The announcement of my contributions says that they will in part concern my recent Keeping the World in Mind: Mental Representations and the Sciences of the Mind. The overriding theme of that book is the proper understanding of “representation” as it occurs in recent cognitive neuroscience, broadly understood. I argue that “representation” is in effect ambiguous, with the sense typically understood in recent philosophy being quite different from that in much of cognitive neuroscience (CNS). In fact, the term seems quite unruly, to judge from dictionary entries such as this:
Full Definition of REPRESENTATION
1: one that represents: as
a : an artistic likeness or image; b (1) : a statement or account made to influence opinion or action (2) : an incidental or collateral statement of fact on the faith of which a contract is entered into; c : a dramatic production or performance; d (1) : a usually formal statement made against something or to effect a change (2) : a usually formal protest
a : representationalism (i.e., sense-data theory); b (1) : the action or fact of one person standing for another so as to have the rights and obligations of the person represented (2) : the substitution of an individual or class in place of a person (as a child for a deceased parent); c : the action of representing or the fact of being represented especially in a legislative body
3: the body of persons representing a constituency
None of the M-W definitions captures the sense of “representation” broadly present in analytic philosophy, where representations have content and satisfaction conditions. In the following post we will look at the thesis that another sense underlies the word in cognitive neuroscience; we will also see an argument that the representations in the philosopher’s favorite senses do not have a place in the recent sciences of the mind. Let us just note for now that there do seem to be alternatives to the philosopher’s sense of “representation” as implying content and satisfaction conditions.
Though we will look at topics from Keeping the World in Mind, my primary intention is to introduce new material or material that is underdeveloped.
Dopamine and its roles
We are going to start, then, with something quite different. And that is the role of dopamine in a range of phenomenon concerned with perception and action. We will look at cases of the following: socially inappropriate behavior, OCD and Affordances.
There are a lot of reasons to bring in dopamine. First of all, it is going to be handy to have an example of research to hand in various discussions. Secondly, I think the topic of dopamine, considered in connection to perception and action, raises some interesting questions about philosophical methodology. The questions taken together ask us about the epistemology of philosophy of mind. To what extent can we explain our experience just by having it and philosophically reflecting on it. For example, the metaphor of a beacon suits Herbert Dreyfus in his descriptions of experience and to a limited extent Susanna Siegel agrees. But is our visual experience in pursuit of a goal really like beacon? More generally, evolution has produced creatures who have to act and react extremely swiftly. Might a very rapid complicated causal cluster seem to us a fairly straightforward linear progression, on which an ordinary metaphor fits?
I am certainly going to simplify my description of what we know about the role of dopamine. I’m following the public discussions of neuroscientists I know or at least have listened to, so I intend to remain accurate if also incomplete. Further, what I’m discussing is a matter of on-going research and there still may be major adjustments in the relevant theories. The theories are certainly under scrutiny, and lots of answers are not yet in. So I am not announcing the truth as we know it.
Theories in philosophy of mind should similarly be understood with the same tentativeness. Everything is up for revision. So the idea in these sections is to try to bring up and start to put together items that are not all entirely common matters of discussion.
There is still a great deal to understand, but it does appear that dopamine plays very important causal roles in the production of our actions. As I indicated above, one important question for philosophers is whether the roles it plays are at odds with some of the results of recent phenomenological analysis, informally understood. We do seem to think that we can concentrate on how things seem to us and how our experience unfolds, with a view to understanding the causation in the buzzing and blooming. But when we do this, do we get it right? The literature on affordances makes it clear that many philosophers take more or less as starting points an understanding of quite causally complicated experiences. I think it is common to think they are right. But should we do so?
One role dopamine plays, probably in conjunction with serotonin, concerns a great deal of very fast modulation of our social interactions. Here dopamine may function as a reward signal and serotonin as a disincentive signal. For these purposes, they are emitted in the anterior cingulate cortex and the insula. When someone functioning cognitively well does not get how to behave acceptably, chances are they have deficits in dopamine or serotonin in these areas of the brain.
Another crucial role of dopamine is in what is called reinforcement learning. Variations in pulses of dopamine signal whether things are going as expected, or worse or better, as one progresses toward a goal. One can think of dopamine as a very important part of a guidance system that tells us about how various things in our environment are to be valued. It is important that while dopamine will be originally set off by a rewarding experience, the burst comes to accompany instead experience of a reliable conjunct of the rewarding experience. A whiff of smoke, for example, may send one off looking to borrow a cigarette. The burst of dopamine will make the whiff stand out in one’s experience and set one off on a motor routine.
In addition, different psychological theories may fit the same dopamine facts. For example, Friston et al. (2014) provide a Bayesian account of goal seeking, and note that on their account the decision making looks a lot like reinforcement learning in which we think dopamine plays a large role. If, following Montague P. Read Montague, King-Casas, and Cohen (2006); R. Montague (2007) we think of humans and other sentient creatures as continuously tracking the value of what they are doing, we might think of them as continuously seeking optimal grip in Merleau Ponty’s sense. That is, they seek what one has when one’s gotten things right. Thus, we might without thinking turn off a stove, pick up our keys, leave, and lock the door. The process is not one we have to reflect about, just because we have gotten it right. We have optimal grip.
Herbert Dreyfus has pursued similar ideas for some time. As he says,:
… when we are ready to leave a familiar room we not only do not need to think that the door affords going out. We need not even respond to the door as affording going out. Indeed, we needn’t apprehend the door at all…Thanks to our background familiarity, when it is appropriate to leave, we are simply drawn to go out. In general, the absorbed coper [one who is coping – ajj] is directly drawn by each solicitation in an appropriate way: the chairs draw him to sit on them, the floorboards to walk on them, the walls may draw him to hang pictures on them, the windows to open them, and the door may draw him to go out. (2013-03-05). Mind, Reason and Being-in-the-World: The McDowell-Dreyfus Debate (Kindle Locations 455-462). Taylor and Francis. Kindle Edition.
If you know about OCD, then you’ll know the very unfortunate victim of this disease will find the scene just described as full of causes for anxiety. Was the stove really turned off? Were the right keys picked up? Perhaps those picked up were really ones needed by one’s partner. These two problems can even prevent a third worry – was the door really locked? – from arising. Rather time is spent checking and rechecking the stove and the keys. Even worse, repetitive and-washing may interrupt and become the dominant behavior.
Dopamine is implicated in OCD, experimental results indicate, and as with other dopaminic disorders, such as Parkinson’s, deep brain stimulation may help. And there may be very important connections here that we can draw between the operation of dopamine and the phenomenology of OCD. Haan, Rietveld, Stokhof, and Denys (2013) provide a phenomenological interpretation of OCD and the results of Deep Brain Stimulation in sufferers who do not respond to standard therapies. According to the authors, at the heart of the compulsive component of OCD is an inability to have a sense of optimal grip. Consequently, one enters a quite terrible loop of hyper reflexive thinking about one’s actions, repetitive actions and failing to achieve a sense of having gotten an optimal grip.
It is a very standard interpretation of OCD that the checker is checking to relieve the anxiety. This seems odd, though, since the checker will certainly be aware that the checking does not work for long at all. The idea that one lacks any sense of having completed something seems to me a more foundational account of what is going wrong. (It is not clear that Haan et al see these two interpretations as wholly distinct.
There are, then, four things that are being put together to, with any luck, give us better insight into OCD: (1) the failure in a sense of optimal grip, (2) repetitive behavior; (3) dopamine mis-functioning and (4) dopamine’s role in guiding us to the achievement of a goal.
The discussion so far bears on a recent philosophical dispute between John McDowell and Herbert Dreyfus. The dispute concerns whether there need be any representation of the goal in many cases of what we would say are goal directed behaviors. Dreyfus claims there need not be:
… in familiar situations there need be no representation of conditions of satisfaction; rather, the absorbed coper [one who copes] behaves like a pilot following a landing beacon. For the pilot there need be no representation of a goal. When things are going well, the beacon is silent. However, when the plane deviates from the optimal path a beeper sounds. The pilot can then remain in flow and directly respond to the signal by “automatically” correcting his course, or he may sense a tension that draws him to correct his course. Only when things are not developing normally and no alternative perspective directly draws the coper to replace the current one, does the coper have to represent a goal and deliberate as to how to reach it. (Op cit.)
In this passage, Dreyfus can seem to echo what proponents of reinforcement learning say. That is, as long as one is on course, the pulsation of dopamine is the steady “Things are fine.” That does not mean that, however, that there is no need for a representation of the goal. The metaphor of a beacon is faulty for several reasons. We do now have an account of how we manage to hone in effortlessly on a goal and it employs neural calculations on, among other things, representations of the goal. And we do not have a better account in the making.
As far as I can understand it, the cognitive neuroscience researchers advance two closely related reasons for wanting a representation of the goal. The first is a quite general one: there are good reasons for thinking goals are standardly represented in the prefrontal cortex, and there are quite acute problems when they are not. As it were, we do not stay on course. Secondly, the computations that result in the goal-tracking dopamine are over goal representations, among other things (P. Read Montague & Quartz, 1999; R. Montague, 2007).
A very common take on affordances has one or more objects in a visual experience attracting us and causing our action toward it to use it. We could see this as captured by the metaphor of a beacon that keeps us on track. I am going to start with two slightly different episodes, both of which are from actual experience, for what that is worth. It will emerge that my family is fond of cats, in case anyone wonders about that.
Walking in Soho with friends, I find myself pulled over to a store window. A bit puzzled by what attracted me, I see there is a dusty stuffed cat in the window. Ugh. I do enjoy looking at live cats in window of stores, but a mistake was made in this case.
Walking in a meadow with my then 3 year old son, I hear him shout “Kitty” and see him charge off with his arm out and pointing. I cannot see any cat, but clearly that’s what he is after.
We have a neat story about what is going on in these cases. Each of us has caught a glimpse of a cat part, the dopamine goes off and it activates an action plan that takes us toward what we glimpsed. The dopamine can also change the visual perception of the part and the full object, if we see it. The part stands out as particularly salient. However, it is the dopamine which is sending one into action, not the visual salience which is a co-effect of the dopamine (see Montague above). The glimpse of the cat part has a causal role in the production of an action but, since that has occurred before the dopamine goes off, it probably is not the sort of thing that counts as an affordance pushing us into action.
As the process resulting in cat pursuit occurs, it is very rapid. The more phenomenologically based theory on which we see the object and then something kicks our body into action posits a slower process, one in which one’s chances of getting the reward go down. The cat may wander away, someone else may beat us to the comfortable chair and the chocolate may get consumed before we get to it. Of course, such disappointments may occur anyway, but if we are in action before the visual processing gets very far, then we may have a better chance at success.
Let me note in general that it seems to me that when speed is important, the effects of the sensory uptake of a relevant object are moved down, as it were, to earlier and earlier stages of sensory processing. Thus, for example, knowing how temporally close a reward is can be very important, and it looks as though that is registered in very early visual processing (Shuler & Bear, 2006). Gender can be important, and it’s a good guess that we register gender signs very early on, even if the full recognition of a person with a gender is the result of a long process
Let me quickly mention those who do seem to hold that the visual experience itself is what urges us into action. These theories tend to be concerned with affordances. There are many different theories of affordances, but I will look briefly at what we can consider a cold version and a hot version. The cold version has the content of a perception of an affordance to include an imperative. The hot version has the affordance getting at one, pulling one in and mandating some actions. With the first we can associate Wheeler and Clark, and with the second Dreyfus and Siegel.
Two Kinds of Theories of Affordances
Clark (1997) takes the experience of an affordance to give us ‘action-oriented representations’. Such representations have a declarative component and an imperative one. These representations say something like ‘This X is before you; use it.’ In such a case, the representations have two satisfaction conditions; they need to fit the world, but the world also needs to fit them, if both conditions are to be satisfied. If there is an X here before you, the first component is satisfied; if you use it, then the second component is also. (I will use the ‘fit’ metaphor in parallel with the ‘corresponds to the facts’ locution; thus, true statements fit the world, while with fulfilled commands, the world fits them. This is in contrast with the more standard use of the ‘fit’ metaphor, which I find intuitively puzzling.)
Wheeler (2005, 2008) focuses quite entirely on the action part of the representation, but it still is a philosophically standard representation and appears to have aboutness and satisfaction conditions. Rather like a shopping list, the action-oriented representation will only be satisfied by the environment’s fitting it.
Imperatives and shopping lists we can see as cool ways to be action oriented. We have seen Dreyfus above maintain that the experience in some way acts on us, perhaps signally like a beacon. Some of Susanna Siegel’s affordances also appear to act quite strongly on us.
… an important class of experiences of affordances… are experiences of the environment as compelling you to act in a way that is solicited or afforded by the environment. I call such experiences experienced mandates. They are generally structured by how you are already acting in a situation—not only by how you can act or are disposed to act in it. From your
point of view, the environment pulls actions out of you directly, like a force moving a situation, with your actions in it, from one moment to the next… the environment invites or solicits an activity, and you experience these affordances by doing, and by feeling moved to do, what they invite you to do. … “(2014-09-05). Does Perception Have Content? (Philosophy of Mind Series) (Page 52). Oxford University Press. Kindle Edition.
Siegel’s affordances come in degrees, with mandated affordances pulling us much more strongly than others. What is common to all her statements about the latter sort is that if a perceptual experience of an affordance involves one acting in accordance with it, then one has been moved by the experience. Thus it is experiences that do the mandating and motivating.
What we have seen of reinforcement learning makes it at least unclear that the perceptual experience is moving us to act, contrary to how many people appear to experience such situations. What sets us off is the dopamine burst that results from the experience of a reliable conjunct of the reward (which may be the reward itself). Thus, the sight of a door might set me off on my way to my very comfortable sofa. Further, in addition to activating a motor plan, the dopamine will alter one’s experience of both the reliable conjunct and finally the sofa itself. This brightened experience with its increased salience is often what philosophers seem to mean by saying that we experience affordances that demand actions, that in some sense call to us. But the motivating force comes from the dopamine, not from the experience.
Experience of the reward – for example, the sight of my super comfortable sofa – may have a causal role to play in the kind of pulsing the dopamine bursts display, but it is the dopamine that has set us on the path and initiated the motivation. And here Dreyfus seems to have it more right than Siegel. That is, we often hardly need to see the targets of our actions. Further, to take the content of one’s experience to include an imperative seems also to mislocate the origin of the movement. It is in us, not in our environment.
It may seem indeed better to say that what it going on is that one really desires something, such as a comfortable resting place or a piece of chocolate cake. If someone is smoking a cigarette in a house, the addicted smoker may find the scent is a very obvious component in the ambient smells, but what is driving her to find the cigarette is her desire. The desire may well be there before any perception of one’s target. But if this is correct, we need to ask how the desire fits in with the conjunction of the dopamine burst, activated motor plans and the heightened perception.
There is a problem with fitting desire in. Desires certainly can look like propositional attitudes, and it is unclear whether an appropriate propositional attitude is on the scene. We have a quite full account of what is moving us and though it is on what is sometimes counted as the sub-personal level, we can understand it in human – or sentient animal – terms. But it is not at all clear that we need a propositional attitude in the picture. For example, we can even assess the action as prudent or not without having to assess a propositional attitude.
In the next section, I will take up the question of whether we can locate a propositional attitude in the scenario described that can be desire that P. This question will lead us to consider mental representations in the sciences of the mind. For now we have raised some questions about whether the claims of cognitive neuroscience about the role of dopamine can challenge claims in philosophy of mind. Dreyfus on representations and Clark, Wheeler, Dreyfus and Siegel on affordances all take certain features of their experience to be accessible to them. However, it may be that causal complications are not easily discerned, and the result may be theories making questionable causal claims. That said, to be fair to Dreyfus, it may well be that those who assert representations are present in reinforcement learning are not actually asserting something he denies. As we will soon see.
Clark, A. (1997). Being there : putting brain, body, and world together again. Cambridge, Mass.: MIT Press.
Friston, K., Schwartenbeck, P., FitzGerald, T., Moutoussis, M., Behrens, T., & Dolan, R. J. (2014). The anatomy of choice: dopamine and decision-making. Philosophical Transactions of the Royal Society B-Biological Sciences, 369(1655). doi: 10.1098/rstb.2013.0481
Haan, S. d., Rietveld, E., Stokhof, M., & Denys, D. (2013). The phenomenology of deep brain stimulation-induced changes in OCD: an enactive affordance-based model. Frontiers in Human Neuroscience, 7. Retrieved from
Montague, P. R., King-Casas, B., & Cohen, J. D. (2006). Imaging Valuation Models in Human Choice. Annual Review of Neuroscience, 29(1), 417.
Montague, P. R., & Quartz, S. R. (1999). Computational Approaches To Neural Reward And Development. Mental Retardation And Developmental Disabilities Research Reviews 5, 86–99.
Montague, R. (2007). Your brain is (almost) perfect: How we make decisions. New York: Penguin Group.
Shuler, M. G., & Bear, M. F. (2006). Reward Timing in the Primary Visual Cortex. Science, 311(5767), 1606-1609. doi: 10.1126/science.1123513
Wheeler, M. (2005). Reconstructing the cognitive world : the next step. Cambridge, Mass.: MIT Press.
Wheeler, M. (2008). Minimal Representing: A Response to Gallagher. [Article]. International Journal of Philosophical Studies, 16(3), 371-376. doi: 10.1080/09672550802113276