Lingvisticheskij eksperiment: Potencial'nye slova v russkom jazyke: Vzgljad iznutri (Linguistic experiment. Possible words in Russian: A view from inside)

What is a possible word in a language? Does an idealized cognitive model exist in a speakers’ grammar or only in the imagination of linguists? What does it mean to be “theoretically possible”? The advent of large electronic corpora and software for database management now enable us to extract attested novel words that are generated ad hoc and therefore lie beyond the scope of “legitimate” lexicon. In our paper we show how cognitive linguistics can account for probabilistic aspects of grammar in terms of radial structures of linguistic phenomena. The cognitive approach has facilitated fruitful insights in the domain of word-formation (Booij 2010; Janda 1986, 2011; Nesset 2010; Onysko & Michel 2010). However, newly generated words that are possible for some speakers but not acceptable to others are hard to account for. We present a large-scale corpus-based study of Russian perfective factitive verbs built from nouns and adjectives like objasnit’ ‘clarify’ (