In the same sort of way, the strategy that tries to explain ownership of cognitive processes in terms of the idea of integration is going to identify a fixed frame of reference-for which there is no question of ownership-in terms of which the integration of cognitive processes can be effected. Traditionally, this frame of reference has been supplied by the input, in the form of sensation or perception, and the output, in the form of action or motor response. However, this strategy will work only if the question of the ownership of these items is unproblematic. However, the question of ownership arises for these sorts of process too. The idea that the ownership of input and output for an organism is unproblematic does not survive the demise of the attempt to understand ownership in terms of spatial containment. The broadly functionalist account of ownership that replaces the spatial criterion, and underwrites the idea that ownership is a matter of appropriate integration, has one obvious but unavoidable consequence. It is not simply that cognitive processes must be appropriately integrated into input and output in order to be owned by a subject. Just as fundamentally, input and output must be integrated into cognitive processes in order to be owned by a subject. Therefore, any attempt to employ the integrationist approach in answering the question of ownership simply begs the question of ownership for sensation and action.
If this is correct, then it makes the most sense to regard the integrationist attempt as attempting to answer a criteriological problem: given the assumption of a fixed frame of reference-the existence of states or processes for which the question of ownership is unproblematic-what are the conditions necessary and/or sufficient for a cognitive process to be integrated into this framework? This question is perfectly legitimate, but it raises an even more basic question that the integrationist approach cannot answer: what does an organism's ownership of its cognitive processes consist in? This question does not assume a fixed frame of reference and then try to integrate into this, but, instead, tries to understand what it is for an organism to own both its cognitive processes and the inputs and outputs to these processes. The question of what it is for an organism to own its cognitive processes cannot be separated from the question of what it is to own its detection of the environment, and what it is to own its subsequent behavioral responses. To answer the constitutive question is to answer all three of these questions.
The rest of the book will focus on the constitutive question. It would, of course, be nice to solve the various criteriological problems too. However, not only would this require a book in itself, but doing so is not necessary for the defense of the amalgamated mind to be developed here. Once we see how to solve the constitutive problem, I shall argue, the theses of the embodied and the extended mind will emerge as natural-indeed obviousconsequences, in much the same way that the thesis of the extended mind emerged from liberal functionalism. Both theses, I shall try to show, are straightforward implications of our ownership of cognitive processes, when this is properly understood. Showing this is the principal task of the rest of this book.
5 Ownership and Agency
To recap: the story so far is a rather winding one, and looks something like this. According to condition (2) of our criterion of cognition, a cognitive process has the proper function of making previously unavailable information available. However, it can do this in two ways: by making information available to the subject, or by making this information available to subpersonal operations. In the latter case, however, this simply postpones the question of ownership: what makes these processes belong to that subject? As we have seen, spatial containment does not provide a realistic criterion of ownership. These subpersonal processes belong to a subject only to the extent that they are themselves appropriately integrated into the subject. And they will be thus integrated only if they, eventually, make a difference to processes to which the subject has conscious access and over which it has conscious control. A subpersonal process is integrated into a subject to the extent that it, together with perhaps many other subpersonal cognitive processes, will ultimately have an impact on consciously accessible, personal-level cognitive processes. The qualification together with perhaps many other subpersonal processes is not insignificant. Many subpersonal processes, taken in isolation, need have no impact whatsoever on consciously accessible, personal-level cognitive processes: either because they are transient or because they are otherwise insignificant. However, in combination with other processes at the same level, subpersonal cognitive operations must reflect representational transitions that occur at the conscious level to the extent that the whole system of conscious and unconscious states and processes must form an at least relatively coherent whole. Once we fall below a certain threshold level of coherence, then we shall encounter, at the conscious level, certain thought- and attitude-transitions that we regard as alien.
If this is correct, then we should regard ownership of subpersonal cognitive processes as derivative upon the ownership of personal cognitive processes. Ownership of the former can be understood in terms of integration. But this project will work only if we have solved the problem of ownership for personal-level cognitive processes. The project we need to undertake is that of understanding the phenomenon of ownership at the personal level. And this I am going to understand as a constitutive rather than a criteriological project.
With respect to the constitutive project, a useful place to start is provided by the observation that, at the personal level, cognitive processes are, fundamentally, things we do: they are activities. Therefore, we might plausibly try to understand ownership of personal-level cognitive processes along the lines of ownership of activities in general. And, in general, we own our activities precisely because, and to the extent, that we do them. So, it seems we need to consider two questions:
1. If personal-level cognitive processes are activities, what sorts of activities are they?
2. In what sense do we perform these activities?
Of course, in one sense, to each different type of cognitive process there corresponds a different sort of activity: the activity of thinking is different from the activity of perceiving, which in turn is different from the activity of remembering, and so on. Nevertheless, this does not rule out the possibility that there is a more general activity-type that can subsume all personal-level cognitive processes.' I shall argue that there is. If this is correct, then understanding the sense in which we do or perform personal-level cognitive processes is a matter of understanding the sense in which we engage in this sort of activity. Both these questions are the subject of the rest of this book. Here, I simply want to provide a preparatory analysis of what it is to do or perform-and thereby to own-activities in general.
Suppose I am building a house, and am doing this on my own. In what sense, if any, might I be the owner of this activity? To the extent this question means anything, it is equivalent to the question: in what sense is it me and not anyone else who is doing the building? Ownership of an activity reduces to being the person who is engaged in it. I own an activity when I am the author of it.
However, this raises as many questions as it answers. To be the author of an activity-to be the person engaged in it-entails having authority over that activity. I have, intuitively, a certain sort of authority over my building activities, an authority that I do not have over, for example, my falling off the roof. I am the author of the former, but the victim of the latter. But what, precisely, does this intuitive sense of authority amount to?
It is common to suppose that my ownership of my activities can be explained in terms of my intentions (and associated intentional statesbeliefs, desires, etc.). I am the author of the activity of building the house because I have the intention to build the house. And the various activities that go into this overall activity can similarly be explained in terms of my intentions to perform the necessary components of the overall activity of building the house. Conversely, I am the victim of my tumble from the roof because this was not the result of my intentions.
The appeal to inten
tions, however, will not help with the problem I have in mind: the appeal would only push that problem back a step. We need also to understand what constitutes ownership of intentions and other relevant intentional states. The problem of ownership in this regard can manifest itself in several different ways. Here is one that is particularly useful for our purposes: the problem of ownership manifests itself by way of a distinction between what we might call practical and epistemic authority. Each of these, I think, shades by degrees into the other; but the absence of a firm distinction is, of course, not the absence of a distinction. Since I am, ex hypothesi, the only person on-site, there is a clear sense in which I, and I alone, am doing the building. However, in building the house, I am using bricks and tiles made by someone else, wood supplied by someone else, tools manufactured by someone else, and so on. So, in what sense, and to what extent, is it I who am doing the building and not these people implicated in the overall process? The appeal to intentions will not help us with this question, for that appeal merely begs the question of how to distinguish those parts of the process with respect to which I do have intentions from those parts with respect to which I do not.
The sense in which I am doing the building, one might think, reduces to the sense in which I have authority over the process and the product. However, there are two different sorts of authority involved here. To see this, consider one part of the overall process: the laying of bricks. There are certain parts of this process over which I have epistemic authority. I am, of course, capable of identifying individual bricks-distinguishing one brick from another, and where necessary reidentifying individual bricks. I am, in short, acquainted with the bricks in all relevant and necessary ways. I can also, let us suppose, recognize the characteristics of good mortar work-I know the optimal amount of cement to put on each joint; I know, given the ambient air temperature, how much water should ideally be present in the cement; and so on. I know these things, and so a failure to satisfy them is a failure on my part-a failure of my epistemic responsibility.
Epistemic responsibility can be distinguished, at least in ideal cases, from practical responsibility. As a way of fleshing out what would be an ideal case, consider a situation where I can be reasonably be expected to have no knowledge of the internal constitution or properties of individual bricks. I am, let us suppose, in a foreign country, where I don't speak the language; there is only one manufacturer of bricks, and he is notoriously tight-lipped about the materials he uses in his bricks and the nature of the manufacturing process. In choosing to proceed with the bricks, I am practically responsible for them-I could have instead abandoned building, for example. However, in knowing essentially nothing about their internal constitution, I am not epistemically responsible for each brick. In such circumstances, we can (almost) neatly divide my authority over the bricklaying process into my practical authority over the bricks, and my epistemic authority over how each brick is laid. (I say "almost" because, of course, precisely the same issues that arose in connection with the bricks will also arise in connection with the cement I use to join them, etc.)
In reality, of course, matters will almost certainly be conceptually a lot messier. Epistemic responsibility shades by degrees into practical responsibility. In more realistic situations, there are certain other facets of the building process concerning which I may have my suspicions, but cannot really be sure. I may know nothing about the internal constitution of bricks. But I do know that factory A has a reputation for making good bricks, while factory B has a reputation for making cheaper but more questionable bricks. My decision to employ shoddy bricks from factory B then may be a partial failure of epistemic responsibility. But it may also be offset by other constraints-notably financial constraints, and/or the unavailability of bricks from factory A. However, crucially, what I am unable to do is identify the internal constitution of each brick in the way that I can identify the characteristics of good mortar work. So, although my employment of bricks of a certain sort may have epistemic aspects, it is largely a matter over which I have practical authority, but not epistemic authority.
So, subject to the vicissitudes of real-world messiness, we can distinguish practical authority and epistemic authority. Each of these senses of authority corresponds to, indeed coincides with, a sense of responsibility. My responsibility for the process and the product divides into my epistemic responsibility over certain parts of the construction, and my practical authority over the rest. But, one might argue, pure practical authority is little real authority at all. If the house fell down because of the inferior quality of the bricks, and if I were in a situation where I couldn't be expected to know this, then the house's demise, we would probably want to say, was not really my fault. If, on the other hand, it fell down because of shoddy pointing work on my part, then it would be my fault. I may be said to own certain parts of the building process, and certain parts of the result of the process, but not others. And this is equivalent to saying that I am responsible-epistemically responsible-for certain parts of the process and product, but not others.
For reasons that will become clear, I do not want to advance epistemic authority as a criterion of ownership of personal-level cognitive processes. What the concept of epistemic authority does provide us with, however, is a tolerably reliable way of distinguishing personal-level cognitive processes that we own from subpersonal processes that we also own (and whose ownership can be understood in terms of integration into the former). This is not, in general, enough for the purposes of this book, and in the final section of this chapter and in the chapters that follow, I shall attempt to push beyond the idea of epistemic authority to identify the deeper roots of ownership. However, even without this further excavation, the idea of epistemic authority does, I think, provide us with enough resources to successfully disarm the fourth standard objection to the amalgamated mind: the problem of cognitive bloat. It is to this that we now turn.
6 Authority and the Problem of Bloat
The concept of epistemic authority provides a useful way of thinking about the difference between personal and subpersonal ownership of cognitive processes. At the personal level, my epistemic authority over a process is a tolerably reliable indicator of my ownership of that process. The same is not true, of course, for subpersonal cognitive processes: these are characterized precisely by the absence of my authority. I am not author of, but hostage to, the processes that, for example, transform the raw primal sketch into the full primal sketch. Authority provides an epistemic criterion of ownership of cognitive processes. It clearly, then, is not the sort of criterion applicable to processes to which we lack epistemic access. Nevertheless, subpersonal processes can qualify as cognitive. And they qualify as cognitive because of the contribution they make to-the ways in which they are integrated into-personal-level cognitive processes to which the authority criterion is applicable. With this in mind, let us return to the problem of bloat.
We have no epistemic authority over the processes occurring inside the telescope, calculator, or computer. Therefore, they do not qualify as personal-level cognitive processes because, at that level, they would be processes owned by nobody. This is compatible, however, with their being subpersonal cognitive processes. They would qualify as such just in case they are appropriately integrated into the subject's (personal-level) psychological life. In this regard, they would occupy a role akin to subpersonal neural or computational processes: they can count as cognitive to the extent, and only to the extent, that they make the appropriate contribution to personal-level cognitive processes.
This, I think, renders anodyne the problem of bloat. First, the thesis of the extended mind is not committed to the idea that the processes occurring inside telescope, calculator, or computer are cognitive processes on a par with perceiving, remembering, reasoning, and thinking. None of these things occurs inside the telescope, calculator, or computer. Second, although the extended mind is committed to the claim that processes occurring inside these items can be subpersonally cogni
tive-on a par with, for example, the operations that transform the raw primal sketch into the full primal sketch-this is true only when the telescope, calculator, or computer is appropriately coupled with a cognitive subject: that is, a subject that satisfies conditions (1)-(3) of the criterion and is thus a subject of personal-level cognitive processes.
The problem of bloat is sometimes presented as if the thesis of the extended mind is committed to an uncontrollably expanding conception of the cognitive domain: as if my cognitive processes continue expanding ever outward into my notebooks, telephone directories, and Internet connections. This conception of the problem is unwarranted. According to the extended mind, the domain of the cognitive need extend no further than systems that are appropriately coupled to cognitive subjects-systems that, among other things, have epistemic authority over their activities. And even then, it is not every type of cognitive process that bleeds into the murky innards of artifacts: only subpersonal cognitive processes bleed in this way. And crucially, they do so only when the artifacts are appropriately coupled with cognitive agents.
Of course, according to the extended mind, personal-level cognition can also extend into the world in the form of my activities-my manipulation, exploitation, and transformation of information-bearing structures in the environment. However, with personal-level cognition also, this cognitive bleed is strictly limited: it extends no further than the activities over which I have epistemic authority. To the extent that epistemic authority is a matter of degree-merging gradually with mere practical authority-then the extension of my personal-level cognitive processes into the environment will also be a matter of degree. However, since the extent of the activities over which I have epistemic authority is strictly limited, so too will be the extension of my personal-level cognitive processes into the environment.
The New Science of the Mind Page 21