334 REPORTS ON THE STATE OF SCIENCE.—1912, 
with, and when some practice in combination has been done, it is 
possible, in so far as the language is strictly phonetic in spelling, for 
the child to attack any word with a fair measure of success: he may 
be able to produce the spoken form unaided, even if he has never seen 
the printed form before. ‘ 
But such ‘ barking at print ’ is not reading. This, as we have seen, 
involves gathering the meaning. And in so far as the child’s attention 
is absorbed with the process of deciphering the words, no true reading 
can occur. But after each act of deciphering the resulting sound- 
combination strikes the child as familiar. It is a word with which he 
has already a meaning associated. And as less and less attention is 
required for the process of deciphering, more and more can be given 
to the meanings aroused. Thus true reading begins. 
But such reading as this would be extremely halting and laborious. 
By whatever method a child is taught, he comes ultimately to read in 
much shorter time than any which would allow of deciphering. Prac- 
tically all the experiments which have been made with the object of 
investigating the processes which take place in reading, in the case of 
those who have already acquired the art, confirm the view that reading 
is by word-wholes—sometimes, indeed, by sentence-wholes. Thus 
Javal, Huey, and Dearborn have shown that the eye only makes two 
or three pauses, each of a small fraction of a second (usually 4, though 
varying much) in each line, the swift movements between these pauses 
being too rapid to enable any distinct perception to occur. By tachisto- 
scopic experiments, Cattell showed that ‘ the shortest exposure which 
would permit the recognition of single small letters and capitals sufficed 
also for the recognition of short words, and that long words needed but 
one-thousandth of a second more.’?4 Additional experiments by 
Zeitler, Erdmann and Dodge, Goldscheider and Miller, Messmer, and 
others confirm these results. Huey further found that four readers, 
whom he tested, proceeding as fast as possible, read 
50 letters in an average of 15:7 seconds 
50 four-letter words 3 a ERS its 
50 eight-letter words % = 19°6; / V5 
In all these results there is hardly a word to suggest that we read 
by the aid of phonic principles. The only inharmonious note is that of 
Goldscheider and Miller, who think that the ‘ determining letters ’ }* 
may suggest the word by reason of their sounds being aroused. But 
‘further investigation has not justified Goldscheider and Miiller’s con- 
clusion. . . . The word-sound seems usually to be suggested as a 
whole.’ 18 
words as seen, they are far from being the constituent units of words heard and 
spoken, the wholes in the latter case being clearly other than a mere aggregation 
of the letter sounds, even when the proper phonetic value is given to these.’ 
' (Sully, op. cit., p. 196). -Meumann, who recommends the Phonic Method, is never- 
theless fully alive to these difficulties, and criticises severely the attempts of 
Olivier (Lautiermethode) and-Spieser (begriffliche Methode) to overcome them 
(op. cit., p. 219), It may be said, however, that these difficulties are not perhaps 
so real as they are made to appear, especially in the case of a language which is 
consistently phonetic in its spelling. 
** Huey, ‘The Psychology and Pedagogy of Reading,’+p. 73. 
” Seo footnote on p. 332. % Huey, op. cit., p. 146. 
