Dwurodzajowe jednostki rzeczownikowe M2/M3 z obocznością -a/-u w D.lp. Modelowanie relacji między końcówkami a alternantami rodzajowymi na podstawie frekwencji

Abstract

In the dictionaries that list M2/M3 double gender noun units (with the morphosyntactic variance in ACC sing.), the variance of the endings -a/-u in GEN sing. is usually recorded synthetically, i.e. without their gender (M2 or M3) specification. In the genitive, there is essentially no syntactic distinction between M2 and M3 nouns, and only one of the endings – -u – is limited to one gender (M3). Still, the article presents various possibilities of qualification of the -a ending in GEN sing. within M2/M3 double gender noun units: 1. assigning it always (regardless of the frequency data) only to the M3 alternate 2. assigning it always to both alternates (frequency data only determines which of the two endings is dominant for M3) 3. assigning it – individually (based on the frequency data) – to one (only M2) or both alternates (also M3) of a double gender unit, with the assumption that the ratio of generic frequency in GEN sing. is: 3.1. either balanced (the same for all units); 3.2. or proportional to the “real” (quantifiable) ratio of generic frequency in ACC sing. A formal model was proposed for interpreting the relation between endings and types according to 3.1 and 3.2. The model was applied to the set of 43 noun units with the -a/-u variance, confirmed (and calculated) by the corpus, as well as with the M2/M3 double gender (in the light of ACC sing.).