Reductionism, Embodiment, and the Generality of Psychology

Lawrence A. Shapiro

University of Wisconsin -- Madison

A central controversy in philosophy of psychology pits reductionists against, for lack of a better term, autonomists. The reductionist’s burden is to show that psychology is, at best, merely a heuristic device for describing phenomena that are, when speaking more precisely, just physical. I say “at best,” because reductionists are prone to less conciliatory remarks, such as: “psychological property P just is physical property N, so scientific explanation might as well focus exclusively on N,” and “psychological property P is nothing other than N, so generalizations about N suffice to say all that there is to say about P,” and “knowledge of all the N facts suffices for knowledge of all the P facts.”

The autonomist’s burden, on the other hand, is to explain why psychology should not be reduced to physics given the naturalistic assumption that any two objects alike in all their physical properties will also be alike in all their psychological properties, and that these physical properties completely determine the psychological properties.¹ If the naturalistic assumption is true, and most autonomists do accept it,² then the task of defending a coherent conception of psychology’s autonomy becomes hard indeed. As matters stand, it certainly looks as if a complete physical reckoning will say all that needs to be said about psychological phenomena.

Traditionally, autonomists have met their burden by appeal to the purported multiple realization of psychological states.^3,4 Because, they claim, psychological properties can be realized by different kinds of physical properties, and disjunctions of physical properties are not themselves physical properties, psychology cannot be reduced to physics. It is false that psychological property P is nothing other than physical property N. Generalizations about N do not suffice to say all that there is to say about P. Knowledge about N facts does not suffice for knowledge about P facts. Psychology is autonomous because psychological kinds are invisible from the perspective of physics. Whatever psychological generalizations are available will emerge from the study of these kinds, and physics has no place in this investigation.

Familiarity with the reduction literature leaves the impression that autonomists see themselves as the good guys. The multiple realization argument is a Declaration of Independence and the autonomists are fighting to maintain the sovereignty of the special sciences. However, this posturing is somewhat mystifying. Reductionist pursuits have not been undertaken with the goal of shutting down valuable scientific programs. Rather, the point of reduction is to gain generality and unification (Nagel 1961). Reduction buys generality, for if it is true, say, that water is H₂O, then whatever is true of one sample of water will be true of all samples of water. Similarly, if water is H₂O we then have an explanation for why water exhibits a variety of properties whose association might otherwise appear accidental (freezing point, boiling point, density, specific gravity, and so on). Surely the generality and unification that reduction provides is valuable, and thus autonomy, if it truly does entail a loss of these things, is not so clearly a desirable end.

Jaegwon Kim (1992; 1998b) especially has emphasized the cost of autonomy. According to Kim, if higher-level properties are multiply realizable, then generalizations at the higher-level are possible only within narrow confines. For instance, Kim denies that there can be psychological generalizations that cross species lines. Human psychology and Martian psychology, assuming that human beings and Martians realize their psychological processes in distinct physical ways, could share no laws. This follows, Kim thinks, because differences in psychological realization prevent similarities at the psychological level. Differences, as it were, trickle up. Kim’s conclusions are important because they directly challenge the presumed benefit of multiple realization. Multiple realization, Kim argues, may bring autonomy, but the autonomy brings with it an acute constriction in domain. Given my previous remarks about the goals of reduction, perhaps it would not be so surprising if the troubles that multiple realization makes for reduction also cause strain in the idea of an autonomous and general psychology.

In this chapter I consider two challenges to the generality of psychology. The first – Kim’s – depends on an argument that I do not think is sound. However, I want also to consider a body of psychological research that suggests an independent reason for thinking that psychology cannot be general. This research, sometimes labeled embodied cognition, may tell against the possibility of psychological generalizations that span species. If certain claims that embodied cognition theorists make are correct, I think the prospects for a general psychology are slim, but for reasons other than those that emerge from the multiple realization thesis. Kim’s vision of a fractured psychology may be accurate, but not for his reasons.

1. Multiple Realization and Autonomy

Autonomy, in this context, is the idea that the special sciences have explanations, kinds, and subject matters that are unavailable or inaccessible to physics. For instance, consider the psychological predicate ‘believes that the lake is frozen.” Defenders of autonomy hold that this predicate is not simply another name for a physical property. The psychological property is not identical to any physical property, and so explanations that appeal to the psychological property have no physical counterpart. Thus, even a completed physics could not replace psychology. Physics is about the physical domain, psychology is about the psychological domain, and never the twain shall meet.

As I mentioned above, the justification for belief in autonomy comes from the apparent multiple realization of psychological kinds. The human belief that the lake is frozen is realized in a brain state, whereas the Martian’s and robot’s beliefs that the lake is frozen are realized in different kinds of physical states. The motivation for claiming that the human being, Martian, and robot all have the same kind of belief is the functional similarity these agents exhibit. Each agent, for instance, if confronted with the same stimuli, forms the same desires, engages in the same actions. Counterfactuals true of one, e.g. if she had not believed that the lake was frozen then she would not have stepped onto it, are true of the others. The three are psychologically identical despite their physical differences.

To summarize, the connection between multiple realization and autonomy rests on the following argument.

1. Psychological properties are multiply realizable.

2. If psychological properties are multiply realizable then psychological generalizations over these properties have no physical counterparts.

3. Psychological generalizations have no physical counterparts.

Therefore, psychology is autonomous.

2. Disintegration from Below

Jaegwon Kim (1992; 1998b) has argued that the multiple realizability on which the defense of autonomy rests is, in fact, a cause for the disintegration of psychology as well as the other special sciences.⁵ His argument begins with two principles:

Causal Individuation of Kinds: Kinds in science are individuated on the basis of causal powers, i.e. objects share a property insofar as they have similar causal powers (1992, 17).

Causal Inheritance: If mental property M is realized in a system at t in virtue of physical realization base P, the causal powers of this instance of M are identical with the causal powers of P (1992, 18).

The first principle, Kim thinks, is “plausible,” and, “in any case, widely accepted,” (1992, 17). Of course, this is not to say that all philosophers would accept it, and those who have defended externalist theses about mental content, e.g. Burge (1989), may well reject it. Nevertheless, the principle does have intuitive appeal. If instances of two different kinds share all their causal powers, one must wonder why it is necessary to distinguish the kinds. Certainly the kinds could not be experimentally distinguished, but then by what other criteria might a scientist draw the distinction?

The second principle also aligns neatly with intuition. If a vase is realized in glass, then it will break just when the glass will break, will melt just when the glass will melt, will weigh just what the glass weighs. The vase, in other words, will have all and only the causal powers of the glass that realizes it. To deny this, Kim thinks, commits one to the existence of “causal powers that magically emerge at a higher-level and of which there is no accounting in terms of lower-level properties and their causal powers and nomic connections” (1992, 18).

I propose to grant Kim his principles. More interesting than the principles is their implication. Kim believes that acceptance of the principles leads to the fracturing of psychology into a collection of individual, species-local, psychological theories. I shall call Kim’s argument for this conclusion Fracture. The first premise of Fracture is this:

(1) Mental kind M is multiply realized by distinct physical kinds P_h and P_m.

This premise is tantamount to a statement of autonomy. Kind M is a kind despite its multiple realization because, presumably, there are generalizations involving kind M that cannot be captured in the language of the realizers’ domains. From the perspective of the science of the realizers, these higher-level generalizations are invisible. This is why psychology is indispensable.

Premises (2) - (4) spell out some consequences of Kim’s principles:

(2) By Causal Individuation of Kinds, the causal powers of P_h must differ from the causal powers of P_m.

(3) By Causal Inheritance, the mental state P_h realizes cannot have all the same causal powers as the mental state P_m realizes.

(4) Thus, again by Causal Individuation of Kinds, P_h and P_m cannot realize the same mental state.

Premise (4) captures the “trickle up” effect. Differences in the causal powers of the realizers trickle up to create differences in the realized kinds. Given the trickle up effect, it follows:

(5) Therefore, “[e]ach mental kind is sundered into as many kinds as there are physical realization bases for it, and psychology as a science with disciplinary unity turns out to be an impossible project” (1992, 18).

(6) Hence, M cannot be multiply realized (from the contradiction between 1 and 5).

As stated, Fracture is a reductio. From the assertion of multiple realization, which, recall, is a statement of autonomy, and Kim’s two principles, it follows that multiple realization of psychological kinds is in fact impossible. Far from a justification for autonomy, multiple realization proves to be the undoing of the special sciences.

As I stated earlier, I do not think that Fracture is sound. This is because premise (4) – the trickle up premise – is false. Underlying this premise is an unidentified principle, which I shall call Relevance:

Relevance: all causal powers of a realizer are relevant to the individuation of the kind that it realizes.

Relevance justifies the trickle up effect. If Relevance is true, then because P_h and P_m differ in their causal powers, the kinds that they realize must differ as well. Clearly, however, Relevance is not true. Consider the corkscrews I discussed in my (2000). The steel and aluminum double-lever corkscrews differ in many of their causal powers in virtue of their difference in realization. But, relative to their end as corkscrews, the fact that one might bend under two thousand pounds of pressure and the other under three thousand pounds, or that one might melt at a higher temperature than another, and so on, is irrelevant. Indeed, one might plausibly hold that no two instances of a functional kind are ever realized identically and therefore adoption of Relevance renders the notion of a functional kind incoherent. For instance, two steel corkscrews, one with a molecular impurity that the other lacks, must, according to Relevance, be different kinds.

But one doesn’t have to make up examples to see the problem with Relevance. It is routine in science that schemes of individuation cross-classify each other. Such cross-classification is possible because the causal powers of interest in some scientific domains can be ignored in others. The chemist might be interested in DNA for its acidic properties. The geneticist’s focus on DNA traces to its suitability as a unit of heredity. Perhaps the chemist doesn’t even know that DNA plays this biological role; and the geneticist is ignorant that DNA is electron deficient. The causal powers that make DNA an acid are of no concern to the geneticist and the causal powers that make DNA the unit of heredity do not matter to the chemist. The point is that not all causal powers of a kind are relevant to the kind’s individuation. DNA can simultaneously instantiate different kinds – acid, unit of heredity – because it has causal powers that are both relevant and irrelevant to its various individuations.

Moreover, it is precisely because Relevance is false that chemists can say of both DNA and HCl that they are acids. The discovery that DNA plays a role in heredity did not force the chemist to jettison it from his list of acids. Acids can and do differ in their causal powers. This is ok and consistent with what we know to be true: objects can instantiate many different kinds all at once.

So, what happens to Fracture once we recognize that the same object may instantiate a property in the domain of one science but not another (e.g. HCl counts as a kind in chemistry but not in genetics) and that the same object may count as one kind in one science but a different kind in another (DNA is an acid in chemistry but not a gene, and it is a coder for proteins in genetics but not an acid)? If Relevance is false then the trickle up premise loses its force. There is no longer a reason to think that differences in realization must trickle up to differences at higher-levels, and so no reason to suspect that the multiple realization of M is impossible.

Without (4), Fracture’s conclusion no longer follows. To deny trickle up is to embrace the possibility that, relative to psychology, a set of distinct physical kinds {R₁, R₂, …, R_n} can realize a single mental kind. Given this, psychology can be at once autonomous and also general. Psychological generalizations are free to range over different species because physical differences in species do not entail psychological differences.

3. Disintegration from Above

Kim locates the fracturing of psychology in the alleged fact of multiple realization. I have just argued that multiple realization does not lead to this conclusion. However, there may be another reason to question whether psychology can be general or, alternatively, whether psychology must face disintegration. In the remainder of this chapter I would like to consider a motivation for psychological disintegration that comes from the field of embodied cognition (henceforth EC). This research suggests its own kind of challenge to a general psychology. I will call the argument that presents this challenge Body Determinism:

(1) EC lends support to the idea that the body and environment are constitutive of cognition.

(2) If EC lends support to the idea that body and environment are constitutive of cognition, psychology must partition its domain into species-specific subject matters.

Therefore, psychology must partition its domain into species specific subject matters.

Body Determinism differs considerably from Fracture. Fracture depends on a couple of metaphysical principles about individuation and causal inheritance, as well as a tacit premise about the role of relevance in individuation. Body Determinism shares Fracture’s conclusion, but, unlike Fracture, it is ultimately empirical. To be sure, philosophical considerations will be prominent in the argument. In particular, questions about individuation remain important. However, unlike in the case of Fracture, the motivation for Body Determinism derives from a corpus (no pun intended) of empirical work.

Also distinguishing the arguments is their intended breadth. Kim thinks Fracture applies not only to psychology, but to any science that includes functional kinds in its domain. On the other hand, Body Determinism is just about psychology. This difference in scope reflects the difference that empirical considerations make in each argument. Fracture has an a priori feel, and confirming this is the fact that the empirical details of a science make no difference to whether Fracture applies to it. Any science, if it has multiply realizable kinds, will end up fractured. However, because Body Determinism takes its motivation from data about how the mind works, it has no application to sciences that are not about minds.

With preliminary remarks behind us, it is time to consider the first premise of Body Determinism. The following quotations from philosophers and psychologists, while not exactly clarifying the first premise, at least give it expression in ways that can be further illuminated.

[T]he peculiar nature of our bodies shapes our very possibilities for conceptualization and categorization (Lakoff and Johnson 1999, p. 19).

If perception is in part constituted by our possession and exercise of bodily skills – as I argue in this book – then it may also depend on our possession of the sort of bodies that can encompass those skills, for only a creature with such a body could have those skills (Noë 2004, p. 25).

The cognitive processing that gives rise to mental experience may be something whose functioning cuts across the superficial physical boundaries between brain, body, and environment (Spivey et al. 2004, p. 178).

I will argue that many cognitive capacities in symbol-using creatures, far from being purely internal, are either enactive bodily capacities, or world-involving capacities. These capacities are not realized by some internal arrangement of the brain or central nervous system, but by embodied states of the whole person, or by the wide system that includes (parts of) the brain as a proper part (Wilson 2004, p. 188, his italics).

Common to these passages is the idea that an organism’s body shapes an organism’s mind or mental capacities. This idea of shaping, however, requires scrutiny. There are many ways that an organism’s body might shape its mind that have no special interest. An empty stomach might cause an organism to desire food. Having two eyes allows depth perception by the detection of disparities in the retinal images. Having a head that can rotate on a neck opens perceptual opportunities that a stationary head would lack.

These examples of shaping, however, all point to the body as a cause of or influence on cognitive capacities. But, of course, it has long been standard in cognitive science to grant the body this kind of role in cognition. A cognitive scientist wedded to the idea that the mind is realized entirely in the brain, or supervenes entirely on the brain, will explain the above examples by appeal to a notion like transduction. The empty stomach causes a desire for food only in virtue of a process by which the stomach’s emptiness becomes represented in a code that the brain can understand. Similarly, the retinal images or the motions of the head can cause or influence perception only by their translation into a neural code. Once head motion, for instance, becomes encoded, it is accessible to the cognitive processes in the brain that result in perception. According to this internalist picture, the body does shape cognition, but only by a process of transduction that converts bodily goings on into a form on which the brain can operate.

Of course, the authors above recognize all this.⁶ They see the body as something much more than just a causal influence on the mind. Lakoff and Johnson, for instance, argue that human beings are equipped with a class of basic concepts that, through metaphorical reasoning, can give rise to more sophisticated concepts. The concept of HAPPY, they speculate, is rooted in the basic concept UP; SAD derives from DOWN. Through application of metaphorical reasoning, UP and DOWN structure the concepts HAPPY and SAD: “I’m feeling up. That boosted my spirits. My spirits rose. You’re in high spirits. Thinking about her always gives me a lift. I’m feeling down. I’m depressed. He’s really low these days. If fell into a depression. My spirits sank (1980, p. 15). Crucially, the reason HAPPY and SAD are structured in terms of UP and DOWN has to do with the nature of our body. “Drooping posture,” Lakoff and Johnson suggest, “typically goes along with sadness and depression, erect posture with a positive emotional state,” (1980, p. 15).

Because my concern is less with the tenability of Lakoff and Johnson’s claims than it is with their consequences for the nature of psychology, I do not intend to discuss them critically here. For present purposes, Lakoff and Johnson’s work suggests a sense in which the body shapes cognition that goes beyond the banal claim that the body has a causal influence on cognition. If Lakoff and Johnson’s account of concept acquisition is correct, it follows that the body informs cognition by constraining how an agent can conceive the world. This is perhaps easiest to see in a fanciful case: “[i]magine a spherical being living outside of any gravitational field, with no knowledge or imagination of any other kind of experience. What could UP possibly mean to such a being?” (1980, p. 57). With no concept of UP available to this being, its concepts of HAPPY and SAD would be nothing like our own. If possible at all, they would be structured by entirely different metaphors.

One might insist that the body’s properties are indeed causes of the concepts HAPPY and SAD, but I hope it is clear that, if so, they are not causes in the same sense that an empty stomach or a moving head contributes causally to cognition. In the latter cases, the empty stomach and the moving head are encoded in a fashion that allows the information they carry to be integrated into a computation resulting in a desire for food or a perception of (perhaps) depth. But in the former case the body is a cause in the way that a spout’s shape is a cause of the volume of liquid it can pour in a given amount of time. The spout’s shape constrains the flow of liquid but it needn’t be encoded to have this effect on the volume of flow. Just as spouts that differ in shape will differ also in the volume of liquid they allow to pass in a given time, so, according to Lakoff and Johnson, organisms that differ in bodily shape will differ as well in the kinds of concepts they can acquire.

The spout analogy is helpful in another sense. Clearly two spouts that differ in the volume of liquid they pour must differ as well in other properties. That is, their pouring capacity supervenes on properties like their size, surface texture, and so on. We might take Lakoff and Johnson’s discussion of the spherical being to imply that concept acquisition supervenes not just on neural properties, but also on bodily properties. This makes clear the difference between the body’s being a causal influence on the mind and the body’s being something more like a constituent of the mind. Concepts like HAPPY and SAD, if Lakoff and Johnson have correctly described the manner of their acquisition, supervene not just on the brain but on the body as well. Of course, this does not imply that differences in body must entail differences in concepts, but it does reveal that differences in body are the kinds of differences that can suffice for differences in concepts.

Noë (2004) too makes a case for conceiving of the body as a constituent of cognition. Noë defends a skill-based account of perception, according to which perception is possible only in virtue of acquaintance with a special class of sensorimotor skills. The difference between a skill-based account of perception and traditional representational account can be brought out with the following example. Consider two strategies for moving through narrow passageways without banging your head on the walls to either side. On a traditional account, one forms a representation of the passageway, a representation of one’s position relative to the passageway, and computes a course that will take one safely through. In effect, once the representations have been constructed, one could proceed with one’s eyes closed. On the skill-based account of perception that Noë favors, one might orient oneself in front of the passage so that the amount of opening to the left and right always appears to be the same. This insures that one is moving toward the center of the opening. This second strategy does away with or minimizes the need for computations over representations, demanding instead the exercise of a particular skill.

On the skill-based model, one forsakes representation-guided computation for world-guided activity. Aligning oneself with the center of the passageway requires constant calibration with the amount of open space that appears to one’s left and right. If more space appears to the right, one must move toward the right in order to create more open space on the left. This kind of interaction with the world constitutes a kind of knowledge. One knows, tacitly at least, that one will pass through the center of the doorway if one evenly balances the amount of opening that appears to the left and right. This knowledge, however, leads to success for purely contingent reasons. For one thing, it would not work for organisms whose bodies were extensively asymmetrical. An organism whose head was fastened to the right side of its body would deploy the strategy only at risk of bruising its left side. Likewise, if the world were such that passageways were vast, the strategy would seldom be of practical value.

I have made up the above example to illustrate the flavor of the theory of perception that Noë wishes to defend (but, for all I know, human beings really do adopt such a skill-based strategy to navigate successfully through doorways). One can take Noë’s project as an effort to see how far skill-based accounts can go in explaining all of perception and cognition. Insofar as perception and cognition are skill-based, Noë believes that this makes the body intrinsic to psychology:

In general it is a mistake to think that we can sharply distinguish visual processing at the highly abstract algorithmic level, on the one hand, from processing at the concrete implementational level, on the other. The point is not that algorithms are constrained by their implementation, although that is true. The point, rather, is that the algorithms are actually, at least in part, formulated in terms of items at the implementational level. You might actually need to mention hands and eyes in the algorithms! (2004, p. 25, his italics and excitement).⁷

It is tempting to think that Noë overstates his case here. Why is it necessary to mention hands and eyes in the algorithms involved in visual processing? I can imagine two ways to resist this claim. First, one might insist that hands and eyes can affect visual processing only by virtue of an encoding of the information they provide. It is something like a category mistake to think that an eye – a spherical chunk of proteins – can be part of an algorithm that takes as input light reflected from surfaces and produces as output a visual scene of the world. How can an eye “get into” such an algorithm? On the other hand, there is no problem imagining that representations of the eye’s properties can participate in an algorithm. If, say, the eye’s motion contributes something to perception, this can be only because this motion is represented in a code that plays a role in the visual processing algorithm.

Second, one might see hands and eyes as merely facilitating visual perception. Movements of the eyes and hands might facilitate the perception of occluded objects, or distance, or shape. But this is not to say that eyes and hands are literally part of visual perception. They are no more part of visual perception than eyeglasses are. They provide opportunities that make possible, or in some way enrich, the processes that yield visual perception (Clark forthcoming). On this view, Noë errs in extending the realm of psychology beyond its proper borders. He mistakes the track for part of the train just because trains can’t move without tracks.

Perhaps neither of these objections to Noë are easy to dismiss. However, it is clear that there is a fair amount of question-begging going on. For instance, to deny that eyes and hands are part of the visual process already assumes something about the nature of visual processing. Noë’s plaint is that vision, as well as other psychological capacities, has been misconceived. Hands and eyes might be part of vision if one takes vision to be a skill-based activity – an activity that requires the use of sensorimotor knowledge. This response seems also to meet the challenge of the first objection. To insist that visual processing can make use only of suitably encoded information is again to pre-judge the nature of psychological processing.

As with all disputes that involve charges of question begging, the winner will be the one who makes his position superior on independent grounds. If Noë’s skill-based account of cognition answers questions that representationalist accounts do not, if it resolves puzzles, makes predictions that would be surprising on more traditional accounts, and so on, then this is reason to take the account very seriously. Of course, I am not in a position to decide this issue. Fortunately, as was the case in my discussion of Lakoff and Johnson’s work, I am less interested in whether skill-based accounts will hold up in the long run than I am with tracing their consequences for the possibility of a unified psychology. I will turn to this issue after some more remarks about how to understand premise (1) of Body Determinism.

The quotations from Spivey et al. (2004) and Wilson (2004) suggest another sense in which the body or environment is constitutive of cognition. I think Wilson is clearest here. Wilson describes processes of wide computation. Computation is wide when it involves representational or informational states that are external to the brain. Examples are easy to come by. Hutchins (1995) describes the navigator’s use of navigational instruments like a sextant to compute location. Wilson discusses the use of pencil and paper to aid in arithmetic tasks. Clark and Chalmers (1998) ask us to consider the role of a pocket diary in an Alzheimer’s victim’s interaction with the world. Each of these cases involves a process in which an external prop plays a computational role.

Consider, for instance, the use of pencil and paper in computing the product of 683 and 548. One could do this computation without benefit of external aids. However, most of us would have an easier time with the task if we could “offload” steps in the computation to a piece of paper. The marks on paper are, in a completely literal sense, parts of the multiplication process. The steps involved in this process are realized in not just the brain, but in the brain and the marked paper.

Clark and Chalmers (1998) provide a similar example. Otto, who has Alzheimer’s disease, relies on his notebook to provide him with the information that MoMA is on 53^rd St. This information in his notebook, Clark and Chalmers claim, can play the same causal role for Otto that Inga’s in-the-head belief that MoMA is on 53^rd St. plays for her. “The moral, they say, “is that when it comes to belief, there is nothing sacred about skull and skin. What makes some information count as a belief is the role it plays, and there is no reason why the relevant role can be played only from inside the body” (1998, p. XX).

The point to which proponents of embodied cognition wish to draw attention is that once one gives up a reductionist view of mind – a view of mind that identifies mental properties with physical properties – and adopts instead a functional conception of mental states, one cannot avoid the consequence that things external to the brain might qualify as mental states. If the criteria for mental states are specified functionally, then there is nothing to prevent bodily or environmental states from being mental. There is, in other words, no reason to align the mental boundaries with the neural boundaries. The supervenience base of the mind, or, at any rate, cognitive processes, can include items as diverse as notebooks, fingers, calculators, PDAs, and memory sticks.

Perhaps at this point I have said enough to convey the gist of premise (1) of Body Determinism. Again, my goal is not to defend premise (1), but to consider its consequences for the science of psychology. With this in mind, it is time to consider premise (2) – the claim that embodiment implies that psychology must fracture itself into as many sciences as there are relevantly different species.

Suppose it is true that cognition is embodied in the way that Lakoff and Johnson (1980) surmise, i.e. that anatomical facts constrain and influence the contents of thought, and that the contents of thought supervene on these anatomical facts. Also suppose that perception is best conceived as intrinsically sensorimotor, requiring skillful interactions between body and environment, as Noë believes. Further suppose it is true that much of cognition is wide in the sense that Spivey et al. (2004), Wilson (2004) and Clark and Chalmers (1998) describe. Beliefs, on such a view, can exist outside the head and, indeed, outside the body. The mind’s realization can extend beyond the brain and into the world. How do these points bear on the nature of psychology?

One might take the claims from EC as supporting a view of psychology very much like the one that Fracture envisions. Just as Kim thinks that psychology must disintegrate into species-specific sciences because differences in realization trickle up to create differences in psychology, a proponent of EC might think that because the kind of body an organism has makes a difference to the kind of psychology it has, psychology must splinter itself into as many sub disciplines as there are types of body.

At this point the difference between the sort of claim that Noë and Lakoff and Johnson make and that which Wilson and Clark and Chalmers make grows in prominence. If differences in body type are to suffice for the fracturing of psychology, then the body must have a role in cognition that is deeper than mere realization. For, consider again Clark and Chalmer’s claim that the entries in Otto’s notebook qualify as beliefs. The justification for this claim is that the notebook entries are functionally identical to beliefs. Hence, if one is a functionalist, then this stipulation makes it trivial that the notebook entries are beliefs. However, functionalism supports the view that psychology can be general. The functionalist is untroubled by the failure of reduction and takes psychological generalizations to range over mental states that are individuated in terms of their causal role in a network of inputs, other mental states, and outputs. Wide computationalists like Wilson and Clark and Chalmers are simply functionalists who have come to appreciate that the states which play “mental” causal roles need not all be in the head.⁸

On the other hand, Lakoff and Johnson and Noë suggest that bodily properties are sometimes constitutive of cognition. To have concepts like HAPPY and SAD that bear any resemblance to these concepts as human beings understand them requires a human-like body. The human body is not just the realization base for these and other concepts, but it is in virtue of having a human body that human concepts mean what they do.⁹ Or, for Noë , perception is the disposition to respond to features of the environment in ways particular to a certain sort of body. Having a different sort of body necessitates a different kind of perception, hence, “only a creature with a body like ours can have experiences like ours” (Noë 2004, p. 26).

An analogy may be useful for clarifying Lakoff and Johnson and Noë’s view. Consider Whorf’s principle of linguistic determinism. Stated roughly, this principle claims that the structure of the language one speaks determines the kinds of thoughts one can have. Individuals who speak languages with distinct grammatical categories may as a result conceptualize the world differently. My interest is not in the truth of Whorf’s principle, but in its implication for psychology. If the principle is true, a psychologist interested in how an individual conceives his world would have to attend to the structure of an individual’s language. Insofar as an individual’s mind bears the stamp of his language, psychology cannot proceed independently of linguistics.

One can usefully think of Body Determinism as linguistic determinism writ large. Insofar as an individual’s mind bears the stamp of its body, psychology cannot proceed independently of an investigation of the body. But Body Determinism has more significant consequences for psychology than does linguistic determinism. Linguistic determinism does not entail the fracturing of psychology because it does not challenge the idea that members of differing linguistic communities may nevertheless share some psychological capacities. Indeed, this must be so, for the thesis of linguistic determinism is interesting only if differences in some psychological capacities, e.g. categorization, are the product of linguistic differences rather than the result of more fundamental psychological differences. Hence, linguistic determinism presumes a level of psychological uniformity.

However, Body Determinism is not so committed. If Lakoff and Johnson and Noë tell a coherent story of the body’s influence on the mind, then there need not be any universal psychological capacities. In such an event, psychology has no choice but to give up hope of generality in favor of species-specific regularities that bear all the idiosyncrasies of the organisms over which they range. Psychology disintegrates into myriad subfields, each dedicated to the psychology of a particular species.

4. Conclusion

The relationship between reduction and autonomy is complicated. Functionalists see autonomy as emerging from the failure of reduction. Reduction fails because psychological kinds are multiply realizable; and because reduction fails psychology is autonomous. However, because reduction carries with it the promise of unification and generality, one might worry that as it goes so go these virtues. Kim’s Fracture argument develops this worry. However, I have argued, Fracture fails because it rests on a relevance assumption that is false. Kinds exhibit a variety of causal powers, and the individuation of kinds needn’t take into account all of them. This means that differences between realizers don’t necessarily trickle up into differences at higher levels. Different chemical kinds can realize the same psychological kinds. Consequently, reduction can be false while a general psychology remains possible.

But what if, as some researchers in EC have argued, psychological processes are embodied in the sense that they intrinsically comprise bodily processes? I have argued that this might lead to the disintegration of psychology. No longer could psychology generalize over differently embodied organisms if these differences are constitutive of differences in psychology. For each species there would be sui generis psychological laws. This consequence, I think, sets a course into uncharted territory. Philosophers have tended to see the choices between reduction and autonomy as mutually exclusive but also as exhaustive. Reductionism is either true or false, and psychology is autonomous if only if it is false. But the kind of disintegration of psychology that EC suggests embraces neither of these options. Here we have reduction of a sort. Psychological identity implies bodily identity. But we also have autonomy of a sort. Organisms that differ physically may nevertheless be psychologically similar if they share the relevant morphological and physiological properties.¹⁰

I wish to emphasize again that I have not attempted to defend the conception of psychology that leads to what I have called Body Determinism. I have merely sought to reveal consequences that EC may have for the unity of psychology. So, are the claims of EC that lead to the disintegration of psychology true? Without access to alien beings, the claims are difficult to test. Perhaps work on non-human animals may someday bear results. On the other hand, the principle of linguistic determinism remains controversial to this day, despite ready access to members of distinct linguistic communities.

References

Block, N. (2005). Review of Alva Noë [Review of the book Action in Perception]. Journal of Philosophy, 102, 259-272.

Burge, T. (1989). Individualism and causation in psychology. Pacific Philosophical Quarterly, 70, 303-322.

Clark, A. (forthcoming). Pressing the flesh: Exploring a tension in the study of the embodied, embedded mind. Philosophy and Phenomenological Research.

Clark, A. and Chalmers, D. (1998). The extended mind. Analysis, 58, 7-19.

Hutchins, E. (1995). Cognition in the Wild. Cambridge: MIT Press.

Kim, J. (1992). Multiple realization and the metaphysics of reduction. Philosophy and Phenomenological Research, 52, 1-26.

Kim, J. (1998a). Philosophy of Mind. Boulder: Westview Press.

Kim, J. (1998b). Mind in a Physical World. Cambridge: MIT Press.

Lakoff, G., & Johnson, M. (1999). Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought. New York: Basic Books.

Lakoff, G., & Johnson, M. (1980). Metaphors We Live By. Chicago: University of Chicago Press.

Nagel, E. (1961). The Structure of Science. New York: Harcourt Brace.

Noë, A. (2004). Action in Perception. Cambridge: MIT Press.

Shapiro, L. (forthcoming). Can psychology be a unified science? Philosophy of Science.

Shapiro, L. (2004). The Mind Incarnate. Cambridge: MIT Press.

Shapiro, L. (2000). Multiple Realizations. Journal of Philosophy, 97, 635-64.

Shapiro, L., & Sober, E. (forthcoming). Epiphenomenalism -- the do's and the don'ts.  In G. Wolters & P. Machamer (Eds.), Studies in Causality: Historical and Contemporary.  Pittsburgh: University of Pittsburgh Press.

Spivey, M., Richardson, D., & Fitneva, S. (2004). Thinking outside the brain: Spatial indices to linguistic and visual information. In J. Henderson & F. Ferreira (Eds.), The Interface of Vision, Language, and Action (pp. 161-189). New York: Psychology Press.

Wilson, R. (2004). Boundaries of the Mind: The Individual in the Fragile Sciences: Cognition. New York: Cambridge University Press.

Notes

1Kim notes frequently that a minimal sort of physicalism requires at least both these claims, as supervenience by itself does not imply dependence. See, for instance, (1998a, 10-12).

2With some codicils about wide content.

3See Shapiro (2004) for some reasons to be cautious about the multiple realization assumption.

4There is another kind of burden that autonomists face: the risk of causal exclusion. I intend to put this to the side in this chapter, but for discussion see Kim (1998b) and Shapiro and Sober (forthcoming).

5Discussion in this section draws heavily on Shapiro (forthcoming).

6But perhaps they are not always as careful as they should be. Block (2005) takes Noë to task for sometimes confusing the claim that the body is a causal influence on the mind with the claim that the body is a constituent of the mind.

7I make a similar point in my (2004), where I draw an analogy between, on the one hand, the doomed effort to pilot an airplane through use of instructions designed for a submarine and, on the other hand, the use of human psychological algorithms to guide action in a non-human body. The analogy was intended to show the sense in which human psychology is tailored to the particulars of a human body.

8I am grateful to Andy Clark for helping me to see this. For valuable discussion of these issues, see his “Pressing the Flesh: Exploring a Tension in the Study of the Embodied, Embedded Mind,” forthcoming.

9Clark (forthcoming) challenges this very strong claim about the body’s role in cognition, arguing that the body is just one facet in the constituents of cognition. Perhaps, he suggests, compensatory changes in environment and internal processing can endow differently embodied organisms with a similar cognitive profile. This seems to me a fruitful line of criticism against Lakoff and Johnson.

10Here I assume that differences between species entail differences between body types sufficient for differences in psychology. Perhaps species with very similar bodies will have very similar psychologies.