a helpful and necessary simplification for the sake of initial progress, also
restricted the study of other patterns of inferential reasoning. Only the Jaina
logicians explicitly tried to develop a theory of extrapolation free from this
restriction.
What licences the inference from Ra toTa? The Nya ̄yasu ̄tra answer is given in
five brief and controversial aphorisms (NS 1.1.34–8):
A reason is that which proves what is to be proved by being like an example.
Again, by being unalike.
An example is an observed instance which, being like what is to be proved,
possesses its property.
Or else, being opposite, is opposite.
The application is an assimilation to what is to be proved “this is thus” or “this is
not thus” depending on the example.Likeness and unlikeness are relative to properties. Something is “like” another
thing if both possess a given property. They are unalike with respect to that prop-
erty if they do not both share it. Now arguably the natural way to interpret these
su ̄tras is as follows. Either the locus of the inference is like the example (in that
both possess the reason property, R) and, since the example has the to-be-proved
property, so does the locus. Or else the locus of the inference is unlike the
example (it possesses the reason property, but the example does not), and since
the example does not have the to-be-proved property, the locus doeshave it. If we
let “b” stand for the example, then we seem to have:
ais likeRba is unlikeRb
Tb ~Tb
\Ta \TaThis formulation actually makes the inference a generalization of the inference
from sampling. The example is a typical member of the class of things having
the reason property. And it has this other property, the to-be-inferred target
property. But the site of the inference is also a member of the class of things
having the reason property. So it too has the target property (the negative for-
mulation is similar). This is a powerful form of reasoning, one which we engage
in all the time. It is not formally valid, but it is a pervasive type of informal rea-
soning. We employ it whenever we infer that an object has a property on the
grounds that it belongs to a type, the typical members of which have that prop-
erty. Compare: this grain of rice is typical of the whole pan of rice, and it is
cooked. So any other grain will be cooked as well. This drop of water is typical
of the entire sea, and it is salty. So this other drop must be salty as well.
We said that in the debating model, rationality is subject to public norms of
correctness. In arguments of the kind being considered, public norms do indeed
have a role to play, for they determine whether the object adduced by the debater
hinduism and the proper work of reason 433