Shattered Myths: The problem with teaching ‘other’ behavior II

So, Pavlov says that once an association is made it doesn’t fade just because you teach something knew. That should trigger a reaction of ‘who doesn’t know that’? Lots of people – everyone from behavioral scientists to modern, scientific dog trainers and educators think that is a false statement. They think that if you see a problem behavior, the way to get rid of it is to teach something else. They have a number of pet names for this process, based on the concept of ‘differential reinforcement.” Differential reinforcement means you select which behaviors you will reinforce and ignore those you don’t like. That is also called “shaping” – a term coined by B.F. Skinner himself. There are several problems with this term and concept.The first thing to realize is that there is no ‘shaping’ with positive reinforcement unless other behaviors are punished. If that word makes you think of abuse, you are likely unintentionally abusing the word. It simply means some event that causes a suppression or stopping of a behavior or simple association. At least for the duration of this article, think of punishment as the brakes on a car that suppresses movement. In order to truly understand behavior you have to be able to say reinforcement and punishment without adding some kind of emotional connotation. In the case of shaping, the full name of its complimentary effect is ‘differential punishment’. This means that you select which behaviors you don’t reinforce. How tough it that? In ‘shaping’, one behavior is reinforced while all other behaviors are ignored – which punishes the likelihood that they will occur.

Now for the pet names. You will now know why I had to define the term ‘differential reinforcement’ These acronyms represent about 75% of the ‘solutions’ offered to virtually every behavior problem.

DRO – Differential Reinforcement of Other Behavior
DRA – Differential Reinforcement of Alternate Behavior
DRI – Differential Reinforcement of Incompatible Behavior
DRL – Differential Reinforcement of Less Behavior

You can see they all describe a means of creating a preference for one behavior over another. The intent is to stop an existing behavior by replacing it with something else. Here is where we starting seeing an oddly unscientific bias. All of these terms describe trying to suppress a behavior. Though there is a term that describes that process – punishment – it is not referenced. Instead, these terms use the term ‘reinforcement’ – the behavioral effect that increases behavior. They want to stop a behavior by suggesting a process that increases behavior. Why would scientists describe a process that increases behavior (differential reinforcement) when they want to shape behavior and use the same term to describe a situation where they want to stop a behavior? That really doesn’t make sense. That doesn’t mean there isn’t reason for this distortion. It’s all about the image.

They can’t say that ‘shaping’ is a process that requires positive reinforcement and negative punishment. That blows the claim of teaching with exclusively ‘positive’ methods. (Why that would matter is another issue) So, they conveniently leave out a reference to the punishing effect of shutting down reinforcement for all other behaviors intended to make them decline. Then, when they decide they want to make a behavior decline, they continue to use terminology that describes the process from the perspective of positive reinforcement. Oy vey! My brain is hurting.

4 thoughts on “Shattered Myths: The problem with teaching ‘other’ behavior II

  1. So now I’m curious,

    I tend to lurk around the YouTube comment sections on training videos, and I see many-a-people arguing about this very subject. Some of the purely positive trainers do mention negative punishment in their methodology… But the thing is, is ignoring an undesired behavior while subsequently reinforcing a new one really sufficient in causing the extinction of the old behavior? Regarding your earlier post on this, and regarding various other observations I made, my answer would be no. It is not a severe enough punishment to warrant extinction (obviously “severe” not suggesting you beat your dog when he jumps up on the couch). Any dog that is variably or heavily reinforced for a behavior will never see the extinction of that behavior without proper punishment. Its like what Michael Ellis said about positive reinforcement being like a gambling addiction for dogs. It applies to the extinction of a behavior as well. This does leave me wondering of the proper way to punish a dog for jumping up on people or pulling on lead… perhaps a spray from a pet convincer for the former and a leash correction for the latter?

    Another topic that I find is hot in the comment section is the problem of punishment “suppressing” behavior (i.e. you pop a dog’s leash when it displays reactivity, makes the dog more internally aggressive and the dog explodes one day and shocks the shit out of everyone). I would think you would want to pair this with some other kind of stimulus conditioning, like B.A.T training, at the very least, so the dog would learn that A) this behavior is unacceptable and B) the stimulus isn’t something to fear. Maybe not use leash corrections at all and just work on stimulus training.

    I’m so sorry for rambling. I’m 19, and I am going to be a dog trainer, and I am going to be the best that I can so I can help as many dogs as I can. I’m doing my best to learn everything possible so I can get it relatively right. Thank you for allowing me to brain puke my poorly formed ideas onto your page, which I am in love with. No nonsense. Real. To the bone.

    • Hannah,
      People who craft training practices to fit an agenda invariably have to fictionalize reality to make things appear consistent. You are correct, simply removing reinforcement for a behavior and reinforcing something else won’t remove the behavior from the existing repertoire. As for things like BAT, it contradicts the literature of the people who created it. The way to stop a behavior is to stop a behavior. The suggestion that punishment triggers aggression is both mindless and specious. It is mindless because punishment stops behaviors – including aggression. You can elicit aggression with pain, but a never-cited paper from behavior analysis makes that specious. Ulrich, Wolfe & Dulaney, Punishment of Shock Induced Aggression, JEAB, 1969. They triggered aggression with shock – and then punished the aggression. Their term was ‘completely suppressed’. So, why did a professor of behavior analysis and a graduate student create BAT? Because they were pushing an agenda, that does not reflect reality. That is profitable (they sold a video series and got paid to give seminars) and elevates status. Too bad for the people who paid the money. It promotes several fictions.

    • Rachel,
      It isn’t me making that connection – it is the confused language of behavior analysis that does not actually describe behavior correctly. To be ‘extinct’ means to remove from existence, like dinosaurs. The process of ‘extinction’ and ‘extinction schedules’ result from removing positive reinforcement that causes a behavior to decline or stop. That is actually called ‘negative punishment’. Meaning ‘extinction’ is a misnomer. There are several problems with the term – it has no influence on instinctive behaviors, it does not prevent the behavior from returning and it leads people further from understanding that reinforcement and punishment are roughly symmetrical opposites that are interdependent. So, it is not I who have connected extinction to punishment, it is the field of behavior analysis that has made them synonymous – but treats them as if the are two separate things.

Leave a Reply to Rachel Cancel reply

Your email address will not be published. Required fields are marked *