Page History

...

From the general principle that one can infer from the reality of x to the 
reality of whatever is, in fact, a necessary condition of the possibility of x, it by no 
means follows, without more ado, that one can infer to the reality of y from the 
reality of x. For that to follow, there must be argument to the effect that y, in 
reality, is a necessary condition of the possibility of x-so that: -y :::>-x. 

From the general principle that one can infer from the reality of x to the 
From the general principle that one can infer from the reality of x to the reality of whatever is, in fact, a necessary condition of the possibility of x, it by no means follows, without more ado, that one can infer to the reality of y from the reality of x. For that to follow, there must be argument to the effect that y, in reality, is a necessary condition of the possibility of x-so that: -y :::>-x. 

But, further, even if this premise is argued for, it is important to be on the lookout for the fallacy of four terms that is likely to occur when the meaning of y in -y:::> -x is different, however subtly, from the meaning of y in x:::> y. Put differently: in a valid reductive argument, the only condition whose necessity can be asserted in the conclusion is the same condition that is asserted in the major premise and therefore also requires to be argued for. 

...