Íslenskt mál og almenn málfræði - 01.01.1979, Blaðsíða 249

Pronunciation of modern lcélandic rövl(a) and slafneskur 229
‘Slavic’ and in the noun slafneska ‘Slavic language, culture, etc.’. One
way of explaining this state of affairs would be to claim that the suffix
-nesk- has the same status as suffixes such as -laus- and -leg-, i.e. that
it has the properties of compound-word constituents. (The /v/ which
ends one compound-word constituent, and the /n/ which begins the
next one, do not undergo rule (5), cf. af-neita [-vn-] ‘deny’.) The follow-
ing speaks against this proposal:
(1) The lexemes latneskur ‘Latin’ and gotneskur ‘Gothic’ are pro-
nounced with [-h(ln-], i.e. with ‘preaspiration’, which would not obtain
if t and n were not within the same simplex. True enough, vit-laus
[-h(ll-] ‘crazy’ and a few other examples show that compound words
can sometimes come to be treated as simplexes, but such cases are rare
and limited to words of frequent usage in the spoken language, which
latneskur and gotneskur are not.
(2) The suffix -nesk- came into being through the re-analysis of
lexemes such as gotneskur, in which n originally pertained to the stem
(notice that Got-i ‘Goth’ is an n-stem noun), as got-nesk-ur under the
influence of the root got- in Got-i (Alexander Jóhannesson 1927:29).
The reanalysis took place in pre-literary Norse (words containing
-nesk-, e.g. jarð-nesk-ur ‘earthly’, occur in the oldest Icelandic and
Norwegian manuscripts), when there was not yet any ‘pre-aspiration’
in gotneskur.2 The suffix -esk- had never been a compound-word con-
stituent, therefore it is not likely that -nesk- became one.
2 The reanalysis may also have been prompted by the tendency sometimes
observed for the perception of a word to become facilitated through the morpheme
boundary moving in order to coincide with a syllable boundary: got$n+esk-~>
got$ + nesk- ($ = syllable boundary).
An example of a similar change might be the neuter noun hunang [-aur|g]
‘honey’ when pronounced trisyllabically as huna-ung (as it sometimes is, Magnús
Pétursson 1978:42). Here a syllable boundary was inserted between the two parts
of the diphthong [au], presumably because its [u] had been analysed as belonging,
together with the following ng, to a quasi-suffix -ung. (Such a suffix does not occur
in Icelandic neuter nouns,' but its existence may have been presupposed by the
speakers on the analogy of -ung and -ung-a in feminine nouns, -ung-ur and -ung-i
in masculine ones.) In short, first a morpheme boundary was introduced in the
middle of the diphthong [au], then also a syllable boundary. The process is typo-
logically interesting because it involves a diphthong, and as such it is relatively
seldom observed.