Grossmann's target article posits a novel “virtuous caring cycle” on which the increased care that is received by more fearful children begets increased cooperative tendencies in those children (target article, sect. 1, para. 3). While this proposal is insightful in many ways, it may overlook an alternative, and potentially complementary, explanation of the unique level of cooperativeness that we find in humans. More specifically, for all that's been said, anxiety remains an equally plausible driver of the ontogenetic changes that Grossmann's proposal aims to explain.

The starting place for an anxiety-focused alternative is the observation that human social life is structured by norms whose complexity and vagueness often leave individuals uncertain about how they ought to behave (Kitcher, Reference Kitcher2011; Sterelny, Reference Sterelny, Sterelny, Joyce, Calcott and Fraser2013). Anxiety, as an emotion triggered by problematic uncertainty of this sort, would then be a predictable and prudent response (Hookway, Reference Hookway1998; Kurth, Reference Kurth and Timmons2015). More specifically, we should expect that through the course of human history, there was a tendency for some individuals to experience increased anxiety as they struggled to navigate these complex and uncertain social norms (Kurth, Reference Kurth2016). Moreover, this anxiety would have brought the emotion's characteristic response: increased risk assessment and risk minimization behaviors (Davis, Walker, Miles, & Grillon, Reference Davis, Walker, Miles and Grillon2010; Kurth, Reference Kurth2018; Öhman, Reference Öhman, Lewis, Haviland-Jones and Barrett2008). In the context of alloparenting, when these anxiety-driven behaviors were experienced by children, they would likely have manifested as, for example, cautious approach, reassurance seeking, and greater deference toward authority figures – behaviors, that is, that could have worked to enhance the cooperative tendencies of children, especially in the more anxious ones. But, contra Grossmann's picture, these enhanced cooperative tendencies would not be the upshot of a virtuous caring cycle. Rather, they would be the upshot of anxiety.

To draw out the plausibility of this alternative explanation, we can move in two steps. First, notice that there's an important ambiguity in Grossmann's rendering of the “fear” that he takes to underwrite the virtuous caring cycle. In particular, his distinction between general fear and the distinct fearfulness of social animals (target article, sect. 3.1, para. 9) mirrors the standard ways that fear and anxiety are often distinguished in emotion science: Fear engages a fight/flight/freeze response in the face of imminent dangers, while anxiety prompts cautious approach in the face of uncertain threats (e.g., Davis et al., Reference Davis, Walker, Miles and Grillon2010; Kurth, Reference Kurth2016, Reference Kurth2018; Öhman, Reference Öhman, Lewis, Haviland-Jones and Barrett2008). But while this common understanding of anxiety is functionally similar to Grossmann's notion of the fearfulness of social animals, the two responses are underwritten by distinct motivations: Risk assessment/minimization motives in the case of anxiety, care/affiliation seeking for Grossmann's fearfulness. Recognizing this suggests that two different mechanisms may underlie the enhanced cooperation seen in “fearful” children.

Second, much of the evidence that Grossmann marshals in defense of his virtuous caring cycle is compatible with an anxiety-focused alternative, thus frustrating our ability to determine which of these two mechanisms (or both) underlies humans' distinctive tendency toward cooperation. Three examples will help draw this out.

1. (1) The paper makes extensive appeal to research using emotional facial expressions, and it does this both as a way of measuring relative “fearfulness” levels and as a way of inducing “fear” so that the resulting neuro/chemical changes and behaviors (e.g., cooperativeness) can be assessed. But using emotional facial expressions in these ways is highly controversial (e.g., Barrett, Reference Barrett2006). Moreover, even if we set these general worries aside, a deeper problem remains: In Grossmann's use of this research, what counts as a “fearful” facial expression is understood so broadly that we cannot say which emotion is actually in play. For instance, in some of Grossmann's appeals to the facial expression research, “fear” is understood expansively so as to include general distress displays (target article, sect. 3.1, para. 9). Moreover, where a more narrow rendering is sought by way of contrasting “fearful” faces with other negative emotion facial expressions, the work typically just compares “fear” and “anger” faces (e.g., Grossmann & Jessen, Reference Grossmann and Jessen2017; Krol, Monakhov, Lai, Ebstein, & Grossmann, Reference Krol, Monakhov, Lai, Ebstein and Grossmann2015). Thus, we do not have evidence that allows us to tease apart our competing fear- and anxiety-based hypotheses. In fact, to the extent that the experiments that Grossmann's argument builds from make use of just still images (thus leaving test participants without contextual cues about what the emotional expression is a response to), they're arguably more likely to cue anxiety (a sensitivity to uncertain threats) than fear (a sensitivity to imminent dangers).

2. (2) The article points to imaging work indicating that neural structures like the amygdala and striatum are engaged in “fear” responses. But because these structures have also been shown to be engaged as part of anxiety responses (e.g., Davis et al., Reference Davis, Walker, Miles and Grillon2010; Kurth, Reference Kurth2018; Lago, Davis, Grillon, & Ernst, Reference Lago, Davis, Grillon and Ernst2017), we do not have findings that provide unique support for Grossmann's fearfulness model. Similarly, the paper cites the longitudinal study of Tuulari et al. (Reference Tuulari, Kataja, Leppänen, Lewis, Nolvi, Häikiö and Karlsson2020) in support of the amygdala's role in the development of fear biases. But it's unclear how much support this work provides for favoring a fear-based mechanism because the Tuulari's study did not test for activity in “anxiety” brain regions like the base nucleus of the stria terminalis (BNST).

3. (3) The research Grossmann cites in support of a connection between fear detection and increased cooperation also fails to be dispositive (target article, sect. 3.2, paras. 1–6): Much of this work implicates not just fear and its neural correlates (e.g., the amygdala), but also anxiety and its neural correlates (e.g., the BNST) (Bosch, Reference Bosch2011; Insel, Reference Insel1990; Marsh, Reference Marsh2015).

In sum, Grossmann is correct to focus on the role that affective traits might play in fostering cooperation; and in looking beyond the usual suspects like empathy, his work advances our understanding of the complex mechanisms that are likely to underlie these ontogenetic changes. That said, it appears that he has not (yet) succeeded in identifying a virtuous cycle of caring as the central driver of the enhanced cooperation that we see in “fearful” children.