Archive for October, 2011

This week’s forum (10/19/11) was about the soul, and the background reading was “Cognitive Science and Christian Theology” by Charlene Burns (Ch. 8 in Soul, Psyche, Brain, Ed. Kelly Bulkeley, 2005). Lucas chose the reading for the issues it raised and for the summaries of some aspects of the history of the concept of soul, rather than for the specific arguments and conclusions in the chapter. Our discussion mostly focused on the overall topic of the soul rather than the specifics of the Burns chapter.

The concept of “soul” can be traced back (at least) to Aristotle and has been tightly connected to some of aspects of intelligence (sentience, agency, reason, will), our overall topic for the semester. “Soul” has been interpreted differently by different people over time, and Lucas summarized some of the (western) history of the concept of soul as well as some of the modern interpretations and their relationship to scientific ideas.

Brief descriptions of some major shifts in philosophy of science are as follows:

    • Positivism (early 19th century): knowledge is based on natural phenomena and their relations, as verified by empirical sciences. Prior to the rise of positivism, analogy/metaphor were considered important ways of asserting things. Positivism rejected the importance of analogy/metaphor.
    • Modernism (late 19th and early 20th century): one privileged way of looking at the universe, concerns what is true (not useful, etc).
    • Popper (~1930s): Idea that science progresses by falsifying statements rather than by building up more and more true/correct statements.
    • Kuhn (1960s): Structure of Scientific Revolutions. Science is about paradigms (networks of connected concepts) and paradigm shifts. Science can be divided into “normal” and “revolution” phases.

Joanna suggested that modern science has returned to the importance of analogy or metaphor, only they are called models in science. If you agree that the only complete description of the entire universe is the universe itself, then all scientific descriptions of things can be understood as models, that is, they are simplified summaries of something more complex that are useful for describing/discovering patterns and regularities.

The metaphysics of Aristotle included matter and form (shape and function). He described the soul as a property of certain things (not as a separable “thing”)

    • Being able to grow characterized the vegetable soul
    • That which gives sensation, movement, desire is the animal soul
    • Being able to think is a property of the rational soul

A competing notion was the Hebrew notion of spirit (breath), which was basically a supernatural concept. It is unclear whether Hebrew “spirit” refers to other animals or only humans.

Both soul and spirit were functional holisms, not a separate thing. Soul does not need to have an everlasting component. Sheol is the Hebrew idea of an afterlife, and could be understood as the echo in the universe of a person/soul. There was a question about the Hebrew word that is translated as “soul” and often used together with “body” (“body and soul”). The group did not know the origin or history of that word.

The Greeks generally valued the immaterial over the material and begin to develop the concept of “soul”. However, this idea of “soul” is not really present in the New Testament.

Paul uses the word “soul” to refer to what makes you alive and “spirit” to refer to what you have after baptism (this is what can have eternal life and is later called the “soul”).

The Anglican tradition has been that spirit and soul are present in all people, whereas some other Christian theologies have more strongly emphasized Paul’s distinction.

Thomas Aquinas was a major foundational theologian of the Catholic Church and he based many of his views on Aristotle. However, Aquinas (unlike Aristotle) conceived of supernatural as something distinct and separate from the natural. He saw the vegetative and motive (animal) soul as natural and the rational soul as supernatural. He claimed that a person had a rational soul starting 40 days after conception for a male or 80 days after conception for a female (apparently corresponding to some false beliefs about when movement of the fetus begins).

Important people in developing the mechanic philosophy included Galielo, Gassendi, and Newton. According to this philosophy, nothing has internal motivation (inclination).

In response to the mechanic philosophy, Descartes moves the soul outside of the body. The concept of the soul as the “ghost in the machine” is a strict ontological dualism. According to Descartes, one possible point of connection/interaction between the soul (mind/seat of intellect/agency) and the physical body is the pineal gland.
The issue of emergence came up in the reading and we discussed this topic in general before discussing how emergence may relate to mind/soul.

The word “emergence” is used differently by different people. Some use it in a sloppy way as a stand-in for some sort of mystical explanation, and this can make it difficult to communicate using the non-mystical meaning of the word.

Lucas proposed that there were three distinct ways in which “emergence” is being used.

    1. An “emergent” property is one that you cannot explain from a complete description of the “individual” properties of the components themselves. People in the forum seemed to have diverse ideas about whether this is a viable conception of “emergence”. If you do not think there is sense in which a description of one thing can be “complete” without including all of its relational properties and all of their relational properties (i.e., everything in the universe), then this idea of emergence does not work. You never could (even in principle) completely describe an isolated thing, so the whole premise of distinguishing “emergent” properties does not really make sense.

    2. An emergent property could be a property that is better described at that level than by decomposing it into lower levels. Often there is a qualitative difference in models (i.e. metaphors) at different levels. An example that was mentioned in the book chapter and in our discussion was pizza. The basic idea is that describing a pizza in terms of all its particles would be correct (the pizza does not defy part-whole supervenience). But this sort of description would not be useful as a tool for gaining insight about pizza. A description which models (has metaphors for) properties at the level of the pizza would be more useful. [I think this is also called “weak emergence” by some.] This is the “respectable” version of emergence that seemed uncontroversial among forum members.

    3. “Ontological emergence”. [I think this also has been called “strong emergence”] This is basically the idea that something new can really appear at higher levels of organization. That is, the properties of the whole are not deducible (even in principle) from the properties of the parts.

The issue of emergence is particularly relevant for mind/brain questions and we briefly discussed the issue of what it means to have really explained something. Cognitive scientists work on explaining mental states (thoughts, feelings, etc) in terms of neurons and their connections. But so far, qualia or the “feeling of what happens” does not have a compelling/complete explanation from neuroscience. That is, why “you” have a subjective experience of anything has not been fully explained and some question whether it is possible to explain this by descriptions of the component parts (neurons) that apparently give rise to subjective experiences. The question came up of whether theologians think that neurologists will ever satisfactorily/completely explain qualia. Lucas said that there is no consensus on this point among theologians.

Another point of discussion was the fact than at least several forum members saw the article as characterizing two separate camps (science and religion) which are fighting each other. Furthermore, the author critiqued others’ positions not only in terms of why a particular point may or may not be correct, but also in terms of why it would not be strategic to concede particular points.

There are many responses followed the Mechanic Philosophy, summarized briefly as follows:

    • One popular idea was that of the particulate mind, developed by de Chardin (John Haught and Dobzhansky are also associated with this idea). De Chardin was a Jesuit of the early 20th century. The idea is a type of orthogenesis—that evolution is progressive and internally directed. Evolution in this view is seen as progressing towards greater consciousness and complexity, eventually culminating in God.
    • Intelligent Design– in this view, evolution is progressive, but progress is due to an external reason (God). ID can be broken into two basic camps: young earth believers and those that think evolution occurred over a long period of time but that God initiated the process and that life has irreducible complexity.
    • Another response is process theology (important people include Whitehead, Cobb) the dynamic tension between God and humans is what really exists, and soul is a part of this. Self and God are inseparable (notably similar to the Hindu idea that “Self is God”).
    • Compatiblist– basically that the cognitive neuroscience perspective and the concept of the “soul” are both correct and are currently the best way of getting at the same thing. They will be reconciled after major advances in cognitive neuroscience and theology. (The Burns article was warning against this perspective)

We briefly discussed the issue of free will. Physicalist positions include the idea that free will could be a (weakly) emergent property or that free will is illusory and not a useful concept. Dualists are able to maintain a notion of free will. The issue of is moral responsibility perhaps more relevant/interesting to us than issue of free will. People often link agency and liability—though not always, as in Calvin (liable for your sins even though you have no choice about it). We will have more discussion later in the semester about free will/agency/moral responsibility.

Returning to the work that the concept of “soul” has been doing during history, Lucas provided the following list:

    • sentience
    • agency
    • reason
    • will

(in short, internal motivations)

These properties had been explained/subsumed in the concept of the “soul”. The mechanic philosophy eliminated internal motivations from physical (all?) reality, leading to the dualist idea that the “soul” is something like a “ghost in the machine”. But now, humanity’s understanding of the “machine” (in particular, the brain) is good enough to start to explain aspects of sentience, agency, etc. in physical terms. So the dualist notion of “soul” would face serious problems. But the status of other notions of the soul is not as simple. The concept of the soul is clearly obsolete for some of the “work” that it has done over the course of history, such as explaining why animals move. But for other reasons, such as salvation, life after death, etc., theologians still find some concept of soul to be needed/useful.

Another aspect of the discussion is that it was also generally thought that these properties (sentience, agency, etc) were irreducible, meaning that they always come from something else with it. But the current scientific worldview suggests that all of these features (sentience, agency, etc) do have simpler precursors and that the same is likely true for life itself (i.e. that life came from nonlife at least once).

Post Contributed by D.S.

Read Full Post »

The topic of this week’s forum was focused sharply on Geoffrey Miller’s hypothesis that sexual selection was the primary evolutionary mechanism that shaped human intelligence (or at least the “exceptional” aspects of human intelligence, such as humor, creativity, language, consciousness, morality.) Our reading was the first chapter in The Mating Mind: How Sexual Choice Shaped the Evolution of Human Nature by Geoffrey Miller, published in 2000. Because Joanna Masel read the whole book (which she highly recommends) she was able to expand on Miller’s arguments and provided many examples throughout our discussion. Lucas Mix directed our attention to the human trait of morality in the latter part of the discussion.

Our discussion began with a primer on Sexual Selection so that the group came to a common foundational understanding of this evolutionary mechanism. Some key points that emerged from this discussion include:

    1. Sexual selection is often neglected as a strong evolutionary force
    2. There are two facets of sexual selection: Mate choice (females choosing males based on some trait) and competition (selection for traits that ensure greater success in male-male competition)
    3. Mate choice is usually (but not always) unidirectional — females choosing males
    4. “Runaway” sexual selection makes exceptionalism possible in any species
    5. Traits that are a “handicap” are reliable signals of good genes, but not any arbitrary trait can become a signal; Signals are constrained by the sensory system of the signal perceiver (e.g., increasing the coloration or pattern complexity of the peacock tail will only be advantageous if the peahen has the visual sensory system that is able to detect the variation)

Turning the discussion to address the evolution of human intelligence, the above key points became relevant:

1. Most explanations of the evolution of human intelligence have focused on survival advantages of problem solving rather than the reproductive advantages of being creative (i.e., “survival of the fittest” vs. “reproduction of the sexiest”). Miller attacks the fields of Evolutionary Psychology and Paleoanthropology for assuming that human evolution was predominantly shaped by natural selection for survival traits, largely ignoring sexual selection. For example, artifacts such as the hand axe are nearly always explained in terms of hunting utility when it is just as plausible that they were sexually selected signals of male fitness. Miller argues that a hand axe is more likely a signal of fitness because of the large cost to learn and perfect the skill to create highly symmetrical axes– greater than necessary for utility, the cost it takes in time/energy, and the risk of injury while creating it. Further, it is hard to explain the survival value of Theory of Mind (ToM) when a) it seems to require an expensive brain, and b) so many species facing a wide variety of ecological challenges do well without it.

2. Sexual dimorphism (where males are larger or equipped more than females) is often an indicator of sexual selection via male-male competition. In humans there is relatively little dimorphism physically and mentally (according to Miller). Clearly Miller favors sexual selection via mate choice rather than via competition as the predominant driver of some aspects of human intelligence.

3. Sexual selection via mate choice is often unidirectional (usually females being choosy about males). However, the more that both parents are involved in parental care, the more both sexes become choosy. Based on sperm competition theory, relative testes size in humans (i.e., testes larger than expected for body size but not as large as in highly promiscuous/polygynous species) indicates that this trait evolved in a socioecology of monogamy/promiscuity mix– perhaps serial monogamy. Given this tendency of some form of monogamy and the associated high investment in offspring by both parents, sexual selection via mate choice may be a bidirectional—females are choosy about male mates and males are choosy about female mates. Though there is often-cited evidence of men selecting women with physical traits that signal high fecundity (e.g., large breasts and buttocks, hip/waist ratio), there is also evidence of women selecting males with physical traits (e.g., shoulder/waist ratio, penis size). Penis size is relevant for “copulatory courtship” whereby females that are positively reinforced by the pleasure of orgasm are likely to copulate more frequently, allowing the male to have a greater chance of copulation during the ‘hidden’ fertile period. Likewise, in humans where biparental care is necessary for offspring survival, females and males may be choosy about mental traits that have no survival value (e.g., humor) but are honest signals of good genes.

4. A unique feature of sexual selection is this “runaway” process whereby the opposite sex finds some trait desirable, the trait then becomes advantageous to produce, though costly. The “handicapping” trait then becomes an indicator of good genes and so having the preference for the trait also becomes advantageous. This circular process develops a greater preference for pronounced versions of the trait very quickly until the cost of producing the trait is equal to the reproductive benefits. (In some cases the survival cost is very high, arguably greater than the reproductive benefits, leading to extinction, as in the Irish Elk.) Runaway sexual selection can lead to highly developed seemingly “arbitrary” traits. Perhaps aspect of human intelligence are just sexually selected traits.

5. Though runaway selection can make any trait desirable, the trait cannot be totally arbitrary. Traits must be perceivable by the selector. It seems reasonable that the principle of evolutionary conservatism would apply and thus the machinery for producing a desirable male trait (e.g., genetic instructions for large brain with linguistic ability) is the same machinery females use for detecting a good male (large brain with linguistic ability). The genetics would be carried by both sexes. So for example, a male using his big brain to exhibit his linguistic trait by using ‘big words’ is matched by a female using her big brain/linguistic abilities to detect misuse of his vocabulary. Similarly, if ToM is a sexually selected trait, it may be necessary for the chooser to likewise have ToM.

By the same principle of evolutionary conservatism, in a species where there is bidirectional mate choice, the trait that one sex uses for mate choice can also be the same trait that is used by the other sex. Again the genetics would be carried by both sexes. For example, instead of males choosing females based on breast size and females choosing males based on penis size, males and females could both select for big brain/intelligence in their mate.

Male care can be a sexually selected trait, especially in monogamous species. One must take into account that male care can be considered either paternal care (care for his offspring) or courtship effort/mating effort. Miller suggests that in contemporary hunter/gatherers, a male caring for his own or step children is often observed and is really mating/courtship effort: From the female’s perspective, it’s handy to have this male around as babysitter.

Regarding Morality:
Since we are dealing with this question about selection for intelligence traits, it is important to recognize that this process happens unconsciously (conscious “selection” not necessary).

If morality is a sexually selected trait in humans, then one would have sophisticated adaptations to detect it and have strong feeling about it, making it hard to think about in a scientific way. The observed difficulty in studying morality seems to match what one would expect if morality is a sexually selected trait.

Another line of evidence that morality is a sexually selected trait is that moral norms seem to match the criterion of a costly signal indicator. For example the ‘altruism’ of Zahavi’s Arabian warblers feeding non-kin is a costly signal. Similarly, religion is a signal of a complex suite of traits that is costly (including a high entry cost).

Regarding multiple forces acting on the Evolution of Intelligence:
There are different forces that may be acting on human intelligence including

    • Selection for status-gaining traits in a complex social group (for survival and mate competition-Dunbar)
    • Selection for arbitrary costly traits (for mate choice-Miller)
    • Selection for complex social groups themselves (group selection)

Selective forces may not be acting in the same direction. Also the relative contribution by sexual selection is greater than the others, Miller would argue.

Regarding human exceptionalism:
Traits that are sexually selected may have no survival value (and indeed often have a survival cost.) Given this runaway process, sexual selection by mate choice makes “exceptional” traits possible in any species (e.g., peacock tail, elaborate species-specific behavior in the bower bird). This exceptionalism is not because of an adaptation to the environment but rather an outcome of the runaway feedback in sexual selection. Indeed if a trait is an adaptation to the environment, then one would expect convergence among species that face similar environmental challenges. If a trait is an adaptation to sexual selection, one would expect uniqueness. In this respect, sexual selection makes human exceptional traits seem highly plausible.

Post Contributed by N.S.

Read Full Post »

This week’s forum revolved around a discussion of “The cognitive functions of language,” by Peter Carruthers. We began by discussing the definition of language. Joanna and Lucas proposed the following definition: “Language is a symbolic syntactic system for conveying propositions.” We all agreed to use this definition for the remainder of the class. This was followed by a brief discussion about what exactly is meant by “intelligence,” however we decided that this was ground that had been covered in prior meetings, and that we could move on, given that for the purposes of the day’s class we would be assuming that intelligence was synonymous with cognition. There was also some preliminary discussion surrounding the interplay between language and cognition, and the presence of non-linguistic (but sophisticated) cognition in species other than humans. We next went over some of the paleolinguistic evidence cited in the paper concerning the appearance of language and symbolic cognition in early humans, which several of us felt was not particularly convincing. It was agreed upon that language is a particularly salient feature of cognition that appears to be fairly unique to humans, and hence the article and subsequent discussion are of particular relevance to the broader themes of the course.

The specific, central claim of the Carruthers article regarding the role of language in human cognition is that language serves as a mechanism for for domain-general cognition. The author claims that it does so specifically by enabling communication across distinct cognitive modules. In order to understand this claim it was necessary to spend considerable time discussing and defining the notions of cognitive modules and cognitive domains. We began with modules. We first noted that there are several possible (but not mutually exclusive) ways of conceiving of modules. For example, one can define a module at an algorithmic level, as an encapsulated subroutine for processing a particular type of information or solving a particular type of problem. Additionally one can develop an anatomical notion of modules as specific portions of the brain that are devoted to a single task. We also noted the difference between a purely methodological claims about modules (which treat them solely as useful conceptual entities for understanding cognition) and factual claims about modules (which treat them as things that actually exist). We concluded that the Carruthers article does indeed make factual claims about modules and modularity.

We then spent some time discussing the various bits of evidence both for and against the anatomical notion of domains. At least at the level of early sensory processing, it appears that there are indeed specific brain structures that are devoted to specific tasks. The LGN (lateral geniculate nucleus) was cited as an example of such a structure. The LGN is a clearly anatomically defined area involved in early-stage visual processing, and does little else but process visual information (albeit at a fairly low-level). The evidence become less and less clear when one looks at increasingly complicated types of cognitive tasks. There is fMRI data showing that specific spatial patterns of brain activity can be correlated with specific cognitive tasks, and that these patterns are repeatable across trials within a given individual, lending support to the notion of a cognitive module being anatomically well-defined. However there are a number of important caveats. Specifically, the observed patterns of activity associated with particular tasks can be different across individuals, and can overlap significantly across different cognitive tasks within the same individual (which seems to contradict the idea of modules as being well-encapsulated and limited to specific types of inputs and outputs). Furthermore, the brain is capable of a great deal of plasticity, and hence the functional role of a particular patch of brain tissue may change dramatically over time as a function of experience. It was pointed out that the notion of modularity commonly used in studying networks of gene interactions is a more statistical one, where the various interacting nodes are grouped into modules based on the density of within-module connections relative to between-module connections. This notion of a module can be applied to cognition and is in many respects easier to defend. Additionally, the central claims of the Carruthers paper regarding the role of language in cross-modular communication are compatible with this definition.

The discussion then shifted to the algorithmic or functional notion of a module. At some point (possibly earlier in the discussion) it was observed that this definition of a cognitive module does not depend upon the existence of anatomically well-defined modules, and remains meaningful even in the absence of clear neurophysiological evidence. We proposed a number of examples of cognitive modules, including vision, innate phobias, folk biology, face recognition, folk physics, cheater recognition, theory of mind, and social cognition. We also placed these modules along a continuum ranging from “more perceptual” (less interpretive) to “more conceptual” (less interpretive), with primary visual processing occupying the perceptual end of the spectrum and social cognition falling closer to the conceptual end of things. It was pointed out that there can be significant interaction between modules at different points along the spectrum, and than even early stage processing can be influenced by higher-order cognition, which can be seen behaviorally as well as anatomically (due to the large number of back projections from late-stage processing areas to early-stage processing areas). On the other hand, certain visual illusions provide evidence for the existence of at least some amount of encapsulation separating early-stage visual perception from higher-order cognition. More precisely, because these illusions persist even when one is convinced of their illusory nature, it follows that higher-order cognitive modules (dealing with beliefs about the images being illusions) are unable to influence the early stage perceptual modules that generate those illusory perceptions.

Links to some interesting illusions:

We also spent some time discussing the idea of a cognitive domain, and how domains are different than modules. To summarize, one can think of domains as referring to various adaptive problems (locating food, mating, sensing the environment, etc.) that an organism must solve, in contrast to modules, which are better described specific subroutines used to solve problems within one or more domains. We concluded that the concept of a domain was perhaps less important for the purpose of understanding the reading than the concept of a module.

Finally, toward the end of class we began to hit upon the central thesis of the Carruthers paper, namely that language serves as a medium for communicating and integrating across multiple cogntive modules, thereby acting as a type of cross-domain cognition. We discussed the key bit of experimental evidence cited in the reading, in which it was found that speech shadowing (repeating spoken language after hearing it) interfered with the ability to combine cross-domain information in a dual-task experiment more than rhythm shadowing (playing back a sequence of beats using one’s hands), even though the demands on working memory were comparable for speech-shadowing and rhythm-shadowing. We fleshed out the idea of language as a central hub in a network of modules, and discussed how this can lead to the formation of increasingly complex and abstract representations that can integrate information across multiple domains and modules.

Images of the white board for this meeting (from left to right):

Post Contributed by D.L.

Read Full Post »