Catching up. Notes about Behavior/Training.
November 6, 2012 at 6:23 pm
It has been a crazy month – major computer issues, health issues, catching up on a huge backlog of work – and poor Boogie’s Blog has been neglected. I don’t have time to go into too many details, so this blog post is a quick summary of stuff that I have found interesting.
1. Two great books!
First of all, I am plugging Grisha Stewart’s Ahimsa Dog Training Manual - I am doing new illustrations for the next edition!
Paul Chance’s First Course in Applied Behavior Analysis is about modifying behavior in people, not dogs, but the philosophical premise and methods are the same as what are used in modern dog training methods, and I enjoyed learning about the same concepts in “human contexts”. Some quotes:
ABA is concerned with using environmental events to change behavior in desirable ways.
Much of the time, a behavior problem means either that a behavior occurs too often or that it does not occur often enough. The task of the parent, teacher, manager or therapist is to increase or decrease the frequency of the behavior.The chief difference between the people who live inside mental hospitals and those who live outside of them is not that we never behave as they do but that we behave as they do less often than they do.
One of the great things about ABA is that is focuses on what people can do rather than on a label or on some mysterious, unseen psychological disorder.
Please note that reinforcement strengthens BEHAVIOR, not PEOPLE. Everyone slips up now and then and speaks of reinforcing a person, as in, “John was studying very hard so I reinforced him”. You don’t increase the strength of people with reinforcement; you increase the strength of their behavior.
Some people think that people with Ph.D’s go around thinking up new terms for everyday words just so what they say will sound more sophisticated. There may be some truth to that, but the word reinforcer is not just a synonym for reward. Here’s the essential difference: Rewards are defined by consensus; reinforcers are defined by results. If an event strengthens or maintains the behavior it follows, it’s a reinforcer; if it doesn’t it isn’t a reinforcer. There is no other defining characteristic of a reinforcer.
This is an important point; one students often miss. People often get the impression that a behaviorist is someone who carries a bag of reinforcers about and whenever behavior needs strengthening, he reaches into the bag and hands out reinforcers. The essence of behavior analysis is not handing out reinforcers but analyzing the effects of antecedents and consequences on behavior. That includes identifying consequences that are reinforcing.
2. The DogRead Yahoo Group
I am subscribed to the DogRead group by email. Every month a new book/author is featured readers can interact directly with authors, ask questions etc. This week, the amazing Kathy Sdao is on DogRead. Kathy Sdao used to train dolphins for the Navy, and I almost included this example in my poster. I love Kathy Sdao’s response to this reader question:“is it always possible to use only positive, gentle reinforcing training methods with dogs?”
My experience has taught me that we tend to presume a false correlation between reliable behavior & the use of aversives. IOW, it seems to make intuitive sense that if the animal “must” do something really important (e.g., locate a bomb, track down a criminal, guide a blind person), we have to use some sort of force to teach the animal he doesn’t have a choice.
Yet, I believe we revert to coercion and force when our skills at implementing structured, clever, careful reinforcement procedures run out. The dolphins I trained for the US Navy reliably located deep-moored mines (as much as 600 feet down), reported the mine’s presence to their trainer on a small boat, then carried a heavy packet of explosives back down to the mine, attached this in a quite specific location on the mooring cable, then swam back to the small boat and leapt onboard for the long ride back to their pier-side pens. This behavior chain was long, cognitively and physically challenging, performed in the presence of huge distractions (including live food-fish swimming all around the work site), and ended with each dolphin choosing to go back to captivity.
It is beyond the scope of this discussion of my book (Plenty in Life is Free) to further discuss details of this training, except for the critical fact that we did not use “corrections.” We obtained accurate reliable real-world performance through the use of careful design of the training environments and tons of R+ over the course of many months of training. Punishment was not part of the program.
When folks say “yeah, but those were dolphins and they’re really smart,” I respond that dogs are every bit as smart. It’s just that their trainers are more often seduced into believing that punishment is necessary to teach “the important stuff.”
My affiliation (as an occasional consultant) with Guide Dogs for the Blind (GDB) over the past few years has allowed me to see how amazingly effective positive-reinforcement training can be when done by skilled and creative trainers. Their implementation of clicker-training over the past ~five years has been amazing to watch. Their results have been so impressive that guide-dog organizations all over the world are seeking GDB’s advice on how to modify their own programs.
Hanging by my desk is this quote, from one of my all-time favorite books, Dr. Murray Sidman’s Coercion & Its Fallout (2001):
“An overworked and incorrect bit of folk wisdom pronounces the carrot to be of no avail unless backed up by the stick. But the carrot can do the job all by itself.”
… which is the point of this poster illustration -
[click on pic for more info]
Entry filed under: Art, Articles, links, Books & DVDs, Reads, Training. Tags: .