The main claim of reductionism is that the goal of science in the long run is to discover reductive explanations of macro phenomena. Reduction is a relation between scientific theories about some natural phenomenon. The macro-level theory describes the phenomenon in high-level terms, and the micro theory aims to characterize the lower-level constituents and their interactions. A reduction is achieved when the macro phenomenon can be explained in terms of the microstructure and its dynamics. According to logical empiricists, this required that the macro theory be logically deducible from the micro, together with suitable ‘bridge laws’. This is generally recognized as inappropriately strong. First, because reductions, like explanations, are usually partial and incomplete; second, because getting the theories to fit explanatorily typically involves conceptual modification to one or the other, or sometimes both. Achieving reductive integration between macro and micro theories does not imply that there is anything wrong or unscientific about the macro level theory, only that it is to be explained in micro terms.
The classical examples of reductive integration with minimal modifications to the theories are two-fold: thermodynamics to statistical mechanics (temperature in a gas is reductively explained as mean kinetic energy of the constituent molecules) and theory of optics to theory of electromagnetic radiation (visible light is explained as an electromagnetic phenomenon). The reduction of macro genetics to molecular biology is incomplete and is a good example of the diachronic aspect of theory integration. That is, the macro and micro level theories co-evolve, each providing constraints and corrections for the other. It is also complicated for many reasons, including the fact that a GENE can be distributed over separate chunks of DNA and a given segment of DNA can be involved in the production of distinct traits. Additionally, traits are the product of many interactions, including epigenetic (see EPIGENESIS) conditions.
In the course of co-evolution of theories, sometimes the conceptual modifications (high and low level) are made. This is easiest to see after a long passage of time and scientific evolution. Thus central concepts in Aristotelian physics, such as natural place and impetus, end up having no role whatever in Newtonian physics. In biology, the notion of ‘vital spirit’ dropped out of the science as cell biology and molecular biology flourished. These sorts of revisions are often accompanied by revisions in the formulation questions in science. For example, at the onset of Harvey’s seventeenth-century investigation of the heart, his question was ‘Where are the animal spirits concocted?’ After his discovery that the heart was actually a mechanical pump, the original question faded from view, replaced by new questions, such as ‘What causes the heart to change beating rates?’
Explanation of psychological phenomena in terms of the underlying neuronal mechanisms is a goal of reductionism in COGNITIVE NEUROSCIENCE. The favoured is the strategy that research should proceed at many levels of brain organization simultaneously. A purely ‘bottom-up’ or purely ‘top-down’ research (see TOP DOWN VS. BOTTOM UP PERCEPTUAL/ NEURAL PROCESSING) strategy is generally considered unnecessarily restrictive. The interaction between experimental psychology and neuroscience has been enabled good progress on a number of questions, including the nature of MEMORY, ATTENTION, perceptual systems, DRIVE and EMOTION, for example FEAR Typical aspects of macro-micro co-evolution are already evident. For example, conceptual revision has been required to accommodate the data. Memory turns out not to be a single undifferentiated function, but a complex set of partially dissociable functions. One area of impressive progress has been the phenomenon of WORKING MEMORY for spatial location. Particular neurons that hold spatial information in the absence of a the stimulus have been identified, and computer models suggest how networks of neurons might function to receive, hold and deliver spatial information. Another problem that has begun to yield concerns the REWARD system and how new information can be used to make predictions about future events. In this instance, bumblebee neurobiology has been successfully modelled to explain FORAGING behaviour and the bee’s ability to predict which type of flower will be a good source of nectar, as well as remember which individual flowers of that type have been already visited. Models of NEURAL NETWORKS have provided insights into various mechanisms, and are important in suggesting experiments for PSYCHOPHYSICS or NEUROPHYSIOLOGY. RECURRENT NEURAL NETS (having both feed-forward and feedback connections) explore the dynamical properties of nervous systems, and suggest mechanisms for such functions as generating representational sequences, decomposing sequences, storing information in the short term, and using stored information to interpret a degraded stimulus.
Much remains to be discovered, and the coevolution of neuroscience and psychology is still in the very early stages. Whether a reduction can be achieved, and how much concep-tual revision will be needed, will depend on the facts of the case.
Bechtel W. & Richardson R. (1993) Discovering Complexity, Princeton University Press: Princeton NJ.
Churchland P.S. (1986) Neurophilosophy, MIT Press: Cambridge MA.
Wimsatt W. (1976) Reductionism, levels of organization, and the mind-body problem. In Brain and Consdousness: Scientific and Philosophic Strategies, ed. G.Globus, G. Maxwell & I.Savodnik, pp. 199–267, Plenum Press: New York.
PATRICIA S.CHURCHLAND
This is the complete article, containing 840 words
(approx. 3 pages at 300 words per page).
