The article demonstrates the importance of subjectification processes in shaping the metatextual layer of language. Using three lexical units (prawda ‘true, right’, pewnie ‘sure, certainly’ and szalenie ‘extremely, madly’) as examples, the author shows how the enrichment of their semantic structure with a subjective component led to the emergence of new propositional functions and ultimately to the establishment of a new meaning. The study is conducted diachronically, drawing on the oldest attestations of the lexemes in question. Based on a contextual analysis, the moment the meanings with a subjective component appeared is identified. The results unequivocally demonstrate that subjectification has its origin in the pragmatic domain, while its consequences are visible on the semantic level. The language material comes from both lexicographic sources and corpora. The analysis shows that subjectification is correlated with formal changes including loss of inflectional endings, loss of morphological properties, recategorization, and syntactic isolation. The paper provides evidence for the need for in-depth comparative diachronic research on subjectification.

In Polish, passive potential adjectives are productively formed by means of attaching the suffix -alny to transitive verbs (Szymanek 2010). They have been shown to project the external argument of their verbs, as well as being able to co-occur with agentive przez-phrases and instrumental phrases (Bloch-Trojnar 2019). Hence, under a syntactic approach to word formation such as Distributed Morphology, they are derived via outer affixation, with their structure containing the vP and VoiceP heads. A small subset of Polish passive potential adjectives are derived with affixes other than -alny. These include czytelny ‘legible, readable, understandable’, strawny ‘digestible’ and zrozumiały ‘understandable, comprehensible’. In this paper, it is demonstrated that while these adjectives behave similarly to -alny adjectives in terms of licensing Voice-related modifiers, they are excluded from a wide range of verbal contexts available to regularly derived passive potential adjectives. As such, czytelny, strawny and zrozumiały offer evidence for the claim that the layer that introduces event implications is distinct from the verbal head that triggers spell-out. Specifically, adjectives such as czytelny, strawny and zrozumiały can be argued to contain the little v head, but not the cyclic vP projection, which is in line with the architecture of grammar as proposed by Embick (2010).