Science-Based Dog Training (with Feeling) - Day 2
Day 2 - Transcending Reinforcement Schedules, Going Way Beyond “Dominance” and Punishment and Rekindling the Relationship
This seminar presents many exciting new concepts: How, by using our voice, we can transcend laboratory-generated learning theory. How to better motivate dogs for happy and willing compliance. How to use life rewards, interactive games and even “behavior problems” as rewards. How to empower mega secondary reinforcers. How to effectively reduce and eliminate undesirable behavior and non-compliance by using just our voice and not even having to raise it. Essentially, how to make training easier, quicker, more enjoyable and much more effective.
Understanding the practical application and constraints of the principles of learning theory is essential for continually improving the reliability, speed and quality of responses. If we don’t effectively apply these basic principles, clients and their dogs will have difficulty learning. And if dogs are slow to learn, their owners become frustrated and often resort to physical and/or aversive “punishment”.
Learning Theory was laboratory-generated from studies in which computers trained captive rats and pigeons using food and shock, which, of course, is a very different from people training dogs off-leash and at a distance, i.e., with many distractions and with many options. Not surprisingly therefore, when people train dogs, learning theory has many constraints in practice; the more powerful reinforcement schedules are impossible to implement and punishment is notoriously ineffective and hardly sufficient for what we need to accomplish.
Continuous, fixed interval and fixed ratio reinforcement schedules are pretty ineffective at maintaining reliable responding and make it difficult to phase out food rewards. Variable interval and variable ratio schedules fare better but none of these schedules improve the quality of behavior. More relevant though, few people can compute ongoing variable schedules and so, they are unusable.
Similarly, whereas punishment works extremely well in the laboratory when consistent computers shock rats in cages (bored and with no option of escape), punishment is largely ineffective in the real world, wherein inconsistent humans administer ill-timed aversives in an attempt to control the behavior of dogs that have many other interests and hobbies (distractions) and are usually off-leash (with options to escape, or to convince the trainer to escape). In the real world, “punishment” rarely reduces undesirable behavior. Instead the dog learns to reserve misbehavior for those times when it cannot be punished, i.e., when the owner is absent, or not paying attention. But if aversive stimuli do not effectively reduce and eventually eliminate undesirable behavior then, by definition, they cannot be defined as punishment — stimuli which reduce the immediately preceding behavior such that it is less likely to occur in the future. On the contrary, depending on their severity, most aversive stimuli are merely harassment or abuse.
These days, many trainers are trying to emulate computers and indeed, a great deal of consequential feedback is given in the form of quantum clicks and kibble and jerks and shocks, wherein instructiveness (effectiveness) depends almost entirely on precise timing and consistency. But people simply do not have the tireless consistency, impeccable timing, or the computing power of computers. Luckily though, many people are smarter and more creative than computers. Didn’t we invent computers? Consequently, we can use differential reinforcement schedules that not only maintain reliable levels of responding but also, cause the quality of behavior to continually improve from trial to trial. Additionally, we can effectively eliminate undesirable behavior without hurting or frightening the dog.
Indeed, when we use our voice as feedback, we may easily transcend many of the constraints of being a computer and radically increase the speed and effectiveness of learning. Binary verbal feedback is both instructive and analogue. In addition to communicating whether behaviors are desirable or not, analogue feedback communicates how well the dog did, or the relative seriousness or potential danger of non-compliance.
Rewarding dogs for desirable behavior is pure and simple and extremely effective. However, we must also teach owners how to effectively reduce and eliminate undesirable behavior and/or potentially dangerous behavior. What are they expected to do when their dog is about to run across the street, or jump on a child? Ignoring undesirable and dangerous behavior just doesn’t work in practice. Instead, owners become frustrated and dogs are “punished”, often by unpleasant means. This, I think, explains why we have seen a resurgence of aversive training techniques during the past decade.
People simply do not know how to eliminate undesirable behavior and of course, effectively punishing dogs for undesirable behavior, even when using non-aversive “punishment”, is woefully insufficient. Inhibiting undesirable behavior is just one part of the puzzle. When dogs are non-compliant and misbehave, our prime directive is to get the dog back on track as soon as possible as well as indicating the urgency for compliance (or the potential danger of non-compliance).
Punishment is what laboratory computers had to do because they are not human. Computers could not explain the task or the solution: all they could do was shock when animals made incorrect choices — breaking rules that they didn’t know existed. We on the other hand have language and in a single word can communicate three pieces of information: 1. The present behavior is undesirable. 2. The precise nature of the specific desirable behavior and 3. The potential danger of non-compliance.
Verbal feedback is instructive and effectively communicates how the dog may get back on track a.s.a.p. A Specific Redirection acts like punishment (by decreasing or eliminating undesired behavior) but also immediately prompts the desired behavior. Verbal feedback is by far the very best means to correct undesirable behavior and achieve compliance without being aversive, without causing fear or pain and without even raising our voice.
By using softly-spoken verbal feedback, we may simply transcend the training ability of any computer, or of any technician trying to emulate a computer. Our language, our voice and our tone can communicate so much more than any click, food treat, leash jerk or shock. By all means let’s use food lures, clicks and food rewards but we must phase them out and check that our dogs will respond promptly and reliably to verbal commands without them, when off-leash and when distracted.