Abstract
Acoustic communication is now recognised as a widespread and functionally important component of fish behaviour. However, its evolutionary interpretation remains fragmented, as different types of acoustic signals are not interpreted with the same level of confidence. In particular, sounds recorded under disturbance or from hand‐held individuals are widely used in comparative studies but remain debated in evolutionary contexts. Here, we re‐examine these limitations by reviewing empirical, mechanistic and behavioural evidence across a wide range of teleost taxa. We show that interspecific acoustic differences are not restricted to reproductive contexts and that sounds obtained under disturbance can reflect intrinsic species‐level properties of neural control, musculature and sound‐producing structures. At the same time, we highlight that acoustic species specificity has not been demonstrated as a general rule across fishes, nor does sound alone constitute a universal prezygotic barrier. By integrating behavioural context, sound‐production mechanisms and recording methodologies, this review provides a framework to interpret acoustic variation in an evolutionary context and clarifies the scope, limits and evolutionary relevance of different acoustic datasets.