Tuesday, December 29, 2015

Pure Positive

The biggest problem with the “Pure Positive” training movement is that they refuse to discuss how to use corrections (punishments).  All corrections are lumped into that giant emotional trash bin called “abuse”.  The Pure Positive folks completely rule out corrections in training, despite the fact that the quadrant of operant conditioning includes positive punishment and negative reinforcement, both of which involve pressure on the dog.  The pure positive movement preaches like the abstinence only sex education classes. 

I think the abstinence only sex ed is a disservice to young people.  I think pure positive training is a disservice to pet owners and pets.  We need to talk about corrections.  We need to accept training in the pure punishment quadrant.  MOST IMPORTANT, WE NEED TO TEACH PEOPLE HOW AND WHEN TO USE CORRECTIONS.  Because if we don’t then someday Fluffy is not going to come when called, or nip the neighbor’s kid, or jump all over mom who is dressed to go out, and it won’t go well for Fluffy.  Those people who you lectured at length on the evils of correction are now frustrated with Fluffy, sure that all correction is abuse, and far more likely to be abusive than if the proper use of corrections had been discussed and practiced.  They’ll correct too hard.  Because Fluffy has never been corrected she’ll see it as a personal attack rather than a result of a behavioral choice.  Then the owners will feel guilty and not correct her the next time.  Reinforcements work on a variable schedule, punishment does not.  You need to be consistent, correcting essentially every time.  But Fluffy’s owners don’t know that.  So Fluffy can’t figure out the pattern and know what to do to avoid the correction.  Fluffy’s owners get more frustrated, and probably correct Fluffy again, harder this time.  Why, because their instructor never taught them that corrections must be consistent, and very seldom need to be harsh.  Their instructor never warned them that escalation of corrections is not a substitute for consistency of corrections, and indeed escalating corrections without being consistent is abusive because Fluffy never figures out exactly what to do to avoid the corrections.  All the instructor ever said was that corrections are abusive, don’t go there.  It doesn’t work for sex ed and it doesn’t work for dog training.

First, let’s get over this abuse thing.  Properly done corrections are not abuse.  They are feedback on a behavioral choice.  And if you say that pet owners won’t correct properly, well whose fault is that?  Since the pure positive movement won’t allow any reasonable discourse on corrections then there is no learning.  Corrections are everything from a pop on the leash, a quick scold to blocking the dog from its goal, etc. 

I’m not even sure the pure positive folks even let themselves learn the basics of correction:

1.       Correction must be consistent, think V1 (if you really are a positive trainer you should know what that is)

2.       Correction needs to be adjusted so that the dog thinks about it.  For some dogs a hard word or body block is enough.  For a snarky pup pinning it to the ground until it gives up is appropriate.  For a pup that loves to bite and is not responding to being redirected to toys, scruff it or give it a smack, hard enough that the dog stops and thinks about it.  The dog should neither run away nor launch back into the behavior. 

3.       Sometimes, in the case of interrupting a dog mentally involved in its behavior (example a keen young dog working stock) you may need to be fairly harsh to get the dog’s attention.  Once you have the dog’s attention you may need to back way down.  

4.       Do not correct when you are angry.  Indeed better to nip problem behavior before you get totally frustrated with it.  Many folks avoid correction, allow the problem behavior to get established, then not only are they frustrated but the behavior is much more difficult to stop.

5.       You need to be calm and quick when delivering a correction.  No emotion, no tantrums.  Sort of the same as 4, but very important.

6.       Correct at the moment the dog is bad, then stop.  The dog learns from when you stop the correction.  If you continue to berate or correct the dog after it has complied then the dog has no chance to figure out what it is you wanted. 

7.       Timing is everything, just like reinforcements.

8.       Use reinforcement when you can.  Recognize where correction is a better option (simple single behavior you want to eliminate) or the only realistic option (you don’t control the primary reinforcements)

9.       For heaven’s sake when the safety of your dog, or someone else’s dog, or your cat, or the sheep are at stake don’t stand on your pure positive pedestal and think you are some almighty savior because you did not administer an aversive.  While the sheep are terrorized in the corner or the cat is living in the basement you are responsible.  No Pontius Pilate washing your hands clean for you.

10.   If you don’t control the primary reinforcements you are not going to solve the problem with positive training.  To train with reinforcements you need to be able to control the dog’s access to those reinforcements.  Much more difficult in the real world than the training class.  Yes, superb and determined positive trainers can proof a behavior extensively so the dog never thinks about the distraction, but the Jones family is not going to do that with their Lab.  And even that fails when you are training a dog in something self-reinforcing like working livestock.  Yes, you can still use positive reinforcements, but you will need some form of corrections to keep the work in a zone where you have something to reinforce.

I’ve seen quite a few trainers who do incorporate corrections looking for euphemisms to describe corrections.  Let’s not change the word.  Corrections fall into the pure punishment quadrant, nothing wrong with that.  Let’s keep the language clear rather than coming up with new terms to avoid ruffling the clicker cult.

For those who now picture me training with a whip and chair, I have lots of clickers.  I train with clickers.  I went to “chicken camp” with Bob and Marian Bailey and loved it.  I had a great conversation with Bob Bailey about the challenges of training in a self-reinforcing environment.  I love training with positive reinforcement.  I’ve brought in a clinician for a clicker seminar I put together.  I’ve gone to other clicker seminars, including travelling quite a ways to attend.  I put lots of agility titles on dogs using positive reinforcement and my trusty clickers back when I was doing agility.  But these dogs did get corrections for certain life behaviors.  I live on a busy road.  Dive past me to get out the front gate and you will be corrected, every single time.  My dogs work stock.  I don’t control all the reinforcements.  I use corrections to keep the work in a zone that can be reinforced.  I used to foster retired racing greyhounds.  I have cats.  Those dogs wore a muzzle and were kept on a leash when they came in my house.  They were corrected for any intent on the cat, corrected hard for trying to go after it.  Within a day or two a cat would walk in the room and the dog would look away…click/treat/praise.  Several of these dogs who were quite aggressive with cats were able to be placed in homes with cats and live in peace together.

No comments:

Post a Comment