In her book, Cecilia Heyes (Reference Heyes2018) puts forward a bold hypothesis: The foundational and unique meta-cognitive capacities underlying the human mind – social learning, imitation, theory of mind and language – are not genetically-based “instincts.” They are culturally evolved “gadgets.” Constructed in the social domain, they are transmitted through cultural inheritance, and acquired through cultural learning. What this means is that virtually nothing was required at the brain level during human evolution apart from a dramatic increase in the overall capacity of the brain for plasticity. Heyes acknowledges some additional biases in the ontogenetic “starter kit”: slightly greater social tolerance and social motivation; some slight perceptual visual and auditory biases; and greater executive control based on our disproportionately larger frontal lobes. But apart from these, Heyes claims that our cognitive uniqueness lies in the simple fact that our brains are much larger. Because our large brains evolved for increased general plasticity, there is no reason to expect the gadgets to be supported by genetically given, gadget-specific brain structures and cognitive biases.
We agree with Heyes that increased plasticity is the hallmark of human metacognition, and we think that Heyes has done an admirable job showing how associative learning can account for what seems like inborn biases. We believe, however, that Heyes’ overall conclusion – although theoretically possible – is implausible. To see why, we have to look more closely at the phylogenetic dynamics.
Incontestable signs of cultural, social, and cognitive uniqueness began to appear in hominin communities around 2 million years ago. From then on, these communities gradually transformed themselves into increasingly efficient, gadget-based collaborative enterprises, which required an entire array of new capacities – exactly those that we don't share with our ape relatives. Individuals thus began to be selected for their capacity to participate in the collaborative efforts. Increased plasticity was rewarded, and so were the biases that Heyes acknowledges. But beyond this, every variation that allowed individuals to extend and put less effort into their learning – whatever its nature – was probably selected, as long as it did not jeopardize other learning capacities too much. Different culturally learned adaptations were gradually genetically accommodated, forming biases for the gadgets. It was this accommodation that allowed the extension of plasticity in specific domains, and the invention of new and more elaborate gadgets, which increased the pressure on individuals, and so on.
As West-Eberhard (Reference West-Eberhard2003) and others have shown, the process of genetic accommodation (as opposed to the narrower process of genetic assimilation) encompasses the evolution of plasticity and the evolution of canalization, which includes the evolution of canalization for plasticity. Consider, for example, the human hand. The hand is an exquisitely evolved “tool of tools,” as Aristotle put it, with specific morphology, sensitivity, innervation, and musculature. It evolved in the context of tool use and social communication in small, highly collaborative hominin groups. Its hyperplastic affordances are based on the canalization of its accommodated structure. Human hyperplastic culturally molded cognition evolved in the very same social and ecological context and involved multiple interacting, selection pressures that resulted in patterns of plasticity-enhancing canalization: organized systems that extended humans’ affordances. Heyes claims that our brains, our “cognitive hands,” are shaped almost exclusively during ontogeny. The alternative is that, just like the hand, the human brain was gradually shaped during phylogeny for specific types of plasticity and enhanced affordances, required for specific types of social behavior.
Heyes doesn't mention genetic accommodation in her book, and her arguments against the role of genetic assimilation are based on problematic assumptions. First, genetic assimilation need not be complete; it is almost always partial, leading to quicker and more efficient context-sensitive responses, so plasticity need not be compromised. Second, the genetic assimilation of one part of a behavioral sequence can facilitate learning that refines and lengthens this behavioral sequence (Avital & Jablonka Reference Avital and Jablonka2000). Third, and most relevant to our argument, genetic assimilation of specific plasticity-promoting strategies increases, rather than decreases, the more general aspects of plasticity (Dor & Jablonka Reference Dor, Jablonka, Larson, Deprez and Yamakido2010). Culturally learned strategies can drive the evolution of the enhanced plasticity that extends the scope and ease of learning of this strategy, leading to its genetic accommodation. In the case of imitation, for example, an enhanced, initially laboriously learned plastic ability to imitate unfamiliar contents is likely to become genetically accommodated in changing environments requiring context and content flexible imitation.
Our interpretation of the observations that fraternal and identical human twins are alike in some of their metacognitive capacities (e.g., mindreading), and that both familiar and unfamiliar tasks can be learned with the same facility, is different from that of Heyes. Whereas she infers from that similarity that no genetic assimilation for these metacognitive capacities was involved in their evolution, we propose that the similarity is the result of strong positive selection for plasticity and canalization in the relevant cognitive domains (e.g., imitation, mindreading, control of social emotions, imagination, additional levels of representation). Selection for plasticity and canalization renders genetic variation cryptic and selectively neutral and expands responsiveness to environmental cues, which is exactly what these studies show.
Heyes argues that genetic assimilation did not occur because the cultural world changes too rapidly. Thus, for example, specific linguistic structures change too quickly to be accommodated. This is correct, and this is indeed why Chomsky's universal grammar project – the flagship of the “instincts” approach – resulted in failure. But the functions of language-specific structures that enhance the overall communicative capacity of all languages are much more stable, much more fundamental: Language requires the capacity for auditory phonetics, which cannot be reduced to general hearing; it requires the capacity for semantic categorization, which cannot be reduced to the ways we and other animals categorize our worlds of experiences; it requires new levels of representation, of words and communicative norms, and an entirely new relationship between episodic memory and the capacity for imagination (Dor Reference Dor2015). It goes without saying that every aspect of our language-related physiology – the innervation and musculature around the mouth, the larynx, and the vocal cords; the unique function of the expanding muscles around the lungs; and so on – has been genetically accommodated. There is no reason to believe that the cognitive system, responsible for the activation and control of this physiology, somehow managed to remain unbiased towards it.
We thus agree (and have argued repeatedly) that the language capacity started, like literacy, as a cultural adaptation (Jablonka & Lamb Reference Jablonka and Lamb2005; Jablonka & Rechav Reference Jablonka, Rechav and Trabant1996). But language, unlike literacy, probably began to emerge, in a rudimentary form, as early as half a million years ago; and again, unlike literacy, it became an obligatory component of human life. Genetic variations contributing to the stabilization, fine-tuning, and enhancing the flexible capacity to culturally acquire a language must have been selected for. As we see it, the dynamic relationship between the evolution of the brain and the evolution of language was reciprocal: Language adapted itself to the brain, as Heyes emphasizes, but the brain also adapted itself to language, and the two came to be entangled in a co-evolutionary spiral. Language, in other words, culturally, and then genetically, molded the brain.
All this means that we should expect to find additional biases and changes in brain organization that are related to the human-specific, gadget-based social environment. At the emotional level, we see it in the capacity to internalize the “social gaze” and thus feel the uniquely human social emotions of shame and guilt, embarrassment and pride. These emotions express themselves, as Darwin (Reference Darwin1872) noted, in the uniquely human blush. The blush and its emotional and cognitive foundations seem to be a part of the human innate starter-kit (Crozier Reference Crozier2006). A second bias, also related to the evolution of social emotions, is the increased executive control of emotions, their social (and sometimes voluntary) regulation, and their linkage with episodic memory and social communication. We see this increased social-communicative emotional control as a specific evolved facet of a more domain-general executive control, which was a precondition for the evolution of the language capacity, and has further increased following language evolution (Jablonka et al. Reference Jablonka, Ginsburg and Dor2012).
Genetic and morphological evidence show variations that lead to both the increase in general aspects of brain plasticity and in specific features of the human brain and presumably human cognition. Examples of the more general features are variations linked to the overall increase in the size of the human neocortex (Florio et al. Reference Florio, Albert, Tverna, Namba, Brand, Lewitus, Haffner, Sykes, Kuan Wong, Peters, Guhr, Klemroth, Prüfer, Kelso, Naumann, Nüsslein, Dahl, Lachmann, Pääbo, Wieland and Huttner2015; Gómez-Roblesa et al. Reference Gómez-Roblesa, Hopkins, Schapiro and Sherwood2015) and to the disproportionately larger human frontal lobes (Berto & Nowick Reference Berto and Nowick2018). One notable variation that seems to enhance particular aspects of human brain plasticity is the increased depth of the superior temporal sulcus in the right hemisphere that is probably related to social communication (Leroy et al. Reference Leroy, Cai, Bogart, Dubois, Coulon, Monzalvo, Fischer, Glasel, Van der Haegen, Bénézit, Lin, Kennedy, Ihara, Hertz-Pannier, Moutard, Poupon, Brysbaert, Roberts, Hopkins, Mangin and Dehaene-Lambertz2015). Many other variations that are associated with specific aspects in the human brain are reviewed by Sousa et al. (Reference Sousa, Meyer, Santpere, Gulden and Sestan2017) and include alterations of the arcuate fasciculus that is involved in audition-based language production; the organization of the superior longitudinal fasciculus (the primary white matter tract connecting lateral frontal with lateral parietal neocortical areas), which is implicated in social learning and tool use; great expansion of two lateralized human frontoparietal networks in the cortical regions; and structural and functional reorganization of corticofugal projection neurons (connecting the neocortex and the subcortical regions) that may be important for motor control and digital dexterity. Together, the distinctive functional attributes of these connections build up the domain-general metacognition of humans.
Genetic accommodation of developmental dispositions and strategies brought about selection for altering the threshold, range, and sensitivity of developmentally acquired and learned responses is ubiquitous. As West-Eberhard (Reference West-Eberhard2003) has shown, processes of genetic accommodation can explain the evolution of general and specific flexible aspects of animal morphology, behavior, and cognition. The key to human evolution is the fact that we constructed the cultural niche collectively, phenotypically accommodated to it individually, and eventually became, through genetic accommodation, even better niche constructors and phenotypic accommodators. As Heyes argues, the ontogenetic tool kit does not include “instincts.” It seems, however, to include a set of interconnected, partial biases and specific structural and functional reorganizations for different cultural gadgets. This set of biases is not the a priori foundation of the gadgets, as the “instincts” approach would have it: It is the a posteriori outcome of the huge and enduring significance of the gadgets in our social lives.
In her book, Cecilia Heyes (Reference Heyes2018) puts forward a bold hypothesis: The foundational and unique meta-cognitive capacities underlying the human mind – social learning, imitation, theory of mind and language – are not genetically-based “instincts.” They are culturally evolved “gadgets.” Constructed in the social domain, they are transmitted through cultural inheritance, and acquired through cultural learning. What this means is that virtually nothing was required at the brain level during human evolution apart from a dramatic increase in the overall capacity of the brain for plasticity. Heyes acknowledges some additional biases in the ontogenetic “starter kit”: slightly greater social tolerance and social motivation; some slight perceptual visual and auditory biases; and greater executive control based on our disproportionately larger frontal lobes. But apart from these, Heyes claims that our cognitive uniqueness lies in the simple fact that our brains are much larger. Because our large brains evolved for increased general plasticity, there is no reason to expect the gadgets to be supported by genetically given, gadget-specific brain structures and cognitive biases.
We agree with Heyes that increased plasticity is the hallmark of human metacognition, and we think that Heyes has done an admirable job showing how associative learning can account for what seems like inborn biases. We believe, however, that Heyes’ overall conclusion – although theoretically possible – is implausible. To see why, we have to look more closely at the phylogenetic dynamics.
Incontestable signs of cultural, social, and cognitive uniqueness began to appear in hominin communities around 2 million years ago. From then on, these communities gradually transformed themselves into increasingly efficient, gadget-based collaborative enterprises, which required an entire array of new capacities – exactly those that we don't share with our ape relatives. Individuals thus began to be selected for their capacity to participate in the collaborative efforts. Increased plasticity was rewarded, and so were the biases that Heyes acknowledges. But beyond this, every variation that allowed individuals to extend and put less effort into their learning – whatever its nature – was probably selected, as long as it did not jeopardize other learning capacities too much. Different culturally learned adaptations were gradually genetically accommodated, forming biases for the gadgets. It was this accommodation that allowed the extension of plasticity in specific domains, and the invention of new and more elaborate gadgets, which increased the pressure on individuals, and so on.
As West-Eberhard (Reference West-Eberhard2003) and others have shown, the process of genetic accommodation (as opposed to the narrower process of genetic assimilation) encompasses the evolution of plasticity and the evolution of canalization, which includes the evolution of canalization for plasticity. Consider, for example, the human hand. The hand is an exquisitely evolved “tool of tools,” as Aristotle put it, with specific morphology, sensitivity, innervation, and musculature. It evolved in the context of tool use and social communication in small, highly collaborative hominin groups. Its hyperplastic affordances are based on the canalization of its accommodated structure. Human hyperplastic culturally molded cognition evolved in the very same social and ecological context and involved multiple interacting, selection pressures that resulted in patterns of plasticity-enhancing canalization: organized systems that extended humans’ affordances. Heyes claims that our brains, our “cognitive hands,” are shaped almost exclusively during ontogeny. The alternative is that, just like the hand, the human brain was gradually shaped during phylogeny for specific types of plasticity and enhanced affordances, required for specific types of social behavior.
Heyes doesn't mention genetic accommodation in her book, and her arguments against the role of genetic assimilation are based on problematic assumptions. First, genetic assimilation need not be complete; it is almost always partial, leading to quicker and more efficient context-sensitive responses, so plasticity need not be compromised. Second, the genetic assimilation of one part of a behavioral sequence can facilitate learning that refines and lengthens this behavioral sequence (Avital & Jablonka Reference Avital and Jablonka2000). Third, and most relevant to our argument, genetic assimilation of specific plasticity-promoting strategies increases, rather than decreases, the more general aspects of plasticity (Dor & Jablonka Reference Dor, Jablonka, Larson, Deprez and Yamakido2010). Culturally learned strategies can drive the evolution of the enhanced plasticity that extends the scope and ease of learning of this strategy, leading to its genetic accommodation. In the case of imitation, for example, an enhanced, initially laboriously learned plastic ability to imitate unfamiliar contents is likely to become genetically accommodated in changing environments requiring context and content flexible imitation.
Our interpretation of the observations that fraternal and identical human twins are alike in some of their metacognitive capacities (e.g., mindreading), and that both familiar and unfamiliar tasks can be learned with the same facility, is different from that of Heyes. Whereas she infers from that similarity that no genetic assimilation for these metacognitive capacities was involved in their evolution, we propose that the similarity is the result of strong positive selection for plasticity and canalization in the relevant cognitive domains (e.g., imitation, mindreading, control of social emotions, imagination, additional levels of representation). Selection for plasticity and canalization renders genetic variation cryptic and selectively neutral and expands responsiveness to environmental cues, which is exactly what these studies show.
Heyes argues that genetic assimilation did not occur because the cultural world changes too rapidly. Thus, for example, specific linguistic structures change too quickly to be accommodated. This is correct, and this is indeed why Chomsky's universal grammar project – the flagship of the “instincts” approach – resulted in failure. But the functions of language-specific structures that enhance the overall communicative capacity of all languages are much more stable, much more fundamental: Language requires the capacity for auditory phonetics, which cannot be reduced to general hearing; it requires the capacity for semantic categorization, which cannot be reduced to the ways we and other animals categorize our worlds of experiences; it requires new levels of representation, of words and communicative norms, and an entirely new relationship between episodic memory and the capacity for imagination (Dor Reference Dor2015). It goes without saying that every aspect of our language-related physiology – the innervation and musculature around the mouth, the larynx, and the vocal cords; the unique function of the expanding muscles around the lungs; and so on – has been genetically accommodated. There is no reason to believe that the cognitive system, responsible for the activation and control of this physiology, somehow managed to remain unbiased towards it.
We thus agree (and have argued repeatedly) that the language capacity started, like literacy, as a cultural adaptation (Jablonka & Lamb Reference Jablonka and Lamb2005; Jablonka & Rechav Reference Jablonka, Rechav and Trabant1996). But language, unlike literacy, probably began to emerge, in a rudimentary form, as early as half a million years ago; and again, unlike literacy, it became an obligatory component of human life. Genetic variations contributing to the stabilization, fine-tuning, and enhancing the flexible capacity to culturally acquire a language must have been selected for. As we see it, the dynamic relationship between the evolution of the brain and the evolution of language was reciprocal: Language adapted itself to the brain, as Heyes emphasizes, but the brain also adapted itself to language, and the two came to be entangled in a co-evolutionary spiral. Language, in other words, culturally, and then genetically, molded the brain.
All this means that we should expect to find additional biases and changes in brain organization that are related to the human-specific, gadget-based social environment. At the emotional level, we see it in the capacity to internalize the “social gaze” and thus feel the uniquely human social emotions of shame and guilt, embarrassment and pride. These emotions express themselves, as Darwin (Reference Darwin1872) noted, in the uniquely human blush. The blush and its emotional and cognitive foundations seem to be a part of the human innate starter-kit (Crozier Reference Crozier2006). A second bias, also related to the evolution of social emotions, is the increased executive control of emotions, their social (and sometimes voluntary) regulation, and their linkage with episodic memory and social communication. We see this increased social-communicative emotional control as a specific evolved facet of a more domain-general executive control, which was a precondition for the evolution of the language capacity, and has further increased following language evolution (Jablonka et al. Reference Jablonka, Ginsburg and Dor2012).
Genetic and morphological evidence show variations that lead to both the increase in general aspects of brain plasticity and in specific features of the human brain and presumably human cognition. Examples of the more general features are variations linked to the overall increase in the size of the human neocortex (Florio et al. Reference Florio, Albert, Tverna, Namba, Brand, Lewitus, Haffner, Sykes, Kuan Wong, Peters, Guhr, Klemroth, Prüfer, Kelso, Naumann, Nüsslein, Dahl, Lachmann, Pääbo, Wieland and Huttner2015; Gómez-Roblesa et al. Reference Gómez-Roblesa, Hopkins, Schapiro and Sherwood2015) and to the disproportionately larger human frontal lobes (Berto & Nowick Reference Berto and Nowick2018). One notable variation that seems to enhance particular aspects of human brain plasticity is the increased depth of the superior temporal sulcus in the right hemisphere that is probably related to social communication (Leroy et al. Reference Leroy, Cai, Bogart, Dubois, Coulon, Monzalvo, Fischer, Glasel, Van der Haegen, Bénézit, Lin, Kennedy, Ihara, Hertz-Pannier, Moutard, Poupon, Brysbaert, Roberts, Hopkins, Mangin and Dehaene-Lambertz2015). Many other variations that are associated with specific aspects in the human brain are reviewed by Sousa et al. (Reference Sousa, Meyer, Santpere, Gulden and Sestan2017) and include alterations of the arcuate fasciculus that is involved in audition-based language production; the organization of the superior longitudinal fasciculus (the primary white matter tract connecting lateral frontal with lateral parietal neocortical areas), which is implicated in social learning and tool use; great expansion of two lateralized human frontoparietal networks in the cortical regions; and structural and functional reorganization of corticofugal projection neurons (connecting the neocortex and the subcortical regions) that may be important for motor control and digital dexterity. Together, the distinctive functional attributes of these connections build up the domain-general metacognition of humans.
Genetic accommodation of developmental dispositions and strategies brought about selection for altering the threshold, range, and sensitivity of developmentally acquired and learned responses is ubiquitous. As West-Eberhard (Reference West-Eberhard2003) has shown, processes of genetic accommodation can explain the evolution of general and specific flexible aspects of animal morphology, behavior, and cognition. The key to human evolution is the fact that we constructed the cultural niche collectively, phenotypically accommodated to it individually, and eventually became, through genetic accommodation, even better niche constructors and phenotypic accommodators. As Heyes argues, the ontogenetic tool kit does not include “instincts.” It seems, however, to include a set of interconnected, partial biases and specific structural and functional reorganizations for different cultural gadgets. This set of biases is not the a priori foundation of the gadgets, as the “instincts” approach would have it: It is the a posteriori outcome of the huge and enduring significance of the gadgets in our social lives.