J. L.

MASSON

PHILOSOPHY AND LITERARY CRITICISM

IN ANCIENT INDIA 1

... a man might maliciously object that Plato, being a philosopher was a natural enemy
of poets. For indeed after the philosophers had picked out of the sweet mysteries of
poetry the right discerning true points of knowledge, they forthwith, putting it in method
and making a school art of that which the poets did only teach by a divine delightfulness,
beginning to spurn at their guides, like ungrateful prentices were not content to set up
shop for themselves, but sought by all means to discredit their masters; which by the
force of delight being barred them, the less they could overthrow them the more they
hated them. (Sydney, The Defense of Poesie)
I f by a philosopher we mean a man capable of abstract thought, of the
logical ordering of ideas, I think there can be little argument that the good
philosopher will, generally, not be a good poet and vice-versa. " I t is not
the function of the poem to propound a solution.., a poet differs from
other kinds of sage by the fact that he makes his poems out of his igno-
rance as much as his certainties" says John Wain in a brilliant essay on
Yeats' " A m o n g School Children".z He quotes Wallace Stevens:
You must become an ignorant man again
And see the sun again with an ignorant eye.
Wain insists that the essential training of the poet is to "unlearn the disci-
pline of abstract thought." He ends by showing what it is that this great
poem achieves: "Instead of circling back on itself, the poem moves for-
ward, in the form of a bridge, then suddenly stops with no opposite shore
in sight. It is not a bridge after all, but a pier. It leads nowhere; its purpose
is to afford us, before we turn and retrace out steps, a bleak and chastening
glimpse into the deep waters."
We do not read poetry in order to extract ideas, and most of the ideas
of even the finest poetry sound appallingly mediocre when baldly stated in
expository prose.
But the question is whether the same is true of literary criticism as well.
Here the answer is not so easily forthcoming. Have the great critics of this
century been philosophers of repute? Generally not, but often the two
disciplines seem to touch each other, even if only for a moment. T. S.
Eliot wrote a P h . D . thesis in the department of philosophy at Harvard

Journal o f Indian Philosophy 1 (1971) 167-180. All Rights Reserved

Copyright 1971 by D. Reidel Publishing Company, Dordrecht-Holland
168 J.L. MASSON

(on Berkeley), and it is clear from many of his best poems (the Four
Quartets for example) that abstract notions of metaphysics had a bene-
ficial effect on his poetry. In his actual criticisms, strangely enough, there
seems less influence. I. A. Richards wrote The Meaning of Meaning which
can, after all, be found in the philosophical stacks of libraries. Empson is
obviously no stranger to philosophical ideas, and has a profound interest
in pyschoanalysis. But the younger critics of today seem to have been less
influenced by philosophical systems, especially those critics who are, as
becomes increasingly more common, themselves poets. 'Systems' of
literary criticism are of course common, and one could hold that under-
lying them are philosophical ideas. Ivor Winters, Elder Olson, Edmund
Wilson, all these critics have systems, informing most of their particular
judgements. On the other hand we cannot deny that some of the best
literary criticism is coming out of the small journals. I think of the reviews
of Poetry Chicago which seem to me to contain much of the best literary
criticism being written. The writers, on the whole, do not write from spe-
cific points of view. Rather they seem to be looking at each poem in and
of itself.
What can we say about ancient India? When we speak of a philosophy
of aesthetics, of underlying philosophical notions in literary criticism,
what precisely do we mean? What did it mean to be a philosopher in an-
cient India ? Is it merely a question of training in philosophical doctrines ?
This all literary critics shared. The very nature of education in ancient
India was such that we cannot imagine a literary critic who was not well-
versed in nydya, mtmd.msd, veddnta and grammar. Poetics uses all of these
disciplines. But does this mean that it is founded on any particular philo-
sophy? Is there, in fact, any system in Sanskrit poetics that could be de-
scribed as philosophical that is more than just the use made of other sys-
tems? Bharata, Bh~maha, Da.n.din, V~mana, Udbha.ta and Rudrat.a, in
short, all the writers before the time of ,~nandavardhana, 3 do not rely on
any philosophical structure. It is only with Ananda and Abhinavagupta
(at least among extant authors) that we find philosophical issues playing a
major role in literary criticism. In our book gdntarasa and Abhinavagupta's
Philosophy of Aesthetics we have tried to lay bare the philosophical scaf-
folding on which Abhinava built his magnificent system. Now I would like
to examine a specific issue in .~nandavardhana's Dhvanyaloka, to show
how the holding of philosophical opinions allowed both ,~nanda and
 PHILOSOPHY AND LITERARY CRITICISM 169

Abhinava to understand certain aspects of aesthetic experience that were

lost to older writers. Ananda, for the first time in literary criticism, began
asking philosophical questions: what is really important (pradhana) in a
poem? What takes place in great poetry? I-Iow are we effected in drama?
What is 'propriety' (aucitya)? When is a work of art vtdgar? What is the
relation between naming (svagabdavacya) and describing in poetry? How
should we define a figure of speech? Why do we 'enjoy' tragedy? I wish to
examine one of these questions, the true nature of a figure of speech, in
some detail to see precisely how it is that Ananda's underlying philosophy
enabled him to solve literary problems. What I have said so far is rather
abstract, and while I believe it is true, if it is to be convincing after a first
reading, it will have to be supported by a closer view of how -~nanda (and
Abhinavagupta) dealt with the problem of what is called rasavadalahkdra.
[Roughly, it is the name of an alahkara 'figure of speech' which possesses
rasa, an imaginative experience. Its nature will be discussed shortly.]
I will therefore do two things now: I will show what opinions were ear-
lier held on rasavat, and then I will give an annotated translation of the
section from the Dhvanydloka dealing with this problem, thereby showing
that it was the underlying philosophical assumptions and seriousness of
Anandavardhana and Abhinavagupta that enabled both writers to correct-
ly evaluate a major element in aesthetic experience that had been incor-
rectly understood by all earlier writers, precisely because of the fact that
they lacked a philosophical foundation for their theories of literature.

PART I -- E A R L I E R WRITERS

Bhgtmaha (III. 6) is the first recorded author to speak of rasavadalahkara.

The definition he provides is as follows:
rasavat dargitaspa~t, a~.rhgarddirasam.
"Rasavat is when such rasas as ~.rhgdra etc. are clearly shown." The exam-
ple he gives is unfortunately obscure, 4 but it is clear that for Bh~maha,
rasavat is simply the inclusion of any rasa. Whenever an imaginative ex-
perience is undergone by the reader, this is art example of rasavat, This is
essentially the position of all writers, until the time of ,~nandavardhana.
Dan..din (II. 275-292) illustrates all eight rasas, under rasavat. Here, for
instance, is his example of~rhgdra:
170 J.L. MASSON

m.rteti pretya sahgantu .myayd me maran, am. matam

sai.sd avantf mayd labdhd katham atraivajanmani
" H o w wonderful! The very Avanti (Vfisavadatt~) whom I thought to be
dead and wished to die myself in order to join, why, she has been found
again by me in this very life !"
Da0.din remarks of this that: seya.m rati.h ~r~gdratd.m gatd. "The love in
this verse has reached the state of 6rhgdra (i.e. has become a rasa, an ima-
ginative experience)."
Udbhat.a however is more interesting, if less logical. He uses, more or
less, the same definition of rasavat as the one given by Bhfimaha above: s
rasavad dar~itaspa.s.ta~.rhgdrddirasodayam
"Rasavadis that in which there is displayed clearly the arisal of rasas such
as g.rhgdra etc."
It is his next phrase that is most interesting:
sva~abdasthdyisagcdrivibhdvdbhinaydspadam
This rasa can be conveyed by one or more of the five elements named,
namely by directly naming the emotion, 6 by mentioning the sthayibhdva,
the vibhavas and the sa~cdribhdvas, or even by acting. 7 He distinguishes
rasa from bhdva, which he has dealt with, rather illogically, under preyas-
vat (IV. 2). 8 He provides an example from his own poem, the Kumarasamb-
hava:
iti bhavayatas tasya samastdn pdrvatFgun, dn
sam. bh.rtdnalpasam, kalpa.h kandarpa.h prabalo 'bhavat
svidyat~pi sa gdtre.na babhdra pulakotkaram
kadambakalikdko~akesaraprakaropamam
k.sa.nam autsukyagarbhi.nyd cinMni~calayd k.san,am
k.sa.na.m pramod:tlasya d.rgasyasyarn abhfi.syata
"As he was thus thinking about all the qualities of P~trvati, many desires
were produced in his mind, and love grew very powerful. His sweating
body with its mass of hairs raised in excitation, bore a resemblance to the
cluster of filament-fibers that proceed from the pericarp of a Kadamba
flower. His face was adorned for one moment with a lazy look of joy, and
the other moment with a look immobile from thought, and for a moment
with a look full of longing."
 PHILOSOPHY AND LITERARY CRITICISM 171

The first thing any literary critic will notice about this verse, is how
pedestrian it is. 9 It is quite likely that Udbhat.a in fact wrote the Kumdra-
sambhara in order to illustrate the figures of speech he gives in such de-
tail. 1 Be that as it may, what interests us here is what this verse has achiev-
ed, in the eyes of its author. For Udbha~a, this verse is an example where
an imaginative experience relating to love takes place in the reader. Now
any reader is likely to ask himself the following questions: precisely what
is this imaginative experience? What does it consist in? How is it brought
about? How do I know that this verse has been successful in its creation?
These are questions that Udbhat.a does not even attempt to answer. In
their absence, it is hard to say precisely what he means by rasa. If he does
mean "imaginative experience", then he would surely admit that it lies
at the root of literature, for literature deals primarily with heightened
responses in its readers. Now if he would agree with this, one is surprised
to find that he has not dealt with the problem at any length. In fact, it
would seem that Udbhat.a does not agree with this. He has given this
verse as an example of rasavadalahkdra, a figure of speech known as
rasavat, i.e., containing rasa.
,~_nandavardhana for the first time asked the crucial question: "if this
is a figure of speech, if it is something that ornaments something else,
precisely what is it that is ornamented?" In other words, what is the real
point behind a figure of speech? For what purpose do they exist? It would
seem that Pratih~renduraja, Udbha.ta's commentator, was aware of.~uan-
da's criticism (he lived not long after ,~nanda), for he says (p. 53), ras&h
khalu tasya (i.e., rasavatkdvyasya) alahk~ra.h. "Rasas embellish poetry that
contains rasas." But this is rather circular. Probably aware of this, Indu-
raja says on p. 54: rasdnd.m bhdvdnam, ca kdvyagobhatigayahetutvat ki.m
kdvydlahkdratvam uta kdvyaflvitatvam iti na ttivad vicdryate granthagaura-
vabhaydt, rasabhdvasvarftpam, c~tra na vivecitam aprak.rtatvdd bahuvakta-
vyatvdc ca. "Because of the fact that rasas and bhdvas are a great source of
beauty for poetry, it can be asked whether they are embellishers of poetry,
or indeed the very essence of poetry. For fear of our work becoming too
long, we have not dealt with the issue here. N o r have we investigated the
true nature of rasa and bhdva, as that is not our main concern, and also
because there would be a great deal to say on this subject." 11
,~nanda, in asking the question, "just what is it that constitutes a figure
of speech," reached the very heart of literary criticism. For he is really
172 J.L. MASSON

asking: what constitutes a good poem, and what happens when a sensitive
reader reads a good poem. By answering this question at length in his
great work, the Dhvanydloka, Ananda has put into proper perspective what
until then had remained largely misunderstood. Thus in the present case,
Ananda was able to see that the definition o f an alahkdra precludes what
Udbha.ta and earlier writers understood by the term rasavat. He shows
that rasavat is in fact nothing but rasa, and rasa is the very soul of poetry.
It can never be described as a figure of speech. But he does not completely
discard the figure. Fie simply redefines it: whenever a rasa is subordinated
in a p o e m to some other element, we can say that a figure of speech known
as rasavat is found. But it will only be an alahkdra, and the real aesthetic
significance will lie elsewhere.
N o w in order to see precisely how ,gmanda arrives at this conclusion,
let me translate in full the section of the Dhvanydloka dealing with rasavat.
It is a fine illustration of 2gmanda's method, in which philosophical assump-
tions as to the nature of aesthetic response are of primary importance.

PART II - THE DHVANY.g, LOKA

The problem of rasavadalwikdra is dealt with in Kdrikd V of the D..~I.,

and the V.rtti thereon. Here is my translation: 12

KARIKX V
"But where rasa etc. are subordinate and especial literary importance is
given to a different meaning, then it is my opinion that in that poem, rasa
etc. is only a figure of speech and nothing more."

VR. T T I
"Although others have defined the domain of the figure of speech known as
rasavat, 13 nonetheless my own position is that when some other sense
being principally conveyed by a sentence as a whole (vdky~rtMbhfita),
rasa etc. being subordinated to it, c o m e under the domain of figures of
speech. For example, in flattering hymns of praise (cdt.u), where flattery
(preyola~kdra) is the major literary experience (vdkydrthatva), rasa etc.
are seen to be subordinate. 14 This use of rasa etc. as a figure of speech can
be either pure or mixed. An example of the first is:
'Why do you smile? You can't get away from me now that at last I've seen you again
after such a long time. Ruthless man, what is this sudden love for travel? What is separ-
 PHILOSOPHY AND LITERARY CRITICISM 173

ating us (i.e., what did I do wrong)? This is the way your enemies' women speak in
their dreams. Then they throw their arms about the necks of their beloveds only to
awaken to the empty circle of their arms. How loud they weep!'15

Clearly this is an example of the figure of speech known as rasavat, for

pure karu.narasa (unmixed with any other element such as a figure of speech)
is subordinated (to the predominating eulogy). A n d thus in similar cases
it is quite clear that other sentiments could be subordinated. An example
of mixed rasa etc. when it is subordinate is:

The women of the Triple City wept from lotus eyes as ~ambhu's arrow-
flame embraced them; but still, though shaken off, the fire caught their
hands, though struck, did pluck their garments' hem, denied, it seized their
hair, and scorned like a lover who has lately loved another, lay before their
feet. May this same fire burn away your sins.1~

Here the aesthetic experience of jealous love, achieved by means of pun-

ning is subordinate to the main purpose which is to describe the extra-
ordinary power of the enemy of Tripura (i.e. Siva)fl 7 Only such is the
proper province of figures of speech like rasavat etc. That is why there is
nothing wrong with including two (usually contradictory) rasas, namely
karun.a and jealous love, in one poem, precisely because they are both
subordinate. For when rasa predominates, how can it be a (mere) figure
of speech? Since everybody knows that a figure of speech is meant to add
beauty to a poem. Surely a thing's own self cannot act as its own beautifier.
The matter can be succinctly put as follows: " W h e n figures o f speech are
used to embellish rasa, bhdva etc. as the main and exclusive concern, then
they truly act as embellishers." Therefore, wherever rasa etc. predominate,
they are not to be included under the domain of the figure of speech
rasavat or any other, but must be considered to fall under the domain of
suggestion itself. Simile etc. act as its embeUishers. On the other hand,
whenever another meaning predominates as the main idea in the sentence,
and rasa etc. contribute to its beauty, rasa etc. should be considered to
fall under the domain of such figures of speech as rasavat. And so there
is a eonsirable distinction to be drawn between dhvani and such figures of
speech as simile or rasavat."
There follows a discussion on the value of describing natural objects
(non-living beings, acetanavastu) in poetry 18 which is not directly relevant
to our discussion, and a somewhat confusing argument on an avak~da,
and then ~,nanda ends by saying:
174 J . L . MASSON

Therefore rasa etc. are figures of speech (only) when they do not predominate. But
when a rasa or a bhava predominates, then by all means it (takes on major literary im-
portance and) should be considered as the very soul of dhvani, worthy of being em-
bellished (alaakarya).19

Abhinavagupta, on p. 204 of the Locana, brings out the final point very
nicely: alahk&yavyatiriktag edlahkdro' bhyupagantavya.h, loke tathd sidd-
hatvdt, yathd gu.nivyatirikto gu.na.h, gu.ndlahkdravyavahdrag ca gu~.iny
alahkdrye ca sati yukta.h, sa edsmatpak.sa evopapanna... "The ornament
must be accepted as something distinct from the thing which it ornaments,
because this is a well-known fact in everyday life, just as a quality must be
accepted as different from that which is qualified (gun.in). And the use of
the words guna and ala/tkdra is with reference to a gun.in and an alahkdrya.
This becomes possible only if one follows our view."
And indeed, the author of the K6rikds, and Anandavardhana go on to
apply the same implicit philosophy to the ancient doctrine of the gugas,
and again completely revolutionise later thinking on this issue, through
the clarity and compelling logic of their philosophical position. What
Ananda has done is really quite simple: he has asked the basic question -
what is essential in a p o e m ? Is poetry to be defined in terms of style 20 and
in terms of figures of speech? Or is there something that lies behind style,
difficult as it m a y be to define and describe empirically, that is really
the essence of poetry? Ananda accepts the latter, and calls it rasadhvani,
that is, the literary creation of an emotional situation that is only suggest-
ed, not directly expressed. It is what a p o e m suggests that moves us, not
what it crudely and directly states. I f this capturing in words of an emo-
tion is the essence of poetry, it follows that it cannot be subordinated 21 to
any other element in the p o e m without losing its power and importance.
It cannot, therefore, be a mere ornament, and at the same time be the
source of an overpowering aesthetic experience. This is ,~nanda's essen-
tial point, made for the first time in Sanskrit literary criticism. It is a point
that owes as much to the clarity of Ananda's philosophy of poetry as it
does to the sensitivity of his literary taste and is a good illustration of the
literary/philosophical method of India's outstanding literary mind. 22

University of Toronto
NOTES

1 I have used the following editions: Bh~,maha's Kavyalankara (ed. and transl, by
 PHILOSOPHY AND LITERARY CRITICISM 175

P. V. Naganatha Sastry), Tanjore 1927; Dan.qlin's Kdvyddar~a (ed. with his own com-
mentary by Pandit Rangacharya Raddi Shastri), B.O.R.I., Poona, 1938; Udbha~/a's
Kdvydla~kdrasdrasa~graha, with the Laghuvrtti of Indur~ja (ed. by N. D. Banhatti),
B.O.R.I., Poona, 1925; ,gmandavardhana's Dhvanydloka with Abhinavagupta's
Dhvanydlokalocana, the Bdlapriyd of ~ri R~tma~xaka (ed. by Pandit Patt~bhir~,ma
~gstri), Kashi Sanskrit Series, 135, Banaras, 1940; Ndtyas~stra with the Abhinavabhd-
rati of Abhinavagupta (ed. by M. R. Kavi and K. S. R. Sastri), 2nd ed. Oriental Insti-
tute, 1956 (vol. I).
Interpretations - Essays on Twelve English Poems (ed. by John Wain), Routledge,
London, 1955.
a It is not impossible, though I think it is unlikely, that Udbhat.a dealt with wider
issues of literary criticism in his lost commentary o n the Ndt.ya~dstra. Works, now lost,
written after ,~'tandavardhana, most certainly did: the Kdvyakautuka of Abhinava's
teacher, Bhal/~tauta, and the Hrdayadarpa.na of his rival Bhat.t.anayaka, both referred
to frequently in the Abhinavabh~ratf, dealt with philosophical issues of importance.
See extracts translated from these authors in ~dntarasa and Abhina vagupta's Philosophy
of Aesthetics by J. Masson and M. V. Patwardhan, Bhandarkar Oriental Research
Institute, Poona, 1969. See also our: Aesthetic Rapture, 2 vols, Deccan College,
Poona, 1970.
4 The example (III. 6) is:
devf samdgamad dharmamaskari.nyatirohitd
I discussed this verse with several Pandits at the Deccan College. They all found it
obscure, but the concensus seemed to be that samdgamadis used in the sense o f " t o make
love to". A Maskarin is one who carries a maskara, a bamboo staff, i.e., an ascetic.
Possibly, though I don't think such a usage is attested, the locative is used in the sense
of the instrumental (perhaps by analogy with other verbs of loving, e.g., mayi snihyati).
So the idea would be that the queen made love with the 'virtuous' (the adjective dharma
will have to be ironic) ascetic openly (atirohitd), i.e., nothing could prevent her. Now
according to all definitions of ~.rftgdra, this could not be real love. It must be ~.rt~gdrdb-
hdsa. Since Bh~maha gives it as an example of ~.rftgdra, we must either assume that he
had no notion of rasdbhdsa (the term is not found in the Kdvydlaftkdra, though we meet
with it in Udbha~a) or else that I have totally misunderstood the verse. Without the
context, it is difficult to be certain what is meant. The translation by Sastry makes no
sense to me whatever: " T h e Devi the receptacle of D h a r m a arrived unobstructed."
s Udbhata's Kdvydlalikdrasdrasaftgraha, IV. 3, p. 52.
a The issue of sva~abdavdcya is an extremely interesting one, but it would require a
monograph in itself. Briefly the problem is this: it would seem that before ,~manda-
vardhana (though see the interesting passage in the Abhinavabhdratf, Vol. I, p. 343),
writers held that emotions could be conjured up by simply naming them, as in the verse
by Udbhat.a. ,gmanda realised that there was all the difference in the world between
declaring something to be the case ( " I feel great longing") and actually enabling the
reader to feel the emotion without baldly stating it. (See D.,~.l.p. 78-83. It has been
mistakenly held that in the second Uddyota of the D./V~I. (pp. 248-249) .~'~,nanda claims
that in asa~lak.syakramavyahgya (i.e., rasadhvani) the vyabhicdribhdvas are directly
expressed. This is incorrect. See "Solution to a Long-Confused Issue in the Dhvanya-
loka" by M. V. Patwardhan and J. L. Masson, to appear shortly, which deals with this
problem. Further on the notion ofsva~abdavdcya, see our book on $dntarasa.
Surely U d b h a t a would not have insisted that svadabdavdcyatva invariably be present.
176 J . L . MASSON

On the other hand, Indttr~ja notes that all five elements are actually found in this verse.
Kandarpa, which stands for 'love', i.e. rati or g.rftgara, is sva~abdavacya. Kandarpa is also
the sthayibhava. The anubhavas are the horipilation and the sweating limbs. He, curi-
ously, also says that sveda and roma~ca are also the sagcaribhavas for the following rea-
son: s6ttvikana~ sthayibhavavasthavi~e.satvena nirvedadivat sa~caritvat. But this is not
very convincing, and generally purely physical sattvikabhavas are not regarded as
vyabhicaribhavas O.e. sahcaribhavas) nor are they necessarily different states of the
sth~yibhava, for, as Abhinava often points out in the Abhinavabharatf, they can be
common to many sth~yibhavas. For example, sveda could come from bhayanaka as
well as from g.rtigara. The vibhavas are of course Phrvati and ~iva. The abhinaya is the
apatigabhinaya, i.e., movements of the eyes. Note that Indur~ja points out that the
sth~yibhava and the sattvikabhavas have been mentioned by name (sva~abdavacya).
Naturally the vibhavas have as well. So in fact, all the dements of rasa in this verse
are svagabdavacya. But the temptation to take svagabda as separate from, and con-
struing with each of the elements mentioned (apart of course from abhinaya) must be
resisted, since Indur~ja quotes from a lost text of Udbhat.a (his commentary on Bhama-
ha? his commentary on the Ng?)where he says:pagcar@a rasab.
8 The bh~tva mentioned is rati. The example Udbhat.a gives is where Phrvati feels
motherly affection for a small deer, pressing it against her bosom, speaking to it, etc.
Now what is curious here is that this is regarded, by later writers, as an example of
devaputradivi.sayakaratibhava, where bhava means a vyabhicaribhava. Indufftja claims
that rati here stands for bh~vas in general, but it seems an odd choice. Generally rati is
the sthayibhava of grhgara. The example of course is not ~.r~gara. Note too that the
definition ofpreyasvat given by Udbha~a will apply equally to rasavat! The Viv.rti (ed.
by K. S. Ramaswami Sastri Siromani, Oriental Institute, Baroda, 1931) is aware of the
difficulties: ratir hi devagurun.rpadivi.saya g.rhayte, k6ntavi.sayay6s tu rateb saeane rasava-
dalatik6ro vak.syate. But Udbhat.a has said in the definition: anubhavadisaeana.h which
seems to suggest that he does not mean rati to be taken as a vyabhicaribhava. And in
fact the Viv.rti says that the bhava is ratyatmaka. Indur~ja calls it rather vatsalyasvab-
hava. The word bhava (in such expressions as bhavadhvani) refers to two different
things: vyabhie6ribhavadhvani (which is ephemeral) and devan.rpaputradivis.ayaratib-
hava, which is a permanent emotion. Sanskrit literary critics are not very clear on how
this latter differs from rasa proper. Since in the Abhinavabharati, p. 341 (Vol. 1), Abhi-
nava refers to a view whereby sneha could be considered as a rasa, a view he does not
accept (ardratasthayika.h sneho rasa iti tv asat), it would seem that he, and therefore
probably most later writers who follow him (e.g., Mamma~a, Vi~van~ttha, Jagann~tha
etc.) would include such examples under g.r~gara, though this is not certain. The problem
of bhavadhvani requires investigation.
9 When I say that Udbhat.a is not a very good poet, I am thinking in particular of verse
8, under IV. 7, p. 56, which he gives as an example of samahita, i.e., the allaying of a
bhava. Not only is the verse not good poetry, it even makes bad sense. It is not clear
whether the anubhavas described belong to giva or P~rvati (though I suppose they must
go with ~iva). Kanta~ d.r~am is moreover bad syntax. If it is P~rvati who is being de-
scribed (unlikely), then to what does romageasvedasahkulam refer? Surely not to her
face, for since when does a beautiful woman have hair on her face? If it is ~iva who is
being described, then mukharagam and prasannam make poor sense. Whoever is being
described, these two words go badly together.
~o This bad habit is not peculiar to Udbhat.a. Bh~maha cites only his own examples,
and with one or two problematic exceptions so does Dan.din. V~.mana alone, of the
 PHILOSOPHY AND LITERARY CRITICISM 177

early writers, cites verses from other poems. Before _~nanda, he is the only literary
critic to do so. No author, before .~nanda, has cited Pr~krit examples to illustrate
literary principles.
11 Note the implication of Indur~tja's remarks, namely that Udbhat.a too has not dealt
with this subject at any length elsewhere, for otherwise, surely Indurgja would have
made a reference to such a work. It is likely then that Udbhat.a, in his comments on the
sixth adhy~ya of the N~!yagastra, did not make any extended remarks on the nature of
rasa.
12 D,/~d. II, 5. p. 191. Here is the text:
pradhane' nyatra vakyarthe yatrahga~n tu rasadaya.h
k a w e tasminn alahkaro ras~dir iti me matib
vrtti
yady api rasavadalafik~rasyfinyair dar~ito visayas tath~pi yasmin k~vye pradh~nta-
yfinyo' rtho vfikygrthibh~tas tasya c~flgabhfitfi ye ras~,dayas te ras~der alafikfirasya
vi.sayfi iti mgmakina~, pak.sal)., tad yath~ cfi~u.su preyolafik~rasya vfiky~rthatve' pi ras~-
dayo' figabhfit~ ~ y a n t e . sa ca ras~dir alaflk~rah guddha.h safikir.no vfi. tatr~dyo
yathfi -
ki.m hfisyena na me prayfisyasi punah, prfiptag cirfid dar~anam
keyar!a ni.skaru~a pravasarucita kenfisi dfirik.rtal).
svapnfintesv iti te vadan priyatamavy~saktakaot.hagraho
buddhvfi roditi riktab~huvalayas t~ra .m ripustrijanah
ity atra karun.arasasya ~uddhasy~figabhfiv~t spa.stare eva rasavadalafikfiratvam, evam
eva .mvidhe vi.saye rasfintar~xT~ .m spa.st.a ev~,figabh~tvat)., saflkirfio ras~dir afigabhfito
yatha -
k.sipto hasfftvalagnah, prasabham abhihato' py ~dad~no ' .m~ukhntam
gr.h.nan kege.sv ap~stag cara~a~ipatito nek.sita.h sambhrame .na
~liflganyo 'vadhfitas tripurayuvatibhih, s~runetrotpal~tbhil).
k~mivgrdfftpar~dhah, sa dahatu duritar0 ~fimbhavo va.h garfignih.
ity atra tripuraripuprabh~v~ttigayasya v~ky~rthatve ir.sy~tvipralambhasya ~le.sasahitas-
y~flgabh~va iti, eva .mvidha eva rasavad~dyalafikfirasya ny~yyo vi.sayat)., ata eva cer.sy~vi-
pralambhakarun.ayor aflgatvena vyavasthan~t sam~vego n a do.sah, yatra hi rasasya
v~kyarthibh~vas tatra katham alafik~ratvam? alafkaro hi c~rutvahetuh, prasiddhal)., n a
tv as~v gtmaivgtmanag c~rutvahetul~., tatha cfiyam atra safik.sepal~.-
rasabhgv~dit~tparyam g~ritya vinive~anam
alafikrtin~.m sarvgs~m alafikaratvas~dhanam
tasm~d yatro rasfidayo v~.ky~rthibhat~i.h sa sarvah., na ras~der alafik~rasya vi.saya.h, sa
dhvaneh, prabhedah, tasyopamgdayo' laflk~rah, yatra tu prgdh~yen~rthfmtarasya
vS.ky~,rthibh~ve ras~,dibhig cgrutvani.spattih kriyate, sa ras~.der alaflkarat~ya vi.sayah.
eva.m dhvaner upamgdinft .m rasavadalafikOxasya ca vibhaktavi.sayatfi bhavati.
13 There is a good discussion on rasavat in 'The Problem of Rasavadala~kara' by M.
Hiriyana, Proceedings of the Fifteenth All-India Oriental Conference, 1949 (Poona). His
point is that rasavadalatikara can never be vacya, and is therefore really a misnomer,
for whatever is vya~gya, i.e. the rasa, cannot be regarded as an alatikara at all, but only
as an ala~kffrin. He refers to the example given by U d b h a ~ under IV. 2., and notes
that it is a good example of ~.rt~g~ra, where the rasa is the ala~k~rya, the gu~in. It must
be pointed out however that ~ a a n d a himself makes this same point, on p. 204 of the
178 J.L. MASSON

V.rtti: yah ptmar aftgi raso bh6vo va sarvakaram ala~karyaO sa dhvaner atmetL Abhinava
too, on pp. 196 and 198 of the Loeana, makes the same point in greater detail.
14 This phrase is .gmanda's general illustration of rasavadala~kara, that is, a situation
in which rasa etc. act as an ala~kara. Now what is interesting about this phrase is the
remarks of Abhinava on pp. 191-192. The words itidam eka.m v~kyam will have to refer
to the interpretation of the V.rtti passage beginning tad yathd and ending a~gabhgtta
d.rdyante. It would seem then that this line is being interpreted. So that bhamahabhi-
praye.na (in the Loeana) will have to mean: "according to the view of (those who follow)
Bh~maha." Abhinava then goes on to say that: bhamahena hi gurudevan.rpatiputra-
vi.sayaprftivar.nana.m preyolatikara ity uktam. But this is incorrect. Bh~aaha does not in
fact defmepreyola~kara at all. At III. 5, he merely illustrates it:
preyo g.rh~gatam k.rs.nam avddit viduro yath~
adya y~ mama govinda j6ta tvayi g.rhKgate
k~lenais~ bhavet pritis tavaivagamanat puna.h
"Preyas is as in the following words of Vidura to K.r.sl~awho had come to his house:
O Kr.sn.a, that pleasure which I feel today at your coming to my house can only be
equalled by your return at some other time."
Clearly what is meant is a simple devavi.sayakaratibhava, namely Vidura's devotion to
IQ..sl)a. We can therefore interpret Abhinava's phrase in one of three ways: either he
had before him an expanded text of Bh~,maha, now lost, or he is referring exclusively to
Bhfimaha's followers, or he has inferred what he propounds from the remark of
Bh~naha. In the phrase: na tv alahkarasya v6ky6rthatva.m yuktam Abhinava recognises
the principal behind rasavat: an ala~kara cannot be predominant, for the very name
implies that it is the alat~kara of something else, namely of an ala~karya. Abhinava says
this most clearly on p. 197-198. Abhinava then gives the interpretation of this same
phrase in the Vrtti according to the followers of Udbha.ta - udbha.tamatanusdri.nas tu
bha~ktva vydcak~ate. His remarks however are obscure, and I cannot follow precisely
what he means. However what does seem beyond controversy is the fact that Abhinava
gives here two different interpretations of a line from the Dhvanyaloka, from two
different schools. The implication is that the Dhvany6loka was being studied (and
refuted?) by followers of both these schools already before the time of Abhinavagupta.
It is not of course certain that the two were rival schools, in fact we find that Udbhat.a
generally follows Bhfimaha, often to the point of using whole half-stanzas from the
latter.
15 This verse contains three different suggestions (dhvani) as far as I can see. There is
karu.narasa, which is gu.nibhata, which makes the verse a case of rasavat. It is gu.nibh~ta
to vastadhvani, which is the narapatiratibhdva (or, if one does not accept this as a vastu,
then it is subordinate to bh6vadhvani). This is the interim dhvani. The final dhvani of the
stanza is nrpativi.sayakaratibhdva. According to Abhinava, what makes this verse so beau-
tiful is the karu.narasa itself. As he says: na hi tvayd ripavo hatd iti ydd.rg analatik.rto" yaw
v~ky~rthas tdd.rg ayam, api tu sundarataribhftto' tra vdkydrthaO, saundarya .mca karu.nara-
sakrtam eveti. "The sentence-meaning, without any embellishment, namely, 'you have
killed your enemies' is not as charming as it becomes through the addition ofkaru.narasa."
Now I cannot agree with this on purely personal grounds. I do not find that the karu.na-
rasa adds to our appreciation of the glory of the king. I find that on the contrary, it
detracts from that appreciation. Abhinava never really tells us precisely bow the
karu.narasa embellishes the bhdvadhvani. Here is what he tells us about the stanza
(Loeana, p. 194): eandradind vastund yathd vastvantara~a vadanddy alahkriyate tadupa-
 P H I L O S O P H Y AND L I T E R A R Y C R I T I C I S M 179

mitatvena carutay~vabhasat tatha rasen@i vastu va rasantara.m vopaskrta.m sundara.m

bhatfti rasasy@i vastuna iv~la~karatve ko virodha~ ? nanu rasena ki.mkurvata prakrto
'rtho 'lahkriyate. nanu tayopamiyate prastuto' rtha.h, rasen@i tarhi sarasikriyate so
'rtha iti svasar~vedyam etat. "Just as one thing, such as a face etc. is beautified by
another thing, such as the moon etc., by acting as a comparison, so also a situation, or
another rasa, can be beautified by another rasa and appear lovely. And so even a rasa,
just like an object, can be an alankara. What harm is there in that? Should you ask
precisely how the subject under discussion is beautified by the rasa, we would respond
by asking precisely how a comparison beautifies something? Should you reply that a
given object is compared by it (with another), we reply that so also is a subject made
more pleasing (saraMkriyate) by a rasa. This can be experienced by anyone for himself."
Note that Abhinava has not in fact answered the question at all. He has only begged
the issue, though in an interesting fashion. I have given him the benefit of the doubt by
translating sarasfkriyate in a non-tantologous fashion, when literally it means: " a
thing is made full of rasa by rasa". But even if we do take it to mean 'more charming',
Abhinava has still not told us anything new. All he has done is to repeat his assertion
that rasa, in a subordinate position, can act as an ala~kara.
16 The translation of this verse is taken from Daniel H. H. Ingalls: An Anthology o f
Sanskrit Court Poetry, Harvard Oriental Series, Vol. 44, Harvard University Press,
Cambridge, 1965, Int. p. 21. See also his explanation of this verse, pp. 21-22. I have
discussed the verse and what it implies for comparative studies of literary criticism, in
my forthcoming translation of the Dhvanyalokalocana of Abhinavagupta.
~7 Here is Abhinava's commentary on this verse, Locana, p. 195: k~mipak.se 'n~d.rtah.,
itaratra dhuta.h, avadhQta iti na pratipsitah pratyhlifiganena, itaratra sarv~6gadh~tna-
nena vi~ar~rfik.rtab., s ~ r u t v a m ekatrer.syay~, anyatra ni.spratyh~atay~t, k~mivety aneno-
pam~nena ~le~nug.rhitener.sy~vipralambho ya ~k.r.st.as tasya ~le.sopam~sahitasy~6gat-
vain., na kevalasya, yady apy atra karu~ao raso v~stavo' py asti tath~pi sa taccarutva-
pratityai na vy~tpriyata ity anen~tbhipr~.yen.a ~lesasahitasyety efftvad evhvocat, na tu
karun, asahitasyety api. " I n the sense of a lover it means not paid any attention to; in
the case of fire, it means shaken off. Avadhata means not greeted by an embrace in
return. In the other case (i.e., in the case of the fire) it means broken to pieces (though
actually, how can a fire be broken to pieces?) by shaking of all the limbs. Sa~rutvam
means out of jealousy in one case, and in the other, out of a loss of hope. Kam~va -
this comparison which is helped by punning, brings in vipralambha based on jealousy.
This vipralambha based on jealously (which is suggested by comparison of the fire
with a lover) is a subordinate element accompanied by ~lesopam& It is not unaccom-
panied. (The idea is that vipralambha is combined with ~le.sopam~, i.e., there is sa~kara
between the two. Both are subordinated to givavi.sayakaratibhava). Although in this
verse karu.narasa really exists, still it does not enhance the beauty of the real meaning
of the stanza. With this in mind he says 'helped by glesa" and not 'helped by karu.narasa'."
I fail to see Abhinava's point. Surely this verse is parallel to the other verse in which
karu.narasa was a help to the major bhava. So here also it should be. The only difference
is that in this verse there is sa~kara as well Perhaps Abhinava's idea is that karu.na here
is only a third, intervening rasa, and the real a~gabh~tarasa is vipralambha, and the
pradhanarasa is givasya mahatmyam, i.e. bhavadhvani. But in any case, this contradicts
what Ananda says on p. 196, that both ~r.syavipralambha and karuna are ahgabhata in
this verse: ir.sy~vipralambhakarun.ayor a6gatvena.
~s ,~manda ends his discussion with a fine insight: yasman nasty evasav acetanavastu-
v.rttanto yatra cetanavastuv.rttantayojana nasty antato vibhavatvena (p. 203): "Because
180 J. r~. M A S S O ~ r

there is really no case of the description of something insentient in which there is not
some connection with the behaviour of a living creature, at least in so far as the descrip-
tion of the insentient things serves the purpose of creating the (appropriate) atmosphere
(for the behaviour of the living creature)." (The phrase antato vibhavatvena occurs
again on p. 495 of the D.,~I.) Abhinava's remarks are particularly good (Locana, p.
203): stambhapulakady acetanam api var.nyamanam anubhavatvac cetanam ~k~ipaty eva
tavat, kim atrocyate, atija4o' pi candrodyanaprabh.rtib svavigranto" pi var.nyamano
'va~ya.m cittav.rttivibh~vatd.m tyaktv~ k~vye" n~khyeya eva sy~t. "Even an insentient
phenomenon such as paralysis or horripilation, when it is described, because of the
fact that it is the outward sign of an emotion (anubhava), suggests its connection with a
living creature. What can we say further? Why even when completely lifeless objects
such as the moon, a garden etc., are being described (apparently) for their own sake
(svavi~ranta), unless they were meant as vibhavas of mental moods, they would have no
place in poetry! (And so they are to be understood as being connected with the citta-
v.rttis of living creatures.)"
19 D.,~I. p. 204: tasmad atigatvena ca rasadfnam ala~karata, ya.h punar a~gi raso bhavo
va sarvakaram alahkarya.h sa dhvaner atmeti.
2o Note what ,~.nanda says in the fourth Uddyota, p. 537: anapek.sitarthavige.salc.sarara-
canaiva bandhacchayeti neda.m nediya.h sahrdayanam. "Style which merely consists in
writing down syllables without paying any attention to a suggested meaning is not
something that is dear to the hearts of sensitive readers." The rest of the passage has
been badly misunderstood by Jacobi and others. See the article by M. V. Patwardhan
and J. L. Masson, ' A Difficult Passage in the Dhvan yaloka' to appear shortly.
al Further on pradhanat6, an important concept in the D.,~I., see D.A1. p. 191; 369;
397; 494 and 506. Cf. N~. XVI (anubandha) 17; XIX, 24, and the interesting verse
X X 76. See also Dan...din IL 132.
22 It follows that figures of speech are not the be-all and end-all of poetry, and accor-
dingly _~nanda, completely reversing the usual trend, devotes comparatively little space
to them.