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

Liesje

Makes perfect sense to me.  I was just trying to keep it simple, since it seems the majority here seem to lump ALL aversive actions into one.

And, depending on how one trains, there are times where it really is a correction (leash pop the dog that is getting dirty, etc).

Interesting that it is the same as a lure.  I agree though, because a lure has to be faded as well.  So someone like me that has a more "real" dog who doesn't get really into drive until he's worked up a bit in defense, well when he gets to trial this doesn't happen.  First thing out, he runs to the hot blind and the helper is standing there still.  There's no pumping up the dogs or anything like that allowed.  So the "lure" of the threat has to be faded, just like if I use a treat in my hand to lure a left finish.  I don't like to get too dependent on lures, but in many cases they are necessary and don't do any harm as long as you can quickly teach the dog and fade the lure.  But that's a whole different discussion I suppose...

I'd say it's the same as a lure, and I am sure I won't get much agreement on this, because it has the same effect on behavior. I don't think either that I will change everyone from lumping things together that shouldn't be. I'd suggest that in order to continue in these discussions, everyone hop on over to amazon.com, and get themselves a copy of Behaviorspeak. Then, when you want to use a term, you look in there and make sure you are using it correctly. It's a pretty hilarious book, with definitions such as, "Sesame Street/The Muppets: the apex of contributions to civilization." I think using terms as they are meant might be the apex of contributions to these discussions.

This thread started as a discussion on punishment, which, in the OC sense, means something that is added to suppress behavior.  So, I think a copy of Behaviorspeak might be just the thing, too.  When using the term "punishment" we mean something that suppresses behavior.   If your actions elicit a particular behavior, they are an antecedent (cue). 

The sequence:  Antecedent (cue) - Behavior (what the dog does) - Consequence (reward or punishment)   The more a dog is rewarded, the more he repeats the behavior, the more he is punished, the less he exhibits the behavior.

Premack:  when a highly preferred activity is used effectively as a reinforcement for a less preferred activity  (Eat your vegetables and you can have dessert.) 

spiritdogs

This thread started as a discussion on punishment

I wish to just add a few notes from "researches" those nasty liberals on poor pay packets that want to take over the world  :)

There is a substantial difference in the action of Aversives. Physical or threat Aversives trigger the adrenalin system and in a dog that is poorly trianed and or poorly prepared as most are cause fight or flight behaviours. There is no room for debate on this it just is. You can train some dogs to have a different response such as Booth and Dildei expect which is to look at you. These dogs are highly confindent high drive dogs, and Dildei suggests that if you use more than one correction your training is at fault. This is a hugely different picture from what we see with CM. The underlying brain chemistry is still that of the Adrenalin system

Time outs are vastly different. Depending on the use and placement they just turn off the SEEKING circuits which reduces pleasure. They can trigger PANIC circuits in dogs that have seperation anxiety histories which is pretty awful.

I prefer to use very little punishment, and if my training has been poor and caused a habit that isn't helpful, i like to do it as early as possible and as short as possible using very short time outs. My dogs have not experienced Physical or threat Aversives because they have been unneccessary .

It is interesting to note that Sheila Booth the co authour of Training in Drive went to write "Purely Positive", a term that many of the hardline school seem to hate with a passion.

spiritdogs

This thread started as a discussion on punishment, which, in the OC sense, means something that is added to suppress behavior.  So, I think a copy of Behaviorspeak might be just the thing, too.  When using the term "punishment" we mean something that suppresses behavior.   If your actions elicit a particular behavior, they are an antecedent (cue). 

 

An antecdent is what immediately precedes the target behavior, because I think we should get picky about our definitions here. I didn't think some people would be thrilled with me putting it in the same category as a lure. 

So, are we only discussing P+, and not P-? Based on the article linked, it seems to be that we are mostly referring to P+, and that's what most people use, but just to clarify these things.

How bout this little compromise that no one will agree on. Use punishment all you want, but be willing to collect data. Be willing to collect data about whether your punishment is actually working, because we are often tempted to continue an using something we believe is a punisher because of the temporary effect it has in stopping behavior. If you're going to use it, because of the pretty big potential for harm (not to say you can't cause harm with reinforcement, but not as bad, and that's a separate discussion), you'd better be sure it works. In addition to this, you should look at the side effects. Look at escape/avoidance, aggression, and all those other things you're being warned about. If it is indeed effective, not because you think it is, but because the data says so, then go ahead with it.

For the record, ethical use of punishment requires far too much effort for me, so I just don't do it. 

griffinej5

An antecdent is what immediately precedes the target behavior, because I think we should get picky about our definitions here. I didn't think some people would be thrilled with me putting it in the same category as a lure. 

So, are we only discussing P+, and not P-? Based on the article linked, it seems to be that we are mostly referring to P+, and that's what most people use, but just to clarify these things.

How bout this little compromise that no one will agree on. Use punishment all you want, but be willing to collect data. Be willing to collect data about whether your punishment is actually working, because we are often tempted to continue an using something we believe is a punisher because of the temporary effect it has in stopping behavior. If you're going to use it, because of the pretty big potential for harm (not to say you can't cause harm with reinforcement, but not as bad, and that's a separate discussion), you'd better be sure it works. In addition to this, you should look at the side effects. Look at escape/avoidance, aggression, and all those other things you're being warned about. If it is indeed effective, not because you think it is, but because the data says so, then go ahead with it.

I totally agree with your comment. I really also want to see evidential data. Show me a dog trained with a lot or a majority of correction  and it is pretty evident.

I want to see videos.Show me a video of the dog learning something new.

It takes ages to undo the damage even with a sound drivey dog .

It shows up in lots of activities such as tracking.

But i am a pretty low IQ  trainer, wouldn't know a whole lot of what i am talking about :)))

I

tenna

And I think the argument that owners don't recognize fear/distrust/unhappiness in their dog trained with punishmet is definitely valid.

I don't think the method of training one uses would diminish or increase your ability to read (fears/distrust/happiness) your dog. Some people are intune with their dogs and others not so much. I'm a positive trainer 99% of the time and I have some difficulties reading my dog. The people who post on this board and incorporate corrections/punishments into their training, IMO are excellent trainers and seem quite intune with their dogs, so when they tell me their dogs do not fear or distrust them, I have no reason to doubt them.

denise m

I don't think the method of training one uses would diminish or increase your ability to read (fears/distrust/happiness) your dog. Some people are intune with their dogs and others not so much. I'm a positive trainer 99% of the time and I have some difficulties reading my dog. The people who post on this board and incorporate corrections/punishments into their training, IMO are excellent trainers and seem quite intune with their dogs, so when they tell me their dogs do not fear or distrust them, I have no reason to doubt them.

I am not saying that positive only trainers are great at reading dogs, I am saying that people don't notice how big of an effect punishment can have on their dog, until they do away with it altogether and/or train a different dog without it. Not that they will necessarily be able to point out WHY it is different, but it's definitely there and it's usually good.

poodleOwned

I totally agree with your comment. I really also want to see evidential data. Show me a dog trained with a lot or a majority of correction  and it is pretty evident.

I want to see videos.Show me a video of the dog learning something new.

It takes ages to undo the damage even with a sound drivey dog .

I use positive reinforcement to make my dog do things that i want him to do (but i am not into teaching him tricks) and i re direct with the leash, my body or my voice when i think he is doing stuff that he should not be doing.

I have a video, but then you will doubt the amount of time that my dog took to learn it. I collected my data first, in 5 minutes i knew that corrections this time were not going to work and i did something different

But then you will also say that is only one dog

PS I think rule number 5 has a lot to do with rule number 3

tenna

I am not saying that positive only trainers are great at reading dogs, I am saying that people don't notice how big of an effect punishment can have on their dog, until they do away with it altogether and/or train a different dog without it.

Sorry, you are still saying the same thing. 'People can only read their dogs accurately when they become positive only trainers'  

denise m

I don't think the method of training one uses would diminish or increase your ability to read (fears/distrust/happiness) your dog. Some people are intune with their dogs and others not so much. I'm a positive trainer 99% of the time and I have some difficulties reading my dog. The people who post on this board and incorporate corrections/punishments into their training, IMO are excellent trainers and seem quite intune with their dogs, so when they tell me their dogs do not fear or distrust them, I have no reason to doubt them.

I do think the method of training would effect your ability to read your dog. It is known that people tend to continue using an ineffective punisher because of its temporary effect. Punishment for one tends to be negatively reinforcing for the person delivering it. If someone tells me they are seeing absolutely zero side effects of punishment, maybe that's true, but it's darn hard to believe. You may think you aren't seeing any side effects, but if you don't have the data, you don't know.

denise m

Sorry, you are still saying the same thing. 'People can only read their dogs accurately when they become positive only trainers'  

No. I am saying there is a difference between a dog's behavior when he is trained purely positively vs. trained with corrections. And that a person MAY NOT NOTICE that he was giving signs of being unhappy until they see a different model (the positive-only trained dog). It could go both ways - someone may not notice how happy their dog is to work until they start using corrections. I am saying that a lot of owners may not notice the fine signs a dog is giving that he is not 100% enjoying working/training until they see a dog that has been trained positively. I am saying that sometimes you cannot see fine body language in your dog until it is no longer there. Please, try again.

espencer

I have a video, but then you will doubt the amount of time that my dog took to learn it. I collected my data first, in 5 minutes i knew that corrections this time were not going to work and i did something different

But then you will also say that is only one dog

Actually the arguement is a bit more subtle. You made a bold claim that you know how to use punishment and that us low IQ trainers don't. If you can put up an example then that is great, we need to  look at redefining how we train and look for more examples.

If you can't well that is usually how it goes and i won't be at all suprised. I believe that you have to put up or shut up. That to a large extent is why i trail in obedience.Oh and the huge fun i have with my dogs because the p word is mostly out of the equation.

poodleOwned

I wish to just add a few notes from "researches" those nasty liberals on poor pay packets that want to take over the world  :)

Corinthian

The only reason they proved that punishment is not that effective and carries unwanted side effects is because they want to take away our assault rifles.  It starts with positive training and before you know it we are living in a gulag.

What is the point of this statement?  Is it designed to encourage debate and to enlighten people?   Fail.

JackieG

What is the point of this statement?