Íslenskt mál og almenn málfræði - 01.01.1999, Side 179

177
Flugur, smágreinar og umræðuefni
would be rather easily perceived because of the acoustic cues given by
formant transitions from the vowel to the velar. In the second set of
examples the velar would be adjacent to a nasal after metathesis, gi-
V)ug intermediate stages like /dankst/, /enkst/, /ynksti/, respectively.
As shown above, it does not vanish without a trace in this position
since it triggers assimilation of the nasal, giving a velar nasal as indi-
cated. In the final set of examples, on the other hand, .vk-metathesis
would put the fkJ next to a liquid and a voiceless one at that (since
s°norants are devoiced before /p,t,k,s/ in Faroese). That would give
(hypothetical) intermediate stages like /norkst/, /írkst/, /dplkst/. It
would seem that the fk/ would be just as hard to perceive in that posi-
hon (formant transitions from a voiceless liquid would not be very
clear!) so the velar would be deleted in that position too (or metathesis
would not apply). In these examples, no velar is perceived nor does it
leave any kind of trace as no assimilation between the liquid and the
Velar takes place.5
We see, then, that those that believe in some sort of functional ex-
Planations of phonological processes like metathesis could say that it
aPplies in Vskt- and N.vkr-contexts “in order for the velar to be pro-
Perly perceived” but it does not apply in Likr-sequences since the ve-
^ar will not be perceived there anyway. Altematively, we can say that
11 applies automatically in all these contexts but the velar is then deleted
lwithout a trace) after the liquid. Under the first account, we would
Pfesumably also want to say that the metathesis does not apply in
bisyllabic words likc fðroyskt ‘Faroese’, ymiskt ‘diverse’ since it has
n° lunctional (or perceptual) role (the velar will vanish without a trace
ar*yway). Under the second account, we could say that it applies in
ihese examples too but the velar is subsequently deleted. The outcome
ls identical.
As is well known, assimilations of liquids to stops are much less com-
ai°n than assimilations of nasals to stops. The reason could be that the nasals
ave more features in common with the stops to begin with, cf. e.g. the dis-
eussion in Itð, Mester and Padgett 1995:600 of prohibition of linking
etween segments that are too dissimilar. Hence it is not clear that we would
exPect the liquid to show assimilation to the velar even if metathesis applied
ln these examples.