Monday, July 25, 2005

The "concept" of a "religion"   posted by Razib @ 7/25/2005 12:51:00 AM

Over at Randy's place a heated discussion about Islam & homosexuality has broken out (via Abiola). One of the issues that (as always) crops up is the "true" Islam vs. Islamism dispute. To which I ask, what is the more birdy bird, a robin or an ostrich? Both are birds by a checklist definition, or a phylogenetic definition, but when posed this question by cognitive scientists, most people assert that the robin is the more exemplar or prototypical "bird." Concepts derive from a host of inputs (induction via examples), a few axioms, a general theory of the concept (ie; often a level of telos, which shouldn't surprise those familiar with evolutionary biology) and the context that the concept is framed within. In other words, most concepts are not derived from a few axioms like the mathematical definition of a triangle. Attempting to use logical methods to falsify someone as a "Muslim" or infer the "natural" implications of being a Muslim are fraught with difficulty because the assumption is that there is a prototypical or ideal Islam that naturally follows from agreed upon axioms. Unfortunately many concepts (most) don't work this way, they are characterized by a continuous distribution about central tendency (or central tendencies if there are multiple exemplars) with variation and outliers. Additionally, context matters, grey hair is generally clustered with white hair, with black hair being the outgroup, but grey clouds and black clouds are clustered together with white clouds being the outgroup.

Of course, aside from ornithologists and hard-core cladists (yes, you GFA) the question of what is a bird is not ontologically significant. But as I noted yesterday, arcane circumlocutions around the unknowable in the garb of logic is a common tendency when you are talking about heartfelt beliefs that are difficult to approach using rationality or empiricism (politics?), but from the attempt at such a feat comes the impression that there is a truist Islam, closer to the ideal as inferred from the axioms of the faith (by learned scholars in the faith). The problem is that language is not expressed through formal logical notation, and squishiness is a feature not a bug. Consider the idea that "there is no compulsion in religion" in Islam. Sayyid Qutb, the evil genius who was the Marx to bin Laden's Lenin agreed with this, for you see, his idea was simply that under and Islamic government everyone could choose to be a Muslim or a dhimmi,1 there was no compulsion, and everyone then had the "freedom" to choose Islam as they should. As Bill O'Reilly would say, "Show me where I'm wrong?"

Recommended: A far more adroit explication of concepts can be found at Chris' blog (I left a lot of stuff implicit obviously, in part because I'm not familiar with this area, it just "makes sense" to me in terms of the religion debates that go on on weblogs), Concepts I: The Classical View, ConceptsII: The Prototypes, Concepts III: The Exemplars, Concepts IV: A Second Revolution and The Importance of Names. If paper is your thang, check out Mind Readings: Introductory Selections on Cognitive Science or Mind: Introduction to Cognitive Science, both have chapters on concepts.

Addendum: Obviously some of the ideas that I hint at about concepts (and categories) have relevance to many socially constructed demarcations with fuzzy boundaries. Longtime readers of this blog might find this from page 102 of Mind Readings amusing: "...sensitivity to correlations of properties with a category enables finger predictions...." Yes Virginia, there is one true reality, and intellectual disciplines are its many faces.

1 - Apostasy is of course not permissible, but one Muslim apologist (see Leaving Islam: Apostates Speak Out) for this position simply said this did not contradict the "no compulsion in religon" edict because apostasy (from Islam) was an act against nature itself, so it really didn't fit the bill. Of course, it all makes rational sense.