The study of self-regulation is a big deal in social psychology, and the results are of interest to many, not just those who struggle to win the good fight against themselves. Roy Baumeister's ego-depletion model is currently the reigning paradigm. According to this model, self-regulation is an effortful act that consumes psycho-physiological fuel in the form of glucose. When effortful self-regulation draws energy faster than it is replenished, further attempts at self-regulation falter. To illustrate, imagine you are out shopping for groceries. Each decision between brands or options depletes the ego a bit. As you finally reach the check-out line, you are no longer able to resist the allure of gum, Ritter Sport, or People Magazine. If, however, you took a quick sip of glucose-rich lemonade before putting your groceries on the conveyor, all is well. You proudly resist the allure of merchandise not on your shopping list.
Although I objected in an earlier post to the wholesale condemnation of short-term gratification, I recognize that impulse control is often a good thing. Another matter is what self-regulation can tell us about human agency and the specter of free will that lurks behind it. Baumeister himself has come to think that successful self-regulation is one signal that free will exists. My reading of his argument is this: The immediate temptation, e.g., the chocolate bar in front of you, represents the pull of the situation. The chocolate bar is the stimulus that "wants to" control your behavior. It wants to be eaten by you. If you oblige, the situation scores a point for control over your behavior and you lose one. In contrast, if you stand firm against the bar's pull, you score a point for control over your behavior and the situation loses one. Now you are the agent; you did not respond to a situational demand, but you created your own behavior (or non-behavior, so to speak, in the example of snack refusal). As there is no external situational stimulus that we can point to to explain your behavior, we refer to a power within you, your strength of will. From there it is a short step to assume that this will was a thing onto itself, uncaused, and hence free.
Voilà, a moral victory over oneself (i.e., one's baser instincts) is now evidence for free will. In another post, I suggested that this idea is illogical and disingenuous. It is illogical because it is asymmetrical. Why should a free will only decide to do one thing, namely resist tempation, and not choose to eat pork rinds and guzzle beer? It is disingenuous because people do punish others for giving in to impulse, saying that they could have resisted because they have free will. In other words, the assumption of free will tends to enter our calculations before we look at the behavior, not after.
There are two other flies in the free-will ointment. One is the strange implication that the will's freedom depends on the glucose level in the brain. It seems to me that a statistical correlation between the level of glucose and achieved delay of gratification is a good example of a deterministic cause and effect (plus error) relationship.
The bigger fly is the implication of human exceptionalism. Non-human animals are typically not granted free will by those who defend the idea for humans. Some mammalian species may have awareness and some may even have self-awareness. But free will? Come on now! If you have a theory about how empirical findings reveal free will in humans (as Baumeister does) and if you want to reserve free will for humans (I don't know Roy's position on this), then you would have to predict that the critical empirical findings will not replicate with non-human animals.
But some do (see also a post on canine personality). In a clever paper published in Psychological Science, Miller et al. (2008) took the self-control paradigm to the dogs. They had their canine participants sit and stay for 10 minutes, which seems like a very long time indeed (even for the owners). The dogs were then praised and given a treat. Then they received a toy from which ordinarily another yummy treat could be extracted, and the dogs had experience doing this. For the experiment, however, the toy was sealed and the dogs struggled in vain. The finding of interest was that after having gone through the taxing sit-and-stay exercise the dogs were quick to give up on the toy. Dogs that did not experiences the prior strain of having to stay in place were more persistent. A second experiment showed that if the sit-and-stay dogs received a shot of a glucose drink, they persisted with the toy as long as the control dogs did.
These data patterns look exactly like the patterns obtained in experiments with humans and human-appropriate stimuli and behaiors. I think this is good news. It shows the cross-species generality of interesting and important findings. It also suggests with we can get away with simpler theories when explaining human behavior. There may be no need to stipulate a distinction between a "reflexive" and a "reflective" system of mind, where the former can only be found in humans. Most critically, the findings can help us suspend the belief in free will. Self-control, like other psychological-behavioral events, are lawful in man and dog. If you wish, however, you can choose the alternative option, that is, the view that both dog and man have free will.
The experiment holds another interesting lesson. Theories of self-regulation, and ego-depletion theory in particular, tend to follow the conventional moral consensus that it is good to resist temptation and that it is good to persist at a task. In other words, effort -- and the pain it entails -- is deemed morally good. The ego-depletion research suggests that an organism exhausted from moral strain will then act less morally. Now, putting aside the question of whether dogs should be judged from a moral perspecive, it appears that in their state of exhaustion, they did the smart thing; they did NOT persist at an unsolvable task. In other words, the state of ego-depletion improved behavioral decision-making instead of making it worse. Once again, performance rationality depends on the interplay of the organism and the nature of the task at hand.
Miller, H. C., Pattison, K., DeWall, C. N., Rayburn-Reeves, R., & Zentall, T. R. (2010). Self-control without a "self"? Common self-control processes in humans and dogs. Psychological Science, 21, 534-538.