Learning by thinking and watching: Tolman and Bandura

Transcript

RyanLeila is eight. For three years she's ridden in the back seat to her Thursday piano lesson, mostly looking out the window. One day her dad, distracted, misses a turn. And from the back seat Leila says, Dad, it was left back there.

SoheeNobody ever taught her the route. Nobody rewarded her for knowing it. And yet she knew it.

RyanWelcome to EPPP Prep, the podcast. I'm Ryan.

SoheeI'm Sohee. Today we go beyond pure conditioning. Two researchers, Edward Tolman and Albert Bandura, showed that thinking and watching matter in learning. And both ideas change how you read a learning scenario.

RyanSome context first. Classical and operant conditioning are part of a tradition called behaviorism. The OpenStax psychology textbook notes that its more extreme adherents once insisted behaviors are triggered by the environment with no intervening thought.

SoheeSkinner was the clearest case. He believed cognition didn't matter so firmly that his ideas were called radical behaviorism. He considered the mind a black box, unknowable, and so not something to study.

RyanTolman was also a behaviorist, but he saw it differently. At the time, the prevailing idea was that reinforcement had to be immediate for learning to happen. Tolman's rat experiments showed organisms can learn even without immediate reinforcement.

SoheeHere's the design. Hungry rats were placed in a maze with no reward for finding their way through. A comparison group got food at the end of the maze every time.

RyanAs the unrewarded rats explored, they developed what Tolman called a cognitive map, a mental picture of the maze's layout. After ten sessions without reinforcement, food was put in a goal box at the end.

SoheeAnd as soon as those rats knew about the food, they ran the maze just as quickly as the comparison group that had been rewarded all along.

RyanThat's latent learning: learning that occurs but isn't observable in behavior until there's a reason to show it. Which is exactly Leila in the back seat. She'd built a cognitive map of the route, and it only showed when it was needed.

SoheeThe textbook has a nearly identical example. Ravi's dad drives him to school every day. One morning his dad has to leave early, and Ravi rides the same route on his bike. He'd learned it, but had no need to show it before.

RyanSo the big lesson from Tolman. Learning and performance aren't the same thing. Something can be learned without any reinforcement, and reinforcement may only decide when it's shown.

SoheeThat distinction comes right back with Bandura. Let's set him up. In observational learning, we learn by watching others and then imitating, or modeling, what they do or say. The people performing the behavior are called models.

RyanThe textbook frames it as extending the reach of conditioning. In classical and operant conditioning, learning happens only through direct experience. In observational learning, you watch someone else.

SoheeAnd it isn't only humans. In one study, a group of chimpanzees dipped straws into juice boxes and sucked the little bit of juice off the end. Another group sucked straight through the straw and got much more. After watching, all the dippers switched to sucking directly.

RyanResearch also suggests this imitative learning involves a specific type of neuron, called a mirror neuron.

SoheeNow Bandura. Like Tolman, his ideas differed from strict behaviorists. He and other researchers proposed social learning theory, which took cognitive processes into account.

RyanBandura's argument: pure behaviorism couldn't explain why learning can happen without external reinforcement. Internal mental states must play a role. And observational learning is more than imitation. In imitation, a person simply copies the model. Observational learning is much more complex.

SoheeHow so? The textbook lists several ways observational learning can happen. You can learn a new response. You can choose whether or not to imitate, depending on what happened to the model. And you can learn a general rule you apply to other situations.

RyanThen the three kinds of models Bandura identified. A live model shows the behavior in person, like a yoga instructor demonstrating a pose.

SoheeA verbal instructional model doesn't perform the behavior, but explains or describes it. Like a soccer coach telling young players to kick with the side of the foot, not the toe.

RyanAnd a symbolic model is a fictional character or real person showing behavior in books, movies, television, video games or online. The model doesn't have to be present.

SoheePause and think. A child learns a dance move by watching a cartoon character do it on a streaming show. What kind of model is that?

RyanA symbolic model. A character on a screen, not a person in the room, and not someone explaining in words.

SoheeNow the heart of it, the steps Bandura said must happen for modeling to succeed. Four of them: attention, retention, reproduction and motivation.

RyanAttention: you have to be focused on what the model is doing. Retention: you have to remember what you saw. Reproduction: you have to be able to actually perform the behavior.

SoheeAnd motivation: you have to want to copy it. Whether you do depends on what happened to the model. If the model was reinforced, you're more motivated to copy them. That's vicarious reinforcement.

RyanAnd if the model was punished, you're less motivated. That's vicarious punishment. The textbook's example: four-year-old Allison watches her older sister play in their mother's makeup and get a time-out. Allison is tempted, but she doesn't want a time-out either.

SoheeNotice how this fits Tolman. Allison may well have learned how to play with the makeup. Motivation decides whether she does it. Learning and performance, separated again.

RyanAnd once you actually do the new behavior, the reinforcement you receive yourself plays a part in whether you repeat it.

SoheeNow the famous study. Bandura, Ross and Ross, 1961. A five-foot inflatable doll, called a Bobo doll. A child watched a teacher act aggressively with the doll, hitting, throwing, even punching it.

RyanAnd the children's behavior depended on what happened to the teacher. When the teacher was punished, the children decreased their tendency to act as she had.

SoheeWhen the teacher was praised or ignored, not punished, the children imitated what she did, and even what she said. They punched, kicked and yelled at the doll.

RyanSo Bandura concluded that we watch and learn, and that this learning can have prosocial and antisocial effects.

SoheeOn the prosocial side, the textbook's advice to parents is direct. If you want your children to read, read to them, and let them see you reading. Children tend to copy what you do, not what you say. So toss out the old line, do as I say, not as I do.

RyanOn the antisocial side, children who grow up watching parents handle anger with violence often learn to behave that way themselves. The textbook's example is a mother who hit her son with a belt and later saw her four-year-old daughter whipping a teddy bear.

SoheeAnd media violence is still debated. Craig Anderson and colleagues found extensive evidence suggesting a causal link between violent media exposure and aggressive thoughts and behaviors. Christopher Ferguson and others argue research hasn't accounted for other risk factors, like mental health and family life.

RyanSo how might this be tested? Expect scenarios where someone learns without doing anything or getting any reward. If the learning is hidden until it's needed, think latent learning, Tolman. If it comes from watching a model, think observational learning, Bandura.

SoheeAnd the tempting wrong answer is often plain operant conditioning. Ask: did this person's own behavior get reinforced? If the reinforcement went to someone else, the model, that's vicarious reinforcement.

RyanOne more pause and think. A new employee sees a coworker praised by the manager for speaking up in meetings. The next week, the new employee starts speaking up too. What's the term, and which step of modeling did the praise affect?

SoheeVicarious reinforcement. The model was reinforced, and that boosted the newcomer's motivation, the fourth step.

RyanLet's recap. Strict behaviorists, Skinner most of all, treated the mind as a black box. Tolman showed rats learn a maze without reinforcement, building a cognitive map. That's latent learning, hidden until there's a reason to show it.

SoheeBandura's social learning theory added cognition to learning. We learn by watching models: live, verbal instructional, or symbolic.

RyanModeling takes attention, retention, reproduction and motivation. Seeing the model reinforced, vicarious reinforcement, raises motivation. Seeing the model punished, vicarious punishment, lowers it.

SoheeAnd the Bobo doll showed children imitate aggression when the model isn't punished, which is why models matter for good and for harm.

RyanOne last thought. In Bandura's terms, we're verbal instructional models right now, describing instead of demonstrating. And you're doing the attention and retention part. Give yourself credit for showing up.

SoheeTry a few practice questions when you're ready, and we'll see you next time on EPPP Prep.

Loading the read-along timings…