This article presents a theoretical account of the morphologicalphenomenon known as parasynthesis in compounding, i.e. a word formation process consisting in the merger of two lexical stems (forming a non-attested compound) with a derivational affix. The present analysis aims at showing that a constructionist account, recently developed within the Construction Morphology framework to explain the relevant phenomenon in some Germanicand Romance languages, cannot be successfully applied to a crucial set of Slavic data. The thesis defended here, although confined to the Slavic domain and especially suggested by some challenging cases of compounding, is that a configurational analysis of these compound-affixed forms, implyinga severe mapping between the morpho-syntactic and semantic structure is not only able to explain the Slavic data, but also improves our comprehension of synthetic compounding in Slavic and other I.E. languages. The implementation of the analysis shares some of the theoretical assumptions and formalsolutions of the morphological model outlined by Ackema and Neeleman (2004).