Jump to content
BC Boards

Cesar's method - disagreement and discussion


INU
 Share

Recommended Posts

  • Replies 153
  • Created
  • Last Reply

Top Posters In This Topic

Sorry, I was away for a few days.

 

This is a little off Cesar Milan, but I'm curious about what you use to communicate with your dog that the dog is doing the correct thing. You said you don't like the syntax of click/treat--does that mean you don't like the syntax of mark/reinforce or does it mean you don't like the "language" with the components that are clicks and treats (on the generative assumption that the syntax is the same regardless of the language)? If it's actually the syntax you don't like, what's the string that you do like?

 

What I don't like is extrinsic or extraneous reinforcements (treats) or promises of same (clicks). I like intrinsic reinforcement -- reinforcement that grows out of what you're doing. That could be sheep in sheepdog training (though sheep are not as easily and precisely granted or withheld as one would think from reading some of the posts here). In training other things, "good dog" is an intrinsic reinforcement. It's an integral part of our joint endeavor -- an answer to her in response to her question to me -- and has no reference to some extraneous treat or reward. I guess it would be marker and reinforcer all in one. It tells the dog he was right, and in my experience these dogs derive satisfaction from being right. They like to work with you, they like to solve problems, they like to be right. It's bred into them. And it's a bred-in trait that I want to encourage and develop and expand, so that it's the basis of what we do together. IMO, click/treat extinguishes it. It doesn't let it develop. It's like offering money to someone who's trying to do you a favor. It eclipses a natural reward with an unnatural one.

 

Generally I do too, but I think it's a mistake to believe that clicker training is what produces such a dog.

 

I think we're just going to have to agree to disagree on this point. It has been my experience that clicker training absolutely produces such a dog!

 

I think you misunderstood me. I'm not saying that offering behaviors won't result from clicker training. I'm sure it will, I'm not at all skeptical about that. If you tell someone, "I'll give you $1,000 if you can figure out what I want you to do," most people will expend at least some effort in offering behaviors. What I'm saying is that offering behaviors is already built into your dog (at least if it's a border collie), and you're misallocating credit for it to the clicker training. It's there already -- you can have it without ever clicker training the dog. You don't need to click to get it. Those of us who don't clicker train get it too.

 

I hate to break it to you, but in that situation I doubt the "no" was actually a correction (or punishment in behavior theory terms) - to you dog it appears that "no" is actually a cue to do something different rather than a punisher. Does your dog find "no" to be aversive at all? At least the way you typed it, that doesn't seem to be the case and if it's not aversive, well then it's not a correction/punishment. In behaviorist terms the punishment must stop or decrease the behavior, yes, but it generally doesn't also encourage a "try again" mentality as it is an aversive.

 

No, it wasn't aversive, and if something that stops or decreases behavior has to be an aversive to qualify as a punishment in behavior theory terms, then it wasn't a punishment in behavior theory terms. But it most definitely IS a correction in MY terms (I don't think behavior theory uses the term "correction," does it?). A correction may or may not be an aversive, but it communicates to the dog that what he did is not what I want. Therefore, in this particular case, where she knew I wanted something because I'd given her a heretofore-unknown command, she continued trying to figure out what I did want. If I used "no" to correct her for barking, and she already knew from experience that that meant I didn't want her to bark, she would fall silent. I guess you could call that "no" a cue to do something different (not bark)? But I don't find behavioral terms all that congenial, and I would simply call both those "no"s corrections.

 

When I had Maggie on sheep, the clicker and treats never figured in, but I used the sheep as a reward just like treats or a toy in another setting; when Maggie did something I disliked, she was prevented from accessing sheep (-P) and when she behaved correctly, she got to interact with the sheep (+R).

 

When my dog chases a squirrel, I prevent her from accessing the squirrel (-P). (Actually the squirrel does that for me, by running up a tree.) But somehow that doesn't keep her from chasing a squirrel the next time she gets the opportunity. :rolleyes: I'm not saying that there's no way in which you can manipulate access to sheep in training to reinforce or punish. But I'm saying that method is very much more limited in its feasibility than you're making it sound, and is a relatively small part of the methodology of sheepdog training. Chasing sheep is rewarding, even if the chase is cut very short by your spectacular reflexes and footwork. Sometimes chasing sheep for two seconds is more rewarding than quietly moving sheep for minutes on end. Sometimes just trying to get past you (and seeing the sheep react to her efforts to get past you) is plenty rewarding in itself. Sometimes she's chasing in response to pressure you're not aware of, and unless you diagnose and fix that pressure no combination of +R, +P, -R, -P is going to help. There's just more to it than that very simplistic behavioral model of "us[ing] the sheep as a reward just like treats or a toy in another setting."

 

And one last clarification - when I said "harsh" in my last post I wasn't trying to imply that something was inhumane or "mean, more that it was "more negative" or "punishment based" than my chosen methods. My questions still stands by the way - if a gentle +R/-P method works, why skip it without giving it a try before going into a -R/+P based method??

 

I don't know what I can say that I haven't already said. My dog and I are in a relationship, and in a mutual endeavor. I recognize the importance of reinforcement to all beings, but I would rather her reinforcement came from something intrinsic to the endeavor or the relationship than from something extrinsic to it. I don't want to distract her, and then have to "fade" the distraction. I think corrections are more direct and efficient as communicators (and less patronizing to her) than "take her far away and begin giving her treats as you move closer." I think training that incorporates correction builds character in a way that all-rewards training does not. I don't want to extinguish her ability to "sing like she don't need the money."

 

For those who imply that +R dogs are only doing things for clicks or treats... couldn't we just as easily say that dogs trained by adversives are only doing things to AVOID the adversives? In neither case is the desire to run through the tunnel a reasoned decision on the part of the dog. Both of these methods are ways of communicating to the dog that it's in his best interest to do what the human says. I've never understood the certain level of derision that exists about "bribing" dogs with clicks when compared to "bribing" them with the promise of no leash tug.

 

But as MaggieDog points out, I didn't use or threaten an aversive to get her to run through the tunnel. There's no need for rewards or aversives to get a border collie to do something like that. Just help her to figure out what you want, and let her know when she's succeeded. For the vast majority of activities like this there's no need for rewards (other than a simple "good dog") or aversives, because the dog has no antipathy to doing it or not doing it. He's happy to join in figuring it out and getting it right. I once saw a video made by a very poor trainer in which he demonstrated how to get a dog to jump in a pickup by jerking it with a leash. I reacted as negatively to that as I would to luring or rewarding with treats to get the dog to jump in the pickup. Both seem cluttered, distracting and diminishing to me. Seems to me it's good for you as a trainer, and for your ongoing relationship with your dog, to show him you want him to jump in the pickup and get him to do it without resorting to treats or aversives, and it just plain ain't that hard.

Link to comment
Share on other sites

I am really totally confused by this whole discussion. Or do I have weird BCs? The ONLY time my dogs are given treats is when I just feel like it. It is never to teach them anything. Y'all know that Skip will chase Jackson, chasing the frizbee/ball. I didn't mind that but he would grab Jackson's neck ruff or cheek or something. I finally started yelling at Skip to "get outa that!" which is a sheep working command, but can be used in other instances. Like this one. I would have to run (that's a sight!) towards Skip and yell it to him. Couple times was all it took. He got good boy when he stopped, get outta that when he did it. He knew what I wanted. He acted accordingly. EVERY thing my dogs know, they know cuz I explained what I expected and that was that. To me, treat training or clicker training is so time consuming! It places so much of the dogs mind on the treat. Treat trainers say they eventually wean them off the treats. Why start? I have not ever spent more than a few minutes teaching them anything I wanted them to do. Not saying I haven't had to reinforce those things because, they are like kids, sometimes they try to get by with not doing things they are suppose to. I guess the thing is with me is that treats for tricks is bribing. I want to tell the dog what to do and he does it because I have asked him to do it. And he likes me well enough to follow through.

Link to comment
Share on other sites

Linda, I tend to agree with you in one way. Missy lives to please me. And she works hard to figure out what I want. The idea of using treats with her is silly IMO. Kipp on the other hand is a bit different. He is a little more independant and self serving. If I ask Miss to jump she'll say "how high?" but Kipp is more of a "why?" dog. However, he is a food hound. So he gets to work for his meals. Right now he's got to earn every piece of food he eats, and he has become a more eager pupil because of it.

 

I prefer a slightly more biddable dog, but right now I'm working with what I've got. I'm hoping that through working for his food it will improve his overall desire to work for me.

Link to comment
Share on other sites

 Share

×
×
  • Create New...