432 Mr DE morgan, ON THE SYLLOGISM, No. V. AND 



what fish you please, it is not any fish : turbot is not trout. This is a slight error, easily 

 prevented by a postulate. 



The idioms of logical quantity have had very little consideration given to them. The 

 word some, although it may have points on which logicians divide, has one case of sub- 

 division upon which all logicians unite against the world at large. The distinction is that of 

 certain some, and some or other: the first has an unknown definiteness, the second is truly 

 indefinite. People in general incline to the unknown definite : the logician demands the true 

 indefinite, but can in many cases follow the usual tendency. Whenever one term of a pro- 

 position is a definite, known or unknown, the some of the other term is the unknown defi- 

 nite. As in 'AH men are [certain some] animals', or as in ' The men he spoke of were 

 [certain some of those who were] here yesterday'. But when one term is truly indefinite, 

 then certain some is not admissible in the other. Thus * any men are [certain some] animals' 

 is not true when 'any"' implies unlimited selection out of 'all'. This is most obvious in the 

 unusual exemplar forms: thus 'some animal is any man' would reduce mankind to an indi- 

 vidual if ' certain' some were intended. In some subsequent parts of this paper the reader 

 must watch himself on this point. 



Logic may take liberties with language for the expression of thought : but she must 

 not declare her alterations to be actual parts of speech. I fully understand and agree to 

 the assertion that complete quantification may be made to allow simple conversion ; that ' some 

 X is not any Y' may infer 'Any Y is not some X'. Nevertheless, this cannot be admitted 

 if subjection and predication remain notions attached to the subject and predicate : for pre- 

 dication is posterior to subjection ; the subject comes first into thought, and the question of 

 predication /bWows. For instance, 'some man is not any animal' is a falsehood: designate 

 the man, and a search through the animals will find him. But ' any animal is not some 

 man' is true : choose any animal, man or not man, and we can then show some man which 

 he is not. In order to make this last proposition as false as its converse, the right of prece- 

 dence must pass over to the second side with the term which originally had it. Of ' any 

 animal', first chosen, 'some man' may be denied: of 'some man', first chosen, 'any animal' 

 cannot be denied. The same thing in every case in which soff*e comes into contact with all 

 cr any. Hamilton saw this, and it made him insist upon enunciation being pure equation or 

 non-equation of subject and predicate, meaning identification- or differentiation of simul- 

 taneously entering terms. But Hamilton had the faculty of fastening upon his whole species 

 any use of language into which he had drilled himself. Thus (IX. ii. 294) he says — •" Why, 

 for example, may I say, as I think, — Some animal is not any man; and yet not say, con- 

 vertibly, as I still think, Any man is not some animal? For this no reason, beyond the 

 caprice of logicians, and the elisions of common language, can be assigned." If he should 

 think it, he may say it : but in common language, and this with no- elisions, ' Any man is not 

 some animal' does not contradict 'Every animal is man', as he intends it should. For 

 though every animal had been man, yet any man would not have been some animal. Com- 

 mon language makes subjects of terms and then predicates of them. 



The word some has three distinct uses. First, as non-partitive; here it is only not-none, 

 some-at-least, some-may-be-all. And this is the old sense of the logicians. Secondly, as 



