212 Sir Frederic Bartlett 



age prefers to have cues for performance and of the course of 

 performance from as many soin'ces as possil)le. In particular 

 it may be that sensitivity for proprioceptive feed-back 

 diminishes. For this however there is no direct evidence as 

 yet, or, in fact, of significantly lowered threshold response 

 in any direction. It is safer, though of course less satisfying, 

 to treat the demand for a greater amount and diversity of 

 confirmiup" evidence as one of the signs of the increased 

 caution which is alleged to accompany advances in age. 



It is more intriguing to consider the possibility that the 

 range and function of anticipation may change with age. We 

 have decisive experimental evidence that far and away the 

 most important character in skill performance is what has to 

 be called its temporal structure. The way in which any item 

 of skill is dealt with depends upon its relation to preceding 

 and succeeding items in the sequence of which it forms a part. 

 If a signal for action is given at an appropriate interval before 

 the moment of action it is the resting time between move- 

 ments in one direction and another, whether of muscle or of 

 mind, that is significantly reduced. It is a reasonable sugges- 

 tion that the general effect of accumulation of experience is 

 to enhance the influence of preceding items in any skilled series 

 and to depress that of succeeding items. There may therefore 

 be a retraction of anticipation range with age and if this is 

 so it would undoubtedly produce just that preference for 

 shorter advances and more and longer halts which older 

 performers are known to show. But only further experiments 

 can decide. 



There is yet another possibility. Every human action 

 exhibits a character which may be most forcibly described 

 as a "point of no return". That is, it reaches a phase beyond 

 which putting in new signals produces no effect whatsoever. 

 Before this stage is reached there is a range within which 

 incoming signals produce actions, but the action is mal- 

 adjusted: a mistake is made. It has become a matter of great 

 theoretic and practical importance to determine both the 

 absolute and the relative "points of no return" for all kinds 



