Íslenskt mál og almenn málfræði - 01.01.1992, Page 121

Phonological Variation in 20th Century Icelandic
119
(19) Accommodation between dialects will tend to go the easy
way. The accommodator will not adopt new phonological
contrasts and not adopt a more complex or less general phono-
logical rule than the one (s)he already has. But (s)he may
eliminate phonological contrasts and simplify or generalize
rules.
our Icelandic context, this predicts the following direction of accom-
modation:
(20) a “Hard dialect” -> “soft dialect” because it involves a sim-
ple rule: Deaspirate all stops that follow long vowels. The
reverse rule (“Aspirate all stops... ”) will not work because
of (loan)words like túba ‘tuba’, stúdent ‘student’, sagó ‘sago
pearls’, etc., that do not have aspirated stops in anybody’s
dialect.16
b “Voiced dialect” -> “voiceless dialect” since it involves a
simple rule: Devoice all sonorants before /p,t,k/. The reverse
rule (“Make all sonorants voiced... ”) will not workbecause/r/
is voiceless before /p,t,k/ in everybody’s dialect and because
/1/ is voiceless before /t/ (in certain contexts) in everybody’s
dialect (cf. Þráinsson 1980; Jónsson 1982).
c [xv]-dialect-> [khv]-dialectbecauseitinvolvesasimplerule:
Tum all instances of initial [xv] into [khv]. Again, the reverse
The “rules” talked about here are intended as rather informal descriptions of the
changes involved in going from one dialect to another. Thus we are assuming that
somebody who speaks the “hard” dialect adopts the “soft” dialect, this change in
's/her grammar can be informally described by a rule that says “deaspirate all stops
jhat follow a long vowel”, whereas a change in the other direction (from “soft” to
hard”) cannot be described by a general aspiration rule since some of the stops that
°Uow long vowels should not be aspirated, as illustrated by the examples in the text.
^onsequently, speakers accommodating in that direction tend to make mistakes like
b u:pha] for [thu:pa] for túba ‘tuba’, etc., whereas those accommodating from
ard to soft do not have to worry about any such mistakes. But the question of how
to describe the consonant systems of these two dialects is not a simple one and we
are largely sidestepping the important issues here. This does not affect the present
^gumentation, however.