Eight Rules for Punishment (and why we shouldn't use them)

corvus

 Here's a video of early "leave it" training with Kivi to compare with espencer's video of Chuck. If anyone wants to.

http://www.youtube.com/profile?user=melstarling#p/u/0/vOaomQvHilg

I haven't had a chance to check it yet. Hopefully it works. 

I think I would have trained it similarly, just issue the treats from the clicker hand, and make them better than what you had in the bait hand. That way, the dog is reinforced not only by getting a treat (which he could learn to anticipate from the bait hand), but he gets an even better one than the one he left.  It is in this way that you can continue to up the ante, so to speak.  Feeding from the bait hand or allowing the dog to take the same treat from the floor is setting them up to fail when the "leave it" item is very cool and something they want very much.  Learning to train so that the dog always expects *your* treat to be better than what he left is the reason that good trainers have dogs that will leave the squirrel alone when they hear leave it.  NOT because the squirrel isn't more enticing than anything else, but because they have been so conditioned, over time, to think that even if it is a squirrel, chances are that you will have something better...  With so many conditioned responses, they also tend to forgive you when you produce roast beef instead of squirrel a la mode;-))

spiritdogs

Feeding from the bait hand or allowing the dog to take the same treat from the floor is setting them up to fail when the "leave it" item is very cool and something they want very much.  Learning to train so that the dog always expects *your* treat to be better than what he left is the reason that good trainers have dogs that will leave the squirrel alone when they hear leave it.  NOT because the squirrel isn't more enticing than anything else, but because they have been so conditioned, over time, to think that even if it is a squirrel, chances are that you will have something better...  With so many conditioned responses, they also tend to forgive you when you produce roast beef instead of squirrel a la mode;-))

I teach pretty much the same, but being poodles they are infinitely more intelligent so i can teach them in five clicks flat NOT... 

Actually i get heartily sick of hearing the dog going to the squirel routine from other quarters as it just isn't true.It just creates demand for e collars.Dogs go for what is certain over what isn't. So do may other animals . It was reported in the mid 80s as an anomaly. If your reinforcement is dodgy or patchy you have no hope.

One stunt that i pull (and used to do with my food crazed Lab) is to set up for a traditional recall, and put high vaule treats on the ground between me and the dog. Without training Cadbury would come in to me like  a steam train. Luci is the same. You can then premack it if you like.

One thing with poodles is getting that instant decision or reflexive behaviours on recalls. Their default behaviour is to follow ground scents and food is often not number one. You have to make yourself interesting and rewarding and play may be the way...

poodleOwned

spiritdogs

Feeding from the bait hand or allowing the dog to take the same treat from the floor is setting them up to fail when the "leave it" item is very cool and something they want very much.  Learning to train so that the dog always expects *your* treat to be better than what he left is the reason that good trainers have dogs that will leave the squirrel alone when they hear leave it.  NOT because the squirrel isn't more enticing than anything else, but because they have been so conditioned, over time, to think that even if it is a squirrel, chances are that you will have something better...  With so many conditioned responses, they also tend to forgive you when you produce roast beef instead of squirrel a la mode;-))

 

 

I teach pretty much the same, but being poodles they are infinitely more intelligent so i can teach them in five clicks flat NOT... 

Actually i get heartily sick of hearing the dog going to the squirel routine from other quarters as it just isn't true.It just creates demand for e collars.Dogs go for what is certain over what isn't. So do may other animals . It was reported in the mid 80s as an anomaly. If your reinforcement is dodgy or patchy you have no hope.

One stunt that i pull (and used to do with my food crazed Lab) is to set up for a traditional recall, and put high vaule treats on the ground between me and the dog. Without training Cadbury would come in to me like  a steam train. Luci is the same. You can then premack it if you like.

One thing with poodles is getting that instant decision or reflexive behaviours on recalls. Their default behaviour is to follow ground scents and food is often not number one. You have to make yourself interesting and rewarding and play may be the way...

 

I know that Poodles are smart, but there's nothing smarter than a stock dog;-)    How many clicks???  (Just kiddin' - couldn't resist lol)

poodleOwned

It is normal for humans to seek the R+ that they get from an Aversive again and again, but most aversives habituate with too much use. Aagin this is well documented

 

the person applying the aversive gets R-. The subject stops doing something you didn't want him to do, and it increases your likelihood of applying the aversive in the future. Therefore, it's R-. 

Liesje

Corinthian

Liesje

I have one dog trained positive only, and another trained positive probably 90%.  The positive only dog is shy, nervy, somewhat neurotic

 

Are you suggesting causality? I hope not.

The nervous dog would not have benefited from punishment, that would only reinforce that the fear because he would have confirmation that the world is a dangerous place. The second dog would also be the same dog, you see if also trained without physical punishment. I also disagree, for a punishment to work, the dog must experience some level of avoidance or else it would not affect behavior.

Let's face it, some dogs can take a beating and walk away psychologically unscathed, others are not so lucky. Unfortunately, you don't always know which is which until it is too late.

I am also never convinced about dogs that have no food drive.  The only dogs with no food drive are dead dogs.

 

Casually?  Lord know!  I do LIVE with this dog.  

Just responding to Tenna's claim.  I know full well what *any* form of punishment would do to this dog, that's not why I posted.  The dog already shows stress and avoidance if you accidentally bump her, or if DH is screaming at a sports team on TV.

The statements I bolded are exactly my point.  You wouldn't believe how many people ask me if the nervy dog was abused.  Ha!  She's could not have grown up in a better environment with a better owner.  But some things you cannot control with training or behavior mod regardless of the method used.  This is just the dog's temperament.  In contrast, my mixed breed who has been passed around several homes and was neglected before we adopted him is a very sound, happy, confident dog. 

NO. CAUSALITY:

Causality "is the relationship between an event (the cause) and a second event, where the second event is a consequence of the first"1

I'm sure you know what it means and just misread it.

Yes, I did misread it.  But the answer is the same....my point was that it does NOT prove causality, and that sometimes (more often than not, I'd wager) a dog's temperament and mental soundness are more important than which specific method you choose.  There is only so much socialization and training you can do before you hit the genetic wall.  So, I do not believe that a dog trained *all* positive will always look and act and behave happier and more freely than any other dog trained with any other method.  To me that goes against everything I've learned about training dogs and my breed of choice in particular.

Liesje

Yes, I did misread it.  But the answer is the same....my point was that it does NOT prove causality, and that sometimes (more often than not, I'd wager) a dog's temperament and mental soundness are more important than which specific method you choose.  There is only so much socialization and training you can do before you hit the genetic wall.  So, I do not believe that a dog trained *all* positive will always look and act and behave happier and more freely than any other dog trained with any other method.  To me that goes against everything I've learned about training dogs and my breed of choice in particular.

Earlier today, I called Luke over in a, "you're in big trouble," sort of tone. I called him Lucas and yelled for him. He raced over to me, butt wiggling and eager. I was thinking, most dogs would run the other way. I never had to get him over any fear of being yelled at in this tone. This is generally the type of tone people use, and if the dog isn't terrified from the get go, he will learn to run away terrified when he hears it. Now, I will continue to reinforce response to this tone so I don't get the terror, but in many breeds, I'd be really fighting to get the dog to come towards me rather than away if I called him like this.

 And the point remains - ignoring all ethical considerations - that using force and aversives is a gamble because you never know if it will break the dog until he lies shattered in pieces or worse if you are a service dog when you demand something from him that because of his training he will just shut down.

No one is an all positive trainer. That doesn't exist. The argument is that punishment does not have to painful or induce fear for it to be effective.

Corinthian

that using force and aversives is a gamble because you never know if it will break the dog until he lies shattered in pieces

Corrected my dog many times, the only time he had the tail between the legs was when he put his head inside a cardbox himself. I was not able to correct the same way my miniature schnauzer so the level of intensity on the corrections were significately less. I've corrected many dogs, abosolutely none of them showed a behavior like the one you described. Have you ever used physical corrections before?

It's all about rule number 5 and 3

Corinthian

that using force and aversives is a gamble because you never know if it will break the dog until he lies shattered in pieces or worse

Meanwhile, I've seen seemingly all-positive dogs lunch people like it was no problem in their mind.  So dogs can correct people with pain/discomfort, but people should not correct dogs with pain/discomfort "because their timing is not as precise as a dog's"?  Unbalanced, imo.

Meanwhile, I've seen seemingly all-positive dogs lunch people like it was no problem in their mind.  So dogs can correct people with pain/discomfort, but people should not correct dogs with pain/discomfort "because their timing is not as precise as a dog's"?  Unbalanced, imo.

We are not talking about what you do if a dog is attacking you, and the fact that a dog was trained positively does not exempt the dog from genetic temperament problems.  However, I think that the dogs that are trained correctly using positive methods are less likely to "lunch" a benign human than dogs treated badly (unnecessary punishment falls into the "badly" category IMO)  We are talking about how you train a dog under normal circumstances.  I think Corinthian's statement about there being no 100% positive trainers is valid, it's just that when we punish (scientific definition) we do so without causing pain or fear whenever we can, and we don't try to find any excuse we can to continue to punish dogs when it is not necessary. 

spiritdogs

it's just that when we punish (scientific definition) we do so without causing pain or fear whenever we can, and we don't try to find any excuse we can to continue to punish dogs when it is not necessary. 

Oh then that means i was always "positive"

spiritdogs

However, I think that the dogs that are trained correctly using positive methods are less likely to "lunch" a benign human than dogs treated badly (unnecessary punishment falls into the "badly" category IMO)  We are talking about how you train a dog under normal circumstances.  I

espencer

spiritdogs

it's just that when we punish (scientific definition) we do so without causing pain or fear whenever we can, and we don't try to find any excuse we can to continue to punish dogs when it is not necessary. 

Oh then that means i was always "positive"

I think from Anne's definition just about everyone would be a positive trainer here.  

miranadobe

So dogs can correct people with pain/discomfort, but people should not correct dogs with pain/discomfort "because their timing is not as precise as a dog's"?