common
to South America and Sumatra, and not traceable to any country of
the Continent of India.

"It remains, however, to inquire into the language of the interior
tribes of Borneo, Celebes, and New Guinea; and, on such inquiry,
should they be found to possess the same primary roots as the rest, I
believe the conclusion must ultimately be arrived at of the existence
of a Polynesian language common to this vast geographical extent,
and distinct from the languages of Asia. In tracing this identity,
we can only, of course, find it in few instances in the cultivated
Javanese and Malayan languages. Discrepancies must naturally be great
from the intermixture, from early recorded times, of all languages
in the archipelago; but, nevertheless, if the radical affinities be
striking, they will be conclusive in establishing the original identity
of all the races before mentioned; for, without this original identity,
how can we account for these affinities of language? It may, indeed,
be urged that this language has gradually crept into the dialects of
Java and Menangkabau. But, in the first place, the affinities will
be found in the very roots of the language--in the expressions for
the primary and necessary ideas, which seldom alter in any people;
in the next, there is a high degree of improbability in supposing
a rude dialect to supplant a substantial portion of a more polished
one; and, thirdly, we must not overlook the collateral evidence of the
similarity of conformation pervading the entire race from Polynesia to
the archipelago--distinct alike from the Caucasian and the Mongolian.

"In tracing the identity of this language, we may reckon the dialects
of the Dyaks of Borneo, &c., as the lowest step of the ladder; those
of the Pacific islands next; and so through the dialects of Sumatra and
Tagala, up to the Malayan and Javanese. For this purpose, a comparative
view of all must be attained; and Eastern scholars should point out,
when possible, the words taken from Sanscrit and other languages. For
my own part, these remarks are made as a sketch to be enlarged on,
and to assist in obtaining the vocabularies of the Dyaks and Arafuras.

"_Dec. 6th._--In looking over Marsden's admirable Introduction to his
Malayan Grammar, I find I have taken many of his views in the foregoing
remarks; but I consider that his opinions may be pushed to conclusions
more extended than he has ventured upon. Having described the 'exterior
circumst