Word formation in the earliest stages of L2 Polish

Abstract

This paper examines word formation strategies in initial SLA, with particular regard to implicit processing of the distributional properties of the input. 163 learners with various L1s and no experience of the target language took a 14-hour L2 Polish course under controlled input conditions. In an oral production task, learners were asked to describe properties of human referents who had never appeared in the input by stating their nationality or profession. In their output, learners most often referred to the target referents by attaching a -k(-) sound cluster to a lexical morpheme borrowed from a known language. Quantitative analysis shows that indeed, within the input considered, the same -k(-) cluster is characteristic of most words referring to human entities. The study concludes that learners can analyse the morphological structure of target words even after minimal exposure to the input, identifying at first the derivational formants characterised by the strongest association to the intended meaning.