Psychology of Learning

Module 01: What is Learning?

Part 2: Biological Constraints on Learning

Looking Back

In Part 1, we established learning as a hypothetical construct—an inferred change in an organism’s mental state that results from experience and influences what the organism can do. We distinguished learning from maturation, reflexes, and temporary states, and explored the critical learning-performance distinction. Now we’re ready to examine whether any organism can learn any behavior—a question that challenges the early behaviorist principle of equipotentiality.

The Challenge to Equipotentiality

For much of the early 20th century, behaviorists dominated the study of learning. Behaviorism is the approach to psychology that focuses exclusively on observable behaviors and their environmental causes, deliberately avoiding discussion of mental states or consciousness (Watson, 1913; Skinner, 1938). Behaviorists believed in equipotentiality—the idea that learning does not vary as a function of species, situation, or response being learned (Pavlov, 1927; Thorndike, 1911).

This belief in equipotentiality had important implications. It meant that researchers could study learning in rats, pigeons, or dogs and expect their findings to apply equally well to humans. It suggested that the same principles governed all learning, whether an organism was learning to press a lever, navigate a maze, or speak a language. It implied that biology placed few, if any, constraints on what could be learned.

However, by the 1960s and 1970s, accumulating evidence suggested that this view was too simplistic. Martin Seligman (1970, 1971) wrote influential papers arguing that organisms are not equally prepared to learn all associations. Instead, evolutionary history has made some things easy to learn, others moderately difficult, and still others nearly impossible.

Seligman’s Three Categories of Learning

Seligman (1970) proposed that learned behaviors fall into three categories based on how biological predispositions affect learning: prepared behaviors, unprepared behaviors, and contraprepared behaviors. This framework, sometimes called the preparedness continuum, recognizes that evolution has shaped not just what organisms do instinctively, but also what they can readily learn.

Prepared Behaviors

Prepared behaviors are those that organisms learn so easily and quickly that they almost appear instinctive. These behaviors typically require minimal training, few repetitions, and little effort to acquire (Seligman, 1970). Often, prepared behaviors are vital to an organism’s survival or reproductive success, which explains why evolution has made them so easy to learn.

Birds learn to fly with remarkable ease. While some maturation of muscles and neural systems is necessary, young birds quickly master the complex motor coordination required for flight with relatively little practice. Similarly, songbirds learn their species-specific songs remarkably easily—but only if they are exposed to those songs during a specific developmental window (Marler, 1970).

Critical Periods & Prepared Learning

Many prepared behaviors can only be learned during a critical period—a specific time window in development when the organism is especially receptive to learning that behavior (Lorenz, 1935). If the learning opportunity is missed during this window, the behavior may be impossible or extremely difficult to acquire later.

The concept of critical periods was famously demonstrated in Konrad Lorenz’s work on imprinting in birds. Imprinting is a rapid form of learning in which young animals form a strong attachment to the first moving object they see, typically their mother (Lorenz, 1935; Hess, 1959). In goslings and ducklings, this occurs within hours after hatching, with peak sensitivity around 13-16 hours after hatching.

Lorenz famously demonstrated imprinting by being the first moving object seen by newly hatched goslings. The goslings imprinted on him and followed him everywhere, treating him as their mother. This wasn’t just cute—it revealed something profound about how evolution has prepared young birds to quickly learn who to follow and trust.

Hess (1959) conducted systematic experiments showing that the strength of imprinting depends critically on timing. Ducklings exposed to a moving object at 13-16 hours after hatching showed the strongest imprinting. Earlier or later exposure resulted in weaker or no imprinting. After about 30 hours, the critical period closes and imprinting becomes nearly impossible.

Interestingly, imprinting has been used practically with endangered species. When breeding programs raise endangered crane chicks, handlers wear costumes that resemble cranes and use crane-shaped puppets to feed the chicks. This ensures the birds imprint on their own species rather than on humans, which is essential for their eventual release into the wild (Ellis, Gee, & Mirande, 1996).

Prepared Behaviors in Humans

Do humans have prepared behaviors? While the evidence is more complex than in other species, several strong candidates exist.

Attachment behavior—the formation of strong emotional bonds between infants and caregivers—appears to be a prepared behavior in humans (Bowlby, 1969). John Bowlby argued that attachment evolved because it kept infants close to protective adults, increasing survival. Human infants seem biologically prepared to form attachments, showing distress when separated from caregivers and using caregivers as a secure base for exploration.

Bowlby (1969) suggested there might be a sensitive period for attachment formation, roughly in the first few years of life. While the evidence for a strict critical period in humans is debated, studies of children raised in severely deprived institutional settings show that early attachment formation is important for later social and emotional development (Rutter et al., 2007).

Language acquisition also shows characteristics of prepared learning (Chomsky, 1959; Pinker, 1994). Children learn language with remarkable ease, acquiring complex grammatical rules without explicit instruction. Moreover, there appears to be a critical period for language learning. Children who learn a second language before puberty typically achieve native-like fluency, while those learning after puberty rarely do (Johnson & Newport, 1989).

Not all psychologists agree that language is a prepared behavior. B.F. Skinner (1957) argued in his book Verbal Behavior that language could be explained entirely through operant conditioning principles. According to Skinner, infants produce random vocalizations, and parents and caregivers selectively reinforce sounds that approximate words in their language. When a baby says something like ‘ma-ma,’ parents respond with excitement, attention, and affection—powerful social reinforcers. Through this process of shaping—reinforcing successive approximations to the target behavior—children gradually acquire language.

Noam Chomsky (1959) famously criticized this view, arguing that operant conditioning cannot explain how children acquire complex grammatical rules so quickly, often producing sentences they have never heard before. The debate between Skinner’s learning-based account and Chomsky’s nativist view—which proposes an innate ‘language acquisition device’—remains one of the most significant controversies in psychology. Today, most researchers believe both learning and innate predispositions play important roles, with biology preparing humans to learn language easily while environmental input shapes the specific language acquired (Pinker, 1994).

Striking evidence comes from cases of children raised in extreme isolation. Genie, discovered at age 13 after years of isolation and abuse, never acquired normal language despite intensive intervention (Curtiss, 1977). Her case suggests that if language is not learned during the critical period, full acquisition becomes impossible.

Even more compelling evidence comes from the creation of Nicaraguan Sign Language. When deaf children in Nicaragua were brought together in schools for the first time in the 1980s, they spontaneously created a new sign language. Remarkably, the youngest children in each generation added more complex grammatical structures that older learners couldn’t master (Senghas & Coppola, 2001). This demonstrates both the prepared nature of language learning and the existence of a critical period.

Unprepared Behaviors

Unprepared behaviors represent the middle of the preparedness continuum. These are behaviors that organisms can learn, but only with moderate effort and repeated practice (Seligman, 1970). Most of what we typically think of as “learning” involves unprepared behaviors.

Learning to read and write are classic examples of unprepared behaviors. Unlike spoken language, written language is a recent human invention (only about 5,000 years old), so humans have not evolved specific adaptations for literacy. Learning to read requires years of instruction and practice. Even adults who are fluent readers had to invest considerable effort to learn this skill.

Similarly, learning mathematics, playing musical instruments, driving a car, or mastering academic subjects all involve unprepared behaviors. These skills are learnable—they don’t conflict with our biology—but they require sustained effort, practice, and often explicit instruction.

Most of the learning we’ll study in this course involves unprepared behaviors. Classical conditioning, operant conditioning, and observational learning typically involve learning arbitrary associations between stimuli and responses. A rat learning to press a lever for food or a dog learning to salivate to a bell are examples of unprepared learning—possible, but requiring training.

Contraprepared Behaviors

Contraprepared behaviors are those that organisms find extremely difficult or impossible to learn because they conflict with evolved predispositions (Seligman, 1970). These behaviors work against biological programming, making learning slow, difficult, or impossible even with extensive training.

The classic demonstration comes from experiments trying to teach pigeons behaviors that conflict with their natural tendencies. It’s easy to train pigeons to peck a key for food—pecking is how pigeons naturally eat. It’s also easy to train them to fly or flap their wings to avoid shock—that’s their natural escape response. However, it’s nearly impossible to train pigeons to peck to avoid shock or to flap wings to get food (Bolles, 1970). These arbitrary pairings violate the pigeon’s evolved association between pecking and eating, and between wing-flapping and escape.

Instinctive Drift: When Biology Overwhelms Learning

Some of the clearest evidence for biological constraints comes from the phenomenon of instinctive drift—the tendency for learned behaviors to drift toward instinctive behaviors over time (Breland & Breland, 1961). Keller and Marian Breland, former students of B.F. Skinner, discovered this phenomenon while training animals for commercial purposes.

The Brelands attempted to train a raccoon to pick up coins and deposit them in a piggy bank for food rewards. Initially, the raccoon learned this behavior easily. However, over time, the raccoon began “misbehaving.” It would rub the coins together, dip them in the bank and pull them back out, and generally refuse to let go of them. The training was breaking down despite continued reinforcement (Breland & Breland, 1961).

What was happening? Raccoons have a natural behavior of “washing” their food (actually manipulating it to assess texture and remove unwanted parts). The coins had become associated with food, and the raccoon’s instinctive food-handling behaviors were overriding the learned response. This is instinctive drift—learned behavior drifting back toward instinctive patterns.

The Brelands encountered similar problems with pigs. They tried to train pigs to pick up wooden tokens and deposit them in a piggy bank. Initially successful, the pigs eventually began dropping the tokens, pushing them with their snouts, rooting around them, and tossing them in the air—behaviors that resemble a pig’s natural foraging and rooting behaviors (Breland & Breland, 1961).

The Brelands also struggled to train chickens for stage performances. They attempted to train a chicken to walk out on stage, stand on a platform, and remain still while the audience watched. Despite initial success, the chicken would eventually begin scratching at the platform—a natural foraging behavior—rather than standing still. The chickens’ instinctive scratching and pecking behaviors, deeply tied to food-seeking, continually interfered with the trained ‘stand still’ response. Like the raccoon and pig examples, the chicken demonstrates that when learned behaviors involve stimuli associated with food, instinctive food-related behaviors tend to emerge and compete with training (Breland & Breland, 1961).

These failures occurred despite the Brelands’ expertise and despite the animals initially learning the tasks. The behaviors weren’t impossible to learn, but biological predispositions eventually overwhelmed the learned responses. This was a significant challenge to the behaviorist assumption that any behavior could be trained with proper reinforcement schedules.

Biological Preparedness in Taste Aversion Learning

Some of the clearest evidence for prepared learning comes from taste aversion learning. John Garcia and colleagues discovered that rats could learn to avoid foods that made them sick, even when the illness occurred hours after eating—a finding that violated established principles of learning (Garcia & Koelling, 1966; Garcia, Ervin, & Koelling, 1966).

According to traditional learning theory, associations should form best when stimuli are close together in time. Yet rats learned to associate taste with illness even with delays of several hours. Moreover, rats readily learned to associate tastes with nausea, but had difficulty learning to associate tastes with shock, or lights with nausea. The associations that formed easily were those that made biological sense—tastes predict food poisoning, not electric shock (Garcia & Koelling, 1966).

This selectivity in what associations form easily reflects evolutionary preparation. In nature, poisonous foods cause delayed illness. An organism that could learn taste-illness associations despite long delays would have a significant survival advantage. Evolution has therefore made this particular form of learning remarkably easy—often requiring just a single experience—while making biologically irrelevant associations (like taste-shock) difficult to form.

The evolutionary advantage is obvious: animals that quickly learn to avoid poisonous foods survive to reproduce. The specificity of the learning—tastes paired with nausea, but not with other consequences—reflects the structure of the natural problem the learning mechanism evolved to solve (Rozin & Kalat, 1971).

Implications for Understanding Learning

The discovery of biological constraints fundamentally changed how psychologists think about learning. The behaviorist dream of finding universal laws that apply equally to all species, all behaviors, and all situations was shown to be unrealistic. Learning mechanisms are not general-purpose devices; they are specialized tools shaped by evolution to solve specific adaptive problems (Timberlake, 1994; Gallistel, 1990).

This doesn’t mean we can’t study learning scientifically or discover important principles. But it does mean we must consider the evolutionary context. What is an organism prepared to learn? What problems did its ancestors need to solve? Understanding the biological constraints helps us predict what will be easy versus difficult to learn.

For practical applications, recognizing biological constraints is essential. Animal trainers have learned to work with, not against, an animal’s natural behavioral tendencies (Pryor, 1999). Educators recognize that some subjects (like spoken language) align with human prepared learning, while others (like reading or mathematics) require extensive explicit instruction because they’re unprepared.

Looking Forward

Now that we understand both what learning is and how biology constrains what can be learned, we’re ready to explore the major theoretical approaches to studying learning in Part 3. We’ll examine how behaviorism, cognitive approaches, and neuroscience have contributed to our understanding, and consider why psychologists have relied so heavily on animal research—and what we can and cannot conclude from it.