Evolutionary psychology and design reincarnation: A reply to Peters

“Design reincarnation”

Prominent evolutionary psychologists have articulated the link between genes and adapted designs as “design reincarnation,” whereby reliably developing species-typical designs are constructed anew during the ontogeny of each organism (Barrett, 2006, 2007; Tooby, Cosmides, & Barrett, 2003). So far as ontogeny is concerned, the naturally selected gene complexes associated with the design elements of organisms effectively also “pre-suppose” a specific multi-dimensional developmental parameter space—a sort of developmental backdrop—with which to interactively construct the elements of organismic design they have been selected to help achieve. This conceptualization has some affinities with the notion of “developmental canalization” (Griffiths & Machery, 2008) first propounded by Waddington (1959). Genes, after all, can be construed as “difference makers” (Sterelny & Kitcher, 1988)—heritable units that can be altered so as to effect different outcomes to the phenotypes that emerge as a result of the developmental matrix of genes interacting with the broader environmental context of relevance.

Although Peters refers to the concept of innateness throughout his critique, the concept is likely to generate confusion when discussing the theoretical foundations of evolutionary psychology. ² In general, evolutionary psychologists eschew use of the concept. Rather, and as highlighted earlier, the notion of “design reincarnation” is a much more suitable conception of the relation of genes to adaptations that evolutionary psychologists posit (Barrett, 2006, 2007; Tooby et al., 2003). Alternatively, one might wish to conceptualize such processes as developmental systems, as per evolutionary developmental biology and developmental systems theory (Oyama, Griffiths, & Gray, 2001; Pigliucci & Müller, 2010).

To a large extent, evolutionary psychologists aim to ultimately investigate the nature of reliably developing species-typical psychological adaptations. The general neurobiological considerations of developmental plasticity that Peters argues belie the theoretical position that evolutionary psychologists hold are deeply problematic. Such general neurodevelopmental facts are not only consistent with the general theoretical framework of evolutionary psychologists, they are in fact more or less what the theoretical orientation expects to be the case, at least in many instances. For, again, it is emphatically not the case that evolutionary psychologists posit the existence of an overarching cognitive architecture comprised of “genetically pre-specified” cognitive modules that are necessarily fully developed at birth or early childhood, or insensitive to experience across the board or even in general. Many of the modular capacities postulated by evolutionary psychologists are also aptly and metaphorically dubbed “learning machines”—modules designed by evolution to flexibly learn about and calibrate in responsive ways to specifically delimited domains. Indeed, as Marcus (2004) puts it, “Built-in doesn’t mean unmalleable; it means organized in advance of experience” (p. 40).

Once again, evolutionary psychologists are interested in the reliably developing targets of naturally selected design that recurrently arise in every generation in all normally developing humans. Importantly, these naturally selected developmental targets are to be construed as either occurring across different stages of ontogenetic time, as specific types of calibrations occurring during specific windows of development, or at various levels of cognitive abstraction. More concretely, and to take just one example with respect to levels of cognitive abstraction, although humans evince various mating patterns, evolutionary psychologists argue that specific types of surface-level variation across individuals and cultures are orchestrated in systematically predictable ways (e.g., Gangestad, Haselton, & Buss, 2006). In the case of such mating patterns, the reliably developing targets of natural selection of special interest to evolutionary psychologists are the underlying cognitive mechanisms systematically governing the surface-level variation. Indeed, it is the existence of these systematic patterns at the surface level which supports the existence of underlying domain-specific adaptations. For one would be at a loss in attempting to explain such systematically predictable patterns occurring at the surface level in merely domain-general/blank slate terms. Once again, it is not a legitimate criticism to claim that such domain-specific adaptations underlying the systematic mating patterns observed at the surface level must also be constructed developmentally and thus require experience-dependent input. For evolutionary psychologists agree that the very developmental targets of design that they are especially interested in no doubt require varying degrees of such developmental support and environmental input.

These above points can also be applied, mutatis mutandis, to Peters’ misguided critique of the treatment of culture by evolutionary psychologists such as Confer et al. (2010). Robust, systematically generated patterns occurring cross-culturally at any level require an explanation.

Peters accuses evolutionary psychologists of question begging for postulating an underlying naturally selected cognitive architecture as governing such cross-cultural patterns at the surface level. But pace Peters, it is actually question begging to assume that cultural factors alone are mysteriously generating the empirical patterns without any specification as to how this can happen cross-generationally at a universal scale. Such large-scale and iterated coincidences require an explanation (e.g., Salmon, 1984)—one articulated in terms more specific than the theoretical blunt instrument of “culture,” which amounts to a sort of hand-waving (i.e., “culture causes it”). In relation to this, one cannot underscore enough the notion of reliably developing design that evolutionary psychologists are particularly interested in. Such organized design certainly does not come for free, and evolution by natural selection is the only known anti-entropic process capable of generating such functional organized design in nature. As Tooby et al. (2003) remind us:

Thermodynamics informs us that, in general, such functional order does not come about spontaneously: Geology does not produce frescoed Tuscan villas, hurricanes do not assemble violins, and shaking up mixtures of sugar, milk, and dirt will not produce komodo dragons or congressmen. More generally, because functional orderings of atoms are a vastly smaller subset of arrangements than nonfunctional ones, they are astronomically more improbable. Because the second law of thermodynamics states that physical systems tend to move toward more probable states, they tend to move away from organization on their path toward maximum disorder. (p. 862)

Evolutionary psychologists are also fully aware of our uncontroversial capacity to co-opt our ancient cognitive adaptations (no differently than, say, our anatomical adaptations) in the service of a manifold of evolutionarily novel ends (e.g., Buss, Haselton, Shackelford, Bleske, & Wakefield, 1998). Relatedly, though distinctly, they are similarly aware of the way in which the mismatch that obtains between our evolved psychology, on the one hand, and our contemporary societies, cultures, and lifestyles, on the other, issues in a variety of maladaptive outcomes (in the evolutionary fitness sense) or otherwise “quaint” behaviors (Kanazawa, 2004; Plotkin, 1998; Tooby & Cosmides, 2005).

Peters (2013) also argues that it is “difficult to say what might have been there at birth, or instead shaped by common environmental experiences that we all share” (p. 311). However, and as this reply has attempted to make salient, it is not a necessary condition for some trait’s being an adaptation that it be present “at birth.” Nor is some trait’s being “shaped by common environmental experiences that we all share” incompatible with its being an adaptation. Moreover, so far as investigating and identifying psychological adaptations is concerned, evolutionary psychologists are able to conduct systematic research to test both adaptive and non-adaptive hypotheses (e.g., Buss, 2005; Schmitt & Pilcher, 2004).

Neurobiological and genomic issues

There are also problems with Peters’ interpretation of the brain development literature. ³ Here I highlight only a few points that are seemingly apropos. To begin with, many cortical structures, such as the organization of the somatomotor cortex, somatosensory cortex (Ramachandran & Blakeslee, 1998), and the intraparietal sulcus’ horizontal segment (which is implicated in object cardinality and audible sequence ordering), are traits not explicable in terms of neural pruning (Piazza & Dehaene, 2004). Ramus (2006) and Goldberg and Weinberger (2009) provide reviews of a sizeable neurobiological literature on human and non-human cortical development, demonstrating its genetic basis. Research on mice has also demonstrated that their brains can develop normally to a significant degree even in the absence of neurotransmitters, and hence in spite of experiential poverty (Verhage et al., 2000). Machery and Barrett (2006) also point out that in ferrets, monkeys, and cats, the highly resolved structure of their ocular dominance columns is initially set before the onset of any visual experience, and that representations of whiskers in the somatosensory cortex of rodents is similarly set before any real-world experience. There is thus evidence that specific genes are implicated in the development of cortical structures, even independent of experience.

Furthermore, since the developmental pathways undergirding the cortex are phylogenetically conserved, there is seemingly no reason to suspect that all highly resolved areas of the human cortex are simply experientially shaped without a significant amount of genetically guided structure. Research also indicates the heritability of the three-dimensional structure of many human cortical areas (Ramus, 2006). As one upshot to these considerations, it stands to reason that there has indeed been selection for relatively specific effects on cortical development in humans, even if such systems have been designed to be both malleably responsive in various ways to experience and dependent on experience for their development (Machery & Barrett, 2006).

Interestingly, Peters (2013) states that “at least part of what we have inherited by natural selection is the ability to flexibly shape our biology, within genetic constraints, to suit the demands of our environment and the symbolic realities defined by our cultures” (pp. 311–312), all the while overlooking that the theoretical foundation of evolutionary psychology fully integrates these theoretical possibilities, and that, in fact, evolutionary psychologists expect many of our psychological adaptations to be so characterized—though, of course, one must examine traits empirically on a case-by-case basis to determine the extent to which any particular trait is flexibly responsive to experience. As one example, Gangestad and Simpson (2000) provide an in-depth discussion of the highly facultative and adaptive nature of human mating.

Peters also claims that the so-called “gene-shortage” argument undercuts the modular conception of the human mind advanced by evolutionary psychologists. But the “gene-shortage” claim is quite controversial. For instance, Miller (2008) notes that even when comparing our genome (Collins & McKusick, 2001) to the genome of our closest evolutionary cousin, the chimpanzee (Olson & Varki, 2003), only 1.2 percent of the 3 billion DNA base pairs diverge (Ebersberger, Metzler, Schwarz, & Pääbo, 2002). And yet despite this underwhelming-sounding amount of change, the inter-specific genomic divergence that has occurred between humans and chimpanzees over the last 5 to 6 million years has implicated alterations to 35 million single-nucleotide polymorphisms (a conservative estimate), 5 million indels (insertions and deletions), substantive chromosomal translocations (Chimpanzee Sequencing and Analysis Consortium, 2005), sizable segmental duplications (Cheng et al., 2005), significant shifts in genetic recombination hot-spots (Ptak et al., 2005), alterations in the activity patterns of gene promoter regions (Heissig et al., 2005), as well as swift genetic changes subserving neurodevelopment in humans (Khaitovich et al., 2005).

And, as Miller (2008) also notes, the cognitive characteristics that distinguish us from chimpanzees are most likely to result from differences occurring via gene expression during ontogeny, which are orchestrated by cis-regulatory elements in the genome (Ochoa-Espinosa & Small, 2006; Stathopoulos & Levine, 2005), rather than by highly phylogenetically conserved elementary structural genes coding for proteins. Marcus (2004) provides an overview of developmental neurogenetics and cogently argues why there are more than enough genes to support the expectation of evolutionary psychologists that the human cognitive architecture is highly modular in constitution. Given these significant genomic changes, there is much reason to suspect that evolution continued to sculpt the human mind after the split with chimpanzees in ways which gave rise to a modular cognitive architecture, and that concomitantly enabled our distinctively human capacities. As Marcus and Rabagliati (2006) put it:

Modularity certainly does not require that all genes or even most of the genes involved in a given process be domain-specific; only that some genes (or even some portions thereof) be differentially expressed. Modularity can arise from the actions of a handful of “upstream” regulatory genes, even if many or all downstream genes are broadly shared across domains. (p. 397)

Psychological adaptations as domain-specific “learning machines”

Peters (2013) also accuses evolutionary psychologists of conducting “research supposing innate biological mechanisms, without seeing much need to study these mechanisms first-hand” (p. 315). However, given the current “black box”-like nature of how a great deal of our higher-order cognitive capacities are instantiated in the brain, investigation at the cognitive and behavioral level in a specifically adaptationist manner is a much-needed theoretical lens and methodological heuristic with which to investigate their neurobiological underpinnings (Platek, Keenan, & Shackelford, 2007; Platek & Shackelford, 2009; Tooby & Cosmides, 2000). For often in science higher-level empirical discoveries can predict the sorts of discoveries that inevitably will be made at lower levels. Relativistic physics and behavior genetics, for instance, can issue forth veracious claims even in the absence of lower-scale sub-atomic and molecular-genetic understandings of those higher-scale phenomena, respectively. Indeed, such higher-level understandings strongly force researchers working at the appropriate lower levels to look for the mechanisms instantiating those higher-scale phenomena. So far as cognitive modularity is concerned, some neuroscientists have already found traces of the dynamic signature of functional modules in the brain (e.g., Meunier, Lambiotte, & Bullmore, 2010; Meunier, Lambiotte, Fomito, Ersche, & Bullmore, 2009).

With respect to adaptive flexibility and plasticity, aspects of the human evolutionary niche that were unpredictably variable may have selected for a variety of modular designs inhering in our overall cognitive architecture that are accordingly less specific in their constituent representational formats and computational algorithms. Lesser degrees of domain-specificity may have likewise also selected for additional and perhaps very specific types of architectural connectivity between modular components, as well as algorithmic processes that govern how functional interconnectivity between modules occurs across ontogeny. This brief canvassing is not intended to provide an exhaustive taxonomy of the architectural possibilities for the human mind, which of course is as much an empirical issue as a theoretical one. Although an in-depth discussion and defense of massively modular cognitive architectures are beyond the scope of the present reply, the interested reader should consult the superb accounts by Carruthers (2006) and Barrett and Kurzban (2006).

At various points in his paper, Peters (2013) also seems to argue that domain-specificity entails rigid specifications that are largely incapable of being modified in response to experience. Additionally, he similarly seems to argue that any cognitive traits that evince more generalized, plastic capacities cannot nonetheless have been structured by design by natural and sexual selection to function within the context of some delimited, albeit more flexibly broad, domain. For instance, he asks:

Would it not make more intuitive sense, for example, to have acquired biological systems that direct us to “fear and avoid environmental threats” (e.g., things that evoke physical or emotional pain or discomfort), versus having separately acquired systems that direct us to “fear and avoid spiders,” “fear and avoid snakes,” “fear and avoid being raped,” and so on? (p. 312)

But the very function of fearing and avoiding environmental threats, despite being a relatively broad specification, is still not truly general—at least in the sense of a capacity vouchsafed by a supposedly tabula rasa-based cognitive engine—but rather specifically circumscribed in its selected-for domain of purposeful functioning. Indeed, evolutionary psychologists would have no necessary qualms with construing a psychological adaptation in this sense—that is, as characterized by a specific yet relatively broad domain of functioning, and as being constituted by at least some computational algorithms and representational formats which are designed to be flexibly plastic and responsive to various dimensions of environmental novelty in specific types of ways (Geary, 2005).

Some have also cogently argued that pure domain-general cognition does not in fact exist at all. To wit, Gallistel and his colleagues have argued that what appear prima facie to be domain-general processes in the paragon cases of associative learning among non-human animals are actually, upon closer inspection, finely calibrated domain-specific algorithms instantiating a “rate-estimation” procedure (Gallistel, 2000; Gallistel & Gibbon, 2001; Gallistel, Mark, King, & Latham, 2001). For, as Peters rightly points out, completely domain-general cognitive processes operating in complete absence of supportive structures and processes with more specific and structured design characteristics are left impotent by dint of the “frame problem.” It is therefore left ambiguous whether many of Peters’ qualms with evolutionary psychologists are the result of semantic misunderstandings or genuine disagreements.

In closing out his critique, Peters notes a number of empirical challenges to the empirical findings of evolutionary psychologists. But as per usual in the sciences, only direct appraisal of all of the relevant empirical evidence can arbitrate between competing hypotheses—which essentially amounts to comparative theory assessment, otherwise known as inference to the best explanation (e.g., Haig & Durrant, 2002; Lipton, 2008; Psillos, 2002; Thagard, 1978). Finally, given Peters’ insistence on the plasticity of the human mind, the following apposite remark by Tooby (2001) captures the need to transcend even this simple observation:

But what for [critics of evolutionary psychology] is the end of the story—our brains are largely learning systems—is for evolutionary psychologists, the beginning. Rather than accepting learning as an omnipotent black box, they insist on breaking open the box to map the circuit logic through which its programs solve adaptive problems. (p. 363)