How dogs learn

Two systems running on every walk

Dog professionals in general, and dog walkers in particular, should have a basic understanding of how dogs learn. When working with dogs, handlers should follow the basic principles of behavioral learning theory. Dogs learn mainly in two ways: by association (classical conditioning) and by consequence (operant conditioning).

Learning by association: classical conditioning

Through classical conditioning, dogs learn that one event predicts another, creating an automatic emotional or physical response, like a reflex. A neutral signal (e.g., a light or a sound) is followed, usually many times, by something that triggers a natural reflex (e.g., food), until the signal alone triggers the reaction. No behavior is needed and nothing is earned: the first event simply comes to predict the second.

This is involuntary, and it drives the dog's emotions, which is why it matters so much. Through classical conditioning, we can shape a dog's view of their environment. This is especially important when working with dogs with phobias, fear and aggression. Change what the scary thing predicts, and the feeling changes.

• Classical conditioning is also known as Pavlovian or respondent conditioning.
• Classical conditioning was discovered by Ivan Pavlov in the 1890s, while he was studying digestion in dogs (more in lesson 2).

Learning by consequence: operant conditioning

Through operant conditioning, dogs learn by trial and error how their voluntary behavior changes what happens next. Voluntary behaviors are shaped, strengthened or weakened by what happens right after them (reinforcement or punishment). The dog chooses the behavior, and the consequence makes it more or less likely next time. This is the basis of teaching cues like "sit" and "stay": something happens (the cue "sit"), the dog acts (sits), and a consequence follows (a treat), so the dog sits when they hear the cue because sitting has earned treats before.

People learn this way too. You keep going back to the coffee shop that gives you a free drink after every ten, and you stop feeding coins to a vending machine that has swallowed your money twice.

• Operant conditioning is also known as instrumental conditioning or consequence-based learning.
• In 1898, psychologist Edward Thorndike published his puzzle-box experiments with cats, which led to his Law of Effect: behavior followed by a satisfying result becomes more likely. In the 1930s, B.F. Skinner built on Thorndike's work and coined the term operant conditioning.
• Classical and operant conditioning are both forms of associative learning.

Other ways dogs learn

Dogs also learn in a few other ways, but almost everything you'll see on a walk traces back to the first two.

• Social and observational learning: Dogs learn by watching people and other dogs. For instance, a puppy learns to use a dog door faster after watching an older dog go through it.
• Cognitive learning: Dogs use memory and trial-and-error reasoning to solve puzzles or find their way around their environment.
• Habituation: Dogs learn to ignore repetitive, non-threatening things, such as vacuum cleaners and kitchen noises, once those things turn out to have no meaningful consequences. It is one of the simplest forms of learning, and it is non-associative: nothing good or bad follows, so the reaction fades (more in lesson 7).

How dogs see the world

In essence, a dog is constantly answering two questions: "What is safe or good for me?" (classical conditioning) and "What works and what doesn't?" (operant conditioning). Consequences create associations too. The two systems are always running together.

Key takeaways

• Dogs learn mainly in two ways, by association (classical conditioning) and by consequence (operant conditioning).
• Classical conditioning is involuntary and drives emotion, so change what a scary thing predicts and the feeling changes.
• In operant conditioning the dog chooses a behavior, and what happens right after makes it more or less likely next time.
• In habituation, a harmless thing has no consequences and the dog's reaction to it fades.