i 5 4 MIND IN EVOLUTION CHAP. 



Our memory of a person or place seen once is seldom 

 perfect. It fades with time, and is strengthened by renewal. 

 In these respects it clearly follows the same law as memory 

 due to repetition, and we may infer that there is no 

 fundamental difference. The real distinction lies in 

 connectedness of elements. If A B c form a rigidly 

 and necessarily connected whole for thought, to re- 

 member A would be to remember B and c as well. If 

 they are utterly disconnected, it requires three distinct 

 efforts to recall them. There is an intermediate case in 

 which they are apprehended in relation, though not in 

 necessary relation, so that the remembrance of one 

 naturally, though not inevitably, leads on to another. 

 Thus I may readily remember a connected conversation 

 or a well put argument, as each point leads naturally to 

 the next. So again the metre as well as the thought 

 help in the memory of verse, while in matters of simple 

 perception, to recall one part, for example, of a once 

 visited place is to have a beginning from which the space 

 relations themselves help one in reconstructing the rest. 

 In short, if instead of A, B, c, three quite separate data, we 

 have A B c so related that one passes into the other 

 or is continuous with it in perception, or leads up to it in 

 meaning, we have quite a different groundwork to go 

 upon. Our memory rests partly on the connection. In 

 the alternative case it rests on association, the bringing of 

 two things together in consciousness without any recognised 

 connection. The difference which is made by connected- 

 ness may be estimated by comparing the number of re- 

 petitions necessary for learning (i) a series of isolated 

 words or meaningless syllables, and (2) a connected piece 

 of writing. Comparing words and phrases, MM. Binet 

 and Henri estimate the advantage of phrases in one case as 

 25 to i. 1 Comparing a series of meaningless syllables 

 with a stanza of Don Juan, Ebbinghaus 2 found that 

 ten times as many repetitions were necessary in the former 

 case. 



If we could say that association only acts by more or 

 less frequent repetition, the problem of determining what 



1 Annee Psychologique, 1894, p. 31. 2 Gedachtniss , p. 69. 



