508 THE BELL SYSTEM TECHNICAL JOURNAL, MARCH 1957 



cabularies. The speed for reading loud and soft was less than for reading- 

 while-keying. This may imply something about the relative efficiency of 

 human beings performing two tasks by using two sets of muscles as 

 against using one set in two different ways. 



It is common experience that we can walk about and carry out other 

 simple tasks while talking or thinking. It is possible though not obvious 

 that some sort of automatic, almost purely reflexive response — ■ as, 

 moving the left hand when the right hand is touched — could with 

 practice be carried out cjuite independentlj^ of a task such as reading. 

 The information rate for such responses would be small, the experimental 

 error would make it difficult to settle the question, and the interpretation 

 of such an experiment would not be entirel}^ clear. 



The Patterns Which Govern Reading Time 



Early in the experiments the question was raised whether readers may 

 not read letter b}^ letter or syllable by syllable. Several findings bear on 

 this. 



Fig. 3 shows clearly that the reading time for a two-syllable word is 

 much less than twice the reading time for a one-syllable word. 



One of us knows a negligible amount of German. German syllables are, 

 however, reasonably familiar. It was found that in reading German 

 aloud he had the same reading rate in syllables per second as a man 

 whose native language was German had in words per second. The two 

 readers had substantial!}'' the same reading speed in English. Presumably 

 in reading German one man recognized syllables and the other recog- 

 nized words. This also reinforces the conclusion that reading rate is not 

 limited by the time taken to utter words. 



Some experiments were done using lists of common Chinese characters 

 and lists of the corresponding English words. Average word rates over 

 three lists for two readers Avho could read both languages are given in 

 Table IV. The slightly lower rate for English is plausibly explained by 

 the fact that Chinese was the reader's native language. All words were 



Table IV 



