Rit (Vísindafélag Íslendinga) - 01.06.1970, Page 516

5J4
predictable, to a certain extent, only on the basis of a minute
contrastive analysis, which cannot wholly exclude extra-
linguistic, above all psychological, factors. Such an external
factor, for instance, makes a Pole put a wrong stress in Russian
words where the stress falls on the same syllable as in Polish,
because he feels that Russian words should be stressed dif-
ferently.
The premise of interference being the possibility of free
choice, enabling a bilingual to transfer automatically the
stronger motor habits (plane of signals) and (mostly
semantic) associations (plane of information) from one system
to another, we must, as already pointed out, reckon with a
greater amount of interference phenomena in the less deter-
mined subsystems of language P the more determined the given
subsystem, the smaller the probability of interference. In this
connection it should be borne in mind that habit formation
in language concerns íirst of all the closed systems. In the
forming of utterances out of known units the speaker is free,
provided that he observes the formal rules of coding. The
only restriction here, for the native speaker, is provided by
the conventionalized syntagms, which, however, can be freely
transferred from one language to another owing to the open-
ness of the lexical system.
The general tendency of interference can be shown by the
diagram in Figure i (in which the shaded area shows the ratio
of interference impediment).
This general tendency, as well as the rules behind it, ought
to be kept in mind when predicting the behavior of bilinguals.
The general rule of interference may also prove of vital
importance for the preparation of material for foreign language
teaching, as well as for dialect study and geographical
linguistics (Zabrocki 1961:17, 1956:170).
xWe should, of course, not overlook the fact that the degree of closeness of the
given subsystem is relative. So, for instance, the syntactical system of English is
more determined than that of Latin. Nevertheless, it is in both cases less deter-
mined than the subsystem of functional morphemes in the respective languages.