Real Training vs. “Operant Conditioning”  - by Roger Hild

In recent years many dog trainers have jumped on the “Operant Conditioning” bandwagon.  The practitioners of this variety of “OC” are quick to make scientific claims about how dogs learn.  Unfortunately, there is nothing new nor scientific in their assertions or methods.  There are more efficient and effective training alternatives available.

To illustrate, allow me to introduce you to Jane and her dog Mocha (not their real names of course).  Mocha is a very young, neutered Chocolate Labrador Retriever.  Jane is a small woman of slight build.  She owns her own business, a shop in town, and Mocha comes to the shop with Jane every day.  When Jane contacted me she was very nearly at wits end and was seriously considering getting rid of her beloved Mocha. She could no longer tolerate Mocha's misbehavior.

When I arrived at her very lovely home, I was told that Mocha was tied out back.  Lately, Jane was keeping him tied a lot and Mocha only got to come in the house to be put directly in his crate.

When Jane let Mocha inside, he'd rush wildly about the house jumping on furniture and destroying things.  If she had guests, the dog would pester them and jump repeatedly on them.  Jane's arms were covered with scratches from Mocha's nails and teeth.  If Mocha wanted something while Jane was busy, he would start nipping her until she stopped what she was doing and paid attention to him.

Whenever she refused the dog something or tried to direct him away from anything he wanted to do, he’d become quite “saucy” with “bark-back,” nips and snarls. Jane could no longer walk him because he was way too strong and pulled her down the street.  She tried to use a head halter but Mocha didn’t like it and fought her so strenuously she had to put it away.  Some of the recent scratches were head halter battle scars. 

I was not the first dog trainer Jane had worked with.  Mocha had gone to a “Puppy Kindergarten” program and graduated at the top of his class.  He been through two socialization classes and (according to the class instructors) had done well.  He had also been through a basic dog-training program.  These programs were from two different schools; both were based on positive reinforcement only and utilized the  “clicker” as their primary tool.  This dog could (when in the right mood and if a clicker and food were present) perform tricks to rival any circus animal.

I wanted to know how long these behaviors had been present and if she had contacted her previous trainers.  Jane told me she had tried to get help from the very start of Mocha’s training classes. At first, she’d been told, “this is just a puppy and all puppies behave like that.”  She was told that under no circumstances was she to scold, correct or punish him.  She was supposed to reinforce Mocha's when his behavior was good and ignore him when he behaved badly.  If he jumped she was to walk away from him and/or remove any attention he was getting.  If she was working with treats or toys, she was to withhold them if he started to act up.  If he pulled while walking, she was to stop and not move until he stopped pulling.  She had repeatedly tried these measures with minimal success and felt the trainers weren't hearing her frustration.

After listening to the first two minutes of her story, I could have listed the rest of it off by heart - I’ve heard it so many times before.  After doing everything with Mocha that she had been instructed to do, after trying all the  “OC” solutions, why were things between Jane and Mocha so bad that Jane was seriously considering terminating the relationship?  What she got from these schools and their trainers is an example (alas, an all-to-common example) of the advice being handed by many who have boarded this “OC bandwagon,” and illustrative of just how destructive this watered down training can be. 

Let’s briefly leave Jane and Mocha (I shall return to them shortly) and take a look at what Operant Conditioning is.  True Operant Conditioning is a reasonably balanced learning theory.  It is a four-quadrant model that attempts to explain learning in terms of the consequences related to an action.  Within that 4Q model are the different contingencies of positive and negative reinforcement as well as positive and negative punishment.  Those who would claim to subscribe to the OC learning theory but who only are willing to work with the “reward” side of the equation are in reality practitioners of a bandwagon variety of OC that might more properly be called “PROC” or Positive Reinforcement Operant Conditioning.  Where true OC offers a reasonable chance of success through balance, “PROC” is a very protracted and unrealistic method for attempting to train.  The results are mediocre at best and more often simply disappointing.  The motivation for “PROC” is not better training but a philosophy (sometimes pursued with almost religious zeal) of abolishing all painful life lessons.  It is indeed unfortunate that this almost narcissistic need (for a “warm and fuzzy,” feel good above all else approach to life) gets packaged and marketed as “animal-friendly” or “more humane.”  It is a selfish approach designed to place the trainer's need to feel good above the learning needs of his student.

In psychological terms, conditioning means, “causing an organism to exhibit a specific response to a stimulus.”1   The conditioned response must be specific, reliable, highly predictable and reproducible.  Any response (other than the “conditioned” response), any randomness or any failure to respond correctly, must be accounted for and explained.  As I said earlier, there are several ways to cause the sought after response, using positive and negative consequences.  By definition conditioning, particularly Skinners Operant Conditioning model, does not acknowledge or take into account any internal events such as thoughts, feelings, or motivations and therein lies its weakness.  If these internal events are not acknowledged as contributing to the conditioning of the behavior, they cannot then be used to explain any “conditioning failures.”  Conditioned performances, (particularly utilizing only positive reinforcement) while often improved, are often not the best that one would hope for or expect.  When performance falters (as it frequently does) MORE CONDITIONING WILL NOT SOLVE THE PROBLEM whereas addressing some of the internal factors (excluded by OC) or looking at relationship related issues, very often does.

The main problem with the theory of conditioning (and particularly with Operant Conditioning) is in the understanding and application of the learning process.  Operant Conditioning is simply one kind of conditioning (made famous by B F Skinner) which seeks to explain all behavior and learning in terms of the associations made between responses to stimuli and the resulting consequences.  Although behaviorists believe all thought processes can be accounted for through associations of stimuli and responses, other psychologists strongly reject such an approach as inadequate to explain many kinds of behavior.

“Real Training,” on the other hand, addresses the whole dog and not just the behavior.  Along with utilizing all four quadrants found in the Operant Conditioning model, it also seeks to deal with all those areas that behaviorists refuse to acknowledge (such as choice, motivation, drive, and various mental/emotional processes).  It acknowledges that, in addition to or regardless of, any conditioning, dogs make decisions and sometimes become contentious.  Real training is about working with the dog to teach him what is expected – what choices to make and how to behave.  It holds him accountable for the choices he makes.  It acknowledges there is a difference between knowing and doing and that difference can sometimes represent a point of contention, rather than a lack of conditioning.

Training (for me) is as much about the interactive dynamics between student and teacher as about what is being taught.  In the learning process (some of which will be a conditioning process) the student also learns about the teacher.  Often the emerging interpersonal dynamics will influence subsequent behavior far more than any single training or conditioning sequence.  At some point, tasks will be performed as taught because a choice has been made to do so - not simply as the result of some stimulus-response reflex (read conditioning) action. 

On the other hand, behaviorism and its tool (Operant Conditioning) is, “neglected by cognitive etiologists and ecological psychologists convinced that its methods are irrelevant to studying how animals and persons behave in their natural and social environment,” according to a recent article on behaviorism published in the Stanford Encyclopedia of Philosophy.  Also of note in the same article: “The deepest and most complex reason for behaviorism's demise is its commitment to the thesis that behavior can be explained without reference to mental activity. Many philosophers and psychologists find this thesis hopelessly restrictive. They reject behaviorism because of it....”2
 

Nature vs. Nurture vs. Nomenclature or Excuses, Excuses, Excuses 

When a “conditioned behavior” begins to break down an explanation is called for.  Since conditioning, by definition, means a specific response to a stimulus, any unreliability that creeps into the conditioned response, must be addressed.  For example, if a dog has been conditioned to sit (specific response) to the cue word “sit,” (specific stimulus), an explanation is required any time the dog fails to sit on cue or does anything else but sit.

In the early 1960’s Keller and Marian Breland published THE MISBEHAVIOR OF ORGANISMS. Marian Breland had studied under Skinner during the war years and met her husband Keller
working on experiments to train animals for the war effort.  Marian became very well known and was one of the first psychologists to utilize operant conditioning commercially.

The Brelands attempted to explain why breakdowns occur in conditioned behaviors. 
“Our first report (Breland & Breland, 1951) in the American Psychologist, concerning our experiences in controlling animal behavior, was wholly affirmative and optimistic, saying in essence that the principles derived from the laboratory could be applied to the extensive control of behavior under nonlaboratory conditions…”

“Emboldened by this consistent reinforcement, we have ventured further and further from the security of the Skinner box. However, in this cavalier extrapolation, we have run afoul of a persistent pattern of discomforting failures. These failures, although disconcertingly frequent and seemingly diverse, fall into a very interesting pattern. They all represent breakdowns of conditioned operant behavior…” 

“These egregious failures came as a rather considerable shock to us, for there was nothing in our background in behaviorism to prepare us for such gross inabilities to predict and control the behavior of animals with which we had been working for years.

The examples listed we feel represent a clear and utter failure of conditioning theory. They are far from what one would normally expect on the basis of the theory alone. Furthermore, they are definite, observable; the diagnosis of theory failure does not depend on subtle statistical interpretations or on semantic legerdemain - the animal simply does not do what he has been conditioned to do.” 3
They go on to label this breakdown as “instinctive drift.” 
 

The “Operant Conditioning” referred to here is the same as the “PROC,” I I referred to earlier.  The conditioning was primarily positive positive reinforcement and shaping along with “negative punishment,” (a euphemism for removing the rewards: “all bribes are off”).  Blaming instinct for the breakdowns presents an interesting theory but it doesn’t answer some important questions:

- Why, for instance, would the sought after behavior “condition” in the first place if “instinct” would dictate otherwise? 
- Why is this “misbehavior” first described only after the animals are moved into the “real world,” and out of the sterile, boring environment of the behaviorists’ laboratory?
- Why call it instinct at all - there's almost nothing that animals do that is entirely genetically programmed, which is the usual requirement of an instinct.


Because of the unreliability of their training approach, PROC practitioners have embraced “instinctive drift” and any other convenient terminology/excuse to explain away training failures.  Thus, after the fact explanations of failure are replacing effective Real Training.  Management advice to cope with unacceptable behavior is becoming more and more the norm and is turning our homes into modified zoos for semi-wild canine companions. 

Both classical and operant conditioning are valuable concepts in training and all trainers use the techniques to one degree or another.  Where Real Training parts company with the Skinner Behaviorists is over their belief that conditioning is the ONLY way of learning.  Their belief that if a dog does not perform some command, then the dog is not properly conditioned is also a mistake.  They do not allow for the possibility that a dog can “refuse” a command. 

Dogs learn from life experiences.  From each behavior, there is a net outcome realized.  ALL consequences, even the painful ones, are a necessary part of that process.  In trying to eliminate undesirable behavior, one can try teaching “alternative incompatible behaviors,” counter-conditioning or one can try desensitizing (sort of the opposite to conditioning) in order to try stopping certain behaviors.  These procedures are usually quite tedious, protracted and the results are inconsistent:  NONE OF THESE PROCEEDURES TEACHES THE DOG “NOT TO.”   Punishment (the term as used by behaviorists) is by definition the only consequence where the lesson is what “not to do.”  Applying a suitable punisher will stop the unwanted behavior (teach the “not to” part of the equation) and then, providing for an alternative behavior which is positively reinforced will get you to where you want to go much faster than not providing punishment.

N.H. Azrin and W.C. Holz’s  research demonstrated that when there was no punishment, the behavior continued unabated.  “When the punished response was the only way to get what was wanted, punishment produced a moderate suppression of the behavior.  The availability of an alternative response for obtaining a positive reinforcement greatly increased the suppressive effects of punishment.4 
 

Now back to Jane and Mocha:

At my request, Jane brought Mocha into the house on leash.  I had already discussed the possible actions I might take, depending on how the dog responded, and she agreed and was noticeably relieved when I suggested a much firmer hand would likely be necessary.  Mocha came in like the proverbial “bull in a china shop,” jumping, mouthing and totally out of control.  Fending the dog off with the leash made no difference, as did nothing Jane said or did.  Jerking the leash also seemed futile.

I had Jane hold Mocha while I put on a more suitable training collar (in this case a medium pinch) and then I took the leash.  On his very next jump, I calmly said “No,” and snapped the leash to the side – the effect was immediate and Mocha stopped jumping and quickly sat.  Using calming words and mild praise, I then started to walk him up and down the hallway.  Next, we went in the living room and sat and talked while Mocha was kept quietly on the floor at my feet.  Midway through, I transferred the leash to Jane and she was able to correct or praise (what ever was necessary to reinforce the calm).  Mocha was then taken for a walk and after less than 30 seconds, he seemed to understand the new rules and he walked like a “perfect gentleman.” 

Due to her previous PROC trainers’ advice, Jane (who was naturally assertive in many areas of her life) had been inhibited in her dealings with Mocha.  I encouraged her to be more assertive in her relationship with her dog.  Jane learned a few training techniques to gain compliance, how to use consequences appropriately, and how to use her praise and approval as rewards.  Jane was more than happy to follow my recommendations and once she made the new rules consistently clear to Mocha, he seemed much happier as well. 

Real dogs live in real families in real communities and perform real functions; they are not the theoretical dogs and laboratory rats that inhabit the behaviorist’s textbook.  Real dogs don’t need explanations or excuses, they don’t need ‘management only approaches,’ real dogs need real training. 

__________________ 
1 "Conditioning," Microsoft® Encarta® 97 Encyclopedia. © 1993-1996 Microsoft Corporation. All rights reserved. 

2 Graham, George; "Behaviorism", The Stanford Encyclopedia of Philosophy (Fall 2002 Edition) 

3 “THE MISBEHAVIOR OF ORGANISMS,” Keller Breland and Marian Breland (1961) 
Animal Behavior Enterprises, Hot Springs, Arkansas.  First published in American Psychologist, 16, 681-684. 

4 W.K. Honig (Ed.) “Operant Behavior: Areas of Research and Application,” published by Appleton Century Crofts, New York, 1966.
 
 


[Other Articles of Interest]

[Back to Home Page]