This page has not been fully proofread.

88
 
taste whether the type should be assigned to the category upama or the
category atiśayokti, for an element of comparison as well as an element
of exaggeration is indeed present. Since the peculiarity of the figure is
best described in terms appropriate to upamā, it is listed there.
 
It will be seen that this method, as well as assuming a system of
figures, defines such a system. The fact that the method works is the
best proof that the figures are not haphazard lists of random verbal
phenomena. Despite occasional lapses, more pronounced in Mammaţa
and later writers, there is a consistency in the treatment not only of
individual figures, but in the kinds of figures deemed important. Each
figure was evidently seen to occupy a place, which was inimitable yet
finely attuned to adjacent figures, in the universe of poetic discourse.
 
We do not, of course, mean to imply that there is but one possible
system of alamkāras. Indeed, this is clearly not the case, and to maintain
otherwise would again deny autonomy to the different systems.
 
d) In the most obvious case, two authors may disagree as to the relev-
ance of a given figure in any system of figures. Bhämaha will not admit
hetu, sūkṣma, leśa, and a few others because he thinks they do not involve
an element of deviation, essential to any figure. 192 Dandin accepts these
figures, but objects to Bhāmaha's utprekṣāvayava, stating that it is only a
special kind of samsrsti.¹9
193 Mammaţa likewise rejects rašanopamā on
the grounds that iteration is an improper discrimination to apply to
the figurative idea.194 We have allowed all these figures, in the terms
required by the authors who accept them, because our purpose is not to
criticize the scope of the figurative idea, but rather to show the develop-
ment of which it is capable.
 
INTRODUCTION
 
e) Such disputes among different authors do not involve the definition
of the figure, but simply whether that definition can be included within
the idea of figuration. Other problems of relating the several systems
are not as clear cut. For example, Rudrața enumerates six kinds of
simile, calling one utpädya in which the object of comparison is hypoth-
esized. 195 Dandin, among his forty-odd kinds of simile, discusses several
which involve "hypothesis" (adbhuta, abhūta, asambhava, etc., but none
of these requires specifically the hypothesis of the object per se (rather,
a transfer of property to the object, or a generalization of the object,
or a predication of an incongruous property in the object).196 The type
193 Bhämaha, 2.86.
 
199 Dandin, 2.359.
 
194 Mammaţa, 134ff. This and later refs. are to serial order of topics, not verses.
195 Rudrata, 8.15.
 
19* Dandin, 2.24, 38, 39.