Quick….. Word association:
Pavlov= Bell. Saliva. Conditioning. Reactions.
Skinner= ________…….. (sounds of crickets)……….
Yeah, that’s what most people do too.
B. F. Skinner. Figured out the other conditioning. Not the “ring a bell- make you salivate” of Classical Conditioning fame. But the WAY less understood “Operant Conditioning”.
Obviously, you can read the links and research this stuff, but I’ll try to break it down.
Classic Conditioning changes how you “react”. (pair a song with something scary enough, and they’ll get scared when they hear it)
Operant Conditioning changes how you “act”. (you can teach a mouse to touch a lever to get food, even though cheese is right in front of them)
The tricky part comes into play in the fact that they both have the others effect eventually. Changing how you feel about something effects how you decide to behave. Doing a rewarding behavior enough times will effect how you feel about that behavior.
Uh oh….. I said “reward”. Crap. I opened a can of worms.
The biggest problem in researching/talking about dog training, or behavior modification, is the “operant conditioning quadrants”
It looks like this:
Now, here the Positive and Negative are more about the “blood test” type of thing. As in Positive means that thing is there, and negative means that thing is not there. But everyone gets hung up in the “feeling” definition of those words. As in positive is good, and negative is bad. The problem is in the most used and understood quadrant (+R), it can be used both ways. Giving a treat to encourage a behavior, is “positive” in the “added to” sense, and also “positive” in the “feel good” sense. But that’s where it falls apart. Because +P (spanking a dog for ex), is positive in the scientific “added to” sense, and definitely negative in the “feelings” department.
So, clearly, I don’t think posts out ahead of time. Because what started as an informative post is turning into the friendship shattering discussion of reward/punishment in training!
Good sense tells me not to write this as it is kind of divisive. In dog training (and people training!) which method to use, and how to use them, starts as many fights as politics or religion.
But, since I care more about knowledge than people liking me, I’ll just go ahead and throw this at you.
There is no such thing as “purely positive” in the feel good sense. Everyone just loves to say that. “we use only positive techniques”, We use only reward based training”. There is no such thing.
For there to be light, there must be dark. That Ain’t my rule. That’s nature. Like it or not, you’re using punishment.
Look at the quadrant again. The diagonal is what’s important. If you use +R, you MUST also be using -P! There’s no way around it. When doing purely positive training, you say “sit” and the dog lies down you say “no…. Sit” and the dog lies down you DON”T GIVE THE FOOD! That’s punishment. -P, but punishment all the same. Honestly, it just makes you feel less guilty. Here’s the rub: when you ask a “reward only” person what to do if your dog isn’t that food motivated….. Oh, that’s easy. Train before feeding. Or reduce their food to increase “motivation”. Really? Wow. The thing is, it’s all in your focus and intention. If you tell your kid “get a good grade and mommy will be sooo proud”, “get a bad grade, and I’ll be soo disappointed”. You’ve essentially said the same thing. One just feels better because of your focus and intention.
What about the other diagonal. Yes, if you use +P, you are also using -R. This is the classic “the beating will continue until morale improves” mindset. Look at abused spouses. The reward is not getting pounded. Obviously, this works, but doesn’t build the best relationships. But, we aren’t talking specific examples of the +P concept. We’re talking the concept itself. Which I use all the time. And before you act like I’m an abusive asshole….. Bear in mind….. SO DO YOU.
Yeah, that’s right I said it. You….. With your clicker and hot dogs…… You use +P!!!!!!
Shut, up. Yes you do.
Ever gotten a stern tone and posture, and said “NO” to your dog?
Yep. +P. You added a bad thing, to prevent that action from continuing.
To a dog you were one step from “touching” them.
Dogs warn in this order:
Stillness (the freeze warning)- Pressure (a hard look, lip lifting, or leaning/moving towards you)- Sound (growling)- Touch (muzzle punch, or warning touch)- Attack (try to hurt you)
You froze. Gave pressure. And, made a sound. They stopped to avoid a touch.
You didn’t mean to, but you added (+), the implied threat pf physical violence to stop an action.
See the problem isn’t with the science. The problem is with our understanding.
And our identity. Soft people will “never” use any punishment. And Napoleonic tough guys think reward training is for sissies.
We all use all of them in one form or another.
The issue is do we understand how. Obviously, you want to de-emphisise all Punishments. You CAN”T not use them, but it’s healthier for the relationship if you emphasize the rewarding part of the diagonal that your using.
You should be aware of all your interactions when you teach. Because, you can’t minimize what you don’t realize you’re doing! Understand the quadrants! Understand how they play together. Realize that punishment in your training is a reality you can’t do anything about. Be AWARE of the punishments you use. Try to focus on punishment little as possible.
ALWAYS try to focus on or stay in the R half.
NEVER punish without intending too. That’s irresponsible.
NEVER punish more than is necessary. It makes them not trust you, and it makes you an asshole.
NEVER punish (+ or -) out of frustration/anger. That’s abuse.
Well. There it is. I picked a fight. My only saving grace may be that no one reads this blog. Guess we’ll see.
Feel free to comment. I’ll respond.
12 thoughts on “No Love For Skinner”
If you say “do good and mommy will be proud,” and assuming one cares whether mommy is proud or not, the consequence of doing good, is a reinforcer, because it increases the likelihood of good behavior occurring. If we say “do bad and mommy will be disappointed,” and again, it matters, we are using the threat of punishment in order to decrease the likelihood that the bad behavior will occur. The outcome may be the same, but the way we get it matters. In one case we have someone working to gain a reinforcer and on the other to escape from a punisher. There are lab tests with animals that show that there is a difference in the future behavior we can expect from an animal depending on which method is used. The threat of punishment increases the chances that an animal, or child, will avoid situations in which the potential for P+ or R- exist. We need only look at drop out rates and how many people dread going home for the holidays, to see it in action.
Thanks, for posting! I totally agree with you!
When this theoretical person works for “mom’s” affection, they are working to get a reinforcer. And when they work to avoid disappointment, they are working to avoid a punishment.
And I completely agree that any being acting to avoid punishment, will fundamentally be in a worse mental state than a being working to get a reward. I’ve seen relationships with dogs absolutely ruined by using poor +P methods.
The main points that I was thinking about were;
1- The “diagonal” on the conditioning quadrant is necessarily true. It’s not theory. The inverse MUST always be true. I just think it’s a naive view for people to try to convince themselves that they “do not use punishment”. I believe that +P in the same terms as self defense for people. As in, I will obviously do it if I need to….. but the fact that I did, means things have gone very, very wrong. And If I had paid attention, or had a better response plan, it wouldn’t have been necessary (yes, teaching humans self defense is my day job). I also believe you can’t minimize what you don’t realize. It’s like when people develop a wheat allergy and suddenly realize that wheat byproducts are in damn near everything. You have to understand “P” both -, and +, in order to consciously reduce it’s use and impact. If you use +R, you are by necessity, using -P. That just is a fact. Now -P is hugely less traumatic than +P, so that’s awesome. And clearly, if the fact is, the +R/-P axis is a fact…… It is absolutely beneficial for the long term psychological health of the being to focus on the “R” side!
2- +P isn’t necessarily “physical” punishment. The problem is I’m geeking out on the science of this stuff. And scientific terminology is almost always, not the same as common vernacular. To most +P is physical discomfort. It’s not. +P is the adding of anything that is likely to decrease an activity. So a sound, a smell, assuming a posture. Those are all +P! Again, the people that like to say “I don’t use punishment”, just don’t understand the concepts. An “AH, AH” sound, or saying “NO” when you dog jumps up is +P! The “purely positive” crowd’s “coins in a waterbottle” noisemaker is +P! Again, if you don’t understand something, you can’t regulate it.
My premise is, that I think we should be completely aware of all the ways in which we use any form of “P”, in order to best regulate it’s use.
And, yes I’m exquisitely aware of the effects of the misuse of both of these methods. My father was not above throwing me a beating. And my mother was emotionally abusive, although, she never laid a hand on me. No idea which one caused me to drop out after the seventh grade, and withdraw from society. But. I can tell you for sure…. Long term….. The -P was by far more hurtful to me. Fortunately, I turned myself around, and now try to dissuade people from using “P” methods in any form without a lot of foresight.
For the record, I’m violently opposed to animal abuse, in all forms, whether mental or physical, including neglect. I say that because, because as I discuss these concepts, people always assume I’m a “Khoeller”, “beat ’em until they comply” guy. I’m not. Just throwing that out there.
Did I also mention I enjoy debate, and almost always play the devils advocate?
Again, thanks for posting, and hopefully this is all taken in the spirit of friendly discussion.
“If you use +R, you are by necessity, using -P. That just is a fact.” One’s use does not mean that the other is occurring at the same time. They each address different behaviors in a stream of behavior. This may be what you mean, but not what some might understand.
“any being acting to avoid punishment, will fundamentally be in a worse mental state than a being working to get a reward.” Will they? Fundamentally? Mental state? I am going to put on a raincoat today to avoid getting wet. I will not need a cocktail to get through the day because of it.
We can talk about behavior. Not so sure about mental states.
I am familiar with the purely positive debate, and not going there 😉
Let me start by saying that the line about the raincoat and cocktails was absolute gold! I will be using that for sure. Although, it seems like was a line that could be used in favor of milder forms of +P. Did I misunderstand that. or is that what you were saying?
I didn’t mean to make it sound like one would use +R AND -P at the SAME TIME. I really just meant that they are linked. That I think, there really is no such thing as reward only. I saw Dr. Dunbar us a verbal correction on a dog that nipped him. +P for sure! I’ve seen well respected marker trainers use a negative marker like “uh, uh”, or “nope” and then withhold the treat, when the dog picks the wrong response. That’s -P isn’t it?
By the way, when I say things like that, I really am asking! I’m not saying it is the condescending “isn’t it” smug way. I’m literally asking you, and am hoping to learn from this discussion. Really.
I’m new to the science of this. Check the “about us” page, it says that I’ve been rescuing, and living with dogs for almost 25 years. I’ve always been “good” with dogs, but a very problematic dog took me to professional. This has started what I can only describe as an obsessive desire to learn about it. So, I really am hoping to learn. I am smart enough to know that I don’t know everything!
“Ever gotten a stern tone and posture, and said “NO” to your dog? Yep. +P. You added a bad thing, to prevent that action from continuing.”
Not necessarily. This is only an example of positive punishment IF the dog’s behavior decreases in the future. Otherwise, it’s just an interrupter.
It could also be negative reinforcement if the dog shows an increase in a behavior that stops or prevents this aversive.
It’s not just whether or not an aversive was added or removed that determines which quadrant is in play, but whether or not the behavior increases or decreases in frequency, intensity, etc.
That was an AWESOME point. The defining aspect of which quadrant an action falls into, is it’s effect, NOT it’s intent!
It’s so funny that I never put that together…… I have a 2.5 year old female who, if she gets too “over-stimulated”, can roll into aggression pretty easily. I’ve had to learn how to say “no” very smoothly and calmly to get e cessation of activity. If I get stern, or pushy, at all, she gets so rolled into drive she can barely contain herself. Blown pupils, hair up. All nails and teeth! It took me forever to figure out that “stern” verbal corrections, got her “more amped”. Even though I’ve been researching this subject incessantly, I never mad the connection, that what I was intending as +P, was actually +R to her! I have to use -P (saying “no” as a mark and then withdrawing attention) to help her calm down.
Thank you so much for the info!
What K9mythbuster said. Would go further to say that if the dog loves attention and is not startled by noise, “NO!” can be R+. “Mom – you’re paying me attention.”
So the only measure of whether something is R+-/P+- is whether you add/subtract and the outcome.
Which means you actually never use R & P at the same moment in time
“’I’ve had to learn how to say “no” very smoothly and calmly to get e cessation of activity” My question, as soon as I read this sentence was, “Why do you have to say “no” at all to get a cessation of activity?” Is the “no” a cue that you have trained your dog to observe, or is it just the normal human response to something they don’t like, which dogs don’t really understand but which humans normally say in such a way as to pose some kind of threat? I trained my dog to “stop” on cue, but it has only one meaning – if you are walking, or stalking, toward something you want, please stop in your tracks and look to me for further instruction. It isn’t used for every “bad” thing she does that I don’t like. The point about “purely positive” is not to actually be so pure that no one can assail you, it’s to ALWAYS seek the most humane way of getting the behavior you want. To me, that means staying in the R+ quadrant as much as possible, and possibly using the P-. But, it also means not drifting into R- if at all possible, and it also means not using P+ intentionally at all. Intention is important. If a person’s mentality is always to “correct” what goes wrong, the nuances of this kind of training will always escape them
Hi Sarah. This just got reposted around FB and I hadn’t seen it before. Agree with you about “purely positive.” But I strongly disagree with you about your points about diagonals. I find them both to be incorrect. Please understand that I am not coming from a place of “Oh no I couldn’t possibly have used punishment!” or “I never use aversives!” I’m transparent about my own methods and their evolution. I’m coming from the science.
Withholding a treat in an R+ session is not punishment. It is an employment of extinction. I just happened to write a whole blog about it. Coincidence–I hadn’t read this at the time. Hope you will allow the link here. http://eileenanddogs.com/2013/10/17/difference-between-punishment-extinction/ We are “not giving treats” all the time when animals perform behaviors. Far more often than we give them. We are not constantly punishing by that. For punishment to have happened, there has to be a consequence to performing the behavior. When one withholds a treat, there is no consequence. You can find it in any learning theory book, especially in the sections about DRA, DRO, and DRI. One behavior is being reinforced, another is going extinct.
Extinction is no picnic if used by itself, and I believe can be employed inhumanely. But it’s not the same as punishment.
Likewise, employing negative reinforcement does not mean one is positively punishing a behavior. I have this from the words of Susan Friedman, PhD (10/17/13 lecture in Living and Learning with Animals Professional Course). The reason is that for punishment to have occurred, a behavior must decrease in frequency. That’s the definition. And generally, the behavior the animal is performing at the onset of an aversive is randomized if one is in a negative reinforcement scenario. They may not always be doing the same thing when the shock (or pinch or pressure or nagging) starts.
Examples: If I decided to teach my soft dog Zani to back up by walking into her whenever we were both in a certain hallway (note: I don’t teach backing up that way to any dog anymore), she would definitely start to avoid being in that hallway with me. Being in the hallway with me would have been positively punished. But let’s say in another training scenario I play a loud obnoxious noise whenever I want Zani to come get on her mat in the kitchen, and the noise stays on until she gets there. (Again, I would not do this.) She could be anywhere in the house when the noise starts. Her behavior at the onset is randomized. So nothing gets punished. But if I started doing it consistently when she was doing one thing, that could get punished. Make sense?
If your point is that for negative reinforcement to occur, there has to have been an aversive present, I’m with ya!
Hi again! Just wanted to apologize for calling you all Sarah! I know now I’m talking to Jay and Amanda. Sorry about that.