11-1: What is Observational Learning?

Psychology of Learning

Module 11: Observational Learning 1

Part 1: What is Observational Learning?

Looking Back

Throughout this course, we’ve explored how organisms learn through direct experience with the environment. Classical conditioning, operant conditioning, & motor learning all share a common feature: they require direct personal experience. This requirement creates significant limitations—learning is slow, potentially dangerous, & inefficient. But humans & many animals possess a remarkable ability that transcends these limitations: learning from others through observation.

Introduction: Learning a Song by Watching

One day a five-year-old girl was trying to figure out how to play a recorder using sheet music. Unfortunately, since she was only five, she did not know how to read music. After a few minutes of frustration, she gave up. Later that day, the girl’s father sat down with his own recorder & began to play “Mary Had a Little Lamb” while following the same sheet music. The five-year-old watched intently as her father played, observing his finger positions & how they related to the symbols on the page. After her father finished, the girl picked up her own recorder & successfully played the song, correctly fingering the notes despite never having received explicit instruction.

This simple episode illustrates the power of observational learning. The girl learned a complex skill—translating visual symbols into finger movements producing specific pitches—not through personal trial & error, but by watching someone else perform the behavior. She didn’t need reinforcement for correct notes or punishment for incorrect ones. Observation alone enabled learning that would have required extensive trial & error through other means.

What Is Observational Learning?

Observational learning (also called vicarious learning or social learning) is a change in behavior that occurs as a result of observing a behavior & its consequences. The learner acquires new behaviors or modifies existing behaviors by watching others rather than through direct personal experience (Bandura, 1977). Related terms include modeling (the process by which observers pattern behavior after a model) & imitation (copying specific behaviors observed in others). These terms are often used interchangeably, though they carry slightly different emphases.

Putting Observational Learning into Perspective

Simple animals rely on reflexes & instincts to adapt to their environments. A moth flies toward light sources through phototaxis—an innate, automatic response. These mechanisms work well if the environment does not change or changes predictably. But complex, unpredictable environments require more flexible adaptation. More complex processes allow for adaptation to complex & changing environments. Habituation enables organisms to ignore repeated innocuous stimuli. Classical conditioning allows neutral environmental stimuli to predict biologically significant events. Operant conditioning enables organisms to discover which voluntary behaviors produce favorable outcomes.

Observational learning represents a qualitative leap in adaptive capacity. Instead of learning only from personal experience, organisms can learn from others’ experiences. This dramatically expands the information available for learning—organisms benefit from the accumulated knowledge of their social group, not just their individual history. Observational learning allows us to benefit from experiences we’ve never witnessed or from people we’ve never met, through cultural transmission of knowledge across space & time.

Ways We Modify Behavior to Fit Environmental Demands

Consider the hierarchy of behavioral adaptation mechanisms: Instincts (useful for unchanging environments) include reflexes, kineses & taxes, & fixed action patterns. Learning (allows behavior change as environment changes) includes habituation, classical conditioning, operant conditioning, & social learning. Beyond social learning comes communication, language, & writing—enabling us to learn from people we’ve never met or who died centuries ago. As evidence of our species’ unparalleled adaptability through these mechanisms, there is no other species that occupies as many different places & climates on earth as Homo sapiens.

Observational Learning Defined

Learning is an inferred change in the organism’s mental state which results from experience & which influences in a relatively permanent fashion the organism’s potential for subsequent adaptive behavior. In observational learning, this change occurs after viewing others rather than actually experiencing the reinforcers & punishers for ourselves.

In the simplest forms of observational learning involving social facilitation & stimulus (local) enhancement, no complex cognitive mechanisms need to be employed to explain the behavioral change. Social facilitation (also called contagion) is a situation in which the observation of another organism’s behavior increases the likelihood that the same behavior, which is already in the current repertoire of behavior, will be exhibited. If a deer hears a snapping twig & runs, other deer will run without knowing why—they already know how to run, & seeing another deer run triggers the same behavior.

Stimulus (local) enhancement is a type of observational learning in which there is an increased likelihood of behavior focused on a particular point or object following the observation of another organism interacting with that point or object. When other animals observe a demonstrator interacting with an object, they become more likely to approach & interact with that same object. In imitation, the behavior change occurs because an animal sees & copies the behavior of others, but does not necessarily understand about the consequences of the behavior. In true social learning, the change in behavior is due to the viewing of the consequences to another with the expectancy that if the learner performs the new behavior the same consequences will occur to them as well.

Thorndike’s Failure to Find Social Learning

Edward Thorndike, the father of operant conditioning, attempted to demonstrate social learning in cats, dogs, chickens, & monkeys—and failed in every case. In his procedure, Thorndike first taught a cat to escape from a puzzle box by performing a specific response (pulling a loop of string or pressing a lever). He then placed a naive cat in an adjacent observation cage where it could watch the trained cat escape repeatedly. After extensive observation, the naive cat was placed in the puzzle box. Despite having watched the solution performed many times, naive cats showed no faster escape than cats who had never observed a demonstration. Thorndike concluded that imitation did not exist in animals & that what appeared to be social learning was actually individual learning through trial & error.

However, Thorndike’s conclusion was premature. His experimental design may have been poorly suited to revealing social learning. The puzzle box apparatus was complex, the relevant behaviors were difficult to observe clearly, & cats are relatively solitary animals that may not be predisposed to learn from conspecifics. Later research with more appropriate species, simpler tasks, & better observational conditions demonstrated robust social learning in animals. Thorndike’s failure revealed more about the limitations of his methodology than about the absence of observational learning in animals.

Noncognitive Forms of Observational Learning: Frida the Octopus

An octopus in a German zoo learned to open jars of shrimp by watching zoo attendants perform the act underwater. Frida, a five-month-old female octopus, opened the jars by pressing her body on the lid & grasping the sides with the suckers on her eight tentacles. With a succession of body twists she unscrewed the lid. This seems like remarkable observational learning in an invertebrate. But before jumping to conclusions, we should consider a more parsimonious explanation.

In the wild, octopi open shellfish by pressing their body against the top, grasping the bottom with their tentacles, then twisting. This sounds remarkably like how Frida opened jars of shrimp. Perhaps this was not true observational learning but rather stimulus enhancement—the attendants drew Frida’s attention to the jars containing food, & she then applied behaviors already in her repertoire. This illustrates Morgan’s Canon: “In no case is an animal activity to be interpreted in terms of higher psychological processes, if it can be fairly interpreted in terms of processes which stand lower in the scale of psychological evolution & development.” We must be careful not to attribute complex cognitive abilities when simpler mechanisms suffice.

Imitation: Bridging the Gap with Cognition

In psychology’s early days, imitation was seen as a very simple thing. In fact, “to ape” meant to imitate something blindly with no cognitive activity involved. In more recent years, however, it has been realized that true imitation is actually a fairly complex skill which may require a theory of mind—the ability to attribute thoughts & feelings to oneself & others, & to understand that others’ thoughts & desires may differ from one’s own. For true imitation to occur, we must understand the other organism as a separate entity from ourselves which thinks & acts as we do. “Some researchers have even argued that true imitation requires that the observer recognize the intentional structure of the actions of the demonstrator” (Zentall, 2006, p. 342).

Imitation as Instinct

More than a century ago, McDougall (1908) noticed that his 4-month-old son would imitate certain facial gestures, particularly tongue protrusion. Meltzoff & Moore (1977) demonstrated four gestures (lip protrusion, mouth opening, tongue protrusion, & sequential finger movement) to 12- & 21-day-old infants whose parents reported that they had never demonstrated these expressions. Even these very young infants imitated readily, with tongue protrusion being the best imitated expression. This is too early for learned imitation—it appears to be an innate capacity, suggesting that humans are biologically prepared for observational learning from birth.

Mice & Joystick Pushing

Collins (1988) performed a study in which mice demonstrated observational learning that probably qualifies as imitation. A demonstrator mouse pushed a joystick for food while other mice observed. Mice watching the joystick pushed to the left for food subsequently pushed the joystick significantly more to the left. These results were replicated by Heyes & Dawson (1990) using rats. Invoking Morgan’s Canon, it seems doubtful that the mice & rats “knew” just by watching that pushing the joystick to the left led to reinforcement. More likely, they were pushed by instinctual drives to imitate other animals, rather than being pulled by cognitive understanding of the behavior-consequence relationship. Nevertheless, the observation followed by imitation clearly sped up the rate of acquisition.

Birds & Milk Drinking

In the early 1900s, milk delivery customers in England experienced a peculiar rash of “thefts”—birds were obtaining milk by piercing the bottle caps with their beaks & drinking the cream that rose to the top. The phenomenon started locally but spread geographically over several decades, eventually involving multiple species of birds (Fisher & Hinde, 1949). This seems to be an example of cultural transmission through observational learning, where one bird’s innovative behavior spread through populations as others observed & imitated the successful foraging technique.

Deferred Imitation

If an animal views a model’s behavior & then reproduces the behavior after some time has passed, the animal has demonstrated deferred imitation. Immediately displayed imitation can often be explained using simpler mechanisms such as fixed action patterns or stimulus enhancement. However, if imitation is deferred, then the behavior seems to require a much more complicated explanation that must invoke memory, representation, & perhaps theory of mind. Deferred imitation represents a cognitively higher form than immediate imitation.

Dorrance & Zentall (2001) had a Japanese quail model a behavior of treadle stepping. An observer quail was given the opportunity to imitate the behavior either immediately or after 30 minutes. Regardless of whether immediate or deferred, the quail imitated the treadle stepping behavior. Clearly, when a delay was added before the quail had an opportunity to imitate, any possible explanation for the behavior which invokes reflexes must be dismissed. Due to Morgan’s Canon, we should be hesitant to assume higher mental processes in quail, but the evidence points in this direction.

Factors Affecting the Probability a Model Will Be Imitated

Mischel (1971) describes the factors of rewardingness, dominance, similarity, & sincerity as determining which models will be imitated. The existence of these factors implies that the observer can calculate some probability of being reinforced, & adjust that probability based on these factors. A model who provides rewards to the observer will be more imitated. A model who holds power or dominance will be more imitated. A model who is similar to the observer in age, gender, or interests will be more imitated. A model who appears sincere will be more imitated. These selective imitation patterns have been demonstrated in humans and, to a lesser degree, in other animals as well. Capuchin monkeys, for example, spend more time observing proficient tool users, suggesting they can identify & preferentially attend to skilled models (Ottoni, de Resende, & Izar, 2005).

Looking Forward

Part 2 examines observational learning in animals in greater depth, exploring Bennett Galef’s groundbreaking research on social transmission of food preferences, mate choice copying in quail, & the effects of environmental stimulation on social learning. These studies reveal that observational learning is widespread across the animal kingdom & serves critical adaptive functions.