Representations in the brain?

From A conversation about the brain
Jump to: navigation, search

One advantage claimed by enactive theories of cognition is that they avoid the need to postulate representations in the brain. For example, where other theories make reference to internal representations of the environment, an enactive theory might refer instead to the claim that an organism’s actions are sensitive to the ways in which patterns of sensory stimulation depend on the organism’s movements. This is seen as an advantage of enactive theories on the grounds that the notion of neural representation is problematic (empirically, conceptually, or both). On the other hand, critics of enactive theories take this feature to be a fatal disadvantage, on the grounds that it commits enactive theories to behaviourism, which is notoriously problematic. (In the standard philosophical terminology, a theory is behaviourist if it reduces cognition to behavior, or if the fully articulated theory eliminates references to cognition, replacing them with references to action or motor behaviour.)

To assess how this dispute bears on the theory discussed here, we need first to distinguish between a couple of different claims about a neural system:

(i) Representation in the system is reducible to motor behavior, or eliminable in favor of motor behaviour.

(ii) The system represents the environment in motoric terms.

The theory discussed here might be committed to (ii) without being committed to (i). For example, the theory might be spelled out in terms of a system of genuine representations – unreduced to actions – which represent the environment in terms of the motor behaviours available to the organism, rather than in terms of a map or 3D coordinate frame (See Discussion).

The theory discussed here does deny the need to postulate a certain kind of representation, namely a map or 3D coordinate frame, but that does not entail denying the need to postulate any representation. To be precise about what kind of representation is being denied, we need to draw a further distinction between the content and the format of a representation. Roughly, a representation’s content is the information which the representation makes available to a system, whereas the format of a representation is the means by which the vehicle of representation carries that information. For example, a map and a sentence might each have the same content – they might tell you the same facts about how objects are arranged, say – though they represent that content in different formats. So for instance the negative idea discussed here, that the brain does not reconstruct a map or 3D coordinate frame, could be understood in either or both of the following ways:

(a) Perceptual representations in the brain have contents which are not expressible in terms of a 3D coordinate frame.

(b) Perceptual representations in the brain have a format which does not include a model of 3D space (that is, it does not include elements isomorphic with the 3 dimensions of the space it represents).

One question to be pursued, then, is whether there are features of the theory of brain function discussed here which really commit it to (i), or to (ii.a), or to (ii.b), or to (ii.a) and (ii.b). Another question is what the relative merits of these commitments are. For instance we can ask: Which of these commitments are consistent with our Perceiving a world beyond ourselves?

On the face of it, the theory of brain function discussed here is committed to (ii.a). For example, consider one piece of motivation for the theory: cases in which observers cannot compare stimulus depth and width accurately, although their depth-depth comparisons and width-width comparisons are accurate. The theory’s explanation of this datum involves (ii.a): if representations of objects’ depth and width had a reconstruction-like content, values for depth and width would presumably be comparable; instead the theory postulates representations with task dependent contents. Furthermore, a commitment to (ii.b) is a likely corollary of this explanation, because where spatial representations have the format of a 3D model of space, this makes all but unavoidable an interpretation according to which those representations have contents which are expressible in terms of a 3D coordinate frame. None of this, however, requires a commitment to (i): the explanation is given in terms of perceptual representations.