Commentary on: P. Andrews, D. Gangestad & D. Mathews (AGM) Adaptationism – How to carry out an exaptationist program. Behavioral and Brain Sciences, in press.


Word Count: abstract = 60, text = 1678, references = 481, total = 2180



Modest Adaptationism:

Muddling Through Cognition and Language


Scott Atran


CNRS – Institut Jean Nicod

1 Bis Avenue de Lowendal

75007 Paris


Tel. 33-4 68 82 51 66 before 01-01-02




ISR – University of Michigan

426 Thompson St.

Ann Arbor, MI 48106-1248


Tel. (734) 936-0458 after 01-01-02



Abstract. Strong adaptationists would explain complex organic designs as specific adaptations to particular ancestral environments. Weak adaptationists don’t assume complex organic functioning represents evolutionary design in the sense of niche-specific adaptation. For some domain-specific competencies (folkbiology) strong adaptationism is useful, not necessary. With group-level belief systems (religion) strong adaptationism can become spurious pseudo-adaptationism. In other cases (language) weak adaptationism proves productive.



Recent cognitive experiments indicate that humans have specialized core mental faculties with privileged access to distinct but overlapping domains of nature, including: folkmechanics (object boundaries and movements), folkbiology (biological species configurations and relationships), and folkpsychology (interactive agents and goal-directed behavior). These plausibly innate (but maturing), domain-specific cognitive faculties are candidates for naturally-selected adaptations to relevant and recurrent aspects of ancestral environments. Under analytic idealization they are “universal” and “autonomous” from other cognitive faculties the way the visual system is universal and autonomous from other cognitive and biological systems (with significant individual genetic variation, and viability only in functional interaction with others faculties).


Folkbiology. A strong adaptationist stance helps to counter claims that folkbiology develops ontogenetically as an “exapted learning mechanism” (AGM’s term) from folkpsychology (Carey 1985). As AGM note, evidence for developmental specificity was lacking. Recent developmental and cross-cultural studies show that the apparent effects of folkpsychology on folkbiological development (e.g., anthropocentric construals of animals and plants) fade or disappear in “nonstandard” populations, that is, human groups other than those (students or children) linked to major research universities (Atran et al. 2001, Ross et al. in press). One interpretation is that nonstandard societies more closely approximate ancestral conditions of intimate interaction with nature. By contrast, standard populations (the near-exclusive focus of most developmental and cognitive psychology) need compensatory learning strategies for lack of sufficient exposure to triggering conditions that enable folkbiological knowledge, including strategies derived from folkpsychology and even folkmechanics (Au & Romo 1999).


Study of standard populations sometimes reveals more about effects of devolutionary cultural processes on innate knowledge than innate knowledge as such – much as study of language acquisition in feral children tells more about how the language faculty degenerates than how it evolved to develop (Medin & Atran in press). Evidence against exaptation in all populations stems from developmental and cross-cultural research; this may be compatible with strong prior or post hoc adaptationism but doesn’t require it.


Religion. To illustrate strong adaptationism’s epistemological stance, AGM cite Sober and Wilson’s (1998) work on social traits as (group-selected) adaptations. S&B reduce complex distributions of human behaviors to artificially coherent bundles of norms, termed “religion” or “culture.” S&B then assume social norms are actual behavioral traits that undergo Darwinian selection. In fact, purported norms are usually summary digests of lone analysts and anthropologists struggling to reduce the flux of observed social experiences into manageable proportions for report. Many social functional accounts (e.g., the Human Relations Area Files used by S&B) purposely exclude information on individual variation. Who holds these norms? An omniscient informant? Some expert(s)? Most people in society? Without such information it is impossible to verify or falsify claims about existence of norms, much less their evolution.


These pseudo-adaptationist accounts are often historically tendentious, sometimes pernicious: for example, D. S. Wilson’s (2002) notion of Judaism as a eugenics program designed to produce intelligent cabals that dominate other groups through intergroup competition (cf. MacDonald 1998). It’s not that religious and cultural cognitions and behaviors lack interesting generalities that can’t be productively investigated through evolutionary approaches (Atran 2002); only, religions and cultures as selectable objects simply don’t exist (except in a loose commonsense way), any more than do whole species (over and above constituent individuals) (Sperber 1996).


Language. Demonstration of biological preparedness doesn’t directly imply “hence, adaptation for language learning,” as AGM suggest. At best, strong adaptationist claims for syntax involve retrodictions of structures previously discovered (mainly through generative grammar). No novel predictions ensue. Reasonable people can argue over whether strong adaptationism has novel predictions or discoveries for any higher-order cognitive process (this excludes – perhaps artificially - “lower-order” cognitions related to sex, kinship and violence; and though I don’t believe there’s evidence for a “cheater-detection module,” Atran 2001, I’m less sure about adaptationist reworking of Kahneman and Tversky’s “biases and heuristics” in human reasoning).


The one seriously strong adaptationist argument for language – natural selection of syntax for serial communication of propositions (Pinker & Bloom 1990) – may be circular, or at least lacks independent empirical support. No example I’m aware of indicates propositional subject-predicate structures in any creature save language-competent humans. Even that stellar bonobo, Kanzi, consistently fails to apprehend such structures; his novel “sentences” are maximally two concatenated arguments with no subjects, such as “chase bite,” which humans shun (Atran & Lois 2001). So, P&B’s proposal may reduce to: language was naturally selected to communicate what only language can formulate (propositions).


Strong and weak adaptationists accept natural selection as the only known (noncultural) explanation for functionally complex design. One possibility consistent with this is that much complex design has no presently known explanation (most human cognitive architecture, Fodor, 2001), and there may be some functional complexity that results largely from more general physical, chemical or biological processes governing complex systems. One alternative, weak adaptationist approach assumes no direct natural selection (no task-specific adaptation to distinctive features of ancestral environments) for language’s “creative core,” that is, the faculty of syntactic recursion that allows potentially infinite production of words and well-formed word-combinations with relatively few and finite means (Chomsky 2000). Putting aside the argument from design as too vague or nearly circular, this “minimalist program” operates on the (huge but bold) assumption that language’s creative core is a recently evolved accommodation to more general physical or biological processes - in ways analogous to the apparent optimization of information flow in a material medium through minimization of “wire length,” as in microchip design, nematodes and human brains (Cherniak 1995).


Perhaps recursion in language is a physically optimal sort of interface (internal accommodation) between two physically suboptimal (but perhaps genetically optimal and adapted) systems of more ancient evolutionary origins: the sensorimotor system (including phonation) and the conceptual-intentional system (including categorization, reference and reasoning). The idea of physical optimality has a distinguished tradition in science generally (Galileo, Newton, Einstein), and in studies of biological form and development (Maynard Smith et al. 1985). For evolutionary biology in particular, the primary objective is to discover and predict, through strictly physical and chemical means, the set of organic forms (molecular, morphological, neuronal) that are likely to emerge from a given starting point. Only then is it worthwhile to inquire into which of those forms might be selected and how. For example, extensive sharing of genomic structure among all vertebrates, and even vertebrates and invertebrates, suggests that many of the same “master genes” program body plan and the control mechanisms of development (Gehring, 1998). Even eyes, which were thought to have evolved analogously and independently in different phyla, may be in each case a homologous derivation from the same DNA (Pax-6). Physical law and mechanical processes appear to be responsible for much of what follows: development of each component of the eye is narrowly constrained by the laws of optics and mechanical contingencies involved in sharply projecting images of three-dimensional objects onto a planar surface of receptors.


In line with Turing’s (1952) vision of biological explanation, much the same organic architecture and behavior may evolve in very different historical environments just as basically similar cognitive architectures and behaviors may be developed in very different physical media (cf. Leiber 2001). If so, then it is plausible to attempt to explain significant aspects of the structure and emergence of these architectures and behaviors without considering how they have been accommodated to (selected for) particular historical environments and physical media. Indeed, further understanding of particular historical and physical accommodations (e.g., the “Cambrian explosion” of multicellular organisms, the “real-time” processing of information) may depend crucially on such nonteleological insights.


Worthiness of this approach depends on success in providing significant and surprising predictions and discoveries. In the minimalist program, these arguably (if controversially) far exceed what its originators previously thought possible. At best, strong adaptationist arguments retrodict old discoveries. This isn’t to deny that adaptationist arguments may ultimately prove insightful into language structure: recent studies identifying multiple genetic loci for language disorders and delays seem to belie any monomutational account for language, and at least one of these genes (FOXP2) seems to have been a target for selection (although this gene concerns speech and processing of morphology, not syntactic recursion, Enard et al., 2002). Nevertheless, novel biological and evolutionary understanding of language (and other cognitive structures) may occur beforehand.


Weak adaptationist (though not necessarily minimalist) investigation of language crucially uses aspects of the strong adaptationist program, especially the comparative approach (Hauser et al. in press). Thus, arguments for natural selection of phonation have involved claims about the uniqueness of categorical auditory discrimination and descent of the larynx in humans. Comparative studies prove otherwise: chinchillas and other mammalian species categorically discriminate human phones; deer and several bird species drop the larynx (possibly to exaggerate size, Fitch & Reby 2001). Perhaps human phonation is itself the by-product of a jury-rigged combination of other by-products and adaptations: the (originally pre-vertebrate) alimentary system and the respiratory system of terrestrial vertebrates interface at the larynx (which drops in humans), hence by chance enabling the production of phones later “exapted” to human language.


Other comparative studies show contrary evidence for prehominid antiquity in parts of the conceptual-intentional system. Intriguing experiments showing subordinate chimps taking the perspective of dominant chimps (Hare et al. 2001) have yet to be replicated in different laboratories (Povinelli 2001). Apparently, chimps can’t repeatedly embed states of mind: [Danny thinks that [Marc believes that [Brian knows that… Etc. Short-term memory typically limits iterated embedding of mental states to 5-7 levels (Barrett et al. 2002); however, as with “center-embedding” of linguistic clauses, computational machinery allows for indefinitely many embeddings (to any apparent limit, add: “You really think that…”). By giving a person more time and external memory, more embedding is interpretable in a unique and uniform way (not predicted by associationist models, connectionist or other). Other parts of the conceptual-intentional system may be more ancient in primates, including perceptually-based reference (Gallistel 1990), categorization (Brown & Boysen 2000) and reasoning (Povinelli 2000).


In sum, combining strong and weak adaptationist strategies might profitably generate evolutionary insights into human cognition. Viewing progress in understanding the emergence of human cognition only (or principally) either through a lens of strong adaptationism or weak adaptationism could lead science into blind alleys. “Muddling through” with modest adaptationism may prove most effective.


Acknowledgements. Thanks to Noam Chomsky, Dan Sperber, Douglas Medin, David Hull for suggestions, and to Marc Hauser, Steven Pinker, Justin Leiber for information.



Atran, S. (2001) A cheater detection module? Evolution and Cognition 7:183-193.


Atran, S. & Lois, X. (2001) Reply to Shanker. Current Anthropology 42:498-500.


Atran, S. (2002) In gods we trust: The evolutionary landscape of religion. Oxford University Press.


Atran, S., Medin, D., Lynch, E., Vapnarsky, V., Ucan Ek’, E. & Sousa, P. (2001). Folkbiology doesn’t come from folkpsychology. Journal of Cognition and Culture 1:3-42.


Au, T. & Romo, L. (1999). Mechanical causality in children’s « folkbiology. » In Folkbiology , ed. D. Medin & S. Atran. MIT Press.


Barrett, L., Dunbar, R. & Lycett, J. (2002) Human evolutionary psychology. Princeton University Press.


Brown, D. & Boysen, S. (2000) Spontaneous discrimination of natural stimuli by chimpanzees. Journal of Comparative Psychology 114:392-400.


Carey S. (1985). Conceptual Change in Childhood. MIT Press.


Cherniak, C. (1995) Neural component placement. Trends in neuroscience 18:522-527.


Chomsky, N. (2000) Minimalist inquiries. In Step by step, ed. R. Martin, D. Michaels & J. Uriagereka. MIT Press.


Enard, W., Przeworski, M., Fisher, S., Lai, C., Wiebe, V.,  Kitano, T., Monaco, A. & Pääbo, S. (2002) Molecular evolution of FOXP2, a gene involved in speech and language. Nature 418:869 – 872.

Fitch, W.T. & Reby, D. (2001) The descended larynx is not uniquely human. Proceedings of the Royal Society, Biological Sciences 268:1669-1675.


Fodor, J. (2001) Review of Evolution of the human mind. British Journal of the Philosophy of Science 52:623-628.


Hare, B., Call, J., & Tomasello, M. (2001) Do chimpanzees know what conspecifics know? Animal Behaviour 61:139-151.


Gallistel, C. R. (1990 The organization of learning. MIT Press.


Gehring, W. (1998) Master control genes in developmental evolution.Yale University Press.


Hauser, M., Chomsky, N. & Fitch, W. T. (in press) The faculty of language. Science.


Leiber, J. (in press) Philosophy, engineering, biology, and history: A vindication of Turing’s views about the distinction between the cognitive and physical sciences. Journal of Experimental and Theoretical Artificial Intelligence.


MacDonald, K. (1998) Separation and its discontents: Toward an evolutionary theory of Anti-Semitism. Praeger.


Maynard Smith, J., Burian, R., Kaufman, S., Alberch, P., Campbell, J., Goodwin, B., Lande, R., Raup, D. & Wolpert, L. (1985) Developmental constraints in evolution. Quarterly Review of Biology 60:265-287.


Medin, D. & Atran, S. (in press) The native mind. Psychological Review.


Pinker, S. & Bloom, P. (1990) Natural language and natural selection. Behavioral and Brain Sciences 13:707-727.


Povinelli, D. (2000) Folk physics for apes. Oxford University Press.


Povinelli, D. (2001) Chimpanzee theory of mind and folk physics. Paper presented at “Innateness and Structure of Mind” Workshop, University of Sheffield, November.


Ross, N., Medin, D., Coley, J. & Atran, S. (in press) Cultural and experiential differences in the development of folkbiological induction. Cognitive Development.


Sober, E. & Wilson, D.S. (1998) Unto others. Harvard University Press.


Sperber, D. (1996) Explaining culture. Blackwell.


Turing, A. (1952) The chemical basis of morphogenesis. Philosophical Transactions of the Royal Society B237:37-72.


Wilson, D.S. (2002) Darwin’s cathedral. University of Chicago Press.