This page has not been fully proofread.

144
 
GLOSSARY
 
In this way, the various figures involving a duplication of the
context are explained and reduced to similes. We need not charac-
terize them more fully here, since at least half of this work is con-
cerned with just that problem. However, some accounting of the
various classifications proposed for simile itself is necessary.
 
The general tendency is for the discriminations or subtypes more
and more to be based on the quadripartite structure we have given.
The earlier writers, however, while obviously recognizing that
structure in defining the basic figure, tended to classify simile in
terms of the end or final cause of the comparison. This is especially
interesting in view of the universal preoccupation with structure,
even among the early writers, in classifying rüpaka. This difference
seems to reflect the character of the problem under consideration.
Since comparison is always a matter of degree, it would appear
appropriate to consider the usage of the various degrees, which is
not a question of structure, but of the kinds of things compared
and the reasons for that selection of things; but metaphor, being
identification pure and simple, is never a question of degree (except
in the sense that the metaphor can be more or less well specified
in its parts, or complete), and the only relevant question concerns
the scope of the identification, which has little to do with the things
themselves, but is entirely a matter of the poet's employing or
not employing the ideal metaphorical type (see rūpaka).
 
In classifying upamā, the non-structural, contextual, tradition
may be said to begin with Bharata himself, for he allows similes of
praise (praśamsā) and blame (nindā), as well as three similes which
differ as to the degree of comparability intended by the poet:
sadrsi, or entirely comparable, that is, where the subject and object
possess the same property to a great degree; kimcitsadrší, where the
same subject shares comparable qualities with several objects
and is therefore partially comparable; and kalpita, where, strictly
speaking, no comparability at all is alleged; that is, no property is
described as common to both subject and object, but rather different
descriptive properties are assigned to both which are, in fact, similar
(the similarity is not literal, but analogical).
 
The other, or structural, tradition may claim almost the same
antiquity, for Bhāmaha, who specifically objects to the classification
by praise and blame as irrelevant (2.37), enumerates only three
kinds of upamā, depending only on the grammatical device by which
the similitude is expressed. We have mentioned previously only the