Relocating attributional phenomena within a language‐cognition interface: The case of actors' and observers' perspectives | Synapse