Pramāṇa · Search · About · API

A History of Sanskrit Literature c25: XVIII. Theories of Poetry

A History of Sanskrit Literature c25

Language
English

Contents

XVIII. Theories of Poetry

1. The Beginnings of Theory on Poetry

c25.s1.1

IT is very possible to exaggerate the effect of theories of poetics¹ on Indian poetry and to ignore the fact that in India as elsewhere the poets set the models on which theory was built, and that it was only gradually that the effect of the textbooks on poetics came to be of ever-increasing importance. It is little short of absurd to imagine Kālidāsa as laboriously striving to conform to rules which in his time were, to the best of our knowledge, only in process of formulation, and which in any case were, as we can see from our extant sources, always being laid down with distinct divergences of emphasis and detail. Of the age of the study of poetics we can say little, but the fact that Pāṇini does not mention Alaṁkārasūtras, while he does recognize Naṭasūtras, certainly suggests that dramaturgy came before a general survey of poetics, even if we do not believe that Pāṇini knew a fully developed drama. With this accords the fact that, beyond vague references to Kāçyapa and a Vararuci, and Yāska’s knowledge of discussions of similes, Upamās,² we have no certain information on poetics until it occurs as a subordinate element in chapter xvi of the *Bhāratīya Nāṭyaçāstra*, which is essentially a treatise of dramaturgy and which may be placed conjecturally somewhat earlier than Bhāsa and Kālidāsa, though there is no strict proof of date. The great merit of this treatise, a compilation unquestionably from previous works, is that it develops the doctrine of sentiment, Rasa, with its eight subdivisions as erotic, comic, pathetic, and those of horror, heroism, fear, disgust, and wonder. Sentiment is a condition in the mind of the spectator of a drama, or, we may add, the hearer or reader of a poem, produced by the emotions of the characters, and the emotions, Bhāvas, are excited by factors which may either be the object of the emotion, as the loved one is in the case of love, or serve to heighten it, as does the spring season. The emotions manifest themselves in effects of various kinds, and they are essentially distinct in psychological character among themselves, while the sentiments, though subdivided according to the emotions which excite them, are nevertheless essentially one in feeling, and this feeling, which later authorities seek more clearly to define, is a special purely aesthetic emotion comparable to the bliss obtained in contemplation of the absolute by the intellect which can comprehend it.¹

c25.s1.2

¹ See S. K. Dé, *Sanskrit Poetics* (1923–5); P. V. Kane, *Sāhityadarpaṇa* (1923); Hari Chand, *Kālidāsa et l’art poétique de l’Inde* (1917); V. V. Sovani, *Bhandarkar Comm. Vol.*, pp. 387 ff.; Trivedī, pp. 401 ff.

c25.s1.3

² *Nirukta*, iii. 13; cf. Pāṇini, ii. 1. 55 f., 3. 72.

c25.s1.4

¹ See Keith, *Sanskrit Drama* (1924), pp. 314 ff.

c25.s1.5

This, however, is not the side of the *Nāṭyaçāstra* which was fated to elicit the chief attention of writers on poetics as opposed to dramaturgy. Poetics developed, if it did not originate, in distinction from dramaturgy, and writers on it were long content to refer merely to that science. The topics which were to engage writers on poetics, however, appear in elementary, though not undeveloped, form in the *Nāṭyaçāstra*. It recognizes four figures of speech, the simile, Upamā, the metaphor, Rūpaka, the Dīpaka, in essence the use of one predicate to many subjects or one subject to many predicates, and the Yamaka, repetition of syllables or alliteration. There is no distinction of figures as those of sound, Çabdālaṁkāra, and of sense, Arthālaṁkāra, and it is significant of early poetry that there are given ten kinds of Yamakas, but only five of Upamās. The Yamakas remain prominent in the older school of poetics, including Bhaṭṭi, Daṇḍin, Vāmana, Rudraṭa, and the *Agni Purāṇa* section on poetics, but Bhāmaha already admits but five kinds and Ānandavardhana and Mammaṭa make it clear that the figure has no real aesthetic importance, though in later as well as older poetry, for instance, the *Ghaṭakarpara*, it is freely used, serving in lieu of rhyme. Further, serving like the figures to bring out the sentiment, are given the ten qualities and the ten defects; it is characteristic of the beginnings of the science that the defects are given positively and the qualities given as the negation of the defects, while in fact it is impossible thus to connect the two lists. Moreover, the details of the lists are obscure, and differently interpreted both by the later writers on poetics and by the commentators on the Çāstra. On one view¹ the defects are: absence of a complete meaning; incongruity with the context; tautology; ambiguity; violation of syntactical regularity; grammatical errors; break of metrical rules as to pause; misuse of long or short syllables in metre; breach of euphonic rules; and inconsistency as to place, time, artistic usage, popular belief, logic, or science. The qualities are: Çleṣa, possibly in the sense of suggested sense; Prasāda, clearness; Samatā, evenness implying ease of comprehension; Samādhi, superimposition of something special in the sense; Mādhurya, sweetness; Ojas, strength arising from the use of compounds with respect to suitable concatenations of letters; Saukumārya, smoothness arising from happy metres and conjunctions; Arthavyakti, explicitness of sense; Udāra, elevation of subject and sentiment; and Kānti, loveliness, delighting the mind.

c25.s1.6

¹ Bhāmaha, iv; logical faults are given in v. For Bharata's list see xvi. 84 ff.

c25.s1.7

Of developments after the Çāstra we know nothing definite, and we can only guess at the stages by which new figures were found out. If we can take Bhāmaha's account as helping us historically—which is a pure assumption not suggested in any way by that author—we may hold² that the first step was to distinguish Anuprāsa, alliteration, from Yamaka, the former affecting only single letters, the latter involving the repetition of syllables. But it is much more dubious if the fact that Bhāmaha mentions after this set of five a set of six has any chronological conclusion, and the figures themselves are rather more complex than can be supposed to have been early. They are: Ākṣepa, paraleipsis, the denial of one thing to imply another; Arthāntaranyāsa, corroboration, the adduction of some instance or principle to prove an assertion; Vyatireka, contrast by dissimilitude; Vibhāvanā, abnormal causation, when something comes about by some unusual reason; Samāsokti, brevity, suggestion by metaphorical expressions; and Atiçayokti, hyperbole. Possibly to this period has been referred the figure Vārttā, which, however, was not generally accepted, though Daṇḍin perhaps,³ treats it as a sort of Hetu, cause. Our trust in the whole theory is seriously undermined when we find that to a third period of development are assigned three new figures: Yathāsaṁkhya, relative order; Utprekṣā, poetical fancy; and Svabhāvokti, description of the nature of a thing in its reality as appreciated by the poetic imagination; and that the fourth period is made to recognize the large number of figures, twenty-four more, in Bhāmaha. What is really clear is that the *Bhaṭṭikāvya*,¹ Daṇḍin, and Bhāmaha all had before them a large number of figures which they treat in slightly different ways, Bhāmaha for instance rejecting the forces of cause, Hetu, Sūkṣma, and Leça, accepted by Daṇḍin. To assert even a common source for Daṇḍin and Bhāmaha as opposed to Bhaṭṭi is beyond our means of proof, and to ascribe to Medhāvin the invention of Utprekṣā is quite invalid.

c25.s1.8

² Jacobi, SBA. 1922, pp. 220 ff.

c25.s1.9

³ If Jacobi rightly refers ii. 244 to it.

c25.s1.10

¹ Cf. on Canto x Nobel, in *Muséon*, xxxvii.

2. The Early Schools of Poetics

c25.s2.1

In Daṇḍin we come, as usual in Indian scientific literature, to an authority who used freely many predecessors whose works are lost, and who, therefore, presents us with a fully developed and elaborate doctrine. Daṇḍin was doubtless the author of the *Daçakumāracarita* and his relation to Bhāmaha has been keenly discussed.² The difficulty of decision rests on the fact that both authors can be made out to be attacking each other's views, but that there is nothing whatever strictly to prove that they are not dealing with views expressed by some predecessor of the other, as we know for certain in the case of Bhāmaha that he used³ Medhāvin, who must have expressed opinions similar to those assailed by Daṇḍin. It is, however, on the whole, probable that Bhāmaha knew Daṇḍin, while Daṇḍin did not use him, and with this agrees the generally less refined views of Daṇḍin as in his enumeration of thirty-two kinds of simile, which Bhāmaha reduces to four. Daṇḍin's rejection of the difference between Kathā and Ākhyāyikā seems thoroughly sound, while Bhāmaha's defence seems specially directed against Daṇḍin. It is striking also that Daṇḍin never notices one of the many verses adduced by Bhāmaha to expound his views. The matter is not, indeed, of the highest importance, for it is not supposed in any case that Daṇḍin lived long after Bhāmaha, who certainly used the works of Uddyotakara (*c.* 650) and probably knew the *Nyāsa* of Jinendrabuddhi (*c.* 700). On the whole, having regard to the facts regarding the *Daçakumāracarita*, which suggests that it precedes Subandhu and Bāṇa, we may place Daṇḍin some generations before Bhāmaha.

c25.s2.2

² Kane, *Sāhityadarpaṇa* (1923), pp. xxv ff.; M. T. Narasimhiengar, JRAS. 1905, pp. 535 ff.; Pathak, JBRAS. xxiii. 19; IA. xli. 236 ff., support Bhāmaha's posteriority against Trivedī, IA. xlii. 258 ff. R.; Narasimhachar, IA. xli. 90 ff.; xlii. 205; Nobel, ZDMG. lxxiii. 190 ff.; Hari Chand, *Kālidāsa*, pp. 70 ff.; Jacobi, *loc. cit.*

c25.s2.3

³ ii. 40, 88; Medhāvirudra, Nami on Rudraṭa, xi. 24. Cf. *Kāvyamīmāṁsā*, p. 12.

c25.s2.4

To Daṇḍin poetry appears under the metaphor of a body of words determined by the sense which it is desired to set out, and that body is ornamented, the term Alaṁkāra here being used in the most general sense to cover anything which lends beauty to the poem as ornaments do to the human body. A poem may consist of verse, prose, or both, as in the drama and the Campū; no Indian writer on poetics allows himself to be led astray into demanding verse form as a condition of poetry. This, of course, was a natural conclusion from the fact that law, medicine, astronomy and astrology, grammar, and philosophy had all been composed in verse, so that outer form was obviously no criterion between the literature of knowledge and that of power. Of verse forms Daṇḍin enumerates the Sargabandha or Mahākāvya, the characteristics of which we have already noted; Muktaka, single verses; Kulaka, groups of up to five verses; Koça, unconnected verses by different authors; Saṁghāta, similar verses by one author. Of prose he mentions Kathā, Ākhyāyikā, and Campū, recognizing as current the difference between the first two, but rejecting it as quite artificial and not even in accordance with practice. The use of different languages is admitted, Sanskrit, Prākrit, Apabhraṁça, and mixtures of these being allowed, seen in the Mahākāvya in Sanskrit, in poems in the Skandhaka metre in Prākrit, in the Āsāra in Apabhraṁça, and the Nāṭaka, drama, in a mixed form.¹ Daṇḍin also recognizes the distinction between a poem to be heard and one to be seen; but refers to works on dramatic art for consideration of the latter.

c25.s2.5

¹ The sense of the terms is not given by Daṇḍin, and is dubious; the last may be poems in one metre. Osara is a *v.l.*

c25.s2.6

Of special interest is the new presentation of the doctrine of qualities. It is clear that before Daṇḍin there had developed the doctrine of schools or paths, Mārga, of poetry, and Bāṇa refers to four of them, as we have seen. Daṇḍin declares for the existence of two types, holding that subvarieties are incalculable, and he sets them against each other as the Vaidarbha and Gauḍa, the former the southern, the latter the eastern, the distinguishing marks being the presence in the former of ten qualities which the other does not usually accept. Daṇḍin shows clearly that these distinctions are not his own, and his descriptions are here and there suggestive of doubt on his own part as to what is meant, a doubt increased by divergences of view among the commentators. One quality, indeed, is admitted to be liked even by the Gauḍas, perspicuity of sense; if the ocean is referred to as red, that requires the addition of the words ‘through the blood of the serpents.’ But the merit of clearness, Prasāda, applicable to the use of words in a natural way, is not attractive to the Gauḍas; they like a phrase such as *anatyarjunābjanmasadṛkṣāṅko balakṣaguḥ*, ‘the white-beamed (moon) has a spot similar to the not-very-white water-born (lotuses)’, where the rare expressions are excused in the Gauḍa view by their being etymologically derivable. Udāratva signifies the presence in a sentence of a distinguished quality, thus giving elevation of style, as in:

c25.s2.7

*arthināṁ kṛpaṇā dṛṣṭis tvanmukhe patitā sakṛt*
*tadavasthā punar deva nānyasya mukham īkṣate.*

c25.s2.8

‘Once the sad eye of suppliants hath fallen on thy face, o king, it taketh there its abode, and gazeth not at the face of any other.’ Another explanation given by Daṇḍin himself makes elevation the result of the use of ornamental epithets such as *līlāmbuja*, toy-lotus, *krīḍāsaras*, play-lake, *hemāṅgada*, gold bracelet. Kānti is the grace of beauty, which is in harmony with nature, as opposed to the exaggeration, Atyukti, of the Gauḍa style; the two are neatly contrasted: the Vaidarbha has:

c25.s2.9

*anayor anavadyāṅgi stanayor jṛmbhamāṇayoḥ*
*avakāço na paryāptas tava bāhulatāntare.*

c25.s2.10

‘O maiden with faultless limbs, there is not space enough between thy creeper-like arms for the expansion of those swelling breasts.’ The Gauḍa exaggerates:

c25.s2.11

*alpaṁ nirmitam ākāçam anālocyaiva vedhasā*
*idam evaṁvidham bhāvi bhavatyāḥ stanajṛmbhaṇam.*

c25.s2.12

‘Surely the creator hath made this world too narrow, foreseeing not so great an extension of thy breasts.’ Samādhi denotes metaphorical expression, and Daṇḍin shows how words normally vulgar can be used if the sense be no longer literal, as in the case of *vam*, vomit, *niṣṭhīv*, and *udgṛ*, spit out.

c25.s2.13

These five qualities are clearly essentially connected with sense, a sixth, Mādhurya, sweetness, is defined as possessing Rasa, which here denotes rather tastefulness than sentiment as taken by Bühler among others, and this is a quality of sense as well as of sound, for it applies to the extent of forbidding the use of expressions suggesting vulgar ideas, requiring that love should be alluded to in decently veiled phrases. It, however, also has to do with tasteful arrangement of sounds, and in this there is a divergence between the styles, for the Vaidarbha likes the combination of harmonious sounds, while the Gauḍa prefers the more obvious and blatant device of alliteration outright. The Vaidarbha also demands Sukumăratā, gentleness, which means the use of syllables which are not rough sounding, while the Gauḍas like harsh sounds when they serve to accord with the sentiment expressed. Thus we have for the Vaidarbha the pleasing if in sense negligible:

c25.s2.14

*maṇḍalīkṛtya barhāṇi kaṇṭhair madhuragītibhiḥ*
*kalāpinaḥ pranṛtyanti kāle jīmūtamālini.*

c25.s2.15

‘Making circles of their tails the peacocks dance in the season of the clouds, uttering sweet cries.’ Contrast the Gauḍa fiery utterance:

c25.s2.16

*nyakṣeṇa kṣapitaḥ pakṣaḥ kṣatriyāṇāṁ kṣaṇād iti.*

c25.s2.17

‘In a moment the host of the warriors was destroyed by Paraçurāma.’ There is again a distinction as to Samatā, evenness: the Vaidarbha style likes the letters to be soft, harsh, or well mixed, but the Gauḍas do not object to unevenness, and the poetry aiming at brilliance or bombast of both sense and ornament (*arthālaṁkāraḍambara*) is recognized as having won fame. The Vaidarbhas also like Çliṣṭa, stability,¹ diction which is not loose, i. e. composed of easily pronounced syllables—while the Gauḍas do not mind the latter defect, provided it be alliterative;

c25.s2.18

¹ In i. 43 this seems the best sense; Lüders in Nobel’s *Indian Poetry*, p. 107, n. 12.

c25.s2.19

thus to express the common idea of a jasmine wreath and its attendant bees the Vaidarbhas say: *mālatīdāma laṅghitaṁ bhramaraiḥ*, the Gauḍas *mālatīmālā lolālikalilā*. Finally both styles like force, Ojas, consisting of lengthy compounds, or rather of a large number of compounds, both in prose and poetry in the Gauḍa view, in prose only in that of Vaidarbhan usage, though the latter would evidently sanction it if it was set off by short words as in:

c25.s2.20

*payodharataṭotsaṅgalagnasandhyātapāṁśukā*
*kasya kāmāturaṁ ceto vāruṇī na kariṣyati?*

c25.s2.21

‘Whose heart is not made lovesick by the sight of the western sky, whose garment, the evening sun, hangs on the slopes of the clouds that are her breasts?’ The poet recognizes that varieties of compounds are made by the mingling of syllables long and short.

c25.s2.22

Daṇḍin insists that to produce the effective poetry he has praised are necessary natural genius, which arises from impressions formed in earlier births, much study, and great application, and, recognizing that the first requisite may be unattainable, allows concentration on the second two. He then proceeds, in Book ii of the *Kāvyādarśa*, to define Alaṁkāras as those qualities which produce charm in poetry, some of which have been already mentioned in dealing with the difference of styles, while those common to both styles are enumerated in ii and iii, the figures of sense coming first, than those of sound, treated from our point of view at absurd length. The early state of Daṇḍin’s views is shown in his failure to distinguish quality and figure, and in his making no effort to explain the poetic effect of figures save by mere generalities. Nor has he any scheme of division of figures, and in a manner somewhat startling we find that he ranks as a figure, the first of all, Svabhāvokti, natural description as a thing appears to a poet. This figure—or rather ornament—is of a quite special kind, for it is classed as opposed to all the rest of the figures of sense, which are classed under Vakrokti, crooked, non-natural, figurative, speech. The meaning of the distinction must be that in the former case the poet, by his discernment, sees the essence of a thing—using that term in the widest sense, be it an individual thing, or a species, or a quality or action—and sets it out in plain speech: in the latter he describes not necessarily with special intuition, but with figurative language. He has already insisted, in his account of the qualities, on the supreme importance to the poet of the use of metaphor.

c25.s2.23

The actual list of figures¹ is a curious mixture including much that we should not reckon figures of a distinct kind, as well as figures more naturally so styled. We have in his order the simile in thirty-two varieties, the metaphor, the Dīpaka, Āvṛtti, repetition in the sense of the use, e. g., of four different verbs with one meaning as a quasi contrast to the Dīpaka, Ākṣepa, Arthāntaranyāsa, Vyatireka, Vibhāvanā, Samāsokti, hyperbole, poetic fancy, and then three figures rejected by Bhāmaha, Hetu, Sūkṣma, and Leça. These express cause, convey a meaning by adroit hint or gesture, or conceal something which has almost come to light; but Daṇḍin gives us an alternative view of Leça, a rebuke or eulogy. Then come order; Preyas, the expression of pleasure; Rasavat, the expression of one or other of the sentiments; Ūrjasvin, that of vigour; Paryāyokta, the expression indirectly of something which cannot openly be avowed; Samāhita, mentioning some fact which has come to afford aid to one’s end; Udātta, description of something noble or elevated; Apahnuti, seeming denial to affirm more strongly; Çleṣa, double meaning; Viçeṣokti, description of a special distinction; Tulyayogitā, putting like things together; Virodha, seeming incongruity; Aprastutastotra, indirect praise; Vyājastuti, praise concealed as censure; Nidarçana, reference to a like result; Sahokti, mention of two things as happening together; Parivṛtti, exchange of objects; Āçis, benediction; Saṁkīrṇa, mixing of figures; and Bhāvika. The latter is a quality applicable to the whole of a composition and expresses the purpose and mind of the poet; it reveals itself in the making of all the different elements of the plot aid one another to their end, the avoidance of needless qualifications, the description of things in their place, and the exposition of even a difficult matter by due regard to orderly exposition. This quality, we can see, would, if Daṇḍin had had any idea of order, have been conjoined with Svabhāvokti; we may compare Aristotle’s¹ ἐνέργεια. It is important to note that Daṇḍin expressly mentions the view of some authors which made a hyperbole implicit in every figure whatever, and he himself lays it down that in every form of Vakrokti the use of the Çleṣa enhances the beauty, thus according approval to the practice of Subandhu and Bāṇa and of himself in his less immoderate action in the *Daçakumāracarita*.

c25.s2.24

¹ Cf. Kane, *Sāhityadarpaṇa*, pp. 1 ff. Nobel (*Beitr. z. ält. Gesch. d. Alaṁkāraçāstra* (1911); *ZDMG.* lxvi. 283 ff.; lxvii. 1 ff.; lxxiii. 189 ff.) deals with some of the figures, but not always satisfactorily; his desire to place Bhāmaha before Kālidāsa leads him to deny the former's obvious reference to the *Meghadūta* (*Indian Poetry*, p. 15), even though he realizes that Kālidāsa was really the older.

c25.s2.25

¹ *Rhet.* iii. 10, 16. On metaphor, cf. c. 2.

c25.s2.26

Book iii of the *Kāvyādarça* develops at great length the doctrine of Yamakas, leading us to the stanza with one consonant, *n*, only; then follow riddles and finally the ten defects of poetry much as in the *Nāṭyaçāstra*. But nothing of real value is here found.

c25.s2.27

The doctrines of Daṇḍin found an echo and completion in those of Vāmana,² who is doubtless to be placed at the end of the eighth century.³ We have in him the emergence, however, of a new idea, that of the soul of poetry as opposed merely to the body. As later than both Daṇḍin and Bhāmaha he has a more developed idea of the nature of Kāvya; it is not merely words and meaning or sense, but there must be qualities and figures as well. But he also seeks to fit all the elements in Daṇḍin into a scheme, based on the doctrine of Rīti, a new word for style. The soul of poetry is style which is a specified arrangement of words, the term specified referring to distinction according to the qualities possessed which are the cause of charm in poetry, while the figures are ranged as things which add to the charm. He admits three kinds of Rīti, Vaidarbhī, Gauḍī, Pāñcālī, so styled because found among the local poets, but not due to local causes. The Vaidarbhī is perfect and has all the qualities. The Gauḍī is accorded the qualities of Kānti and Ojas, understood here in the sense of many compounds which are of great length, and high-sounding words, a statement illustrated by a famous stanza of Bhavabhūti. The Pāñcālī has sweetness and gentleness, Mādhurya and Saukumārya, like the style of Purāṇas. The Vaidarbhī is strongly insisted on, the other two disparaged, and a pure form of Vaidarbhī is expressly commended which uses no compounds, thus allowing full play to the qualities of sense. The qualities in Vāmana are now rearranged as qualities of sound and of sense, each having two aspects, with results far from satisfactory as regards clearness, and disadvantageous as departing from the normal use of the terms established in Daṇḍin. Under the quality beauty Vāmana includes the feature of implying sentiment, which Daṇḍin places in the figures Preyas, Rasavat, and Ūrjasvin, and perhaps in the quality Mādhurya, while the quality of perspicuity covers the Svabhāvokti of Daṇḍin. Under the qualities also room is found for the odd figure Bhāvika, whose awkward position in Daṇḍin's view has been noted.

c25.s2.28

² *Kāvyālaṁkāra* with *Vṛtti*, ed. KM. 15, 1895; Vāṇīvilāsa Press, 1909; trans. G. Jhā, IT. iii and iv.

c25.s2.29

³ Minister of Jayāpīḍa of Kashmir (779-813); Jacobi, ZDMG. lxiv. 138 f.

c25.s2.30

Vāmana's treatment of figures is important for his reduction of their importance as elements in poetry; the qualities are vital, the figures not: they are related rather to the body, word and meaning, of poetry than to the style which is the soul. Further, he insists that the simile lies at the bottom of all figures and to achieve this result has to omit various figures, in addition to those above mentioned, such as Udātta, Paryāyokta, and Sūkṣma, while others he defines differently. Vakrokti to him is a special mode of metaphorical expression, not the generic term for all figurative speech as in Daṇḍin.

c25.s2.31

As opposed to Daṇḍin we find in Bhāmaha's *Kāvyālaṃkāra*¹ a decided preference for a system which insists on the figures as the essential feature of the poetry whose body is word and sense. Bhāmaha definitely rejects outright the distinction of two styles, and the qualities which he does recognize are connected generally with poetry, not with any special style. Moreover, he shows the reduction of qualities to three, which is characteristic of later thought, though he does not specifically deal with the matter as do the later writers, who reduce Daṇḍin's ten to their categories. He mentions, however, as sweet, a poem which is agreeable to hear and has not too many compounds, and a clear poem is one which can be understood by even women and children; strength he understands as usual as connected with long compounds, and he implies that this is incompatible with clearness as well as sweetness. He has, however, no clear marking line between qualities and figures; he mentions clearness and sweetness in close proximity to his account of figures, and he describes Bhāvikatva as a figure or quality indifferently. He definitely insists on the distinction of figures into those of sound and sense, and he more or less vaguely is conscious of the doctrine which regards the essential feature of poetry to be figurative expression, Vakrokti. For the two-fold division of subject-matter of poetry favoured by Daṇḍin, which recognizes traditional matter and invention, he substitutes one admitting also foundation on the arts or sciences. His division of classes of poetry is five-fold, the Sargabandha, drama, Ākhyāyikā, Kathā, and detached verses, and he defends the distinction between Kathā and Ākhyāyikā on quite worthless grounds. But he insists that there is a common element in all poetry, Vakrokti, while he denies, accordingly, to Svabhāvokti the right to be styled a figure at all. This figurative expression he identifies with hyperbole, which is explained as an expression surpassing ordinary usage, meaning no doubt a poetical conception as opposed to the prosaic everyday conception of facts. Bhāmaha examines the various figures from this point of view, and his work in this regard was carried on by Udbhaṭa, the contemporary of Vāmana, whose *Alaṃkārasaṃgraha*¹ deals with forty-one figures, including three varieties of alliteration. His *Bhāmahavivaraṇa* is lost, and from Pratīhārendurarāja, pupil of Mukula, who wrote *c.* 950, and commented on Udbhaṭa, we learn little of importance. Of no historical importance is Bhāmaha's treatment of defects, in which he gives a new list of ten additional to the tradition alone (Book iv), while in Books v and vi he describes logical and grammatical errors in poetry.

c25.s2.32

¹ *Ed.* as App. viii to K. P. Trivedi's ed. of *Pratāparājayacobhūṣaṇa*, BSS. 1909.

c25.s2.33

¹ *Ed.* Jacobi, *JRAS.* 1897, pp. 829-53; BSS. 1925.

c25.s2.34

There are in Udbhaṭa hints of new views which later had some effect. The ascription to him of the doctrine that sentiment is the soul of poetry is due to an error, a verse cited by Pratīhārendurāja being wrongly ascribed to him. But he did lay some stress on the element of sentiment in poetry and he added to the list of eight of the *Nāṭyaçāstra* a ninth, the calm. Further, while he ignored, like Bhāmaha, the styles of Daṇḍin, he introduced a new classification based entirely on sound effects, primarily alliteration, in the shape of the theory of three Vṛttis, manners, classed as elegant (*upanāgarikā*), ordinary (*grāmyā*), and harsh (*paruṣā*). In treating figures he adds Dṛṣṭānta, exemplification, and Kāvyaliṅga, poetical causation, divides simile according to the grammatical form of expression, as by suffixes like *vat*, and starts the investigation of the relations of double meaning to other figures, which is later developed, as well as the complex issue of the different kinds of blending of figures, Saṁsṛṣṭi and Saṁkara.

c25.s2.35

Rudraṭa, who wrote before 900 and probably in the earlier part of the ninth century, the *Kāvyālaṁkāra*,¹ in sixteen chapters of Āryā verses, makes no innovation in theory, but belongs essentially to the school which, without scientific investigation, accepted as its duty the enumeration of figures. He seeks to divide figures on the base of sound and sense, and then to subdivide on principles of his own; under those of sound he classes figures on the basis of equivocation (*vakrokti*), paronomasia (*çleṣa*), pictorial effects (*citra*), alliteration and Yamakas; those of sense are based on reality, similitude, hyperbole, and coalescence. This results in a repetition of some figures under different heads, and his plan of division received no general acceptance, though Mammaṭa adopted some of his figures, and his new interpretation of Vakrokti as an equivoke based on paronomasia or intonation (*kāku*), though rejected by Hemacandra, prevails from Mammaṭa onwards over the wider sense of Daṇḍin or the narrower interpretation as a figure based on similitude of Vāmana. He generalizes and extends the manners of Udbhaṭa, in whom they seemed to be restricted to alliterative effects, by laying down five manners of letters (*varṇa*), sweet, harsh, pompous, dainty (*lalitā*), and excellent (*bhadrā*). But he accepts also the styles, Rītis, of Vāmana, though under the influence of Bhāmaha we find them looked at in a new light. They now number four, and the distinction is based on the use of compounds. The Vaidarbhī has none, verbal prefixes not ranking as compounding elements. Pāñcālī compounds up to three words, Lāṭīyā five to seven, and Gauḍīyā any number. His debt to Daṇḍin is seen in his dealing at great length with Yamakas and developing the idea of Citra, tricks in poetic form, such as Māgha declares to distinguish poetry in his day, but which Bhāmaha and Udbhaṭa ignore, while Udbhaṭa also passes over Yamakas. A novel feature is the introduction in four chapters of the theory of sentiment which, however, is in no wise brought into vital connexion with his subject, but stands in a merely formal collocation. He recognizes ten sentiments, adding the feelings of calm and friendship to the traditional list.

c25.s2.36

¹ Ed., with the comm. of Namisādhu, a Jain (1068), KM. 2, 1909. Rudraṭa is son of Vāmuka and is also called Çatānanda. His difference from Rudra Bhaṭṭa was proved by Jacobi, WZKM. ii. 151 ff.; ZDMG. xlii. 425. Rudra Bhaṭṭa is known to Hemacandra (p. 110); his *Çṛṅgāratilaka* is ed. Pischel, Kiel, 1886.

c25.s2.37

Still less important from the point of view of theory is the *Kāvyamīmāṅsā* of the dramatist Rājaçekhara (*c.* 900) which is a work in other regards of no small interest and originality. He conceives of the Kāvyapuruṣa, the spirit of poetry, son of Sarasvatī, and the Sāhityavidyā, science of poetics, who becomes his bride, the term Sāhitya being derived, we may believe, from the old doctrine of the union of word or sound and sense to make a poem, as laid down by Bhāmaha, Māgha, and others. He distinguishes carefully science, Çāstra, and poetry, and analyses the divisions of the former; he discusses at length the relation of genius, poetic imagination, culture, and practice in making a poet and classifies poets on this score. A further classification is based on the fact that a poet may produce a Çāstra, or a poem, or combine both in varying proportions, and of poets in the narrower sense he makes eight illogical groups. His own conception of poetry appears traditional; he defines it as a sentence possessing qualities and figures, and he accepts Vāmana's doctrine of styles which are the outcome of Sāhityavidyā's wanderings in diverse lands. The sources of poetry are touched on, and the subject-matter as concerned with men, divine beings, or denizens of hell is investigated. Very interesting is the discussion of borrowing from earlier works; it is recognized as justified by freshness of idea and expression, and elaborate illustration is given of thirty-two different ways of evading improper plagiarism. Important also is the consideration of poetical conventions, and we are given a geography of India and many remarks on the seasons with their appropriate winds, birds, flowers, and action. Rājaçekhara also gives curious details of the likings of different parts of India for certain languages and their mode of mispronouncing Sanskrit. The Magadhas and others east of Benares are blunt in Prākrit, good at Sanskrit, but the Gauḍas are thoroughly bad in Prākrit, the Lāṭas dislike Sanskrit but use Prākrit beautifully, the Surāṣṭras and Travaṇas mix Apabhraṅça with Sanskrit, the Draviḍas recite musically, Kashmirian pronunciation is as bad as their poetry is good, Karṇāṭas end up sentences with a twang, northerners are nasal, the people of Pañcāla sweet and honey-like. Women poets are recognized, and sex barriers despised, while of the ten grades of poets the rank of Kavirāja, held by Rājaçekhara, comes seventh even above the Mahākavi himself. Great stress is laid on the assemblies at which poets were judged and where the prize given by the king included crowning with a fillet and riding in a special chariot. The poet's paraphernalia is given, chalk, a board, palm leaves, birch bark, pen and ink.¹ More important is the insistence on the equal rights of all four forms of speech: Sanskrit; Prākrit, elegant, sweet, and smooth; Apabhraṅça also elegant, as loved in Marwar, Ṭakka, and Bhādānaka; and Bhūtabhāṣā current in Avantī, Pāriyātra, and Daçapura, while the people of the Madhyadeça used all equally well. The people of that land show also their admixture by their colours, brown like the easterners, dark like the southerners, white like the westerners, while the northerners are fair. When we add that he quotes extensively including the *Mahimnaḥstotra*, gives many fine verses and some anecdotes, and is usually lively if pedantic, the merits of his work can be appreciated.²

c25.s2.38

¹ On these matters see Bühler, *Indische Palaeographie*; Hoernle, JASB. lix. pt. i. no. 2; on the use of paper, Waddell, JRAS. 1914, pp. 136 f.; Haraprasād, *Report*, i. p. 7; on the claim of Indian writing as indigenous, not of Semitic origin, see Bhandarkar, POCP. 1919, ii. 305 ff.

c25.s2.39

² Ed. *Gaekwad's Oriental Series*, 1916. Many stanzas on poets by Rājaçekhara probably came from some lost work, perhaps the *Haravilāsa*; cf. Bhandarkar, *Report, 1887-91*, pp. ix ff; Peterson, JBRAS. xvii. 57-71; for an exposure of forged verses adduced to support an attack on Bhāsa's authorship see G. Harihar Sastri, IHQ. i. 370 ff.; K. G. Sesha Aiyar, 361; a bad case invites worse arguments; cf. Keith, BSOS. iii. 623 ff.; T. Gaṇapati Śāstrī, 627 ff.

3. The Doctrine of Dhvani

c25.s3.1

Rājaçekhara lived at a time when a new doctrine, that of Dhvani, tone, had been steadily winning its way to power. It is represented for us by the metrical Kārikās preserved in the *Dhvanyāloka*³ of Ānandavardhana of Kashmir (*c.* 850) with its super-commentary by Abhinavagupta, *Locana*.¹ The Kārikās assert that the doctrine is old, but if so we must assume that it had not won much success, and it may be that the author referred really to some not distant predecessor, justifying himself by the view that the doctrine was implicit in the older writers. His name was possibly but not certainly Sahṛdaya, which at best is merely an epithet, and he must have written early in the ninth century. At any rate by the ability of his commentators and by the adoption by Mammaṭa of the doctrine the new view won on the whole a dominant position in Indian poetics.

c25.s3.2

³ Ed. KM. 25, 1911; trans. H. Jacobi, ZDMG. lvi and lvii.

c25.s3.3

¹ Ed. KM. 25 (i–iii); Dé, Calcutta, 1923 (iv).

c25.s3.4

The theory finds its origin in the analysis of language and meaning. The phrase, a herdsmen's station on the Ganges, is obviously as it stands absurd; the denotation (*abhidhā*) gives no sense, and we are obliged to find a transferred sense (*lakṣaṇā*) which gives us the sense of a station on the bank of the Ganges. This shows the incompatibility of the literal sense as one factor, and the possibility of giving an allied meaning as another. But this is not all; there is brought to us by such a phrase deliberately used in poetry a sense of the holy calm of such a station situated on the sacred stream with all its associations of piety. This, it is contended, is not given by implication, but by the power of suggestion which is derived from the poet's purpose (*prayojana*) in applying the phrase. This doctrine of suggestion which the grammarians did not accept could be based on a philosophical opinion of the grammarians themselves. They recognized the Sphoṭa,² a mysterious entity, a sort of hypostatization of sound, of which action sounds were manifestations, and the same idea of the revelation of something inherent (*vyañjanā*) is found in the Vedānta, where all is a manifestation of the underlying reality, the Brahman or absolute. There were common-sense people³ who held that all could be put down to denotation; a word might be regarded like an arrow which could pierce armour and slay the foe in a single movement, without inventing new phases of operation, while yet others⁴ claimed that the signification, Tātparya, resulting from the taking of words together in a sentence explained all that was required, and others¹ again held that even this idea of Tātparya was needless, because the words had the power *per se* of conveying their relations with other words to make up a whole. A further school, which became more insistent later, declared that suggestion was not real, and that what was explained by suggestion ought to be accounted for by inference. From the mention of the station on the Ganges one at once inferred the intention of the speaker to convey the ideas of purity, &c.

c25.s3.5

² E. Abegg, *Festschrift Windisch*, pp. 188 ff.; ZDMG. lxxvii. 207ff.

c25.s3.6

³ Dīrghavyāpāravādin school, dubiously ascribed to Lollaṭa (Dé, *Sanskrit Poetics*, ii. 192, n. 16).

c25.s3.7

⁴ Abhihitānvayavādin school of Mīmāṁsā.

c25.s3.8

¹ Anvitābhidhānavādin school.

c25.s3.9

But the holders of the doctrine of Dhvani remained unconvinced, and on the basis of their theory they declared that the soul of poetry was not style nor sentiment, but tone, Dhvani, by which they meant that an implied sense was the essence of poetry. What was suggested might be threefold, either a subject, or a figure, or a sentiment and, while these three possibilities are admitted by the more orthodox members of the school, including Ānandavardhana and Mammaṭa, Abhinavagupta went much further and declared that in reality all suggestion must be of sentiment, holding that in the long run suggestion of subject and figure reduced themselves to this. Viçvanātha, author of the *Sāhityadarpaṇa*, followed his lead, but this never became the accepted doctrine, for the writers realized that, by attempting thus to limit suggestion, they would cut out a good deal of admitted poetry. Suggestion, however, can be expressed in two ways, for it may rest on the metaphorical sense of words, in which case we have the species of Dhvanikāvya where the literal meaning is not intended at all (*avivakṣita-vācya*), thus making provision for the ordinary view which attached great importance to metaphor or simile as the base of poetry. Or, again, the literal sense may be intended, but a deeper suggestion implied, in which case we have the type where the literal sense is meant but ultimately comes to something deeper (*vivakṣitānyapara-vācya*). Here, again, we have two different cases, for the process of apprehension may be instantaneous (*asaṁlakṣya-krama*), which is the rule in respect of suggestion of sentiment, or due to a perceptible process (*saṁlakṣya-krama*), as in the suggestion of subject and figure. The process of apprehension of sentiment is comparable to the piercing of a hundred lotus leaves with one needle; there is a process by which the factors induce the sentiment, but it is so rapid as to seem instantaneous. It is clear also that the rising up of sentiment is not the result of inference; it can come into being only in a person who has had in previous lives experience which gives him aesthetic susceptibilities, makes him a feeling heart or connoisseur (*sahṛdaya*), and in him it arises as a perfectly unique emotional experience, comparable only to the bliss of cognition of the absolute,¹ a transcendental (*alaukika*) joy. He who sees on the stage, or reads in poetry, the factors which are connected with sentiment presented, does not regard them as external to himself, whether as the property of the actor or of the hero of the play or poem, nor does he appropriate them to himself; he sees them under the aspect of universality, and this causes the sentiment to be unique and pleasurable, whatever the corresponding emotion, as a personal possession, would be. What in real life would be horror, thus as a sentiment is exquisite joy. We have, it is clear, a real effect to explain the nature of disinterested aesthetic pleasure arising from literature.

c25.s3.10

¹ This is, we must remember, identic with the bliss which is part of the absolute as one, being, thought, and joy.

c25.s3.11

But the system does not deny the right to rank as poetry of poetry which contains only in a secondary degree suggestion (*guṇībhūta-vyaṅgya*). This head helped them to find a place for the doctrines of the older writers who accepted in certain figures the expression of sentiment, as in the Preyas, Rasavat, and Ūrjasvin of Daṇḍin. Moreover, it served to include cases in which these writers found that one figure lay at the base of others, as when Vāmana found simile in all, and Bhāmaha held that in all figures there lay hyperbole, a view mentioned by Daṇḍin also. Finally the system, though not its sterner advocates, confessed that they must permit the kind of poetry called Citra, picture, in which there was mere beauty without any suggested sense at all. The beauty may be of sense or sound.

c25.s3.12

It remained to seek some way of dealing with the qualities and figures and the styles or manners of the earlier writers, so as to find them a just place. One great simplification was effected by reducing the number of qualities, restricting their extension to sound effects, and by merging in them both the Rītis of Vāmana and the Vṛttis of Udbhaṭa, which were at the same time practically identified. This became possible through the adoption of a new doctrine as to the relation of qualities to the poem; the sentiment being regarded as the vital element, the qualities are related to it as the soul of the poem, in the same way as heroism is an attribute of the soul of man. This fact, however, precludes us from regarding qualities as stereotyped in the old fashion; everything depends on the sentiment, and what relatively to sentiment would be a quality might in the abstract be a defect. If, then, we admit qualities, they must be such as are never defects, and they must be positive in nature, not mere lack of defects, and distinct in character. On this score we can dismiss Vāmana's Çleṣa, Samādhi, and Udāratā as merely forms of Ojas, strength; Saukumārya and Kānti are no more than the absence of the faults of harshness and vulgarity; and Samatā, evenness, is in some cases positively a blemish. We have thus left just three qualities, and these of sound only, there being no need in the views of the school, which Mammaṭa in special develops very clearly, to allow of qualities of sense. These are strength, which is regarded as causing, or as Viçvanātha insists, coincident with an expansion (*vistara*) of the mind, and which has its proper place in the sentiments of heroism, horror, and disgust; sweetness, which stands in a like relation to a melting (*druti*) of the mind, and is normally present in the sentiment of love-in-union, but appears also, rising in degree in order, in pathos, love-in-separation, and calm; and clearness, including the older Arthavyakti, which corresponds to an extension or pervasion (*vikāsa*) of the mind. The idea of these psychological equations was probably borrowed from Bhaṭṭa Nāyaka who in his theory of the enjoyment (*bhoga*) of sentiment spoke of these three conditions of the mind. In concrete terms the characteristics of the three qualities of sound are given by Mammaṭa as depending on arrangement of letters, compounds, and style of composition; thus sweetness depends on the use of all the mutes (save linguals) with the corresponding nasals; *r* and *ṇ* with short vowels; and no compounds or short compounds; strength arises from the use of double consonants, or consonants followed by the corresponding aspirate; conjuncts of which *r* forms part; lingual letters save *ṇ*; the palatal and lingual sibilants; long compounds; and a formidable, loaded, composition; no special rules are given for clearness. It is obvious that Mammaṭa is here incorporating much of what Udbhaṭa taught regarding his Vṛttis, the characteristics of the Upanāgarikā and Paruṣā forms being closely similar, and thus it is possible for Mammaṭa to bring the Vṛttis under qualities. Nor, as he includes the use of compounds in his treatment, does he find it difficult to include the styles of Vāmana, as brought into close relation to compounds by Rudraṭa. It is, of course, all rather artificial, and very much of an effort to harmonize without real care for the facts, but it is normal and plausible enough.

c25.s3.13

In the case of the figures a definite line is drawn between them and the qualities. The figures are only of importance in so far as they seek to enhance the sentiment; they do not, however, act directly on the sentiment, but they aid it by decorating the body, sound and sense, just as the soul of a man has as attributes the qualities, while ornaments such as a necklace affect his body directly. If figures do not aid the sentiment, then they are merely forms of speech, and their place is in poetry of the third type, Citra, pictorial poetry which Viçvanātha denies outright the name of poetry.

c25.s3.14

Ānandavardhana give much else of great interest, and his remarks on compounds are sensible and just; he allows them freely in Ākhyāyikās, but he points out that even there where pathetic, or love-sorrow effects are aimed at such compounds are not suitable, and in the Kathā they should be employed in moderation. The doctrine of Doṣas, defects, is treated from the same point of view as that of qualities; tautology, for instance, may become an excellence if the suggested sense is made more effectively felt by means of it. But as with qualities, there may be real faults which are always such; the Dhvanikāra insists that in love there is always a defect in using unmelodious (*çrutiduṣṭa*) expressions, though such are in good taste in the heroic or the horrible sentiments.

4. The Critics and Supporters of the Doctrine of Dhvani

c25.s4.1

The idea of suggestion did not pass unchallenged. Bhaṭṭa Nāyaka in the *Hṛdayadarpaṇa*,¹ perhaps an independent work though there is some evidence of it having been a commentary on the *Nāṭyaçāstra*, who wrote before Abhinavagupta, insisted on his own theory of the effect of words. In addition to denotation, he ascribed to them the faculty of generalization, Bhāvakatva, which consists in making the meaning intelligible as universal to the audience, while a third power, Bhojakatva, results in the audience relishing the enjoyment of the poem. This condition is one of an enjoyment which cannot be described, but which is marked, as we have seen, by the melting, expansion, and extension of the mind. The loss of his work makes it very difficult to appreciate what Nāyaka exactly intended to convey.

c25.s4.2

¹ Cf. M. Hiriyanna, POPC. 1919, ii. 246 ff., who regards him as expressing the Sāṁkhya view of aesthetic joy as arising beyond nature to something finer if not real, while the Vedānta view rests on the revelation of the absolutely real which is joy.

c25.s4.3

More fortunate is Kuntala, probably a contemporary of Abhinavagupta, whose *Vakroktijīvita*¹ is an effort to present in a new and improved form the idea vaguely present to Bhāmaha and those who laid stress on figures as the essential feature of poetry. He insists that Vakrokti, crooked or figurative speech, is the life of poetry, distinguishing it from science and any merely ordinary or natural mode of expressing facts of any sort. It is, therefore, a deviation from the ordinary language of life in order to produce a certain striking effect (*vicchitti*), or an imaginative turn of speech (*bhaṅgī-bhaṇiti*). Poetry, therefore, is to be defined as embellished sound and sense, the embellishment being figurative speech, and as this is the only Alaṁkāra possible, and as it is essential to poetry, it is absurd to have any definition which omits figures or makes them subordinate. He goes in great detail through all the forms of poetry in order to show that the principle of Vakrokti covers adequately all developments, citing copious examples from the poets, especially Kālidāsa. It is to the imagination or skill of the poet, his work (*kavikarman*), that we owe the presence of Vakrokti in any poem, and this work can be classed according as he exhibits it in regard to the letters, to the base or termination of words, to a sentence, a particular topic, or a treatise as a whole. It is clear that we have here in part a reminiscence of the doctrine of an element of hyperbole in all poetry asserted by Bhāmaha; a poem attains at best a transcendental charm (*lokottara vaicitrya*), which can be judged in the long run only by the man of taste, a result in which Kuntala agrees largely with the theory which he attacks.

c25.s4.4

¹ Ed. S. K. Dé, Calcutta, 1923.

c25.s4.5

The strength of this position is clearly the room it finds to allow of accepting figures on their own merits, and not as ancillary to a sentiment as essential features of poetry; we have their cause in the poet's imagination (*kavipratibhā*), and their effect is a definite fact, a species of charm. Mammaṭa gladly accepts this fact and, when figures do not affect sentiment, still declares that they have charm (*vaicitrya*), and Ruyyaka built up his treatment of figures on this basis. To complete his theory Kuntala naturally endeavours to bring both sentiment and suggestion under the scope of his principle, with just as much success as the opposite effort achieved.

First 68 of 108 passages. Read on

Linked texts (36)