________________
NATURE AND FORMS OF KNOWLEDGE 51 the inference : 'Whatever is smoky is fiery; this object is smoky ; therefore this object is fiery. If this conclusion be false, then its contradictory, this object is not fiery' should be true. But the latter proposition is found to be absurd by the following tarka. If in the case of this object smoke is not related to fire, then it cannot be an effect of fire. But it must be due either to fire or to not-fire There is no third alternative here. We do not find it to arise out of not-fire Hence if it is not due to fire, it must be either an uncaused effect or a non-existent phenomenon The first alternative contradicts the law of universal causation and is therefore untenable. The second alternative becomes self-contradictory, since it commits us to the proposition that the smoky object is smokeless For, if A (smoke) be a mark of B (not-fire), and B (not-fire) were a mark of C (not-smoke), then A (smoke) would be a mark of ( (not-smoke) In view of such absurdities involved in the contradictory of the original conclusion we must reject it as false and accept the original conclusion as true and as based on a valid inference.
It is to be observed, however, that the Nyāya division of tarka into five different kinds is logically unsound This division has reference to the different kinds of reasoning which may be tested by an argument like tarka. But the classification of tarka should not be based on the kinds of reasoning that may be tested by it, because these are unlimited and quite external to the nature of tarka as a type of argument. A classification of tarka must be based on the logical character of the arguments employed in different cases. Now having regard to its logical character, we find that tarka is fundamentally of one kind. In every case in which it is employed it has the form of an inconsistent argument (anıştaprasanga) developed out of the conclusion of a given reasoning or its contradictory. If
