16 COMMON SENSE ( AXD SdENXE) 



catory concepts that teach us how to find in nonidentical things a 

 significant identit)' in t\'pe. Mill writes: 



We compare phenomena with each other to get the conception, and 

 we then compare those and other phenomena witli the conception. 

 . . . and if it [a particular instance] agrees with that general con- 

 ception, we include it in the class. The conception becomes the 

 type of comparison. 



There is this rattle— \\ith its own color, shape, weight, and sound. 

 There is this other rattle, with its own distinct set of qualities. And 

 there is the general concept "rattle." A member of the class must be, 

 in some degree, graspable and shakable; and it must gi\e a "rattling 

 sound," though this may vary greatly in \olume and timbre. In 

 forming the class some of the most striking qualities (e.g., color and 

 shape) are adjudged irrele\ant, and e\'en the key qualities are per- 

 mitted to \ar\- within broad limits. The discrimination so expressed 

 represents appreciation of the function of rattles. 



Function seems eenerallv decisi^•e in our understandino; of clas- 

 sificatory concepts, bodi in common sense and in science. Often— 

 though we all understand a concept in much the same wa\'— we can 

 find no specific criterion or group of criteria that establishes the 

 class. In such cases we deal far less with rigid specifications than 

 with a general pattern. Failure to fit the pattern in some respects is 

 disregarded if a good fit obtains in other respects that most concern 

 us at the moment. Our grasp of meaning is strengthened by our sense 

 of purpose— a point stated in its broadest application by Popper: 



. . . an>' two things which are from one point of view similar may 

 be dissimilar from another point of view. Generally, similarity, and 

 \\ith it repetition, always presuppose the adoption of a point of view: 

 some simihirities or repetitions will strike us if we are interested in 

 one problem, and others if we are interested in another problem. 

 . . . points of view, or interests, or expectations, are logical!)' prior, as 

 well as temporally (or causalh' or ps>'choIogically) prior, to repetition. 



Each of a third \'ariety of common-sense concepts highlights some 

 one attribute of a great many things. Though often unlike in all 

 other respects, these things can meaningfully be compared with re- 

 spect to this one "property." Such quality-concepts, abundantly rep- 

 resented in common sense, express recognition that for a gi\'en pur- 

 X^ose a certain quality is of overriding importance. In getting the first 



