“NO!” and unwanted behavior. (Notes from Clickerexpo Part 2)
August 21, 2013 at 5:50 am
*This blog post was written a few months ago, following on from Part 1.
From “How To Live With a Neurotic Dog” (1960)
“Eli, No!” – a picture book (2011)
The pictures above are from two books that I bought only for the artwork. The first book is from the 60’s; and the second book is recent.
I think we live in exciting times because there is a cultural shift in thinking about dog training, dog behavior and human-dog relationships, and there is now more information available on the internet and in books on understanding why dogs do what they do and how they learn, and adjust our training methods to be smarter and kinder.
From “How To Live With A Neurotic Dog” (1960)
Traditional dog training, being based in punishment and behavior suppression put a lot of emphasis on the “No!’s”.
To (loosely) quote Sarah: “There are people who want their dogs to be seen and not heard and they believe that this is how the relationship should be between human and dog. And then there are those of us who want our dogs to be able to express themselves, communicate, initiate things, and feel empowered in a relationship”. The inspiring thing about Clicker Expo is that everyone is there to focus on the “Yes’s” – how to train in ways that encourage more behaviors, not less.
At the January Clicker Expo, I went to three Ken Ramirez seminars. Ken Ramirez is an expert trainer at Shedd Aquarium who has worked with many species of exotic animals and all his training stories (mistakes AND successes) were very insightful and heartwarming.
In his seminar on “What To Do When The Animal Makes Mistakes-“, “No!” is technically referred to as a No Reward Marker or Punisher.
Ken Ramirez said that he will not judge people who use punishment or negative reinforcement in their training methods, but he personally finds no need to say “No” to any student animal that he is training. He is able to accomplish any training goal with Positive Reinforcement.
“We are more creative trainers if we don’t have a way to say NO. You don’t want to say NO to an elephant.”
To Ken, trust is one of the most important aspect of any training plan, and what defines a good relationship between trainer and trainee is a strong positive reinforcement history.
One of the most common trainer mistakes is requesting a difficult behavior from an animal when he/she is not yet fluent in that behavior. This messes up the trust relationship. He shared several examples when trainers got too greedy and asked for too much too soon. Trainers might get too caught up in their egos, push the animal too far, the animal is unable to do the behavior he was trained to do (or doesn’t feel totally comfortable with it yet), has some sort of breakdown, stops doing that behavior… and then there is a lose-lose situation where it can take YEARS to re-train that behavior.
Ken Ramirez outlined the various ways that we have learned to deal with unwanted behavior. He cited the 8 methods in Karen Pryor’s classic book “Don’t Shoot The Dog” and agrees with Karen Pryor that “Changing the Motivation” is the most humane positive method. However, this is not always the most practical solution. There has to be a more immediate way to deal with unwanted behavior or mistakes.
Common ways of changing unwanted behavior that are aversive to the animal in some way:
PUNISHMENT – self explanatory. We add something aversive to make the behavior stop. The animal does not learn what is right. Punishing is addictive too to the trainer, and can get out of control. There is a high risk of fallout and loss of trust in the relationship. Dogs as a species may be more forgiving than killer whales but this is beside the point. Physical punishment works but is the least humane and most intrusive method.
NO REWARD MARKER (NRM) – basically a word like “No”, or “Wrong”, or “Oops!” or “Stop it”. Which to Ken, are essentially still punishers (Conditioned Punishers or Secondary Punishers). Even if we use “Oops!” as a warning (“Watch out, this is your last chance or else…”) if we lack self-control and use it too frequently, it could definitely become something aversive and will damage the relationship between human and animal. Even the kindest version of a “No” when overused will lead to frustration.
TIME OUT or Negative Punishment (eg, the trainer leaving the room with all the food, when the animal misbehaves) This is another response that causes frustration and anxiety, mainly because the animal does not have a clue which behavior he did was wrong and does not learn what is right. The information is not clear. Time Outs are also ONLY effective if the animal likes you in the first place or finds the training reinforcing. The animal may be relieved to be away from you.
KR gave a great example of a classic Time Out mistake: Dog does something ‘naughty’, person picks up dog, puts him in the crate and person leaves the room.
NEGATIVE REINFORCEMENT = Training a behavior where the animal gets to avoid or escape something aversive. Ken Ramirez says that even though Negative Reinforcement always involves some sort of aversive, this doesn’t mean it is inherently evil. For example we use alarm clocks (aversive noise) to wake us up in the morning. We hate the noise but it makes us wake up. Another example is when we learn to turn off the car lights because every time we forget, there’s that loud beeping noise. The biggest issues with Negative Reinforcement are that overuse will lead to frustration and anxiety. The severity of the aversive can be hard to control and can become inhumane. It takes a lot of skill to use Negative Reinforcement effectively.
I am thinking of training protocols for reactivity/aggression like BAT , which include some element of Negative Reinforcement. To paraphrase what Grisha Stewart said at the BAT seminar – yes, there is the use of an aversive in BAT set ups (presenting the dog with a scary dog/person) but we can’t avoid this aversive in real life. “It’s not like we can sit on the couch with our dog and calmly chat with him about his fear of other dogs”. I see that the humane use of Negative Reinforcement involves working in a controlled environment, not adding any artificial aversives, and always letting the dog feel safe and in control.
I like these Susan Friedman quotes:
Control the environment not the animal.
Control is a primary reinforcer, to deprive an animal of control is akin to depriving them of water, food.
To the greatest extent possible all animals should be empowered to exercise personal control over significant environmental events.
In addition to whether or not a method is effective, we have to consider what is the least harmful or least intrusive technique for teaching or changing behavior, and Ken Ramirez referenced the “Heirarchy of Effective Procedures” chart by Dr. Susan Friedman.
Here is Dr. Susan Friedman’s original article which is a must-read.
Example: If your dog runs around in the yard all day and growls at people on the street, you could change his environment/antecedent arrangements (level 2) – bring him indoors, or put him in the backyard out of view of the street – this would be less intrusive than actively changing his behavior with reinforcers or punishers (levels 3-5).
Ken Ramirez’s favorite least intrusive method for dealing with mistakes is what he calls the LEAST REINFORCING STIMULUS (LRS).
The LRS method was developed and used in zoos but it is also useful with pets. Top priority:
1. not reinforcing unwanted behavior, and
2. not adding any stress or frustration to the relationship.
This is how I understand it.
1. First of all, there must be already a strong Positive Reinforcement history (ie, good relationship) and high rate of reinforcement.
2. When the animal makes a mistake or does the wrong behavior (not the one you asked for), be NEUTRAL for 3 seconds. Stay calm and DO NOTHING for three seconds. One thousand, two thousand, three thousand…
3. Immediately ask for another behavior that you know is easy, that the animal can do. Give reinforcement.
Assuming that 3 seconds is the right amount of time, the mistake/unwanted behavior won’t be reinforced, and it won’t be too long of NOTHING HAPPENING for the animal to develop frustration.
Emily Larlham (advocate of Progressive Reinforcement training) reiterates that the problem with “No!” and other conditoned punishers is that:
- it suppresses your dog’s behavior (overuse leads to a shut down dog)
- you create bad associations for your dog with yourself, your dog will do bad behaviors when you are not around.
Emily’s method of dealing with unwanted behavior is by using a Positive Interrupter – a sound (eg, kissy noise) and conditions this sound with a treat &/or petting, so that whenever the dog does an unwanted behavior, she uses the noise to redirect the dog away from doing the unwanted behavior, and then asking for a desired behavior that can be reinforced.
**IMPORTANT WARNING: Always give attention to your dog when he is doing good behavior and reinforce this good behavior. The Positive Interrupter is “attention” so if you use this ONLY when dog does unwanted behavior, then your dog will purposely repeat bad behavior just to get your attention.
Oh yes, I learned this the hard way 🙂
How Boogie learned that barking is awesome.
Entry filed under: Articles, links, Events, Reads, Training.