What is weakness of will? And why do I think I can make progress on it?
Why do people do things they know they shouldn't do? We all do it. At the benign end of things, we reach for that second (or third!) brownie when we really know we shouldn't. In more pernicious instances, we often fail to help or even harm people all the while knowing that this is not the right thing to do.
This phenomenon has bothered philosophers since the time of Socrates, and they have called it the problem of akrasia, or weakness of will. The problem may be characterized as, Can individual genuinely act against her better judgment? If she can, how? And perhaps most importantly, why does she do this?
Philosophers have broadly adopted what I call an 'introspective approach' to weakness of will. They have looked inside of us to see what knowledge, intentions, desires or even feelings must be in place for a moment of weakness of will to occur. For a look at how contemporary philosophers of action examine weakness of will, please visit here.
My research takes a different approach. Influenced by empirical findings from experimental psychology and the cognitive neurosciences, my work starts out from the ugly fact that standard theories of action just don't square with our basic empirical understanding of how human beings think, make decisions, or behave. Accordingly, my current projects takes on two tasks: 1. criticizing standard philosophical theories of weakness of will, and 2. formulating an empirically-consistent account of weakness of will.
Evidence from artificial intelligence and cognitive neuroscience suggests that three complementary, goal-directed and habit-based mechanisms help the brain encode and predict values (Daw et al. 2005). The difference between the first two mechanisms can be illustrated using the example of planning a trip from work to a favorite restaurant. The goal-directed mechanism proceeds by imagining all the possible routes one could take and then trying to gauge which one would be the most efficient. By contrast, the habit-based mechanism accumulates experiences over time and relies on memory to gauge which route is generally the most effective (Fig. 1, adapted from Dayan et al. 2008).
This phenomenon has bothered philosophers since the time of Socrates, and they have called it the problem of akrasia, or weakness of will. The problem may be characterized as, Can individual genuinely act against her better judgment? If she can, how? And perhaps most importantly, why does she do this?
Philosophers have broadly adopted what I call an 'introspective approach' to weakness of will. They have looked inside of us to see what knowledge, intentions, desires or even feelings must be in place for a moment of weakness of will to occur. For a look at how contemporary philosophers of action examine weakness of will, please visit here.
My research takes a different approach. Influenced by empirical findings from experimental psychology and the cognitive neurosciences, my work starts out from the ugly fact that standard theories of action just don't square with our basic empirical understanding of how human beings think, make decisions, or behave. Accordingly, my current projects takes on two tasks: 1. criticizing standard philosophical theories of weakness of will, and 2. formulating an empirically-consistent account of weakness of will.
Evidence from artificial intelligence and cognitive neuroscience suggests that three complementary, goal-directed and habit-based mechanisms help the brain encode and predict values (Daw et al. 2005). The difference between the first two mechanisms can be illustrated using the example of planning a trip from work to a favorite restaurant. The goal-directed mechanism proceeds by imagining all the possible routes one could take and then trying to gauge which one would be the most efficient. By contrast, the habit-based mechanism accumulates experiences over time and relies on memory to gauge which route is generally the most effective (Fig. 1, adapted from Dayan et al. 2008).
Fig. 1: The goal-directed (left) and habit-based (right)
mechanisms
Because the goal-directed and habit-based value mechanisms use different computational algorithms, each mechanism works well under different kinds of uncertainty. Goal-directed predictions are more accurate in new situations, but struggle if there are too many options to consider. Habit-based predictions rely on extensive experience, but can make accurate choices even in very complex situations (Dayan et al. 2008). I argue that the tradeoffs between these two value attribution mechanisms hold the key to explaining weakness of will. Specifically, I contend that distinctive kinds of uncertainty cause the brain to abandon the use of the goal-directed mechanism for the sake of the seemingly more reliable, but actually suboptimal, habit-based mechanism. This suboptimal substitution causes the agent to experience weakness of will, and thus provides an answer to the puzzle of how an agent can act intentionally against her better judgment.
