296 Journal of the Asiatic Society of Bengal. [NS., XIV, 
The statement that “inference is not a means of know- 
ledge ” is of no use to an expert who is aware of the invalidity of 
inference. It may be advanced usefully before a suspicious or 
an erring person. Now, how are we to know that the person 
is in suspicion or error? It is certainly not through perception 
but through inference. 
- “Inference is not a means of knowledge.” Is this 
statement valid or invalid? If it is valid we must admit ver- 
bal testimony to be a means of knowledge. If on the other 
hand it is invalid, inference is to be admitted as a means of 
knowledge. Both the alternatives are incompatible with the 
- Opponent’s view. 
4. 
there could not arise any doubt about its validity. 
e 
Inference is therefore a means of knowledge, and percep- 
expressions as pervasion, inseparable connection, perpetual 
attendance, constant co-presence, etc. 
aif eRz | 
Five Provisional Definitions of Invariable Concomitance. 
__ 1. Invariable concomitance is the non-presence of “ 
rine, Tm in the locus of the non-existence of the maj® 
ns Se The hill is full of fire, 
because it is full of smoke. 
Diagram I. Diagram I. 
EG! 
gs eo, 
NAMEABLE - 
KNOWABLE 
ZS) 
u 
oO 
wo 
< 
3 
h 
Ny 
-E 
ONSIRE “o> 
Jig 40 
_,_ Here smoke (the middle term) is absent from all regions 
side the region of fire (the major term) as shown in diag” 
out- 
1. 
| 
| 
| 
