Kissing Cousins: Mental Representation and Biological Mimicry

Kent Van Cleave
History and Philosophy of Science, IU Bloomington

     Full text: Not available
     Last modified: February 15, 2005
     Presentation date: 07/14/2005 4:00 PM in ROZH 105
     (View Schedule)

Abstract
The angler fish dangles a writhing pseudo-worm from a rodlike appendage sprouting from its forehead. A smaller fish approaches for a worm lunch and ends up as lunch for the angler fish. Biologists might say the pseudo-worm “represents” worms in general; better, it naturally represents the fact that appropriate prey for the angler fish habitually approach
wormlike things.

In general, mimicry arises when natural variations and selective forces combine to create a structure that is responded to by another kind of organism in a way that is advantageous to the mimic. Given the conflicting interests of the two kinds of organism for this variety of mimicry,
anti-mimicry countermeasures could spoil everything.

It should be easier for a single kind of organism to evolve a structure (better, a device _producing_ structures) designed to be responded to by one or more of its own subsystems – not as a _sign_ of something external (Millikan’s _intentional icon_), but as a _proxy_ for it – as the direct object of response. No need to extract, reconstruct, process, or transduce _information_ about the
environment – or know that it _is_ information. The environment is modeled on-the-fly, and the animal responds directly to the models.

Unlike matched-mode models (as mimics or statues visually resemble their subjects), these are realized in patterns of neuronal activation. Imaging studies show small neuronal populations scattered across the cortex respond selectively through training to natural kinds (at various levels), familiar individuals, properties, features, positions, etc. “Semantic network”
theories seek to explain mental content by positing semantic/conceptual correlates to this stuff of perception, with the _concepts_ linking into useful semantic structures. Posit instead direct reciprocal links among trained representational populations (effectively proto-concepts), and you get _phenomenal_ networks in which populations of neurons reinforce or inhibit one another’s activity via feedback. The entire network activated in response to occurrent sensory input is a many-dimensioned proxy world (or world fragment) that simply _is_ the world for the naive
organism.

Instead of, “How do we get and process information about the external world?” we should ask, “How do we generate our own world of experience that’s an adequate model of the external one?” “Information talk” is largely superfluous. The angler fish’s lunch doesn’t derive the mistaken semantic content _worm_ from sensory information; instead, photons striking its
retinas begin a causal chain that adds a _wriggling thing to eat there_ to the dynamic flow of its rudimentary proxy world.

Proxy worlds are likely structured as _driven, damped oscillators_. Feedback among nodes is insufficient to sustain mutual activation for long; but driven by inputs from the senses and/or attentional mechanisms, they generate richly detailed and persistent “perception”, memory, and imagery (the latter two driven by attention, often in spite of conflicting sensation).

I will sketch some supportive arguments, drawing on neuropsychological research and philosophical considerations. I hope for questions regarding how my view satisfies the central motivations behind teleosemantics, indicator semantics, internalist, interactionist, and other popular views.

Multiple Paper Session:
Other papers in this session:
The Justificatory Role of (Evolved) Moral Intuitions
Evolutionary Biology and the Epistemology of Metaphysical Modality
Sexual Selection and the Evolution of Moral Belief
How Evolutionary is Evolutionary Economics?