Rit (Vísindafélag Íslendinga) - 01.06.1970, Side 95

93

by the author himself, ibid.). In particular, the definition of
environment as meaning ‘one phoneme to the left and right,’
though without doubt operationally convenient for the elec-
tronic computer, involves a simplification that is bound to
render the value or usefulness of the entire procedure ques-
tionable indeed.
There is one additional factor, especially, which would seem
to be of significance in assessing the value of the notion of
functional load, but of which no account is taken in the above
procedure. This is the relationship between the phonological
opposition concerned and the morphological system, or in
other words, the morphological function assumed by the pho-
nological opposition or its status on the morphophonemic
level. As an example we may take the mergers of short and
longj with i in Icelandic. The merger oi\ongý+í>í appears
to form an integral part of the chain of vowel mergers in
Icelandic, that is, to be simply an individual manifestation
°f the structural principle governing these mergers (for which,
see § 4.4 below). The merger of short y + i > i, on the other
hand, cannot be integrated into the process on the same basis;
one would expect a merger ofy+u>u (H.Benediktsson 1959:
299)- It was therefore suggested that this merger was pre-
vented by the high functional load of the oppositiony: u (ibid.
3ri). However, the above calculation indicates that L(y,u)
was much lower than L(y,i), which seems to speak strongly,
ln this case, against any significance of the functional load in
diachronic phonology (King 1967 (a) 1840). But the two sig-
mficant points that must not be forgotten are: (a) there are
traces of the merger y+u>u, e.g. Mod. Icel. spurja ‘to ask’,
kjussa ‘to kiss’, kjurr ‘quiet’ <spyrja, kyssa, kyrr (while there are
no such traces in the case otý+í> í)\ and (b) the opposition
y'-u took part in morphophonemic alternations such as pret.
sukí-flytti, flyttum, ‘he, we carried',fyndum ‘we found’, etc., vs.
md.77utti,fiuttum,fundum, etc., or comp.yngri ‘younger’ vs. pos.
nngri (dat. sg. fem.), while the opposition y: i (or ý:í) took
part in no such alternation (H. Benediktsson 1959:299-300,