Eduardo J. Fernandez, a behavioural psychologist, and I discuss the differences between capturing behaviours and superstitious behaviours. Capturing behaviour is when you spontaneously reinforce behaviour the animal you work with shows. For example if you want to teach the animal to make a vocal sound and there is a moment the animal does it, you reinforce right away to make it happen more often to the point where you can connect it to a signal. Capturing is a very cool way to teach animals new behaviour. As with many techniques it does have it’s advantages and disadvantages but it can be a lot of fun!
The animal does not know that you, whether by accident or on purpose, reinforced this behaviour.
Superstitious behaviour is a behaviour that is accidentally reinforced, the delivery of the reinforcer occurs close together in time with the behaviour, which increases the likelihood of the behaviour occurring again. Often you don’t realise you’ve reinforced it. But from the animal perspective this might be the same as capturing behaviour now the animal doesn’t know if it was intentionally or unintentionally.
Dr. Eduardo J. Fernandez received his Ph.D. in Psychology (minors in Neuroscience and Animal Behaviour) from Indiana University, where he worked with the Indianapolis and Cincinnati Zoo. He received his M.S. in Behaviour Analysis from the University of North Texas, where he founded and was President of the Organization for Reinforcement Contingencies with Animals (ORCA). He has worked with close to 50 species of animals, with a focus on marine animals, carnivores, and primates. He is currently a Visiting Professor in the School of Behaviour Analysis at Florida Institute of Technology. Eduardo also continues to run the Animal Reinforcement Forum (ARF), a former listserv and now Facebook group (facebook.com/groups/AnimalReinforcementForum/), which is dedicated to group discussions on animal training and behavior from a scientific perspective.
Zoospensefull: Eduardo could you explain to us a bit more about who added the term ‘superstitious behaviour’ in the animal behaviour modification dictionary?
Eduardo J. Fernandez: Sure. The concept came directly from Skinner’s (1948) paper, “Superstition in the Pigeon”. The idea that Skinner had and wrote about anecdotally is that behaviour could potentially be rewarded by contiguity alone, meaning in the absence of an actual contingency, and just by the mere fact that some potential reward followed some response.
Zoospensefull: Can you explain more about the meaning of superstitious behaviour?
Eduardo J. Fernandez: So, as I noted above, the idea Skinner had was that rewards reinforce behaviours in the absence of actual contingencies, and simply by following some response in time. Skinner saw his own pigeons doing strange, wall-hugging responses when put on fixed-time schedules (about 15 seconds), and surmised the behaviour occurred because it was somehow being rewarded. He went on to describe the behaviour as no different than the human superstitions we see, particularly in places where random, non-contingent rewards may be prevalent, like during gambling or sporting events.
Zoospensefull: Considering that capturing behaviour is an intentional action by the trainer, would this mean that capturing is the opposite of superstitious?
Eduardo J. Fernandez: I avoid using “intent” when describing any behavioural procedure, because it removes the contingency from occurring between an organism’s behavior and the environment and instead places it within the organism itself. That said, this is one of the many difficulties encountered with the concept of superstition; the control of that supposed non-contingent, adventitiously reinforced response now existing at least at some level outside of an environment-behaviour contingency. It’s why we require contingency just for defining reinforcement! There are ways around this when focusing on a contiguously-based definition of reinforcement, but it’s at least worth noting as one of the many difficulties in using adventitious reinforcement (and, thus, superstition) as an explanation of anything.
But, I’m digressing a bit here. I would say that what you and the training community are more interested in is not superstition or adventitious reinforcement, but rather, contingent, ACCIDENTAL reinforcement. For instance, you unintentionally but continuously (and contingently) reward some behaviour again and again without necessarily knowing it. You’ve created a contingency for the response that, regardless of your intent or knowledge, exists. That’s not superstition or adventitious reinforcement. That’s just accidental reinforcement. And that accidental reinforcement very much would be different than capturing, which may be accidentally OFFERED behavior on the part of the trainee, but certainly intentionally rewarded on the part of the trainer.
Zoospensefull: Will there be a negative outcome of superstitious behaviour?
Eduardo J. Fernandez: I guess it’s not clear in terms of actual superstitious behaviour, but there are certainly unintended consequences for inaccurately identifying the function of some behaviour as supposedly a result of adventitious reinforcement, particularly when lacking any empirical support!
Instead, if we’re talking about accidental reinforcement, I would say there are always potential problems with not understanding the operant contingencies involved in any behaviour. If you’ve created a contingency that you’re unaware of, like accidentally rewarding barking or jumping on unwanted surfaces, then it would be hard to get rid of that behaviour without first knowing what it was that you have been accidentally reinforcing. And that, really, is what a large amount of the consulting community is about; identifying accidental contingencies that otherwise have been missed by owners/keepers/trainers, etc.
Zoospensefull: How can we avoid this happening?
Eduardo J. Fernandez: Well, first, we need to be clear on this distinction between adventitious reinforcement (superstition) and accidental reinforcement. And I think most, if not all, of what we’re talking about here, and what the training community is primarily interested in, is not superstition, but rather, behaviour accidentally reinforced by the trainer, but nonetheless, still consistently and contingently rewarded as such.
Regardless, the best way to understand all behaviour is to understand its function, and that ultimately requires being empirical about assessing that function. Behaviour Analysis itself is necessarily a quantitatively driven endeavor. That’s part of why I spend so much time talking to trainers and other applied animal behavior people about the importance of data. There are plenty of simple tools for incorporating actual data into all our assessments, and any and all data go a long way to avoiding the pitfalls of not properly identifying ALL contingencies, operant and other contingencies alike!
Zoospensefull: From the animals point of view, is there a difference between capturing behaviour and superstitious behaviour?
Eduardo J. Fernandez: I think what most trainers are really talking about is not superstition caused by adventitious, contiguous reinforcement, but rather, accidental reinforcement that is still contingent. That said, I think regardless of the intent of the trainer or trainee, reinforced behaviour is reinforced behaviour. If it produces a reward, then I would expect it to have the same effect on increasing that response, and similarly, on how an organism feels about having that behaviour rewarded. The real trick is in identifying what contingencies are in place and adjusting them accordingly. Shifting rewards from undesired responses to desired ones, and thus, focusing on increasing the welfare of that animal you’re working with.
Thank you very much for your participation Eduardo and your insight into superstitious behaviour. As you might have discovered through this article already the challenge is not necessarily whether it is superstitious or captured behaviour but more what we as trainers are doing. Many times when we train, we come across accidental reinforcement, it happens. This is only a small part we have to problem solve afterwards, and asses what has happened in the past to solve future behaviour.