388126_2372821642069_1836219881_n
(click on to enlarge)

I was informally schooled under the tenets of operant conditioning using marker-training, whereby behaviors are typically shaped or captured using a marker (i.e. bridge), like a clicker, whistle or utterance of the word “good.” This is called reward-based training, but it is composed of more than just the use of positive reinforcement.

579685_1402550693321791_1449467493_n
(click on to enlarge)

Indeed, positive reinforcement is not the only conditioning tactic that can be used to describe operant conditioning, nor is it the only reward-based conditioning tactic. In fact, various training methods are not necessarily mutually exclusive.  In dog training reward-based management is often supplemented or integrated with other tactics or techniques.

In the behavioral management of zoo and sanctuary animals, I relied primarily on shaping or capturing behavior. Today, such “positive reinforcement” has become synonymous with reward-based training. However, negative reinforcement (whereby a new behavior is essentially elicited to replace another) is also technically a form of operant conditioning, just as positive punishment and negative punishment are forms of operant conditioning. If you are already familiar with reward-based training, you too, may have used these other training techniques without  being aware of it.

Many people like myself first learned to train marine mammals before any other captive wildlife species or even domestic animals.  Whether we worked in a zoo, aquarium, marine park, sanctuary or rescue facility, we were encouraged to read Karen Pryor’s Don’t Shoot the Dog. This important and highly acclaimed literary contribution to the animal training community suggests that operant conditioning is a form of training that uses neither “punishment or force.” This is not the case and has led to much confusion.

Because of the piscivorous (i.e. fish-eating) nature of marine mammals and the circumstances surrounding an aquatic lifestyle, marine mammal trainers adopted a conditioning practice, which almost exclusively caters to reward based training program for their subjects (e.g. dolphins and whales and other marine mammals). However, the notion that operant conditioning is based solely on “positive training” is a misnomer. To help clarify, I provided a glossary below for your reference.

Reward-Based Training techniques as categorized by the American Association of Professional Dog Trainers (APDT). The following definitions are also described in greater detail in Pamela Reid’s book Dog Insight and on the APDT website.

Operant Conditioning — A behavioral cue or command is used to signal a specific behavior.

Negative Punishment  — Something desirable like food rewards are taken away or an animal is removed from an activity or circumstance. Extinction is the elimination of behaviors through negative punishment. In this case, an unwanted behavior is ignored until the elicitation of the behavior is eliminated. Manipulation of rewards is different from administering an aversive form of punishment like positive punishment (see below).

With negative punishment, whether it be through the extinction of a behavior (by removing a reinforcement), time out, counter-conditioning with an incompatible behavior, etc., it is very hard to eliminate the behavior.  There is also a time delay before a trainer may see a reduced frequency of the presentation of unwanted behaviors.

Positive Punishment — A negative consequence occurs following undesirable behavior like correction from an electronic collar, leash jerking or the administration of citronella spray, etc. There is little time delay in achieving the desired result of eliminating unwanted behaviors.  However, there are inherent dangers with the misuse of aversive control aids/agents. In addition, some trainers like Dr. Pamela Reid, whose work is cited on this page, assert that the biggest danger to using aversive control aids like electronic collars is abusive use from “trigger” happy trainers. Remember that it is up to trainer’s discretion whether or not positive punishment should be administered through aversive control methods.

Positive Reinforcement — Something desirable is offered when a desirable behavior is initiated like a reward being given for responding to a recall command

Negative Reinforcement — Something undesirable is removed when a desirable behavior is presented. For example, when a dog finally responds to a recall command corrections from an e-collar cease

ADDITIONAL TERMINOLOGY:

Classical Conditioning (Associative Learning)

Counter-Conditioning — conditioning of a new and incompatible behavior. Sometimes when a trainer attempts to extinguish a behavior (see below), they counter-condition the behavior to sustain the extinction or elimination of the unwanted behavior.

Conditioned Reinforcer (bridge) — links the primary reinforcer (e.g. food, tactile stimulation) to the cued behavior in reward-based conditioning. Essentially this bridge links the reward to the behavior performed. “Loading” the conditioned reinforcer or “bridge” refers to making the first association between the conditioned reinforcer and the reward. A bridge such as a whistle, clicker or the word “good” can be generalized to multiple types of primary reinforcers or specific reinforcers.