Willpower Meets the Computational Theory of MindPublished 1 September, 2011
Today is the official release date of Willpower: Rediscovering the Greatest Human Strength by Roy Baumeister and John Tierney. As I mentioned in a previous post, I’m fairly confident – but do not know for sure – that the book argues that willpower relies on a resource, and that the resource in question is glucose. My view is that “energy,” “glucose” and the metaphors used in discussions of this model are the wrong sort of explanation for the sorts of phenomena that come under the term “willpower.” Previously I reviewed some empirical reasons to doubt the idea; here I discuss some theory.
If, say, you’re using your computer and you find that your spreadsheet application is slow, it wouldn’t occur to you that the problem might be that your computer isn’t getting enough electricity. Of course all computing devices require energy in some form to function. However, the right place to start looking for explanations for understanding computational devices – natural or artificial – is, well, computations. If your computer is slow, you might want to close other applications, not check the power cord. In contrast, my electric razor, a mechanical device, slows down when the battery is nearly drained. A low battery won’t explain why my computer, a computational device, is slow.
Stepping back, even though debate continues, the cognitive revolution ushered in a new approach to understanding the mind, viewing the brain as an organ that processes information. The mind is not like a modern personal computer in all particulars, of course. But the reason that the brain consists of neurons is that these cells are good for processing information; muscle cells, in contrast, are good at moving things around.
The explanation for behavior, then, for things with brains, anyway, is to be found in the way that brains process information, just as with computers. In How the Mind Works, Steve Pinker answers his own question, “Why should you buy the computational theory of mind?” this way:
Because it has solved millennia-old problems in philosophy, kicked off the computer revolution, posed the significant questions of neuroscience, and provided psychology with a magnificently fruitful research agenda. (p. 77)
Can computations explain why people don’t persist on certain tasks? To take one example that illustrates the basic idea, consider a baby looking at a novel scene. Soon, all the new information is extracted from the scene, and looking at it feels (I imagine) “boring.” Eventually, she loses interest, and gives up. This phenomenon, looking away when all the information has been extracted, is a key aspect of the work using habituation methods in developmental psychology. When the baby turns her head, it’s not because her brain has run out of some sort of “mind stuff” or “attention stuff,” some resource needed to continue to direct her attention. It’s because of a computation in the baby’s head that the benefits of looking have gone down – no new information – and she’s better off looking elsewhere.
This isn’t to say that energy (or “power”) doesn’t matter at all. The brain can’t work without fuel. Moreover, how much energy an organism has available influences decision making. The polar bear in Disney’s earth took risks it otherwise wouldn’t as its food situation became increasingly dire, an effect mirrored in humans (see, for instance, Wang & Dvorak, 2010). This doesn’t mean that sugar suppresses risky decision-making. It means that how much energy you have on board is taken into account by the information-processing systems that are designed to make good choices about how risky, or patient, to be.
The notion of “willpower” or “energy,” then, is the wrong kind of explanation for why persisting on various tasks is hard. The right sort of explanation has to do with information processing, and the language of information processing can help explain not just decision making, but the other interesting element in this area of research, the sensation of effort.
Suppose, for the moment, I am a squirrel. I am sitting comfortably on a tree branch, but it has been some time since I’ve eaten. I’m not aware of it, but sensors in my body are taking measurements, assessing how much sugar is in my blood, how full my stomach is, and so on. These sensors send their information to the brain, and a summary is generated of how urgent the need is for me to eat. This summary is experienced by me, the squirrel, as hunger. As the measurement of my present caloric state gets smaller and smaller, I feel a greater and greater urge to look for a patch of food and eat.
Motivated by my hunger, I scurry to the base of a tree where I recall having seen some nuts ready to be harvested. In short order, I find them, and begin picking them off the ground and eating them. As I eat the nuts, the sensors in my body re-compute my current caloric state.
At first, I find the eating quite enjoyable. Every time I see a nut, I experience a little burst of satisfaction, and the taste of the nut itself is rewarding. (Disclosure: I have no idea, really, what it’s like to be a squirrel or, indeed, if it’s like anything at all to be a squirrel. Anyway…) My tiny squirrel brain is computing how good an idea it is for me to stay on this patch. However, sadly, the patch I’m on doesn’t have many nuts, and soon the time between finding nuts is long. No nuts for a minute, then two… processes I’m not conscious of are measuring how good a patch I’m on, and I experience the outcome of this measurement. Soon, I’m finding staying on the patch frustrating, even, perhaps, effortful… this is evolution’s way of telling me to move along… and I find that I want to explore the local area, maybe find a denser patch… (For an explicit version of this idea, see Hockey, 2011.)
Ok, I’m me again. A squirrel foraging in a patch of nuts has to make a decision which is, with apologies to The Clash, “Should I stay or should I go?” The way to solve this problem is through various sorts of computations: how good is this patch, particularly compared to how good the next patch I’m likely to find might be? The brains of organisms, such as squirrels, perform these computations, guiding adaptive behavior.
Van den Berg (1986), who I quoted in a previous post, noted that eighteenth century scientists thought that the mind was “like a muscle,” equating “the efforts of the spirit with those of the body.” These worthies were writing before Darwin, Turing, and the cognitive revolution, before anyone had the idea that the mind was a computational device with evolved functions (Pinker, 1999). The sensation of “effort,” and the disinclination to persist on certain kinds of tasks is going to require some sort of explanation in the language of information processing.
To get to the right explanation, we’re going to have to move past the ones that are wrong, and my reading of the current state of affairs is that this is unlikely to happen any time soon.
Hockey, G. R. J. (2011). A motivational control theory of cognitive fatigue. In P. L. Ackerman (Ed.), Cognitive fatigue: Multidisciplinary perspectives on current research and future applications (pp. 167-187). Washington, DC: American Psychological Association.
Pinker, S. (1999). How the mind works. New York: Norton.
Van den Berg, C. J. (1986). On the relation between energy transformation in the brain and mental activities. In R. Hockey, A. Gaillard, & M. Coles (Eds), Energetics and Human Information Processing(pp. 131-135). Dordrecht, The Netherlands: Martinus Nijhoff.
Wang, X.T. & Dvorak, R. D. (2010). Sweet future: Fluctuating blood glucose levels affect future discounting. Psychological Science, 21, 183-188.