THE LANGUAGES OF THE INDIAN ARCHIPELAGO* 205 
cause, that the Malay is in the condition which we have indi¬ 
cated, and he has neglected the colloquial language, and the 
laws of sound under the influence of which it grew to its pre¬ 
sent state. 
It will be seen how unsuited to such languages in par¬ 
ticular, any arrangement must be which arbitrarily isolates 
substantives, adjectives, verbs, and adverbs from each other. 
Those who know how soon an illiterate native becomes 
wearied and impatient when communicating words for a 
vocabulary, will understand the difficulty of maintaining for a 
few hours together and of renewing from day to day, his at¬ 
tention and interest, when the object sought is not a list of a 
few hundred words, but a whole larguage comprising several 
thousands. Now an alphabetical and grammatical arrange¬ 
ment increases the difficulty tenfold by rendering the task 
as abstract, unnatural and fatiguing to his mind as it is pos¬ 
sible. No sooner has his attention, often with much trouble, 
been directed to a certain idea, and the word expressive of 
a particular modification of it written down, than his mind, 
still occupied with the connected ideas and desirous of com¬ 
municating them, is required to discard them, and fix itself 
on something else with which they may have no associations 
whatever. Painful experience has taught us that the only 
successful method is to give free play to the natural current 
of ideas, and make the writing down of words as nearly as 
possible a full copy of the pictures "which our inquiries suc¬ 
cessively evoke, in the mind whose verbal riches we seek to 
extract. 
Impressed with these facts, we proceeded to devise a more 
natural system of classification, but soon found that what seem¬ 
ed sufficiently easy in principle abounded with difficulties of 
detail. We sought in vain for any dictionary constructed in 
accordance with such a system, and the imperfections of our 
first draughts satisfied us that a thoroughly scientific arrange¬ 
ment of a whole language, was a work demanding much more 
time and labour than we could command. One difficulty is 
indeed insurmountable. Under any conceivable system many 
words will appertain to more than one class, and as, for com¬ 
parative philology, it would be only a waste of space to repeat 
them, they must be arbitrarily confined to one class, which 
will leave one or more of the remaining ones somewhat maim¬ 
ed. Words which have distinct meanings, as unfortunately 
too many have in most languages, must of course be repeated 
in different classes or subdivisions. After recasting our clas¬ 
sification several times, we adopted one which we have found 
