1.6 Rationality and Extrapolation
The discovery of identities among the contents of one’s perceptions is a core
function of reason. There is another. This is the capacity to extrapolate from
what one has perceived to what one has not. That extrapolation is a key concept
in the early history of Indian logic is clear from some of the examples Va ̄tsya ̄yana
gives under NS 1.1.5.^26 Seeing a rising cloud, one infers that it will rain. An inter-
esting variant is: seeing the ants carrying their eggs, one infers that it will rain.
Seeing a full and swiftly flowing river, one infers that it has been raining. Seeing
a cloud of smoke, one infers the existence of an unseen fire. Hearing a cry, one
infers that a peacock is nearby. Seeing the moon at one place at one time and at
another place at another time, one infers that it is moving (even though one
cannotseeit move). The medical theorist Caraka^27 has some other examples:
inferring impregnation from pregnancy; inferring the future appearance of fruit
from the presence of seeds. In the ancient Buddhist logical text, The Essence of
Method,^28 we also find: inferring from a child’s special mark that this person is
that child, now grown up; inferring from the salty taste of one drop of sea water
that the whole sea is salty. The Ts’ing-mu^29 (a commentary on Na ̄ga ̄rjuna’s Middle
Stanzas) has a similar example: inferring that all the rice is cooked on tasting
one grain. And the Vais ́es.ikasu ̄tra mentions another sort of extrapolation – the
inference of an entire cow from the perception only of its horns.^30
Extrapolation from the seen to the unseen can take place in any of the three
dimensions of time – past, present, and future. Our interest is in the Indian
theory of rationality, and for this we want to look at answers given to the ques-
tion: on what basis, if any, ought the extrapolation be made? For while dice-
throwing, guesswork, and divination are ways of extrapolating, they are not
rational ones. Extrapolation, like critical investigation, must be done “on the
basis of reasons,” and a theory of such “reasons” is a theory of that in virtue of
which an extrapolation is warranted. So we can discover Indian theories of ratio-
nality in their explanations of why the extrapolations in the examples mentioned
above are warranted. Rationality now is the search for extrapolative license.
While Va ̄tsya ̄yana says only that there should be a connectionbetween what
is seen and what is inferred,^31 many of the early writers have a definite interest
in prediction and scientific explanation, and assume that extrapolation is war-
ranted when underwritten by a causal relation. On the other hand, it is clearly
recognized too that not all warranted extrapolation is causal. The Vais ́es.ikasu ̄tra
lists, in addition, the relations of contact, inherence, coinherence in a third, and
being contrary (VS 3.1.8, 9.18), while the early Sa ̄m.khya s.as.t.itantra has an over-
lapping list of seven.^32 Take the inference from a drop of salty sea water to the
conclusion that the whole sea is salty. This is not an inference based on any
causal relation between the drop of sea water and the sea as a whole; rather, the
relation between them is mereological. One would say that it is an inference from
sampling a “typical” member of a group. This is a very common and useful form
of reasoning (witness the example of checking that all the rice is cooked by
hinduism and the proper work of reason 429