Positive and Negative Reinforcement in Context: A Look Inside Relational Frame Theory
Relational frame theory (RFT) posits that human actions never take place in a vacuum. There is always something preceding and something following each action. It is among these contextual factors—those that precede and those that follow—that the behavioral analyst looks for answers to questions about what governs behavior. If someone, in a certain context, turns his eyes to me with a specific expression on his face, I might address him by saying something like “Can I help you?” My utterance is followed by a new occurrence: The person replies. So my action is followed by a consequence, in this case that someone answers me. The core principle in operant conditioning is that the consequences following a behavior (a response) influence the probability of the behavior being repeated. Let’s speculate a bit using two rather different consequences in this everyday example.
Behavior in Context: An Example
Imagine that what follows upon my utterance “Can I help you?” is that the person gives me a friendly smile and tells me what he wants. An alternative would be the response “Mind your own business, you jerk!” It is hard to say exactly how each of these consequences might influence the probability of me asking if I could help if someone else were to look at me with this specific facial expression in the future.
What is essential is that earlier consequences do have an influence. “Once bitten, twice shy” is a well-known saying. It is not always easy to determine what kind of behavior to expect based on previous consequences—but previous consequences do have an influence. This is the core of operant learning. In operant psychology, the different influences of consequences are categorized based on whether they increase or reduce the probability of an earlier behavior being repeated. If I get friendly responses to utterances like the one above and thereafter more often address people who exhibit the same type of facial expression in similar situations, the friendly responses would be said to have had a reinforcing effect on this specific behavior of mine.
A consequence that increases the probability of the preceding behavior being repeated is thus termed reinforcing. In this case, the reinforcing consequence is that I receive something: a friendly response. Something is added. This kind of process is called positive reinforcement. A behavior can also be reinforced through a consequence consisting of something being removed. This is illustrated by the behavior of the person who responded to my question by saying, “Mind your own business, you jerk!” Let’s assume that when this is uttered, I become quiet and turn my attention away from the speaker. I do what I am told to do. This consequence could increase the probability of the other person repeating the utterance “Mind your own business” in similar situations in the future. The consequence— that I became quiet and turned away—has in this case become reinforcing to this person’s behavior of telling off jerks. This time, though, the reinforcement consists of something being removed, namely, the attention from a jerk. When a behavior increases because something is removed, it is termed negative reinforcement.
Distinguishing Between Positive and Negative Reinforcement
Distinguishing between positive and negative reinforcement (which are both processes that increase the probability of a certain behavior) is not always essential. These two concepts can be said to describe two different sides of the same thing (Michael, 1975). If the behavior of telling off jerks was reinforced by me becoming quiet, then one factor is that my annoying questions came to an end. This is negative reinforcement. But another way of describing the same thing is to note the condition that resulted: silence, for example, or anything else that was added. That would be positive reinforcement. It is often convenient to distinguish between positive and negative reinforcement, even though the difference may not be clear-cut from a theoretical perspective. It may sometimes be more obvious that something is removed than that something is added. By speaking about this as negative reinforcement, we clarify the process. And the distinction is often practical in clinical situations.
When a consequence reduces the probability of a certain behavior being repeated, it is known as punishment. If I receive an unfriendly response in the example above and I subsequently refrain from addressing people in that specific social context or do so less often, then the earlier consequence has been punishing. Punishment, too, can be separated into positive punishment—in which something has been added—and negative punishment—in which something has been removed. Remember, however, that there is no way of determining what is reinforcing versus punishing based on any intrinsic quality that signifies the consequence as such.
Some consequences more often function as reinforcing to people, for example, certain types of social attention. But this is not always the case. To be recognized and addressed in a kind way is usually reinforcing for human behavior, but we can all think of situations when this is something we want to avoid. Likewise, certain consequences usually function in a punishing way, like being hit, yet this is not always the case. There are situations when being hit reinforces the behavior that preceded this consequence. A child who encounters only indifference, despite several actions meant to attract attention, may repeat a behavior that leads to getting smacked, simply because the smack involves attention. It is the function of the consequence that provides the definition. When the consequences increase the probability of a certain behavior, this is reinforcement; and when the consequences reduce the probability of a behavior, it is punishment.
Learn more about RFT and the roots of ACT at this invaluable two-day training.
This article has been adapted from Learning RFT by Niklas Törneke, MD.