Polger, Thomas - Realization and the Metaphysics of Mind

10/05/2007

Australasian Journal of Philosophy, Vol 85 No 2 June 2007

Author's main target is Gillett's account of Realization and Reduction. Author has numerous objections to the family of realization accounts (Kim, Shoemaker, Gillett), as well as Gillett's specifically, which is called the 'dimensional' view. Author claims that these accounts of realization will destroy the distinction between Realization Physicalism (RP) and the original account (the account that RP was supposed to be an alternative to), identity theories. Author does not want to defend RP, but instead make sense of Realization so that RP can be properly evaluated.

Realization is not the same as 'instantiation'. Author gives multiple examples of use of the term 'realization' (pg 235), and suggests that the paradigm 'textbook' case is something like:
My computer currently realizes Microsoft Word; or
Memory fixation is realized in humans by long term potentiation of neurons.

"Certain electrical states of the device realize computational states such as, say, storing the contents of the last copy operation. The electrical activity of the device is not identical to any program state of Microsoft Word, but it implements or realizes such program states" (pg 236)

Author takes the view of Gillett et al to be that causal powers of properties of objects individuate the realization of the function in question. Author considers this the 'causal view' of Realization. Gillett's view differs from Kim & Shoemaker's in that their view is of a 'flat' causal theory where realizer and realized properties are in the same object, at the same level, in virtue simply of causal powers. Gillett's view is that the realizer properties can take place at a lower level, or a horizontal one, or as part of the structure of the object (or at the object level). This is considered the 'dimensional' causal view. (pg 238)

Author's major reply to the entire causal view is that it fails to capture objects that realize abstract processes, like machines that realize 'addition' or 'Microsoft Word'(pg 240). The claim is that an abstract function like addition is a formal and not a causal relation, and so you can't use causal powers to individuate them. Thus the causal approach fails to capture textbook cases of abstract, formal, or algorithm realization, since these things get realized but the objects/properties that realize them are not doing so causally but in form instead. Author extends this to intentional and etiological (historical) realized properties too (e.g. A US dollar is whatever the US Gov't says is a dollar). Author predicts there will be numerous objections to his attack, which he considers: (pg 243-6)
1) The computational/functional model of cognition (RP) is over
Author: so? We should still try to get the Realization relation right

2) The project of abstract realization in general is defunct
Author: wrong! 'We cannot dismiss abstract realization out of hand' (pg 245) [important!]

3) Ok, maybe a machine can't cause 'addition', but it can cause 'adding things', and that's all you need for Realization. Thus the causal view is saved.
Author: A) there are other realization relations that I hope that doesn't work for B) We can still say that an 'adder' that 'adds things' stands in a particular relation to addition, and thus name that Realization. (pg 246) [this makes no sense]

4) Realization of abstract functions is not Realization proper.
Author: But then we have no room for the special sciences, or for functionalism.

Author turns to the specific criticism of Gillett's dimensional causal view. Author claims (pg 248-50) it essentially destroys Realization and makes it into an identity theory. Author then proposes his own theory of Realization:

'to realize a property or state is to have a function'. (e.g. 'something realizes the property of being a heart iff it has the function of pumping blood') (pg 251)
Author leaves open what kinds of functions will be realized.

Author then considers a final attack on Gillett. This is a discussion about Multiple Realization (MR). Gillett thinks that before we have an idea of MR we need an account of R. But author denies this. Author says that MR is an argument for R, so we can't have R figuring in the explanation of MR-- that would beg the question. Author claims that MR is a theory about explanation and explanatory kinds, whereas R is a metaphysical theory about properties. (pg 255) This leads to a 'paradox' where something might be MR across species/objects/etc. but not actually realized (R) in the special, irreducible sense, since the multiple species/objects' properties are the same across those species/objects/items. Author puts forth that we must first use MR for kinds as an explanation, then investigate the relation between the properties realized (e.g. an eye) and the physical objects doing the explaining (e.g. retina, cornea). If the relation isn't one of identity across the MRs, then maybe we have the Realization relation instead.