AN GAOḊAL
749
This is no mere theory, for the Wuerzburg MSS.,
quoted by Zuess, gives us the genitive cindas per-
sine, (6b) quatitas personae, and the St. Gall MS.
cen torant persine, without signification of person,
also the dative h persin with an alternative form
in Windisch (ib. ut sup) i persuind, which bridges
over the chasm to modern pearsain.
That the n belongs to the root may be also prov¬
ed from the other words of this declension, thus cu
gen. con, is connected with the Greek kuon, Latin
canis, Sanskrit root, choan uille with G. olene,
Lat. ulna, Gothic aleina; meanma, Old Irish, men-
me with Greek root men, man reduplicated in mem-
ona, Skt, man, Latin reduplicated, memini,* all
showing clearly that their present form is the re-
sult of the loss of an original n. The
words Sacsain and leiġeann, nomina-
tives themselves, point to the fact that
n is not a case ending; and the two
nominatives, cara and carad, even in
Modern Irish, prove the same for Dr.
Joyce's d-nouns.
The examination, then, of this fifth declension
shows that it possesses no case-terminations, prop-
erly so called, distinct from the second declension.
In fact it has no right to be called a separate de-
clension than the ordo ordinis class in Latin has
to be distinguished from princeps principis. It
may, indeed, have claims to be considered as a
separate class under another declension, but cer-
tainly the grammarians who refuse to
take caṫair, caṫaraċ, out of the third
declension, are by no means consistent
in emancipating pearsa from the sec-
ond. Father Bourke's argument §55 does not bear
on the question at issue. Number proves nothing
in declension, and if there were only two nouns
which possessed distinct case endings, they would
be entitled in strict philosophy to be
placed in a separate declension. Pear-
sa, however, as we have shown, does
not possess such case endings, and therefore (I
think we are justified in saying of it and of the
whole class, that they are in a very unsatisfactory
condition.
2. From the fifth we pass to the second declension.
Here we have two classes of nouns, broad and slen-
der. Taking peist as an example of
the slender class and comparing
with cleas, a noun of the third declen-
sion, we get ;—
Sing.
N. peist
G. peist-e
D. peist
cleas.
cleas-a.
cleas.
Pl.
N. peist-e
G. peist
D. peistiḃ
cleasa.
cleas.
cleasaiḃ.
Leaving out the ḃ of the dative plu-
* Grundzuege der Greichischen Etymologie von
Georg Curtins (Windisch’s Ed.) cu p.158, uille p.
377, menme p. 312.
ral which is common to all the declen-
sions, we find that the only termina-
tions used, are a for the broad third de-
clension noun, e for the slender second
declension noun. A priori, or applying
the rule caol le caol, we would suspect
the identity of the terminations, for
with such a word as peist a should be-
come e. Following the principle of au-
thority, we are led to the same resuit. Thus Dr.
Joyce says of the second declension, “When the
characteristic vowel is broad, the nominative plu-
ral is formed by adding a; when the
characteristic vowel is slender by add-
ing e." Giving this rule as he does
give it, for one declension, he recog-
nizes that a and e arise from the one
cause, and thus that they are substantially the
same termination. Canon Bourke recognizes this
still more fully and in a broader sense, for when
giving rules for the formation of the plural in what
he calls imparasyllabic nouns, he says, it is form-
ed “by annexing e or a to the final syl-
lable — e when the preceding vowel is
slender, a when broad.” Hence, as
far again as a principle of declension is concerned,
these slender feminines have as much right to be-
long to the third as to the second declension. The
broad feminines, however, though in
the plural they are perfectly assimilat-
ed to the cleas-type, give in the sing-
ular some reason to suspect a dividing line — but I
shall return to this again.
3. The third dec., like the second, has also slender
nouns, but they differ from the slender nouns of
the second in two respects, 1, the gen-
itive is in broad a not e as we might
expect, and 2, contrary to analogy the
plural ends in a consonant iḋ. These
variations, however, are more apparent than real,
for in the spoken language, the sounds
of the terminations, a. e, iḋ, hardly dif-
fer from one another, and as far as the
written language is concerned, the
grammarians who class cleas and
slanuiġṫeoir in the one declension, ap-
pear to admit that iḋ is only variant
of a. Take now the formula obtain¬
ed from this class, viz.
Sing.
N.
--- --
Pl.
N.
-----iḋ
G.
-----a
G.
—
D.
—
D.
—iḃ.
and apply it to the nouns of the fourth declension
thus —
Sing.
N.
tiġearna,
Pl.
N.
tiġearna-iḋ.
G.
tiġearna-a,
G.
tiġearna,
D.
tiḋearna,
D.
tiġearnaiḃ.
