Tuesday, 16 October 2012


‘Reductionism’ is the name of a family of related approaches within epistemology and ontology. That is, it can be a view of how best to seek knowledge of the world by subsuming either higher level theories or explanations under lower level, and more general, theories or explanations. An often cited example is the reduction of Kepler’s Laws to Newtonian mechanics. But it can also be a view about the nature of the world itself, captured in the slogan ‘the physical is all there is’.

In the early part of the twentieth century, epistemological and ontological forms of reductionism were combined in the thesis that the sciences could, ultimately, be unified. The Logical Positivists gave a specific logical interpretation of the conditions for unification. They construed reductionism as a logical or semantic thesis about the relation between theories: in particular the relation between terms and laws in different scientific theories. Broadly, one theory could be reduced to another if its terms could be defined using the terms of the reducing theory and its laws be explained by the laws of the reducing theory. ‘Bridge laws’ would connect higher level types to lower level types. This sort of relation was supposed to have been exemplified by theories of thermodynamics to statistical mechanics [Nagel 1961: 338]. It was also supposed to be the general tendency of scientific progress that ‘higher level’ theories would be reduced to lower level or more basic theories.

Given the construal of reductionism as a claim about the eventual relation between theoretical terms one can see how it also dovetailed with an epistemological approach popular at the time and which still has an echo in contemporary mental healthcare. This involved the separation of theory and observation encapsulated in the ‘two language’ model [Feigl 1970]. If observation is to play a foundational role for assessing theories it has to be neutral between competing theories. The Logical Positivists attempted to ensure this by separating observation and theoretical languages. An epistemological thesis that one can know the truth of an observation statement independently of knowing the truth of any theory was underpinned by a semantic thesis about the meaning of observational and theoretical terms. This required that observation-language concepts could be defined by basic connections to experience presupposing no theory. Theoretical concepts were then supposed to be defined in observational terms.

This is a version of epistemological reductionism. Theoretical language can, and should, be reduced to a more basic and theory-free observational language. The atheoretical aspirations of DSM III and IV are recent instances of this approach. But one of the morals of recent philosophy of science is, however, that theory and observation are of necessity inter-woven [see eg Churchland 1979; Hanson 1958].
Whilst epistemological reductionism has fallen from favour, ontological reductionism and the general aim of the unity of science has not. Whilst not universal, the assumption is still widespread that psychological sciences will eventually be reduced to biology (which might be construed as physiology or evolutionary biology), biology to chemistry and chemistry to physics.

Although the emphasis on a semantic or logical characterisation was replaced instead by a more ontological approach, Paul Oppenheim and Hilary Putnam expressed this view in their classic 1958 paper ‘Unity of science as working hypothesis’.

It has been contended that one manifestly cannot explain human behaviour by reference to the laws of atomic physics. It would indeed be fantastic to suppose  that the simplest regularity in the field of psychology could be explained directly – ie. “skipping” intervening branches of science – by employing subatomic theories. But one may believe in the attainability of unitary science without thereby committing oneself to this absurdity. It is not absurd to suppose that psychological laws may eventually be explained in terms of the behaviour of individual neurons in the brain; that the behaviour of individual cells – including neurons – may eventually be explained in terms of their biochemical constitution; and that the behaviour of molecules – including the macromolecules that make up living cells – may eventually be explained in terms of atomic physics. If this is achieved, then psychological laws will have, in principle, been reduced to laws of atomic physics… [Oppenhiem and Putnam 1991: 407]

Oppenheim and Putnam go on to make two important and distinct kinds claims. Firstly, they argue that the unity of science is served by ‘microreductions’. These are reductions in which:

The objects in the universe of discourse of [the reduced science or theory] are wholes which possess a decomposition into proper parts all of which belong to the universe of discourse of [the reducing science or theory]. [Oppenhiem and Putnam 1991: 407]

In fact they argue more strongly that microreduction is the only method seriously available for the unity of science [ibid: 408]. They then go on to explore the consequences of this view by examining the preconditions for successfully attaining unity via microreduction (see below).

Secondly, they characterise the unity of science as a ‘working hypothesis’. In other word it is not a piece of a priori metaphysics (or corresponding conceptual analysis). It is, instead a high level scientific hypothesis to be tested by its success over time.

That is, we believe that it is in accord with the standards of reasonable scientific judgement to tentatively accept this hypothesis and to work on the assumption that further progress can be made in this direction, without claiming that its truth has been established, or denying that success may finally elude us. [ibid: 408]
Since microreduction is construed as the only serious possibility for the unity of science, and since its success rests on a number of other things being the case, the goal of unification has a number of presuppositions which are then outlined. The list begins:

1.       There must be several levels.
2.       The number of levels must be finite.
3.       There must be a unique lowest level…
4.       Any thing of any level except the lowest must possess a decomposition into things belonging to the next lowest level… [ibid: 409]

This list suggests the following view of nature. The world is made up of basic building blocks or atoms which display regularities that can be described in the law statements of the most basic science. The basic atoms also combine to constitute larger structures which display characteristic regularities of their own. These can in turn be codified in the law statements of higher level sciences. But the higher level regularities do not emerge out of nothing. They can be explained as the consequences of the more basic patterns of behaviour of atoms. So the structure of the world and the structure of science can be seen as two isomorphic hierarchies of levels.

With this picture in place, the most basic level can assume a metaphysical role as a touchstone for what is really real. A good example of this relevant to psychology is an explicit argument for reductionism about mental content offered by Jerry Fodor in his book Psychosemantics [Fodor 1987]. Fodor’s project is to explain the intentionality of mental states. (In the context of the philosophy of thought and language, the term ‘intentionality’ generally means the aboutness or world-involving nature of both mental states and linguistic utterances. This scholastic term was reintroduced to philosophy by Franz Brentano (1838-1917) a philosopher and psychologist.) Fodor puts forward an argument for reductionism about intentionality thus:

I suppose that sooner or later the physicists will complete the catalogue they’ve been compiling of the ultimate and irreducible properties of things. When they do, the likes of spin, charm and charge will perhaps appear upon their list. But aboutness surely won’t; intentionality simply doesn’t go that deep. It’s hard to see, in face of this consideration, how one can be a Realist about intentionality without also being, to some extent or other, a Reductionist. If the semantic and intentional are real properties of things, it must be in virtue of their identity with... properties that are neither intentional nor semantic. If aboutness is real, it must be really something else. [Fodor 1987: 97]

This passage summarises the motivation for reductionism about the mind, and especially mental content, but it applies more generally. If the full extent of what exists in nature is describable by physics (by a future physics, that is), then it should be possible in principle to show how any genuine feature or property will either be found in a list of basic physical properties or can be reduced to some underlying physical description. If not, then it is not a genuine feature of the world. Since meaning or intentionality is a genuine aspect or feature of the natural world it must be reducible to more basic terms which make no essential use of – do not themselves presuppose - the notion of meaning.

It is, however, one thing to argue that reduction of mental to more basic properties must be possible and another to show how that might be done. The most influential version of reductionism about the mental is called ‘representationalism’ of which Jerry Fodor’s Psychosemantics was also a clear statement. He combined the idea that the systematicity and compositionality of thought is explained by structured mental representations or symbols in a language of thought with a variant of a causal theory of how the symbols come to have worldly content or reference. Mental representations mean what causes them.

An alternative form of representationalism accounts for intentionality not merely through a complex causal mechanism, or what typically causes a representation, but through the idea of biological or proper function and thus what the representation is biologically designed to represent. This approach – teleosemantics – came to prominence through the work of Ruth Garrett Millikan who distinguishes between biologically mechanisms which produce representations, such as perceptual systems, and mechanisms which ‘consume’ them [Millikan 1984]. These might include mechanisms designed for predator evasion. It is the contribution that the representation makes to the consumer mechanism that determines its content: such as representing the presence of a predator.

This is an ongoing research programme [Macdonald and Papineau 2006]. But it has recently received a substantial and sustained criticism which threatens to undermine the key idea of mental representations deployed to explain the everyday intentionality of mental states.

In a summary of attempts to naturalise content through the idea of biological representations, Peter Godfrey-Smith expresses pessimism. ‘I doubt that teleosemantics, or any theory like it, will deliver the direct, reductive, puff-of-papal smoke solution that the 1980s literature envisaged’ [Godfrey-Smith 2006: 66]. His reason is that 30 years of philosophical theorising has delivered something in the same area but more basic.
One of the intuitions that has driven teleosemantics is the idea that rich biological concepts of function pick out a special kind of involvement relation between parts of organisms and their environments. Edging even closer to the semantic domain, there is a kind of specificity or directness that an evolved structure can have towards an environmental feature that figures in its selective history… But this relation is found in many cases that do not involve representation or anything close to it. [ibid: 60]

In other words, reductionist accounts of meaning simply have not been forthcoming. A more general critique is provided by William Ramsey in his book Representation Reconsidered [Ramsey 2007]. Ramsey points out that a lesson from the history of the philosophy of mind is that, as Daniel Dennett noted, it is nearly always possible to describe physical processes in representational terms but it is never necessary. Just as one can avoid biological descriptions by describing biological systems in lower level physical and chemical terms, so representational terms need not be used. What then justifies the use of representational terms for complex systems? What is the explanatory benefit? Ramsey argues that all the dominant approaches to explaining intentionality fail this test.

Take the case of a Venus fly trap which, according to Fred Dretske’s teleosemantic analysis, is supposed to have an internal trigger which responds to movement and thus signals the presence of insects [Dretske 1988]. Drekske says ‘there is every reason to think that this internal trigger was selected for its job because of what it indicated, because it told the plant what it needed to know.’ [ibid: 20]. But Ramsey asks why we need think of this in representational terms in addition to thinking that because of the lawlike connection between movement and plant closure plants with the trigger would be selected. ‘[T]here is no reason to think that structures recruited because their states have the property of being nomically dependent on some condition are also recruited because they carry information about that condition’ [Ramsey 2007: 135]. This mirrors the case of the firing pin in a gun which bridges the gap between pulling the trigger and firing the round. That lawlike connection is why the pin is part of the design. But there is no reason to think that the firing pin is a representation of anything.

Such criticism threatens the prospects for a reductionist account of intentionality or mental content. But that is not yet to diagnose a mistake in Fodor’s argument that, in principle at least, a reductionist account must be possible. Anti-reductionists may seem to have to deny the reality of mental states in general and intentionality in particular [cf Churchland]. But there is another option which is to reject the implicit premiss that physics alone sets limits to what is real. John McDowell, for example, argues that that whilst scientific method has been a genuine achievement of the modern era, the assumption that scientific descriptions of the world also exhausts its nature is ‘not the educated common sense it represents itself as being; it is shallow metaphysics’ [McDowell 1995b: 164; 1998b: 182 cf McDowell 1994: 82]

The cost, however, of such an anti-reductionist view of the mental is that it can seem to make the relation between mind-presupposing person-level descriptions and more basic sub-personal mechanisms mysterious in principle. Wilfrid Sellars (1912-89), for example, argued that the central task for philosophy was to reconcile the ‘manifest image of man in the world’ with the scientific image, derived from ‘postulational theory construction’ [Sellars 1963: 19]. Sellars is not alone in taking there to be a key distinction between person-level descriptions and underlying natural scientific accounts. Building on Sellars’ work, McDowell, for example, contrasts the logical space of reasons with the realm of law or of natural science [McDowell 1994]. Donald Davidson argues for the ‘constitutive ideal of rationality’ which has ‘no echo in physical theory’ [Davidson 1980: 223, 231]. All three authors share an assumption about the central importance of the normativity of person-level descriptions and thus face, in different ways, the challenge of explaining how two such distinct levels can both be a part of the same natural world.

One frequent claim is that, whilst distinct, the mental supervenes on the physical. The idea of supervenience was originally articulated by GE Moore (1873-1958) to relate moral properties to physical properties given the assumption that moral properties cannot be analysed into physical descriptions.

[I]f a given thing possesses any kind of intrinsic value in a certain degree, then not only must that same thing possess it, under all circumstances, in the same degree, but also anything exactly like it, must, under all circumstances, possess it in exactly the same degree. … it is not possible that of two exactly similar things one should possess it and the other not, or that one should possess it in one degree, and the other in a different one. [Moore 1922: 261]

More recently, it has been used most often in the philosophy of mind. It claims that determining the physical properties of a system determines its mental properties but not vice versa. It was most famously asserted by Davidson to further characterise his token identity theory which claims that every particular mental event is a physical event even though types of mental event cannot be identified with types of physical event [Davidson 1980: 207-27]. But as John Haugeland pointed out, it need not be associated with a position even as apparently weakly reductionist as a token identity theory [Haugeland 1982]. It could be used to characterise a yet weaker relation between mental and physical descriptions as a whole even where there is no claim that the same individuals (entities or events) can be identified in both. Haugeland calls this ‘weak supervenience’ (although that ambiguous label is also used for modally constrained versions of supervenience). There remains, however, some doubt as to whether supervenience is a stable middle point between full blown reductionism and some form of dualism [Evnine 1991, Thornton 2009].

Accommodating mental content into a scientific account of nature is not the only opportunity for a reductionist account of clinical psychology. A distinct second area concerns the nature of pathology itself. Can facts about illness be explained in underlying and purely biological terms or are they essentially evaluative and thus resistant to such a reduction as, for example, Bill Fulford argues [Fulford 1990]? The most promising weapon in the reductionists’ armoury is – like teleosemantics – the idea of the biological function of a trait or mechanism. Thus Jerome Wakefield argues that a disorder is a biological dysfunction which can, itself, be explained in value-free biological terms using evolutionary theory [Wakefield 1999]. He himself retains a role for the value harm in picking out the composite notion of illness or disease as a harmful dysfunction. It remains a matter of debate, however, whether this reductionist analysis is correct [eg Bolton 2008, Thornton 2000].


Bolton, D (2008) What is Mental Disorder? Oxford: Oxford University Press
Churchland, P. M. (1979) Scientific realism and the plasticity of mind, Cambridge: Cambridge University Press.
Davidson, D. (1980) Essays on Actions and Events, Oxford: Oxford University Press.
Dretske, F. (1988) Explaining Behavior, Cambridge, Mass.: MIT Press
Evnine, S. (1991) Donald Davidson, Oxford: Polity
Feigl, H. (1970) ‘The orthodox view of theories’ in Radner, M., Analysis of Theories and Methods of Physics and Psychology, Minnesota Studies in Philosophy of Science Vol. 4 Minneapolis: University of Minnesota Press.
Fodor, J. A. (1987) Psychosemantics: the problem of meaning in the philosophy of mind, Cambridge, Mass.: MIT Press.
Fulford, K.W.M. (1990) Moral theory and medical practice Cambridge: Cambridge University Press
Godfrey-Smith P. (2006) ‘Mental Representation and Naturalism’ in Macdonald, G. and Papineau, P. (eds) Teleosemantics, Oxford: Oxford University Press pp. 42-68
Hanson, N. R. (1958) Patterns of Discovery, Cambridge: Cambridge University Press.
Haugeland, J. (1982) ‘Weak Supervenience’ American Philosophical Quarterly 19.
Macdonald, G. and Papineau, P. (eds) (2006) Teleosemantics, Oxford: Oxford University Press
McDowell, J. (1994) Mind and World, Cambridge, Mass.: Harvard University Press.
McDowell, J. (1998) Mind value and reality, Cambridge, Mass.: Harvard University Press
Millikan, R.G. (1984) Language, thought and other biological categories. Cambridge, Mass.: MIT Press.
Moore, G.E. (1903) Principia ethica. Cambridge: Cambridge University Press.
Oppenheim, P and Putnam, H. (1991) ‘Unity of science as a working hypothesis’ in Boyd, R. Gasper, P. and Trout, J. D. (eds.) Philosophy of Science London : MIT Press.
Ramsey, W.M. (2007) Representation Reconsidered, Cambridge: Cambridge University Press
Sellars, W. (1963) ‘Philosophy and the Scientific Image of Man’ in Science, Perception and Reality, London: Routledge & Kegan Paul
Thornton, T. (2000) ‘Mental Illness and Reductionism: Can Functions be Naturalized?’ Philosophy, Psychiatry and Psychology 7: 67-76
Thornton, T. (2009) ‘On the interface problem in philosophy and psychiatry’ in Bortolotti, L. and Broome, M. Psychiatry as Cognitive Neuroscience, Oxford: Oxford University Press: 121-136
Wakefield, J.C. (1999) Mental disorder as a black box essentialist concept. Journal of Abnormal Psychology 108, no 3: 465-472.