Puppy Forum and Dog Forums banner

Does Operant Conditioning for Dummies Exist?

5K views 58 replies 13 participants last post by  Pawzk9 
#1 ·
I must be on my 10th book on dog training that dives off the deep end into the details of operant conditioning. I guess I need to figure this out, and I seem to be having trouble especially with the 4 types and how they are different.

I guess I must be a little slow on the uptake, as I still can't fix it in my brain. Anyway is there some book, pdf, image that caused you to get operant conditioning?

Yes, I am looking into dog training as a career.
 
#3 ·
I think the problem is that operant conditioning uses words in weird ways. What confused me is the use of the words "positive" and "negative". Unlike normal conversation, where "positive" means "good" and "negative" means "bad", in OC, "positive" means "adding" and "negative" means "taking away". Once I figured that out, it started to click.

Actually, "Operant conditioning for dummies" is a popular google search. This article seems fairly good.
 
#5 ·
It helps me if I break the words down individually.
Negative = Taking away a stimulus
Positive = Adding a stimulus
Punishment = A stimulus the dog finds unpleasant (aversive)
Reinforcement = A stimulus the dog finds rewarding

With operant conditioning you are always either trying to increasing the likelihood of a behavior reoccurring or decrease the likelihood of it reoccurring by providing consequences the dog finds either aversive or rewarding.

One of the confusing things for me is that it is what the DOG finds aversive or rewarding so it can vary between dogs. For example, one dog might find being squirted with a squirt bottle aversive while another might enjoy it and find it reinforcing.
 
#7 ·
When you would use them:

To increase the likelihood of a behavior reoccurring you could use:
a) positive reinforcement ex. giving a treat when a dog does something desired
b) negative reinforcement ex. relieving pressure on a collar when a dog does something desired

To decrease the likelihood of a behavior reoccurring you could use:
a) positive punishment ex. giving a corrective leash pop when a dog does something undesired
b) negative punishment ex. taking a treat away when a dog does something undesired
 
#10 · (Edited)
Yep.

Operant conditioning isn't that hard. I think people complicate it more than it really is.

In essence, you're either:

Giving the dog something he likes (+R)
Taking away something he likes (-P)
Taking away something he DOESN'T like (-R)
Giving him something he DOESN'T like (+P)

That's really it. Honestly. Even including the "it depends on the dog part" because what the dog likes or doesn't like depends, well, on the dog.

The real key is timing and knowing what behavior you're applying that to so that he's receiving what you think you're telling him and not something completely different. That's where the real "explaining to the dog" happens.
 
#8 ·
I think another part of the confusion is that many situations could be classified as more than one quadrant, depending on how you look at it. For instance, say you have a prong collar on a dog (I wouldn't) and the collar gigs the dog. If it works, is it positive punishment (the act of pulling has been punished by the collar and so is less likely to occur) or negative reinforcment. (the act of walking with a loose leash is rewarded by the collar discomfort being relieved)? It can be seen as both or either, and opposite quadrants (+P/-R or +R/-P) are frequently used together to make the point clearer. (you jump on me, the attention goes away (-P) and then you keep your feet on the ground, you get attention AND cookies! (+R)
 
#12 ·
The book to read on Operant condition is "how dog's learn" by Dr. Bausch and Dr. Bailey. That's all the book is about is operant conditioning.

Also, like Pawz said. It's not how the trainer precieves the training that make things fall into once sector of the OP quadrant...it's how the dog views it. And Postive punishment is not positive punishment unless: the addition of the stimulus stops the behavior....So you do not know after the fact.
 
#13 ·
Another piece of the puzzle is Extinction - a behavior that is ignored will fade away, potentially going through an extinction burst.
Eating, barking, biting, chasing, digging, sleeping, sex are self rewarding.

Ian Dunbar - http://www.dogstardaily.com/free-downloads comes up with clever operant applications and solutions.

A rough history: Pavlov discovered Classical conditioning, Thorndike applied it to learning, Skinner did the definitive work on 'modern' operant conditioning, and I believe that Bailey studied under Skinner.
(Noam Chomsky did the research that helped re-direct to Cognitive Psychology, and Gagne helped define education psychology.)
 
#16 ·
Another piece of the puzzle is Extinction - a behavior that is ignored will fade away, potentially going through an extinction burst.
Eating, barking, biting, chasing, digging, sleeping, sex are self rewarding.
...And the self-rewarding behaviours will not fade away simply by ignoring them.


And Postive punishment is not positive punishment unless: the addition of the stimulus stops the behavior
What would you call hitting a dog on the head while he's doing a self reinforcing behaviour (let's say digging a hole) when the stimulus (hit on the head) does not stop the behavior (digging) ?
Hitting on the head is called "punishment" if it works...and if it doesn't, it's called "cruelty" or "abuse" ?
 
#17 · (Edited)
What would you call hitting a dog on the head while he's doing a self reinforcing behaviour (let's say digging a hole) when the stimulus (hit on the head) does not stop the behavior (digging) ?
Hitting on the head is called "punishment" if it works...and if it doesn't, it's called "cruelty" or "abuse" ?
It means the intensity of the signal did not match or overcome the drive/focus/emotional intensity of the dog. If he's eating off the ground and I just tap his head lightly, I'm not abusing him, but my signal is not intense enough to make him pay attention to it. The same principle as when a cued behavior is not maintained (or performed) in high-drive or high-emotion situations.

If I wanted to, I could hit Wally on the head hard enough to make him stop doing whatever, probably even eating. However, the question is then, is that intensity or that signal one I should use on him?

Of course, hitting him is out of the question for me. But I could give his recall cue, a behavior that's conditioned so strongly that it cuts through his drive and is able to steer him back to me. You could argue that is a punishment unless I reward the act of him coming to me. He obviously wanted to chase the rabbit and I denied it with the recall (you could argue it's negative punishment, I withdrew him from something he liked, though most just call it 'redirection'), but then the act of recalling was rewarded (positive reinforcement), so the end result is that he's apt to recall since it leads to a reward, even if initially it means turning away from something he liked.
 
#18 ·
It means the intensity of the signal did not match or overcome the drive/focus/emotional intensity of the dog. If he's eating off the ground and I just tap his head lightly, I'm not abusing him, but my signal is not intense enough to make him pay attention to it. The same principle as when a cued behavior is not maintained (or performed) in high-drive or high-emotion situations.

If I wanted to, I could hit Wally on the head hard enough to make him stop doing whatever, probably even eating. However, the question is then, is that intensity or that signal one I should use on him?

Of course, hitting him is out of the question for me. But I could give his recall cue, a behavior that's conditioned so strongly that it cuts through his drive and is able to steer him back to me. You could argue that is a punishment unless I reward the act of him coming to me. He obviously wanted to chase the rabbit and I denied it with the recall (you could argue it's negative punishment, I withdrew him from something he liked, though most just call it 'redirection'), but then the act of recalling was rewarded (positive reinforcement), so the end result is that he's apt to recall since it leads to a reward, even if initially it means turning away from something he liked.
Naw, I think trying to call cuing an incompatible behavior as punishment is a pretty far stretch. Does Wally consider a recall as an aversive worth avoiding? I doubt it. And as long as you have plenty of money in the bank, and have spent a some time on varying your reinforcement schedule in some way, he's likely to see it as a possible opportunity to earn reward. (and, I would hope if you called him away from something really difficult, you'd give some sort of reinforcement, even if you don't have food on you..
 
#15 ·
On a trick training forum, I just stickied a thread about this. It's a confusing concept for most. but it's been well explained here :)

Both books are very good at explaining OC :)
 
#20 ·
I've always had trouble with the concept of "it's only punishment if it works" thing. If someone smacks the snot out of their kid for doing something, and he runs right out and does it again, it's not punishment? Then what IS it called? Yeah, that's confusing for me.
 
#22 ·
I think there's effective punishment and ineffective punishment, but it's still punishment. . .
 
#23 ·
Yet sometimes the punishment is enjoyed, then it's not punishment. same with rewards, it's only a reward if the dog (or child) wants it. Sometimes what we think is a reward, might be a punishment for the dog, or vice-versa.

eg.1- dog acts up when he sees another dog. you think that removing the dog from the situation is punishment (-P) because he wants to play, in reality it's a reward (+R) because the dog is afraid of the other dog. In this instance, you've rewarded the dog for acting up, and your dog will continue to act in this way as he was rewarded for it, when you thought you were decreasing the behaviour by punishing them for it.

eg2- You reward (+R) your dog for sitting by vigorous petting, when in reality, the dog does not enjoy the contact and feels uncomfortable and therefore punished (+P). The dog is going to be less likely to sit next time because you punished him for it, when you thought he would be more likely to sit because you thought you were rewarding him.
 
#24 ·
But what if the punishment IS clearly aversive, obviously not enjoyed, but isn't effective for whatever reason?
 
#29 ·
It's still punishment. Really, as someone who was emotionally damaged by physical punishment, which I found VERY aversive and extremely unpleasant, but never once changed my behavior (in the desired manner anyway), I can't agree that ineffective punishment needs to be escalated or that it's not punishment. It needs to be changed but not escalated, and it may be a waste of time/energy/emotional health, but it's still punishment.
 
#30 ·
I'm sorry for what's happened in your past, but what KBLover is trying to stay that in the theory of operant conditioning (which is what this treat is about), what we might consider a punishment is not a punishment within the context of operant conditioning unless it reduces the frequency of the behavior. That is how it is defined within the context of operative conditioning, which has its own set of rules and definitions outside of how those words are used in day-to-day life. Just like "positive" and "negative" are also defined slightly differently within the same theory than normally used.
 
#31 ·
Yes, the secret geek-speak that was previously denied to exist :p. Really, it's super confusing to the average dog owner.
 
#33 ·
LOL, I could probably understand it if it was rocket science. But no matter how much I think I understand the principles of dog training, I have never successfully trained a dog to do/not do anything (except potty training, which is more of a management/guiding of natural behaviors kind of thing).
 
#36 ·
No, I've tried and tried. I took Willow to 4-H classes. I got pretty good at making her offer a "submissive response" (4-H teaches that all "corrections" must be severe enough to elicit a submissive response), but no actual trained behaviors (and she didn't trust me). I took Penny to the Kennel Club training, and I can make her shriek and thrash around when she feels a prong collar, but no trained behaviors (and now she doesn't trust me). I bought all the "recommend reading" books on positive training and tried some of that, but no success with trained behaviors (but at least the boys still trust me). I either have the worst timing in the world or there is something about training I'm incapable of understanding.
 
#38 ·
I don't click on videos/downloads from my phone so I don't know what they're about. . .but thanks. I'm just being whiny. My dogs are old and there's no reason they need trained behaviors now. Although it would be nice to have a dog that walks on a loose leash. . .I've never had that happen. But basically we're all comfortable with each other, no reason to shake up the status quo. When they're gone and I get a new dog I'll be scared to death, though.

To keep it on topic, yes, I would like an "Operant Conditioning for Dummies"! Hopefully it could explain the more confusing things in a way I could understand (ya know, more like rocket science ;)).
 
#40 ·
Dog training is confusing to me too. I used to dog sit to get more experience with dogs, in preparation for the time when I would actually have my own (i.e. now), and I had no idea what I was doing. It wasn't until we started going to a training class that it really started to click. For me, I understood the theory, but I didn't understand how to properly apply it. I'm still not very good at applying it to situations outside of what we've learned in class, and I intend to keep taking classes as long as I can afford them and I feel like we (both me and Snowball) are learning from them. Hopefully one day I'll be able to train my future dog all on my own! :biggrin1:
 
#42 ·
1. One of the reasons that OC is confusing is b/c it is not all encompassing, it is only a small portion of the psychology and the learning that is going on. And, in everyday life you recognize that, even though you may not be able to articulate it. For example, if you give the dog a treat and he "thinks" 'I'm a good dog!' then that is NOT OC, b/c the dog did not present an observable change in behavior .... which is a defined pre-requisite for Behaviorism and operant conditioning (Hey, it's psychology, not necessarily life :) ).
2. There are other types of psychology: Learning theory, Education Psychology, Cognitive Psychology, and so on; which encompass more aspects of 'life' than Behaviorism, but behaviorism was there first :)
3. For this thread and for Behaviorism and OC only... Punishment must result in a change in behavior... The same is true with Reinforcement .. If you don't get an observable change in behavior (Stopping a behavior or Continuing a behavior), then "Nothing" happened !!! This is a limitation of behaviorism, b/c as dog owners many of us can see that 'something' happened ... and then a pure Behaviorist would say, "You got an observed change in behavior" and that discussion might result in a circular path that doesn't get you anywhere... But that's OK, b/c as long as there is an observable change ... :)
4. I hope that's not too confusing ... Behaviorism is effective, but limited and incomplete.
5. If you hit your dog, that may not be punishment ... ask any Lab, Rott, or Pit owner... especially if we hit the dog with our hand... And the dog turns, looks at you, and says, "wanna play, wanna play !!!"
6. I still like Ian Dunbar as a practical OC for Dummies, as well as Karen Pryor's original "Don't Shoot the Dog!" from the early 1980s.
7. One method for Sit: http://www.youtube.com/watch?v=q1W_3CDVVqo
8. and Down: http://www.youtube.com/watch?v=0I0O6ktNJuA&feature=relmfu
 
#43 ·
1. One of the reasons that OC is confusing is b/c it is not all encompassing, it is only a small portion of the psychology and the learning that is going on. And, in everyday life you recognize that, even though you may not be able to articulate it. For example, if you give the dog a treat and he "thinks" 'I'm a good dog!' then that is NOT OC, b/c the dog did not present an observable change in behavior .... which is a defined pre-requisite for Behaviorism and operant conditioning (Hey, it's psychology, not necessarily life :) ).
Depends. If I give Wally a "random" treat:

-He may actually keep doing whatever he was doing. If he was just standing there - he's going to keep standing there with his eyes wide and ears out. No doubt alert for more "food from the sky".

-He may get wound up, likely expressing his happiness and it coming out through increased energy and activity. Getting playful, barking, trying to draw attention, etc.

-He may offer something else. "If that worked, maybe this will, too." Perhaps an indication that we might be starting a "shaping game".

-He'll likely start sniffing the area of none of the above happen "for too long" (how ever he determine that), or if the above happen, but then stop "for too long".

Those are changes in his behavior patterns, so it there was something to observe, but all of those may not fall under OC, but still resulted in changes in behavior pattern (i.e. if I didn't give a "random" treat, he wouldn't have done anything differently than he usually does in that situation).

2. There are other types of psychology: Learning theory, Education Psychology, Cognitive Psychology, and so on; which encompass more aspects of 'life' than Behaviorism, but behaviorism was there first :)
How do you use them to train/communicate with a dog, though? I absolutely want to go beyond "just" OC, but I have not found any methods to using, say, Education Psychology, on Wally. So I stick with using OC and CC in a lot of varied ways as I can think of, or just making him figure stuff out with shaping. The only other non-OC/CC method I've seen and have been able to utilize with some success is calming signals. Not OC/CC, but can and do change behavior.

3. For this thread and for Behaviorism and OC only... Punishment must result in a change in behavior... The same is true with Reinforcement .. If you don't get an observable change in behavior (Stopping a behavior or Continuing a behavior), then "Nothing" happened !!! This is a limitation of behaviorism, b/c as dog owners many of us can see that 'something' happened ... and then a pure Behaviorist would say, "You got an observed change in behavior" and that discussion might result in a circular path that doesn't get you anywhere... But that's OK, b/c as long as there is an observable change ... :)
I don't see that as a limitation of behaviorism. Maybe a limitation in observing behavioral changes on the part of us owners/trainers. Something always happens. The questions are, imo, "What happened?", "Can I use it?", "Can I replicate it?", "Is it what I wanted/something that is desirable/useful, and if not, where did I go wrong?", "How did the dog interpret the consequence and to what did behavior did he see it applied? (perhaps I decreased another behavior instead, but since I wasn't looking for that behavior to change, I conclude that 'nothing happened', when the real conclusion is, 'my observation skills still suck and need more work')"
 
#44 ·
Not sure if this is nitpicking, but I think that "He may actually keep doing whatever he was doing. If he was just standing there " may fall under the category that nothing changed... when you know that he's 'thinking' more "food from the sky???". To me, this is a difference between behavioral vs. Cognitive.

I think you're already advancing beyond pure behavioral methods, b/c Wally has learned how to learn ... And you're having 'trouble' figuring out what else to teach and how else to challenge Wally, as he learns increasingly complex task, more and more quickly. As I warn folks with border collies: Don't leave Wally alone with the keys to the car. :) Three suggestions:

1. Sue Ailsby's Levels of training (Doggies Zen?) http://sue-eh.ca/ strikes me as an excellent application of of using clicker training to teach a dog how to learn increasing complex rules, effectively learning how to learn ... which is a foundation of Cognitive Psychology and education psychology.

2. Rico and Chaser http://www.nytimes.com/2011/01/18/science/18dog.html?pagewanted=all AND
http://www.youtube.com/watch?v=_6479QAJuz8 are examples of Fast Mapping, which can't be explained under any Behavioral theories.... Although we 'know' that our dogs think , proving it takes researchers a bit longer.

3. Calming Signals - I've observed and reacted to my dog's calming signals, ever since I learned about them. Altho Turid originally implied that they were 'pre-aggression' and not trainable, she no longer teaches this idea in her courses, and had recommended that we observe and try to understand what we see. I did, and my dog used to wink at me with 'anxiety'. But now he may wink at anticipation of play, and may yawn as a 'request' for a bit more belly rubbing (or an indication of his dissatisfaction that i stopped :) )
My point is that enhanced communication (like teaching an infant Sign Language), provides an additional level of interaction... I don't have the time and patience for a more rigorous, randomized, double-blind study with a control group. But, anecdotal, wishful thinking is adequate for me right now :)
 
#45 · (Edited)
Not sure if this is nitpicking, but I think that "He may actually keep doing whatever he was doing. If he was just standing there " may fall under the category that nothing changed... when you know that he's 'thinking' more "food from the sky???". To me, this is a difference between behavioral vs. Cognitive.
Could be. Won't say you're wrong or even really disagree. Where does anticipation/pattern recognition or "wait and see what might happen" fall in? Cognitive sounds as good a place as any. Of course, if he stands there instead of the more typical lying down, would that also not be a behavioral change? The "random" treat came while he was still standing - so now he keeps standing. A case could be made for both. In other words, "I got a treat for standing, and just in case it happens again, I'll keep standing in this spot." You could even argue that's how +R actually does work, or the why of "why positive reinforcement works". Perhaps it's why learning tends to accelerate under +R methods, it might combine both behavioral and cognitive more readily.

I like to say "behavior is a window to the mind", which is why I'm primarily became interested in behaviorism - because I REALLY want to get in his mind. I found that if I look close enough (a skill I'm still learning), I can "see" what he's thinking based on what he does and when his behavior changes, especially during learning something new or a different spin on something, what he's thinking about changes, especially once he recognizes the difference in the scenario. He's like "oh now I see!" and his behavior changes in response. I see the behavior change, it's lining up towards the correct answer, and I can say 'Yep, he's starting to get it'.


I think you're already advancing beyond pure behavioral methods, b/c Wally has learned how to learn ... And you're having 'trouble' figuring out what else to teach and how else to challenge Wally, as he learns increasingly complex task, more and more quickly. As I warn folks with border collies: Don't leave Wally alone with the keys to the car. :) Three suggestions:

1. Sue Ailsby's Levels of training (Doggies Zen?) http://sue-eh.ca/ strikes me as an excellent application of of using clicker training to teach a dog how to learn increasing complex rules, effectively learning how to learn ... which is a foundation of Cognitive Psychology and education psychology.

2. Rico and Chaser http://www.nytimes.com/2011/01/18/science/18dog.html?pagewanted=all AND
http://www.youtube.com/watch?v=_6479QAJuz8 are examples of Fast Mapping, which can't be explained under any Behavioral theories.... Although we 'know' that our dogs think , proving it takes researchers a bit longer.

3. Calming Signals - I've observed and reacted to my dog's calming signals, ever since I learned about them. Altho Turid originally implied that they were 'pre-aggression' and not trainable, she no longer teaches this idea in her courses, and had recommended that we observe and try to understand what we see. I did, and my dog used to wink at me with 'anxiety'. But now he may wink at anticipation of play, and may yawn as a 'request' for a bit more belly rubbing (or an indication of his dissatisfaction that i stopped :) )
My point is that enhanced communication (like teaching an infant Sign Language), provides an additional level of interaction... I don't have the time and patience for a more rigorous, randomized, double-blind study with a control group. But, anecdotal, wishful thinking is adequate for me right now :)
Ailsby's levels are one of the first things I ran across when looking for things to teach Wally and a sort of progression :D (which is nice for me since I'm more...I guess scatter-brained LOL. I want to teach this, and oh hey this looks cool, I'll start this and so on. Amazing he's progressed as he has under such "wild west" type training!). I should revisit them.

2. Fast Mapping. Now I have a term for it to search google with. Thanks!

3. I LOVE calming signals. I think discovering those really gave me a "push" to see another side of things and what else I might have been missing with Wally's body language. (A lot, it turns out. I noticed other ways he tries to show me what he wants along with behaviors he developed from basic ones he's learned in an effort to communicate).

Heheh I hear you about the rigorous methods. Anecdotal is good enough for me, too. After all, if it works with Wally, does it really matter if it was 'disproven' in a fuller study? I already know I have the "10th dog" (i.e. 9 out of 10 dogs do something one way, Wally is the 10th dog that's not 'doing it right' LOL)
 
#47 ·
@Gingerkid - I think 'calming protocol' is a method to cue your dog to 'settle.' "Calming Signals" are an informal but rigorous (?)observation of dog's body language theorectically related to wolf body language. They are frequently overlooked expressions which more subtle (and may be precursors to) tail wagging, snarling, barking, etc. There are 30 basic calming signals, and I believe there are about 20 more 'variations.' Google: "Calming Signals" "Turid Rugaas" to get a good taste. Take what you read with a grain of salt, because they aren't rigorous science, but more at the observation/categorization level of research. The Knowledge will improve your observation skills.

@KBLover - OK, I'm wrong and I disagree! :) Practical implementation is different than limited theory. And from a definition perspective, behaviorism requires a blackbox, no theory of mind, and change of behavior is the ONLY indication of learning. Of course, we know better in fact... But are we discussing fact or theory :) Floor wax or dessert topping?

>>Where does anticipation/pattern recognition or "wait and see what might happen" fall in? Cognitive
I Agree, but I want to distinguish between the two approaches. I think folks have trouble with OC b/c they can see their dogs thinking... "getting it" ... and OC requires a change in behavior. OC is proven to work (I agree with this) and the research in cognitive methods used in dog training is not yet proven to work. However, these methods are used in school. Researchers just haven't had the funding to create a Cognitive curriculum for dog training IMHO.

Initially, I believe that an 8 week old puppy (or maybe a little younger) really is close to a 'mindless' black box and the behavioral methods work mainly as described... But, the light grows brighter, the puppy learns to adapt (better than we do), and 'reasons out' that he can get "food from the sky" by pushing the right button. Ever watch Wally, early on, throwing a Sit or a Down... trying to teach you to give him a treat : -) At that point, I think Cognitive methods would be more effective, if we knew what to do... And, I believe that the Thinking Dog gives a behavior without treats, b/c he is thinking. So, +R is needed early on from a kinda black box perspective, but I think it is only a marker ("Yes, that's what I mean") as opposed to ("Good Dog, Click, Here's your reward")... as the dog learns how to learn, learning that words have meanings. Early on, if you say Sit, you get a Sit. Then you teach Down... and you also get a Sit... and with enough cues, I believe the pup discriminates.... Cognitively... thinking ... rather than just reacting behaviorally...

I think you are already well into Wally's mind... and if he is like Shep ( or better!!!!) then he already knows more than you recognize :)
 
#48 · (Edited)
@KBLover - OK, I'm wrong and I disagree! :) Practical implementation is different than limited theory. And from a definition perspective, behaviorism requires a blackbox, no theory of mind, and change of behavior is the ONLY indication of learning. Of course, we know better in fact... But are we discussing fact or theory :) Floor wax or dessert topping?
Which ever one is dessert topping because I'm hungry LOL :D

>>Where does anticipation/pattern recognition or "wait and see what might happen" fall in? Cognitive
I Agree, but I want to distinguish between the two approaches. I think folks have trouble with OC b/c they can see their dogs thinking... "getting it" ... and OC requires a change in behavior. OC is proven to work (I agree with this) and the research in cognitive methods used in dog training is not yet proven to work. However, these methods are used in school. Researchers just haven't had the funding to create a Cognitive curriculum for dog training IMHO.
True, but, on the road to "getting it" the behavior is constantly changing. Maybe because I'm so shaping addicted I see it that way since shaping, you hardly ever get the whole thing but see little 'changes in behavior' that signal he's "getting it".

Or maybe shaping is purely cognitive with the OC part just being the reward marker?

Initially, I believe that an 8 week old puppy (or maybe a little younger) really is close to a 'mindless' black box and the behavioral methods work mainly as described... But, the light grows brighter, the puppy learns to adapt (better than we do), and 'reasons out' that he can get "food from the sky" by pushing the right button. Ever watch Wally, early on, throwing a Sit or a Down... trying to teach you to give him a treat : -) At that point, I think Cognitive methods would be more effective, if we knew what to do... And, I believe that the Thinking Dog gives a behavior without treats, b/c he is thinking.
If I had a dollar for every time Wally does that, I'd be rich and probably could fund that research for a Cognitive training method :)

I mean, all I have to do is have something he'd like to have and he's going to town. If I wait long enough, he'll start looking around for things to do! I showed that to my mom and she was like "that's kinda scary, he's looking for things to do something to". Then with the throwing, she was like "he's trying to think what object to throw - literally looking at each one and picking one out, like he saying 'this one is more likely to be the right one'." And then I really freaked her out when I took away each thing as he threw it (which also prompted a bark each time like "HEY, I can't throw it if you taking it!"). Then when there was nothing left, he started looking around the room. That just freaked her out. "He's looking for something to throw!" He'll be looking around and whining ("I can't find anything!" behavior?). Eventually, he tried to throw the throw rug (hey, it's a throw rug, right? So does that mean it's a rug I throw? *rimshot*)

I notice he almost gets annoyed (at the object) like "why isn't this stupid thing doing right to make him give me a treat??!!" He shakes it harder, slaps his paws on it like he's pissed. LOL

So, +R is needed early on from a kinda black box perspective, but I think it is only a marker ("Yes, that's what I mean") as opposed to ("Good Dog, Click, Here's your reward")... as the dog learns how to learn, learning that words have meanings. Early on, if you say Sit, you get a Sit. Then you teach Down... and you also get a Sit... and with enough cues, I believe the pup discriminates.... Cognitively... thinking ... rather than just reacting behaviorally...
I agree. He's definitely thinking about what the word means. I also agree with your R+ idea that it is a marker that shows "that's what I mean" or "that's what that means". Then after that, there's definitely thought going on. If I had a dollar for his "pondering moments", I'd be rich. The staring at an object trying to figure out what 'game' I'm playing, the stop and look back when he's not quite sure what I said "did he really say 'abajo' or something else? I should check." Those "is this it?" looks at me during shaping. Or the "HEY, LOOK AT THIS AND SEE IF ITS RIGHT" bark/noises if I don't notice quite on time he's tried something else (I look away as my no-reward marker during shaping).

I think you are already well into Wally's mind... and if he is like Shep ( or better!!!!) then he already knows more than you recognize :)
LOL I wouldn't be surprised. :D

And, yeah, I will take your "don't let the border collie have the car keys" to heart :D
 
#49 ·
We can add something that the dog likes when he does what we want.
We can remove something the dog likes when he doesn't do what we want. WE can add something the dog doesn't like when he does what we don't want.
We can remove something the dog doesn't like when he does what we want.

The consequences have to meaningful to the dog and they have to be timely. If I give him food for sitting and he is not hungry it is not meaningful. If I take him in the house after training and give a hungry dog a steak, even though meaningful it is not timely.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top