Íslenskt mál og almenn málfræði - 01.01.1984, Blaðsíða 152

150 Kristján Arnason
forms are spelled with double letters and the English ones not is less
relevant than the fact that they are spelled with the symbols p, t,
k. The fact that the orthography uses the symbols p, t, k and not
b, d, g both in Icelandic and English would then have to be seen
as the crucial factor. Also, English forms like block (bloc) that are
pronounced [þbhg], do not provide spelling geminates to motivate
the assignment to the checked open category.
The facts just mentioned are interpreted by Thráinsson (1978) as
showing that a rule of preaspiration is active in Modern Icelandic.
This rule is, in his analysis, based on an underlying distinction be-
tween tense and lax consonants, exemplified respectively by vakna
[vahgna] and vagna [vagna]. In the present framework the question
is what category the speakers decide to assign the forms to and why
they choose this or that category. Indeed, in the more traditionally
generative framework assumed by Thráinsson, the question is basic-
ally the same. A distinction between forms with preaspiration
([vahgna]) and forms without it ([vagna]) is a classificatory one. If
preaspiration is said to occur on underlying stops, as Thráinsson as-
sumes (a similar analysis is suggested in Árnason 1977a), an under-
lying classification is assumed between forms with ‘tense’ stops and
forms with ‘lax’ stops. What the speaker has to do is to ‘decide’ which
category to assign the borrowed form to, the ‘tense’ one (the one
‘to be preaspirated’) or the ‘lax’ one (the one ‘not to be preaspirat-
ed’).
A basic question in this context is the question of the status of
segments in the phonological structure of Icelandic. In the foregoing,
the emphasis has been on the prosodic aspect, and it is suggested
that approaching Icelandic phonology in this way may make it possible
to make intuitively plausible statements about the phonological struc-
ture of Icelandic. But it is equally true that many of the characteristics
of Icelandic phonology may be accounted for in terms of segments
rather than larger units or prosodies with suprasegmental (or ‘infra-
segmental’) scope.
A phenomenon indicating the significance of the segment is for ex-
ample metathesis of /ðr/ to [rð] in forms like blarða for blaðra and
conversely of /rð/ to [ðr] in form like snuðrulaust for snurðulaust. In
the analysis suggested above, forms like blaðra are parsed s[ o[]?l]o
r[ N[að]N. c[r]c ]r]s a, but blarða would be parsed s[ o[fe>]o r[ N[ar]N