cted those displacements which separated, or were supposed to
separate, the Rajmahali, Kol, and Khond dialects from each other. They
formed the _sea_ of speech, in which those tongues were _islands_.

Now what is the inference from these per-centages? from such a one as
the Bengali, of ninety out of one hundred? What do they prove as to the
character of the language in which they occur? Do they make the Sanskrit
the basis of the tongue, just as the Anglo-Saxon is of the English, or
do they merely show it as a superadded foreign element, like the
Norman--like that in kind, but far greater in degree? The answer to this
will give us the philological position of the North-Indian tongues. It
will make the Bengali either Tamul, with an unprecedented amount of
foreign vocables, or Sanskrit, with a few words of the older native
tongue retained.

If the question were settled by a reference to authorities, the answer
would be that the Bengali was essentially Sanskrit.

It would be the same if we took only the _prima facie_ view of the
matter.

Yet the answer is traversed by two facts.

1. In making the per-centage of Sanskrit words it has been assumed that,
whenever the modern and ancient tongues have any words in common, the
former has always taken them from the latter,--an undue assumption,
since the Sanskrit may easily have adopted native words.

2. The grammatical inflections are so far from being as Sanskritic as
the vocables, that they are either non-existent altogether,
unequivocally Tamul, or else _controverted_ Sanskrit.

Here I pause,--giving, at present, no opinion upon the merits of the two
views. The reader has seen the complications of the case; and is
prepared for hearing that, though most of the highest authorities
consider the languages of northern India to be related to the Sanskrit,
just as the English is to the Anglo-Saxon, and the Italian to the Latin;
others deny such a connexion, affirming that as the real relations of
the Sanskrit are those of the Norman-French to our own tongue, and of
the Arabic to the Spanish, there is no such thing throughout the whole
length and breadth of Hindostan as a dialect descended from the
Sanskrit, or a spot whereon that famous tongue can be shown to have
existed as a spoken and indigenous language.

But, perhaps, we may find in Persia what we lack in India; and as the
modern Persian is descended from the Zend, and as the Zend is a sister
to the Sanskrit, Persia may,