Íslenskt mál og almenn málfræði - 01.01.1980, Blaðsíða 209

On the Dental Accretion
207
that non-basic morphological variants tend to become similar to their
respective basic morphological variants, under favourable conditions.
(See section 1.3 for a similar discussion of Icelandic.) The next, and
last, step in the Old English development consisted, I think, either in
the gradual analogical spread of st from the monosyllabic clipped forms
(e.g. gœst) to the polysyllabic short forms (e.g. démes > démest), or in
the generalisation of the clipping of u to more-than-dissyllabic long
forms (e.g. to démestu). The analogous step in the Old High German
development consisted, I assume, either in the analogical spread of st
from the sg. pres. ind. (e.g. from tuost, salböst) to the remaining 2. p. sg.
forms (e.g. salbötös > salbötöst), or in the generalisation of the clipping
of u to non-present-indicative long forms (e.g. to salbötöstu). I cannot
choose among these altematives. At any rate, the process was aided,
as earlier, by the universal tendency of non-basic morphological vari-
ants to become similar to their respective basic morphological vari-
ants.17
3.
To sum up, it is necessary to distinguish between long and short 2. p.
sg. verbal forms in Icelandic: the long forms end in the suffixed personal
pronoun þú (e.g. krefurðu of krefja), the short forms do not (e.g.
krefur). Some monosyllabic short forms have developed a dental accre-
tion, e.g. lest of lesa, ferð of fara, fœrð of fá. This accretion arose in
the last stage of the development shown sub (3), q.v. In Stages 1 and 2
17 A marginal problem is the situation in Northumbrian, where the jt-forms
Were indeed used (and spread even into the preterite of the non-strong verbs), but
did not eventually prevail over the s-forms (Brunner 1962:177). I assume that this
must somehow be connected with the well known, although not well understood,
Northumbrian replacement of almost all present indicative endings with a new
ending ,r. If this assumption is correct, one would expect more .«-forms in the
Northumbrian preterites, where no comparable spread of the new s took place,
than in the corresponding presents. This prediction seems to be borne out: Sievers
(1951:298) mentions that in the two main Northumbrian texts, the Lindisfarne
Gospels and in the Durham Ritual, í/-forms occur more often in the preterite than
in the present. This fact has hitherto been unexplained.
I assume that Gothic did not develop the clipped íí-forms because it lacked the
long j/«-forms (the latter are not mentioned in any handbook on Gothic).