Heintz & Scott-Phillips (H&S-P) present an evolutionary and cognitive account of the uniquely open-ended and enormously rich expressive diversity of humans, facilities which, they contend, are underpinned by an interrelated suite of cognitive capacities that serve in the production and recognition of informative intentions. They maintain that while natural selection in nonhuman species constrains communication to “narrow domains of statistical mutual benefit,” the cognitive facilities they identify “unleash” human expression. We suggest that these same cognitive devices may be relevant to the human expansion of several classes of nonverbal vocalizations used in far more restricted contexts in other species.

Screams are a particularly interesting category of vocalizations to explore some of the ideas presented in the target article. Across a broad range of species, these vocalizations show remarkable evolutionary stability with respect to acoustical form and function. When in the grasp of a predator, many species of animals scream (Caro, Reference Caro2005; Högstedt, Reference Högstedt1983; Wise, Conover, & Knowlton, Reference Wise, Conover and Knowlton1999). Screams appear to have evolved, originally, to startle the predator and increase the probability of the caller's escape. As a natural example of the well-studied acoustic-startle response (Davis, Reference Davis and Eaton1984), the sonic features of screams employed in this context elicit a particular affective state in the predator. The vocalizations are, at this stage, expressive, in H&S-P's sense, because they serve to generate a “psychological reaction” in the predator.

However, in some highly social nonhuman primate species, the contexts in which screams occur, and their acoustical forms, have diversified over the course of evolution, shifting from limited predator–prey interactions to more complex and nuanced agonistic conflicts relating to the dominance relationships among conspecific group members (reviewed in Gouzoules, Reference Gouzoules and Namy2005; Gouzoules & Gouzoules, Reference Gouzoules, Gouzoules, Campbell, Fuentes, MacKinnon, Bearder and Stumpf2011). These screams recruit aid from allies, usually matrilineal kin. The vocalizations are communicative, in H&S-P's sense, in that they are stimuli (more precisely, signals) that function to generate a reaction by the means of stimulating complementary mechanisms of interpretation within the agonistic context. Evidence points to the conclusion that monkeys appear to recognize that production of vocalizations like screams follows rules of sender–receiver directionality that correspond to the dominance hierarchies in their social groups (Cheney & Seyfarth, Reference Cheney and Seyfarth2007, p. 268). The cognitive processes these nonhuman primates use to perceive conspecific screams are considerably more complex than in the predator-startle context (reviewed in Schwartz, Engelberg, & Gouzoules, Reference Schwartz, Engelberg and Gouzoules2020). Chacma baboons (Papio ursinus), for example, show particular interest when they hear played back sequences of dominance grunts from a subordinate individual followed immediately by submission screams from a dominant group member, an occurrence that would characterize a rank challenge (Bergman, Beehner, Cheney, & Seyfarth, Reference Bergman, Beehner, Cheney and Seyfarth2003). Upon hearing a conspecific scream, monkeys seem to make use of a mental representations of the identity of the vocalizer and the nature of a social interaction, and they contextualize this information within their knowledge of kinship and dominance relationships among individuals (for additional details and examples, see Gouzoules, Reference Gouzoules and Namy2005).

In our species, an evolutionarily conserved mammalian vocal–emotion system is augmented by a speech articulation complex that includes direct neural connections from motor cortex to the musculature controlling the larynx, yielding a “dual pathway vocal production” system (reviewed in Bryant, Reference Bryant2021). The ramifications for human affective signaling are immense because this system allows substantial emancipation of emotion from the evolutionarily original production context of vocalizations. In humans, uniquely, screams are associated with a much greater variety of emotional contexts including fear, anger, surprise, and happiness (Anikin, Bååth, & Persson, Reference Anikin, Bååth and Persson2018; Engelberg, Schwartz, & Gouzoules, Reference Engelberg, Schwartz and Gouzoules2021; Frühholz, Dietziker, Staib, & Trost, Reference Frühholz, Dietziker, Staib and Trost2021). These emotional displays do not ineluctably represent automatic embodiments of internal states: Humans are endowed with remarkable capacities for the volitional and flexible control of expressions, abilities aligned with those allowing speech (Fitch & Zuberbühler, Reference Fitch, Zuberbühler, Altenmuller, Schmidt and Zimmerman2013). As a result, signalers are likely to produce natural expressions in a variety of socioemotional contexts, some of which derive primarily from internal states, while others are driven more by cultural norms or strategic motivations, including manipulation (Scherer, Reference Scherer2013). Humans can use screams voluntarily and deliberately to attract attention and, also, spontaneously and involuntarily when the caller is threatened, startled, or even experiencing sudden joy, and it is not always possible to distinguish these two situations (Engelberg et al., Reference Engelberg, Schwartz and Gouzoules2021). Of note, the degree to which listeners accurately perceive different emotions from screams varies across contexts which, given the evolutionary saliency of screams (they almost inevitably attract attention), places a premium on social vigilance of the sort proposed in the target article and by other authors (e.g., Bryant, Reference Bryant2021).

The added social complexity of screams that is seen in monkeys and, in humans additionally, the enhanced level of intentionality involved in their production, promotes social vigilance. Listeners must attend to the acoustic variation among screams to make inferences about the caller's intentional actions and subsequent behaviors (in humans), even if, in the case of monkeys, they cannot infer informative or communicative intentions. In primate screams, then, we see a communicative context in which, relative to their evolutionary precursors, both the cognitive capacities for goal-directed communication and correlated capacities for social vigilance have advanced along the graded scales described by H&S-P.

Thus, human scream production and usage suggests novel functions in this evolutionarily conserved call type, make them a fascinating subject for understanding nonlinguistic emotional communication in our species. We focused here on screams, but laughter (e.g., Gervais & Wilson, Reference Gervais and Wilson2006) and tearful crying (e.g., Vingerhoets & Bylsma, Reference Vingerhoets and Bylsma2016) represent additional examples where human expression has expanded beyond that shown by any other species. We suspect that H&S-P might not agree with the extrapolation of their term “unleashed” to include the expansion of human nonverbal expression, but they might agree that, in the case of screams, the leash has been loosened.