Lee and Schwarz (L&S) propose grounded procedures of separation as a proximate mechanism for cleansing effects. This same mechanism is proposed to underlie other grounded procedures of separation (e.g., enclosing and avoiding contact), with a variety of psychological consequences (e.g., sympathetic magic and positive contagion). Thus, they claim that the mechanism producing cleansing effects is domain-general, and that only a grounded procedures account can explain how cleansing effects occur in an array of contexts. In this commentary, we argue that this claim can be tested by investigating the ontogenetic origins of grounded procedures of separation. Overall, the developmental evidence casts doubt on the existence of the proposed domain-general mechanism.

The development of these “separation” (or, conversely, “connection”) procedures should include some key elements. First, children must form a mental representation of the act of separation (or act of connection). The idea of cleansing, for example, suggests that there is something that must be purified or removed from one's person. In other words, children must come to understand that there are contaminants (visible and invisible) in their environments, and that such contaminants are threatening (i.e., they harbor disease or may result in other deleterious effects). Second, children must come to construe the procedures as acts of both physical separation (i.e., can remove a physically present contaminant such as dirt) and psychological separation (i.e., can remove an imagined contaminant such as bad luck). Finally, children must start behaviorally displaying the “separation” and “connection” effects. If grounded procedures of separation are the proximate mechanism behind all of these domains, then one may predict that children will display cleansing effects and other “separation effects” at similar developmental time points. Children should start cleansing themselves of dirt and germs at the same time they begin to separate themselves from social outgroup members (we suggest that this can be understood as a form of separation as one is avoiding contact with outgroup members because of negative views or expectations of the outgroup; see Table 2 in the target article). If the trajectory of development is constant across domains, then the developmental evidence supports the domain generality of grounded procedures of separation. However, if the different domains follow different trajectories, then this complicates L&S's claim that all these phenomena involve the same mechanism.

In fact, although the domain-generality of grounded procedures of separation and connection has not been explicitly studied with developmental populations, the present literature suggests that such effects may not have consistent developmental trajectories. In some domains, the concepts of separation and connection and subsequent avoidance behaviors appear quite early in life, but in others they emerge much later. For example, infants in their second year of life have a concept of both connection (two foods touching connects them) and separation (removing the disliked food from the plate alleviates some concern) when it comes to foods. Eighteen-month-olds will refuse to eat a preferred food that has been “contaminated” by touching a disliked food on the same plate, and many even call for the disliked food to be entirely removed from the plate (Brown & Harris, Reference Brown and Harris2012). However, when it comes to germs and illness, an understanding of contamination has a much more protracted development. Preschoolers do not differentiate between eating a clean versus germ-contaminated food (DeJesus, Shutts, & Kinzler, Reference DeJesus, Shutts and Kinzler2015) and do not avoid contact with someone who is “sick” (Blacker & LoBue, Reference Blacker and LoBue2016). In fact, it is not until age 5 or 6 that these capacities reliably emerge (for a review, see Rottman, DeJesus, & Greenebaum, Reference Rottman, DeJesus, Greenebaum, LoBue, Pérez-Edgar and Buss2019). It is important to note that some conceptual causal knowledge of germ contagion is relatively early-emerging (Blacker & LoBue, Reference Blacker and LoBue2016; Raman & Gelman, Reference Raman and Gelman2008). For example, 3-year-olds can accurately provide contamination-based explanations for illness when prompted (Legare, Wellman, & Gelman, Reference Legare, Wellman and Gelman2009). However, as explained above, it is not until kindergarten or later that children reliably display avoidance behaviors, whereas even 18-month-olds will avoid liked foods “contaminated” by disliked foods. Indeed, children who do not have knowledge of germ contagion will nevertheless engage in avoidance behaviors toward foods, animals, and core disgust elicitors (Stevenson, Oaten, Case, Repacholi, & Wagland, Reference Stevenson, Oaten, Case, Repacholi and Wagland2010).

Comparing the domains of food and illness suggests that perhaps children have a harder time understanding invisible contaminants such as germs, viewing them as more abstract than clearly visible foods. Indeed, this suggests that abstract forms of separation and cleansing (e.g., removing germs or an “essence”) may develop later than concrete forms (e.g., removing dirt or a disliked food). Yet, there are invisible elements of separation and connection that children seem to understand even earlier than germs and illness, such as the connection between people and their objects. As early as 4 years, children value authentic objects, rate objects owned by celebrities as worth more than others, and search for traces of ownership (Frazier & Gelman, Reference Frazier and Gelman2009; Gelman, Frazier, Noles, Manczak, & Stilwell, Reference Gelman, Frazier, Noles, Manczak and Stilwell2015; Gelman, Manczak, Was, & Noles, Reference Gelman, Manczak, Was and Noles2016; Hood & Bloom, Reference Hood and Bloom2008). As discussed by L&S, these so-called “sympathetic magic” effects should be undergirded by grounded procedures of connection.

Moreover, although some procedures of separation (e.g., desiring foods to be separated) seem to mature spontaneously, it seems children must be socialized to perform some procedures of separation – in particular, cleansing procedures (Oaten, Stevenson, Wagland, Case, & Repacholi, Reference Oaten, Stevenson, Wagland, Case and Repacholi2014; Stevenson et al., Reference Stevenson, Oaten, Case, Repacholi and Wagland2010). Understanding the ontogenetic precursors to grounded procedures of separation will be a crucial complement to understanding the proximate mechanisms that produce these procedures in real time.

Taken together, children are precocious separators in some domains, but the prototypical act of separation – cleansing the body of contaminants – appears to be relatively late-developing and is not immediately understood as an act of separation. The current developmental evidence presents a complex but intriguing picture of how cleansing effects may emerge in childhood, and we challenge researchers to further investigate the ontogenetic roots of cleansing effects and grounded procedures of separation and connection more broadly.