Abstract
Reciprocity represents one of core research issues in current linguistics. However, identification of inherently reciprocal predicates, i.e., predicates that bear the feature of mutuality in their lexical meaning, have proven to be a challenge, as the criterion of the so called strong reciprocity, adopted from formal semantics and defined with respect to a predicate in all its instances, appears to be too strict for natural language predicates.
In this paper, inspired by Gleitman & Partee (2022, ms.), we redefine this criterion, applying it not to all instance of a predicate but to its individual instances. This weaker definition allows us to group predicates of a natural language in a more appropriate way, as we illustrate on description of mutuality in the VALLEX lexicon, where the classification of verbs with respect to mutuality reliably predict syntactic behavior of predicates in simple reciprocal constructions.