the role of punishment

I'm in agreement that punishment isn't a needed part of a training protocol.  I don't know where the other thread was, but I must have missed it.  Or not wanted to dive in to yet another argument!

I personally prefer to set a pup up to suceed and manage the environment.  I do use verbal clues such as eh eh, but that's about it.  Praise for doing something RIGHT seems far more effective to me than scolding for doing something WRONG.

Are we talking positive punishment here?

I've been asked on a couple occasions "how I punish my dogs" and I never really know quite how to answer because if we're talking positive punishments, I don't on purpose. Sometimes it happens and one of my dogs is really soft so he can be punished even when the punisher didn't mean it, but it's never my explicit intent.
 

yeah, positive punishment only.

mudpuppy

In another thread I suggested it was possible to raise and train a dog without ever using punishment and was astounded to hear that no one agreed with me [...] I believe it is not only possible to never use owner-delivered punishment but desirable and more effective in all situations to not use owner-delivered punishment.

I agree with you that it's possible to raise and train a dog without ever using owner-delivered punishment. I don't believe, however, that it's more desirable, more effective, with all dogs and all owners and in all situations.

I appreciate your "probable results", but probabilities are just that. It doesn't always happen that way. These are not computers we're dealing with.

I believe punishment (which should never be issued with frustration or anger) is as informative to the dog as reward. It's the other side of the coin. And I believe each owner should make the decision to use punishment or not, as they see fit, just as they decide whether or not to use other training techniques and tools.

I would assume you can raise and train a dog without using reward, either, but I wouldn't recommend it.  

Punishment is informative, but it doesn't always inform the dog of what you think you're informing. It's a risk you take with reinforcing as well (lots of owners reinforce things accidentally, or reinforce the wrong thing) but because punishment relies on things the dog, by definition, does not like, you can accidentally make a connection for the dog that makes the dog feel dislike towards something or someone totally innocuous. When you mess up reinforcing, you make a connection for the dog to like something that maybe you'd rather they not like. But given a choice, I'd rather have my dog liking too many things than disliking too many things.

Jumping is up something that a lot of owners mistakenly reinforce. What you wind up with is an enthusiastically jumping dog, which to be sure can be dangerous, especially with kids or the elderly. But even if you are spot on with your timing of a punishment with jumping, the dog may wind up connecting anyone coming to the door with getting a punishment. This is not a dog I want coming to meet me at the door any more than I want one who has been reinforced to enthusiastically jump up.

Kobi never was much of a jumper. His greeting was to nip at ankles, run circles around the feet and paw the leg. Herding behaviour??? Anyways, we tried for months to ignore the behaviour with little success. I eventually, out of desperation began putting a toy in his mouth and putting him between my legs when I came in the door. The toy saved me from those sharp little puppy teeth and having him between my legs stopped the pawing. By about a year Kobi was quite calm at the door. I also think as he matured he did just out grow some of that puppy excitement and the excitement from the visitors also deceased as he grew out of that  "ohhhhh, cute, cute puppy!!" stage. But the alternative/ default behaviours remain. Most times now, Kobi will check out whose at the door, wander over with his head down, butt wiggling, pick something up in his mouth and if given the opportunity will push his head between their legs for a good butt rub but is just as happy with a scratch on the head. There are still a few people that really get Kobi excited at the door. I know it is because of the way they react to him, but they have been much harder to train than the dog. LOL

I do think in some cases, some kind of punishment is necessary.  Sometimes, the extrinsic reward a dog receives for misbehavior, is more desirable than the reward he receives from you for good behavior.  For example, my dog is extremely food motivated.  The reward she would get by successfully counter-surfacing beats anything I can reward her with for staying off the counter.  She would get food either way, and she'd get MORE food from the counter than the itty bitty treat I would give her (I need to watch her weight!) The booby trap you mentioned, mudpuppy, is technically a punishment, right?  And it's a punishment that works because the dog does not learn to associate the punishment with you, and will stay off the counter whether you are or are not around.

This is hillarious. 

I'm sorry, but by removing how a dog learns from their environment, you have just removed the foundation of Skinner's theory. By removing a "no reward marker", you have also removed social information to help a dog succeed, which doesn't belong inside the Skinner box in the first place because social learning was not part of Skinner's experiments. You are using a scientific term "positive punishment", and discounting the science by telling us we cannot use it in a way proven by science, as one of the quadrants of operant conditioning.

Since you repeatedly point out "everything is operant conditioning whether we want to admit it or not" in post after post in this forum, I suggest you not attempt to rewrite the science of operant conditioning. And please, give us a clear and agreed upon definition (by a group of scientists) of which quadrant a "no reward marker" fits into.

You might want to step outside the box and look up some papers on social learning, since that's what we are using in harmony with operant and classical conditioning when we interact with an animal which does not live alone inside of a Skinner box.

mudpuppy

Definition: for the purposes of this discussion, punishment is something the owner directly applies to the dog in an attempt to get the dog to stop doing THAT forever. Not environmentally-applied punishment, and not no-reward markers, which are intended to suggest the dog try a different behavior but feel free to continue to perform that behavior again in future (as, for example, you're working on sit and the dog offers a down. You say uh-oh to indicate that wasn't what you wanted; but certainly at times in the future you will want the dog to offer a down).

It is handy for you to skew the definition to suit your agenda, but that is not the classic definition of punishment in psychology.  

Punishment, positive or negative,  reduces the likelihood of that behavior occurring in the future. Reduces. Not eliminates.

When my puppy is jumping at his crate, excited to get out, and I turn my back to him, that is punishment. Not R-, though some people consider it to be. I am applying social isolation, and in group animals isolation is scary and bad. A more solitary animal might find being ignored to be R-, but not a puppy.

Furthermore, studies have found that puppies who never receive P+ before a certain age (normally applied by the mama) grow up to be untrainable. I don't have the reference handy, but can sort through my dog behavior books if anyone needs it.

Mudpuppy, I am sorry you once over used physical corrections. I'm glad you do things differently now. But for the love of doG, not everyone shares your past or your experience with what mild, appropriate, and effective punishments are.

In terms of behavioral psychology, no I do not believe you can train a dog without punishment. I believe you can train a dog without physical punishment, without yelling, without leash pops and alpha rolls. But without punishment? No. 

I'm not looking for a theoretical argument; I want some solid, specific examples of where and when people think OWNER-inflicted punishment is THE MOST EFFECTIVE method of dealing with a particular behavior. I'm quite fond of judicious use of environmentally-inflicted punishment, of -P, and no-reward-markers can be handy at times. I want someone to come up with actual examples of effective use of owner-inflicted punishment so we can discuss them.  Maybe we'll all learn something.

Like corgipower's example, which has been the only one offered so far: it's interesting and informative. The dog was punished for trying quite vigorously to obtain the reward he felt he had earned- I can't argue with that, dog is basically attacking you, seems your best response is to stop the behavior now somehow. I find it interesting that the owner used a toy to teach the stop-jumping "sit"; most trainers recommend using "getting attention from the person being greeted" as the reward, since that is, presumably, what the dog wants from jumping up. The dog gets the same reward, he just has to perform a different behavior than the one he came up with on his own. So one might conclude that the reward one offers should be chosen with care to avoid unpleasant incidents of this sort? 

 If you asked me to tell you that owner inflicted punishment works, I can show you lots of examples. If you want an example of when owner-inflicted punishment is THE MOST EFFECTIVE method, I cannot, because IMO it is simply wrong to believe that there will be no negative consequence to the dog from that.  Sure, it will work in a lot of cases, but almost invariably, so will the other scenarios that you mention, probably with less damage to the dog. 

Also, I do think that many average pet owners have a bias toward punishment.  Why?  Because generally speaking, they feel better and are better at punishing than they are at the other techniques, which require a bit more finesse and understanding to be implemented.  And, they don't know what to do if a positive technique isn't working as well as they think it should.  They almost always want fast results - not everyone spends as much time or energy on their dogs as we do.  So many times, pet owners muddle along without EVER going to training class, and if they do, they go to beginner class only and then never again.  How can they expect to be effective at this if they don't learn to apply it properly?  Somehow, they seem to think that punishment is an ok form of trial and error, but positive reinforcement is not. Yet, more damage is done to dogs by punishment, more dogs become aggressive because of punishment, and more behavior is unchanged after repeated punishment than I can tell you.  I see it all the time. 

BTW, it is not the fault of trying to use positive methods that failed in the example about the jumping dog.  The dog decides what the valuable resource is - it can be your face (to greet), or it can be a toy {go grab it).  The idea is that the toy should not have been there.  In that case, I would have told the person to put a door between themselves and the dog without acknowledging the dog AT ALL.  No eye contact, no voice, no nothing.  Dogs do get more boisterous (after all, their behavior always worked before, so they ramp it up, assuming that we just "didn't notice them" yet).  But, if the human, instead of giving up, rides that out (in body armor perhaps LOL), the extinction burst will come, and the behavior will extinguish.  Failing to do that is a human problem, not a dog problem - they just do what works, and the dog grabbing the toy just exhibited a giant "gotcha" because he got something for being obnoxious.  Dogs repeat what is reinforced.  What you allow, you teach. 

mudpuppy

I find it interesting that the owner used a toy to teach the stop-jumping "sit"; most trainers recommend using "getting attention from the person being greeted" as the reward, since that is, presumably, what the dog wants from jumping up.

I started out the example in general terms ~ I have seen similar responses in other dogs trained not to jump usig PR methods. Somewhere in the middle of the post I transitioned to Nyx specifically, who I didn't use attention as a reward with because attention from me ~ at that stage in her life ~ was the least interesting thing for her. I used a toy and I used treats, with the same overzealous results. Having given her about a half dozen firm corrections for various things, I am not only seeing a huge improvement in her behavior, I am also seeing a huge increase in her desire to have attention from me.