Cross-species acoustic codes for yes and no in human nonverbal vocalizations

Abstract

Cross-specifically interpretable acoustic codes in mammal vocalizations have so far been evidenced for inner state-triggered, self-referential contexts only. To reveal if such codes exist more broadly, we tested whether humans can intentionally modulate their voice to nonverbally signal encouragement (yes, i.e., do it) or objection (no, i.e., don't do it) to their dogs' impending movement towards different locations (the vocalizer's self, i.e., here or external, i.e., there). In vocalizations following which dogs performed the intended action (approached/avoided the referenced location), yes/no was encoded in multiple acoustic parameters, with similar patterning across reference locations and vocalizer-sexes. Specifically, yes was reflected in higher f₀ mean and HNR mean, lower f₀ range, PPJ, entropy and power, shorter call length, more frequent bout formation, and earlier intensity peak. Acoustic distinctions for here vs. there were typically analogous to those for yes vs. no, but for many parameters reference location code was restricted to either yes or no. These findings reveal in human nonverbal vocalizations (1) the presence of cross-specifically decodable, robust acoustic patterns for yes and no; and (2) the same semantic primes behind yes vs. no and here vs. there distinctions. Universal acoustic codes in mammal vocalizations, previously reported only for inner-state triggered, self-referential messages, may thus extend to general, context-independent meanings conveyed during intentional communication.