Heyes’ (Reference Heyes2018) cognitive gadgets theory (henceforth CGT) constitutes a significant advancement over the standard neo-Darwinian account in evolutionary psychology. According to the latter, the mind is a collection of innately specified domain-specific units, which operate through inner representations and algorithms and are shaped by natural selection because of their causal contribution in adapting the organism to critical issues in the environment (e.g., Barkow et al. Reference Barkow, Cosmides and Tooby1992; Buss Reference Buss2012; Cosmides & Tooby Reference Cosmides and Tooby2013). Culture, in turn, is the physical derivative of such modules and stands as instrumental to the solution of these adaptive problems. The CGT argues that the standard view “does not line up with the evidence from cognitive science” (p. 15). In contrast, it aims to replace the standard view with a “cultural evolutionary psychology” (p. 16), according to which cultural practices create cognitive functions by building upon internal and biologically selected domain-general mechanisms. The CGT thus conceives the mind not as a collection of cognitive modules but rather of culturally constructed “cognitive gadgets.”

Although we praise the constructivist lean assumed by the author, we intend to argue that the CGT does not completely free itself from a variety of problematic neo-Darwinian assumptions. Despite the fact that many accounts in social anthropology (Ingold Reference Ingold2007), cognitive archaeology (Iliopoulos & Garofoli Reference Iliopoulos and Garofoli2016; Knappett Reference Knappett2005; Malafouris Reference Malafouris2013), postphenomenology (Ihde Reference Ihde1990; Ihde & Malafouris Reference Ihde and Malafouris2018; Verbeek Reference Verbeek2005), and enactive cognition (Fenici & Garofoli Reference Fenici and Garofoli2017; Hutto Reference Hutto2008; Hutto & Myin Reference Hutto and Myin2013) have extensively opposed these assumptions over time, the author has not considered such criticisms in the current formulation of the CGT. Within this commentary, we intend to bring to the fore (and suggest) possible solutions to these critical issues that also affect the CGT. In particular, we will discuss and criticize its assumptions concerning (1) the computational and semantic-information theory of mind and culture, (2) the adaptionist view in cognitive evolution, (3) the completeness of cognitive functions as units of selection, and (4) the fixed link between brain regions and cognitive functions.

To start, the CGT shares with the standard neo-Darwinian view a computational and representational conception of mind and culture. Indeed, like modules, domain-general systems remain in any case representational components of a computational architecture. Their content is determined by teleosemantics (Heyes, Reference Heyes2018, pp. 27–30; see Millikan Reference Millikan2004), and natural selection alters their representational capacities by acting on the minimal constituents of their physical vehicles – that is, their genetic bases. At the same time, although the author states that “a cognitive mechanism certainly is not a pellet of information that can be copied inside your head, sent through the air, and planted wholesale in my head” (p. 44), she keeps adopting a model typical of cultural evolution theory, according to which culture is semantic information, traveling across human minds through social learning (e.g., Mesoudi Reference Mesoudi2011; Richerson & Boyd Reference Richerson and Boyd2005). The novelty of the current model thus lies in the proposal that different codes inhabiting respectively the biological and the cultural channels are now integrated, thereby generating the cognitive gadgets through “wide computations” (cf. Wilson Reference Wilson1994). However, these representational and computational foundations remain highly controversial at present. Representationalism has indeed been considered metaphysically inconceivable (Hutto & Myin Reference Hutto and Myin2013; Reference Hutto and Myin2017), the information-semantic view of culture utterly disembodied (Malafouris Reference Malafouris, Etzelmüller and Tewes2016; Walls Reference Walls2019), and the great success of these views primarily associated with biases in computer-based societies rather than empirical validation (Penny Reference Penny2017).

The CGT exploits these conceptual bases also to maintain a hard adaptationist stance in cognitive evolution – although modified in order to accommodate its constructivist amendments – according to which “cultural evolution has the potential to explain the adaptedness of distinctively human cognitive mechanisms” (Heyes Reference Heyes2018, p. 37). The core point is that “genetic and cultural evolution are based on the same, fundamental heuristic – variation-and-selective-retention” (p. 36). Thus, wide computations appear as elements of a broader code that are generated, modified, replicated, and transmitted downstream along the dual-inheritance system just described. This implies that we can analyze cognitive gadgets with the tools of the neo-Darwinian theory – although presented in the form of a “‘Campbellian’ selectionist approach” (p. 36). The modern human mind, allegedly shared by all human populations, therefore appears as a collection of adaptive gadgets that survived selection, in line with the idea that evolutionary principles ultimately justify why the mind has the shape we observe in the present.

In response to such a neo-Darwinian take on cognition and culture, we note that over time many critics from different domains have repeatedly challenged these assumptions. During the 1980s in archaeological theory, the post-processual movement strongly reacted to the idea from the New Archaeology and its heirs that human behavior and cognition are determined by adaptive laws, which could be assembled in the present through ethnographic studies and deductively applied to the study of the past (Bednarik Reference Bednarik2013, p. 21; Hodder & Hutson Reference Hodder and Hutson2003, chap. 2; Shanks & Tilley Reference Shanks and Tilley1987; Trigger Reference Trigger1998). Post-processual critics argued that human culture is created and maintained within a psychological and ideological dimension, which can overcome strict adaptive reasons. Similarly, critics in social anthropology have contended that human life is a shared narration and an embodied way of being in the world. Thus, they have rejected metaphors depicting humans as computational machines, and culture as the outcome of transmission and selection processes (Ingold Reference Ingold2004; Reference Ingold2007; Ingold & Palsson Reference Ingold and Palsson2013; Tallis Reference Tallis2011). If these criticisms are correct, the human mind cannot be reduced to the evolutionary selection of wide computations acting on the cognitive level, as the author suggests, but rather cognitive gadgets are likewise embedded within a social narration. Thus, they can actively be formed and maintained because of what humans believe, think, desire, and emotionally experience independently from supposed adaptive reasons.

The previous discussion hints against a third important neo-Darwinian aspect of the CGT, namely, the idea that gadgets are distinct units of selection provided with specific adaptive values. Following the principles of material engagement theory (Iliopoulos & Garofoli Reference Iliopoulos and Garofoli2016; Malafouris Reference Malafouris2013; Renfrew Reference Renfrew, DeMarrais, Gosden and Renfrew2004), we argue that gadgets are not only integrated with a broad gamut of conceptual and affective states but are also materially extended through their hybridization with artifacts and features of the external world. Furthermore, they are constitutively related to one another, and the boundaries between them are hard to identify. For instance, similarly to Heyes, Everett (Reference Everett2012) has proposed that language is a cultural construct that capitalizes on domain-general systems capturing perceptual regularities, thereby appearing as a cognitive gadget. However, language is not an isolated cognitive function, because it is constitutive of and simultaneously constituted by other cognitive gadgets. To mention one, many scholars defending a constructivist perspective believe that meta-representational mindreading is realized through linguistic practices (Fenici Reference Fenici2017; Fenici & Garofoli Reference Fenici and Garofoli2017; Gallagher & Hutto Reference Gallagher, Hutto, Zlatev, Racine, Sinha and Itkonen2008; Hutto Reference Hutto2008). Mindreading in turn allows language to bend on itself, and bootstrap meta-linguistic awareness: namely, the ability to understand the abstract regularities behind the organization of language (Taylor Reference Taylor2012). Thus, language and mindreading coexist with one another.

The difficulties in isolating units of selection also speaks against the possibility of ascribing decontextualized adaptive values to gadgets, which could be modified and enhanced independently from the rest of the mind/world complex. In contrast, any variation within a gadget can cause cascade effects in a series of intermeshed and materially extended cognitive functions so that the adaptive value of a gadget is given from how it alters the whole cognitive world it lies within. Overall, the massive interconnection and coexistence of factors suggests that selection-based approaches cannot be applied to gadgets in a reductionist fashion and raises the possibility that neo-Darwinian methods are incommensurable with the domain of cognition and culture.

Finally, we argue that the CGT maintains a conservative idea of the relation between brain regions and functions – though milder than the standard evolutionary psychology account – because it contends that brain regions are selected by culture for mediating particular cognitive functions. However, following the theory of neural reuse (Anderson Reference Anderson2014; Raja Reference Raja2017), brain regions are deployed and redeployed in multiple cognitive tasks, depending on their internal constraints and relations to bodily and material structures. By acting on such physical constraints, natural selection makes some neural networks more or less suitable to hosting particular functions, but these networks have no original functional exclusivity of any sort. Thus, brain regions are evolved by culture without being for anything in particular, including domain-general processes.

The criticisms of the neo-Darwinian view that we have expressed shape a radically different conception of the mind in relation to the brain, the body, and culture. Rather than a computational device constructed through cultural selection, this view depicts the human mind through the metaphor of a “brainwave” that invades reality and resonates with external structures (Gibson Reference Gibson1966, p. 5; Raja Reference Raja2017 and references therein; Robbins Reference Robbins2006). Cognition in this sense stands as a dynamic system incurring between particular structural patterns in the world and neural territories showing appropriate conditions for resonance but does not involve a brainbound computational architecture. Furthermore, material structures are non-neutral, because they incite the mind, foster the creation of new patterns of resonance, and therefore continuously reshape cognition (Ihde Reference Ihde1990; Reference Ihde2009; Ihde & Malafouris Reference Ihde and Malafouris2018; Knappett Reference Knappett2005; Verbeek Reference Verbeek2005). Consequently, culture cannot be conceived as a disembodied code of operations and algorithms; rather, it implies the socially negotiated alteration of such structures, and the creation of new affordances for cognitive transformation that are inherited by the new generations (see the concept of a “landscape of affordances” in Rietveld & Kuverstein Reference Rietveld and Kiverstein2014 and in Rietveld et al. Reference Rietveld, Denys, Van Westen, Newen, De Bruin and Gallagher2018). Within this view, cognitive evolution is intended as a metaplastic trajectory, whereby the mind emerges only at the nexus of neural and cultural plasticity, and does not identify an autonomous level for natural selection (Aston Reference Aston2019; Iliopoulos & Garofoli Reference Iliopoulos and Garofoli2016; Malafouris Reference Malafouris2010; Roberts Reference Roberts2016; Woodward Reference Woodward2019).

This conception remains friendly with the constructivist inclination of the CGT, and yet renounces the idea that some cognitive constructs (i.e., the gadgets) are fixed units of adaption and selection, and with this avoids the serious reductionist pitfalls of the neo-Darwinian theory. In contrast, cognitive functions are one and many, intermingled with motivational and emotional aspects, and deeply embedded within materiality (cf. Spivey Reference Spivey2007). They are prone to developmental change and inherently incomplete (Malafouris Reference Malafouris, Etzelmüller and Tewes2016), so that any transformation in one of them potentially affects the whole structure of the mind as well as the relational entanglement of agents and the world (Garofoli Reference Garofoli and Prentiss2019). Such resulting conception predicating the coalescence of cognitive properties, the radical interconnection of beings, and the fluid transformation and “rebirth” of the mind and the world seems closer to a form of naturalized Buddhism rather than a neo-Darwinian theory of cognitive evolution (Varela et al. Reference Varela, Thompson and Rosch2017; Vogd Reference Vogd, Runehov and Oviedo2013). We hope to have convincingly shown that this conception allows the CGT to maintain its constructivist foundations while renouncing unnecessary and controversial assumptions.