Higher-Order Conditioning: Did it Happen To My Dog?

Mar 16, 2018 | Animal Behavior, Learning Theory, Training


tan dog with black muzzle and smaller black dog stare at the camera expectantly The other day I was sitting in my bedroom with Clara and Zani and the doorbell rang. And there was dead silence. This pierced my heart. If you follow the blog, you know that I lost my dear dog Summer suddenly in August. She was wonderful beyond compare. She also barked reactively at delivery trucks, the mail carrier, anyone but me on the porch, and the doorbell. So for me, this silence was one of those dozens of daily moments where my heart ached. There was a hole where my dog used to be. But I was also in the here and now. I couldn’t help noticing that when the doorbell rang, something else did happen.  Clara and Zani’s heads both swiveled in my direction and they both fixed me with their gaze. Why?

Classically Conditioning Another Dog Barking

A sable dog with a purple collar with her cheeks puckered, about to bark. I classically conditioned another dog to respond positively to this. Here’s a hint. Those who have followed the blog for a really long time may remember something else. I classically conditioned Clara to have a positive emotional response to Summer barking. I’ve always been proud of this bit of training. It worked out with no downside. I didn’t want Clara to “catch” Summer’s reactive barking, so I gave Clara her favorite treat whenever Summer barked. I was very consistent and it worked, even to the point that Clara drooled when Summer barked. That demonstrated the actual respondent behavior: the body preparing for food. Then she ran to me for her treat. You can read about the details in the following posts. But please come back, because the plot thickens.

Higher-Order Conditioning

OK, back to the two dogs staring at me in the silence after the doorbell rang.

Higher order conditioning: A form of classical conditioning in which a conditioned stimulus CS1 is first paired with an unconditioned stimulus, in the usual way, until CS1 elicits a conditioned response, then a new conditioned stimulus CS2 is paired with CS1, without the unconditioned stimulus, until CS2 elicits the original conditioned response.–Oxford Reference

So in plain English, you condition a response to a previously neutral stimulus. Then you use that stimulus, in turn, to condition another previously neutral stimulus. Pavlov and his buddies did some of this, using lights and sounds as first and second stimuli. For instance, they conditioned a light turning on to predict food. Then they used a buzzer to predict the light. They used his ghoulish apparatus to collect saliva from the dogs. They found that the buzzer came to make the dogs drool even when there was no food at the end of the sequence (until the association started to break down). This is called second-order conditioning. The umbrella term is higher-order conditioning since there can be more than two “layers.” Exactly how many can pile up is determined by many factors and still being studied. Higher-order conditioning also happens with fear-inducing stimuli, by the way. Does it ever! We are more used to it, I think. My friend Debbie Jacobs of Fearful Dogs puts it this way:

[thinking about higher-order conditioning] could get us aware of why our dogs might not want to go certain places or be around certain things, even if those themselves were not the source of fear or distress. I use the example of hearing screeching tires or smelling burning rubber after being in an accident and the adrenalin dump they can produce, even though it was the windshield that your head made contact with that was the source of injury.

Another example is what happens when we go to a scary movie. Do we sit there, calm and relaxed, until the monster jumps out and scares the crap out of us? Most of us don’t. Any director worth his salt has been loading in conditioned stimuli that have been previously associated with scary stuff happening. The squeaking hinge, the power going off, birds suddenly going silent, even certain types of camera work. We aren’t born being afraid of those things. But our very physiologies have learned what they mean in the movie theater. And predictable precursors of scary or painful things can initiate physical fear responses in our dogs as well.

My Example at Home

So what happened over the years at my house with barking and doorbells? metal doorbell on bricks. Doorbells can easily become classically conditioned or have higher-order conditioningFor Clara: The food conditioned Summer’s bark. In turn, I believe Summer’s bark conditioned the doorbell. Doorbell -> Barking -> Preparation for food. And when the barking fell out of the picture, we witnessed second-order conditioning. Doorbell -> Preparation for food. I mention in the previous posts that Clara actually drooled and smacked her lips at the sound of Summer barking. The other day when the doorbell rang, I didn’t see any drooling. But I got the perked-up reorientation that often signifies a conditioned response. I can’t prove that I got higher-order conditioning with Clara. But I’ll note that there is a lot of research that indicates that if two stimuli are perceived by the same sense, it’s easier to get higher order conditioning 1)Barry Schwartz, E. A. Wasserman, and S. J. Robbins. (2002). 5th ed.  Psychology Of Learning And Behavior.  p. 63   Barking and the ring of a doorbell are both auditory stimuli. The research indicates that they would be easier to pair than, say, barking and a light turning on or barking and a smell. For Zani: I also got a reorientation and a stare from Zani, but I am going to argue that the process for her was different. I did not classically condition Zani to Summer’s barking. She lived with Summer for two years before Clara came along. Over those years she heard plenty of barking that was not paired with food. I started systematically pairing Summer’s barking with food for Clara as soon as Clara arrived. When doing that, I didn’t do the same for Zani, not systematically anyway. But Zani is keen to notice opportunities for food. She started to show up in these situations, so I gave her a treat. So through observation and perhaps some social learning, Zani learned that when Summer barked, she could run to me and get food. So I believe Zani’s response was primarily operant. Summer’s barking was a cue that running or reorienting to me would be reinforced. These two types of responses are intertwined. Cues acquire an element of classical conditioning. Respondent behaviors such as the body preparing for food are quickly followed by operant behaviors to get to the food. So I’m not going to argue that Clara’s response was 100% higher-order classical conditioning, nor that Zani’s was 100% operant. But I think those are the main processes that happened.

The Next Puzzle To Solve

Whenever I leave the house, I put the dogs in their respective parts of the house first. Zani in the den. Summer, when she was still with me, in the front room. Clara in the bedroom. Then I would give them each a cookie that I got out of a particular bag in a particular cabinet. You can see Clara’s response in this video.  a tan dog with black muzzle drooling In the video, she was in the kitchen watching me. You can see that she didn’t respond when I opened the cabinet and got out something else. She sure did when I reached for the cookies on the correct shelf, though. Nowadays she doesn’t see any of this because she is waiting back in the bedroom. She listens intently to the auditory evidence of the progress towards her cookie. By the time I arrive back in the bedroom to give her her cookie, she is doing some serious drooling. I have been trying to stealth video her to see when the drooling starts. My “deliver the cookie” behavior chain is something like: open the cabinet, open the package of cookies, break apart a cookie, put away the cookies, toss Zani’s cookie onto the daybed in the den and tell her to go get it, shut the kitchen door, walk down the hall, and give Clara her cookie. By the time I get there, she sometimes has serious strings of drool hanging down. The question: how far back in my process did the drooling start? Her cookie is a sure thing by the time I walk down the hall with it. How much earlier does her body “believe” it’s coming? Do we have higher order conditioning in this case as well? Or does she start drooling when I open the package in the other room and keep on drooling through my other activities? I’ll keep trying to figure it out and let you know! Do you see examples of higher order conditioning–either appetitive or aversive–with your dogs?

Related Posts

Doorbell photo credit: Wikimedia Commons Text and all other photos copyright 2017 Eileen Anderson Notes   [ + ]

1. Barry Schwartz, E. A. Wasserman, and S. J. Robbins. (2002). 5th ed.  Psychology Of Learning And Behavior.  p. 63