Negative reinforcement is not withholding a reward in the face of bad behavior. The Skinner-observed/invented concept is that a negative stimulus is avoided by not doing the “bad” behavior. Connecting the removal of the food with the “bad” behavior in the case of a dog not sitting when told is problematic when the dog would have to stop eating (or has not yet started) to obey the command.
The point of observing the dog interaction is not to copy dog behavior to communicate with dogs. Watching the dogs communicate with each other should reveal the simple fact that dogs do not fall to pieces and become neurotic wrecks in the face of a negative physical consequence delivered by another dog. When a young dog play-bites an older dog too hard, the older dog typically corrects it. The older dog does not use a treat or a long conversation to do this, and happy play typically continues. The problem behavior does not persist for weeks or months and no experts are consulted.
Eh, many people are dumb. If they have no intuition with the animal, probably no other training technique would work for them either. That isn’t the fault of the person showing what works for him.
Maybe not, but these are typically cases where human incompetence has created the situation in the first place. If he is fixing the problem and ending up with a happy, safe dog instead of one that mauls someone or ends up euthanized, I can’t complain too much about his methods. Working with a dangerous dog is really not something anyone should attempt based on watching someone else do it on a TV show. At least people watching might learn that these are not all hopeless cases, even if they are not the ones who should attempt to fix them.
A person who watched a different show and thought a pocket of dried liver bits and a clicker was going to get them through a training session with a “red zone” dog would also be wrong.
I’m not sure how it’s confusing to the dog. He sits, he is released to eat his breakfast. He does not sit, the opportunity for breakfast goes away until he does as asked. No, he doesn’t have much of a concept of time, but he does understand that if he wants his reward, he will need to perform his cue.
I see and hear a lot of “the dog will only do it for a treat” in regards to positive reinforcement training, and that’s simply not true. A good trainer will fade the treat out as the dog comes to know his cues and substitute real life rewards. All of my dogs are clicker trained, yet I rarely pull the clicker out and rarely reward them with treats. They do have to go to their places and stay for meals, other than that I generally only use food to teach new cues.
They know what our body language means for them. Your dog doesn’t understand that eating cat poop is “bad”, he understands that when you approach him while he has his head in the litter box = something scary happens. I would use (and have used. I live with 4 dogs and 2 very poopy cats) my inferior treats to train a strong leave it and block access to the cat box when I am not there to enforce it.
Do you survey the owners of every well-behaved dog you meet? None of them use treats? Only the out of control dogs were trained that way?
I disagree. I know literally hundreds of dogs who were trained without physical guidance and the majority of them are nicely behaved. The dogs with owners who do not do the work or “spoil” their dogs are obviously going to end up with bratty dogs, the same way that dogs with owners who take a class or two of choke chain training and then never train their dog again are likely to end up with ill-mannered dogs.
I no longer have cable, so I haven’t seen his show in years, but I have seen him kick and pinch dogs on pretty much every episode I watched. Rolling, flooding, intimidating. All common methods for Millan and not only dangerous for the at home audience to repeat, but detrimental to the dogs he’s abusing. There is a clip of him kicking a dog that later bites him that gets taken down every time someone uploads it to YouTube, I wish it were up now so I could link it.
I think you are misunderstanding positive reinforcement training, with all due respect. It is not just a treat free-for-all where cookies are doled out with wild abandon. There are consequences for not responding to a cue, they just aren’t physical corrections. The food is used to teach the cue and reinforce it until the dog knows it solidly in a variety of situations, it’s not a bribe. I know a lot of people miss that part and continue to use cookies for their dog’s whole life. That’s not ideal, but I would rather have them bribing their dog than hurting it.
There is also the question of timing. If you are delivering physical corrections to the dog, and your timing isn’t fabulous, the dog isn’t going to get the message. The same is true of rewards, but a mistimed cookie is less wear and tear on the dog than a mistimed collar pop. It also doesn’t put the dog into a state of fear where they aren’t learning anything but to avoid.
I don’t train every dog I see, but I train every dog that comes through my office door.
<checks forum> IMHO, I am inclined to agree with AnaMen, that being physical with a dog (both with discipline and reward) is not the same as being mean to them, or hurting them.
When our dog was young and still learning, he’d excitedly jump up on me, and others, as we entered the house. I employed the raised-knee method (which knocked him in the chest and he fell down a couple of times) maybe 5 or 6 times, and he learned not to do that. Was it physical? Yes. Did he learn the preferred behavior? Yes. Did he get hurt? No. Shouting at him “down, down, down” is about as effective as parents who are always counting “1, 2, 3…” to warn their toddlers about something (those parents are always counting and the kid has learned to ignore).
He is a very good boy now, but we have a rule that he cannot beg at the table, and must sit on his bed across the room, until we are finished. Unfortunately, the kids tossed him a scrap here or there, so he learned there is a non-zero chance he could score again. Like all dogs, he is an expert observer of people, and he knows when we are nearly done eating, and starts to inch toward the table. When caught someone barks at him, and if he persists, he ends up crated (I know that isn’t necessarily right). But, he still follows the same pattern nearly every night.
My point with these two examples: sometimes getting physical with them is needed and is effective, and sometimes being non-physical is totally ineffective. I am not saying you need to club your dog or hurt them or be violent - that is mean, but again, IMHO, dogs respond positively to physical discipline and reward, as long as both are used moderately.
Also, dogs need your response if they did something wrong immediately if they are going to understand what you want. They have no sense for what happened yesterday, or even 30 mins ago.
I wouldn’t teach a dog to not jump on people by shouting “down down down”. I also wouldn’t knee him in the chest. That runs the risk of hurting him more than you intended (puppies don’t have the greatest coordination) or teaching him that approaching people might result in punishment. It’s just needlessly heavy handed. Would you shove an exuberant toddler down if she came running at you to wrap her sticky hands around your leg? Or would you teach her that if she wants you to pick her up, she asks nicely? All of my dogs came to me as jumpers, all of them were trained without the use of physical correction, all of them now greet politely.
Your non-physical example: teaching a solid down stay on his bed with distraction would serve you (and him) better than having to shout at him and put him in time out nearly every night. This is ineffective in part because of the timing issue you mention. If you aren’t marking the undesired behavior as it happens, he has no idea why he’s in the crate by the time you get him there.
Yes, dogs need immediate feedback. Both good and bad behavior needs to be marked. I don’t think you should reward OR punish for things that happened yesterday or 30 minutes ago. If I can’t catch behavior as it’s happening, I can’t reinforce or discourage it except through management. That’s actually a big problem with physical corrections; they are rarely delivered with good timing and/or enough force to make a difference.
I do want to make clear; I don’t think that all physical guidance or corrections are abuse. Pushing a dog’s butt down isn’t abuse, neither is putting a knee up when he jumps (different from kneeing him). I don’t believe that either of them teach correct behavior, but I don’t think someone who does those things to their dog is an abuser.
[QUOTE=araminty]
I think the shelters are full of untrained dogs because some jerks like making puppies.
[/QUOTE]
Uh, the jerks are the people breeding puppies.
There’s a lot of ignorance in this thread about positive reinforcement training. Maybe its opponents could save the rest of us some time and energy by, I dunno, reading a book about it before dismissing it and arguing against it?
The confusing part is the departure of said “opportunity.”
Can a puppy really be expected to understand a figurative concept like “opportunity”?
I see a similar problem with your examples from a previous post. You say you “prevent the dog from proceeding through the door I’ve asked him to sit at”–but the sitting also prevents the dog from proceeding through the door, so you are either preventing him with his cooperation or without it, not preventing versus not preventing. You say you “stop the walk.” Obviously though, you’ve already walked some distance away from your dwelling. How do you get back now that the walk is “over”? Unless you pick it up and carry it home, not possible in many cases, I’m going to have to assume you walk it there. “This walk is over, so now we will walk some more” is a confusing concept, isn’t it?
Now I don’t doubt that these tactics result in improved behavior, but I think the thought process is probably more along the lines of “sometimes I think she’s going to give me breakfast, but then she just dicks around for a while” and “sometimes I think we are going to go through the door, but then she won’t get out of the way”. When you act this way, the dog learns that if it sits and waits, you will start acting normal again. The dog detects that you are not pleased and some dogs care enough about that for that to be enough.
We can’t criticize Cesar for people failing to execute his methods correctly but look only to positive-only trainers rather than the pet owners trying to use their techniques to judge the efficacy of positive training. Most pet owners never get past the treat stage and end up with a dog that doesn’t do what it is asked unless a treat is involved AND it wants the treat at that moment. If it actually performed the requested behaviors reliably for treats, even that would be something, but when there is an enticing distraction, the treat isn’t enough.
Which would then cause him to avoid repeating the behavior that resulted in the scary thing. Scary things happen to dogs near the litter box, so it is an area best avoided. This is a simple concept.
Here’s where “positive only” trainers always get very vague and hand-wave away all concrete information with the worn “train.” The truth is revealed in the next part: you block access to the cat box. Doesn’t that mean the dog is actually NOT trained to leave it alone?
It is actually very rare these days for me to encounter a well-trained dog. I do see dogs that behave well enough for the situation at hand, but no one is asking them to do anything they don’t want to do anyway. I spend time at the dog park daily. Daily I see dogs that “won’t get out of the car”, “won’t get in the car”, run away when their owners say it’s time to go, won’t come out of the shrubbery when told, won’t stop playing with another dog when the owner thinks play has gotten too rough, won’t stop humping other dogs while the owner bitches at them or waves treats around to get them to comply, etc. When anyone with an actually trained dog gives their dog a command, other owners talk about how mean the person is being to their dog, so they don’t come back after a while. I don’t have to survey anyone to know how they train their dogs, because it’s happening right before my eyes.
I know how positive reinforcement works and I do think it works: for encouraging desired behavior. I am a huge fan of positive reinforcement, BUT when it comes to discouraging undesired behavior, I prefer to couple it with more effective tactics.
Positive-only trainers love to talk about what NOT to do, so owners are left confused and helpless in the face of bad behavior, lecturing and nagging at the poor dog and failing to get the dog’s attention or communicate effectively. When the dog launches into a dangerous behavior–pouncing on a smaller animal or charging in front of a car for example–the person who thinks a treat is enough to stop it is seldom correct, while the person who physically yanks it away may save a life and discourage future repetition of the same behavior.
The out-of-control dog gets hurt when it loses its home. I would rather someone subjected their dog to a few ill-timed leash pops than never walk it because it pulls so much or give up on it entirely and take it to a shelter.
[QUOTE
There is also the question of timing. If you are delivering physical corrections to the dog, and your timing isn’t fabulous, the dog isn’t going to get the message. The same is true of rewards, but a mistimed cookie is less wear and tear on the dog than a mistimed collar pop. It also doesn’t put the dog into a state of fear where they aren’t learning anything but to avoid.
I don’t train every dog I see, but I train every dog that comes through my office door.[/QUOTE]
When it comes to dog training, timing and consistency are key for sure, although most dogs are not so fragile that they can’t handle a few mistakes. If you are using effective training techniques–meaning they work for you–that’s great. It doesn’t mean they will work for everyone and every dog and that others are inferior though. Dogs used to be expected to perform actual tasks (and some still are, of course), but now most pet-owners seem to believe that they are like human toddlers, to be restrained, redirected, and taken care of, while expecting little more than docility and occasional cooperation. My dogs have always happily taken on household duties, not needed to be micromanaged leash-walked, etc. for years. I trained and guided them effectively and they were happy and functional, not fearful.
He knows there was a bowl of food present one moment and gone the next. I verbally mark the undesired behavior as well as the desired behavior so he knows his actions (lunging for the food bowl) are what made the reward go away. It’s pretty black and white. Sit, and you can eat. Don’t sit, you cannot.
I admit to be confused about what you’re saying here. If the dog sits at the door, he is verbally marked and allowed to go out the door. He understands that the sit is what earned him the right to proceed. If he doesn’t sit, I verbally mark the non-correct behavior and step in front of him or shut the door. He understands that moving forward without permission is what lost him the opportunity to go outside.
By “stop the walk” I mean until the dog has given me slack in the leash, given me his attention, or responded to his “heel” cue, depending on what I can reasonably expect of that particular dog. The walk continues once the dog has given me a rewardable behavior. It’s another black and white situation. Walk with slack in your leash, we can keep going. Leash goes tight, walk stops immediately.
We disagree on this. The thought process is “I know how to get what I want”, whether what the dog wants is breakfast, access to the outdoors, or to keep walking. They know how to earn those rewards because I have taught them to exchange behaviors for goodies.
Something else we disagree on. I can and will criticize Millan for putting dangerous methods in front of the public, who then thinks that “rolling” their dog is a valid training method.
It does happen that lots of people don’t get past the treat stage. Lots of people don’t get past the collar pop stage, either, and spend years yanking at their dog’s throat and pushing their butts down. If I have to chose between a lifetime of carrying cookies in my pocket or a lifetime of choking my dog, I choose treats. Neither is ideal.
Or he avoids the litter box when you’re around but sneaks a treat when you’re not, or he associates your approach with punishment, or he becomes hand shy because hands do painful things to him, or he becomes phobic about the room the litter box is in, or a hundred different outcomes.
What am I being vague about? I have trained my dogs to leave the litter box alone by teaching them a firm leave it cue and I block access to the box when I am not there to enforce the cue. I don’t expect them to leave it alone when I am not there, because I have not taken the time to train the cue to that extent. It’s not a priority to me because the amount of effort it would take to train it to that level exceeds the hassle of shutting a door. Could I train them to leave the cat shit alone even if it were the only thing in the house to eat for days on end? Yeah, I could. I don’t care about it enough to do so. That’s not a failure in the method, it’s laziness on my part.
No offense, but I don’t think you do. You keep repeating some version of “the dog will only do it for a treat” and that’s simply not how effective PRT works. The treat is often used to train the behavior initially, because food is something most dogs will work for, and then faded as the cue becomes solid. I have trained plenty of dogs with low food drive by using other things as rewards, and I rarely use food to work with my own dogs. The same way you said we can’t blame Millan for people executing his method incorrectly, you can’t dismiss PRT because some people don’t do it properly. People stop at the first step and don’t move on, that’s true. But that’s like saying exercise doesn’t work for weight loss because lots of people buy the fancy sneakers and then leave them in the closet after two spin classes. It works if you follow through, just like everything.
I don’t know how many positive-only trainers you know, or how many positive reinforcement classes you’ve taken, but we absolutely do talk about and teach non-physical corrections and consequences. PRT is NOT about cookies and coddling, this is where I am getting the impression that you don’t understand it. Obviously if a dog charges in front of a car I’m going to yank it back, but I don’t assume I’ve trained anything in that moment, I’ve just prevented an accident.
And I would rather they used science-based methods to train the proper behavior than to jerk away at the dog until he gives up or shuts down.
Most dogs aren’t that fragile, some are. Dogs with aggression or other behavioral issues (as opposed to dogs who are mentally solid and just need basic obedience training) often are fragile. They wouldn’t be displaying those behaviors if everything was A-OK upstairs. Add the anticipation and existence of pain to a dog that is likely already fearful and you have a recipe for disaster. Add rewards, effective communication, clear leadership and patience and you will get a dog who is more confident, more obedient, and mentally healthy.
Regardless, the fragility of the dog isn’t a deciding factor in whether or not I am willing to hurt or scare a dog. The blank slate of a sturdy puppy doesn’t give me license to start smacking him or stringing him up to teach him or to punish him.
One more thought; training incompatible behaviors, making those behaviors rewarding, and removing the reinforcement for the undesired action is how I discourage “bad” behavior. Dogs who bark at visitors can be trained to present them with a toy instead, a dog who is reactive on leash can be taught to watch his handler instead of scanning for other dogs, jumpy greeters can learn that going to their bed and lying down when the doorbell rings is more rewarding than leaping up. If they are on their bed, have their eyes on me, or have a toy in their mouth they can’t do the thing I don’t want them to do.
Preventing “bad” behaviors by reinforcing incompatible “good” behaviors is something smart trainers teach.
Yes, people still do. However, you are comparing real life to a movie (if there is reference to the act then most likely it has happened, is happening now or will happen again. Generally directors won’t allow the abuse or even any harm of animals, but for entertainment in Mexico horses are brutally abused to perform for entertainment purposes (mostly by wealthy, disgusting men who would kill even another man for disrespecting him). Times and places change the scope of acceptable mean of behavior and are influenced by surrounding culture and level of education as well as religious belief and economic structure.
Positive reinforcement and assessment is the best way to train a dog. Sometimes a dog needs discipline, but training a dog to work for a human proves that human is lazy, lonely, incompetent and mean. Dogs are not capable of taking care of themselves. So to have one for work purposes should imply their deserving reward and gentle care. My dog follows me where I go and tries helping me with what he is capable of. He can open doors, open bottles, herd other animals, responds to questions and several commands, remains calm unless the need for guarding/protecting arises, and has befriended cats, donkeys, kids of all ages, and others smaller or bigger than himself. He truly is an AMAZING animal.
My dad raised sled dogs in Alaska pre WWII. Yes, good "mushers’ used a club (hard rubber was thought to be the best), but only if the dog attacked them.
I dont think that’s being cruel, especially as they werent pets, they were necessary working dogs. A matter of life or death.
