The New Science of the Mind

by Mark Rowlands

  However, the embodied mind and the extended mind are united by one central idea: both are ontic theses of the (partial) composition of (some) cognitive processes. Accordingly, it is this idea that lies at the heart of the non-Cartesian conception of the mind. Some cognitive processes are composed, in part, of structures and processes that are located outside the brain of the cognizing subject. Cognitive processes are an amalgam of neural structures and processes, bodily structures and processes, and environmental structures and processes. We can subsume the theses of the embodied mind and the extended mind into one: the amalgamated mind. The new science will be based on the idea of mental processes as amalgamations.

  The idea that cognitive processes are extended can easily conjure up the wrong sorts of images. The root of the difficulty is that extension is a spatial concept and, so, is closely tied to that of location. And the issue of the location of cognitive processes can easily sidetrack us into concerns that we should not have. Thus, the idea that cognitive processes are extended can easily tempt us into thinking of mental processes as somehow stretching outside of the brain, and thus having a definite spatial location but simply one that incorporates expanses of the extracranial world. This is, I think, precisely how not to think about extended cognition. In general, cognitive processes have a location that is, at best, vague, and may be genuinely indeterminate. That is, there might be no fact of the matter with regard to where a given token cognitive process occurs. It is not, as the metaphor of the extended mind might tempt us to suppose, that cognitive processes have determinate extended boundaries-like an elastic band that has been stretched outside the brain or skull. It is rather that they have no determinate boundaries at all. The thesis of the extended mind might more accurately be called the thesis of the spatially indefinite mind. But, let's face it, that's nowhere near catchy enough.

  The thesis of the amalgamated mind is a little better in this regard. What is important for the new science is the composition or constitution of cognitive processes and not, in the first instance, their location. Claims about the precise location of cognitive processes do not follow, in any straightforward manner, from claims about their composition. That would depend on whether the things that compose cognitive processes have precise spatial location. However, claims about where such processes are not located do follow. Thus, the thesis of the amalgamated mind-the claim that (some) token cognitive processes are amalgamations of neural, bodily, and/or environmental structures and processes-entails that not all cognitive processes are located inside the heads of cognizing organisms. But what is of primary importance to this thesis is the issue of composition, not location.

  Thus, AM, the thesis of the amalgamated mind-cognitive processes as amalgamations of the neural, bodily, and environmental-subsumes both theses of the embodied and the extended mind. Amalgamation of these apparently disparate realms is based on the concepts of exploitation, manipulation, and transformation. When the extracranial structures and processes thus amalgamated are bodily ones, the idea of exploitation is likely to be at the forefront of the possibility of amalgamation. The brain exploits the distance between the organism's ears in its processing of the distance and direction of sound sources. However, when the extracranial structures and processes thus amalgamated are also extrabodily, then manipulation and transformation can come to the fore. In manipulating environmental information-bearing structures, the organism can transform the information they contain from the merely present to the available. In doing so, it can now exploit that information in its subsequent processing operations.

  The concept of amalgamation thus incorporates both the thesis of the embodied mind and that of the extended mind. The possibility of such incorporation will come as a surprise to many-who have supposed, for reasons we will explore later, that the theses of the embodied and extended mind are mutually inimical. The next chapter begins defense of the amalgamated mind.

  1 The Amalgamated Mind: Challenges

  The thesis of the amalgamated mind is the combination of the theses of the embodied mind and the extended mind. Both of these are versions of a more general idea: cognitive processes are partly composed of, or constituted by, extraneural processes. They are made up, in part, of bodily and/or wider environmental processes. The thesis of the amalgamated mind (henceforth simply "the amalgamated mind") is this claim of extracranial composition. Since the amalgamated mind incorporates both theses of embodiment and extendedness, the principal challenges facing it can be divided into three categories:

  1. Challenges to the thesis of the extended mind.

  2. Challenges to the thesis of the embodied mind.

  3. Challenges to the combination of these theses.

  According to the third type of challenge, even if there is nothing wrong with the embodied mind and the extended mind taken individually, these two theses cannot be combined together in the way envisioned by the amalgamated mind since they are, in fact, mutually incompatible.

  I shall discuss the challenges in this order.

  2 The Extended Mind and Its Discontents

  The thesis of the extended mind has recently been the target of a number of objections, broadly divisible into four kinds:

  1. The differences argument This type of objection points to the significant differences between internal cognitive processes and the external processes that the thesis of the extended mind also claims are cognitive. This casts doubts on the claim that both processes should be regarded as examples of a single psychological kind. This sort of objection has been vigorously championed by Rupert (2004) and also by Fodor (2009).

  2. The coupling-constitution fallacy This type of objection claims that the thesis of the extended mind confuses real cognition with its extraneous causal accompaniments. More precisely, it confuses those structures and processes constitutive of cognition with structures and processes in which cognition is (merely) causally embedded. We have already encountered this type of objection on several occasions. It is an objection that has been developed by Adams and Aizawa (2001, 2010) and also, in a somewhat different way, by Rupert (2004).

  3. The cognitive bloat objection According to this objection, the admission of extended cognitive processes places us on a slippery slope. Once we permit such processes, where do we stop? Our conception of the cognitive will become too permissive, and we will be forced to admit into the category of the cognitive all sort sorts of structures and processes that clearly are not cognitive.

  4. The mark of the cognitive objection According to this objection, the thesis of the extended mind should be rejected on the grounds that it is incompatible with any plausible mark of the cognitive. A mark of the cognitive is a criterion or set of conditions that specifies when a process is to count as cognitive (and perhaps also when it is not to count as cognitive). The objection is that any plausible criterion of the cognitive will disqualify the sorts of extended processes invoked by the thesis of the extended mind from counting as cognitive. This objection is developed by Adams and Aizawa (2001, 2010).

  In the next few sections of this chapter, I shall argue that the first three objections all reduce to the mark of the cognitive objection. They do so either in that they presuppose this objection or can be answered by way of the provision of a satisfactory mark or criterion of the cognitive.

  3 The Differences Argument: Parity and Integration in the Extended Mind

  The thesis of the extended mind is often thought to be grounded in the concept of parity: roughly speaking, the similarity between the external processes involved in cognition and internal processes that are widely accepted as cognitive. The extended mind's reliance on this notion of parity is often thought to be embodied in, and demonstrated by, Clark and Chalmers's deployment of what they call the parity principle:

  If, as we confront some task, a part of the world functions as a process which, were it done in the head, we would have no hesitation in accepting as part of the cognitive process, then that part of the world is (so we claim) part of the cogn
itive process. (Clark and Chalmers 1998, 8)

  Critics of the extended mind have, seemingly without exception, understood the parity principle as introducing a similarity-based criterion of when an external process or structure is to be understood as cognitive-that is, as a genuinely cognitive part of a cognitive process. Very roughly, on this understanding of the parity principle, if an external process is sufficiently similar to an internal cognitive process, then it too is a cognitive process.

  It is this interpretation of the role of the parity principle-as introducing a similarity-based criterion for when cognition can legitimately be regarded as extended-that underwrites the differences argument. For example, in one recent development of this argument, Fodor criticizes Clark and Chalmers on the grounds that the alleged parity between the cases of Otto and Inga is merely apparent, and rests on Clark and Chalmers inaccurately describing the scenario:

  Surely it's not that Inga remembers that she remembers the address of the museum, and having consulted her memory of her memory then consults the memory she remembers having, and thus ends up at the museum.... It's untendentious that Otto's consulting "outside" memories presupposes his having inside memories. But on pain of regress, Inga's consulting inside memories about where the museum is can't require her first to consult other inside memories about whether she remembers where the museum is.... There is, after all, a built-in asymmetry between Otto's sort of case and Inga's sort. Otto really does go through one more process than Inga. . . . Inga's "consulting her memories" her memories is a fake, and a particularly naughty fake because ... it makes Inga's case look more like Otto's than it can possibly be. (Fodor 2009, 15)

  In other words, Clark and Chalmers's argument does not work because the parity between internal and external on which it relies is only apparent and not real.

  In a similar vein, Rupert, in connection with an argument for extended memory developed by me (Rowlands 1999), outlines his strategy as follows:

  I argue that the external portions of extended "memory" states (processes) differ so greatly from internal memories (the process of remembering) that they should be treated as distinct kinds; this quells any temptation to argue for HEC [hypothesis of extended cognition] from brute analogy (viz. extended cognitive states are like wholly internal ones; therefore, they are of the same explanatory cognitive kind; therefore there are extended cognitive states). (Rupert 2004, 407)

  Later on, we shall look at one specific application of this strategy. For present purposes, however, let's just focus on the general strategy rather than its specific applications. The assumption underlying both Fodor's and Rupert's objections is that the function of the parity principle is to introduce a similarity-based criterion of when a cognitive process such as remembering can be extended into the world: if an external process is sufficiently similar to internal cognitive processes, then it too is a cognitive process. Fodor and Rupert then argue that since external processes involved in memory are, in fact, not sufficiently similar to internal cognitive processes, then they do not qualify as cognitive processes. Presumably this is intended as an inductive argument: since the internal and external processes involved in cognition are not sufficiently similar, the parity principle provides no reason to regard the latter as cognitive.'

  The differences argument, however, fails to properly understand the arguments for the extended mind.' This thesis does not, in fact, rely on a similarity-based criterion of when an extended process may legitimately be regarded as cognitive. The notion of parity is indeed, I shall argue, an important one for the extended mind. However, equally important is the notion of integration: the meshing of disparate types of process that, precisely because of their disparate character, can enable a cognizing organism to accomplish tasks that it would not be able to achieve by way of either type of process alone (Menary 2006, 2007; Sutton 2010).

  From this integrationist perspective, the differences between internal and external processes are as important as, or even more important than, their similarities. The reason cognition extends into the environment is precisely because, with respect to the accomplishing of certain cognitive tasks, external processes can do things that internal processes cannot do (or, depending on how you want to understand the thesis of the extended mind, simply, in certain cases, do not do). External structures and processes possess quite different properties from internal ones; and it is precisely this difference that affords the cognitive agent the opportunity to accomplish certain tasks that it would not be able to accomplish purely by way of internal cognitive processes. Without these differences, the external processes would be otiose.

  For example, in The Body in Mind (Rowlands 1999), I emphasized the relative permanence and stability of external forms of memory storage, and I examined the implications of this for the development of biological memory strategies during the process of enculturation. It is because this permanence and stability have no real echo in internal, biological processes that the development of external forms of memory was both useful and had marked implications for the character of biological memory (see also Donald 1991). More recently, O'Regan and Noe (2001, 2002) and Noe (2004) have emphasized the importance of the permanence and stability of the external world in constituting visual perception. The world, by providing a stable and relatively permanent structure that can be probed and explored at will by the visual modality, obviates the need for at least certain sorts of visual representations as these were traditionally understood.'

  In The Body in Mind I also emphasized the distinctive structure of external systems-for example, the recursive, combinatorial structure peculiar to linguistic systems-structure that has no real echo in internal, biological processes. I argued that if an organism is capable of appropriately utilizing worldly structure in the performance of its cognitive tasks, then this sort of structure need not be replicated internally. Thus, once again, the importance of the world, and by extension the processes occurring in it, derives precisely from its possessing structure that internal processes lack, and so allows us to do what we cannot do in the head alone.4

  Therefore, it is a mistake to suppose that the case for the extended mind is built on parity alone. Equally important are the differences between internal and external processes. It is precisely the different properties of external structures that allow the cognitive agent to accomplish things that it either could not, or in fact does not, accomplish by way of internal processes alone. Given the central role played by the notion of integration in the thesis of the extended mind, one cannot base, as does the differences argument, an objection to this thesis simply by citing differences between internal and external processes. The thesis of the extended mind, properly understood, both predicts and requires such differences. Understood on its own terms, therefore, the differences argument fails.

  The importance of the role of integration in the extended mind should not, however, blind us to the role of parity. The extended mind requires both parity and integration. And it requires both of these to be properly understood and allotted their proper place and appropriate role. Suppose we tried to develop a case for the extended mind based purely on the notion of cognitive integration. This would leave the thesis open to another objection. If the thesis of the extended mind requires significant differences between internal processes and the external processes that it regards as cognitive, what reason is there for supposing that the latter are really part of cognition rather than a merely external accompaniment to real, internal cognitive processing? That is, using a distinction between the extended mind and the embedded mind we developed in the previous chapter, given that there are significant differences between internal cognitive processes and extended processes involved in cognition, why not simply suppose that the latter are part of the extraneous scaffolding in which the real, internal cognitive processes are embedded? In other words, reliance on the differences between internal and external processes might give you the thesis of the embedded mind. But it will not, it seems, give you the thesis of
the extended mind (Rowlands 2009a).

  Given the integrationist's emphasis on the differences between the internal and external processes involved in cognition, it is not possible to establish the cognitive status of the latter simply by analogical extension from the former. Therefore, if we are to defend the cognitive status of the extended processes, we need some other way of defending the claim that these external processes are cognitive ones. One way of doing this-and it is not clear that there is another way-would be to provide an adequate and properly motivated criterion or mark of the cognitive: a criterion that would allow the thesis of the extended mind to justify the claim that the external processes involved in cognition are indeed cognitive processes, rather than merely an extraneous scaffolding in which cognitive processes are embedded. In short, the integrationist response deflects the differences argument only by leaving the thesis of the extended mind vulnerable to the mark of the cognitive objection.

  In this context, the provision of an adequate and properly motivated mark of the cognitive would accord due importance to the differences between internal and external processes but, at the same time, reintroduce a sufficient degree of parity between internal and external to avoid the charge that we have no reason for supposing that the latter are really cognitive. The parity reintroduced here is of an appropriate sort. It is not parity that obtains between particular processes: we have to safeguard the idea that internal and external processes can be quite different (for that is, in many cases, the point of the external processes). Rather, it is parity with respect to certain abstract, general features of cognition that will be identified by the mark of the cognitive. These are features that are sufficiently abstract to be possessed by both internal and external processes, no matter how different these may be with regard to more concrete features. It is these abstract features that warrant the claim that the external processes involved in cognition are, in fact, cognitive (Rowlands 2009a,b,c).