5. CHAPTER TWO EARLY ALAMKARASASTRA AND THE PROBLEM OF RASA

CHAPTER TWO EARLY ALAMKARASASTRA AND THE PROBLEM OF RASA

The Alamkarikas who predate Anandavardhana have, by and

large, fared poorly at the hands of modern scholars. Because

Anandavardhana's theory of rasa-dhvani has been so widely seen

as the crowning triumph of Indian aesthetic theory, the works

of earlier Alamkārikas have tended to be viewed almost

exclusively through the lens of this later development. These

works have been scoured for any stray observations which may

be seen as anticipations or primitive formulations of elements

of the dhvani theory, while the task which absorbs most of the

attention of these Ālamkārikas -- the definition,

exemplification, and categorization of those phonetic and

semantic features which add beauty to poetic language and of

those features which detract from it (and are thus to be

avoided) -- has been for the most part denigrated and seen as

marginal to the "real" concerns of aesthetic theory.

There are five pre-Anandavardhana Ālamkarikas whose works

survive: Bhamaha (c. 700 AD), Dandin (c. 750 AD), Udbhata (c.

800 AD), Vamana (c. 800 AD), and Rudrata (c. 825 AD). Of

these, Bhamaha, Udbhata, and Rudrata are usually labelled as

"The Alamkara School" (because they are thought to regard

figures of speech [alamkäras] as the primary determinant of

beauty in poetic language) 1, and Dandin and Vamana as "The

1Cf. Kane, History of Sanskrit Poetics, op. cit., pp. 372ff, and De, History of Sanskrit Poetics, op. cit., Vol. II, pp. 32ff.

36

37

Rīti School" (because they are thought to ascribe a central

role in creating poetic beauty to riti ["style", understood as

a particular constellation of poetic qualities (gunas) --

tenderness, strength, clarity, and so on] ) .2

Although these writers are rigidly divided into "schools"

based on what each appears to regard as the fundamental

principle of poetic beauty (alamkāra or rīti), they are

generally seen as paying insufficient attention to these

fundamental principles, and as overly concerned with the

minutiae of definition and classification, to the exclusion of

properly aesthetic or poetic questions (in contrast to

Anandavardhana, who is seen as showing a proper concern for

his central aesthetic principle, dhvani, and as relatively

unconcerned with such "external" features as alamkaras)

Krishnamoorthy sums up the conventional appraisal of these

authors quite well:

Most of these early theorists are fond of formalism and attention to minute details. In their treatment of individual excellences we miss a philosophical approach. They are content with enunciating rules of rhetoric and illustrating them with examples taken at random ... Their tone smacks of dogmatism while their discussion is scholastic, aesthetic analysis remaining perfunctory ...

2Cf. Kane, ibid., pp. 378ff, and De, ibid., pp. 75ff. All these authors deal with many factors in poetry and, with the exception of Vamana, do not explicitly declare any one of these elements to be the "soul" or essence of poetry, so there is some disagreement regarding the "schools" to which certain authors should be assigned; Dandin is sometimes said to belong to the "Alamkara-" rather than the "Riti-school" (cf. Kane, op. cit., pp. 373-374), and Rudrata is sometimes grouped with Bharata (reputed author of the dramaturgical Natyasastra) as

11). a member of the "Rasa-school" (cf. Sharma, op. cit., pp. 10-

38

Though we cannot say that they lacked in literary insight, we feel the lack of depth in their formal categories. The value of these theories is in the good promise they held out for a future writer, the author of the Dhvanyaloka, to rise on their shoulders and perfect a theory free from their drawbacks. "3

The effort these authors devote to the careful differentiation

and definition of figures of speech (alamkāras) and poetic

qualities (gunas) is seen as marginal to those questions which

ought to be of most concern to literary aestheticians.4 It

is recognized that these authors' approach to poetic analysis

differs markedly from that of Anandavardhana and his

followers, but this approach is not considered as a serious

3Krishnamoorthy, The Dhvanyaloka and its Critics, op. cit., pp. 21-22. Cf. also De on Bhamaha (History of Sanskrit Poetics, op. cit., Vol. II, pp. 38-9), "We must not yet look to his work, nor to that of his follower Udbhata, for a thoroughly critical system. conceptions, but We can deduce certain broad the practical object, underlying the speculation of this school in general, and its more or less normative character did not allow sufficient scope for purely theoretic treatment"; and on Rudrata (Ibid., p. 66), "Indeed, the practical nature and scope of his work, like that of Udbhata's, leave hardly any room for discussion of general principles or of speculative aspects of the involved. questions Rhetoric rather than Poetics appears to be his principle theme, as it is of most writers of this system who concern themselves entirely with the elaboration of rhetorical categories in which they suppose the whole charm of poetry lies."

The occasional observations these authors make in passing on what appear to be fundamental aesthetic principles are sometimes treated as their most important contributions to the development of Alamkarasastra. E.g. Bhamaha's mention (in Kāvyālamkāra 2.85) of vakrokti ("oblique expression") as an element common to all figures (cf. De, History of Sanskrit Poetics, op. cit., pp. 47ff, 102, and below, Chapter Six, pp. 276-277) and Vamana's description (Kavyalamkārasutra 1.2.6) of rīti as the "soul" (atman) of poetry (cf. Kane, History of Sanskrit Poetics, op. cit., pp. 380-381 De, History of Sanskrit Poetics, op. cit., Vol. II, pp. 102-3; also above, Chapter One, pp. 7-9 and below, Chapter Four, pp. 144-146).

39

alternative. It is treated only as a failure to hit on the

"right" approach -- i.e. Anandavardhana's approach.

One of the few serious attempts to circumvent this

tendency can be seen in the works of Edwin Gerow.5 Gerow sees

the formalist poetic of the early Alamkārikas as a useful and

viable approach to literary analysis, one that he argues is

entirely appropriate to the stanzaic poetry typical of

classical Sanskrit. He criticizes the standard view of the

early Ālamkārikas as skewed by a focus on later developments:

Contemporary critics have too long been dependent on the reformulation of the alamkara theory proposed by the dhvani writers, Anandavardhana and Abhinavagupta, for whom the alamkārika6 viewpoint is deprived of its independent status and reduced to a moment in a more all- embracing view. Histories of Indian poetics have been little more than attempts to justify this ex post facto subordination. The theory should be considered in its own terms and not for what defects may be predicated of it in terms of another aesthetic.7

Gerow argues, rightly, that if we wish to fully understand the

works of these Alamkärikas we must make an effort to view them

in their own terms, not in those of later theorists; I will

5Chiefly in his Glossary of Indian Figures of Speech (The Hague: Mouton, 1971) and Indian Poetics (Vol. V, fasc. 3 in A History of Indian Literature [ed. Jan Gonda], Wiesbaden, Otto Harrassowitz, 1977) .

6Gerow generally uses the term alamkārika to refer only to the pre-Anandavardhana aestheticians (because, in comparison with later authors, they stress the importance of figures of speech -- alamkāras -- in poetry). restricted usage has no precedent in the Sanskrit tradition, But this

in which the term Alamkarika refers to any practitioner of Alamkarasastra, regardless of his views on the relative importance of alamkāras in poetry.

Gerow, A Glossary of Indian Fiqures of Speech, op. cit., p. 70.

40

argue further that we cannot properly make sense of the

radical transformation of Alamkārasāstra brought about by

Anandavardhana's Dhvanyaloka unless we understand the internal

logic of his predecessors' approach to poetic analysis. These

early Ālamkārikas were not misguided or ineffective

practitioners of discipline later perfected by

Ānandavardhana; they were engaged in a quite different

enterprise.

The Formalist Paradigm of Literary Aesthetics

The early Alamkarikas were not striving for, and failing

to achieve, a holistic theory of poetic beauty; their

understanding of the nature of poetic language made the

articulation of general or underlying principles largely

superfluous to the business of literary analysis. Again,

Gerow is instructive:

Indeed, if the older alamkārika poetic is, as we assert, a study of the capacities of poetic diction, little purpose would be served in expounding a universal theory, (such as "ambiguity, ") for it is only in variety that figuration constitutes a "universe" of discourse Bhamaha's mention of vakrokti is a summation, 'theory' ... Put in another way, telling a poet he must nota

speak "deviously" does not help us understand what in fact he says. The interest, both for poet and for audience, is in how he does it in practice, not in his intention.ª

@Gerow, Indian Poetics, op. cit., p. 235. Yet even Gerow cannot resist the temptation to see in these authors' works a more "general" concern than the definition and exemplification of particular figures and qualities. He sees the analysis of these particular elements as driven by the need to find criteria for the differentiation of poetic from non-poetic discourse: "Having situated the poetic problem in a formalistic context, we must now pose the question (following the Indian tradition) : What is the specific characteristic of poetic usage, its visesana ?... Indian writers, from the

41

It is with the analysis of specific features of poetic

language that these Alamkarikas are concerned; their interest

is in specifics as specifics, not as instances of some more

general phenomenon. Subsuming these particular formal

features of poetic language under some general label (such as

vakrokti ["oblique expression"]) contributes nothing useful to

the enterprise of defining and categorizing them. Thus, when

these Ālamkärikas do refer to more general principles, they

tend to do so only in passing, quickly returning to the

analysis of specifics.

Underlying this particularistic approach to poetic

analysis is a certain understanding of literary aesthetics --

one which differs markedly from that of Anandavardhana and

most of the later Alamkarikas and which has been poorly

articulated by modern scholars. This aesthetic is revealed in

the basic metaphor in terms of which the early Alamkārikas

beginning, conceived their problem as that of 'poetic differentia', the visesana, which both establishes the poetic genre as a proper subject matter and which, by normative application, distinguishes the finest examples of poetry." (A Glossary of Indian Figures of Speech, op. cit., p. 16 -- cf. also Indian Poetics, op. cit., pp. 223-4, 233). Gerow provides no textual backing for this claim, and I can see nothing in the works of any Alamkarikas (pre- or post- Anandavardhana) to justify it. The differentiation of poetic from non-poetic discourse is treated as unproblematic, certainly by the early Alamkärikas; to describe it as what the Ālamkarikas "from the beginning, conceived as their problem" seems bizarre. The task to which these authors apply themselves is not to develop criteria for determining what is poetry and what is not but, taking this distinction for granted, to determine what specific, isolable linguistic features add to the beauty of poetic texts, and what features detract from it.

42

describe the relation between the poem and those linguistic

features which render it beautiful: that of bodily

ornamentation. The words and meanings which make up a poem

are described as its "body" (sarira)9, while those phonetic

and semantic factors which, when introduced into poetic

language, render it more beautiful are called "ornaments"

(alamkāras) .10 Discussions of the "body" of poetry are, for

the most part, perfunctory;1l for all of these Ālamkārikas

attention is focussed primarily on those formal features which

ornament and beautify this "body". A poem, for them, is a set

of words and meanings, a text, into which have been introduced

certain phonetic and semantic patterns each of which adds some

element of beauty to this text; poetry is languge made

beautiful by the inclusion within it of a variety of

individually beautiful elements. The object of their poetic

analysis, therefore, is to isolate, define, classify, and

exemplify as many such elements as possible, thus providing

9Cf. Bhāmaha, Kāvyālamkāra 1.23; Dandin, Kavyadarsa 1.10; Vāmana, Kavyalamkarasutravrtti on 1.2.6 and 2.1.1.

10The treatment of alamkaras and gunas as separate classes of beautifying elements does not really disrupt this metaphor. The only one of the early Alamkarikas to draw a definite categorical distinction between the is Vāmana (Kāvyālamkārasūtra 3.1.1-3), and even he allows that all two

beautifying elements (both figures and qualities) may be covered by the general label "ornamentation" (alamkāra) (Ibid., 1.1.1-3). Dandin, who lays great emphasis on gunas as the basis for differentiating poetic styles, and has often on those grounds been seen as a forerunner of Vamana's, indicates unambiguously that gunas are Kāvyādarśa 2.1-3. a kind of alamkara -- cf.

11Cf. Bhamaha 1.16-30; Dandin 1.10-39; Vāmana 1.3.21-32.

43

something approaching a complete inventory of those concrete

linguistic devices by which poets make their poetry beautiful

and hence desirable.

Once it is understood that for these Ālamkārikas the

beauty of a whole text is a function of the individual beauty

of its component elements, their approach to literary analysis

makes perfect sense. It explains the relentless "formalism

and attention to minute details" which has so troubled modern

analysts of their works. If it is the addition to it of

beautiful phonetic and semantic elements which makes a

literary text beautiful, then it is on the analysis of these

individual elements that literary analysis must concentrate.

There is no need to dwell on the analysis of literary

compositions as wholes; only take care of the details, and the

whole will take care of itself.

What really separates this kind of aesthetic analysis

from that of Anandavardhana and those who come after him is

the total irrelevance of contextual considerations within it.

Anandavardhana, drawing (as we shall see) on a Mimamsa-derived

understanding of the nature of textual coherence, formulates

a model of literary aesthetics according to which the

individual elements of a poem are valuable only insofar as

they subserve a single, overriding purpose: ideally, the

communication of some particular rasa (aestheticized emotion --

cf. below, Chapter Four, pp. 140ff). Thus, for him, the value

of alamkäras and other formal elements of a poem is not

44

intrinsic; it is context dependent. Where a particular figure

would enhance the capacity of the whole text to convey its

intended rasa, it should be used. Where it would detract from

this capacity, it should be avoided. For the early

Alamkārikas, however, alamkāras and gunas are intrinsically

valuable. A properly constructed metaphor is beautiful in and

of itself. It will add to the beauty of a text whenever and

wherever it is used. Its context has no bearing whatsoever on

its aesthetic value. Thus for them the core of literary

analysis is the precise analysis of individual figures and

qualities, rather than of the higher order textual units in

which they are employed.

This particularistic, formalist aesthetic, with its

consequent de-emphasis of context, is common to all the pre-

Anandavardhana Ālamkārikas. In this respect, the tendency of

modern scholars to divide these authors into the alamkära and

rīti "schools" has obscured the basic unity of their

understanding of the nature of poetic beauty and of the

imperatives of aesthetic analysis. The "Riti School",

especially as exemplified in the work of Vamana, is generally

portrayed as a decisive "advance" over the "Alamkara School",

because Vamana speaks of riti ("style") as the "soul of

poetry" (thus foreshadowing Anandavardhana's use of the same

phrase to describe his dhvani), and because he regards certain

poetic features (gunas ["qualities"], the constituent elements

of riti ["style"]) as essential to good poetry, and others

45

(alamkāras) as inessential.12

Yet what is really striking about Vamana's categorical

distinction between gunas and alamkāras is that it has no real

effect on his analyses of elements of either category. In his

examination of both gunas and alamkāras Vamana follows exactly

the same procedure as his predecessors. His analysis of each

consists of the same three components: definition,

classification, and exemplification. The labeling of elements

of the first category as "essential" and those of the second

as "inessential" produces no noticeable alteration in the

treatment of either. Both are beautiful and hence desirable

in poetry, and both are so intrinsically. There is still no

hint of context as a factor governing the value of particular

features in particular settings. The "inessential" status of

alamkāras does not necessitate that their selection be in any

way guided by their appropriateness to particular gunas or

rītis, or to anything else. They are still treated as

elements aesthetically valuable in and of themselves.

Thus Vamana's emphasis on guna and riti does not really

entail any significant departure from the formalist aesthetic

of "The Alamkara School". Whatever their differences with one

another on matters of detail, all of the pre-Ānandavardhana

Alamkārikas are basically at one in their understanding of the

nature of literary aesthetics and in their approach to

12Cf. Vāmana, Kavyalamkarasutra, 3.1.1-3; also below, Chapter Four, pp. 144-146.

46

literary analysis. For all of them, poetic language is

beautiful because it incorporates a series of individually

beautiful phonetic and semantic elements and the primary

object of poetic analysis is to isolate, classify, and

exemplify as many such elements as possible.

This particularistic model of literary aesthetics leads

naturally to the view that where alamkāras and other

beautifying elements of poetic language are concerned, more is

better. If such elements are to be valued for their intrinsic

beauty, rather than for their appropriateness to particular

contexts, then one should incorporate as many such elements as

one can into one's poetry. This attitude can be seen in

Bhamaha's discussion of samsrsti, the combination of multiple

alamkaras in a single verse: "The best ornament [vibhūsā] is

samsrsti, because of its combination of many alamkāras; thus

it is said to be constructed like a string of gems."13 I£

each alamkära is desirable in and of itself, then the

combination of several of them will be preferable to any one

singly. The same idea can be seen in Dandin's and Vamana's

comparative evaluation of regional styles of poetry; for both,

the Vaidarbhi style is superior to all others because it is

characterized by a greater number of gunas.14 This attitude

is the logical outcome of these Ālamkarikas' model of literary

13Bhamaha, Kavyalamkāra 3.49.

14Cf. Dandin Kāvyadarsa 1.40ff and Vamana, Kāvyalamkarasutra 1.2.9-14.

47

aesthetics. If each guna or alamkāra has a certain intrinsic

beauty which it imparts to those compositions in which it is

employed, then the way to maximize the beauty of any

composition is to maximize the concentration of gunas and

alamkāras within it.

The Problem of Rasa

That feature of literature to which Anandavardhana would

later assign a preeminent role, rasa (aestheticized emotion),

receives comparatively scant attention from Bhamaha, Dandin

and Vamana. It is given somewhat greater prominence by

Udbhata and becomes a major topic in Rudrata's Kavyalamkāra.

Yet, for all of these authors, rasa poses something of a

theoretical problem. It represents a particular kind of

thematic content, and cannot easily be dealt with in terms of

the formalist aesthetic of these Alamkārikas. The anomalous

status of rasa becomes increasingly evident in their works,

culminating in that of Rudrata who effectively abandons the

effort to present an integrated poetic; Rudrata elaborates two

separate modes of literary analysis, one formalist and

alamkära-based and one based on rasa, and makes no attempt

whatsoever to relate the one to the other. As we shall see,

one of the great strengths of Anandavardhana's

reconceptualization of Alamkārasāstra is that it succeeds in

integrating a rasa-based approach to literary aesthetics with

the formal categories of the early Ālamkārikas.

The idea of rasa did not originate in the realm of

48

Alamkāraśāstra. It was first developed in the field of

dramaturgy, specifically in the Natyasastra (ascribed to

Bharata) .15 According to the Natyasastra, there are eight

basic emotions (sthayibhāvas) 16 which may be presented in

drama and, corresponding to them, eight rasas17 (literally,

"flavors"; aesthetic moods or aestheticized emotional

states) 18 which may be created by a performance. The concept

of rasa plays a central role in the dramatic theory developed

in the Natyasastra, and comes ultimately to play a pivotal

role in Alamkārasāstra as well.

15The work is encyclopaedic, and is presumably a compilation rather than the work of a single author. It is impossible to date it with any precision, but the text probably took on something like its present form in the mid- first millennium AD. The earliest known commentator on the Natyasastra is Udbhata (c. 800 AD -- although there are stray references in the works of Abhinavagupta and others to early writers on dramaturgy who may have commented on it: cf. Kane, History of Sanskrit Poetics, op. cit., pp. 48ff. and De, History of Sanskrit Poetics, op. cit., Vol. 1, pp. 31ff. The earliest extant commentary is the AbhinavabhāratI of Abhinavagupta (c. 1000 AD) .

16Desire (rati), humor (hāsa) , (krodha), daring (utsaha), fear (bhaya), disgust (jugupsa) , grief (śoka) , anger

and astonishment (vismaya) -- cf. Natyasastra 6.17, 7.8ff (ed. M. Ramakrishna Kavi, Gaekwad's Oriental Series 36 [Vol. I], 1956 [second ed.], pp. 267, 350ff.).

17Erotic (śrīgāra), comic (hāsya), pathetic (karuna) furious (raudra) , heroic (vira) , frightful (bhayanaka), disgusting (bibhatsa) , and wondrous (adbhuta) -- cf. Nātyasastra 6.15, 6.45ff. (ibid., pp. 266, 299ff.).

18The exact nature of rasa, as well as its locus (i.e., whether the term designates an emotional experience of the character portrayed, the actor, or the spectator), were topics hotly debated by later writers on dramaturgy; the most important such views are summed up by Abhinavagupta in his AbhinavabharatI (ibid., pp. 272-286). Also cf. below, Chapter Nine, pp. 455ff.

49

Bhamaha and Dandin are aware of the Natyasastra (or, at

least, of the discipline of dramaturgy) and refer to it in

passing, but neither seems to see it as especially relevant to

poetic analysis or devotes any significant attention to it.

Having mentioned drama (nātaka) as part of his list of

literary genres, Bhamaha says, "This is said to be [a work]

whose meaning is to be enacted [abhineyārtha]; a full analysis

[vistara] of this has been made by others. "19 Dandin refers

on one occasion to several of the elements of drama discussed

in the Natyasastra (sandhis ["plot junctures"] and so on); he

states that these "have been explained in another text" and

notes, without further elaboration, that they should all be

regarded as alamkäras.20 Thus, while both are aware of the

dramaturgical tradition and its potential bearing on literary

analysis, neither wishes to devote much attention to it; they

seem content to allow for a division of labor, referring

readers to the Natyasastra itself for a detailed discussion of

dramaturgical questions. Both Bhamaha and Dandin find a place

for rasa in their taxonomy of poetic figures, yet both seem to

have difficulty explaining rasa in formalist terms, and their

rasa-based alamkaras occupy an anomalous position in their

systems .

Most of the meaning-based alamkaras dealt with by Bhamaha

and the other early Ālamkarikas consist of the use of

19Kāvyālamkāra 1.24.

20Kavyadarsa 2.364.

50

particular propositional structures: e.g. upama ("simile" -- in

which one entity is compared with another on the basis of

shared properties), rūpaka ("metaphorical identification" -- in

which one entity is identified with another on the basis of

shared properties), or samāsokti ("condensed expression" -- in

which the attributes ascribed to one entity bring to mind a

second which is not explicitly mentioned) .21 These

propositional structures impose no restriction on content; in

principle, anything at all may be made the subject of a

comparison or a metaphorical identification. However, in the

beginning of the third chapter of his Kavyalamkara, Bhamaha

deals with a small group of figures which do not conform to

this pattern: figures which consist not in the relation of

unspecified entities in a particular propositional structure,

but in the selection of a particular kind of subject matter --

which specify not the manner, but the content, of expression.

These figures are preyas ("affectionate"), rasavat

("possessing rasa"), urjasvin ("forceful" or "haughty"), and

samāhita ("completed" or "brought together") .22 The

problematic status of these figures can be seen in Bhamaha's

atypical (and rather cryptic) treatment of them. Departing

from his usual procedure, Bhamaha provides no definition at

all for three of these figures (preyas, ürjasvin, and

21Cf. Bhāmaha, Kavyālamkāra, 2.30, 2.21, 2.79.

22Ibid., 3.5, 3.6, 3.7, 3.10. All four alamkaras are explained in terms of rasa and its permutations by Udbhata and by the later Alamkara tradition -- cf. below, pp. 52-58.

51

samahita); in each case, we are given only the name of the

alamkāra and a brief example.23 For the fourth figure,

rasavat, the definition he provides is purely tautologous

("rasavat is that in which the rasas, śrīgara etc., are

clearly shown"). Bhamaha's treatment of all these figures is

extremely terse and opaque, but it is clear that each involves

a particular kind of emotional or dramatic content (e.g. the

presentation of affection, of haughtiness, or of one of

Bharata's eight rasas) .

In Dandin's treatment of the same four alamkāras, their

anomalous, content-based character is likewise apparent.

Dandin treats three of these figures (preyas, rasavat, and

urjasvin) as a group, defining them in a single verse. The

definitions he provides for these figures are (like Bhamaha's

definition of rasavat) purely tautologous: "Preyas

['affectionate' ] is a very affectionate statement; rasavat

23This creates interpretive problems for later Alamkārikas who base their treatment of these figures on that of Bhamaha. In at least one case, that of samahita, it is not at all clear what Bhamaha's understanding of the alamkara is. Bhamaha's example of the figure is as follows: "Narada appeared before the ksatriya-women going to be united with Rama" (in the absence of context, the point of the example is not entirely clear, but it would appear that Narada is to take or direct the women to Rama's location). Dandin and Udbhata have very different definitions of samähita, but both seem designed to fit this example. Dandin (2.296) takes samahita to mean "serendipity" -- the fortuitous "coming together" of several factors so as to produce a desired result; Udbhata (4.7) understands samahita as the "completion", i.e. the subsidence, of some rasa or rasa-like emotional state. Bhamaha's example can be seen as meeting both definitions: the meeting of the ksatriya women with Narada is fortuitous for them, and puts an end to their anxiety to find Räma.

52

['possessing rasa' ] is adorned with rasa; urjasvin ['haughty' ]

is that in which there is strong pride -- and these three are

joined with intensity. "24 All three alamkäras consist in the

depiction of a particular emotional content. As Dandin

explains it, the figure samahita is not specifically concerned

with the emotional content of a verse, but is nevertheless

content-based: "For a person undertaking some task, the

attainment of the means [to achieve] that [task], by the power

of fate -- that is what they [the Alamkārikas] call samāhita

['coming together']."25 This alamkära, then, consists in the

depiction of a particular kind of narrative event; the

fortuitous achievement, by some character, of the means to a

desired end.

It is Udbhata, in his Kāvyalamkārasarasamqraha, 26 who

first explicitly explains all four of these alamkāras in

connection with the dramaturgical concept of rasa. The end of

the eighth and the beginning of the ninth century saw the

beginning of a renaissance in Kashmir, in both Alamkārasāstra

24preyah priyatarākhyānam rasavad rasapeśalam/ ūrjasvi rūdhāhamkāram yuktotkarsam ca tat trayam// Dandin, Kavyadarsa, 2.273 (unless otherwise noted, all translations are my own).

25Kāvyādarsa 2.296.

26Udbhata's only other known work on Alamkārasāstra, his commentary on Bhamaha's Kavyalamkara, survives only in a single, extremely fragmentary, manuscript (cf. Raniero Gnoli ed., Udbhata's Commentary on the Kavyalamkara of Bhamaha, Rome, Instituto Italiano per il Medio ed Estremo Oriente, 1962). That portion of the commentary which would have dealt with this group of alamkaras unfortunately, lost. (on Bhamaha 3.5-10) is,

53

and Natyasastra, and Udbhata played an important role in both

areas; he was, as far as is known, the first person to write

on both Alamkārasāstra and Nātyasāstra. 27 In his work, and

in those of the later Alamkarikas of Kashmir, we see an

increasing erosion of the distinction between poetic and

dramaturgical theory; questions originally raised in the realm

of dramaturgy, particularly the question of rasa, come to play

an increasingly prominent role in Alamkārasāstra during this

period.

Udbhata's analysis of preyas (which he calls preyasvat,

"possessing affection"), rasavat, ūrjasvin, and samāhita

represents the most sophisticated attempt to "formalize" these

apparently content-based (and therefore anomalous) figures.

As he explains them, each of these figures entails not the

depiction of a particular emotional content, but a particular

mode or aspect in which any emotional state may be presented.

He defines preyasvat as follows: "The wise call that poem

27Udbhata was a major figure (sabhäpati -- leader of the royal assembly of scholars) in the court of JayapIda of Kashmir (r. c. 776-807 AD) -- cf. Rajataranqini 4.495 (on Jayapida's date, see Stein's translation of Rajatarangini, Introduction, pp. 94-96). Udbhata's dramaturgical work, a commentary on Bharata's Natyasastra, is lost, but is known from references in Abhinavagupta's commentary on the same work (cf. Kane, History of Sanskrit Poetics, op. cit., p. 49). He appears to have written on Carvaka (Materialism) as well; he is presumably the same Udbhata mentioned by Cakradhara (in his commentary [Granthibhangal on Jayantabhatta's NyayamañjarT) and in Devasūri's Syadvadaratnākara commentary Brhaspati's Nyayamanjarī ed. Gaurinath Sastri, Sivakumarasāstrigranthamalā on (lost) as the author of a Lokāyatasūtra -- cf.

5, Vol. 1, p. 100, Vol. 2, pp. 195, 257, and 262; also, NyamanjarIgranthibhanqa (ed. Nagin J. Shah, L.D. Oriental Series 35, Ahmedabad, 1972), Introduction, pp. 5-7.

54

preyasvat which is composed with [aesthetic factors] such as

anubhavas, 28 which are indicative of the emotions [bhavas],

attraction [rati] etc. 29 The dramaturgical tradition

distinguishes between bhavas, emotional states, and rasas, the

aesthetic "moods" developed from them in a dramatic

performance, and Udbhata explains the relation between rasavat

and preyasvat in terms of this distinction.

In the figure rasavat, one of the nine30 possible

aesthetic moods (rasas) is fully developed by means of the

appropriate factors (vibhavas and so on) ;31 in the figure

preyasvat, an emotion (any emotion) is depicted, but not

developed into a full-fledged rasa. Udbhata does not explain

28anubhavas are those visible symptoms (crying, laughter, etc.) which are indicative of particular emotional states. The other major factors recognized by the dramaturgical tradition as involved in the portrayal of emotional states are vibhavas ("stimulants", factors which are understood to create an emotional response in a character -- e.g. a beloved person [in the case of love] or a dangerous animal [in the case of fear] ) and vyabhicaribhavas ("variable emotions", those transient emotional states associated with a more persistent underlying emotion -- e.g. [in the case of love] eagerness, anxiety, and so on).

29ratyādikānām bhāvanam anubhavādisūcanaih/ yat kavyam badhyate sadbhis tat preyasvad udahrtam// Udbhata, Kåvyalamkarasarasamqraha, 4.2 (Bombay Sanskrit and Prakrit Series No. LXXIX, 'ed. N.D. Banhatti, Poona, Bhandarkar Oriental Research Institute, 1982 [reprint], p. 50).

30To the eight rasas listed by Bharata (above, p. 44), Udbhata adds a ninth, santa ("quiescence") . This ninth rasa is accepted by most of the later Ālamkārikas.

31Ibid., p. 52.

55

the basis for differentiation,32 but his examples for these

alamkäras make the nature of the distinction clear, at least

as applied to srīgara- ("erotic") rasa and its underlying

emotional state, rati ("attraction"). Udbhata's example of

rasavat33 depicts Siva's love for Parvati -- a clear case of

srīgāra-rasa; his example of preyasvat34 describes Pārvatī

affectionately playing with a young deer, treating it as if it

were her own child. The latter example depicts attraction,

but attraction which, unlike that in the example of rasavat,

is not directed toward an appropriate sexual partner, and thus

cannot manifest itself as full-fledged srīgāra. 35 Thus the

bhava rati is conveyed without developing into a rasa. As

Udbhata's commentators explain, his definition of the figure

32His commentator Pratiharenduraja explains that only when there is "full development" or "ripening" (pariposa) of rati and other bhavas do they become rasas -- ibid., p. 52.

33Ibid., p. 53.

34 Ibid., p. 52.

35This explanation of the relationship between sexual and non-sexual love is accepted by Mammata and, following him, most of the later Alamkärikas. According to Mammata, when love is directed toward such objects as a god, a king, a parent, or a child, the bhava rati will be suggested; only when love is directed toward a "beloved" (kanta) -- i.e. an appropriate sexual partner -- is the rasa śrngära suggested (cf. Kavyaprakasa, Mysore: Oriental Research Institute Series 120, Vol. I, pp. 284-5). A few Ālamkārikas, most notably Rudrata and Bhoja, take a different approach, making preyas ("affection") a full-fledged rasa in its own right, rather than simply of Kāvyālamkāra a permutation Rudrata, 15.17-19 srngara (cf.

Śrīgāraprakāsa [Madras, Punarvasu, 12.3, and V. Raghavan, Bhoja's 1978 (third ed.) ], pp. 402ff; also V. Raghavan, The Number of Rasa-s [Madras, Adyar Library and Research Centre, 1975 (third ed.) ], pp. 119ff.).

56

preyasvat demands that we understand the word preyas

("affection") in the figure's name as an indicator

(upalaksana) of emotional states in general.36 Any bhava

which, for whatever reason, fails to develop into a rasa, will

fall within the scope of this figure.

Similarly, Udbhata's treatment of ūrjasvin makes this

figure consist not of the introduction of a specific emotional

content, but of a particular mode in which any emotion at all

may be presented. He defines the figure as follows: "The

composition of rasas or bhavas which are manifested

inappropriately due to some cause such as desire or anger is

called ūrjasvin. "37 His example of the figure describes the

love-struck Siva's momentary impulse to seize Pārvatī by

force. Ūrjasvin, then, entails the presentation, not of any

particular emotional state (such as pride), but rather of any

kind of emotional excess; any rasa or bhäva, if manifested in

defiance of accepted norms and limits, will constitute a case

of ūrjasvin.

Udbhata explains samahita too as an aspect in which any

36Cf. Pratiharendurāja's Laghuvrtti (ibid., pp. 51-2) and Udbhatalamkaravivrti (printed in Kavyalamkarasarasamgraha, ed. K.S. Ramaswami Sastri, Gaekwad's Oriental Series 55, Baroda, 1931), p. 32. The latter commentary is anonymous but, as Ramaswami Sastri demonstrates (ibid., intro., pp. 34-38), it is almost certainly the work of Ruyyaka's father Tilaka (occasionally referred to in J Ruyyaka's Alamkarasarvasva) . Jayaratha's commentary on

37Anaucityapravrttānām kāmakrodhādikāranāt/ bhavanām ca rasanam ca bandha urjasvi kathyate// Udbhata, Kavyalamkarasarasamgraha 4.5 (op. cit., p. 54).

57

emotional state may be presented: "Samahita is the composition

of the cessation of the operation of rasa, bhava, or the

semblance of these [tadabhasa], in a form devoid of the

anubhavas of any other [rasa, bhava, or abhasa] . "38 The

portrayal of the subsidence of an emotional state of any sort,

if it is not accompanied by the depiction of conditions which

indicate the arising of some new emotional state, will fall

within the scope of this alamkāra. Udbhata's example

describes Siva's casting off of the disguise he had taken on

while observing ParvatI; as the commentator Pratiharenduraja

explains, this verse shows the subsidence of "concealment"

(avahittha), one of the variable emotional states (vyabhicāri-

bhavas) associated wich srīgara-rasa.39

So, Udbhata explains all four of these alamkāras in terms

of rasa and other closely related emotional phenomena. Yet he

clearly tries to avoid tying any of these figures to a

specific emotional content -- preyasvat consists not

specifically in the depiction of "affection" (preyas), but in

the depiction of any bhäva at all in such a way that it does

not give rise to rasa; likewise ūrjasvin and samāhita consist

38rasabhāvatadābhāsavrtteh prasamabandhanam/ anyānubhāvanihsūnyarūpam yat tat samāhitam// Kavyalamkarasarasamgraha 4.7 (ibid., p. 56) . The "semblances" or abhasas of rasas or bhavas are those rasas or bhavas which are inappropriately manifested, as in the alamkāra ūrjasvin. The categories of rasa- and bhava-abhasa play an important role in the rasa theory of the later Alamkärikas; both the term and the concept seem to have been coined by Udbhata in his explanation of urjasvin.

39Laghuvrtti, ibid., p. 56.

58

in the depiction of the inappropriate manifestation and the

cessation (respectively) of any emotional state at all.

However, even as they are explained by Udbhata, these

alamkāras are still content based; Udbhata greatly expands the

scope of these alamkaras, but each still consists of a

statement about something in particular (e.g. an

inappropriately manifested emotion, the cessation of an

emotion), rather than a particular kind of statement (e.g. a

simile) which may be used to describe anything at all.

The categories of emotional experience Udbhata elaborates

in connection with these alamkāras (rasa, bhava, rasabhasa and

bhāvābhasa, rasa-, bhava-, and abhasa-prasama) are adopted by

Ānandavardhana and subsequent Alamkārikas,40 but Vamana and

Rudrata, Udbhata's successors among the pre-Anandavardhana

Ālamkārikas, do not accept his treatment of these alamkāras --

both abandon altogether the attempt to account for rasa and

related emotional phenomena within the framework of the

alamkāras.

Vamana (c. 800 AD) was, like Udbhata, active at the court

40Cf. Dhvanyaloka 2.3 (Kashi ed., pp. 175ff). Abhinavagupta expands the list of categories to include the "arising" (udaya) , the "mixing" (sandhi), and the "juxtaposition" (sabalata) of emotional states; some later alamkara-theorists introduce new figures corresponding to these additional categories (cf. Ruyyaka, Alamkarasarvasva, Kāvyamāla 35, pp. 190ff; Visvanātha, Sahityadarpana, ed. Sālagrāmasāstrī, Delhi, Motilal Banarsidass, 1982 [reprint], pp. 367-368; Appayadīksita, Kuvalayananda (Vidyabhavana Samskrta Granthamala 24, pp. 272-274) . 105-107

59

of Jayapida of Kashmir,41 so the two must have been at least

roughly contemporary. Unlike Udbhata, Vamana has very little

to say regarding rasa. In his taxonomy of the poetic figures,

Vamana simply omits rasavat, preyasvat, ūrjasvin, and

samahita. Rather than treating rasa and other varieties of

emotional content as figures of speech, Vamana incorporates

them in his scheme of semantic qualities (artha-gunas) .

Vamana defines the semantic quality of kanti ("splendour") as

"possession of resplendent rasas" (diptarasatvam) 42 He has

nothing more to say regarding rasa, and nothing at all to say

regarding those rasa-related phenomena (bhāva, rasābhāsa,

etc.) described by Udbhata ..

While Vamana shows little interest in the role of rasa in

poetry, his work serves in one important respect to make the

status of rasa-theory in Alamkārasastra a more pressing

question -- Vamana essentially obliterates the categorical

distinction between drama and poetry. We have seen that

Bhamaha and Dandin, while they acknowledge in passing the

literary status of drama, allow for a division of labor

between poetic and dramatic theory, and recognize the analysis

of drama as the proper domain of Natyasastra, rather than

Alamkarasastra. While Udbhata, in his commentary on Bhamaha's

Kāvyalamkāra, quotes many example verses from existing poetic

41Cf. RajataranqinI 4.497.

42Kavyālamkārasutra 3.2.15 (Kashi Sanskrit Series 209, Varanasi, Chaukhambha Sanskrit Sansthan, 1989 [fourth ed.]), p. 116.

60

works, he cites no verses from drama. Vamana, by contrast,

draws freely from plays43 as well as non-dramatic poetry to

find examples for the poetic features he explores. For

Vamana, drama is not merely one type of poetry among others;

it is, in fact, the very best kind of poetry.44 By breaking

down the genre distinction between dramatic and non-dramatic

poetry, Vamana helps to erode the basis for the separation

between the two modes of critical analysis devised to explain

them; if there is no fundamental distinction between dramatic

and non-dramatic poetry, there is no reason why the role of

rasa should be any less significant in the latter than in the

former. Vamana's practice of freely mixing examples drawn

from dramatic and non-dramatic poetry is adopted by

Anandavardhana and most of the later Alamkarikas; for these

authors, rasa-based analysis is no less relevant to non-

dramatic poetry than to drama, and alamkara- and guna-based

analysis is no less relevant to drama than to non-dramatic

poetry.

The increasing prominence of rasa in the analysis of non-

dramatic poetry can be seen clearly in the work of Vamana's

nearest successor, Rudrata (c. 825 AD) .45 Rudrata is the

43E.g. Kālidāsa's Abhijñānasākuntala (1.2.11, 3.2.6, 4.2.14) and Vikramorvaśīyam (3.1.10, 4.3.29), Bhavabhūti's Mahavīracarita (1.2.12) and Uttararamacarita (4.3.6).

44Cf. Kāvyālamkārasūtra 1.3.30.

45Rudrata gives no information on the time and place in which he was writing, but he is generally thought to have been a Kashmiri, largely because of his name (the suffix -ta being

61

first of the Ālamkarikas to make rasa a primary category of

analysis (rather than incorporating it within the framework of

alamkaras or gunas). Like Vamana, he omits the alamkāras

rasavat, preyasvat, urjasvin and samahita from his discussion

of the figures. The first eleven chapters of Rudrata's

Kavyalamkara deal with alamkaras, poetic flaws (dosas) and

other formal aspects of poetry, very much in the manner of

earlier works on poetics. Having concluded his discussion of

these matters, he introduces rasa as an altogether separate

topic. The remaining portion of his work (apart from a brief

concluding discussion of genre-typology) deals with the role

of rasa in poetry, along with related questions of plot and

characterization, largely (thought not exclusively) modelled

on the treatment of the same topics in the Nātyasastra.

Rudrata makes no attempt of any kind to connect his

analysis of rasa and related factors with his earlier

discussion of alamkäras and other formal poetic features. The

disjunction between these two sections of Rudrata's

Kāvyālamkāra is absolute; so much so that if its first eleven

most common in Kashmiri names) -- cf. Kane, History of Sanskrit Poetics, op. cit., p. 153, De, History of Sanskrit Poetics, op. cit., Vol. I, p. 85, and Gerow, Indian Poetics, op. cit., p. 239). In any case, he Anandavardhana. seems to have been known to In the vrtti on Dhvanyaloka 2.21 (Kashi ed., p. 239), Anandavardhana refers in passing to vakrokti as an alamkāra involving the use of punning expression; Rudrata was the only one of the early Alamkarikas to define the figure vakrokti in this way (cf. Rudrata, Kavyalamkāra 2.14-17) . Thus it would appear that Anandavardhana was familiar with Rudrata's treatment of this figure (contra De -- cf. History of Sanskrit Poetics, Vol. I, p. 85).

62

chapters were severed from its last five, there would be

nothing to indicate that either segment was part of a larger

work. What Rudrata presents, then, are two entirely separate

schemata in terms of which poetry may be analyzed; one dealing

with alamkaras and other formal aspects of poetry and modelled

on the work of Bhamaha and the other early Ālamkarikas, and

one dealing with rasa and related elements of narrative

content and drawing largely on the dramaturgical theory

embodied in Bharata's Natyasastra. Faced with the problem of

integrating content-based rasa-analysis with formal alamkāra-

and guna- analysis, Rudrata simply abandons any effort at such

integration. In this way he avoids the difficulties faced by

the early Alamkärikas in their attempts to explain rasa in

terms of gunas and alamkaras, but only at the expense of

introducing a certain incoherence; he delineates two distinct

modes of poetic analysis, and gives absolutely no indication

as to how, if at all, these two can be related to one another.

Conclusion

Even in the earliest works of Alamkārasastra, the status

of rasa is already problematic; this can be seen in Bhamaha's

anomalous treatment of rasavat and the other emotional-

content-based alamkāras. The problems associated with the

attempt to treat rasa in formalist terms become more

pronounced in the works of the early ninth-century Kashmiri

Ālamkārikas, as the division between poetic and dramaturgical

theory begins to erode and the question of rasa's role in non-

63

dramatic poetry becomes more pressing. This trend reaches its

height in the work of Rudrata, who altogether abandons the

attempt to formulate an integrated approach to poetics,

leaving formalist and rasa-based analysis as two separate and

unrelated analytic modes.

It is against this background that Anandavardhana's

revolutionary reconceptualization of Alamkārasāstra must be

viewed. The most obvious strength of Anandavardhana's

approach to literary theory in comparison with that of his

predecessors is that it successfully integrates formalist and

rasa-based analysis in a single, coherent framework. It does

this by abandoning the particularistic aesthetic which had

informed the work of the early Ālamkārikas. For

Ānandavardhana, poetry is not simply a text made beautiful by

the incorporation within it of a variety of individually

beautiful elements. The individual phonetic and semantic

components which make up a poem are to be valued not for their

intrinsic beauty, but for their ability to contribute to the

overall purpose of the text of which they form a part; this

overall purpose, at least ideally, will be the effective

communication of some particular rasa. Thus, in

Anandavardhana's work, rasa is not simply one more poetic

element among others; the need to convey rasa becomes the

central organizing principle which guides (or ought to guide)

poets in their selection and deployment of alamkāras and other

formal poetic elements, and provides a clear criterion in

64

terms of which the aesthetic usefulness of the specific

features of any poetic work may be judged.

This re-envisioning of the poem as a functionally unified

text with the communication of rasa as its overriding goal

does not merely provide a useful means of integrating rasa-

based poetic analysis with the analysis of alamkāras and gunas

which had so far been the primary focus Of interest in

Alamkāraśāstra. It demands an entirely new understanding of

the nature of poetic texts and of the imperatives facing

practitioners of Alamkārasastra. If poems are to be seen not

as texts made beautiful by the incorporation within them of

individually beautiful phonetic and semantic features, but as

functionally unified wholes in which all elements operate

together to achieve single, coherent aesthetic effect, then

poetic analysis cannot limit itself to the study of

particular, isolable, linguistic elements employed in poetic

texts, but must concern itself with the functional relations

between these individual poetic components, and the factors

which govern their ability to further (or detract from) the

overall aesthetic objectives of the works in which they are

employed.

As has already been noted, Anandavardhana's view of

unifunctionality as the basis of textual coherence, while new

to Alamkāraśāstra, is not without precedent in the world of

Sanskrit linguistic theory. The practitioners of Mīmāmsa (a

school of textual exegesis concerned chiefly with the

65

interpretation of the Vedas) had long before developed a

system of textual interpretation which took unity of function

as the principle which integrates the individual elements of

any meaningful text into a coherent whole. As we shall see,

Anandavardhana brings into the realm of literary analysis the

Mīmamsakas' general understanding of textual coherence, along

with many of the specific interpretive techniques and concepts

they derive from it. Before we turn to examine

Ãnandavardhana's work and its impact on the subsequent

development of Alamkārasāstra in Kashmir, we must briefly

explore the interpretive theory and method of the MImamsakas.