• No results found

Functionalism, CTM and collective mentality

At this point I’ve collected all the tools necessary for constructing a theory of collective mentality that shows how they are analogous to the mental states of other cognitive systems. I’ll start to develop this theory of collective mentality by considering an argument offered by D. H. M. Brooks (1986) in his paper “Group minds”. Brooks argues that accepting functionalism entails at least some cases where we would be warranted in attributing mental states to collectivities. I’m inclined to think that the sort of story he tells is insufficient in the end, however it does provide a foundation, and with some elaboration this view becomes incredibly plausible.

If materialism is true,26 then there will be some supervenience relation between the mental and the physical. What this relationship is, however, is far from immediately obvious. One way of looking at this relationship is in terms of type identities between mental states and brain states. However, alternatively, and far more plausibly, we can adopt a functionalist view with token identities between mental states and brain states.27 Type identity seems too strong to adequately capture our standard understanding of the mental. After all, we should expect no psychological change in a person where one of her neurons is replaced with an artificial neuron (Brooks 1986, 456). But if replacing one neuron with an artificial neuron maintains psychological identity, there is no reason to suppose that a system consisting exclusively of artificial neurons having the same functional properties as neurons and the same relational properties as your neurons would have different psychological properties from you.28 But, artificial neurons can be built in a variety of ways given that the relevant function of a neuron is merely to be on/off-signaling devices. So, anything that can function as an on/off-signaling device can be used as an artificial neuron. A person can function as an

26 I think that you can get away without buying this supposition provided that you’re willing to concede a

supervenience relation between the mental and the physical (which even Descartes was willing to do). There are, of course, deep questions here that would turn on a particular sort of reading of Descartes dualism—and answering these questions would lead us quite far a field of my main line of argument. However, this is Brooks’ (1986) argument and he presupposes materialism.

27 I have serious reservations about the truth of this claim. To begin with, there is reason to think, with Burge

(1979 and 1986), that the content of mental states depends not just on these molecular identities but also on facts about the social histories in which the relevant concepts were learned. Moreover, if the extended mind thesis pushed by Clark (1997), Clark and Chalmers (1998), Clark and Wilson (forthcoming), and Wilson (1995, 2004) is correct, then the environment (both physical and social) of an entity will also be relevant to the identity of mental states because according to this view mental states are not to be understood just as states of brains but instead as the computational states that facilitate various sorts of actions by a system—some of which may be extended or embedded in various ways that are not wholly dependent on the neural state of the system. I’ll return to these worries throughout the thesis.

28 This argument, of course, comes far too quickly. Although a commitment to functionalism does entail the

truth of the proposition that a system whose functional characteristics were identical to yours would be psychologically identical to you, Brooks claim only follows if you suppose that there would be no difference in the functional characteristics of a system that consisted exclusively of artificial neurons. This, of course, has not been established—and Brooks does not attempt to establish it.

on/off-signaling device, so a person can be an artificial neuron. So, there is no reason to suppose that a system that consisted of a number of persons arranged so as to replicate the relational and functional architecture of the neurons in your brain would have different psychological properties from the individual that such a system replicates.

As Brooks (1986, 457ff) puts the point, if we were to collect everyone in the nation of China and create a Brain City that replicated the functional architecture of your brain, there is no reason to suppose that Brain City would lack any of the psychological properties that you possess. In fact, Brain City could even have a ‘drink’ that had the same effect of it that alcohol has on us provided that we could introduce runners into the system that would “dash through the appropriate parts of the city doing the analogue of whatever it is that alcohol molecules do, damping down neurone response levels or changing the signal relations or whatever” (Brooks 1986, 457). The thought is spelled out simply as follows: If functionalism is the correct account of the mind, then it is not the biological properties of neurons that allows them to realize mentation, rather it is their functional properties; and if a group of people can have the same functional properties as a person then they can realize a group mind that has all of the same psychological properties as that person does.

Sure enough, this is one way to defend collective mentality. However, there are other ways of spelling out functional analysis of mentality that provide equally plausible accounts of the mentality of collectivities—indeed accounts that will diverge in interesting ways from this model. While we might choose, with Brooks, to look to the implementation level for an account of the relation between psychological states and their realizers (In fact, this is the way that much of the literature on collective intentionality has gone),29 there is reason to

think that there might be some systems were what we should pay attention to the functional organization of a system at the level of computational structures and the algorithms that they carry out. While looking to the implementation level might suggest some places where collective mentality is possible, in order to get to the most interesting cases of collective mentality in the actual world it will help to look at a variety of other levels of explanation. In order to get a handle on how this might be possible, it will help to look to a distinction between functional roles and the occupants of those roles, between programs and realizers, or between software and hardware. Now, we might rest content with a picture of the mind that operates just in terms of two-level explanations, looking to see which sorts of physical structures underlie which sorts of psychological functions and then trying to find analogous structures to see if a system can have mental states. The worry, however, is that this picture fails to capture all of the relevant similarities between mental states.30

Consider what it takes for something to count as the program that we know as Mozilla Firefox. To begin with, we can find a LINUX version of Firefox, a MAC OS X version, a Windows version, etc., and each of these programs counts as Firefox in virtue of the sorts of computations and algorithms that the program is running. Even though the program is running on different platforms, we identify the program in virtue of what the system can do. Now, a functionalist might assume that there is exactly one way to make sense of the relation between the mental and the physical. The functionalist might claim that the only relevant analogies between cognitive systems are at the implementation level—but doing so is surely a mistake (cf., Lycan 1987). Brooks supposes that the best way to defend

30 I do not, here, mean to claim that Brooks (1986) is committed to two-levelism. What I intend to suggest here

is that there is no reason to think that this is the only way in which collective mentality can be realized. Numerous attempts to explicate the relations between collective intentionality and individual psychological states have focused on this sort of explanation—I find none of these stories compelling. I turn to some of the problems with this approach in subsequent chapters.

the hypothesis of collective mentality is by showing that the functional role of belief, for example could be realized in terms of artificial neurons. However, this is not the only level of explanation to which we might look if we are to develop a story of collective mentality. In fact, for many of the collectivities to which we might be willing to attribute mentality, there is a far more promising story to be told. As Bill Lycan has correctly pointed out, it only makes sense to say that something is a role as opposed to an occupant modulo a particular level of explanation. Thus, where an individual person is concerned, it might make sense to say that a language of thought is the software and the wetware is the hardware, there might be also be reason to say of a corporation that the marketing plan is the software and the individuals in various departments at the corporation are the hardware. All of this will depend on what sort of explanation we are looking for. Let me, then, turn to an elaboration of the claim that there might be another way to look at the realization relationship in collectivities.

In the remainder of this thesis, I adopt the sort of position that Lycan (1987) refers to as homuncular functionalism, or homunctionalism. The key claim of homunctionalism is best put in terms of Marvin Minsky’s (1988) notion of a society of the mind. The thought is that the functional architecture of the mind is best understood as having a similar structure to a corporate hierarchy. There are various different divisions of the mind, each of which is dedicated to a particular sort of computational task. Consider the visual representation of a beer bottle being thrown at your face as you watch Patti Smith play the last show ever at CBGBs. In this case, there will be a system dedicated to detecting motion, a system dedicated to constructing representations of objects (presumably out of the representations of edges and colors that have been constructed by simpler systems), and a system connecting these

perceptual representations to an action system, and probably a whole host of other systems beyond these. On the homunctionalist theory of mind, the way to explain the functioning of each of these homunculi is by an appeal to the more specialized homunculi that constitute it, and by detailing the behaviors of these homunculi to explain how they produce a corporate output rather than going all the way down to the level of neurology. Lycan (1987) contends that mental states are type-identical with the property of having such-and-such an institutionally characterized state obtaining in one (or more) of one’s appropriate homunctional departments or sub-agencies.31

Now, suppose that we find cases where we are willing to ascribe mentality to collectivities from the standpoint of commonsense psychology. Further, suppose that we develop an account of what it is to have a particular sort of mental state in terms of the functional characteristics of such a state. At this point, the important questions focus on the how this function is actually realized in the system. We have garnered some tools at this point for thinking about these functions in terms of the sort of information that they are able to process. What we will need to look for is a functional decomposition of the particular sort of mental state that we are attributing to the system. What we will do at this point is try to spell out a sort of boxology, or corporate hierarchy for the relevant sort of belief.

Assuming that it is possible to develop such a picture, we will then look not to the level of physical mechanisms, but instead to the computational processes that need to take place in order for a particular sort of mental state to be possible. This story will be told in terms of the passing of information from homunculus to homunculus. Provided that there are the right sort of homunculi in a system and provided that these homunculi are able to pass the

31 Lycan spells this out in the case of pain: To be in pain of type T is for one’s sub…sub-personal

Φ-er to be in a

right sort of information to one another, this will yield a system that can possess genuinely mental state. At this point it will be a further question what the realizers are for the function of each of the relevant homunculi. This question will either be answered in terms of a further homuncular decomposition, or at some point just a story about the brute physical primitives.

This is the important point: for something to count as a legitimate psychological process, all that needs to be the case is that the process is realized on some sort of

representational mechanism that is running the right sorts of computations. For collective mentality to be possible, all that will need to be the case is that there is an isomorphism between the computational processes that occur in individuals when we attribute a psychological state ψ to them and the computational processes that occur in collectivities when we attribute ψ to them. Here also lies the key to the story about the autonomy of psychological explanation from neurology. Neurological explanations are irrelevant to psychological explanations except (and this is a very important consideration) in so far as the neurological explanations suggest that the psychological story can’t be realized in the system in question. So long as the homunctionalist decomposition comes out right, psychology should be happy to concede mentality regardless of how the computations happen to be realized. Thus, we should concede that a collectivity has mental states to the extent that it exhibits the right sort of homuncular decomposition, regardless of the sorts of states that are possessed by the individuals that compose the collectivity.

If we have reason to suppose that the right sorts of computations are being carried out by some subsystem (or sub-subsystem) of a collectivity, then we have good reason to say that our ascriptions of collective mentality are warranted. Alternatively, if the system is behaviorally identical to a system that has a particular mental state but there are no such

computations going on, then we have no reason to think that the system in question is a cognitive system, and thus we have reason to suppose that our attributions of collective mentality are not warranted. It at least seems possible that there could be a collectivity that engaged in computations that were functionally equivalent to the computations that individuals engage in. Thus, it seems like collective mentality is possible.

1. 4. Concluding remarks:

There are three points that I want to make clear about this preliminary sketch of the view I wish to defending. First, what I am concerned about is the functional architecture of cognitive systems. While there are a number of views that have been defended in recent years about the nature of collective intentionality (e.g., John Searle 1990a, 1995; Raimo Tuomella 1992; Margaret Gilbert 1987, 1989), these views have typically been concerned with explanations at the level of the implementation of collective intentions by individual psychological states of various sorts. I do not want to argue that such explanations are unimportant. In fact, I am inclined to think that it will, at some point, be quite important to determine how it is that the individuals who compose a collectivity can implement collective intentions. This is true in just the same way that it will be important to know how the brain realizes mental states in individuals. However, we do not, at least for the purposes of doing psychology, need to be committed to any particular story about the implementation of collective mental states.

Second, we need to be careful with the sorts of states with which we concern ourselves in the analysis of collective mentality. If we start with the sorts of cases that are typical of the collective intentions literature, we find ourselves entangled in a number of

debates that are better left to the side. Margaret Gilbert, for example, often starts with cases like deciding to take a walk with someone. However, if we start with these sorts of cases, we will focus on facts about the individuals and what it takes for them to intend to walk together rather than facts about the functional organization of the system in question. This leads to a number of serious worries about how it is that you can get from facts about the psychological properties of aggregations of individuals to facts about genuine collective states. However, if we start with more complicated cases where the corporate architecture is more salient, the functionalist picture that I want to defend becomes much clearer—we see how the psychological properties of the group are actually isomorphic in the right way to the psychological facts about individuals. This being done, it is easier to see that the actual psychological states of individuals aren’t really what are at issue in defending collective mentality. My contention is that we need to have the right sort of functionally characterized system in any case where we find collective mentality. The problem with the previous literature on the topic is that it’s just harder to see this sort of organization in simpler systems.

Finally, I’ve argued mental states form a natural kind from the standpoint of psychology and cognitive science. We should not expect there to be one account of individual mental states and another completely distinct account of collective psychological states—at least not so long as we are attempting to develop a theory of collective psychology. Once we know what it is to have a mental state, we can apply this theory either to the individual or to the group and the explanation will be the same sort of explanation in both cases. In the same way that we don’t need a story about the properties of individual neurons