In her book, Heyes (Reference Heyes2018) sets up a dichotomy between two potential sources for mechanisms of distinctly human cultural learning: biological specification and inheritance via genetic programs versus acquisition of these mechanisms via their own cultural learning processes. Referring to these two possibilities in shorthand, she asks if the processes that support cultural learning are cognitive instincts or cognitive gadgets.
If there are only two possibilities, then eliminating Option 1 allows for settling on Option 2 by default. Much of Heyes’ case for cognitive gadgets relies upon this strategy, as she focuses on ruling out the possibility that the mechanisms she investigates are cognitive instincts. She presents less conclusive evidence, however, that the process of acquiring these mechanisms involves cultural learning, defined as “social learning involving cognitive processes that are specialized for cultural evolution” (Heyes Reference Heyes2018, p. 86).
Heyes does not fully consider a plausible third option: that small, quantitative changes in genetic predispositions for social tolerance, motivation, and attention (i.e., the changes laid out in her “starter kit”) could collectively create the opportunity for developing humans to acquire novel social cognitive processes in the absence of any cultural transmission process or selection for cultural evolution and accumulation. To express this idea in another set of Heyes’ terms, she lays out how the genetic inheritance of the human starter kit may promote the social learning of “grist,” including social behaviors such as infant gaze following. I suggest that this genetic starter kit could also support the social acquisition of novel “mills” (i.e., cognitive processes that support learning), including one featured among Heyes’ case studies: imitation.
Heyes makes a convincing case that human imitation should not be considered a cognitive instinct. Setting aside the debate regarding the possibility of some neonatal imitation based on innate perceptual-motor mapping of a limited range of facial movements (Meltzoff et al. Reference Meltzoff, Murray, Simpson, Heimann, Nagy, Nadel, Pedersen, Brooks, Messinger, De Pascalis, Subiaul, Paukner and Ferrari2018; Oostenbroek et al. Reference Oostenbroek, Suddendorf, Nielsen, Redshaw, Kennedy-Costantini, Davis, Clark and Slaughter2016), she presents strong evidence that many “vertical” links between perceptual and motor representations of specific actions are acquired (and can be altered) through associative learning, and that we learn the ability to execute perceived sequences of actions motorically. Our comparative expertise at topographically matching others’ movements is thus not the product of a genetically specified mechanism, but is rather an acquired cognitive skill.
But is the acquisition of the ability to imitate an example of cultural learning, or does it fall in the space of noncultural social learning? Heyes defines cultural learning not by the presence of particular transmission mechanisms (e.g., teaching) or outcomes (e.g., selective learning), but by the selection of the learning mechanism for its role in promoting cultural evolution. Thus, to answer this question we need to ask if imitation is learned because it promotes the faithful transmission of and selection over cultural variants. Are the necessary experiences of correlated action execution and perception (e.g., parents’ imitation of their infants), and reward for imitating (e.g., social partners’ positive responses to being imitated), specialized for the transmission and evolution of a cultural learning strategy? Or is the prevalence of these elements in human social interaction the product of the incremental increases in the human genetic predisposition for social motivation and attention?
Research on the nature and role of imitation in social interaction suggests the latter. Human interaction partners imitate many aspects of one another's behavior regardless of the potential for social or cultural learning (Chartrand & Lakin Reference Chartrand and Lakin2013). As in the studies on synchrony and social bonding cited by Heyes (e.g., Tunçgenç & Cohen Reference Tunçgenç and Cohen2016), experiments find that imitative interactions elicit liking and rapport (Chartrand & Bargh Reference Chartrand and Bargh1999; van Baaren et al. Reference van Baaren, Holland, Kawakami and van Knippenberg2004). Moreover, the deployment of imitation in social interaction is sensitive to the need for social inclusion and the imitated social partner's ability to perceive that she is being imitated, indicating that users understand its role in creating rapport (e.g., Bavelas et al. Reference Bavelas, Black, Lemery and Mullett1986; Lakin et al. Reference Lakin, Chartrand and Arkin2008; Over & Carpenter Reference Over and Carpenter2009). A preference for imitators also precedes the robust use of imitation for cultural learning (or any other purpose) both phylogenetically and ontogenetically (e.g., Agnetta & Rochat Reference Agnetta and Rochat2004; Carpenter et al. Reference Carpenter, Uebel and Tomasello2013; Paukner et al. Reference Paukner, Suomi, Visalberghi and Ferrari2009; Powell & Spelke Reference Powell and Spelke2018b).
The preference for being imitated and for those who imitate their social partners cannot be explained by reward learning following one's own imitative behavior, nor can it be explained by a preference for contingency, which is typically equated for in control conditions. It is also not well explained by approval of the imitator's capacity for cultural learning, as might be the case for positive responses to a learner's skill at reading printed text, as it extends to the imitation of useless behaviors and to infant observers who do not yet imitate for cultural learning purposes themselves. Instead, preferences for imitation likely stem from what imitation indicates about the imitator's disposition as an attentive and affiliative social partner (Meltzoff Reference Meltzoff, Cicchetti and Beeghly1990; Powell & Spelke Reference Powell and Spelke2018a; Reference Powell and Spelke2018b).
This social value of imitation explains why young human learners receive both the experience necessary to solve the correspondence problem and the rewarding feedback that reinforces imitative behavior, without the need for cultural learning processes that have been selected based on the role of imitation in cultural evolution. Parents and other caregivers may imitate their infants (Kokkinaki & Kugiumutzakis Reference Kokkinaki and Kugiumutzakis2000; Pawlby Reference Pawlby and Schaeffer1977), providing correlated perceptual and motor experience, simply because that is a regular component of their interaction with valued social partners. Meanwhile, infants and young children may be rewarded for their early imitative behaviors as a result of the same social motivations that reinforce imitative interactions among adults. Finally, the relative prevalence of these factors in human social experience, compared to that of other primates, can be explained by the incremental shifts in motivation for social affiliation and attention described in Heyes’ starter kit. (This is not to say that, once acquired, the capacity for high-fidelity topographic imitation is not used for effective cultural learning, just that the forces that promote its acquisition do not depend on this use.)
What is at stake in the distinction between imitation as a culturally learned cognitive gadget versus merely a socially learned capacity? There are at least three implications. The first involves the necessary relationship between the imitative capacities that promote cultural learning and those that are acquired by learners via social interaction. As Heyes notes, for imitation to be shaped by cultural selection as a cognitive gadget, the imitative mechanism would need to be transmitted from model to learner with relatively high fidelity. In contrast, if imitation is learned due to its socially rewarding nature, along with the increased prevalence of social attention and engagement, then these forces can continue to shape imitation in similar ways across generations of learners without the need for strict fidelity across individuals.
Second, the distinction changes predictions about the impact of cultural disruption. Heyes notes that, like all cultural knowledge and skills, cognitive gadgets could be lost when cultures are disturbed by conflict or natural catastrophe. If, however, imitation is the product of the human starter kit, then it should recur regardless of the maintenance of a preexisting cultural repertoire.
This leads to the final implication, which is that the acquisition of imitative capacities via social, but not cultural, learning could serve as the starting point for the accumulation of culture, including true cognitive gadgets. The strength of selection for cultural learning mechanisms should depend on the size of the pool of cultural knowledge and skill that such mechanisms could potentially tap. When little cultural knowledge exists, accumulation may depend on mechanisms, like imitation, that could be acquired without the benefit of shaping via cultural selection. With the growth of information stored within a culture's traditions, dedicated cultural learning mechanisms will become more valuable, leading up to our current state of affairs, in which the information that can be accessed by cultural learning is now so vast that it is considered beneficial for virtually any human to engage in the immensely effortful, years-long process of learning to read. Thus, taking one item off her list of cognitive gadgets and assigning it to a third category may ultimately provide necessary support to Heyes’ innovative theory of cultural learning.
In her book, Heyes (Reference Heyes2018) sets up a dichotomy between two potential sources for mechanisms of distinctly human cultural learning: biological specification and inheritance via genetic programs versus acquisition of these mechanisms via their own cultural learning processes. Referring to these two possibilities in shorthand, she asks if the processes that support cultural learning are cognitive instincts or cognitive gadgets.
If there are only two possibilities, then eliminating Option 1 allows for settling on Option 2 by default. Much of Heyes’ case for cognitive gadgets relies upon this strategy, as she focuses on ruling out the possibility that the mechanisms she investigates are cognitive instincts. She presents less conclusive evidence, however, that the process of acquiring these mechanisms involves cultural learning, defined as “social learning involving cognitive processes that are specialized for cultural evolution” (Heyes Reference Heyes2018, p. 86).
Heyes does not fully consider a plausible third option: that small, quantitative changes in genetic predispositions for social tolerance, motivation, and attention (i.e., the changes laid out in her “starter kit”) could collectively create the opportunity for developing humans to acquire novel social cognitive processes in the absence of any cultural transmission process or selection for cultural evolution and accumulation. To express this idea in another set of Heyes’ terms, she lays out how the genetic inheritance of the human starter kit may promote the social learning of “grist,” including social behaviors such as infant gaze following. I suggest that this genetic starter kit could also support the social acquisition of novel “mills” (i.e., cognitive processes that support learning), including one featured among Heyes’ case studies: imitation.
Heyes makes a convincing case that human imitation should not be considered a cognitive instinct. Setting aside the debate regarding the possibility of some neonatal imitation based on innate perceptual-motor mapping of a limited range of facial movements (Meltzoff et al. Reference Meltzoff, Murray, Simpson, Heimann, Nagy, Nadel, Pedersen, Brooks, Messinger, De Pascalis, Subiaul, Paukner and Ferrari2018; Oostenbroek et al. Reference Oostenbroek, Suddendorf, Nielsen, Redshaw, Kennedy-Costantini, Davis, Clark and Slaughter2016), she presents strong evidence that many “vertical” links between perceptual and motor representations of specific actions are acquired (and can be altered) through associative learning, and that we learn the ability to execute perceived sequences of actions motorically. Our comparative expertise at topographically matching others’ movements is thus not the product of a genetically specified mechanism, but is rather an acquired cognitive skill.
But is the acquisition of the ability to imitate an example of cultural learning, or does it fall in the space of noncultural social learning? Heyes defines cultural learning not by the presence of particular transmission mechanisms (e.g., teaching) or outcomes (e.g., selective learning), but by the selection of the learning mechanism for its role in promoting cultural evolution. Thus, to answer this question we need to ask if imitation is learned because it promotes the faithful transmission of and selection over cultural variants. Are the necessary experiences of correlated action execution and perception (e.g., parents’ imitation of their infants), and reward for imitating (e.g., social partners’ positive responses to being imitated), specialized for the transmission and evolution of a cultural learning strategy? Or is the prevalence of these elements in human social interaction the product of the incremental increases in the human genetic predisposition for social motivation and attention?
Research on the nature and role of imitation in social interaction suggests the latter. Human interaction partners imitate many aspects of one another's behavior regardless of the potential for social or cultural learning (Chartrand & Lakin Reference Chartrand and Lakin2013). As in the studies on synchrony and social bonding cited by Heyes (e.g., Tunçgenç & Cohen Reference Tunçgenç and Cohen2016), experiments find that imitative interactions elicit liking and rapport (Chartrand & Bargh Reference Chartrand and Bargh1999; van Baaren et al. Reference van Baaren, Holland, Kawakami and van Knippenberg2004). Moreover, the deployment of imitation in social interaction is sensitive to the need for social inclusion and the imitated social partner's ability to perceive that she is being imitated, indicating that users understand its role in creating rapport (e.g., Bavelas et al. Reference Bavelas, Black, Lemery and Mullett1986; Lakin et al. Reference Lakin, Chartrand and Arkin2008; Over & Carpenter Reference Over and Carpenter2009). A preference for imitators also precedes the robust use of imitation for cultural learning (or any other purpose) both phylogenetically and ontogenetically (e.g., Agnetta & Rochat Reference Agnetta and Rochat2004; Carpenter et al. Reference Carpenter, Uebel and Tomasello2013; Paukner et al. Reference Paukner, Suomi, Visalberghi and Ferrari2009; Powell & Spelke Reference Powell and Spelke2018b).
The preference for being imitated and for those who imitate their social partners cannot be explained by reward learning following one's own imitative behavior, nor can it be explained by a preference for contingency, which is typically equated for in control conditions. It is also not well explained by approval of the imitator's capacity for cultural learning, as might be the case for positive responses to a learner's skill at reading printed text, as it extends to the imitation of useless behaviors and to infant observers who do not yet imitate for cultural learning purposes themselves. Instead, preferences for imitation likely stem from what imitation indicates about the imitator's disposition as an attentive and affiliative social partner (Meltzoff Reference Meltzoff, Cicchetti and Beeghly1990; Powell & Spelke Reference Powell and Spelke2018a; Reference Powell and Spelke2018b).
This social value of imitation explains why young human learners receive both the experience necessary to solve the correspondence problem and the rewarding feedback that reinforces imitative behavior, without the need for cultural learning processes that have been selected based on the role of imitation in cultural evolution. Parents and other caregivers may imitate their infants (Kokkinaki & Kugiumutzakis Reference Kokkinaki and Kugiumutzakis2000; Pawlby Reference Pawlby and Schaeffer1977), providing correlated perceptual and motor experience, simply because that is a regular component of their interaction with valued social partners. Meanwhile, infants and young children may be rewarded for their early imitative behaviors as a result of the same social motivations that reinforce imitative interactions among adults. Finally, the relative prevalence of these factors in human social experience, compared to that of other primates, can be explained by the incremental shifts in motivation for social affiliation and attention described in Heyes’ starter kit. (This is not to say that, once acquired, the capacity for high-fidelity topographic imitation is not used for effective cultural learning, just that the forces that promote its acquisition do not depend on this use.)
What is at stake in the distinction between imitation as a culturally learned cognitive gadget versus merely a socially learned capacity? There are at least three implications. The first involves the necessary relationship between the imitative capacities that promote cultural learning and those that are acquired by learners via social interaction. As Heyes notes, for imitation to be shaped by cultural selection as a cognitive gadget, the imitative mechanism would need to be transmitted from model to learner with relatively high fidelity. In contrast, if imitation is learned due to its socially rewarding nature, along with the increased prevalence of social attention and engagement, then these forces can continue to shape imitation in similar ways across generations of learners without the need for strict fidelity across individuals.
Second, the distinction changes predictions about the impact of cultural disruption. Heyes notes that, like all cultural knowledge and skills, cognitive gadgets could be lost when cultures are disturbed by conflict or natural catastrophe. If, however, imitation is the product of the human starter kit, then it should recur regardless of the maintenance of a preexisting cultural repertoire.
This leads to the final implication, which is that the acquisition of imitative capacities via social, but not cultural, learning could serve as the starting point for the accumulation of culture, including true cognitive gadgets. The strength of selection for cultural learning mechanisms should depend on the size of the pool of cultural knowledge and skill that such mechanisms could potentially tap. When little cultural knowledge exists, accumulation may depend on mechanisms, like imitation, that could be acquired without the benefit of shaping via cultural selection. With the growth of information stored within a culture's traditions, dedicated cultural learning mechanisms will become more valuable, leading up to our current state of affairs, in which the information that can be accessed by cultural learning is now so vast that it is considered beneficial for virtually any human to engage in the immensely effortful, years-long process of learning to read. Thus, taking one item off her list of cognitive gadgets and assigning it to a third category may ultimately provide necessary support to Heyes’ innovative theory of cultural learning.