Hearing reliability: a common prosodic code automatically signals confidence and honesty to human listeners
The success of human cooperation crucially depends on mechanisms enabling individuals to detect unreliability in their conspecifics. Yet, how epistemic vigilance is achieved from naturalistic sensory inputs such as speech remains largely unknown. Here we show that a common prosodic code signals both confidence and honesty to listeners. Using a novel data-driven method, we decode the prosodic contours driving social attributions of confidence and honesty from speech across pitch, duration and loudness. We find that these two kinds of judgements about the reliability of a speaker rely on a common prosodic code that is independent from individuals’ conceptual knowledge and native language. We also demonstrate that listeners extract these prosodic signatures of reliability automatically, and that this impacts the way they memorize new spoken words. These findings shed light on a unique communicative adaptation that enables human listeners to quickly detect and react to unreliability during linguistic interactions.