Rit (Vísindafélag Íslendinga) - 01.06.1970, Side 204

202
with Hjelmslev—by means of a commutation test that gives
the same result no matter which of the above-mentioned
manifestations is chosen as the starting-point.
If, on the contrary, we make use of distinctive features,
we are all at sea if distinctive is not firmly based upon the
commutation test, and this the concept is obviously not in
generative phonology. I am by no means well versed in this
view of language, but as I understand it (and my knowledge
derives principally from my colleague Professor Eli Fischer-
Jorgensen’s treatment of the subject), unlike the linguistic
theories of Bloomfield and Hjelmslev, generative phonology
does not describe a certain corpus, a given material, but is
generative, predicting, is a plan or a set of rules for an infinite
number of examples. While Bloomfield’s linguistic theory is
monoplane, Hjelmslev’s glossematics is built on commutation,
which presupposes a plane of content and a plane of expres-
sion. In a certain sense, generative phonology also works with
two planes: the object is to describe the speaker’s competence
(his intuition), and this is put in contrast to performance, a
distinction which corresponds roughly to Saussure’s langue
and parole. What is implied in competence, however, is not
an inventory of words, but a set of rules, which can be used
creatively. The stress is put upon the rules of manifestation,
to use Hjelmslev’s terminology, even though a kind ofstructure
must be assumed to form the basis of competence. How this
structure is arrived at, we are not told. The rules of manifest-
ation are based on universal distinctive phonetic features.
However, these distinctive features must become ambiguous
when they are not fixed in a structure underlying the manifest-
ation. These distinctive features are, as far as I can see, some-
times purely phonetic differences, sometimes important, easily
distinguishable, phonetic differences (Otto Jespersen’s ‘speech
sound’), and, apparently, sometimes also linguistically
relevant differences. It is understandable that Moulton gets
into difficulties here, because distinctive in all cases means a
phonetic ‘either-or’, and this does not agree with Moulton’s