Assuming that we have some reasonable way of characterizing the notion of "actual" word, then, Aronoff's suggested index of productivity might be formalized as (1), as suggested in Baayen [Baayen, Harald, 1989, A corpus-based approach to morphological productivity. Statistical analysis and psycholinguistic interpretation, Unpublished dissertation, Vrije University, Amsterdam]: (1) "I = V/S'" where "I" = index of productivity, "V" = the number of types, and "S" = the number of types the WFR in question could have given rise to.
- Baayen & Lieber (2004), a pag.64