If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

Main content

Learned behaviors

Habituation, imprinting, classical conditioning, operant conditioning, and cognitive learning.

Key points

  • Habituation is a simple learned behavior in which an animal gradually stops responding to a repeated stimulus.
  • Imprinting is a specialized form of learning that occurs during a brief period in young animals—e.g., ducks imprinting on their mother.
  • In classical conditioning, a new stimulus is associated with a pre-existing response through repeated pairing of new and previously known stimuli.
  • In operant conditioning, an animal learns to perform a behavior more or less frequently through a reward or punishment that follows the behavior.
  • Some animals, especially primates, are capable of more complex forms of learning, such as problem-solving and the construction of mental maps.

Introduction

If you own a dog—or have a friend who owns a dog—you probably know that dogs can be trained to do things like sit, beg, roll over, and play dead. These are examples of learned behaviors, and dogs can be capable of significant learning. By some estimates, a very clever dog has cognitive abilities on par with a two-and-a-half-year-old human!1
In general, a learned behavior is one that an organism develops as a result of experience. Learned behaviors contrast with innate behaviors, which are genetically hardwired and can be performed without any prior experience or training. Of course, some behaviors have both learned and innate elements. For instance, zebra finches are genetically preprogrammed to learn a song, but the song they sing depends on what they hear from their fathers.
In this article, we'll take a look at some examples of learned behaviors in animals. We'll start with simple ones like habituation and imprinting, then work our way up to complex cases like operant conditioning and cognitive learning.

Simple learned behaviors

Learned behaviors, even though they may have innate components or underpinnings, allow an individual organism to adapt to changes in the environment. Learned behaviors are modified by previous experiences; examples of simple learned behaviors include habituation and imprinting.

Habituation

Habituation is a simple form of learning in which an animal stops responding to a stimulus, or cue, after a period of repeated exposure. This is a form of non-associative learning, meaning that the stimulus is not linked with any punishment or reward.
For example, prairie dogs typically sound an alarm call when threatened by a predator. At first, they will give this alarm call in response to hearing human steps, which indicate the presence of a large and potentially hungry animal.
A photograph of 4 prairie dogs. Two of the prairie dogs are poking out of a hole and the other 2 are near the hole.
Image credit: Black-tailed prairie dogs by Mathae, CC BY-SA 3.0
However, the prairie dogs gradually become habituated to the sound of human footsteps, as they repeatedly experience the sound without anything bad happening. Eventually, they stop giving the alarm call in response to footsteps. In this example, habituation is specific to the sound of human footsteps, as the animals still respond to the sounds of potential predators.

Imprinting

Imprinting is a simple and highly specific type of learning that occurs at a particular age or life stage during the development of certain animals, such as ducks and geese. When ducklings hatch, they imprint on the first adult animal they see, typically their mother. Once a duckling has imprinted on its mother, the sight of the mother acts as a cue to trigger a suite of survival-promoting behaviors, such as following the mother around and imitating her.
A photograph of a duck with ducklings following along behind it.
Image credit: Behavioral biology: Figure 6 by OpenStax College, Biology, CC BY 4.0
How do we know this is not an innate behavior, in which the duckling is hardwired to follow around a female duck? That is, how do we know imprinting is a learning process conditioned by experience? If newborn ducks or geese see a human before they see their mother, they will imprint on the human and follow it around just as they would follow their real mother.
An interesting case of imprinting being used for good comes from efforts to rehabilitate the endangered whooping crane by raising chicks in captivity. Biologists dress up in full whooping crane costume while caring for the young birds, ensuring that they don't imprint on humans but rather on the bird dummies that are part of the costume. Eventually, they teach the birds to migrate using an ultralight aircraft, preparing them for release into the wild.2,3
The photograph on the left is a biologist in an all white cover up that also covers his head, and he is holding a whooping crane puppet. The photograph on the right is 8 whooping cranes flying in a straight line with a ultralight aircraft flying in front of the cranes.
Image credits: left, Costumed human by Steve Hillebrand/USFWS, CC BY 2.0; right, Whooping crane ultralight migration by USFWS, CC BY 2.0

Conditioned behaviors

Conditioned behaviors are the result of associative learning, which takes two forms: classical conditioning and operant conditioning.

Classical conditioning

In classical conditioning, a response already associated with one stimulus is associated with a second stimulus to which it had no previous connection. The most famous example of classical conditioning comes from Ivan Pavlov’s experiments in which dogs were conditioned to drool—a response previously associated with food—upon hearing the sound of a bell.
As Pavlov observed, and as you may have noticed too, dogs salivate, or drool, in response to the sight or smell of food. This is something dogs do innately, without any need for learning. In the language of classical conditioning, this existing stimulus-response pair can be broken into an unconditioned stimulus, the sight or smell of food, and an unconditioned response, drooling.
A picture of a dog sitting in front of a bowl of food. The dog is salivating and the image is titled Unconditioned response: the dog salivates in response to seeing food.
Image credit: Behavioral biology: Figure 7 by OpenStax College, Biology, CC BY 4.0
In Pavlov's experiments, every time a dog was given food, another stimulus was provided alongside the unconditioned stimulus. Specifically, a bell was rung at the same time the dog received food. This ringing of the bell, paired with food, is an example of a conditioning stimulus—a new stimulus delivered in parallel with the unconditioned stimulus.
A picture of a dog sitting in front of a bowl of food and above the bowl of food is a bell that has sound waves coming from it. The dog is salivating and the image is titled Conditioning: every time the dog sees food, a bell is rung.
Image credit: Behavioral biology: Figure 7 by OpenStax College, Biology, CC BY 4.0
Over time, the dogs learned to associate the ringing of the bell with food and to respond by drooling. Eventually, they would respond with drool when the bell was rung, even when the unconditioned stimulus, the food, was absent. This new, artificially formed stimulus-response pair consists of a conditioned stimulus, the bell ringing, and a conditioned response, drooling.
A picture of a dog looking at a bell that has sound waves coming from it. The dog is salivating and the image is titled Conditioned response: the dog salivates in response to a bell being rung.
Image credit: Behavioral biology: Figure 7 by OpenStax College, Biology, CC BY 4.0
Is the unconditioned response, drooling in response to food, exactly identical to the conditioned response, drooling in response to the bell? Not necessarily. Pavlov discovered that the saliva in the conditioned dogs was actually different in composition than the saliva of unconditioned dogs.

Operant conditioning

Operant conditioning is a bit different than classical conditioning in that it does not rely on an existing stimulus-response pair. Instead, whenever an organism performs a behavior—or an intermediate step on the way to the complete behavior—it is given a reward or a punishment. At first, the organism may perform the behavior—e.g., pressing a lever—purely by chance. Through reinforcement, the organism is induced to perform the behavior more or less frequently.
One prominent early investigator of operant conditioning was the psychologist B. F. Skinner, the inventor of the Skinner box, see image below. Skinner put rats in boxes containing a lever that would dispense food when pushed by the rat. The rat would initially push the lever a few times by accident, and would then begin to associate pushing the lever with getting the food. Over time, the rat would push the lever more and more frequently in order to obtain the food.
A diagram of a setup for a rat. The rat is shown in a box, and on the floor of the box is the label electrified grid. The rat is facing the left wall of the box and there are 5 different structures on the wall. At the top of the box is a structure labeled loudspeaker, below that are 2 structures labeled lights and then there is a rectangle labeled response lever and a square labeled food dispenser.
Image credit: modified from Skinner box by Andreas1, CC BY-SA 3.0; the modified image is licensed under a CC BY-SA 3.0 license
Not all of Skinner's experiments involved pleasant treats. The bottom of the box consisted of a metal grid that could deliver an electric shock to rats as a punishment. When the rat got an electric shock each time it performed a certain behavior, it quickly learned to stop performing the behavior. As these examples show, both positive and negative reinforcement can be used to shape an organism's behavior in operant conditioning. Ouch! Poor rats!
Operant conditioning is the basis of most animal training. For instance, you might give your dog a biscuit or a "Good dog!" every time it sits, rolls over, or refrains from barking. On the other hand, cows in a field surrounded by an electrified fence will quickly learn to avoid brushing up against the fence.4
As these examples illustrate, operant conditioning through reinforcement can cause animals to engage in behaviors they would not have naturally performed or to avoid behaviors that are normally part of their repertoire.

Learning and cognition

Humans, other primates, and some non-primate animals are capable of sophisticated learning that does not fit under the heading of classical or operant conditioning. Let's look at some examples of problem-solving and complex spatial learning in nonhuman animals.

Problem-solving in chimpanzees

The German scientist Wolfgang Köhler did some of the earliest studies on problem-solving in chimpanzees. He found that the chimps were capable of abstract thought and could think their way through possible solutions to a puzzle, envisioning the result of a solution even before they carried it out.
For example, in one experiment, Köhler hung a banana in the chimpanzees' cage, too high for them to reach. Several boxes were also placed randomly on the floor. Faced with this dilemma, some of the chimps—after a few false starts and some frustration—stacked the boxes one on top of the other, climbed on top of them, and got the banana. This behavior suggests they could visualize the result of stacking the boxes before they actually carried out the action.5

Spatial learning in rats

Learning that extends beyond simple association is not limited to primates. For instance, maze-running experiments done in the 1920s—maze shown below—demonstrated that rats were capable of complex spatial learning.6,7
A diagram labeled T-maze is shown. On the left is a vertical box labeled start and from the box the maze branches into a T. The lower portion ends and the bottom of the image, the upper portion branches again into another T. The left portion ends, the right portion branches into a vertical T. The upper portion of the vertical T ends, the lower portion continues downward and branches into a horizontal T. The left portion ends, the right portion continues and branches into a vertical T. The lower portion ends, the upper portion branches into a final horizontal T. The left portion of the final branch ends, and the right portion ends at a box labeled Food.
Image credit: modified from Behavioral biology: Figure 9 by OpenStax College, Biology, CC BY 4.0; based on original publication by Blodgett6, reproduced by Tolman7
In these experiments, rats were divided into three groups:
  • Group I: Rats got food at the end of the maze from day one.
  • Group II: Rats were placed in the maze on six consecutive days before receiving food at the end of the maze.
  • Group III: Rats were placed in the maze for three consecutive days before receiving food at the end of the maze.
Not surprisingly, rats given a food reward from day one appeared to learn faster—had a more rapid drop in their number of errors while running the maze—than rats not given an initial reward. What was most striking, however, was what happened after the Group II and III rats were given food.
A horizontal phylogenetic tree extending from brackets is shown. The lines end on the right side of the diagram and are labeled from top to bottom Finch, Alligator, Chimpanzee, Frog, Tuna, Lancelet. The bottom line of the first bracket extends to the end of the diagram and is labeled Lancelet, and the top line of the first bracket branches into a second bracket. From there, the bottom line of the second bracket extends to the end of the diagram and is labeled Tuna, and the Top line of the second bracket branches into a third bracket. The bottom line of the third bracket extends to the end of the diagram and is labeled Frog and the top line of the third bracket branches into a fourth bracket. From there, the bottom line of the fourth bracket extends to the end of the diagram and is labeled Chimpanzee, and the top line of the fourth bracket branches into a fifth bracket. The bottom line of the fifth bracket extends to the end of the diagram and is labeled Alligator, and the top line of the fifth bracket extends to the end of the diagram and is labeled Finch.
Image credit: modified from Behavioral biology: Figure 9 by OpenStax College, Biology, CC BY 4.0; based on original publication by Blodgett8, reproduced by Tolman9
In both groups, the day after the food had been provided, the rats showed a sharp drop in number of errors, almost catching up to the Group I rats. This pattern suggested that the Group II and III rats had, in fact, been learning efficiently, building a mental map, in the previous days. They just didn't have much reason to demonstrate their learning until the food showed up!
These results show that rats are capable of complex spatial learning, even in the absence of a direct reward, in other words, without reinforcement. Later experiments confirmed that the rats make a representation of the maze in their minds—a cognitive map—rather than simply learning a conditioned series of turns.

Want to join the conversation?