On the accusative with participle: Typological and cognitive aspects

This paper deals with the the complements of the verbs of visual and auditory perception in Old Church Slavonic: Accusative with participle (AP) and clause. The two types of complements are semantically differentiated by evidentiality: AP serves for the firsthand evidentiality and the clause for the non-firsthand evidentiality. Since AP is attested in Old Russian, Old Czech as well in some other old Slavonic languages, it is evident that it was an indigenous Slavic construction. It belongs to the Indo-European syntactic inheritance - the appositive double accusative. Since in early Indo-European the accusative was a general adverbial case, it expressed both types of evidentiality. With the typological drift of Indo-European and its daughter languages toward a nominative language type, which meant the development of syntactic transitivity, the AP was reanalyzed as an object, but only in the cases of the firsthand evidentiality (where the subject has control over the information). For non-firsthand evidentiality, another strategy, inherited also from the proto-language, was used: a sentence with delimitative connective(s). This process was finished by the end of Proto-Slavonic, as testified by Old Church Slavonic. In the process of the further strengthening of transitivity, which gave a prominent role to the predicate as the centripetal core of the sentence, the other predicative center - the active participle - had to be removed, while the passive participle was reanalyzed as an adjective. This led to the loss of the AP in the early history of Slavic languages and the development of hypotactic structures. It was a long process, marked by the competition of different particles and deictic forms which were on the way to be grammaticalized into conjunctions. It ended with the formation of the two types of conjunctions for the two types of evidentiality, e.g. jak - e in Czech, da - e in Bulgarian, kako - da in Serbian, kak - to in Russian etc. This shows not only the importance of evidentiality in a diachronic perspective but also that its formalization is based on the language type.