Íslenskt mál og almenn málfræði - 01.01.1979, Qupperneq 251

Pronunciation of modern Icelandic rövl(a) and slafneskur 231
lisation rules operate across a derivational morpheme boundary. The
following altemative explanations come to mind:
(1) I claim above, by implication, that the operation generating the
‘preaspiration’ of gotneskur cannot be a phonological mle. However,
the insertion of the ‘preaspiration’ can be a part of the word-formation
mle that creates gotneskur. For such examples only the mle might be,
add -nesk- and insert the ‘preaspiration’ before the root-final plosive.
(2) It can be claimed that the ‘preaspiration’ of gotneskur came into
being when the preaspiration mle was not yet a neutralisation mle, and
that the form with ‘preaspiration’ was then passed on from generation
to generation of speakers in toto. (The feeling that Got-i ‘Goth’ and
gotneskur are related stems from the semantic and phonetic similarity
between the two words, and this similarity need not be formalised in any
special way, say, by the embedding of got- (of Got-i) into the under-
lying representation of gotneskur.) Incidentally, I believe that this is the
correct explanation for gotneskur.
(3) It could be that gotneskur were deliberately created from the
elements got- and -nesk- in response to a felt need for such a lexeme,
by linguistically minded people who know that tn is normally preceded
by ‘preaspiration’ in the pronunciation. The pronunciation with the
‘preaspiration’ of the new lexeme would then be passed from one
generation to the next in toto.
Such hypotheses as stated in (l)-(3) above are in principle verifiable,
for it can be ascertained, in principle, when and how a word came into
being.
(4) Specifically for the preaspiration mle, one could have recourse
to the distinction (alluded to at the beginning of the present paper) be-
tween automatic and non-automatic mles, and say, following Höskuldur
Thráinsson 1978, that the preaspiration mle is automatic, and thus
exempt from Kiparsky.’s universal. This too would account for the
‘preaspiration’ of gotneskur.
5. A modification, equal to mine in its effects, of Kiparsky’s universal
has been proposed independently by Harry van der Hulst (personal
communication, August 1979), see Hulst (forthcoming).