Íslenskt mál og almenn málfræði - 01.01.1980, Blaðsíða 203

On the Dental Accretion
201
are conjugated in the active voice singular of both tenses and moods,
except in the indicative present of the non-preterite-present verbs
(roughly speaking), in such a way that the 2. p. sg. is longer than the
first and third persons. E.g. the indicative preterite of the strong verb
lesa is, 1. las, 2. last, 3. las; of the weak verb kalla, 1. kallaði, 2. kall-
aðir, 3. kallaði. Etc. The clipped forms such as lest, ferð follow this
prevailing model: the old paradigm was, 1. les, 2. les, 3. les, the new,
1. les, 2. lest, 3. les.—When lest, ferð had prevailed in the spoken
language, the written language also accepted them, as unofficial variants
to begin with, and only in the twentieth century as the normal official
forms (Arpi 1904).
To the best of my knowledge, only four modem Icelandic mono-
syllabic short second persons singular—not including the short impera-
tive—remained without a dental ending after lest, ferð had prevailed in
the spoken and written language: skín of skína, hvín of hvína, hrín of
hrína, and vex of vaxa. I cannot offer any explanation of this state of
affairs. The situation is all the more perplexing because the same verbs
can form the clipped imperative, according to my informants: skínd,
etc. (e.g. skínd þú, blessuð sóliri).
An interesting question is why the clipping of u has not spread to the
2. p. sg. verbal forms ending in -urðu, so that -urð would result, e.g. a
clipped krefurð beside the long form krefurðu of krefja. (Words in urð
are commonplace in Icelandic, cf. megurð, fegurð, lipurð, so that the
prohibition cannot be due to restrictions on the pronunciation.) The
following explanation can be suggested. In all those pairs of long and
corresponding short verbal forms that originally participated in the
formulation of the clipping rule, the long form was dissyllabic and the
short form monosyllabic. In contradistinction, the long forms in -urðu
are at least trisyllabic. The implication is that the statement concerning
the number of syllables is a part of the clipping rale.10
The oldest example of the lest type, from the former half of the
eighteenth century (see § 1.1 above), shows that the clipping rule must
have begun to spread outside imperatives around 1700 at the latest. I
have nothing to say about why the clipped forms of the lest type began
to appear precisely in the eighteenth century. The oldest example of the
10 This suggests that morpho(no)logical rules can be sensitive to the number of
syllables in the words on the basis of which such rules are formulated.