Criteria for determining the degree of ‘danger’ associated with false friends

False friends (also known as ‘deceptive cognates’, ‘treacherous twins’, etc.) are a source of lexicosemantic interference. They can be defined as lexemes in different languages with formal congruency (i.e. similarity or regular correspondence) that do not provide the best semantic equivalent for each other, e.g. R. stol ‘table’ vs. Bg. stol ‘chair’. While some false friends cause a lot of trouble for almost all bilinguals, other pairs are hardly ever confused. What determines how ‘dangerous’ a pair of false friends is? The degree of ‘danger’ can be analyzed on the basis of two factors: the probability of confusion and the seriousness of the potential consequences. This paper shows that the former factor is much more important than the latter. Among the various items discussed, the risk of confusion is increased especially by an overlap of the aggregate denotations of the words in question and by a semantic connection (proximity or antonymity) between their individual definitions. With criteria like these at hand, the ‘danger’ of any pair of false friends can be estimated, and certain particularly nasty pairs can be singled out in order to compose effective ‘false friends’ exercises for learners or to mark them in bilingual dictionaries.

Translated from “Kriterii dlja opredelenija stepeni ‘opasnosti’ psevdo-analogonimov (‘ložnyx druzej’)”. Submitted in 2000 for publication in a Bulgarian volume that never saw the light of day and self-published online in 2001 (DOI: 10.18716/bun/krit).