Íslenskt mál og almenn málfræði - 01.01.1999, Blaðsíða 175

173
Flugur, smágreinar og umræðuefni
norn.sg.fem. donsk [donsk]. When neuter/-t/ is added, the underlying
/dansk+t/ changes through an intermediate step */dankst/ to [daqst]
where the cluster combination /nk/ has undergone assimilation (and
^evoicing) to [q].
The metathesis in /skt/ sequences is discussed by Rischel
(1961:XXXIV, 1972:490). His examples include assimilation exam-
Ples like danskt [daqst] ‘Danish’.2 Rischel also points out (1972:490)
Ihat this metathesis “does not apply to sk preceded by r or in un-
stressed syllables”. He then gives examples like the neuter forms
norskt ‘Norwegian’ and ymiskt ‘diverse’ in the phonological transcrip-
twns /norsk+t/ and /imisk+t/, respectively, and about examples of this
s°rt he says (1972:490); “In these environments the velar stop is dele-
ted”. In his earlier paper (1961:XXXIV) his examples include neuter
forms like írskt [ust] ‘Irish’, eingilskt [ojicþllst] ‘English’ and fðroyskt
[fpaist] ‘Faroese’. In connection with these Rischel points out the
deletion of the velar and attributes it to the nature of the consonant
clusters (that they include a liquid before the /skt/ cluster) and/or lack
°f stress.
While Rischel’s discussion is very brief in both cases, it is very
much to the point as he manages to point out the most important types
°f examples. But the situation is not quite as straightforward as he
makes it out to be. First, examples like eingilskt ‘English’ are in fact
arnbiguous in the sense that the velar could be deleted from the /lskt/
duster because of the lack of stress or because of the presence of the
Ijquid /1/. But if we look at examples like dfilskur [dœlskuj] ‘apathe-
tlc , which show up in nom./acc.sg.neut. with no [k], that is d0lskt
Ldœlst], we see that the liquid /1/ does in fact seem play a role in
c°nditioning deletion of the /k/ (or blocking the metathesis). In
e*arnples like norskt and írskt the role of a liquid is not as simple since
ese are not pronounced with a clear liquid (such as [j] or [j]) but
rather with a retroflex [s], which is in fact the typical outcome of /rs/
■ lu his 1961 paper Rischel transcribes danskt with a [k], [daij.k.st] (the [.]
lcates half-long consonants in Rischel’s transcription) but later (1972) he
s me velar stop in parentheses (and does not indicate the voicelessness of
e nasal); [darj(k)st]. Here it it maintained that forms of this sort only con-
ln a velar nasal and no [k].