Íslenskt mál og almenn málfræði - 01.01.1984, Qupperneq 149

147
Model ofModern Icelandic Syllable Types
seems to have a tendency to devoice certain nuclei. Thus /r/ in vors:
[vojs]; but this ís not the case with all nuclei, since for example /l/
stays voiced in heils [heils]. These matters need further investigation,
and the question of their significance relative to the present analysis
will have to remain open for the time being.
An important question at this point is whether the same type of
account as suggested in Section 3 for the nucleus is appropriate for
the coda. In this brief presentation, the coda has been discussed in
terms of segments, that is, the units referred to are the traditional
holistic segments. It seems clear that the articulatory parameters are
the same as in the nucleus, apart from the one of vowel quality, but
it is possible, and indeed very likely, that the hierarchy of these para-
meters would be different. For one thing, it seems that voicing vs.
voicelessness is less central in the structure here than in the nucleus.
For example, codas that are voiced when followed by vowels may
be devoiced when occurring before pause: lána [lau:na] vs. lán [Iau:n],
vatna [vah^na] vs. vatn [vah^g], borða [þorða] vs. borð [þorþ].
The importance of the distinction between simple and complex codas
depends on what importance is to be assigned to the segment in this
context; I will leave open the question for example whether or in
what respect [s] of the coda st in skarst [sgajs$] should be seen as
phonologically equivalent to the [s] in bars [þajs] or that in bás [þau:s].
Similarly, I will leave open the question whether [n] in baun [þœy:n]
is equivalent to [n] in vatn(ið) [vah^n(lþ)] etc.
4. What is gained (and what is lost)? — Some afterthoughts
The foregoing can be seen as a preliminary experiment in applying
a sort of prosodic approach to a part of the phonology of Modern
Icelandic. I will end this expression with some general remarks as
to the possible gains or losses of approaching Modern Icelandic
phonology in this way.
The aim is to try and form a coherent model of the phonological
patterns of Icelandic. It looks as though a relatively complete account
of the phonology would not be overly complicated. This in itself must
be considered as a great asset. The model carries considerable infor-
of the nasal. This is in no way presupposed by the model, and a different analysis
of forms of this sort is possible within this framework.