Gillett, Carl - Understanding the New Reductionism: The Metaphysics of Science and Compositional Reduction

09/21/2007

Journal of Philosophy, Vol CIV No 4 April 2007

Author begins by discussing the recent move by Kim, a reductionist, from semantic reduction to metaphysical reduction. This eschews the long-standing 'Nagelian' approach of reducing entities in the 'special sciences' (basically, any science that isn't particle physics) to more elementary ones by way of 'bridge laws'. Instead, the focus is on the metaphysics and ontological relations involved in 'mechanistic explanation'. Mechanistic explanation is, basically, an explanation of how a device or entity works by description of its parts, how each part functions, and how the parts work together. One of the suspicions is toward the term 'metaphysics', but author claims that this is the 'metaphysics of science': a careful, abstract investigation of ontological issues as they arise within the sciences-- (e.g. not a priori). (pg 194-5)

Author discusses the main problems with previous reductionist attempts: there was an unavailability of 'bridge laws' or other descriptors of the lower-level entities such that the higher-level 'emergent' properties couldn't be explained in terms of the lower-level entities. Thus the antireductionists argued that the entities, terms and properties of the special sciences (higher-level) were ineliminable for proper scientific understanding and experimentation.

Author launches into an example of neuroscience where many diverse lower-level entities compose higher-level ones (pg 198). What is interesting in the example is that the higher-level changes that take place aren't caused by lower-level changes-- it is part of the structure of the entities involved that the changes take place. Author argues that this isn't causation but instead noncausal determination (pg 199-200), and that a familiar kind of scientific practice is going on here: explanation by describing composition.

Focusing on the compositional nature of higher-level entities reveals the following: (1) the powers ascribed to the higher-level entity does not have a lower-level analogue. This was the point made by the antireductionists. (2) The lower-level entities are qualitatively different in kind, making up the higher-level ones. This indicates a 'many-one' relation, not an 'identity-identity' relation.

Earlier author put how he individuates properties using a Shoemaker-type 'causal theory' attempt: a property is individuated by what power it gives to the objects that instantiate it. A 'power' is an entity whose possession allows an individual to enter into a certain process (pg 201). Basically, the powers in the lower-level entities comprise the powers in higher-level entities iff the activated powers of the lower-level entities all taken together activate the powers of the higher-level entities (pg 202), and not vice versa. There is some weird discussion/usage of 'manifestation grounds'. Given that there is asymmetry in the 'manifestation grounds', there is room to argue for reduction on an ontological level while leaving the semantic aspects of the special sciences intact and ineliminable.

Author formulates his 'Argument from Composition': (pg 204-5)
1- Properties are individuated by what powers they grant to their instantiators
2- The properties belonging to the lower-level entities are sufficiently efficacious for manifesting higher-level powers
3- Higher-level properties grant no special powers to their higher-level entities, therefore are ontologically dispensable.

The remainder of the paper, author shows that though ontologically dispensable, higher-level entities and properties are not semantically or epistemologically dispensable, being necessary for science and scientific practice. Thus it seems to be granting some of the antireductionist points (and perhaps main contentions) while remaining reductionist in what matters most: ontology.