Rit (Vísindafélag Íslendinga) - 01.06.1970, Page 466

464
ð, y] is a classic problem in Danish phonology. The matching
[p- -b, t- -d/t, k- -g, d- -ð, g- -y] suggested by the feature
definitions above reveals a skewness in the pattern since there
is no counterpart of [b-].
As stated above the consonants should be studied in their
occurrence in various types of strong and weak positions
rather than just initially and finally, at least if this be under-
stood as absolutely word-initial and word-final position. An
inspection of the behaviour of the entities in other environ-
ments reveals a very complex pattern.
Contrasts like [p] vs. [f] make it plausible that there is
a classificatory distinction between segments that are [-gon-
tinuant] or [ + interrupted] and segments that are [ + con-
tinuant] or [ —interrupted], according to one’s terminology.
One may now ask quite generally whether [d, g] of the strong
positions and [ð, y] of the weak positions have the same
phonemic status, i.e., the same definition in terms of distinctive
features.
A proof of the relatedness of the two series is fumished by
alternations in foreign words with and without stressed suf-
fixes, cp. [filolo'gi:’], [meto'dig], with [g, d] in strong position,
vs. [filo'lofy, filo'lofyisg], [mc'toiðo, me'toiðisg], with [y, ð]
in weak position.3 (Note that there is no trace of juncture
before the last vowel of Jilologi, metodik.) This does not, how-
ever, prove that all occurrences of voiced fricatives must be
derived from underlying stops.
In order to round off the picture of the relationships
between initial and final, or more generally, strong and weak,
positions one must consider also the semivowels [i, u]. In an
autonomous phonemic analysis the former must obviously be
paired with the [j] found in strong positions (phonetically it
is not very easy to distinguish [j] and [i] in Danish at all),
but a study of morphological alternation (variation) imme-
3Pronunciations with the stop alternants generalized to weak position also
occur (Andersen 1954:344).