This post is about the difficult concept of combining negative and positive reinforcement.
I’ve always wondered about actions having multiple consequences. I recently learned more about that possibility and found a good example in Lewis.
So-called “composite reinforcement” occurs when multiple reinforcers are the result of the same behavior. Although this phenomenon has been recognized since at least 1969 (Osborne), it has only been named and systematically studied in the last decade. The synthesized reinforcement is recorded in functional evaluation and used for functional analysis.
Synthetic reinforcement and escape to appetite
Sometimes escape just means running away. Must stay away from scary monsters and hot stoves now. But sometimes reaching for something better is a function of escape. This topic is often debated among humans. This combination of reinforcement is called “reinforcement.” synthetic reinforcement.
Some scientists suggest that synthetic enhancement is common and that looking for just one contingency in functional assessment or analysis is artificial. Synthesized reinforcement may include the consequences of aversion and appetite, multiple aversions, or multiple appetites. For example, multiple appetites may be the reason your dog is more happy when you play with a toy than when he is alone. Human examples abound. Consider that for sports fans, entering a stadium for a game can be enhanced in many ways.
Returning to the interesting combination of R- and R+. A classic (albeit old) example is watching something and waiting until a commercial starts before getting up to grab a snack. Escape from boring commercials (escape) and gain access to food (appetite). Most of us hate commercials, but most of us don’t walk away from the screen without doing anything for 30 seconds.
Another example is doing extra work to miss class. The classes aren’t too bad, but if you have time, your friends will probably take you water skiing that day.
Synthetic reinforcers are also kind because they provide another source of appetite when engaging in unpleasant rearing activities. This is much kinder than simply saying, “Hey, my dog can leave if he wants.” If you have nothing else to do in the room and are managing your only food source, simply leaving the room isn’t that great.
Here’s my example.
Scene setting (precedent)
I walk my dog individually every day, unless I’m sick or the weather is really bad. I’ve always gone by seniority, so young Luis goes last.
In the cooler months, walk late in the evening. My partner makes dinner for the dogs while I walk with Luis. Lewis knows the routine. When I get home from my walk, dinner is waiting for me. Did I mention he’s excited?
In contrast, in hot weather, walking your dog will be much slower. It was nearing dusk and it had been a long time since we had eaten dinner.
Action: Harness removal
Lewis is looking forward to taking off his harness (turquoise harness in the last photo) when dinner is waiting. But his GPS collar tracking unit is bulky. This means that if he moves while you’re manipulating the strap, the harness can get caught on the collar. So I set up a contingency for unharnessing. Do not remove the harness until he is still.
Our system starts by undoing two buckles. Next, he has to be extra still while I manipulate the harness over the collar’s transmitter. Once that’s done, I give him the release signal. With lightning speed, he completely removed his head from the harness and dashed off to get dinner.
Usually I give my dog a treat after putting the harness on and taking it off. It’s not the most fun thing to have someone play with your straps or snaps around your body. But in this situation, Lewis has no interest in that one treat. His dinner is waiting for him.
I explained the course of action and its consequences. Lewis’ behaviors include staying still, pulling his head out of the harness when backing up, and running toward dinner. Let’s simplify the scenario a little. Let’s focus on his “self-removal” of the harness, or escape behavior.
A. The harness is in an uncomfortable position (wrapped around the neck, half on, half off)
B. Lewis wriggles out of his harness.
C. The harness is off (no more discomfort and you are free to leave)
This is a negative reinforcement scenario. But for part of the year there is another big result associated with breaking out of the harness. That’s a bowl of food. How might that affect Lewis’ harness behavior? There was a great way to find out.
Evidence of positive reinforcement
As soon as Lewis turns his head, the first thing he can do is escape from the harness. Otherwise, I wouldn’t be whipping my head on the way to get dinner. But in the summer, Lewis’ attitude changes completely. If I switch my dog walks after dinner instead of before, there’s no meal waiting for him after the walk, and he knows it. When his dinner isn’t waiting He never sticks his head out of the harness. He “helps” me take off the harness, but often with a delay, with a light twist or wiggle. My attitude changed from “Get me out of here!” “Hmm, could you pull your head back a little so I can take this harness off?”
So the positive reinforcer (dinner) seems to play a big role. When I don’t have a harness, Luis just helps me take it off. The terrain of action is different and there is no discretionary effort.
Can I do something different regarding removing the harness?
My tentative conclusion is that, at least with the harness I use, there is always an element of negative reinforcement in the act of removing the harness. Even if the dog abandons the contingency of being still before breaking out of the harness, there is still automatic negative reinforcement for the wiggling out behavior. And even staying absolutely still can become an escape behavior if they realize that is the most efficient way to remove the harness.
But I recently realized something obvious. Instead of removing the harness when he enters the room, you can simply remove the leash and leave the harness on. No arguments! You may want to wait to remove the harness until after Louis has eaten dinner and experienced something interesting going on around the house.
If you remove the harness after he gets excited, there’s still an element of automatic negative reinforcement, but he’ll be much less frustrated. And he accepted the kibble I offered him.
big picture
The idea of synthetic reinforcement is appealing, but also cumbersome. It can overturn many assumptions. This again ruins our hopes that we live in a tidy, dualistic world. If you follow the right formula and pay attention to functional assessment, you can teach your dog successful pet behaviors without creating aversion. You can reduce every situation to one contingency, so you know which ones to use and which to avoid. That doesn’t always happen for me.
If you think in terms of contingency tables (“quadrants”), synthetic reinforcement also doesn’t quite fit (and I still do). Modern behavior analysts increasingly categorize behaviors by function. Typically, 4 to 6 possible functions are identified depending on the source. The most frequently cited behavioral functions are social, escapist, tangible, and sensory (Cooper et al., 2014, p. 511). Lewis’s harness behavior is described as a flight into tangible objects (some systems may further specify edible objects). In the example of synthetic reinforcement above, playing with a dog with a toy adds social reinforcement to access to a tangible object.
But I live in the dog training world and we still struggle with contingency tables. If you think I’m in the middle of promoting R-, you haven’t read much of my articles. Escape to tangible things is very common in the human world. However, what I have seen in training and other interactions with dogs is that the escape response (R-) is primarily performed by humans in ways that are uncomfortable for the dog. And even events that seem trivial to us can have negative effects. I’ll discuss things primarily in a contingent way for now as I learn more about functional categories.
It’s very difficult to avoid R- completely when you live with a dog. I accept that R- exists sometimes, even if I don’t want it to. I try to be transparent about it. I was using the R- contingency for safety when I asked him to stay still while I pulled the harness over Luis’ head. But I figured out a way to change that by removing the harness after the meal instead of before. By observing small moments of aversion, you can practice analysis and dig deeper to pursue your goal of making your dog’s life better.
References and resources
Note: Most of these resources pertain to the use of ABA with children who are deaf or autistic. Some people may not want to check it. See my statement on ABA on my Training Philosophy page.
Cooper, J. O., Herron, T. E., and Heward, W. L. (2014). Applied Behavior Analysis, 2nd edition. pearson.
Garmagami, M., Hanley, G. P., Jin, S. C., and Vanserow, N. R. (2016). Confirmation of control by multiple reinforcers through progressive treatment analysis. behavioral intervention, 31(1), 70-86.
Holehan, KM, Dozier, CL, Díaz de Villegas, SC, Jess, RL, Goddard, KS, Foley, EA (2020) Comparing isolated and synthesized contingencies in functional analysis. Journal of Applied Behavior Analysis, 53(3), 1559-1578.
Smith, SC, Arroyo Antunes, Belgium, J. DeBartello, Sullivan, WE, Roan, HS, Craig, AR (2024). Combined alternative reinforcement and resurrection. Journal of experimental analysis of behavior, 122(2), 195-206.
John G. Osborn (1969). Free time is a reinforcing factor for behavior management in the classroom. Journal of Applied Behavior Analysis, 2(2), 113-118.
Copyright 2024 Eileen Anderson
