AN GAOḊAL
451
PROF. ROEHRIG ON THE IRISH LANG-
UAGE.
Continued from page 438,
In Irish, particularly, the initial letters under-
go certain changes according to the grammatical
position of the words, and their use in composi¬
tion. The relation which words sustain to others
in a phrase or sentence, is thus indicated by mut¬
ations in their initial parts, and those changes are
phonetically adopted to the final letter which pre¬
cede them. Such a peculiar modification of the
initial characteristic seems to point far back to
antiquity, — perhaps to a state of speech, a condit-
ion of language, prior even to that of Sanskrit.
The direction taken by these changes, in the Cel¬
tic group is just the reverse of that which occurs
in the other Indo-Euporean languages. On the
whole, however, the phonetic system in Celtic re¬
sembles, generally speaking, that of Sanskrit.
Vowels undergo changes according to certain rules.
The euphonic laws in Sanskrit are, moreover, so
plainly recognized in the Celtic languages, that we
are obliged to conclude that these languages were
already much developed when the two forms of
speech separated. The Sanskrit phonetic changes
called guna and vriddhi (by prefixing a to certain
vowels, and producing, by fusion, e and o in one
case, and at and au in the other), can be shown to
exist in the Celtic tongues, where it is sometimes
to be clearly recognized, and, at other times, seen,
more or less modified, altered or obscured, but re¬
mains, nevertheless, discernible for the practised
eye of the philological observer. So we can also,
in Greek and other Aryan languages, discover
such guna and viriddhi changes, though, perhaps,
somewhat distorted in the vocalic elements; e.g.,
i (as in imen and eimi; fug (as in efugon) and
feugo; lip (as in elipon) and leloipa etc. Guna
of u. viz., o (=a x u) which is characteristic of
the genitive singular in the Sanskrit declension of
stems ending in short u, as, for instance, bhanu,
dhenu, where it affects, however, only the end-
vowel (u) of the stem or base, — e. g., bhanos,
dhenos. — occurs also in Irish, but bears on the
root-vowel itself as, for instance, —
sruṫ (sruth), gen. sin. sroṫa (srotha),
cruṫ (cruth, croṫa (crotha; lus lus),
losa (losa); gruṫ gruth, groṫ (grotha;
gul (gul), gola (gola); guṫ (guth), goṫa
(gotha), &c.
A great many Celtic roots are identical with
those of Sanskrit; and the Irish language possess-
es, also, very many words that are to be derived
from, or connected with, such as Sanskrit roots
as have been, hitherto, standing isolated, and
in no wise be analyzed, classified, or accounted for
in our dictionaries. The Celtic roots are, moreo¬
ver, for the greatest part, monosyllabic, like those
of Sanskrit and the Indo-European languages.
These roots are, in Irish, as well as Sanskit, al¬
ways (at least in their original or primitive condi¬
tion) of the nature of a verb. Also many substan¬
tives in Celtic (Gaelic and Kymric] are closely a¬
llied to Sanskrit roots. The system of derivation
and composition of words is analogous, and often
the same in both Celtic and Sanskrit — mostly by
prefixes and suffixes, simple or compound, and
which are often, in both forms of speech, identi¬
cal. A large number of Celtic compounds are
such as can be explained only by Sanskrit, and
must have existed already before these languages
branched off from the common parent-stock. The
whole system of grammatical forms in Celtic is
closely connected with Sanskrit [notwitstanding
some mutilations which have ocurred in the long
process of time). The anomalies in Celtic can
often, find their full explanation only through
Sanskrit, and also their elements can be derived.
in the last analysis, only from Sanskrit. In the
system of conjugation, the affinity between Irish
and Sanskrit becomes particularly apparent. The
power and facility of forming compounds is very
great in Irish, and may fairly be compared with
the Greek, German and Sanskrit. These com¬
pounds display the riches, elegance, and flexi¬
bility of the Irish language and it is especially
in poetical productions that we meet in Irish with
combinations of nouns, which come very near the
much admired Sanskrit compounds. Thus, to give
an instance or two out of many, night has some¬
times received the epithet glan-realt shoilseach,
a Buhuvrihi compound, meaning having stars with
pure, or bright, effulgence. Thus, of a certain
young man it may be said [to write it all in one
word, as is done in Sanskrit, and make the resem¬
blance more conspicuous in external appearance),
oighfheargruaighfhinshiodfhaindgualscaineogach
viz., a young man, whose beautiful, silken hair
falls, scattered, in ringlets down over his shoul¬
ders]. Such compounds have nothing analogous,
except in Sanskrit, not even in Greek, not, at least
to the same extent. It is also worthy of remark
that the other Celtic languages [here and there
Welsh excepted] possess nothing of this, compared
with Irish. As already stated, the whole phonet-
ic system of the Celtic group is intimately related
with that of Sanskrit. The consonantal arrange¬
ment corresponds accurately with the Sanskrit or
Zend, Greek or Latin; and Grimni's law is gen¬
erally speaking, not strictly applicable in Celtic.
Vowel changes remain within the limits of analogy.
Sauskrit euphonic laws have left an unmistakable
impress on the Celtic languages. The intimate
relation of the Celtic tongues to Sanskrit extends
in fact, to all the parts of grammar. The final
vowel of the oblique cases becomes, aften, affected
by attenuation, just as we find is the case in Zend
or Old Bactrian, where an i is introduced into the
preceding syllable, and this i of the attenuation
