Part 3 - VP’s Definition of anumAna
anumAna: VP defines anumAna as the instrument of
inferential knowledge, which is itself called anumiti
(anumiti karaNam anumAnam). anumiti is the knowledge
that follows (anu = after and miti = knowledge). I.e.
it follows another knowledge, namely knowledge of
some data. The knowledge that follows has to have
some bearing upon the knowledge that preceded it.
Hence the later knowledge is gained only because of
its inherent relation to the former knowledge. If
the inherent relation is not known, then the later
knowledge will not take place. Hence, the later knowledge
is produced as a result of the knowledge of the invariable
concomitance or vyApti. The nature of the later knowledge
therefore depends not only on the knowledge of the
data that is perceived but also on the exact knowledge
of the vyApti or the invariable concomitance. Hence
the knowledge that follows, anumiti, is not an attributive
objective knowledge but a knowledge that is purely
based on logical deduction, which in turn is based
on the knowledge of the invariable concomitance.
Invariability means the universality of the relation,
implying that there is no exception to the rule. Taking
the example of smoke and fire, perception of the smoke
is direct objective knowledge which is attributive.
I.e. the vRRitti that is formed has in its contents
the attributes of the smoke, and smoke is recognized
directly and immediately. I.e. it is based on pratyakSha
pramANa. anumiti follows if we have knowledge of the
vyApti that relates the smoke to fire. That is, wherever
there is smoke there must be fire, and this is a universal,
invariable concomitance or vyApti. Hence, the inferential
knowledge that follows depends on the exact knowledge
of this invariable concomitance.
Suppose that we conclude, based upon the vyApti,
that there must be fire. Fire, in this case, is not
an objective knowledge like that of the smoke. If
smoke were related to dust, we could say that ‘there
is dust, based on the invariable concomitance that
wherever there is smoke there is dust.’ Thus,
the knowledge that it is fire or dust depends on the
nature of the vyApti. These examples illustrate the
fact that inferential knowledge is as good as, or
as valid as, the knowledge of the invariable concomitance
and nothing more. The validity of inferential knowledge
depends on the validity of the vyApti only. Hence
vyApti or invariable concomitance forms the core of
When the perception of the smoke occurs, as we discussed earlier,
a vRRitti forms based upon the perception of the attributes
of the smoke. The resulting knowledge of the smoke is direct
and immediate. Along with knowledge of the smoke there is immediate
cognition of the knowledge ‘I know this is smoke’.
Thus, ‘this is smoke’ occurs first via the vRRitti
and is then followed by what is known as the apperception ‘I
know this is smoke’. We mentioned that, according to
Advaita, apperceptual knowledge occurs because knowledge is
self revealing. I.e. we do not need another means to know that ‘I
know.’ If another means were required then we would end
up with an infinite regress, since the ‘knowing that
I know’ would require another means of knowledge, and
so on. Apperception is not inferential knowledge, but ‘self-revealing’ knowledge.
VP says that, even in the case of the inferential knowledge
that ‘there is fire there since I see smoke’, that
inferential knowledge of the
fire is self revealing and does not depend again on another
vyApti, since that would lead to infinite regress. Hence VP
says that, once the inferential knowledge through vyApti is
known, that inferential knowledge is self-revealing and does
not depend on another concomitant relation. Similarly, the
recollection of the vyApti (that wherever there is smoke there
must be fire) is not based on another concomitant relation.
The vyApti is based upon previous experience of the cause-effect
relationship, which is itself established by pratyakSha pramANa
only. Thus, the vyApti ‘wherever there is smoke there
must be fire’ is established by direct observation of
the relation between smoke and fire. The knowledge of the vyApti
has to be acquired from past observations. Now, when I see
smoke on a distant hill, the vyApti or the concomitant relation
between the smoke and fire is recalled from memory. VP says
that recollection of the vyApti relating smoke and fire is
not based on another vyApti or relation, since such requirement
would lead again to infinite regress. Hence, it is said that
a vyApti is not based on another vyApti for its operation.
We can, however, have a sequential logical deduction
before inferential knowledge takes place. The vyApti
chain can be of the form: A is related to B; B is
related to C; C is related to D; and therefore A is
related to D. Here, each one is a definite and precise
relation. Ultimately, A is related to D and that vyApti
involves interlinking to B and C via secondary vyApti-s
or concomitant relationships, each being universally
applicable, in order for doubt-free knowledge to take
place. From the hetu A, in order to arrive at the
sAdhya D, one has to have complete knowledge of the
vyApti chain. If there is any missing link in the
chain, one would not arrive at inferential knowledge
of D. Ultimately, the A to D relation forms one compound-vyApti
involving interlinking logical deductions, which are
needed in order to arrive at the inferential knowledge
of D from A.
A vyApti is not a postulation but a universal concomitant
relation between two things. In addition, ambiguity
in the knowledge may lead to doubt if the concomitant
relation is not universal. This means that, if there
are many exceptions to the relation, then the inferential
knowledge will not be free from doubt. As we discussed
before, doubts are different from errors. For example,
if I am not sure whether the object in front of me
is a rope or a snake, this is considered to be a ‘doubt’.
However, if I am sure that is a snake, even though
it is actually a rope, then it is an ‘error’.
In the case of a doubt, the knowledge is subject to
verification by the doubter. However, if one has concluded
that an error is the truth, there is no desire to
enquire into the real truth. In the case of the world,
we have concluded that what I perceive is real and
therefore the world is real in our mind.
There is no desire to inquire into the absolute reality
of the world. Scripture points out that our conclusion
about the world is in error, which we will discuss
with reference to shabda pramANa.
Coming back to the anumAna, both advaitins and naiyAyika-s
agree that inference, as a means of knowledge, operates at
a) svArtha, inferring for oneself and
b) parArtha, logically deducing for others.
In inferring for oneself, he remembers the concomitant
relation with what he sees, and deduces what he does
not see. When he sees the smoke, he remembers that
smoke cannot exist without fire, and therefore infers
that there is fire, although he cannot see the fire.
But when presenting these facts to others, he has
to provide a formal statement of reasoning (syllogism)
in order to convince them of the fact there is a fire
even though they cannot see it. The syllogism involves
three steps, according to advaitins while naiyAyika-s
feel that five steps are required to convince others.
A detailed discussion of this can be found in ‘Methods
of Knowledge’ by Swami Satprakashananda of Advaita
Ashrama. The direct and necessary parts consist of:
a) proposition or pratij~nA as in ‘there is fire on the
hill, although we do not see’;
b) the reason (hetu) this proposition is made: ‘because
we see smoke on the hill’ and
c) justification with example – vyApti with dRRiShTAnta: ‘wherever
there is smoke there must be fire, as in the kitchen’.
In western logic, the vyApti is considered to be
premise’ and the current observation is called
the ‘minor premise’, based on which a conclusion
is made. In the example, the major promise is ‘wherever
there is smoke there must be fire’; the minor
premise is ‘the hill has smoke’ and the
conclusion is ‘therefore the hill is on fire’.
The naiyAyika-s subscribe to a five step process and
the way in which they differ from Advaitins will be
discussed next. VP highlights these differences.
Proceed to the next