It is a great privilege to welcome guest blogger Randi Rossman! This article was originally posted on the Fearful Dogs Facebook page.
“All CC/DS does is bribe the dog.”
I hear this a lot lately. And not just from typical pet owners, but from professionals as well.
I’ll say this – if a dog training professional says that CC/DS is cookie tossing or “bribing” the dog, please hire another professional. They do not understand the science of behavior. It should be regarded as big of a red flag as going to the doctor with a pus-oozing sore and having the MD say, “Antibiotics will only bribe the germs.”
Hopefully by the end of this post you’ll agree. Warning, I use geek talk. Most of you are already more knowledgeable about what respondent behavior is and how to do counter conditioning than most professionals, so are ready for some geek talk. Geek talk helps create more precise definitions around what we are saying.
First – what is respondent/classical behavior?
Second – what is respondent /classical conditioning and counter-conditioning?
Third – how does respondent differ from operant conditioning?
Fourth — do I ever use bribes?
What is respondent behavior?
Geek answer: A respondent behavior is one that is automatically elicited by a stimulus.
Example: Eye blinks because a puff of air hits it. That is a reflex. That is classical or respondent behavior. The puff of air is an unconditioned stimulus creating an eye blink that is an unconditioned response. All reflex – you don’t need to learn this.
These are unlearned responses. Some stimulus appears; the body responds.
Loud noise: startle.
Hand on stove: pull back hand.
Emotional responses fall under respondent/classical behavior because the body is automatically responding to a stimulus. Bird sees a lion, bird is fearful – there are physiological responses that add up to what we would call “fear.”
What is classical conditioning and counter conditioning?
Geek answer: Classical conditioning pairs one stimulus with another to change the function of a previously neutral stimulus to one that elicits a respondent behavior.
Example: Clap hands (neutral stimulus) → puff of air (unconditioned stimulus) -→ eye blink (respondent behavior).
Do that enough times and clapping hands will elicit the eye blink. You physiologically respond the same way as you would with a puff of air in the eye.
In geek terms, the unconditioned stimulus of air puff has been paired with the conditioned stimulus of clapping hands, so that now the conditioned stimulus produces a conditioned response of eye blink. (Yep, it’s a mouthful).
And you get:
Clapping hands (conditioned stimulus) –→ eye blink (conditioned response)
I hope you can see that we are not “bribing” an eye blink by clapping our hands. I purposefully used a non-food example so you can see how it works with any respondent behavior.
This is science. Irrefutable because there are thousands of studies replicating this phenomenon. That respondent/classical conditioning works is not just Debbie’s or my opinion.
With counter conditioning, there is already a negative association with a stimulus.
Dog = scary.
Man in hat = scary.
People carrying things = scary.
Noise = scary.
Since there is an association that exists that is scary, we then want to pair something that has a positive, good feeling about it in order to change the respondent behavior that is elicited by the previously scary thing.
We start with Man in hat (stimulus) -→ fear (respondent behavior)
Now, just like in the clapping hands, puff of air and eye blink, we want to change the respondent behavior elicited by the man in the hat.
First, we want to use desensitization to avoid the fear response — staying far enough back so that the dog is aware of the scary thing but not yet afraid of it.
It looks like this:
Man in hat (stimulus that is far enough away to not elicit fear but close enough to be noticed) –→ food (unconditioned stimulus) -→ yum, happy feeling because of the food (respondent behavior).
Do that enough times and the man in the hat will elicit yummy happy feelings.
Man in hat (conditioned stimulus) —> yummmm, happy (conditioned response).
In CC/DS, the dog does not have to perform any behavior to get the food. The very positive stimulus of food is just getting paired with the scary man in the hat, in order to change the reflexive response to the scary man in the hat.
I think some of the confusion as to what we are doing with CC/DS is because we ALSO use food in operant conditioning. While it looks like the same thing, they are serving completely different functions in the two types of conditioning.
Bribing? You are not getting the dog to do anything, so by definition of the word bribe, if you are doing CC/DS, you can’t be bribing.
How is respondent conditioning different than operant conditioning?
Respondent conditioning deals with reflexive behaviors only – emotions and responses that we are born with. Sit, down, bark, lunge – none of these are respondent. What we are doing is pairing one stimulus with another to change the function of that stimulus. If you pair a stimulus with the function of “fear” with a stimulus with the function of “happy yum,” in the right way, you change the function of the scary man from “fear” to “happy, yum!!”
Operant conditioning deals with behaviors that we have voluntary control over and that are evoked through reinforcement (behaviors will increase as a result of being reinforced after a behavior) and abated through punishment (behaviors will decrease as a result of being punished after a behavior). Dog learns that their behavior has consequences and behave in the future according to those consequences.
I ask dog to sit.
I give dog a reward.
This is positive reinforcement, and the dog will be more likely to increase or maintain sitting behavior in the future.
Note – neither operant nor respondent conditioning is a bribe. A bribe is giving something valuable to someone in order to get them to do a behavior. You give a bribe (or promise of one) BEFORE the behavior to get someone to do it.
Do I ever bribe?
Sure, sometimes I use a lure to get a behavior, and then I quickly fade the lure so I am doing operant (providing a consequence after the behavior). If luring is the most efficient way to initially get a behavior, I do it. Hold a treat above a dog’s head to get them to sit, then reinforce the sit with the treat. I will typically fade a lure within 3 or 4 trials.
If I misjudge or get surprised and a dog goes over threshold, I may bribe by tossing treats all over the ground to get them to drop their head and eat the treats instead of barking and lunging. And (hang on, this may blow you away a bit), you are doing some counter-conditioning without desensitization there because scary thing brought treats. You are not reinforcing the barking and lunging behavior, you are calming them down. But that’s not a long-term solution; that is an emergency management situation.
For some reason unknown to me, many people use the emergency management situation and claim that’s all CC/DS is doing.
Remember, the number one goal with these dogs is to keep them feeling safe. That means we strive to manage their environment so that while they are learning a different response (through CC/DS) and different behaviors (through R+), they always feel safe and never go over threshold.
I’m more of a visual teacher, so I hope this makes sense while I’m in the process of re-doing my website and creating more visually interesting ways to present this information. It’s a mouthful and there are a lot of science terms. The more you understand what they mean, the more what you are doing with your dogs makes sense.