Using Behaviorism to Train the Intuitive System

In his landmark book Beyond Freedom and Dignity (1971), psychologist B. F. Skinner argued that the concepts of free will and responsibility for one’s own actions had outlived their usefulness. Instead, he suggested that human behavior could be explained entirely by the intersection of environmental contingencies, instinctive reactions, and learned responses. Together, instincts and conditioned responses made up a “behavioral repertoire” of pool of all possible responses for that individual at that point in time. Two Minds Theory incorporates this perspective in suggesting that the Intuitive System produces behavior, not the Narrative System. The Narrative System has been historically associated with ideas of “freedom and dignity” based on rational choice. Furthermore, TMT posits that the Intuitive System can be “trained,” adding potential responses to the behavioral repertoire and making some responses more likely than others by way of deliberate practice.

Skinner articulated basic principles for training the Intuitive System in the 1960s, and they have been confirmed through decades of research. Behaviorist studies have yielded well-established principles and broadly replicated findings (e.g., in promoting organizational development, preventing adolescent substance use, addressing children's behavior problems, changing self-care behaviors to address chronic pain, or reducing suicidal behaviors, among many other applications). Behaviorist principles can be shown with a rat in a Skinner box (pictured above) and also work well in efforts to change human behavior.

Some well-established “training” strategies based on behavioral principles can be easily included in interventions based on TMT:

- Rewards work better than punishments. In experimental studies, rewards produce the desired behavior much more consistently than punishments do. In any animal, punishments can provoke a fight-or-flight response that actually interferes with successful learning at the neurochemical level. And in humans, punishments can also induce feelings of resentment or anger against the person administering them, which further decreases willingness to perform the desired behavior. A reward strategy can be done poorly or done well, but at this point there is little question that punishment is the weaker tool in the behavior-change toolbox. So, for example, it is better to give my daughter a small treat when I notice her starting her homework right after school than for me to threaten removal of all electronic devices on days when she doesn't. It should be noted in passing that negative reinforcement -- the situation in which someone performs a behavior to escape from some negative stimulus -- is a type of reward and can also be a powerful motivator. The difference in this case is that the punishing stimulus is seen as simply a part of the background environment, and the behavior is effective in reducing it. An example might be one's willingness to pay extra money to avoid hassles at the airport, where a small fee can seem like a reasonable trade for an escape from unpleasantness. If the same hassles were applied as a punishment -- "cooperate or else we will move you from the exit row to the back of the plane" -- they would be much less successful in shaping people's behavior.

- Extinction is hard, so substitute an alternate behavior instead. The only situation in which extinction happens quickly is when a behavior is always associated with a reward (a.k.a., continuous reinforcement). In that case, the reward becomes psychologically indistinguishable from the act of pressing the lever, and breaking the connection makes it fade out quickly. For instance, if you secretly substitute decaf coffee for caffeinated in my kitchen, I am likely to stop drinking coffee as soon as I recover from caffeine withdrawal. But in most situations, we receive rewards like a compliment or a paycheck only periodically, not immediately after performing a behavior. In that context, old habits die hard, and they are quick to reassert themselves as soon as a new reward is encountered. Even worse, a well-established behavior may actually increase when you try to extinguish it -- in my example I might drink several cups of the decaf before giving up! The paradoxical increase of undesirable behaviors is familiar to any parent who has tried to shape a child's behavior by "just ignoring it"; the behavior, which has in the past resulted in an amusing outburst from its target, actually tends to increase before the instigator gets tired of it. In these situations, or when removing the reward is impractical (e.g., even if you empty my stores at home I know that Starbucks still sells coffee with caffeine) it is better to substitute a new behavior instead, and to reward people for performing that behavior. So, for instance, if you were to remove my coffee and stock my kitchen instead with lower-caffeine green tea, I would be more likely to maintain the switch even when coffee again became available.

- Make low-frequency behavior contingent on high-frequency behavior (the Premack principle). Often we want to train the Intuitive System to do something that it isn't used to. But it's hard to reward something that almost never occurs. The solution is to find something related that does occur relatively often, and use that as a cue to prompt the desired behavior. A strategy that helps many people exercise is to own a dog: I don't normally think to interrupt my day to go take a walk, but if I don't walk my poodle by about 10:30 in the morning he starts to become restless. And when he gets restless, he lets me know it! Low-frequency behavior (me going for a walk) is then prompted by high-frequency behavior (my dog chewing on my leg when he is restless). Similarly, during coaching sessions on motivational interviewing, my colleagues and I train people to use a high-frequency behavior (clients denying they have problems) to prompt a low-frequency behavior (reflecting back the client's statements, which is something most of us don't do naturally in everyday conversations). For any behavior that you would like to see more of, the trick is to think of something else that does happen routinely, and to use that as a cue for the desired behavior. Once the desired behavior occurs, you can sustain it through rewards as described above.

- Use variable-ratio reinforcement to strengthen desired behaviors. The best way to help someone establish a behavior and to maintain it over time is to keep them guessing about when they will receive a reward. This means not giving a reward every time the behavior occurs (a "ratio" reward schedule) and keeping the actual schedule unpredictable ("variable ratio," as opposed to a "fixed ratio" like 1 reward every 3 times the behavior occurs). I once counseled a young woman who had broken up with her boyfriend and wanted him to stop calling. But every now and then when he called, she took pity and listened to him talk about how much he missed her, how things could be different in the future between them, etc. This would inevitably remind her why she didn't want to be with him, but unfortunately it also provided him with a variable-ratio reward (validation, listening) that was just enough to maintain his problematic pattern of phone calls. The same principle can be used in a positive way: For example, I want my dog to come when I call, but if I really want him to do it when there are competing reinforcers (a squirrel, an open gate), I need to give him a treat for coming only some of the time, or maybe sometimes a higher-value treat like a fish crunch and other times a boring treat like a Milkbone biscuit. The unpredictability of the reward is what makes the behavior stronger. If you give someone a gift card as a thank-you every time they do you a favor, they come to expect it -- the gift card is now fair market value for services rendered. But if you reward them only occasionally they actually will be much more likely to help you in the future. This is in part because they are motivated to keep trying in order to finally get a reward, but also in part because the reward component is less obvious and they feel more ownership for their own actions.

- Build up new behaviors by rewarding small steps (shaping). Another major strategy to get new behaviors started is to break them into small steps, and then offer rewards whenever any of those steps is made. This strategy is familiar to many people who work with children, or coach students, or try to motivate ourselves to complete items on a to-do list. What may be less familiar is the great success that Skinner had when using this method with animals: He was expert at looking for small steps that might add up to a larger pattern of behavior, even if those steps had nothing to do initially with that desired behavior. For example, he taught pigeons to play ping-pong using the method of shaping. If you watch the video, you can see that the pigeons actually seem to understand the rules of the game, that they will receive a food reward only if they manage to score a point by getting the ball past their opponent. Skinner developed this game-playing behavior gradually, by first rewarding the pigeons when they just approached the ball, then only when they pecked at it, then when it moved in response to a peck, and finally rewarding them only when they actually managed to score a goal against their opponent. Did the pigeons understand that they are playing ping-pong? Skinner's answer would be "it doesn't matter" -- if the desired behavior is achieved, the shaping method was a success.

As a final note, one of TMT's central ideas is the idea that behavior is produced by the Intuitive System rather than by way of rational, language-based choices made by the Narrative System. However, this does not in our view mean that behavior is not “free.” In other words, acceptance of TMT does not require a philosophical position of determinism. The success of Skinner's methods is indisputable as a behavior-change technology, but we do not believe that accepting these empirical results means we must automatically move "beyond freedom and dignity" in our understanding of human beings. We suggest only that any choice that occurs before a behavior is made at the Intuitive level, and not by the conscious mind. These ideas deserve further explanation, so watch for more in a future blog.