Many thanks to the reviewers and especially to @fuhsejan.bsky.social for his feedback on the paper! Many of the ideas in the paper build on his work.

This framework opens a new perspective on the structure of action in social relationships, some aspects of which I explore in the paper. Among other things, I propose a formalization of relational ambiguity and explore some of the conditions under which relationships are more or less ambiguous.

Through this lens, RFs become probability distributions over actions while any specific relationship is modeled as a distribution over frames.

RFs emerge from regularities in the duality of dyad and content. Rather than a priori asserting ‘types of ties,’ I show how RFs can be modeled and inductively inferred from the action content of 1.2 million character relationships in fiction (“what actions often go together in dyads?”).

I argue that we should approach and study such types as cultural structures. One promising way to do so is to conceptualize them as ‘relationship frames’—cultural blueprints that stabilize relational expectations by helping people make sense of the interaction in relationships.

Through this lens, RFs become probability distributions over actions while any specific relationship is modeled as a distribution over frames.