Papex

Open problem · Neuroscience · Open

Modularity of mind

Is the mind made up of distinct, specialised modules, each evolved for a particular task?

Summary

Some mental processes appear to run as specialised, partly independent systems, but how far the mind is divided into innate modules, and even what should count as a module, is disputed.

Why it is open

The modern debate starts with Jerry Fodor's book The Modularity of Mind (1983) [1]. Fodor listed nine features of the systems that interested him, among them domain specificity, mandatory and fast operation, a fixed neural architecture and characteristic patterns of breakdown [15]. A system counts as modular if it has most of them to an appreciable degree, and informational encapsulation, being unable to use information stored elsewhere in the mind, is "more or less essential" [15]. Fodor argued that input systems such as perception and language are modular, but that the central systems that fix beliefs are not, because in principle anything one knows can bear on what else one should believe [15].

The massive modularity hypothesis, advanced by evolutionary psychologists including Leda Cosmides and John Tooby and defended most fully by Peter Carruthers in 2006, holds that the mind is modular "through and through", including high-level cognition [15]. It also changes what a module is. Carruthers drops encapsulation, and H. Clark Barrett and Robert Kurzban define modules as functionally specialised mechanisms [15, 5].

The problem is therefore partly about definitions. Philip Robbins and Zoe Drayson note the worry that, with no consensus on which features make a system modular, the two sides may be "talking past each other" [15]. David Pietraszewski and Annie Wertz argue that the massive modularity debate is "ill-posed and confused", because it mixes up levels of analysis [14]; in 2024 Jonathan Egeland disputed their account of what each side believes [15].

No body sets a bar for a resolution. The empirical questions that remain are those Nancy Kanwisher called "as-yet unanswered" in 2010: how much of the mind and brain is made up of functionally specialised components, and how they arise in development [9].

What is known

Reasoning about exchange (2010). Cosmides and Tooby reasoned that exchange can evolve only if cooperators can detect cheaters, and so predicted a system specialised for reasoning about social exchange [8]. People do poorly when asked to find violations of most conditional ("if … then") rules, but much better when the rule specifies an exchange and a violation is cheating [8]. Critics proposed that the advantage extends to all rules about permitted actions; in 2010 Cosmides, Barrett and Tooby reported experiments in which such rules as a class did not elicit a search for violations, and in which people looked for violations of exchange rules only when these were likely to reveal someone intending to cheat [8]. Robbins and Drayson note that deficits in social exchange reasoning appear to come with other social-cognitive impairments, and find "little in the way of neuropsychological evidence of high-level modularity" [15].

Face areas (1997–2006). With functional magnetic resonance imaging (fMRI), which tracks changes in blood flow linked to neural activity, Kanwisher, Josh McDermott and Marvin Chun found a region in the fusiform gyrus of 12 of 15 people that responded more to faces than to common objects and passed further tests of face specificity; they called it "a module in human extrastriate cortex specialized for face perception" [2]. Isabel Gauthier and colleagues found that bird and car experts recruit the same area for their category of expertise, and proposed that "level of categorization and expertise, rather than superficial properties of objects", determine its specialisation [3]. James Haxby and colleagues could identify which category a person was viewing from activity patterns in ventral temporal cortex even after excluding the regions that responded most to it, and concluded that representations of faces and objects are "widely distributed and overlapping" [4]. In two macaque monkeys, Doris Tsao and colleagues recorded single neurons in the largest face-selective region found by fMRI: 97% of the visually responsive neurons were strongly face selective [6].

How far specialisation goes (2010–2013). Kanwisher reviewed regions selective for faces, places, bodies and visually presented words, and one engaged when thinking about another person's thoughts; each of the first three responds about twice as strongly to its preferred category as to any other [9]. Screening 20 further object categories, from tools and food to musical instruments, her group found no new category-specific regions [9]. Evelina Fedorenko, Michael Behr and Kanwisher, defining language regions individually in each person, found little or no response in them to arithmetic, working memory, cognitive control or music [10]. Other regions are general: in frontal and parietal regions known as the multiple-demand network, Fedorenko, John Duncan and Kanwisher found the same voxels (the small volumes fMRI measures) engaged by seven diverse demanding tasks, and these regions often lie directly next to domain-specific ones [11].

Origins (2010–2022). Reading is a test case. Humans have read for only a few thousand years, which Kanwisher notes is not thought long enough for a complex structure to evolve, yet a small region selective for written words exists; its selectivity, she concludes, "is derived from the individual's experience, not their genes" [9]. Michael Arcaro and colleagues raised three macaques without any sight of faces: they did not develop face patches, although they developed domains for other categories, so experience must be necessary for the formation, or maintenance, of face domains [12]. In human infants aged 2 to 9 months, Heather Kosakowski and colleagues found face-, scene- and body-selective responses in the locations of the adult areas [13].

Main approaches

Massive modularity. The mind consists largely of specialised mechanisms shaped by selection for recurrent adaptive problems [15]. Strongest evidence: the exchange experiments, in which the search for violations follows the likelihood of a cheater rather than rule content in general [8]. Main difficulty: after weighing the general arguments for it, from design, from animal minds and from computational tractability, Robbins and Drayson conclude that compelling ones "are hard to come by", and note the "paucity of empirical evidence directly supporting the hypothesis" [15].

Modest, Fodorian modularity. Input systems are modular and central cognition is not [15]. Strongest evidence: illusions such as the Müller-Lyer illusion persist after one has measured the lines and knows them to be equal [15]. Main difficulty: cross-modal effects such as the double-flash illusion, and context effects in speech perception, tell against encapsulation, and regions once thought dedicated are recruited for other tasks [15, 3].

Functional specificity in the brain. Some functions have their own cortical regions, found in more or less the same place in virtually every neurologically intact person [9]. Strongest evidence: face, place, body, word and language regions with large selectivity [9, 10], and a macaque face region made almost entirely of face-selective cells [6]. Main difficulty: few such regions have been found [9]; category information is also spread across other regions [4]; and a selective region need not be innate [9, 12]. Arcaro and colleagues propose that domains form on an inborn map of the visual field through viewing behaviour, without category-specific templates [12]. Whether specialised regions run one shared computation on different inputs is the subject of the Papex page Canonical computation of cortical columns.

Neural reuse. Michael Anderson argues that circuits established for one purpose are often reused, in evolution or development, for new purposes, often without losing their original functions [7]. Strongest evidence: most cortical regions of any size are engaged in tasks across different domains, as Robbins and Drayson summarise Anderson's case [15]. Main difficulty: when language regions are defined in each individual, they respond little to the non-linguistic tasks earlier imaging had linked to them [10].

Open sub-questions

  1. A sensitive period for faces. If face-deprived macaques first see faces later in life, do face domains form at the usual locations, and up to what age [12]?
  2. Before 2 months. Are face-selective responses present in human infants younger than those Kosakowski et al. scanned, and does their strength track how long each infant has looked at faces [13, 12]?
  3. Cheater detection in the brain. Does reasoning about social exchange engage a region selective for it, separable from the region used for thinking about others' thoughts [8, 9]?
  4. Where general meets specific. Do multiple-demand regions differ in their connections from the domain-specific regions that lie next to them [11]?
  5. One experiment, two definitions. Can a candidate module be tested so that functional specialisation and informational encapsulation predict different results [14, 15]?
  6. Language at the start. Is the language network's selectivity [10] present when infants begin to learn language? See the Papex page How infants acquire language.

Work on this in Papex

If you have a testable answer to one of the sub-questions above, you can draft it as a private hypothesis in Papex, or prepare the draft from Claude or ChatGPT with the Papex connector. You review the exact version, sign it and decide whether to publish it; a public record that names this problem is listed on this page. A Papex record shows who signed which version and when. It is not peer review and does not mean the hypothesis is correct.

Key references

  1. Fodor, J. A. (1983). The Modularity of Mind. MIT Press (Bradford Books). 10.7551/mitpress/4737.001.0001 · Checked 15 Oct 2026
  2. Kanwisher, N., McDermott, J. & Chun, M. M. (1997). The fusiform face area: a module in human extrastriate cortex specialized for face perception. Journal of Neuroscience 17(11), 4302–4311. 10.1523/JNEUROSCI.17-11-04302.1997 · Checked 15 Oct 2026
  3. Gauthier, I., Skudlarski, P., Gore, J. C. & Anderson, A. W. (2000). Expertise for cars and birds recruits brain areas involved in face recognition. Nature Neuroscience 3(2), 191–197. 10.1038/72140 · Checked 15 Oct 2026
  4. Haxby, J. V., Gobbini, M. I., Furey, M. L., Ishai, A., Schouten, J. L. & Pietrini, P. (2001). Distributed and overlapping representations of faces and objects in ventral temporal cortex. Science 293(5539), 2425–2430. 10.1126/science.1063736 · Checked 15 Oct 2026
  5. Barrett, H. C. & Kurzban, R. (2006). Modularity in cognition: framing the debate. Psychological Review 113(3), 628–647. 10.1037/0033-295X.113.3.628 · Checked 15 Oct 2026
  6. Tsao, D. Y., Freiwald, W. A., Tootell, R. B. H. & Livingstone, M. S. (2006). A cortical region consisting entirely of face-selective cells. Science 311(5761), 670–674. 10.1126/science.1119983 · Checked 15 Oct 2026
  7. Anderson, M. L. (2010). Neural reuse: a fundamental organizational principle of the brain. Behavioral and Brain Sciences 33(4), 245–266 (published online 2010). 10.1017/S0140525X10000853 · Checked 15 Oct 2026
  8. Cosmides, L., Barrett, H. C. & Tooby, J. (2010). Adaptive specializations, social exchange, and the evolution of human intelligence. Proceedings of the National Academy of Sciences 107(Supplement 2), 9007–9014. 10.1073/pnas.0914623107 · Checked 15 Oct 2026
  9. Kanwisher, N. (2010). Functional specificity in the human brain: a window into the functional architecture of the mind. Proceedings of the National Academy of Sciences 107(25), 11163–11170. 10.1073/pnas.1005062107 · Checked 15 Oct 2026
  10. Fedorenko, E., Behr, M. K. & Kanwisher, N. (2011). Functional specificity for high-level linguistic processing in the human brain. Proceedings of the National Academy of Sciences 108(39), 16428–16433. 10.1073/pnas.1112937108 · Checked 15 Oct 2026
  11. Fedorenko, E., Duncan, J. & Kanwisher, N. (2013). Broad domain generality in focal regions of frontal and parietal cortex. Proceedings of the National Academy of Sciences 110(41), 16616–16621. 10.1073/pnas.1315235110 · Checked 15 Oct 2026
  12. Arcaro, M. J., Schade, P. F., Vincent, J. L., Ponce, C. R. & Livingstone, M. S. (2017). Seeing faces is necessary for face-domain formation. Nature Neuroscience 20(10), 1404–1412. 10.1038/nn.4635 · Checked 15 Oct 2026
  13. Kosakowski, H. L., Cohen, M. A., Takahashi, A., Keil, B., Kanwisher, N. & Saxe, R. (2022). Selective responses to faces, scenes, and bodies in the ventral visual pathway of infants. Current Biology 32(2), 265–274.e5. 10.1016/j.cub.2021.10.064 · Checked 15 Oct 2026
  14. Pietraszewski, D. & Wertz, A. E. (2022). Why evolutionary psychology should abandon modularity. Perspectives on Psychological Science 17(2), 465–490 (published online 2021). 10.1177/1745691621997113 · Checked 15 Oct 2026
  15. Robbins, P. & Drayson, Z. (2025). Modularity of mind. The Stanford Encyclopedia of Philosophy, first published 1 April 2009, substantive revision 8 July 2025. Source · Checked 15 Oct 2026

Page reviewed 15 Oct 2026

Source

Open, as listed by Wikipedia · Unsolved problems in neuroscience. Checked 08 Oct 2026.

The status is the source’s. Papex does not decide whether a problem is solved.

Records on this problem

No public record addresses this problem yet.

More in Life sciences