“When in Rome, do as the Romans do.” For the traveler, it’s valuable advice. For the student of learning, it’s something to ponder. What does happen in an unfamiliar environment where you don’t know the local customs—in other words, an environment where you don’t know which behaviors are reinforced and which are punished? How exactly do you learn how to do as the Romans do?

Operant conditioning provides one answer. From this perspective, when entering an unfamiliar cultural setting, you’re like a rat on its first day in a Skinner box. The rat, at first, hasn’t the foggiest idea how to behave. Its initial actions are essentially random. It learns how to act only due to a slow process of trial-and-error learning in which reinforcements or punishments gradually shape its behavior.

Is this what it’s like for you when you enter a novel environment—for example, if you participate in a student exchange program that sends you to a foreign culture? Do you behave randomly and await reinforcements and punishments—smiles from strangers when you do something right, and angry stares when you do something insulting—to shape your behavior? Or do you learn in some other way?

Most people learn to “do as the Romans” by watching the Romans. In unfamiliar settings, we often observe how others act, gain information about appropriate and inappropriate types of behavior from our observations, and then base our behavior on what we observed. This human ability to learn by observation enables us, quickly and easily, to fit in with the group rather than stick out from the crowd.

297

Curiously, this human capacity for learning by observation received little attention from psychologists studying operant conditioning. The psychologist Albert Bandura, however, paid it a lot of attention and pioneered research on observational learning. (Bandura used his findings about observational learning as a foundation for his theory of personality; see Chapter 13.)

In observational learning, people acquire new knowledge and skills merely by observing the behavior of other people (Bandura, 1965; 1986). Those other people are called psychological models, and observational learning thus is also referred to as modeling. In practice, the terms observational learning and modeling are used interchangeably.

LEARNING WITHOUT REINFORCEMENT. Observational learning is efficient. Through observational learning, people can acquire complex skills all at once, without tedious trial-and-error shaping processes.

Bandura (1965) documented the power of observational learning in research (also discussed in Chapter 4). He showed children a short film in which an adult acted aggressively toward an inflated clown doll, or “Bobo doll.” The adult performed specific behaviors (such as hitting the doll in the nose with a hammer) that the children were unlikely to have performed before. Later, when the children were observed in a playroom, Bandura found that the children spontaneously performed many of the same behaviors they had seen in the film (see photos).

Note that the children performed these behaviors even though they never had been reinforced for performing them. There was no trial-and-error learning in which the children’s behavior gradually was shaped by its consequences, like that of a rat in a Skinner box. Instead, instantaneous learning occurred: Children correctly performed the behavior—they hit the clown doll right in the nose with the hammer—the very first time they had a chance. They learned this behavior merely by watching it on film. As Bandura emphasized, this result shows that people have a capacity to learn that is not explained by the principles of operant conditioning. Observational learning is thus a third type of learning, distinct from operant and classical conditioning.

SUBPROCESSES IN OBSERVATIONAL LEARNING. Bandura recognized that his experiment had revealed a distinct form of learning that had been largely overlooked by operant conditioning researchers: To a much greater extent than the rats and pigeons in a Skinner lab, humans have the ability to learn complex acts through observation. To explain this ability, Bandura (1986) analyzed subprocesses in observational learning—in other words, the psychological processes that are required for a person to get from the performance of a behavior by a model to the performance of a similar behavior by an observer. He identified four subprocesses (Figure 7.16):

298

The fourth step in Bandura’s analysis, motivational processes, underscores a point that received relatively little attention from Skinner. It is the distinction between acquisition and performance. To perform a behavior is to engage in that behavior. To acquire a behavior is to have the potential to engage in a behavior. As Bandura emphasized, people can acquire the skill to perform a behavior, yet not be motivated to do it. By watching a crime show on TV, you might learn ways to commit a criminal act. Nonetheless, you likely choose (wisely) not to perform those behaviors. Learning, then, cannot be measured solely by studying what individuals do because they may have learned actions that they choose not to perform.

In summary, research on observational learning substantially expanded psychology’s understanding of learning in two main ways. First, it highlighted a phenomenon that had received little attention from previous learning theorists (i.e., Pavlov, Thorndike, Skinner): people’s ability to learn behaviors rapidly through observation alone. Second, it explained that phenomenon through a novel set of conceptual tools. Rather than referring merely to simple stimulus–response mechanisms, Bandura explained human learning by referring to complex processes carried out in the mind: attention to models, retention of memory of their behavior, and the use of that memory to guide behavior.

299

Sometimes models send contradictory messages. They model one type of behavior, but tell others that this behavior is not allowed. An example of this is spanking by parents who are trying to control their children’s unruly, aggressive behavior. The spanking is intended as a punishment. An operant conditioning analysis would thus suggest that the spanking would make children less unruly. Yet the parent who is spanking is providing a role model of aggressive behavior; the spanking itself is an act of physical aggression toward the child.

What is the effect of spanking? To find out, researchers conducted a large-scale, long-term study with more than 2000 families (Taylor et al., 2010). When children were 3 years of age, researchers asked mothers to report whether, and how frequently, they spanked their child. Parents differed: About 45% reported no use of spanking; about one-fourth said they spanked one or two times a month; and one-fourth reported a higher spanking frequency. The researchers also measured the children’s level of aggression, specifically, whether each child was defiant, had angry moods, and hit others. Then, two years later, mothers reported their 5-year-old child’s current level of aggressiveness, measured in terms of arguing, bullying, meanness to others, and fighting. With the age 3 and age 5 measures, the researchers could relate earlier spanking to later levels of aggressiveness.

The results suggest that spanking increases children’s aggressiveness (Taylor et al., 2010). Mothers who spanked their children more frequently at age 3 produced children who, at age 5, were more aggressive. This surely wasn’t their goal; mothers who spanked were trying to establish a strict household that produced children who were well behaved. But their efforts backfired. Note that spanking predicted age 5 aggressiveness even after the effects of age 3 aggressiveness were accounted for statistically. The results, then, could not be explained in terms of the child’s personality trait of aggressiveness alone. Instead, they strongly suggest that parental behavior affected the child. Psychological modeling (children saw their parents acting aggressively) was more powerful, in this case, than operant conditioning via response consequences (being spanked for unruly behavior)—just as Bandura would have predicted.

300

People are exposed to a wider range of psychological models in contemporary society than they were in the past. Centuries ago, people in rural villages learned social behaviors and professional skills by observing others who lived in their town (Braudel, 1992). By the mid-twentieth century, most people in rich countries observed people who lived all over the world, thanks to TV. Today, the Internet provides access to an unending stream of expert models on instructional videos. At any time of day or night, you can see a model changing a spark plug, riding a skateboard, or making a cheese soufflé.

Thanks to contemporary digital technology, researchers can overcome a problem that, in the past, limited the effects of psychological models—that expert models sometimes have minimal impact on individuals because the expert seems to differ so much from them. “Sure, Tony Hawk can skateboard,” you might think, after watching an instructional video, “But that’s because he’s Tony Hawk. I could never do it.” How can psychologists provide people with psychological models who seem similar to them?

One way to do this is through virtual representations of the self (VRSs; Fox & Bailenson, 2009). A VRS is a modeling display in which people observe an image that looks just like them. Here’s how it’s done. If you participate in a VRS study, psychologists first take a digital photo of you. They then enter the image into computer software that creates a three-dimensional visual image that looks like you. This image is displayed within a larger computer-generated environment. The image, then, is a “virtual” self. You watch this virtual self expertly engaging in actions within the computer-generated environment. This solves the problem; the expert actions are modeled by someone who looks highly similar to you: your virtual self.

In one study (Fox & Bailenson, 2009), VRSs were used to help people to improve their health by getting more exercise. Participants were assigned to one of three modeling conditions. In one, they watched a computer-based display showing someone else running on a treadmill for 5 minutes. In another condition, participants saw a virtual representation of themselves running on the treadmill for 5 minutes. In a third condition, which controlled for exposure to a VRS, participants saw a VRS just standing around for 5 minutes. (The second and third conditions are displayed in Figures 7.17a and 7.17b.) The experimenters measured the amount of exercise participants did in a 24-hour period subsequent to these video displays. They found that people who saw their VRS exercising engaged in more exercise themselves—spending more time running, playing sports, or working out in gyms (Fox & Bailenson, 2009).

301

People’s ability to learn by observing others’ behavior can be explained not only at a psychological level of analysis, but also at a biological level. Our brains contain mirror neurons, which are cells in the brain that fire when an organism takes action and also when it observes another organism take that action (Rizzolatti, Fogassi, & Gallese, 2001). Their firing contributes to the imitation of the other organism’s behavior.

Mirror neurons are located in the brain’s motor cortex, which contains neural systems that control the movement of the body (see Chapter 3). Whenever you make a movement—for example, when you reach out to pick up a can of soda—neurons in the motor cortex become active and send signals to the relevant body parts. Surprisingly, some of these same motor neurons—the mirror neurons—also become active merely when you observe someone else make that same movement. Their activity provides the brain with an automatic “copying” mechanism (Iacoboni et al., 1999). When you observe someone else perform a motor movement, mirror neurons automatically prepare your body to make that same movement.

Mirror neurons originally were discovered in research with monkeys (Rizzolatti et al., 2001). Experimenters recorded neural activity in monkeys’ brains during two activities: (1) when they themselves picked up a piece of food, and (2) when they observed someone else (either a human or another monkey) pick up a piece of food. The activities—taking action oneself and observing someone else perform that action—are quite different. Yet, remarkably, researchers discovered that a set of neurons—the mirror neurons—fire in the same manner during both activities (Figure 7.18).

Importantly, these mirror neurons did not fire when the monkeys merely observed a piece of food, or observed a hand making a grasping movement when there was no food in sight, or when they observed a piece of food being picked up with pliers instead of by hand. The mirror neurons were highly selective. They fired only when the monkey observed the action of a hand picking up the food. This shows that mirror neurons are highly specialized. Each mirror neuron is designed to mirror a particular motor movement (in this case, the picking up of an object with one’s hand).

Further research revealed mirror neurons at work in humans (Iacoboni et al., 2009). Researchers observed people’s brain activity while they performed a simple task: lifting their index finger. They told participants exactly when to lift their finger by showing either of two displays on a video screen: (1) a hand in motion, lifting one of its fingers, or (2) a symbol (a letter X) that appeared on a static picture of a hand (Figure 7.19).

Although the task that participants performed in the two conditions was the same (raising a finger), brain activity in the two conditions differed. Brain activity was higher when participants moved their finger while observing a moving hand than when they looked at a static picture of a hand (Iacoboni et al., 2009). Observing a motor movement by someone else (the hand lifting one of its fingers) automatically activated the participants’ mirror neuron system. This result thus showed that there is a system in the brain, the mirror neuron system, specifically dedicated to copying observed actions.

302

Mirror neurons may explain a variety of common, and sometimes puzzling, behaviors in animals and people. Why do birds instantly fly away when observing one bird taking flight? Why do people yawn when they observe someone else yawn? In these and innumerable other cases, observing an action may automatically activate neurons that mirror the action being observed.