Language interference as a linguistic phenomenon

Abstract

Language interference under conditions of globalization and educational multilingualism is viewed as a dynamic cognitive-linguistic process that generates persistent deviations from the norms of the target language and leads to error fossilization when neglected in teaching methodology. The aim of the study is to describe the mechanism of interference, to trace quantitatively its evolution at different stages of learning, and to identify zones most resistant to pedagogical correction. The material comprises written and spoken language products of 843 students from three humanities universities who studied English as their first foreign language and German/French as their second; a corpus of over 2.5 million characters and more than 350 hours of audio recordings was analyzed in a four-year longitudinal experiment. Errors were collected by continuous sampling, tagged in a database, and classified by linguistic level, type (interlingual/intralingual), and communicative effect; contrastive analysis, longitudinal measurements, correlation-regression procedures in R, calculation of the interference density coefficient, factor analysis, and Pearson’s χ² test were applied. At the phonetic level, a nonlinear reduction of segmental deviations was revealed (phoneme substitution 18.432%→4.127%; palatalization 31.228%→8.745%) alongside high persistence of prosody (intonation K=0.657; rhythm K=0.716), with segmental errors being eliminated approximately 2.3 times faster. At the grammatical level, articles (145.67 errors per 10,000 words) and prepositional government (134.21) proved most problematic, reflecting difficulties in conceptualizing definiteness and idiomatic collocability. In lexis, false friends of the translator yield faster but erroneous decisions (1845 ms; Err 0.458), whereas non-equivalent vocabulary requires more time with fewer errors (2678 ms; Err 0.215). The integral index demonstrated a stepwise decrease in interference with plateaus of interlanguage restructuring and a shift in dominance: at the C1 level, the interlingual share drops to 23.4%, while the intralingual share increases to 45.7%. The findings substantiate cyclical learning trajectories, early work with prosody, and the development of annotated error corpora and adaptive digital diagnostic tools aimed at preventing covert interference and individualizing correction. Of practical significance is that accounting for learners’ psychotypes and cognitive strategies improves the accuracy of error prediction; therefore, the authors recommend diagnostic questionnaires, individual profiles of interference risk, and regular feedback that prevents avoidance of complex constructions and sustains motivation at advanced stages of learning.