Against an Evolutionary Account of Modularity in Peripheral Cognitive Abilities

A common way of looking at the mind is to say that it is modular, i.e. divided into units, each responsible for a different cognitive ability or domain. A criticism that can be made against this idea comes from its connection to evolutionary psychology. The genetic predetermination of modules as explained by evolutionary pressures is not easily squared with brain plasticity. Here I defend modularity by trying to separate it from evolutionary psychology.


Fodor’s non-classical account (1983) stipulates that a module is a 1) domain specific, 2) information encapsulated, 3) innate cognitive structure with its own input and output transducers, which makes it inaccessible to the rest of cognition and efficient in operation (in Carruthers, 2003: 294; also Pinker, 1997: 21; Buller & Hardcastle, 2000: 308; Tooby & Cosmides, 1997; Currie & Sterelny, 2000: 147, 154; Samuels, 1998: 578). For such a system to be 1) domain specific, its computations, inputs and outputs should be unique to the system, such that its operations are not had by any other parts of the mind (Carruthers, 2003: 295). Its 2) encapsulation component tries to capture that there ought to be independent information processing inside the module while it has no access to the entirety of information in the human brain (Carruthers, 2003: 295; Buller & Hardcastle, 2000: 309). For Fodor, this was the most salient feature of modularity in the way that he originally constructed the idea to describe vision (1983: 37), which is information encapsulated since its operation is unaffected by conflicting beliefs from the rest of cognition (Currie & Sterelny, 2000: 148). Recently it has been argued that this account can be expanded to include all aspects of cognition, including theory of mind, cheater detection, central reasoning and consciousness (see Carruthers’ 2003 “global broadcasting hypothesis”). As a result, the modularity account comes in different modal strengths contingent on the number of cognitive processes we want to attribute to it, ranging from non-modular and entirely domain-general architectures of the mind, via Fodor’s moderate view where only some peripheral cognitive systems are modular, to “massively modular“ minds which are entirely made up of these constructs.

For my purposes, I shall concentrate on peripheral cognitive abilities, such as speed estimation, mate selection or social cognition, that are sophisticated and non-sensory, but not central to cognition as reasoning or consciousness are. Let me give an example: One can argue that social cognition includes a module that is 1) domain specific by providing “intentional markers to the inputs which help to fix belief” unique to the system (Currie & Sterelny, 2000: 158; see Tooby & Cosmides, 1992: 42). It can also be argued to be 2) information encapsulated since, for instance, stage acting produces empathy, even if it is in conflict with beliefs about the situation not being real (Segal, 1996: 147; in Currie & Sterelny, 2000: 150, who disagree). Furthermore, poverty-of-the-stimulus type arguments posit that an 3) innate social cognition device must be present in infants since their abilities exceed what can be expected from an inductive learner. Worse than that, inferential learning as well as classical conditioning might themselves be subserved by a distinct computational module (Gallistel, 2000; in Carruthers, 2003: 306). In essence, modules are “isolable functional sub-components” (Carruthers, 2003: 295) that analogously to the pieces of a Hi-Fi system are “task-specific information processing devices” (Atran, 2001: 8, see Samuels, 1998: 577).

There is strong empirical evidence in favour of this peripheral modularity. For instance, many cognitive functions are prima facie brain-localisable, such as social cognition, which is strongly associated with the right temporo-parietal junction of the neocortex (Saxe & Wexler, 2005). There is pathological evidence from patients with brain lesions that only suffer loss to cognitive functions associated with the damaged region, such that evidently “everything dissociates from everything else” (Shallice, 1988; in Carruthers, 2003: 304). Patients with full low IQs, or those suffering from Williams syndrome can still perform well in social cognition tasks (Samuels, 1998: 596), while on the other hand there are people with autism, for whom general reasoning may function normally while social cognition is impaired. Interestingly, Williams disease is caused by the deletion of about 26 genes that results measurable distortion in the right parietal and left frontal cortex (Donnai & Karmiloff-Smith, 2000), which arguably points towards modules being genetically 3) innate.

In the literature, modularity theories and evolutionary psychology usually come hand in hand. A paradigmatic example is Carruthers’ characterisation of modularity, which relies heavily on the notion of fitness in our evolutionary past (Carruthers, 2003: 299, 300, 301, 306; see also Tooby & Cosmides, 1992, 1997; Thornhill, 1992). His argument goes something like this: P1) In the past, there was some problem-based evolutionary pressure. P2) An evolved specific, reliable, fast-solution module gives a fitness advantage to the organism that possesses it in countering that pressure. P3) Fitness advantages over time become population general. P4) All evolution proceeds this way in countering pressures with specific solutions. (Please note that it is not clear whether Carruthers would embrace such a strong stance on empirical adaptationism, rather an explanatory account of evolution that states that it is the first best explanation) C) Therefore the mind has a modular architecture. In the same way that there is no “general-purpose sensory organ” (Carruthers, 2003: 300), we ought to expect the mind to be separated task-specifically in order to be efficient and survival advantageous. More drastically,  since natural selection is allegedly the the only serious candidate for giving explanation for functional complexity (Dawkins, 1986; in Carruthers, 2003: 294), evolutionary psychologists might argue that because of the conjunction of P2) and P4), evolution must result in modular architectures (Tooby & Cosmides, 1997; in Woodward & Cowie, 2003: 313; also Carruthers, 2003: 293). I have previously mentioned that I shall follow Samuels in his characterisation of innate features as those who have no psychological explanation and are primitive (2002). In this sense modularity theorists that cling to this account of evolutionary psychology will not deny the importance of environmental inputs in bringing about the modular architecture, however, they are merely “triggers […] of a developmental programme”, that is already genetically specified (Tooby & Cosmides, 1992: 82, 87).

In its commitment to this pre-specification of modules, the theory has faced harsh criticisms from neuroscientists of plasticity. Those argue that the premise that brain architecture is genetically encoded for and already coarsely specified at birth is simply wrong in light of evidence of the brain’s capability to undergo drastic changes during all stages of life. Specifically, areas of the brain thought responsible for certain functions can be altered with their inputs in higher vertebrae to gain new functional characterisations (Elman et. al., 1996: 26; in Samuels, 2002: 260). Alternatively, the auditory cortex can be “recruited” for sign-language processing in deaf subjects (Nishimura et. al.: 1999, in Woodward & Cowie, 2003: 319). These and many other studies seem to suggest that modules cannot be genetically specified (Samuels, 2002: 262), since an innate structure assigned to a specific brain region should either be permanently destroyed by a lesion, or by not being used for its innate purpose simply lie dormant. That it is not, counts as evidence against modularity for Panksepp & Panksepp, who assert that this plasticity shows how function in brain areas is assigned contingent on requirements for integrating information, thereby violating the evolutionary aspect of modularity theory. Let me give an example: The creative aspects of speech production may be located in Wernicke’s area, not because it is a module, but because that bit of the neo-cortex is the most “multimodal area for integrating information” (2000: 111). Consequently, even if there is a degree of modularisation, it comes as a result not of natural selection, but instead of developmental plasticity that shapes the brain according to its life-time needs (Buller & Hardcastle, 2000).

Let me summarise where we are now. We have started by looking at cognitive modules as structures that are 1) domain specific, 2) encapsulated and 3) innate and thereby genetically specified, and have looked at empirical evidence for such modules in peripheral abilities. Then we took note of the interrelation between modularity and evolutionary psychology in bringing about 3), and motivated the intuition of modularity by arguing that it is the best evolutionary explanation for the complexity of the mind.  This intuition we saw challenged by evidence from plasticity. Let me now try to argue that that this line of reasoning only attacks the evolutionary component of modularity theory and that maybe we ought to separate ourselves from it in order to conserve the theory as a whole.

Allow me to start by elaborating on my indisputable premise that if a trait stems from natural selection, then it is innate and must be genetically specified (Pinker, 1997: 21; in Buller & Hardcastle, 2000: 308, 319). Therefore what is an innate module must have a basis in DNA, which seems highly unlikely considering the relatively small number of about 50 000 genes in relation to the massive number of cortical synaptic connections, especially considering that their number in the prefrontal cortex increases throughout life                       (Gould et. al., 1999; in Buller & Hardcastle, 2000: 311). Even considering the genetic “combinatorial possibilities” (Woodward & Cowie, 2003: 319) through introns and exons of given segments in the genome that increases the number of genes by an order of magnitude, there seems to be blatant genetic underdetermiaton of brain activity. I am with Buller and Hardcastle who think that this is particularly dramatic considering that genes mostly code for proteins and hence genetic influences must go through molecular structure to affect cognitive modules (2000: 319; also Fodor 2000: 88). I call this the reverse-poverty-of-the-stimulus argument that shows how the high level of sophistication in cognitive abilities is unlikely to be genetically coded for by protein specification alone. Carruthers might immediately respond to this line of criticism by invoking the “triggering” analogy that aims to show the importance of environmental input in bringing about modules, however, this may be misguided since the environment in neural development actively shapes the mind and its modules through its interaction (ibid: 316), rather than serving as alarm-clock to when a module needs to be formed.

There is one more avenue that proponents of modularity might take to recover their 3) innateness component against these critiques: Epigenetics posits that structural changes in the genome are still possible post conceptionem throughout life through environmental inputs. For instance, attachment behaviour in infants as function of parental care can leave measurable genetic variations. It is likely that in fact             many behavioural tendencies and neuronal phenomena “emerge as functions of individual experiences” (Panksepp & Panksepp, 2000: 110). A brain lesion or loss of sensory input that triggers reassignment of brain-regions to cognitive abilities may hence be an epigenetic change in the genome that alters Tooby & Cosmides’ (1992) pre-specification of brain localisation previously attacked by plasticity. Modules could hence be “epigenetically derived” according to life-long environmental input (Panksepp & Panksepp, 2000: 110). Unfortunately, I find this speculative proposal unpromising. Although prima facie intuitive, I doubt whether epigenetic alterations can bring about plasticity for the following reason: The epigenetic process of methylation appealed to in the literature simply turns genes on and off, thereby not creating any new information. How this epigenetic activation of protein creation is supposed to explain the movement of sign-language understanding facilities to the auditory cortex on a molecular level is highly conjectural and still suffers from the same underdetermination that regular genetics did.

In essence, if innate traits are genetically specified, and modularity is 3) innate, then it is in serious jeopardy, since genetic explanations for plasticity suffer from poverty in giving an account for plasticity. Let me argue that the way out of this predicament for the modularity theorist is to abandon 3) and move to a mostly environment-shaped mind that is modular in character, even if not pre-specified. We can then still argue that a moved auditory system is 1) domain-specific and 2) encapsulated, which is enough to bring about Fodorian modularity. To underline the motivation for this move, let me bring forward a few more  general objections against the evolutionary component of modularity.

The assertion that “form follows function” (Tooby & Cosmides, 1997: 13) is simply wrong in light of modern evolutionary theory that posits “spandrels” as the architectural feature of San Marco necessary to accommodate arcs and dome, now used for storage, but never designed for that function and rather was a “necessary concomitant to another decision in design” (Atran, 2001: 4; see Gould and Lewontin, 1979). Similarly, many cognitive abilities could be such spandrels and Carruthers’ assumption in P2) and P4), that a single problem will always lead to the evolution of a new, distinct single solution, is at odds with how evolution operates (Woodward & Cowie, 2003: 316). There are rarely entire new systems “bolted on” (Carruthers, 2003: 300), and the a priori attempt by evolutionary psychologists to give a story for the adaptation value of an ability in the past (see Thornhill, 1992 for a story on rape; in Woodward & Cowie, 2003: 318) is in many cases deeply misguided (Gould 1997: 10754). For instance, language might have evolved as spandrel of extra computational space available without evolutionary pressure (Panksepp & Panksepp, 2000: 111; on Clark, 1997). The project that Carruthers pursues in motivating modules consequently suffers from “just-so-story” critiques (Schlinger, 1996) and “creative excess” (Panksepp & Panksepp, 2000: 109). The past was not always a Carruthers-type fitness-studio and the backward engineering he appeals to is not an effective empirical argument for modularity.

Moreover, there is no consensus as to whether evolutionary pressure would at all have favoured a modular architecture, or that in P2), modularity is cognitively superior in computational capabilities (Samuels, 1998: 587). It is perfectly conceivable that evolutionary pressure for flexible, creative, large scale information assimilation resulted in a non-modular architecture (Sober, 1994; in Woodward and Cowie, 2003: 313). In fact, that is no less likely (Symons, 1992: 142; in Buller & Hardcastle, 2000: 310) and there is no biological contradiction in asserting either modular peripheral systems or their domain-general counterparts were chosen (Samuels, 1998: 598).

Quintessentially, the modularity theorist’s appeal to evolutionary psychology might backfire quickly. For all those reasons, I propose giving up a strong stance on modularity’s 3) innateness component. Still, the problems for modularity in peripheral systems continue especially as worries about 2) information encapsulation in light of plasticity. Many have argued that we need to look at Marr’s computational level, rather than neural correlates for information encapsulation (Tooby & Cosmides, 1992: 64; in Buller & Hardcastle, 2000: 318; also Woodward & Cowie, 2003: 325; Samuels, 1998: 580), and hence give up the so far hidden premise in my argument that modules correspond to specific brain regions. The next step is developing an account without 3) innateness to incorporate that computational. That task however must be left for another post.


Atran, S. 2001. The Case for Modularity: Sin or Salvation? Evolution and Cognition 7: 1-10.

Buller, D. J. and Hardcastle, V. G. 2000. Evolutionary Psychology, Meet Developmental Neurobiology: Against Promiscuous Modularity. Brain and Mind 1: 307-325.

Carruthers, P. 2003. The Mind is a System of Modules Shaped by Natural Selection. In Hitchcock, C. R. Ed. Contemporary Debates in Philosophy of Science. New York. Blackwell.

Chomsky, N. 1967. Recent Contributions to the Theory of Innate Ideas. Synthese 17: 2-11.

Clark, A. 1997. Being There: Putting Brain, Body and World Together Again. Massachusetts. MIT University Press.

Currie, G. and Sterelny, K. 2000. How to Think about the Modularity of Mind-Reading. The Philosophical Quarterly 199: 145-160.

Dawkins, R. 1986. The Blind Watchmaker. New York. Norton.

Donnai, D. and Karmiloff-Smith, A. 2000. Williams Syndrome: From Genotype Through to the Cognitive Phenotype. American Journal of Medical Genetics 97: 164-171.

Dupré, J. 2001. Human Nature and the Limits of Science. Oxford. Oxford University Press.

Elman, J., Bates, E., Johnson, M., Karmillof-Smith, A., Parisi, D. and Plunkett, K. 1996. Rethinking Innateness: A Connectionist Perspective on Development. Massachusetts. MIT University Press.

Fodor, J. 1983. The Modularity of Mind. Massachusetts. MIT University Press.

Fodor, J. 2000. The Mind Doesn’t Work That Way: The Scope and Limits of Computational Psychology. Massachusetts. MIT University Press.

Gallistel, R. 2000. The Replacement of General-Purpose Learning Modules with Adaptively Spcialised Learning Modules. In Gazzaniga, M. Ed. The New Cognitive Neurosciences. Massachusetts. MIT University Press. pp. 179-191.

Gould, S. J. and Lewontin, R. C. 1979. The Spandrels of San Marco and the Panglossian Paradigm: A Critique of the Adaptationist Program. Proceedings of the Royal Society of London B205: 581-598.

Gould, S. J. 1997. The Exaptive Excellence of Spandrels as a Term and Prototype. Proceedings of the National Academy of Sciences USA 94: 10750-10755.

Gould, E., Reeves, A. J., Graziano, M. S. A. and Gross, C. G. 1999. Neurogenesis in the Neocortex of Adult Primates. Science 286: 584-552.

Marr, D. 1982. Vision. New York. Freeman.

Nishimura, H., Hashikawa, K., Doi, K., Iwaki, T., Watanabe, Y., Kusuoka, H., Nishimura, T. and Kubo, T. 1999. Sign Language “Heard” in the Auditory Cortex. Nature 397: 116.

Panksepp, J. and Panksepp, J. B. 2000. The Seven Sins of Evolutionary Psychology. Evolution and Cognition 6: 108-131.

Pinker, S. 1997. How the Mind Works. New York. Norton. Samuels, R. 1998. Evolutionary Psychology and the Massive Modularity Hypothesis. British Journal of the Philosophy of Science 49: 575-602.

Samuels, R. 2002. Nativism in Cognitive Science. Mind & Language 17: 233-265.

Saxe, R. and Wexler, A. 2005. Making Sense of Another Mind: The Role of the Right Temporo-Parietal Junction. Neuropsychologia 43: 1391-1399.

Schlinger, H. 1996. How the Human got its Spots: A Critical Analysis of the Just So Stories of Evolutionary Psychology. Skeptic 4: 68-76.

Segal, G. 1996. The Modularity of Theory of Mind. In Carruthers, P. and Smith, P. K. Eds. Theories of Theories of Mind. Cambridge. Cambridge University Press.

Shallice, T. 1988. From Neuropsychology to Mental Structure. Cambridge. Cambridge University Press.

Sober, E. 1994. From a Biological Perspective: Essays in Evolutionary Philosophy. Cambridge. Cambridge University Press. 50-71.

Symons, D. 1992. On the Use and Misuse of Darwinism in the Study of Human Behavior. In Barkow, J. H. Ed. The Adapted Mind: Evolutionary Psychology and the Generation of Culture. New York. Oxford University Press. 137–159.

Thornhill, R. and Thornhill, N. 1992. The Evolutionary Psychology of Men’s Coercive Sexuality. Behavioural and Brain Sciences 15: 363-421.

Tooby, J and Cosmides, L. 1992. The Psychological Foundations of Culture. In Barkow, J. H. Ed. The Adapted Mind: Evolutionary Psychology and the Generation of Culture. New York. Oxford University Press. 19–136.

Tooby, J and Cosmides, L. 1997. Evolutionary Psychology: A Primer. URL=[]. Accessed 27.02.2013.

Woodward, J., Cowie, F. 2003. The Mind is Not (Just) a System of Modules Shaped (Just) by Natural Selection. In Hitchcock, C. R. Ed. Contemporary Debates in Philosophy of Science. New York. Blackwell.