Chap. VII, ANTIENT METAPHYSICS. 



79 



thought himfelf a philofopher, even though he had joined to 

 the knowledge of geometry, the fcience of numbers, which was alfo 

 very carefully taught in the Pythagorean fchooL It was only by 

 the ftudy of morals, natural philofophy, metaphyfics, and theology, 

 that a man in that fchool could deferve the name of a philofopher. 

 Thefe ftudies, however, of geometry and arithmetic, were held to 

 be very proper preparatives for philofophy ; and, I think I may add, 

 for logic, though even logic by thofe philofophers was not held to 

 be, properly fpeaking, philofophy, but only an organ of philofophy. 



As to the utility of logic, I need only repeat what I have faid in 

 more than one place of this work. That no man can ever know what 

 fcience is without ftudying the logic of Ariftotle, and muft reafon as a 

 child reafons, or as an unlearned man fpeaks, without knowing the 

 principles of the art, or being able to tell why one argument is con- 

 clufive and another not *. It is, therefore, furprifing, that any man 

 fhould pretend to be learned in any fcience, who does not fo much 

 as know what fcience is. 



But, in the progrefs from fenfations to fcience, there is a ftep 

 which is neceflary, and has been made by all men before they at- 

 tained to fcience, and that is opinion^ which is not like the conclu- 

 fions of fcience neceffarily true, but may be either true or falfe, as 

 it happens. All men, when they firft begin to think, muft form o- 

 pinions^ particularly concerning what is good or ill in human life. 

 And by far the greater part of mankind, as they never attain to 

 fcience, have only opinions by which they are governed. And, 

 therefore, when Polybius has faid, that man is Zuov ^o^oToinTiKov^ 

 that is an opinion-forming animal, he has given a very good defini- 

 tion 



• See the paflage above quoted from the preface to third volume of this work. See 

 alfo vol. I. book V. chap. IV. where the nature of the fjllogifm and its ufefulnefs are 

 explained at great length. See alfo vol. VI. of Origin of Language, p. 47. and follow. 

 ing. 



