ON THE COMPARISON OF ASIATIC LANGUAGES. PA i 
Vambéry, and of many other distinguished scholars, have 
established a comparative study of dialects and languages, 
reaching from Siberia to Hungary, which, though less perfect 
than that of the mcre-studied Aryan languages, is equally 
based on sound scholarship and research. The number of 
roots to which the vocabularies are reduced is even smaller 
than that of the Aryan system, because they are more easily 
divided from their added suffixes, and are found to be almost 
entirely monosyllabic. Vambéry enumerates about 200 roots 
for Turkic speech, and these recur in the other divisions of 
the group. The third category of Aryan roots, which are 
bisyllabic, appears to be almost entirely absent, and the dis- 
tinction of letters and of vowels is much less perfect in 
Mongolic languages than in those of our own ancestors. 
Another peculiarity which marks these languages, and 
which is distinctly traceable in Akkadian, is that of “vowel 
harmony,” by which is meant that the vowel of the weaker 
root in a compound varies in accordance with that of the 
stronger root to which it is attached. We may, perhaps, con- 
jecture that the same harmony once existed also in Aryan 
speech, and that itis still discoverable im the parallel instance 
of roots having the same meaning’ but different vowels (as In 
the case of WA and WI, “to weave,” already cited) ; but if so 
it ceased at an early period to be a law of languag e. 
The fact alr eady cited that a simpie root may be reversed, 
as in the case of AR and RA, is also important for comparative 
purposes. The modern Turkic dialects generally prefer, in 
such cases, to put the vowel first, as easiest of pronunciation ; 
but in Akkadian we constantly meet with both forms. 
Among the peasantry of Palestine this inversion of the 
syllable is very clearly to be remarked. Thus, for instance, 
‘he word which means “ wells” is at pleasure Siydr or A biar ; 
and in the same way the Hebrew Len, “ son, > becomes Jbn in 
Arabic. It is well known that both Arabs and Hindoos find 
it difficult to begin some words with the letter S, so that, in 
the mouths of both, Mr. Smith invariably becomes Lsmit. 
We must touch in passing on the relation of Chinese to 
Mongolic speech, though the question is one full of special 
difficulties—Chinese being a very decayed language, in 
which sounds originally distinct have become much confused. 
Its vocabulary, however, still represents a recognisable con- 
nection with that of its western neighbours ; and attempts 
have even been made to compare C hinese directly with 
Akkadian. We suffer, however from the fact that we have 
