Íslenskt mál og almenn málfræði - 01.01.1982, Side 308

306
Ritdómar
is, regrettably, not much of an improvement. It dcpcnds crucially upon thc notion of
syllable boundary and amounts to a claim that vowels are lengthened when no more
than one consonant appears before syllable boundary, i. e.:
(I) V —> [ + long]/___C'$
Of crucial importance is, of course, the assignmcnt of $ and it is here that Arnason
appears most arbitrary. Hc assumes that the coda of a stressed syllable is assigned
all following consonants with the exception of those segmental clusters which induce
lengthening, where the syllable boundary separates them. Thus bidja ‘ask’, aldrei ‘ne-
ver’, vinstri ‘left’ should have the $ before the final vocalic nuclcus while in the case
of nepja ‘cold weather’, titra ‘shiver’ the $ would split the medial cluster. It is hardly
surprising that the length rule will apply correctly to words of this type. What Árnason
has really achieved is to translate a traditional statemcnt into a new one without any
change of substance. Thus the traditional claim that vowels are short if two or more
consonants follow translates into Árnason’s instruction to assign all consonants follow-
ing a stresscd vowel to the coda of that syllable; the traditional specification of the
cxceptional clusters translates into Árnason's violation of or departure from the above
syllabification principle. I should add here that Árnason attempts to provide some addi-
tional evidence in favour of the second clause of his syllabificaton rule (see esp. pp.
33—42) but unfortunately, I find the evidence vague and tenuous.
The basic question is, of course, the notion of the syllable and its role in phonology.
A number of authors have tried to use thc syllable for phonological needs (most recently
perhaps thc Bcll & Hooper (1978) volume bears testimony to it) and Árnason is thus
in good company. Just like others of the samc party, he is unable to solve the dis-
crcpancy between what might possibly be the phonetic syllable and what functions as
a syllable for phonological needs. Some way of reconciling the two concepts must be
found; in other words, the phonological syllable must be related in some way to the
phonetic syllable for othcrwise the syllable in phonology will remain an arbitrary marker
justified ad hoc. There is little in thc phonology of Icelandic, as discussed by Árnason,
to make us believe that the syllable boundary is superior to something we might wish
to call, for example, the length rule boundary, i. e. to something openly arbitrary.
Statements such as ,,[f]rom the point of view of the simplicity of the length rule, a
syllabification like hest-ur, biðj-a and Es-ja seems to be optimal“ (p. 34) and ,,[t]he
main advantage of this syllabification (if it can be called syllabification at all) is that
the environment of the length rule, if defined in this way, will be exactly the same
in monosyllables and polysyllables“ (p. 35) clearly indicate that Árnason is aware of
the weakness of his position. He is, however, consistcnt in accepting only formal argu-
ments. This is questionable since the syllable seems to be a real unit to the speaker
and I am surprised, despite Árnason’s avowedly formal approach, that he does not
invoke his native speaker’s intuitive judgemcnts. Its rejection leads to dubious rcsults.
Thus Árnason is forccd to syllabify the compound þjóðvísa ‘folk song’ as þjóðvSísa
in order to get a short diphthong in the first element of the compound (pp. 49-50).
This is, of course, consistent but consistency is hardly a clinchcr. The whole point of
introducing the syllable boundary and reinterpreting the lcngth rule in its terms was
to rationalise thc traditional and, admittedly, pedestrian formulation. Boundary assign-