related texts

Comments on (MITECS), Culture, Cognition and Evolution

[ 17 Jul 2010 ] By now it is not online anymore. If you find here something that is interetsing and you want to see the actual text in MITECS, you can try to google part of the quote that I give. With some luck you can find the text online. Note also that these commenst are pretty old by now.

The text below is intended to be read side by side with the MITECS text. Indented text is always an exact quote from MITECS, and the user can use search to find the quote in the MITECS text. Where the quote ends with several dots, the comment is about all the paragraph starting with this text.

Note from the MITECS Executive Editor:

Keep in mind that the current mitecs site is a developmental, unedited site. The final site will be posted this spring.

This page contains comments on the Culture, Cognition and Evolution domain. Other pages contain comments on the other domains:


General Comments: There are some cases in these texts where the authors talk about evidence of things in the brain. Invariably, this is coming from overinterpretation of behavioural data. The anatomy and neurobiology of the brain is totally ignored.

The other point that is missing here is that learning is not discussed explicitly at all. Since cultural transmission can happen only by some process of learning, it is essential to understand learning to understand culture.


General: In section 2, the authors confuse between domain specific capabilities and evolutionary adaptations, and between abilities and devices. Domain specific capabilities can be learned, so cannot simply assumed to be evolutionary adaptations. In the same way, a specific ability can be based on a general purpose device, so specific abilities cannot be assumed to be based on specific devices.

The reason that evolutionary adaptations and devices seem unlikely is that the main organ of mental operations, the cerebral cortex, has a stochastic connectivity, and hence cannot contain devices or evolutionary adaptation which are much more complex than a single neuron. Therefore, any claim for evolutionary adaptation or a device must have a very strong supportive evidence before it can be taken seriously.

This approach (described either under the rubric of MODULARITY or DOMAIN-SPECIFICITY) seeks to investigate the nature and scope of these specific abilities, their evolutionary origin, their role in cognitive development, and their effect on culture.
Example of the confusion. Specific abilities don't necessarily have evolutionary origin.
The most important domain-specific abilities are evolved adaptations and are at work in every culture, though often with different effects.
How do the authors know this? This statement is not based on evidence.
The first domain-specific mechanisms to be acknowledged in the cognitive literature were input modules and sub-modules (see Fodor 1982).
Fodor did not 'acknowledge' input modules and sub-modules, he postulated them, without any supporting evidence. His argument was based on the assumption that cognition is a symbolic system, and input systems are not, so must be separated.
Typical examples are linked to specific perceptual modality. They include devices that detect edges, surfaces, and whole objects in processing visual information; face recognition devices; speech parsing devices; abilities to link specific outcomes (such as nausea and vomiting but not electric shock) to specific stimuli (such as eating but not light) through rapid, often single trial, learning.
There isn't any evidence for the existence of any of the devices in this list. There is evidence for the abilities, but that is different thing. You can have an ability to perform something, without having a specific device to do it.
More recently, there has been a growing body of evidence suggesting that central (i.e., conceptual) mechanisms, as well as input/output processes, may be domain-specific.........

These devices are described as providing the basis for competencies that children use to think about complex phenomena in a coherent manner using abstract causal principles. Cultural competencies in these domains are seen as grounded in these genetically determined domain-specific dispositions, though they may involve some degree of CONCEPTUAL CHANGE.

In this paragraph the, authors seems to do the confusion intentionally. They start with evidence for domain-specificity, and then switch to talk about devices and genetically determined dispositions, without trying to justify the switch. Either the authors intentionally mislead the reader, or they are incapable of considering learning.
Evolutionary psychology interprets these domain-specific competencies as evolved adaptations to specific problems faced by our ancestral populations.
Here the authors seem to realize that domain-specific competencies are not necessarily evolved adaptations. However, in the following discussion it is clear that the authors believe that once they can get over the cultural variety, these competencies can be assumed to be evolved adaptation without further evidence.
For instance, even in a case such as fear of snakes and other predators, where a convincing argument can be made for the existence, in many species, of evolved mechanisms that trigger an appropriate self-protection response, the danger cues and the fear are not necessarily directly linked. Marks and Nesse (1994: 255), following Mineka et al. (1984), describe such a case in which fear does not emerge instinctively but only after a specific sort of learning experience:
"Rhesus monkeys are born without snake fear. Enduring fear develops after a few observations of another rhesus monkey taking fright at a snake . . . Likewise, a fawn is not born with fear of a wolf, but lifelong panic is conditioned by seeing its mother flee just once from a wolf."
At least in the case of rhesus monkeys and snakes, this is a clear example of learning, rather than innate response. Rhesus monkeys can learn to fear any moving object in this way, and to less extent non-moving objects too. The authors here and of the original studies are simply blind to the possibility of learning.

The case of fawn and wolf is less clear, and I don't know enough about it.

Let us give three more illustrations -- the case of color classification, that of mathematics, and that of social classifications -- of the relationship between a domain-specific competence and a cultural domain.
As usual, these authors use the typical example of colour, which is different from other kinds of percepts in that we distinguish in already in the tranducers (cones in the retina).
It has been shown that human infants and some other animals can distinguish collections of objects according to the (small) number of elements in the collection. They also expect changes in the number of objects to occur in accordance with elementary arithmetic principles.
Strong overinterpretation of the data. Some of the behaviours of young animals and young humans can be interpreted, by biased researchers, as supporting these conclusions.
In North America and Europe one of the earliest-emerging social concepts is "race." Surprisingly, given the adult belief that the physical correlates of "race" are extremely attention-demanding, the child's initial concept of "race" contains little perceptual information. ....

Conceptual development of this sort -- in which specific concepts are acquired in a singular fashion and contain information far beyond what experience affords -- are plausibly the output of a domain-specific disposition................

An example of blatant nonsense. The way children interpret the word "race" is based on the way that older people around them use it (rather than what the older people believe about it) . That possibility, however, is too obvious for the authors, and they simply ignore it.
Children literally transform the contingent and incomplete cultural form into a noncontingent and fully articulated form. This happens because children are equipped with an evolved device for acquiring language (Bickerton 1990).
A blatant nonsense. It clearly happens because the children try to communicate with each other.


In the study of naive biology, disagreement arises over whether higher-order principles evince strong or weak NATIVISM; that is, whether they reflect the innate modularity and DOMAIN-SPECIFICITY of folkbiology (Inagaki and Hatano 1996), or are learned on the basis of cognitive principles inherent to other domains, such as NAIVE PHYSICS or FOLK PSYCHOLOGY (Carey 1995).
This is very misleading:
  1. It present as options only domain-specific principles and mechanisms (which may be from other domains), and ignores the possibility of generic mechanisms and principles.
  2. It presents 'innate modularity' and 'DOMAIN-SPECIFIC' as if these always go together. This is false, because domain specific rules can be learned by generic mechanisms.

Attribution Theory

General: The most important point to note about this section is how the researchers never lose their belief in the existence of an 'Attribution Theory', i.e. a theory that explains how all humans perform attribution. In other words, they are doing the 'sameness error' (Reasoning errors). When they find contradicting evidence, they try new theory, rather than admitting the possibility of variability between individuals.

Because humans are social animals, a person's survival and success depends greatly on the extent to which she can predict and influence the behavior of other persons. For this reason, it is human nature to watch other people closely, not merely recording another's observable actions but actively interpreting why the person behaved this way.
This does not explicitly states it, but it clearly intends to say that the 'human nature' is innate, and that the way most readers would read it. However, at the moment we don't have any evidence that this is innate, rather than learned.

Domain specificity

General: the most important question about these domains is whether they have innate domain specific features in the brain or not, because if there isn't than these domains are not reflections of human psychology, only of human environment. In this section, the author presents the distinction as if it is not significant. Most importantly, the way in which non-innate capabilities arise, i.e. learning, is totally ignored. The term 'learn' appears twice, once when referring to Chmosky's work, and once as part of the term 'animal learning'.

innate cognitive capacities in infants (Spelke 1994), evolutionary arguments (Cosmides and Tooby 1994),
See in Evolutionary Psychology, sections [3.2.2] and [5.1] for comments on the research of cognitive capacities in infants, and section [2] for comments on the evolutionary arguments.
Evidence for the status of syntax as a module was its innate, biologically driven character (evident in all and only humans), its neurological localization and breakdown (the selective impairment of syntactic competence in some forms of brain damage), its rapid acquisition in the face of meager environmental data (abstract syntactic categories are readily acquired by young children), and the presence of critical periods and maturational timetables (see Pinker 1994).
This is quite mazing heap of misleading statements.

1) There isn't any evidence for the ' innate, biologically driven character' of syntax.
2) There is no neurological localization of syntax.
3) There is no 'selective impairment of syntactic competence in some forms of brain damage'. There are cases of selective impairment, but these are not associated with any specific brain damage.
4) Children do not learn 'abstract syntactic categories'. They learn how to use language for communication, and that does not require abstract categories.
5) There is no critical period and maturation timetable for learning syntax. Adults learn syntax much better than children. It is phonetic skill (both comprehension and production) and lexical knowledge which are difficult for adults.

Some of these (2,3 and 5) may be explained if the author, when saying syntax, actually meant linguistic ability.

Fodor (1983) extended the logic of modules to cognitive abilities more broadly.....
None of Fodor's or Marr's ideas are based on any neurobiological observations. See here for a discussion of Marr's ideas.


General: This discussion concerns 'Altruism' in plants and animals, and is totally irrelevant to altruism in humans, because the latter is based, at least partly, on culture (i.e. individuals are altruistic because they learn that it is a positive attribute). It is relevant only if it is assumed that altruism in humans is also biological trait, and hence this text implies the latter without saying it. It is not obvious if this is intentional or not.


General: This is a good example of a 'detached interpreted model' error (reasoning errors). 'Dissonance' does not correspond to anything in the brain, and there is no way to measure it in any other way. Nothing can be predicted from this theory that is not already quite clear from simple common-sense. In short, it is totally useless.
Total Dissonance = (Dissonant Relations) / (Dissonant + Consonant Relations)
Obviously it is nonsense to express as a mathematical equation the relation between variables that cannot be measured.
Consider a prototypic case: a cigarette smoker, circa 1960, encountering the first medical reports linking smoking to lung cancer, emphysema, and heart disease.....
This and the following paragraphs are the 'interpretation' step of the 'detached interpreted model'. All the analysis can be done without a reference to dissonance, and the different paradigms are independent from each other. In short, 'Dissonance' is used as a buzz-word rather than actual theory.

Cooperation and competition

In addition to these purely behavioral mechanisms (perhaps the product of CULTURAL EVOLUTION), there is also evidence to suggest that there may be dedicated "cheat-detection" modules hardwired in the human brain.
Piece of nonsense, as the evidence does not show anything about hard-wiredness. It shows that humans are good at detecting cheaters, but not that it is hardwired. See for fuller discussion in Evolutionary Psychology, section 4.1.

Cultural Evolution

1. Are cultural memes like genes? ......
This discussion is ridiculously out of touch. The main difference between biological and cultural evolution is that cultural evolution can be 'transmitted horizontally', i.e. between unrelated individuals, in an unconstrained manner. This has many fundamental consequences.

The most important is that cultural evolution is much faster than biological evolution, in many orders of magnitudes. As a result, if a trait is beneficial and can be evolved culturally, it will evolve culturally very fast. Thus biological mutations will not be advantageous anymore, and the trait will not evolve biologically. As a result, any trait that can evolve culturally will never evolve genetically.

Other differences following from unconstrained horizontal transmission are:

  • Group selection become feasible, because a whole group can acquire a trait long before any other process can interfere.
  • There is no physical linkage between traits, as opposed to link between genes.

    Additional point is that cultural evolution cannot be really separated from intentional changes.

    first type of explanation can be found in Donald's account of the appearance of cognitive plasticity as a crucial evolutionary change. The primate mind became modern by developing a powerful learning device without constraining restrictions as to the range of mental contents that could be learned (Donald 1993). An explanation in terms of more specific capacities is Tomasello et al.'s account of cultural learning, as distinct from social learning based on IMITATION and found in higher primates. Cultural learning requires mind-reading and perspective-taking capacities (Tomasello et al. 1993). Obviously, these capacities would have been boosted by the appearance of verbal communication.
    The author mentions here the idea of general learning device, but, with one exception, 'argues' against it simply by ignoring in the rest of the text.

    Additional tricks here are the implications that the 'learning device' must be working by imitation, and that mind-reading and perspective-taking are in contradiction with the 'learning device', even though these capacities can be learned.

    One may push this further and argue that the appearance of culture depended, not on a more powerful general learning capacity, but on a multiplication of specialized capacities (Rozin 1976; Tooby and Cosmides 1989). This would have been less costly in evolutionary terms. It only required gradual addition of small modules rather than the sudden appearance of a general and flexible mind.
    Am example of theory-driven blindness, as it is in contradiction with the facts:
  • At least on earth, evolution hasn't figured out how to gradually add small modules. All evolution on earth is done by modifications of existing structures.
  • The anatomy of the brain does not show any new small modules. It does show enlargement of existing modules (most importantly, the cerebral cortex).

    Nevertheless, the author has a theory, and mere facts do not bother him.

    Also, it makes more computational sense. An unbiased all-purpose learning capacity could be overloaded with many adaptively irrelevant facts and correlations in the environment.
    That argument is irrelevant, as evolution does not try to make computational sense. In addition, it is not based on anything real, as we don't actually have any ideas about the design of really intelligent systems.
    Is culture constrained by genes?
    The rest of the discussion is based on the assumption that genes do constrain the culture directly, but without giving any supportive evidence. All the following examples are about knowledge about the world around the person (which includes other humans), so they all can be learned. The author 'argues' against this possibility by simply ignoring it.

    Evolutionary Psychology

    EPs consider their field methodologically analogous to reverse engineering in computer science....
    Nonsense. 'Reverse engineering' of some device means looking at the internals of this device, and evolutionary psychologists don't do that.
    EPs consider it likely that cognitive architectures contain a large number of evolved computational devices that are specialized in function, such as FACE RECOGNITION systems, a language acquisition device, navigation specializations, animate motion recognition (Gallistel 1995). They are skeptical that an architecture consisting predominantly of content-independent cognitive processes, such as general-purpose pattern associators, could solve the diverse array of adaptive problems efficiently enough to reproduce themselves reliably in complex, unforgiving natural environments that include, for example, antagonistically coevolving biotic adversaries, such as parasites, prey, predators, competitors, and incompletely harmonious social partners.
    This is doing the normal confusion between processes and devices. There is nothing to stop the processes from being content-dependent, even if they are executed in a general purpose device, and this is the most likely explanation of the human brain, based on its anatomy. The author tries to argue for specialized devices by arguing against general-content processes.

    The unnatural switch from 'devices' in the first sentence to 'processes' in the second sentence suggests that the author himself is aware of the inconsistency, and is intentionally mislead the reader.

    Selection drives design features to become incorporated into architectures in proportion to the actual distribution of adaptive problems encountered by a species over evolutionary time.
    That is true only for features that have already arisen. Because mutations are random, not every features arise, and that factor must be taken into account.
    There is no selection to generalize the scope of problem-solving to include never or rarely encountered problems at the cost of efficiency in solving frequently encountered problems.
    Intentionally misleading. 'Generalization' is not done to include a specific problem, but to improve the performance on all problems. Since some of these are going to be often-encountered, the generalization is selected for.

    Naive Sociology

    Evolution prepares all living things to resolve (or attempt to resolve) recurrent problems facing the organism.
    Wrong. Evolution selects individuals that are better equipped to cope with recurrent problems. However, the better equipment has to arise by random mutations. The point is that you cannot assume that if some feature is useful, it is going to evolve.
    It is extremely likely that evolved adaptations emerged in response to recurring social problems that our ancestral populations faced (Baron-Cohen 1995).
    Wrong again, because these adaptations has to be very complex, and hence have a very small probability of arising by chance. In addition, there is no sign in the brain for this kind of adaptations.
    Relevant evolved adaptations involve specialized mechanisms in both humans and nonhuman animals (particularly primates), including THEORY OF MIND; domain-specific devices for the recognition of faces, voices, and affective states; cheater detectors; and capacities for representing social dominance.
    There is no evidence for evolved mechanisms for any of this list. It is all based on overinterpretation of the data. See for for example the discussion of THEORY OF MIND in the mitecs-psychology.
    Other capacities that evolved to coordinate information processing of non-social phenomena may have also been recruited to treat social group-level phenomena.......
    It is quite amusing how the authors and the people he quotes takes evidence that shows the generic nature of human intelligence as evidence for transfer between evolved domain specific mechanisms. This is a typical example of theory-driven blindness.
    Hirschfeld (1995) and Jackendoff (1992) argue that mental representations of human groups are also governed by a distinct cognitive faculty of social cognition or naive sociology.....
    None of this is based on any evidence that can be used to reject the hypothesis that human capacities are based on a generic learning mechanism.
    The complexity of the social environment led Hirschfeld (1996) to propose the existence of specialized knowledge structures dedicated to social group understanding.
    A blatant nonsense. There is no way to deduce from the complexity of a problem/situation the nature of mechanisms that are used to cope with it. In particular, just because something is complex does not mean that humans (including children) cannot learn to cope with it.

    Numeracy and Culture

    Studies with diverse species show that animals are sensitive to number (see Gallistel 1990 for a review). This literature suggests that there are innate capabilities that have evolved to support numeracy in humans (Gallistel and Gelman 1992). A variety of sources of evidence point to early numerical abilities in human infants (see INFANT COGNITION). For example, infants as early as the first week of life have been shown to discriminate between small numbers (Antell and Keating 1983). The possibility that this discrimination is carried out by a cognitive mechanism encompassing several modalities has been debated (Starkey, Spelke and Gelman 1990). In addition, studies have shown that infants have some knowledge of the effects of numerical transformations such as addition and subtraction (Wynn 1992).
    All these studies are heap of overinterpretation, spiced with irreplicable results. For example, when the authors say that one-week old infants "discriminate between small numbers", they actually refers to studies that show the infants looking a different amount time at pictures with different items. Since these pictures also differ in other aspects (at least the amount of details must be different), we don't know what the infants respond to. The researchers simply assume that it is the number that the infants respond to.

    Sexual Attraction, Evolutionary Psychology of

    General: This author seems to be completely incapable of considering the possibility of learning anything. Even when he brings evidence that points directly to learning, he cannot see it.
    Evolutionary psychology is evolutionary biology applied to the brain's adaptations.
    That is simply false. Evolutionary psychologists do not investigate the brain at all. They do behavioural research.
    Humans make aesthetic judgments in numerous domains, and for theoretical and empirical reasons, these judgments are viewed as reflecting domain-specific psychological adaptations, not general-purpose ones (Thornhill 1997b).
    The author feels the need to defend domain-specificity, but does not try to defend the assertion that these are 'adaptations', by which he means genetic (innate) properties of the human brain, as opposed to learned properties.
    In human evolutionary history, individuals who saw as sexually attractive body traits of health out reproduced other individuals because the formers' preferred mates survived better, had genetic health, and were better able to provide investment.
    That may be intended as an argument for genetic evolution, but it is not. It would be so only if can be shown that humans had genetic mutations that caused them to see as attractive body traits of health. In the lack of this evidence, this is just hot air.
    The adult human form is an array of sex-specific, sex-hormone-mediated secondary sexual traits that signal health.........
    A blatant nonsense. Most of this traits are functional (e,g, athleticism in men, large hips in woman) , and large number of the rest are probably accidental (e.g. the size of the lower part of the face). The reason that we find them attractive is because we learn that they show a good functional body of the right sex. The author seems to be unable to even consider this possibility.
    For example, the value placed on physical attractiveness in choice of a long-term mate in each sex correlates positively with the prevalence of parasitic diseases across human societies (Gangestad and Buss 1993).
    This correlation is obviously not based on genetic variation, so it must be learned. This is an example of a theory-driven blindness.
    Symmetry is a component of sexual attractiveness in humans.
    Yes, but also in other kind of attractiveness. Since the brain and the sensory organs are themselves symmetric, it is quite possible that the preference for symmetric objects is effectively built-in. However, there is no evidence that this preference is sexual-specific, or more important in sexual attractiveness than in other cases.