Browse Our Dictionary
Type: Learning Theory

300 Peck

A method for teaching behaviors used by clicker trainers, taken from a study on pigeons. The method uses variable ratio reinforcement schedule.

A researcher taught pigeons to peck a bar for a reward.  Slowly extending the number of pecks between rewards, the researcher was able to train the pigeons to reliably peck the bar 300 times before they were rewarded.

This method is used to teach long-duration behaviors in dogs, such as stay.

ABC

ABC stands for Antecedents, Behavior, Consequences.  It refers to the complete “picture” of a behavior in context.  Behaviorists look at an animal’s behavior as being determined by triggers in the animal’s internal and external environment, which are referred to as the behavior’s antecedents.  The behavior itself can be analyzed in terms of the function it has for the animal, bearing in mind the antecedents that were there.  Whatever changes in the animal’s environment happen as a result of the behavior are called the consequences of that behavior.

Dog behaviorists keep the ABCs of behavior in mind because they can change both the antecedents to, and the consequences of a behavior to make their interventions more likely to succeed.

Abolishing Operations

An abolishing operation is a change in the environment, or the dog, which results in a reinforcement or a punishment being less effective.

For example, deciding to do a clicker training session just after the dog has had dinner can make him less interested in the treats and more likely to lose focus, or refuse to participate.  Being full of food (satiated) can therefore be considered as an abolishing operation in this case.

Adjunctive Behavior

In operant conditioning, an adjunctive behavior is one that occurs in the context of a reinforcement schedule – it happens during training – but is not directly controlled by any features of the reinforcement schedule itself. Schedule-induced aggression is an example of an adjunctive behavior. These behaviors are usually stereotypical and are thought to be a kind of displacement – they tend to happen when the probability of reinforcement is low, possibly to relieve anxiety or boredom.

Antecedent Arrangements

A way to describe how the environment that the dog is in has been set up, deliberately or not. The antecedent arrangements determine which behavior the dog is most likely to perform.

For example, a dog might only perform a’ “down” when the handler says “down” if he has seen that there is a cup of treats next to her on the counter. Both the cue for “down” and the cup are relevant parts of the environment for the dog. Antecedent arrangements are an important part of all behaviors, not just ones that the owner has put on cue – for example, a dog might get up onto a counter if there is a nearby chair he can use as a step, but not otherwise.

The dog’s emotional state can also be considered as part of the antecedent arrangements – if a dog is stressed, he is less likely to be able to listen to his owner and perform a behavior on cue, but if he is relaxed, he is more likely to do so.

Appetitive

Anything that is considered good or desirable for the dog.  It can be given to the dog by his owner, like treats or praise, or discovered by the dog on his own, like a dead groundhog to roll in or a soft pile of sand to dig.

Appetitive and Positive are considered to be synonymous; their opposites are Aversive and Negative.

Applied Behavioral Analysis

Synonymous with “behavior modification”.  ABA is a systematic approach to the study of, and application of interventions to, an animal’s behavior.

Associative Learning

When a dog learns that one thing causes another by being repeatedly exposed to the cause and effect. For example, a dog learns that when his owner opens a certain drawer, he is going to get a treat. If his owner does this every day, the dog might become excited whenever he hears his owner opening that drawer after a few days. Both classical and operant conditioning are examples of associative learning.

Attachment Theory

A theory in psychology that looks at how humans and animals form connections with each other, most often how parents and their children become bonded and the significance of that bond for the social, emotional and cognitive development of both the parent and the child.

Research has suggested that attachment theory also applies to the relationship between dogs and their owners, showing that there is a deep-seated emotional connection between the two that, ideally, contributes to the emotional wellbeing of both parties.

Research has also suggested that dogs are better able to learn new tasks when they are more securely attached to the person teaching them, suggesting that the attachment is also providing some cognitive benefits.

Aversive

An aversive is anything that makes a given behavior less likely to happen in the future; usually an unpleasant experience.

For example: a dog tries to eat a bee, and gets stung. The next time he sees a bee, he leaves it alone.  This means the bee sting was aversive for the dog.

Although we can usually have a pretty good idea of what kinds of things will count as aversive for dogs, the only way to tell for sure in an individual situation is to analyze the dog’s behavior. Being spiked by a porcupine’s quill, for example, would probably be an aversive to an inquisitive puppy, but not to a highly aroused dog in full prey drive.

Avoidance

According to Domjan’s “Principles of Learning and Behavior”:

An avoidance procedure involves a negative contingency between an instrumental response and the aversive stimulus.  If the response occurs, the aversive stimulus is omitted. By contrast, punishment involves a positive contingency: the target response produces the aversive outcome.

The most common use of the term “avoidance” in dog training is in snake training.  Dogs are taught that if they move away from the sight, sound, and smell of a rattlesnake, they can avoid an aversive stimulus – usually a shock.  Avoidance is taught by pairing the cue – in this case, the snake – with the shock until the dog performs the desired behavior – moving away.  When the dog reliably moves away from the snake, thus avoiding the shock, he has learned the desired avoidance behavior.

Behaviorism

The behaviorist approach to psychology is concerned with measuring and analyzing the outward expression of behavior, and not emotions, thoughts or perceptions.  Behaviorists interpret what an animal or human does in terms of patterns of stimulus and response.  The stimulus for a behavior always comes from the environment, and the response is  determined by the way the stimulus acts on the subject.

Dog behaviorists use many of the principles of behaviorism, like classical and operant conditioning, to modify the behavior of dogs, although not all dog behaviorists believe that we ought to ignore what the dog is thinking and feeling as a way to explain what he is doing.

CER

A CER is a Conditioned Emotional Response.  CER+ stands for Positive Conditioned Emotional Response, and CER- stands for Negative Conditioned Emotional Response.

A CER is when a dog learns to associate the presence of a stimulus with a (pleasant or unpleasant) consequence, and changes his behavior toward the stimulus as a result. One example is, a dog wags his tail when he sees the leash.  This is a CER+.

Classical Conditioning

Also known as Pavlovian, or respondent conditioning.

There are three stages to classical conditioning:

1. Before conditioning, an unconditioned stimulus produces an unconditioned response.  For example, a dog salivates when food is put in front of it.

2. During conditioning, a neutral stimulus – something that does not usually elicit any behavior at all in the dog – is presented before the unconditioned stimulus. For example, a bell is rung immediately before food is put in front of the dog.  This is done many times, so that they dog comes to be able to use the neutral stimulus to predict the arrival of the unconditioned stimulus.

3. After conditioning, the dog has learned that the neutral stimulus predicts the unconditioned stimulus.  He will start to salivate when he hears the bell because he anticipates the arrival of food.   The bell is now a conditioned stimulus, because it makes the dog respond in the same way he would to the unconditioned stimulus.

The key difference between classical and operant conditioning is that in classical conditioning, the dog does not have to do anything on purpose, only to react to the stimulus in the way it normally would.

Classical Conditioning can be positive or negative, building a pleasant or unpleasant emotional association.

Conditioned Punisher

An object, behavior, or sound that is punishing to a dog because he has been conditioned to associate it with unpleasant consequences. One example is the dog’s own name – if an owner only calls her dog when she wants him to come in from playing in the yard, the dog’s name can become a conditioned punisher because it has been associated with being taken away from something fun.

Conditioned Reinforcer

An object, behavior or sound that is positively reinforcing to a dog because he has been conditioned to have a positive emotional reaction to it through associative learning. One example is hand clapping – if an owner claps whenever the dog performs well, and gives the dog treats or the chance to play with a toy, the dog will begin to enjoy the clapping even when it is not paired with anything.

Conditioned Reponse

A response that becomes associated with something that does not usually cause that response.  For example, when the dog has learned that Pavlov always rings the bell before he delivers food, he will begin to salivate as soon as he hears the bell.  Salivating is a conditioned response to the sound of the bell ringing, since the dog would not have done this if he hadn’t learned that there was an association between the bell and the food.

Conditioned Stimulus

Something the dog has previously ignored or not noticed becomes relevant to the dog and elicits a response after it has become associated with a consequence.

For example, Pavlov’s bell becomes a conditioned stimulus that elicits the salivating response, after the dogs have learned that the bell is associated with being fed.

Consolidating Stage

In learning theory, the consolidating stage of learning is when the dog understands the behavior but the behavior is not yet reliable; the dog may not have generalized it or may still make some mistakes.  It is mid-way between the initial acquisition stage, and the final maintenance stage.

Contingency

A complete description of the way the environment is arranged, the stimulus, and the behavior the dog performs.

An operant contingency is a description of the conditions under which a behavior produces a response – either a reward or a punishment. A respondent contingency is a description of the conditions under which an unconditioned stimulus is paired with a conditioned stimulus to create an association.

A zero contingency is where the unconditioned and conditioned stimuli occur randomly, so no association is made.

Counter-Conditioning

Counter-conditioning (CC) is a type of classical conditioning used for behavior modification. The dog is exposed to something that usually causes him to become frightened or aggressive at a distance where he sees it, but the bad reaction is not triggered. As soon as the dog sees the trigger, he is given a food reward. The reward is not dependent on the dog performing any behavior. The aim of counter-conditioning is to change the dog’s emotional response to his triggers from negative to positive. A dog that feels positively about something will no longer feel the need to behave aggressively or fearfully towards it.

Counter-conditioning is often paired with desensitization because both happen at the same time during a successful CC setup.  This is why trainers often use the acronym “DS/CC” to refer to this method of behavior modification.

Cue

Anything – visual, tactile, auditory, or otherwise, that signals to a dog that a given behavior should be performed.  Putting a behavior on cue means teaching the dog to associate a cue with a desired behavior.

A “command” is another word for cue, more closely associated with the trainer using a word to elicit the behavior.

Default Behavior

A default behavior is something that a dog has learned and can do without needing a verbal or gesture cue from his owner.

For example, if an owner asks her dog to sit every time they approach a closed door, the dog will eventually start to sit without the owner having to ask.  The door becomes the cue.

In heelwork, the dog is expected to sit as a default behavior any time his owner stops moving.

Differential Reinforcement

Reinforcing one choice much more heavily than others when the dog is presented with several choices about which behavior to perform.

Dogs will always choose the behavior with the strongest history of good consequences, so choosing to reinforce one possible behavior in a situation and not others will lead the dog to perform that behavior more often than the others.

This is especially useful when one of the other possible behaviors is self-reinforcing, such as chasing a cat.

Differential Reinforcement of an Incompatible Behavior

In behavior modification, DRI is used to prevent a dog from performing undesirable behaviors, by reinforcing something else the dog does in the presence of the same stimulus that makes it impossible for the dog to perform the undesirable behavior.

For example, a dog jumps on guests, so the owner chooses to reinforce him for sitting whenever he sees a guest.  Sitting is incompatible with jumping up, so the dog is prevented from performing the undesired behavior.

Discriminative Stimulus

In operant conditioning, a stimulus that determines whether or not the response will be reinforced. A positive discriminative stimulus (S+) tells the dog that reinforcement is available. For example, an owner opening the cupboard where the dog’s food is counts as an S+. A negative discriminative stimulus tells the dog that reinforcement is not available. For example, a dog might learn that a particular man never gives him pets or treats even when he does a trick, and so chooses to spend time with other family members instead.

Dishabituation

The opposite of habituation.  A dog can become dishabituated to a stimulus after the dog has already become habituated to that stimulus, if something happens that causes the dog to associate the stimulus with a bad experience.

Errorless Learning

Protocols for training behaviors that set the dog up so that he cannot, or is extremely unlikely to, make any mistakes in learning.  When the dog does make mistakes, they are not acknowledged or punished in any way.

Errorless learning techniques are most commonly found in teaching dogs to discriminate between two stimuli. The idea is to teach the dog to perform the desired behavior without needing him to make mistakes.  For example, if a trainer wanted to teach a dog to pick his dinosaur toy out of a box of other toys, she would start by teaching him to pick up his dinosaur toy, and make doing this highly reinforcing for the dog.  She could then add in other toys, once the dog is already confidently performing the desired behavior.

 

Establishing Operations

An establishing operation is a change in the environment, or the dog, which causes an increase in the effectiveness of a reinforcement or a punishment.

For example, doing a clicker training session with a dog just before his dinner time is likely to make him more motivated by the treats, and therefore more focused on successfully performing the behaviors his owner is teaching him.

Extinction

The process by which a behavior happens less and less frequently, until it no longer happens at all.  A behavior can be made extinct by no longer reinforcing it and sometimes also reinforcing other, alternative behaviors.  Some behaviors are very difficult to extinguish because they are enjoyable for the dog without any involvement of their owner.

Fading the Lure

Luring is a kind of prompting used by some positive reinforcement-based trainers.  In order to make sure that the dog is learning that his behavior is causing the reinforcement to happen, the trainer should start to switch to a different kind of prompting as soon as possible, which is called fading the lure.

This is because the presence of food can overshadow other salient things, like what the dog is doing with his body to get that food, and make the process of learning slower.

Forward Conditioning

A type of counter-conditioning, which is used to change the emotion a dog associates with something from fearful to pleasant by pairing the appearance of the scary thing (the trigger) with a reward, usually food.

In forward conditioning, the trainer gives the dog the food soon after the dog sees (or hears, or smells) the trigger, and continues giving the food until a short amount of time after the trigger has gone.

 

Generalization

(1) In basic obedience, a dog has generalized a cue when he reliably responds to it in every environmental situation.

(2) In behavior modification, a fear is generalized when a dog responds with equal anxiety or *aggression to all instances of an *aversive. For example, if a dog is treated badly by a tall woman with grey hair, it may cause him to become afraid of all tall women with grey hair. The fear has become generalized.

Habituation

“Getting used to” something.  After consistent exposure to a stimulus at a level that does not cause it concern, a dog starts to see that stimulus as neutral and no longer responds to its presence in the environment.

For example, a dog might initially react with concern to a Christmas tree that has been left on the side of the road, and thoroughly investigate it.  After a few days of walking the same route, the tree is still there, but the dog no longer notices it.

Inhibition

When a behavior decreases in frequency and/or severity.  This can refer to a specific behavior, or to all of a dog’s behavior being generally subdued.

For example, promoting bite inhibition is a key part of training a puppy.  A puppy learns to bite with less force and frequency when his owners or littermates stop playing with him when he bites.

Intermittent Reinforcement Schedule

When a trainer reinforces a cued behavior some, but not all, of the times the behavior is performed.  Also called a variable interval schedule.

Trainers often change from a continuous reinforcement schedule, where the dog gets a reward every time he performs the behavior, to an intermittent schedule when the dog has demonstrated that he has reliably learned the cue and is not at all confused about what it means.

Latent Learning

The psychological phenomenon where a dog becomes more skilled at a behavior he is learning from one training session to the next. This is because the learning from each session is being consolidated with the neural pathways of the brain in the time between sessions.

Learned Irrelevance

When a stimulus is repeated over and over again with no consequences, a dog will eventually stop reacting to it.

Similar to habituation, although unlike habituation, once a dog has learned that something is irrelevant, he will not spontaneously begin seeing it as relevant again.

Learning Theory

A part of the discipline of psychology, learning theory encompasses the conceptual frameworks used to describe how dogs learn and retain information.

Matching Law

In applied behavioral analysis, the Matching Law states that the relative proportion of times an animal will perform a behavior in response to a cue is equal to the proportion of times that animal is reinforced for doing so.

When a behavior is reinforced more often, the animal will choose to perform it more often.  Experiments showed that the relative proportion of responses matched the relative proportion of reinforcement almost perfectly.  In dog behavior, this means that behaviors that are most often rewarded will be most likely to occur in the future.  Behaviors that are consistently rewarded, using a high rate of reinforcement, can be expected to be consistently performed.

References:

Herrnstein R. J. Relative and absolute strength of response as a function of frequency of reinforcement. Journal of the Experimental Analysis of Behavior1961;4:267–272.

Reed D.D, Kaplan B.A. The Matching Law: A Tutorial for Practitioners. Behavior Analysis in Practice. 2011;4(2):15-24.

Negative Behavioral Contrast

If a dog is expecting a high value reward for a behavior and receives a low value reward instead, we can expect subsequent performances of that behavior to be slower and/or less accurate because the dog has lost motivation for the task.

For example, if a dog is used to being rewarded with steak for performing a retrieve and gets a piece of kibble instead, we can expect his next retrieve to be slower and the dog less aroused.

Negative Contingency

When a conditioned stimulus signals a decrease in the liklihood that the unconditioned stimulus will occur.

If Pavlov rang a bell every time he took food away from the dogs, this would be a negative contingency – the dogs would learn that the bell predicts the food going away.

Negative contingencies produce inhibitory conditioning, meaning that when a conditioned stimulus is presented to a dog, they inhibit their responses.  The dogs would stop salivating when they heard Pavlov’s bell, if they associated it with food being taken away from them.

Negative Punishment

One of the four quadrants in learning theory, shortened to P-. When a dog performs a behavior, something the dog values is taken away with the aim of producing a decrease in that behavior.

For example, when a puppy bites too hard in play, the owner gets up and leaves the room. The puppy values the owner’s presence, associates the bite with the owner leaving, and does not bite again.

Negative Reinforcement

One of the four quadrants in learning theory, shortened to R-. A consequence to a behavior that takes something negative away from the dog, therefore making the behavior more likely to occur under the same circumstance.

For example, an owner says “Sit!” and pushes down on a dog’s back end, which feels uncomfortable. The dog sits, thus relieving the pressure. Next time the owner says, “Sit!”, the dog is more likely to do so.

Non-Associative Learning

When a dog learns that one thing causes another without the need for repeated exposure. For example, a dog might remember where in the park he nearly caught a squirrel, and go back to that location next time he visits the park.

Operant Conditioning

Sometimes called Skinnerian conditioning, after psychologist B.F. Skinner.  A type of learning where an animal must perform a behavior in order to obtain a reward.

For example: a rat is placed in a box with a lever. When he presses the lever, food is dispensed.  After enough repetitions, the rat learns that if he wants food, he should press the lever.

Operant conditioning forms the basis of almost all dog obedience, whether taught using positive reinforcement or compulsion.

Overshadowing

When there are two different stimuli in the environment that a dog is paying attention to, and one of those stimuli has a much stronger importance to the dog than the other one, so that the dog only pays attention to the most important stimulus and not to the less important one.

For example, if an owner only trains her dog when she is wearing a bait bag full of smelly treats, the presence of the bait bag may overshadow the verbal cue, so that the dog only listens to the owner when she is wearing her bait bag and doesn’t learn that the verbal cue is important.

Partial Reinforcement Extinction Effect

If a trainer wants to extinguish a behavior in a dog, she will usually stop reinforcing it until the dog realizes it is no longer associated with a reward and stops performing it.

If, however, the behavior the trainer wants to extinguish was being maintained by an intermittent reinforcement schedule, the dog is already used to not receiving a reward every time he does the behavior.  It can take a great deal longer for the dog to realize he will never be rewarded for that behavior again, if he realizes this at all.  The name given to this is the Partial Reinforcement Extinction Effect, or PREE.

Place Learning

Refers to the fact that dogs tend to learn behaviors in a specific environmental context. A dog may reliably sit on cue in his owner’s front room, but will not do so on a hiking trail, because he has only learnt that “sit, Fido” means “sit when I am in this room”. Dogs must be taught to generalize their learning by a process of proofing.

Poisoned Cue

(1) If a dog has a history of associating a cue with punishment, the dog may have developed a negative emotional response (CER-) to that cue. If a trainer uses that cue in training, the dog may become anxious and be less responsive to reinforcement.

(2) If a dog has been presented with a cue many times with no consequences, it will be more difficult to create any associations with that cue.

Positive Behavioral Contrast

A dog’s behavior will change if he expects a low value treat and receives a high value treat.  He will be more motivated to perform cued behaviors and usually display a higher level of arousal.

For example, if a trainer usually trains her dog with kibble rewards for a retrieve, and rewards her dog with a piece of steak instead, she can expect the next retrieve to be faster and more accurate.

 

Positive Punishment

One of the four quadrants of learning theory, used for behavior modification and shortened to P+. When a dog performs a behavior, a stimulus is administered with the aim of producing a decrease in that behavior.

For example, when a dog walks over the boundary of an invisible fence, he is given an electric shock. If the shock is strong enough, he will associate it with walking to that place, and choose not to walk there again.

Positive Reinforcement

One of the four quadrants in learning theory, shortened to R+.  A consequence to a behavior that adds something desirable to the dog, and therefore makes the behavior more likely to occur under the same circumstance.

For example, an owner says “Sit!” and raises a treat above a dog’s head. As the dog’s head comes up, his backside goes down and he sits. His owner gives him the treat, which makes him more likely to sit the next time his owner raises her hand in the same way.

Positive reinforcement-based training is training that sets the dog up so that he performs desirable behaviors, which can then be rewarded.  When the dog performs undesirable behaviors, the trainer focuses on changing the environment or training an alternative behavior rather than using punishment.

Pre-Exposure Effect

If a dog has already been exposed to a stimulus many times before training begins, training may be interfered with and made more difficult.

For example, if a dog’s owner has used “Come!” repeatedly with no effect and no reinforcement or punishment for the dog, it will be a lot more difficult to train the dog to recall using the word “Come!”.  This is because the dog has already learned that the word has no meaning.

Or, if Pavlov’s dogs had often heard the bell without food being presented, they may not have salivated when they heard the bell even after it had been paired many times with food.

Premack Principle

A concept in learning theory, first outlined by the psychologist David Premack. When applied to dogs, the Premack Principle states that a trainer can use a more probable behavior as a reward for a less probable behavior. For example, being allowed sniff a fire hydrant can be used as a reward after a period of loose leash walking.

Primary Reinforcer

Reinforcers are used in operant conditioning to make a desired behavior more likely.  A primary reinforcer is something that a dog finds naturally rewarding, without any need for learning. For example, drinking water or eating food, especially when the dog is thirsty or hungry.

Prompting

When a trainer manipulates the environment – changing the antecedent arrangements – so that the dog is more likely to perform a desired behavior.  For example, putting a food treat on a mat is a prompt for the dog to go and stand on the mat, which can then be marked and rewarded.

Proofing

When a dog first learns a behavior, he may only perform it in the context he learned it in, for example a dog might sit reliably in the bedroom where all his training sessions happen, but he may not sit in the yard, or around the neighborhood. He also might not perform the behavior if there are cars, or other people, or other distractions.

The process of teaching a dog to reliably perform a behavior everywhere is called proofing.  It usually involves re-teaching the dog a behavior in many different locations, so that he comes to learn that it is the verbal or gestural cue that is important, not any feature of the environment.

Punishment

Something is a punishment for a dog if it is an unpleasant consequence for a behavior that decreases the likilhood that that behavior will happen again.  Punishments are intended to be aversive.

Punishments can be divided into two types, according to the quadrants of learning theory. Positive punishment means applying something to the dog that is punishing for him, like a kick in the ribs or a jerk on the leash.  Negative punishment means taking something away from the dog that he enjoys, like putting a puppy in a time-out or shutting a dog in a crate.

Quadrants

In learning theory, the four quadrants are a way to categorize different consequences of a behavior. The quadrants are split into positive and negative, and reinforcement (sometimes referred to as “reward”) and punishment, and are usually labelled R+, R-, P+, and P-.

Rate of Reinforcement

For a behavior, the rate of reinforcement (Rf, sometimes written ROR) is how often a behavior is reinforced in a given time.

For example, asking a dog to sit twenty times in a minute, and rewarding after every sit, is considered a high rate of reinforcement.  Asking for a sit three times in a minute, and rewarding after every sit, is considered to be a low rate of reinforcement.

Reinforcer

Anything that increases the frequency of a desired behavior in a dog.  For example, a treat, or a toy, or the sound of his own barking.  Reinforcers are sometimes categorized into primary and secondary, which refers to whether the reinforcer is something that all dogs find naturally reinforcing, or something that has specific value for this dog through coming to associate it with pleasurable experiences.

Reinforcer Sampling

When a trainer motivates her dog to respond by offering a small amount of a reinforcer, in order to make the dog anticipate more of the reinforcer as a reward for performing the desired behavior.  For example, a trainer might motivate her dog to come when called by showing him that she has his favorite toy in her pocket.

Reward

A consequence for a behavior that the dog sees as desirable, such as a food treat, a game of tug, or a chance to chase squirrels.

Schedule-Induced Aggression

Researchers in operant conditioning observed that animals would display aggressive behaviors when they are given access to positive reinforcement on an intermittent basis, even if there was no previous aggression. This aggression was shown not to be a learnt behavior, to be unrelated to the behavior that was being reinforced, and to follow a pattern. Most aggression happened immediately after reinforcement.

References

Cohen, PS and Looney, TA (1982) Aggression induced by intermittent positive reinforcement Neuroscience & Biobehavioral Reviews 6:1,pp.15-37

Secondary Reinforcer

Reinforcers are used in operant conditioning to make a desired behavior more likely.  Secondary reinforcers are things that a dog has come to value through their being associated with another reward.

For example, a dog may come to enjoy the sound of an alarm clock because it is associated with his owners waking up and giving him affection and food.  The alarm clock sound can be used as a reinforcer for desired behaviors, even at other times of day.

Self-Reinforcing Behavior

Behaviors are considered to be self-reinforcing if they are enjoyable in themselves, rather than because of the impact they have on the dog’s environment.  For example, many dogs enjoy the sound of their own barking, even if nothing in their environment changes when they bark.

Dogs that repeatedly or persistently perform self-reinforcing behaviors to excess can be considered to be suffering from obsessive compulsive disorder.

Signaling

When an owner creates a stimulus that tells the dog that something unpleasant is about to happen.  A safety signal, for example, tells the dog that something unpleasant is not going to happen.  For example, the collar component of an invisible fence will beep when the dog is approaching the boundary area, where he will be shocked.  The dog learns that the beep precedes the shock, and can then avoid the shock.  According to the safety signal hypothesis, the beep could become positively reinforcing because it is associated with avoidance.

In an experiment, psychologists found that rats strongly preferred to be in a situation where an electric shock was signaled by a sound, even when the rats couldn’t avoid the shock.  The experimenters suggested that this was because the sound was a signal the rats could use to predict safe and unsafe periods of time, and the predictability was something the rats preferred.


 

References:

Badia, P and S. Culbertson (1972). The relative aversiveness of signalled vs. unsignalled escapable and unescapable shock. Journal of the Experimental Analysis of Behavior 17(3), pp.463-471.

Simultaneous Conditioning

A kind of conditioning where the reinforcer or punisher is delivered at exactly the same time as the *stimulus.

For example, when conditioning a dog to accept having his coat brushed, a trainer can deliver constant food treats all the time she is using the brush, and stop as soon as the brushing stops.

Skinner Box

A chamber developed by B.F. Skinner to test how animals learn by operant conditioning by isolating them from the kinds of distractions found in a natural environment.  The box contains one or more items that the subject can manipulate, like a lever a rat can press, or a colored key a pigeon can peck.  It also contains lights and sounds that can be used as cues.

Skinner, B.F.

One of the “fathers of modern behavioral science”, Skinner was a professor who pioneered the experimental analysis of behavior. He distinguished between classically conditioned and operantly conditioned behavior, which have become the cornerstones of behavior modification in dogs.

Operant conditioning is sometimes called Skinnerian conditioning, to contrast with Pavlovian conditioning.

 

Social Facilitation

When a dog is motivated to perform a behavior because someone else – another dog, or a human – is doing it.

Different from modeling and social learning, where the dog learns the behavior from someone else doing it, social facilitation is about the reason the dog is performing a behavior he already knows.

Stages of Learning

When a dog is being taught a new behavior, he goes through several stages before he can be said to have completely learned the behavior.

The first stage of learning is acquisition, where the dog is first making the association between the stimulus, behavior, and consequence.  The second is consolidation, where the associations are made stronger and the dog gets “better” at performing the behavior on cue.  Finally, there is maintenance, where the dog can be said to have completely made the association, and it only needs to be prevented from extinction.

Different reinforcement schedules are recommended for the different stages of learning.

Stereotypical Behavior

Stereotypies are repetitive movements of the dogs body, which are able to be controlled and serve no useful purpose other than being reinforcing to the dog. Stereotypical behavior in animals is generally thought of as an indicator of poor welfare, particularly of lack of opportunities for enrichment. Dogs that are kept in kennels often show repetitive behaviors like bouncing, barking, and pacing, some of which can be considered as stereotypical.

Reference:

Denham, DC, Bradshaw, JWS, Rooney, NJ (2014) Repetitive behavior in kennelled domestic dogs: stereotypical or not? Physiology & Behavior 128,pp 288-294

Stimulus

Anything that causes a behavior – conditioned or not – to happen.  For example, a slice of pizza is a stimulus for a dog to begin drooling; the word “sit!” is a stimulus for a dog to sit.

Stimulus Control

A dog is said to have a behavior under stimulus control when he only performs it on a cue the trainer has paired with the behavior, not on any other cue. For example, a dog who immediately puts his front paws on any book after a session of platform work does not have the behavior under stimulus control. If the dog only put his feet on a book when his owner said, “Mark”, for example, that dog could be said to have the behavior under stimulus control.

Stimulus Relevance

Some combinations of conditioned and unconditioned stimuli tend to result in much faster association than others. For example, a dog will learn not to eat something if it makes him feel ill, even if the taste and illness are far apart in time and it only happened once. In contrast, it may take several trials for a dog to learn not to pass the barriers of an invisible fence because location and shock are less relevant a combination.

Superstitious Behavior

Sometimes, dogs make associations between a stimulus, their response, and a consequence that are not quite accurate.  For example, a dog might bark incessantly at the door after his owner has left, because in the past, his owner has happened to come home just as he has been barking.  He has come to associate the two in his mind, even though there is no connection between them.

B.F. Skinner describes this phenomenon in his book, Science and Behavior: 

In superstitious operant behavior…the process of conditioning has miscarried. Conditioning offers tremendous advantages in equipping the organism with behavior which is effective in a novel environment, but there appears to be no way of preventing the acquisition of non-advantageous behavior through accident.

Dog trainers tend to agree that superstitious behavior can be very difficult to extinguish.

Suppression of Behavior

When a dog’s behaviors – either one, or a variety of different behaviors –  become less frequent or less varied as a result of some stimulus acting on the dog.

Positive and negative punishment are intended to suppress undesired behavior.  Excessive and/or improper use of positive punishment in particular can lead to high levels of behavioral inhibition and learned helplessness as the dog becomes unwilling to try new things out of fear that he may be punished.

Thorndike’s Law

The Law of Effect proposed by Edward Thorndike in his 1911 book, Animal Intelligence, is

Of several responses made to the same situation, those which are accompanied or closely followed by satisfaction to the animal will, other things being equal, be more firmly connected with the situation, so that, when it recurs, they will be more likely to recur; those which are accompanied or closely followed by discomfort to the animal will, other things being equal, have their connections with that situation weakened, so that, when it recurs, they will be less likely to occur. The greater the satisfaction or discomfort, the greater the strengthening or weakening of the bond.

This law forms the foundation of operant conditioning, as it describes how a behavior is made more likely to happen when it is positively reinforced, and less likely to happen when it is punished.

Throwing Behavior

Another term for random sampling.  When a dog tries various different known behavior in order to work out how to get his owner to respond.  A dog might throw behaviors as part of a training session, or as a way to get his owner’s attention.

Umwelt

“Umwelt” is a German word, translated as “self-world” or sometimes, as “worldview”.

The term was coined by biologist Jakob von Uexküll, and has recently been popularized by the canine behavioral researcher Alexandra Horowitz, who uses it in her book “Inside of a Dog”.   She uses the term umwelt to describe the way that dogs perceive the world through their senses, and relate to it using their cognitive abilities.

The umwelt of a dog is different from that of a human, because dogs have different sensory and cognitive abilities; scent, for example, is very important for dogs and allow them to perceive things that humans cannot. Humans, by contrast, can see things in more colors, and recognize complex patterns in the environment in a way that dogs cannot.  This means the same object can be endowed with completely different significance, depending on the species perceiving it.

Unconditioned Punisher

An unconditioned punisher is something that a dog is scared of or finds otherwise aversive without first having to be trained to have that response.

For example, the citronella used in bark collars is unpleasant to a dog from the first time he is exposed to it.

Unconditioned Response

A UCR is an innate response that is elicited by a stimulus with no prior conditioning. For example, a dog salivating in the presence of food.

Unconditioned Stimulus

In classical conditioning, a US is a stimulus that elicits the desired response without any conditioning. In Pavlov’s experiment, dogs were conditioned to salivate at the sound of a bell by the bell always being followed by the presentation of food. The food caused the dogs to salivate – it was the unconditioned stimulus – after enough repetitions the bell became a conditioned stimulus and also caused the dogs to salivate.

Value

The value of something refers to how motivating it is for a dog. Items are classed as high value or low value. A high value treat, like a meatball, is a stronger reinforcer than a low value treat like a piece of carrot. That means if a dog expects a meatball, it will be more likely to perform whatever behavior you are giving the cue for than if it expects a carrot.

Value is usually talked about with regards to food and toys, but everything that can be reinforcing for a dog has value, for example social attention like petting, or the chance to go out for a walk, or chasing squirrels. Choosing the right value of reinforcer is a key part of successfully using differential reinforcement in behavior modification.