When is punishment (potentially) justified?

Lies - I, and many, understand that you come from a very specific background with dogs who are serving a very specific purpose.   Can you comment on what you would do with Coke, for instance, or even Kenya, if you were to teach the same behaviour? Would you also feel like you would use pressure (R-) to teach it, or would you teach it differently because of the fact that you're dealing with different dogs with different goals?

Sure.  Most of the stuff Nikon and Pan do, Coke and Kenya won't do.  They don't have the right nerve and enjoy it so what's the point?  It's not an issue of what training method is best but what I'm asking the dog to do and whether they actually enjoy it.  Coke has no drive (very little drive of *any* kind - prey, hunt, pack, fight, no social aggression....no GSD characteristics and I do not think there is GSD in his mix or if so no more than 1/4).  Kenya has the right pedigree for it but just came out wrong I guess.  She lacks confidence and nerve, doesn't have the power or the fight in her, and her lack of self-confidence severely inhibits the drive.  As far as SchH goes the only phase she enjoys is tracking, and she has a great pedigree for it (Lord Gleisdreieck being a great tracker and having produced one of the greatest trackers).  They just prove the point that I personally think it's ridiculous to say "you should only ever use these methods for ALL dogs".  All four of the dogs have had very different training because they are all very different temperaments based on very different genetics.  Nikon and Pan's training looks the most similar on the surface but it really isn't.  They DO the same things, but aren't trained the same way at all.  Nikon possesses more social aggression (I'm using this term the way working GSD people do, it is a *good* thing, a desirable thing) and he balances out more in defense than prey. Pan is almost all prey drive right now, very little defense, and far more social than Nikon, no social aggression (he is not naturally wary of strangers, not as aloof, and would not protect me).  Nikon is lower prey drive and has a higher threshold, meaning it takes more for something to trigger his drive (both defense and prey).  This makes him more of a "thinking" dog and more clear-headed.  As such he's a really fun dog to freeshape.  He enjoys the process of figuring stuff out without lures or -R.  His retrieve was 100% freeshaped with a marker and reward.  Pan is high prey drive and low threshold.  He tends to "load up" in drive and get over-excited just at the possibility of training.  He's almost the opposite of Nikon, I can't think of how I'd have two strong, stable male GSDs with temperaments any more different!  He is a nightmare to freeshape because the longer you wait for him to figure something out, even the smallest baby step, the more he loads up and the more nuts he gets.  It's like he's screaming (literally) "just SHOW me what you want so we can play and move on!"  I use more -R with him but given his age it's still in small increments for short periods (like no more than half a dozen reps during a single short training session) and *always* paired with +R.

I am not one of those people that says a dog is not reliable without corrections or escape training.  I don't believe that's true at all because Nikon has proven that wrong in several instances and he was High Obedience and High SchH1 at his most recent trial so I know the freeshaped behaviors (just as his retrieves, his send out, his blind search, and his article indications) hold up just as well as those that were built using -R (heeling, down out of motion, long down).  But for the type of work we do and why we do it, it is ridiculous and unrealistic to train completely devoid of any pressure or stress, both mental and physical since the nature of a lot of our work is to expose the dogs' strengths and weaknesses and ensure that a dog used for breeding or as an ambassador of the breed has the right power, confidence, and control to back up the pedigree.  Some will argue I can get away with it because I'm training possible breeding dogs and not "just pets".  I'm not sure whether I agree with that or not but I don't think that's the point.

I do believe that +P and -R don't work as effectively and often not at all when not paired with +R.  I never use pure aversives and expect the result to hold up; there is always a release/reward to encourage the appropriate behavior and cement that into the dog's head. For example, I've used a bark collar on Nikon a few times and it works insofar as he is wearing the collar, but I can't say it "trained" him anything because he's collar smart.  He was never rewarded for shutting up or keeping quiet because of proximity (which is why he ended up having to wear the collar in the first place).  In this case it doesn't really matter, because in the situation where we have the problem with the barking he can just wear the collar, I don't need him to generalize because the barking is either perfectly acceptable anywhere else or he doesn't bark in other contexts, but I won't say that the use of +P "trained" him not to bark in that specific context where the collar was used.  He doesn't bark because he knows the bark collar is on!  If I really wanted to proof this so he wouldn't bark in other contexts, just the collar (+P) would not be enough.

 Given that punishment doesn't teach behaviors, only suppresses them, it definitely isn't effective if you want to teach the dog what to do, at least purely speaking. You've got to have the reinforcement in order to get new behaviors, or to increase another behavior in place of the problem. However, it can be useful in order to stop the problem behavior, and allow you to jump in and show the dog what to do. This is fairly commonplace in teaching puppies not to bite people. We jump in when they do it with a verbal punisher. It  startles them a bit, and they briefly stop, allowing you to give them something appropriate, and reinforce that behavior.

Luke can very much be in the over-excited mode you describe in Pan. He goes from the Quaalude dog, as a friend jokingly called him, to barking nut-job in a few seconds. He gets hyped up just hearing another dog work if he's in his crate. 

I was thinking a bit, and I need to go to bed and not really get far into this thought. I know resistance to extinction has to do with learning history in large part, but I wonder how, all things held constant, what would be the effect of genetic factors relating to stamina? Of course, you can't  hold all the factors constant, but if you could, that would be interesting to see. Of course, as punishment goes, certainly genetic factors relating to how stimuli are perceived are going to effect this, like the simple factor that my 14lb dog is probably going to feel the same stimulus as much more painful  than a German Shepherd,  but even two dogs of the same breed are going to feel the same thing differently. Sorry,  need to go to sleep, the rest of what this is supposed to be has been lost somewhere in my head. 

 If -R falls under the umbrella of "punishment", I respectfully disagree.  I've seen a great number of dogs trained a great number of behaviors using -R.  As I said, I always pair with +R, but a lot of people don't, and while I might not use their methods, I can't say for certain that what they are doing is *not* working because I can clearly see that it is, I just choose not to do it that way.

 No, R- is not under the umbrella of punishment by definition (because reinforcement increases and punishment decreases).  Remember, we're talking pure textbook definitions here. By the definition, punishment only decreases behavior, thus you technically cannot use it to teach. The same stimulus, however, could be applied as both reinforcement and punishment. You could use the same e-collar to apply as R- to teach sit, stimulation turns off contingent upon  barking, and to reduce barking, stimulation turns on contingent upon barking. Some might argue the barking example is also negative reinforcement, since the absence of barking turns off stimulation. If you do argue this, I might argue back that you are just trying to make yourself feel better about the whole thing, since not barking is not a behavior (I suppose we have to invoke a new  version of the dead man test, the  dead man test being, "if a dead man can do it, then it ain't  behavior," which I will call the dead dog test). A dead dog can "not bark," therefore, "not barking" is not a behavior, you are decreasing barking, and applied in this way, it is positive punishment.

 Gotcha.  In that case I agree.  The bark collar stops Nikon from barking when he's wearing the collar but doesn't train him not to bark.

Liesje

 Gotcha.  In that case I agree.  The bark collar stops Nikon from barking when he's wearing the collar but doesn't train him not to bark.

I'm hoping you don't mean it exactly as it says. I hope it reduces the frequency of his barking. It doesn't teach him what he should do other than barking. It's possible, and one of the reasons it would be advised to also reinforce another behavior, that the animal can replace the behavior with something else to serve that function on his own. If you've got a dog who displays escape motivated growling and teeth baring, and you punish that without giving him another way to get out of situations, he might choose another one himself, and it is highly likely you (and I mean the generic you, not you specifically Lies) will like his new behavior even less. This is why, typically, if I have a kid who gets aggressive in order to get out of work, in addition to putting the aggression on extinction, or maybe punishing it, I'd be ill advised not to also teach him how to ask for a break.

I mean what I said, I don't follow?... Yes it stops him from barking (or "reduces the frequency" if you will, but honestly it stops him from barking at all, anything above a whine triggers the collar and he knows it).  He can choose to do any behavior he wants other than barking, when the collar is on.  I don't consider it training, or "learning" because there's no feedback from me, nothing is being reinforced.  I'm not trying to train him to do an incompatible behavior, I just can't have him barking while the collar is on his neck.  He can whine, sleep, chew a bone, or do cartwheels as long as he's not barking.

So, those who insist that punishment doesn't work must be wrong...

I don't get all hung up in the terminology.  Either something works quickly or it doesn't.  Using force to introduce the hold, using a bark collar....things things worked and they worked fast with no negative effects (Nikon's barking when it is called for is absolutely not the slightest bit inhibited by the fact that he occasionally wears the bark collar).

Liesje

So, those who insist that punishment doesn't work must be wrong...

I don't get all hung up in the terminology.  Either something works quickly or it doesn't.  Using force to introduce the hold, using a bark collar....things things worked and they worked fast with no negative effects (Nikon's barking when it is called for is absolutely not the slightest bit inhibited by the fact that he occasionally wears the bark collar).

You got it... if you think it doesn't work, you are wrong. The definition  says it does. If it doesn't work, then technically you didn't do it. 

Nikon clearly can discriminate collar on from collar off,which is good. I'm trying to teach Luke to discriminate between his Rally collar and his regular collar (which would help if I remembered to always put him in his rally collar when we are practicing at home). When he's wearing the Rally Collar, and I will probably put the same collar on him when we start doing therapy, that means he needs to walk in a heel, when we stop, you sit, this sit is straight and at my left side, etc. If he's wearing something else, he can walk out in front, sit where he can get direct eye contact, etc.

 I don't think anyone here ever suggested that punishment doesn't work.  It does, if it's at a sufficiently high level to reduce the intended behavior.  Effectiveness, however, can equally be had from +R applied correctly.  So, in my mind, it comes down to philosophy, not whether or not the process is effective - because either is.  I've simply decided, however, that the relationship I want with my dog, is not one of "do it or I will hurt you."  It's "do as I ask, and I'll make it worth your while."  I guess I've evolved, and maybe if you all read this, you will think about doing it, too.  From Suzanne Clothier: http://www.suzanneclothier.com/blog/i-had

spiritdogs

 I don't think anyone here ever suggested that punishment doesn't work.  It does, if it's at a sufficiently high level to reduce the intended behavior.  Effectiveness, however, can equally be had from +R applied correctly.  So, in my mind, it comes down to philosophy, not whether or not the process is effective - because either is.  I've simply decided, however, that the relationship I want with my dog, is not one of "do it or I will hurt you."  It's "do as I ask, and I'll make it worth your while."  I guess I've evolved, and maybe if you all read this, you will think about doing it, too.  From Suzanne Clothier: http://www.suzanneclothier.com/blog/i-had

Anne, the, "do this or I will hurt you," use of punishment is not one that I think most of us would ever say warrants any justification. I liken the all positive all the time among dog trainers to the Positive Behavior Support, of Behavior Analysis and treatment of humans. It just doesn't work all the time. I link to the Judge Rotenberg Center, which I understand is highly controversial. Yes, they do have a very vested interest in the justification of physical punishment, but they also, because of that interest, have the best compilation of sources of information regarding the topic.

http://www.judgerc.org/canthesameresults.html

griffinej5

I've simply decided, however, that the relationship I want with my dog, is not one of "do it or I will hurt you." 

I would argue that the majority of pet owners do not want that type of relationship either. Most times, they simply don't know a better way, or are so frustrated with their dog that they respond in anger instead of with clear-headed thought.

Keep in mind, though, too, that the term punishment is not synoymous with pain!  If I turn my back on a jumping dog, if I shut the door when a dog gets up from the sit to bolt out the door, if I back up when the dog pulls me towards something (and only approach if the leash is slack), if I simply withhold a treat because I didn't get the response that I want - they also ALL serve as punishment if they reduce an unwanted behaviour.  There are lots of things our dogs experience as punishers - in the scientific definition of punishment - in relationships with us, that I will argue don't lessen our relationship with them in any way. An NRM serves as a punisher, even if only a temporary one. A lost-opportunity marker is a conditioned negative punisher. Your tone of voice can be a punisher, if it lessens a behaviour in that moment.

Punishment doesn't have to include pain, nor should it!  The fact is, we can't totally avoid punishment in our daily lives with dogs. Nor do I think we *should* (I can hear some saying...."WHAT? Did Kim really say that!?!?";).  I do think that dogs benefit from understanding that their behaviour matters - to bring good things about, but also to learn that some behaviours cause good things to go away (it doesn't mean they learn that bad things happen TO them). Both are important, I think, to creating a behaviourally and emotionally healthy animal. And I don't think you can do exclusively one or the other - both are necessary to some degree.

I think we do need to be thoughtful about our interactions, though, to do our best to minimize the use of punishments, and be honest about the kinds of punishments that we do use, and how they work in our relationships with our dogs. Also, to be thoughtful about how we "might" be able to replace a certain punishment with a reinforcement-based teaching method.

Kim, that is a great post.  We both know I'm willing to go quite a bit farther in one direction when it comes to training but I completely agree with your main points.

If anyone feels the need to question my relationship with my dogs they are fully welcome to observe any of our training sessions.

Personally I don't feel the need to constantly re-hash the same disagreements about what punishment is with those people who admittedly have no experience with it.