smere 'dhuna tava mukhe taralāyatāksi/ ksobham yad eti na manāg api tena manye suvyaktam eva jalarasir iyam payodhih// Anandavardhana gives the verse (which he identifies as his own) as an example of rupaka-dhvani (cf. Dhvanyaloka 2.27 [Kashi ed., p. 261]). The verse also contains a pun on the words jala ("water") and jada ("inert").

374

explain such cases within the framework of Udbhata's

alamkāras:

An alamkära which is understood [without being directly expressed] does not for any reason cease to be an alamkāra; thus [in this verse] there will be the alamkāra rūpaka in an understood form. Or, because here the rūpaka is arrived at by an indirect means of expression [paryayoktya], there is the alamkāra paryayokta.99

Any proposition which conforms to the definition of the figure

rupaka is a case of that figure, whether it is directly stated

or conveyed by implication; the same principle would hold for

any indirectly expressed alamkāra. This indirect mode of

expression is itself the defining feature of the figure

paryayokta, so this and similar examples can alternatively be

explained as instances of this figure.

As a case of an unstated alamkāra conveyed by a verse

whose directly expressed meaning is avivaksita, the following

example is given:

Three people pluck the golden-flowered earth: a hero, one who has achieved learning, and one who knows how to serve. 100

99Ibid., p. 87.

100Tbid., p. 87. suvarnapuspām prthivīm cinvanti purusas trayah/ śūraś ca krtavidyaś ca ya ca jānāti sevitum/7 (Mahabhārata 5.35.64 [Poona ed.]) It is not clear that Anandavardhana would regard this, as PratIharenduraja does, as a case of alamkara-dhvani. In his discussions of alamkara-dhvani (Dhvanyaloka 2.21 [pp. 235-47] and 2.25-27 [pp. 257-78]), Anandavardhana deals only with vivaksitavācya examples, suggesting that he does not admit avivaksitavācya alamkāra-dhvani as a possibility. He elsewhere (cf. Dhvanyaloka, p. 137) quotes this verse as an example of avivaksitavacya-dhvani, but does not discuss his interpretation of it more specifically than this. The point of the verse is that these three kinds of people achieve

375

The sense of the is verse that these three types of people

attain success in the world. This attainment of worldly

success is referred to metaphorically as "pluck [ing] the

golden-flowered earth." Thus we understand from the verse an

unstated simile (upama) between the plucking of the golden-

flowered earth and the attainment of worldly success ;

PratIhārendurāja therefore takes it as an example of

avivaksitavācya upama-dhvani. Such examples, he contends, can

likewise be included in Udbhata's alamkāras:

Here, because plucking which has the golden-flowered earth as its object cannot be intelligibly construed with a hero, etc., figurative indication [laksanā] of the gaining of great there is, by virtue of similarity,

wealth, which is similar to this [plucking] and thus becomes a subject of comparison [upameya]; by means of this laksana there is seen to be [the alamkāra] vidarsana, which contains a relation of standard and subject of comparison [upamanopameyabhava] and which has an expressed meaning which is impossible.101

Udbhata defines the figure vidarsana ("illustration") as that

"in which a connection with something, either real [bhavan] or

unreal [abhavan], causes one to postulate a relation of

standard and subject of comparison. "102 Here the connection

of the hero, etc. with the literally impossible act of

plucking the golden-flowered earth causes one to postulate a

wealth and worldly success. This understanding obviously rests on an unstated comparison or identification of such success with "pluck [ing] the golden-flowered earth"; it is not clear, however, that this comparison itself (understood as an unstated upama or rupaka) should be taken as the primary suggested meaning of the verse.

101Ibid., pp. 87-88.

102Tbid., p. 66.

376

relation of similitude between this action and the unstated

action of gaining great wealth (which is possible for these

agents); thus it is an example of the figure vidarsana as

defined by Udbhata.

What Anandavardhana would call alamkara-dhvani based on

the power of word (sabdasaktimula) Pratiharenduraja includes

within the figure slesa ("pun") . He cites the following

example:

Honor him who is the one abode of all, imperishable/not an abode, the lord of thought, the supreme lord/not the lord of thought, Hari/yellow, Krsna/blue, fourfold self/having a having a capable nature, free of karma/inactive, bearing the discus/ wheel, enemies/destroying the wheel.103 destroying

Here a variety of double-meaning epithets are applied to the

god Visnu. If these terms are taken in their contextually

appropriate senses ("imperishable", "the supreme lord", "Hari"

and "Krsna" [proper names], "having a fourfold self", "free of

karma", "bearing the discus", "destroying enemies"), they will

be properly applicable to Visnu and the verse will convey a

coherent sense. If they are taken in their alternate

meanings, however, they will create a series of apparent

contradictions: he is the abode of all, but he is not an

abode; he both is and is not the lord of thought; he is both

yellow and blue; he is both capable and inactive; he both

103Ibid., p. 88. sarvaikaśaranam aksayam adhīśam īsam dhiyām harim krsnam/ caturatmanam niskriyam arimathanam namata cakradharam// Anandavardhana (Dhvanyaloka, p. 246) cites the verse (which he identifies as his own) as an example sabdasaktimūla-dhvani.

377

supports and destroys a wheel. Therefore, according to

Anandavardhana, the verse suggests the alamkara virodha

(" [apparent] contradiction"), and this suggestion, depending

as it does on double entendre, is said to be based on the

power of word (sabdasaktimūla) .

Udbhata's definition of the alamkāra ślesa104 requires

that examples of this figure involve the "composition of

multiple words which are pronounced with a single effort, or

appear to be so"105 and that these words should "produce an

impression [pratibha] of another alamkāra. "106 All cases of

double entendre will meet the first condition; this example,

because it brings to mind an unstated virodha alamkāra, meets

the second. All cases of what Anandavardhana would call

sabdasaktimūla-dhvani would similarly meet Udbhata's

definition and so may be included in this figure. We have

104Udbhata's gives the name of the figure as slista, but virtually all other Ālamkārikas call it ślesa; Pratīharenduraja uses both terms synonymously in his commentary.

105Udbhata, like many other Ālamkārikas and language theorists in the Sanskrit tradition, does not admit the possibility of true polysemy; each word must have one and only one literal meaning. When a word appears to bear two or more independent literal meanings, these theorists would argue that there are in fact two or more distinct words (each having a single meaning) which happen coincidentally to be phonetically identical. For Udbhata, slesa occurs when such identical words are pronounced (and their separate meanings understood) by a single utterance, or when a single word or group of words, when pronounced, calls to mind a second word or group of words which is phonetically very similar (but not absolutely identical).

106Tbid., p. 58.

378

seen (above, Chapter Four, pp. 166ff) that Anandavardhana

tries to draw a sharp distinction between slesa and

sabdasaktimula-dhvani, insisting that only cases of double

entendre in which some explicit verbal cue prompts the hearer

to look for a second meaning should be considered examples of

the figure slesa; there is nothing in Udbhata's definition of

ślesa, however, which would limit its scope to such cases.

Through his discussion of the previous three example

verses, PratTharenduraja has shown that all possible types of

alamkāra-dhvani may, like vastu-dhvani, be included within one

or another of Udbhata's alamkāras. Examples of dhvani which

produce an awareness rasas or rasa-like phenomena (bhava,

rasa- and bhāva-abhāsa, rasa- and bhava-prasama, etc. -- cf.

above, Chapter Two, pp. 52-58) in an audience may likewise be

so included.

Pratiharenduraja gives the following verse as an example

of rasa-dhvani:

When, in bed, the slender-bodied woman heard herself addressed by the wrong name, she considered turning away; she decided to begin; she almost accomplished what she had set out to do, gently drawing away one slender arm; but she was unable to draw her breasts away from her lover's chest. 107

The verse begins with a description of factors suggestive of

vipralambha-srīgāra ("love in separation") -- the woman hears

107Tbid., p. 88. yate gotraviparyaye śrutipatham sayyam anuprāptayā nirdhyatam parivartanam punar api prarabdhum angikrtam/ bhūyas tat prakrtam krtam ca sithilaksiptaikadorlekhayā tanvangya na tu krastum priyasyorasah// pāritah stanabharah

379

her lover address her by another woman's name and resolves to

turn away from him -- and culminates in the description of a

factor suggestive of sambhoga-srgära ("love in union") -- she

finds herself unable to draw away and remains in her lover's

embrace. Udbhata's definition of the figure rasavat ("having

rasa") requires only that rasas such as srngära should be

"clearly shown, 108 and thus plainly applies here. Where

rasa-like phenomena such as bhava, rasa- or bhava-abhasa, and

rasa- or bhava-prasama are conveyed, Udbhata's definitions of

the figures preyasvat (4.2), urjasvin (4.5), and samahita

(4.7) will similarly apply.109

Ānandavardhana's criterion for distinguishing rasa-dhvani

(where rasa is predominant) from the alamkara rasavat (where

the suggested rasa is subordinated to some other meaning) 110

finds no support in Udbhata's definition, which plainly

applies wherever rasa is present. Anandavardhana's claim that

all types of dhvani, including rasa-dhvani, have been

overlooked (rather than merely mis-analyzed) by earlier

Alamkārikas would require that the figure rasavat, as these

Ālamkārikas understood it, be confined to instances in which

108Tbid., p. 52.

109Ibid., p. 89 (also, cf. above, Chapter Two, pp. 52-58 and Chapter Four, p. 133). Anandavardhana's typology of these rasa-like phenomena is itself based on Udbhata's treatment of them in defining this group of alamkaras, so the close parallelism between his own categorization and Udbhata's is hardly surprising.

110Cf. above, Chapter Four, pp. 172ff.

380

a suggested rasa is subordinated to some other meaning.

PratIhārenduraja rejects this interpretation: as he sees it

Udbhata has not dealt with subordinate rasas to the exclusion

of predominant ones, but has simply formulated his definition

of rasavat without regard for this distinction (thereby

including examples of both types) .

"Because rasas and bhavas are causes of the heightening of poetic beauty, are they, properly speaking, ornaments of poetry [kavyalamkara] or the very essence of poetry [kāvyajīvita] ?" -- this question is not considered [by Udbhatal for fear of overlengthening his work. And the nature of rasas and bhavas is not analyzed here, because it is not relevant and because a great deal would have to be said about it. 111

Udbhata's purpose is to provide a complete catalogue of those

formal and semantic features which give poetic language its

beauty; whether these features are, in any given case,

essential or non-essential, primary or subordinate, is a

question which simply does not concern him.

So, examples of vastu-dhvani, alamkāra-dhvani, and rasa-

dhvani can all be accounted for as examples of one or another

of Udbhata's alamkāras; there is no need to posit dhvani as

something separate from these alamkaras (and, consequently, no

grounds for accusing Udbhata of failing to consider the most

crucial feature of poetry in his work). The point is not that

there is no such thing as dhvani, but rather that

Änandavardhana is mistaken in imagining that the phenomena he

deals with have been overlooked by previous Alamkārikas. We

111Ibid., p. 54.

381

have seen (above, Chapter Four, pp. 122ff) that Anandavardhana

is anxious to show that his dhvani theory is not merely a

redescription of poetic features already dealt with by his

predecessors but deals with a range of phenomena which,

despite their pivotal role in the creation of aesthetically

effective poetry, had so far escaped the notice of poetic

theorists. It is this contention of Anandavardhana, and only

this contention, which PratIharenduraja's analysis seeks to

undermine.

Pratiharenduraja does not challenge the possibility or

usefulness of distinguishing, as Anandavardhana does, between

cases in which a suggested meaning is predominant with respect

to the directly expressed meaning which accompanies it (i.e.

dhvani) and cases in which it is subordinate (i.e.

gunībhūtavyangya); again, his point is only that examples of

both types fall within the scope of Udbhata's alamkaras. In

fact, he explicitly accepts the existence of dhvani, and

offers his own categorization of it. 112

We have seen (above, Chapter Five, pp. 198ff) that

Anandavardhana presents two separate schemata on the basis of

which dhvani may be classified; one based on the nature of its

suggested content (vastu, alamkära, or rasa), and the other on

the relation between the expressed and suggested senses

(vivaksita-vācya and avivaksita-vācya, with their respective

subdivisions). PratIhārendurāja's typology of dhvani is

112Tbid., pp. 89-92.

382

basically a merging of these two classificatory schemata. He

notes that the various subtypes of dhvani examined by

Anandavardhana are restricted as to the type of suggested

meaning they may convey. In that type of vivaksita-vācya

dhvani which is based on the power of word (sabdasaktimūla --

e.g. "Honor him who is the one abode of all ... ") only

alamkäras may be suggested; in that type where there is no

perceptible sequence between the expressed and suggested

meanings (asamlaksyakrama -- e.g. "When, in bed, the slender-

bodied woman ... ") only rasas or rasa-like phenomena may be

suggested (and all cases where rasas and rasa-like phenomena

are suggested must belong to this type). In vivaksita-vācya

dhvani which does not belong to either of these subtypes, the

suggested meaning may be either a vastu (as in " ... [Visnu],

who, by a forceful command ... ") or an alamkāra (as in "O you

with the long, trembling eyes ... "), but not a rasa; likewise

in avivaksita-vācya dhvani either a vastu (as in "I am Rāma,

whose heart is very hard ... ") or an alamkāra (as in "Three

people pluck the golden-flowered earth ... ") may be suggested.

Thus, by combining considerations of content and process,

Pratiharenduraja arrives at a single typology of dhvani, whose

six subdivisions are illustrated by the six examples he has

already cited in his argument for the inclusion of dhvani in

Udbhata's alamkāras.

PratIharenduraja not only admits the existence of dhvani,

but also, following in Anandavardhana's footsteps, argues for

383

the central importance of rasa in of poetry. Indeed, he goes

further than Anandavardhana is prepared to do, insisting that

the evocation of rasa is not only the primary object of most

great poetry, but is an essential property of poetry.

Ānandavardhana, as we have seen (above, Chapter Six, pp.

282ff), accepts the existence of a class of poetry (citra-

kavya) which does not convey rasa at all, even as a

subordinate factor. PratIharenduraja, by contrast, contends

that all poetry must necessarily convey rasa. His argument

for the essentiality of rasa is based on his analysis of the

role of gunas ("qualities") in poetry, an analysis which

combines elements of Vamana' s and Anandavardhana' s

understanding of the nature of guna.

Vamana argued that gunas (and the "styles" [rītis] based

upon them) are the sources of poetic beauty and must

necessarily exist in all poetry, in contrast to alamkāras

which, while they may heighten the beauty created by these

gunas, are not similarly indipensible. PratIharenduraja

likewise contends that gunas are necessary and alamkāras

optional in the composition of good poetry, and quotes Vamana

(Kāvyalamkārasūtra 3.1.1-3) in support of this view. 113

However, like Anandavardhana (and unlike Vamana), he sees

gunas not as properties of the words and expressed meanings of

a poem, properly speaking, but as properties of the rasas

113Ibid., p. 82; cf. Vamana, Kāvyālamkārasūtra, Kashi Sanskrit Series 209, pp. 82-82 and above, Chapter Four, pp. 144-146.

384

these words and meanings convey. 114 Verses which convey

"sweet" rasas (such as srngara or karuna) are said to possess

the quality madhurya ("sweetness") ; those which convey rasas

such as raudra ("fury") are said to possess ojas ("strength") .

Prasāda ("clarity") is a quality common to all rasas, and

verses which clearly communicate any rasa may be said to

possess it. PratIharenduraja, again like Anandavardhana,

accepts only these three gunas (rather than Vamana's ten).

All three are, as he understands them, invariably associated

with rasa, so that poetry which possesses any of these gunas

must, by definition, convey some rasa. By combining Vamana's

view that gunas are essential with Anandavardhana's view that

they are based on rasa, Pratiharenduraja arrives at the

conclusion that all poetry must necessarily convey rasa. 115

It is not only in this insistence on the central role of

rasa in poetry that Anandavardhana's influence can be seen in

Pratīharenduraja's work. Central to Anandavardhana's

reconceptualization of the nature of poetry and the demands of

poetic analysis was his importation into the realm of

Alamkārasāstra of a Mīmāmsa-derived, teleogical model of

textual coherence and, along with it, some of the key terms

and concepts116 developed by the MImamsakas as tools for the

114 Tbid., p. 81.

115Ibid., p. 83-84.

116Primary vs. subordinate meanings, vidhi vs. anuvāda, vivaksita- vs. avivaksita-artha, etc.

385

analysis of Vedic discourse. Pratiharenduraja continues this

"Mimamsification" of Alamkārika method and terminology by

recasting many of Udbhata's definitions and categories in

Mīmāmsā terms. He shows that the figure dīpaka

("illuminator" -- in which syntactic parallelism is used to

reveal an unstated simile) is based on a kind of anusanga. 117

Like Vamana before him (but unlike Udbhata), he explains the

figure slesa ("pun") as a kind of tantra. 118

Udbhata (5.11-13) mentions four possible types of

blending (samkara) between figures: (1) that in which it

cannot be conclusively determined which of several alamkāras

is communicated by a verse, (2) that in which a single verse

contains both sound-and meaning-based alamkāras, (3) that in

which two or more alamkäras are conveyed by the same word, and

(4) that in which the understanding of one alamkara depends on

another. PratTharendurajall explains these four kinds of

relation between alamkāras in Mīmāmsa terms: in type (1) there

is vikalpa ("option") between two or more mutually exclusive

alamkāras, 120 in type (2) there is vyavastha ("differential

i17Ibid., p. 16. Anusanga is the process through which an elliptical sentence is made comprehensible by understanding the repetition of a word or phrase from a preceding or following non-elliptical sentence -- cf. Mimamsāsūtra 2.1.48.

118Ibid., p. 58. Cf. above, Chapter Five, pp. 196.

119Ibid., p. 71.

120For the MImamsakas, vikalpa must be accepted when there are two or more contradictory injunctions pertaining to the same act; for instance, if one injunction requires that the substance to be used for a particular sacrificial offering

386

implementation") of two alamkaras which function independently

in their respective domains, 121 in type (3) there is

samuccaya ("aggregation") of multiple alamkāras, 122 and in

type (4) there is an angangibhava ("primary/subordinate

relation") between one alamkāra and another.

Pratiharenduraja's reformulation of Udbhata's categories does

not really add anything to his analysis of samkara, but it

shows an effort to "translate" Ālamkarika language into

Mīmāmsa terms.

Apart from this terminological shift, the influence of

Mīmamsa discourse analysis on PratIharenduraja's thought can

be seen in his efforts to explain the cognitive processes

involved in interpreting the alamkaras described by Udbhata.

should be rice and another requires that it should be barley -- cf. MImamsasutra 12.3.10-15.

121Vyavastha is the assignment of distinct roles to several elements of the sacrifice (e.g. the various ritual utensils) on the basis of their natural capacity. For example, there are several injunctions regarding the division of substances to be used for sacrifice: "One should divide with the ladle, " "One should divide with the knife," "One should divide with one's hand." Although these injunctions do not specify which substances are to be divided by which of these three implements (ladle, knife, and hand), they are not to be used indiscriminately; liquid substances (e.g. ghee) should be divided with the ladle, solids (e.g. meat) with the knife, and collected but separable substances (e.g. grains) with the hand. These utensils are thus "differentially implemented" (vyavasthita); each functions in its own proper domain. Cf. Mimamsasutra 1.4.25 and 3.1.7-10.

122In Mīmamsa, samuccaya is the performance in combination of two ritual elements when there is no blocking (badha) of one by the other and the two need not be viewed as exlusive alternatives (as in a case of vikalpa) -- cf., for example, Sabarabhasya on Mimamsasutra, 10.3.23-29, 10.4.1ff, and 12.3.17.

387

Udbhata, like all of the early Ālamkarikas, is relatively

uninterested in processes of interpretation; the main concern

of his alamkāra-analysis is to specify the necessary formal or

logical constituents of the figures he describes, any

sequential or other restrictions on the presentation of these

constituents, and the conditions under which one or more of

these constituents may be implied rather than directly

stated. 123 PratIharenduraja accepts and defends Udbhata's

treatment of these topics, but pushes his own analysis in a

new direction; in explaining Udbhata's alamkaras, he generally

offers detailed accounts of the interpretive processes through

which they are understood, often in terms of the distinction

(first outlined by the Mīmāmsakas) 124 between avantara-

vākyartha ("intermediate sentence meaning") and paryavasāna

(the "completion" of the interpretive process) . 125

Ānandavardhana, by his introduction into Alamkāraśāstra

of a Mimamsa-derived understanding of the nature of texts and

Mīmamsa-derived techniques of text analysis, began for the

123As in an "elliptical simile" Kavyalamkārasarasamgraha 1.17-21 [op. cit., pp. 17ff]). (luptā upamā -- cf.

124Cf. above, Chapter Three, pp. 105-109

125Cf., for example, his treatment of the figures äksepa ("denial", in which one overtly denies what one actually wishes to say -- ibid., pp. 31-34), vyajastuti ("disguised praise", in which a statement which is overtly a rebuke covertly conveys praise -- p. 66), vidarsana ("illustration" [cf. above, pp. 375-376] -- pp. 66-67), and upameyopamā ("comparison with the thing compared", in which each of two things is said to resemble the other so as to suggest that no third thing is comparable with either -- pp. 71-72).

388

first time to seriously bridge the gap between the conceptual

vocabulary of literary theory and that of philosophical

language-analysis. In Pratīhārenduraja's work we can see a

continuation of this process; there appears to be a concerted

effort on his part to describe and analyze Udbhata's alamkāras

in Mīmāmsā terms.

Conclusion

While they come at the problems raised by

Anandavardhana's work from different angles, Mukulabhatta and

PratIharenduraja are pursuing very similar projects. Each

mounts a defense of existing language theory in the face of

Anandavardhana's charges of inadequacy, while at the same time

seeming to agree with Anandavardhana's revolutionary view of

the nature of poetry and of the crucial role of unstated

meaning, especially rasa, within it. Their response to

Ānandavardhana's critique of previous language theory is in a

sense conservative but, given their apparent acceptance of his

teleological, rasa-centered model of literary aesthetics, this

conservative stance requires that they revise, or at least

significantly elaborate, the theories of language (whether

general or specifically literary) they aim to defend, so as to

demonstrate that these theories can satisfactorily account for

the full range of poetic phenomena dealt with by

Anandavardhana without the introduction of any such new

semantic mode such as vyañjana or dhvani.

In both Mukulabhatta's and Pratihārenduraja's works we

389

see an increasing erosion of the distinction between the

theory of specifically poetic language and the theory of

language in general. Mukulabhatta totally rejects any such

distinction, arguing that all the semantic phenomena

observable in poetic language, in particular those

Anandavardhana classifies as dhvani, can be satifactorily

explained via the already well-established process of

figurative expression. For him, the ways in which poetry

conveys meaning are fundamentally no different from those in

which all language conveys meaning, and there is thus no need

for a special theory of poetic signification.

PratIhārenduraja does not directly challenge the

differentiation of poetic from general language-theory; he

steers clear of the question which most preoccupies his

mentor, namely the legitimacy of Anandavardhana's claims for

vyañjanā as specifically poetic mode of expression.

Nevertheless, his effort to redescribe Udbhata's categories in

Mīmamsaka terms, and to outline the series of interpretive

stages involved in understanding the alamkaras Udbhata

describes, serves in practice to facilitate the merging of

Ālamkārika modes of description and text-analysis with those

of linguistic philosophy, a trend which continues in the works

of later Ālamkārikas.