Abstract The complexity of valency class systems remains an underexplored dimension of typological variation. Drawing on data from BivalTyp, a database cataloging valency patterns of 130 bivalent verbs across 120 languages, this paper proposes three metrics to quantify this complexity. Two of these metrics – the ratio of intransitive verbs among bivalent verbs and the entropy in the distribution of bivalent intransitive verbs across valency classes – are strongly positively correlated, reflecting different aspects of the overarching valency class complexity. The third metric, the entropy of all bivalent verbs across valency classes, captures both of these facets. Languages tend to gravitate toward moderate complexity values, avoiding overly simplistic or excessively complex valency class systems. However, significant variation exists, primarily driven by areal and genealogical factors. Valency class complexity, as measured and discussed in this paper, is empirically interwoven with numerous other parameters of morphosyntactic typology, suggesting that it is a fundamental factor in shaping linguistic diversity. Specifically, languages with large case inventories, a high prevalence of non-verbal predicates, and a preference for satellite-framed constructions in motion events are hypothesized to favor more complex valency class systems. Furthermore, complex argument-encoding devices tend to favor postverbal positions, aligning with the principles of incremental processing and the minimization of processing effort. Overall, this study contributes to understanding how efficiency pressures influence typological distributions.
Sergey Say (Sat,) studied this question.