Pragmatic language use encompasses three distinct skills: understanding social conventions, intonation processing, and causal reasoning

Successful communication requires frequent inferences. Such inferences span a multitude of phenomena: from understanding metaphors, to detecting irony and getting jokes, to interpreting intonation patterns. Do all these inferences draw on a single underlying cognitive ability, or does our capacity for non-literal language comprehension fractionate into dissociable components? Using an approach that has successfully uncovered structure in other domains of cognition, we examined co-variation in behavioral performance on diverse non-literal comprehension tasks across two large samples to search for shared and distinct components of pragmatic language use. In Experiment 1, n=376 participants each completed an 8-hour battery of 20 critical tasks. Controlling for general cognitive ability, an exploratory factor analysis revealed three clusters, which can be post-hoc interpreted as corresponding to i) understanding social conventions (critical for phenomena such as indirect requests, conversational implicatures, and irony), ii) interpreting contrastive and emotional intonation patterns, and iii) making causal inferences based on world knowledge. This structure largely replicated in a new sample of n=400 participants (Experiment 2, pre-registered) and was robust to analytic choices. This research uncovers structure in the human communication toolkit and can inform our understanding of pragmatic difficulties in individuals with brain disorders. The hypotheses put forward here about the underlying cognitive abilities can now be evaluated in new behavioral studies, as well as using brain imaging and computational modeling, to continue deciphering the ontology of the component pieces of linguistic and non-verbal communication.