Review - Boundaries of the Mind
The Individual in the Fragile Sciences
by Robert A. Wilson
Cambridge University Press, 2004
Review by Ray Rennard, Ph.D.
May 14th 2005 (Volume 9, Issue 19)

Boundaries of the Mind is the first of three volumes in a series by Robert A. Wilson exploring the role of the individual in the cognitive, biological, and social sciences--what Wilson refers to collectively as the "fragile sciences."  The second volume, Genes and the Agents of Life, has recently been published by Cambridge University Press, and a third volume on the social sciences is in preparation.  In this first volume, Wilson challenges individualism about the mind, which is a metaphysical and methodological thesis about what mental phenomena are and how they ought to be characterized and studied within the cognitive sciences.  According to the individualist, mental states, processes, and capacities should be individuated solely by appeal to properties that supervene on the intrinsic properties of individuals--typically, states of the individual's brain or nervous system.  In its place, Wilson motivates and defends an anti-individualist or externalist conception of cognitive states, processes, and capacities, and he even offers an externalist account of consciousness.  Most externalists subscribe to some form of content externalism, which characterizes mental phenomena in terms of their externally determined intentional contents.  Content externalism is consistent still with the claim that mental states are "in the head"--understood as a claim about the location of a mental state's causally efficacious properties.  Such a view would be taxonomically wide but "locationally" narrow.  Wilson is one of a growing number of cognitive scientists who think that content externalism does not go far enough.  He contends that some mental phenomena literally extend beyond the skin of thinkers, incorporating features of the wider environment in which they are located.  Some cognitive states are locationally wide as well. 

Boundaries of the Mind consists of four parts.  In addition to introducing the main topic, Part I contains an interesting historical tour of the notion of individualism in psychology and a discussion of the nativism debate in psychology.  Notable here is Wilson's claim that scientists and philosophers have been guilty of "smallism"--metaphysical discrimination in favor of the small, hence, against the not-so-small.  This preference for explanation and reduction in terms of smaller things and their properties, Wilson claims, has made individualism and nativism seem more compelling than they otherwise would.  Wilson favors a two-dimensional approach to mapping the various positions on the existence of innate cognitive capacities.  Radical nativists accept both that the internal structures important to the development and acquisition of some cognitive capacity are rich, and that the structures that are external to an individual are secondary to the development and acquisition of a capacity.  Strong anti-nativists (e.g., behaviorists) reject both the internal richness and external minimalism theses.  This leaves ample room for positions between these two extremes, especially views that take seriously the idea that external structures might play a constitutive role in the development and continued use of certain cognitive capacities. 

The first chapter of Part II of Boundaries of the Mind contains a masterful summary of the individualism/externalism debate in philosophy.  [Those looking for even more detail should read Wilson's Cartesian Psychology and Physical Minds (Cambridge University Press, 1995), which includes a terrific examination and critique of the motivations for individualism in psychology.]  Important to Wilson's case against individualism is his claim that the standard metaphysical view of the relation between mental and physical states--viz., that mental states are realized in physical states--is biased in favor of individualism.  To say that a physical state realizes a mental state is to say that the physical state is metaphysically sufficient for the mental state properties it realizes.  There is a temptation to take the realizers of a mental state to be constituted by the intrinsic, physical states of the individuals who have those mental states.  (Wilson suggests that part of the attractiveness of this physical constitutivity thesis may be due to the widespread acceptance among cognitive scientists of the notion of functional decomposition, which involves the analysis of a complex capacity of a system into its constituent sub-capacities.)  The standard view of realization, says Wilson, combines the metaphysical sufficiency thesis and the physical constitutivity thesis.  In Chapter 5, Wilson argues that these two theses are actually in conflict with each other.  If we assume the constitutivity thesis, we can find "core" states of a system that are not metaphysically sufficient to realize the associated mental state.  It's only the core states considered as part of a greater total realization and in concert with the appropriate background conditions that are sufficient to realize the associate mental state.  However, the background conditions (and sometimes the total realization itself) required to realize a mental state often fall outside the intrinsic, physical states of individuals, Wilson argues.  He proposes an alternative context-sensitive account of realization, according to which "realizations are located within systems, and those systems in turn are located within broader environments" (117).  A number of properties from the biological sciences are understood to have wide realizations--e.g., being a predator, being highly specialized, and the property of fitness.  These wide properties, moreover, play crucial roles in biological theorizing and cannot be eliminated in favor of properties that are intrinsic to individuals.  Why think that there are mental properties that have wide realizations?  At this point, Wilson has merely argued that wide mental properties cannot be ruled out on metaphysical or methodological grounds.  Later, in Part III, Wilson maintains that some mental states and processes do in fact have wide realization bases. 

After a separate chapter in which he further motivates his context-sensitive account of realization, Wilson turns his attention to the nature of representations.  According to the dominant approach in cognitive science, mental representations are discrete states of individuals that have both semantic content--by encoding information about some object or state of affairs--and causal powers--by being realized in states of an individual's nervous system.  The widespread acceptance of the computational (or information-processing) approach in cognitive science informs this individualistic conception of representations.  Wilson claims that the early philosophical motivations for individualism (e.g., Jerry Fodor's methodological solipsism and Stephen Stich's principle of autonomy) were mere gestures toward certain features of cognitive science rather than detailed examinations of actual cognitive scientific practices.  This assumption did not go unchallenged.  Things heated up when the philosopher Tyler Burge argued that David Marr's influential theory of vision is not individualistic after all because it adverts to wide contents.  Wilson reviews some of the main contributions to the debate over Marr's theory, and, in the end, concludes that there may be no fact of the matter whether it employs narrow or wide content.  Rather than follow philosopher Frances Egan, who contends that Marr's theory is individualistic because it is computational (i.e., formal), Wilson favors anti-individualistic conceptions of representation and computation.  According to wide computationalism, some parts of a computational system fall outside the body of the organism.  Some of the representations in wide computational systems are what Wilson calls "exploitative representations," which are constant and regular information-rich aspects of the environment that an organism can exploit without encoding that information internally.  Andy Clark, Mark Rowlands, and Ron McClamrock, among others, have offered versions of wide computationalism and exploitative representation.  Wilson's development of these ideas is well organized, precise, and accessible to those new to this kind of approach. 

Part III is the most daring part of the book.  In Chapter 8, Wilson argues 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 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" (188, italics in original).  We are encouraged to ask not what is in the mind, but what the mind is in.  To support his claim that the mind is both embodied and embedded, Wilson looks to recent research on memory, cognitive development, and the so-called "theory of mind" capacity.  Recent work on memory suggests two externalist alternatives to the traditional "storehouse" model, which conceives of memory as the internal encoding and storing of information about one's experiences.  One approach to memory, the correspondence metaphor, is content-externalistic because it characterizes memories by their contents and their (causal) relations to past experiences.  Wilson endorses another account of memory--one that is locationally externalist because it views memory as an enactive process involving aspects of one's body and physical environment, and even one's social and linguistic community.  This approach, developed by the cognitive psychologist Ulric Neisser among others, employs the notion of remembering as doing.  On this account, internally encoded information is just one kind of resource used in acts of remembering; one also relies on environmental cues or props, language (e.g., the rhyme and cadence of spoken language), and bodily states (e.g., the orientation of one's body in relation to one's immediate environment). 

Following the discovery that children have more knowledge and at earlier ages than was previously thought, developmental psychologists have attributed to children even richer internal and innate cognitive structures.  Accounting for all this knowledge through internal enrichment flies in the face of common sense and earlier research on cognitive development, says Wilson.  But this nativist account can be augmented by the mediational approach of Russian psychologist Lev Vygotsky.  According to this approach, the cognitive capacities of children develop with the aid of mediational tools, such as maps, numerical systems, and language.  These mediational tools do not merely cause changes in the internal mental structures, but literally come to be part of a wide cognitive ability.  Wilson concentrates on the development of explanatory abilities.  He draws on work he has done with developmental psychologist Frank Keil, which points to three features of our everyday explanatory practices: they are ubiquitous, they are informed by (folk) theories, and yet they are astonishingly shallow.  For example, most of us think we have a working understanding of flush toilets, but few of us can give adequate explanations of the mechanism through which they actually work.  We tend to overestimate our ability to explain the inner workings of many common devices.  Keil and Wilson call this "the illusion of explanatory depth."  The shallowness of our own explanations is overcome by our reliance on a division of cognitive labor: if we cannot produce a "clockworks" explanation of the flushing mechanism of a toilet, we know that someone can.  Keil and others have found the same capacity in preschool children.  The child's mind is built to exploit the knowledge of others, and so does not require stores of internally rich representational structures.  The theories and information that underwrite our explanations are locationally wide, since they incorporate the expertise found in other individuals and resources.

Wilson's case for construing our capacity for mindreading as a locationally wide, bodily enactive skill is thinner than those for memory and cognitive development, but no less provocative.  He distinguishes between a bare-bones conception of folk psychology, which concentrates only on the propositional attitudes, and a full-blown conception, which includes, in addition, affective states and sensations.  Because it involves attributing to individuals intentional states with wide contents, full-blown folk psychology is taxonomically wide.  Moreover, because it involves the attribution of affective states--which are knowable by first-person bodily experience of those states--full-blown folk-psychology is bodily enactive, and so locationally wide.  However, Wilson takes the wideness of mindreading one step further, arguing that our most sophisticated uses of folk-psychology--occasions when we follow a complicated psychological narrative or manipulate the mental states of others--are world-involving.  Rather than merely encoding information about the mental states of others, our meta-representational states "index" the states of others.  In a sense, we interact and engage with the mental states of others, exploiting their richness in the course of navigating through a complex social world.   These are intriguing ideas, and I would be very interested to hear more about the world-involving aspects of mindreading.   

In Chapter 9, Wilson considers the implications of his externalist framework for consciousness, understood both as processes of awareness and as phenomenal states.  He develops an original conception of consciousness: the view that consciousness is Temporally Extended, Scaffolded, Embodied, and Embedded (TESEE).  The TESEE approach to processes of awareness--introspection, meta-representation, and attention--is continuous with the embedded and embodied approach to memory, cognitive development, and theory of mind.  As with these other capacities, simple cases of introspection or reflection seem at best taxonomically wide (wide because their intentional contents are wide).  However, when we consider the ubiquity of complicated processes of awareness--e.g. cataloguing one's beliefs about a complex issue, navigating a complicated physical or social environment, or concentrating on a difficult task--it appears that one relies a great deal on factors that extend beyond the immediate subject in space and time.  Not only are such processes extended in time, but they exploit information-rich external scaffolds (such as bits of language and navigation equipment) and rely on dynamic relations between the subject's body and the environment in which it is located. 

The TESEE approach can be extended to phenomenal consciousness as well.  Representationalists about phenomenal consciousness, such as Fred Dretske, William Lycan, and Michael Tye, argue that a phenomenal property is not an intrinsic property of experience but rather a feature of the representation of its objects.  As such, phenomenal properties inherit their width from the intentional contents to which they apply.  According to these representationalists, phenomenal consciousness is externalistic.  Wilson thinks that this global externalism goes both too far (especially with respect to pains and some bodily sensations) and not far enough.  Citing the groundbreaking work of the philosophers Alva Noё and Susan Hurley, and psychologist J. Kevin O'Regan, Wilson discusses recent TESEE conceptions of vision and visual consciousness.  Noё and O'Regan, in their sensorimotor theory of visual consciousness, argue that vision, like touch, involves active and dynamic exploration of the contingent features of the environment.  [Noё's recent book, Action in Perception (MIT Press, 2004), explores in detail the exciting consequences of this "enactive" approach to vision.]  Vision, on this approach, is not passive and internal; rather, it essentially involves the active exploration of objects.  Hurley has developed a related account of perception.  Contra the traditional view that perception and action are the input and output (respectively) of central cognitive processes, Hurley contends that perception and action are involved in dynamic cycles that cut across all three domains.  For example, when engaged in a complicated manual task, our tactile, visual, and motor systems are coupled in a dynamic feedback loop.  The visual component of this process is temporally extended and embodied, as it essentially incorporates aspects of the motor component.  Furthermore, this "horizontal" module may include aspects of the wider environment, making it locationally wide as well. 

In Chapter 10, Wilson considers recent attempts to use the purported narrowness of phenomenal properties and their purported inseparability from intentional properties as a reductio of externalism in the philosophy of mind.  The intimate connection between vision and action on TESEE accounts raises interesting questions about the coherence of thought experiments involving disembodied thinkers, such as brain-in-a-vat scenarios.  Especially problematic, according to Wilson, are arguments against externalism that presuppose that one can have a brain-in-a-vat phenomenal doppelganger.  If vision is embodied and embedded, as TESEE accounts claim, it is hard to see how a brain-in-a-vat--which cannot have haptic or proprioceptive experiences--can have the same type of visual states as a normal thinker.  To assume that such subjects can have the same phenomenal experiences just is to presuppose that TESEE views about phenomenal consciousness are false.  So, one cannot rely on such thought experiments to defeat (or bolster, for that matter) externalist accounts of consciousness.  I think that Wilson is right about this.  Such thought experiments rely on our intuitions and introspective awareness of our own phenomenal states.  If the phenomenal aspects of our visual states are as the TESEE approach contends, then such brain-in-vat cases would involve the illegitimate attribution of widely-determined properties to a narrowly construed system.  It's not clear that we can trust our intuitions when it comes to conceiving our phenomenal lives as detached and disembodied.

In Part IV of the book, Wilson considers claims made by some biologists and sociologists about the plausibility of group minds.  This is, in my opinion, the least successful part of the book.  I understand the rationale for considering the group mind hypothesis, but I was a little disappointed by the cautionary ending.  It seemed to me to be disconnected from the radical and provocative spirit of the rest of the book.  I sense that the subject will be given better treatment in the third volume in this trilogy, and perhaps its inclusion here was intended as a prelude of things to come.  I would have preferred to read about the prospects of preserving the philosophical notion of the self in light of a locationally wide conception of the mind.  There are tantalizing bits here and there about the self as the locus of control of the extended mind, but even this strikes many as controversial and problematic. 

Despite this minor stylistic criticism, I think that this book makes a major contribution to our understanding of the nature of cognition and consciousness.  Wilson does an exceptional job making complicated and abstruse issues in metaphysics and the philosophy of mind accessible, even to the novice.  His writing is characteristically clear, and the arguments and overall structure of the book are straightforwardly mapped.   Although some of the main arguments of the book have been presented by Wilson and others elsewhere, there is plenty here that is new for those well versed in the externalism-individualism debate in philosophy and the cognitive sciences.  This book should be required reading for anyone interested in cognitive science and the philosophy of mind. 

© 2005 Ray Rennard

Ray Rennard will join the University of the Pacific this fall as Assistant Professor of Philosophy.