ON ETHNOLOGY. 
331 
have been adopted by those dialects. I shall abstain at present from enter¬ 
ing into any discussion upon the origin of those words which do not belong 
to the Indo-Gerrnanic family, and of which Dr. Stevenson has given some 
comparative lists, tracing analogies in the Mongolian, Celtic and Hehren 
tongues. I do not think that the affinity of different languages in any 
country c-in be proved by a mere comparison of similar words, and it seems 
to me that by producing analogies from languages so difleront as Hebrew, 
Celtic and Alongolian, one proves nothing by proving loo much. 
e must have studied the individuality of different languages, we must have 
acquired an intimate knowledge of the particular distinguishing character of 
each of them, we must have enteretl into the Bpirit of every idiom, and have 
acquired a kind of feeling »o as to be able to identify ourselves with the lan¬ 
guages of other people, before we can venture to decide upon analogies 
which may exist between them. Afterwards it makes no difference whether 
these analogies consist in words or in terminations of words, whether they 
be etymological or grammatical analogies, provided lliat the one and the other 
be based, not upon the mere sound, but upon the organisation of the words. 
U is on that account that I innat declare myself decidedly, as far as the 
Bengali is concerned, against Dr. Stevenson's theory. Ur. Stevenson says 
that there exists a great resemblance in the gnimmaiical structure of the 
chief modern Janguages in the north and in the south of India, proofs of 
which he produces from iho Himli, Bengali, Gmeratlil, Marathi on the one 
side, and from rdngii. Carnatica, Tamil anti Singhalese on the other. 
Supposing that for none of these characteristic points they are indebted to 
the Sanscrit, he thinks it impossible to account for such a aimilarity of 
grammatical structure in languages, spoken by people having so little inter¬ 
course with one another, as, according to his opinion, the Hindu inhabitants 
of the north and south of India have had, unless vve suppose it to arise from 
their all being originally of otio family, and possessing one primitive lan¬ 
guage, the grammatical system of which may he in sonio measure gathered 
from these their points of agreement. Pr. Stevenson admits however that 
Braliminical influence has modilied the grammatical structure, and intro¬ 
duced into the northern languages some affixes for those in former use, 
especially in the inflexion of nouns, but he says that the general structure 
of all has remained unaffected, and that upon the whole there is more agree¬ 
ment in the construction with the 'J'ur/iinh than with live Sanscrit, so that 
he thinks it likely that ibe original language of India may be the eonnect- 
iog link between what the Germans have called the Indo-Germanic family 
and the Turkish family of languages. 
Now the whole questiun, as far as I can sec, resU upon these two points; 
ta Jt likely or not that the northern langiiagca of India, which are so 
much connecAd with the Sanscrit, that while the Bengali and Hindi, 
which probably contain the moat, have nine-tenths of their vocables of 
Sanscrit origin, and while even the Marathi, which, according to Dr. 
Stevenson a estimation, runtains the fewest, has at least four-fifths of its 
words derived Irorn the name source, the same languages should have 
derived their inflexioniil suffixes from an aboriginal language, which ex¬ 
ercised so little influence upon those modern dialects, that proofs of its 
very existence can only be gathered from gome few words, which, denoting 
things connected with the daily occupations of the working classes, were 
likely to remain in the mouth of the people, and to get by this way intro¬ 
duced into the language of tlie higher classes? Dr. Stevenson himself 
admits that the Braliminical influence has modified the graminatica] struc- 
