Exploring A Concept

[Rom]

Following on from some thoughts in raised in this thread, I'm thinking another concept through that I'd like to throw out there for thoughts. That is:

All learning/behavioural change begins with punishment, positive or negative.

A wise friend once said to me, and I think I agree, that "People don't change until the pain of not changing becomes greater than the pain of change". So, is this true for all living organisms including our dogs?

At this point in time, I think it is....well at least I haven't yet been able to come up with an example where I couldn't identify a punisher that prompted learning/behavioural change.....

Thoughts?

[Tonymc]

Rom, in my experience pain causing change per the example you gave of a Human does not apply to that level to all lving beings.

Usually Humans will make change when the pain is great enough in most cases.There are however those who have linked pleasure to pain and until the link is dissipated will hang onto pain.

Between Humans and Animals the big difference of course is the Human intellect.

I am not sure of the context of your post regards Dogs.Are you talking in all contexts or in association with a trainer or handler? Tony

[jbbb]

Animals do what makes survival easier.

If they find a way to do something that is simpler or easier, they will do it that way.

In regards to training, if they discover that doing something a particular way means they get food or attention (both necessary for dogs in packs) then they will continue to do it that way. They may try to do it another way to see if they get a better result, but if they dont get a better result they will do it the way they know.

But animals learn in the wild without punishment from humans....

IMO punishment is a human concept - animals think in consequences.

Edited February 24, 2008 by jbbb

[Cosmolo]

What do you consider to be a punisher? Not having food and making the dog hungry is not a punisher, its a stimulus that will then cause the dog to do XYZ behaviour and get XYZ result- some kind of consequence that will then determine if that behaviour will be repeated next time the same stimulus is presented. I don't consider dogs to be in a constant state of 'punishment' but maybe my definition of punishment is different to yours.

[Rom]

Rom, in my experience pain causing change per the example you gave of a Human does not apply to that level to all lving beings.

Usually Humans will make change when the pain is great enough in most cases.There are however those who have linked pleasure to pain and until the link is dissipated will hang onto pain.

Between Humans and Animals the big difference of course is the Human intellect.

I am not sure of the context of your post regards Dogs.Are you talking in all contexts or in association with a trainer or handler? Tony

Hi Tony,

I'm talking in terms of operant conditioning and referring to both positive punishment and negative punishment....I probably confused things by including the quote from my friend as it refers to humans, but I believe the concept behind the quote is the same for all learning/behavioural change.

Animals do what makes survival easier.

If they find a way to do something that is simpler or easier, they will do it that way.

In regards to training, if they discover that doing something a particular way means they get food or attention (both necessary for dogs in packs) then they will continue to do it that way. They may try to do it another way to see if they get a better result, but if they dont get a better result they will do it the way they know.

But animals learn in the wild without punishment from humans....

IMO punishment is a human concept - animals think in consequences.

Think about punishment in the behavioural science sense though Jbbb, punishment is anything that weakens a behaviour. Yes animals do learn in the wild, but I don't believe that they do it without punishment. What motivation would an animal have to eat if he didn't feel hunger pains for example? Are hunger pains a punishment that motivate an animal to act to remove that discomfort? Is it the need to remove that discomfort quickly that helps an animal learn to become more efficient at hunting? Is it the effect of negative punishment that weakens all those behaviours that prevent a dog from hunting efficiently?

What do you consider to be a punisher? Not having food and making the dog hungry is not a punisher, its a stimulus that will then cause the dog to do XYZ behaviour and get XYZ result- some kind of consequence that will then determine if that behaviour will be repeated next time the same stimulus is presented. I don't consider dogs to be in a constant state of 'punishment' but maybe my definition of punishment is different to yours.

Its not what I consider a punisher that is important, its the effect that the stimulus has on behaviour that is.

I'm thinking in terms of the scientific definition of punishment......

I agree with you in that dogs are not in a constant state of punishment.....at the moment my dog is sleeping, she's getting exactly what she wants/needs at this time. Also, if a dog has just eaten his fill, he's less likely to want to hunt for food....so not hunting/eating is not a punisher at that time.

ETA: Or....am I way off the mark?

Edited February 24, 2008 by Rom

[Cosmolo]

Stimulus or motivation is different to punishment IMO

[Ah Ngau]

Hi all, newbie dog owner here but I am very interested in operant conditioning.

I think the statement "All learning/behavioral change begins with punishment, positive or negative" only presents half of the equation. Punishment, be it +ve or -ve, only decrease the probability of a behavior from occurring. We can extinguish an establish behavior, say counter surfing or pulling on leash, with punishment.

But then again, what if we want the dog to sit in front of the counter to wait patiently for food, or we want the dog to heel? Punishment can never give you those because it can't INCREASE the probability of a behavior.

"People don't change until the pain of not changing becomes greater than the pain of change" and "People do change if the gain of changing becomes greater than the gain of not changing"

ETA: My wife always says that men are like dogs. I think she is right, at least in operant conditioning.

Edited February 25, 2008 by Ah Ngau

[Diva]

Following on from some thoughts in raised in this thread, I'm thinking another concept through that I'd like to throw out there for thoughts. That is:

All learning/behavioural change begins with punishment, positive or negative.

Not sure why you would only focus on + and - punishment, and exclude + and - reinforcement? The way I think of it, punishment, both + and -, works to decrease the likelihood of a behaviour occuring; reinforcement, both + and -, works to increase the likelihood of a behaviour occuring. So if I want to instill a behaviour, rather than shut one down, I'll go the reinforcement route. Works for me too, I respond very well to reinforcement

[Rom]

Hi all, newbie dog owner here but I am very interested in operant conditioning.

I think the statement "All learning/behavioral change begins with punishment, positive or negative" only presents half of the equation. Punishment, be it +ve or -ve, only decrease the probability of a behavior from occurring. We can extinguish an establish behavior, say counter surfing or pulling on leash, with punishment.

But then again, what if we want the dog to sit in front of the counter to wait patiently for food, or we want the dog to heel? Punishment can never give you those because it can't INCREASE the probability of a behavior.

"People don't change until the pain of not changing becomes greater than the pain of change" and "People do change if the gain of changing becomes greater than the gain of not changing"

ETA: My wife always says that men are like dogs. I think she is right, at least in operant conditioning.

I agree Ah Ngau, that punishment only presents half of the equation (which is why I have highlighted the word 'begins' in the above).....but it is half of the equation, or a portion of it at least. By your example above, how did you teach the dog to wait patiently?......was it, in part at least, by withholding the food until you got the behaviour that you wanted? If so, that portion of the learning process was under the influence of negative punishment. The punishment weakened all of the behaviours that were not about sitting and waiting patiently. So the learning began with a punisher. Then when you got the behaviour you were looking for, you then gave positive reinforcement.

"People do change if the gain of changing becomes greater than the gain of not changing"

Even when the quote is worded this way, I can identify a punisher.....the gain has to be something that the person wants, something that they don't have now to motivate them to change or do something differently in order to get it. You don't get the gain until you get the change what you do right, or, the gain is withheld until you change. When you get it right, the gain is delivered. So you have been through the process of negative punishment > positive reinforcement.

Who was it that said "The definition of insanity is doing the same thing over and over again and expecting different results"? You don't get different results unless you change what you do.

Not sure why you would only focus on + and - punishment, and exclude + and - reinforcement? The way I think of it, punishment, both + and -, works to decrease the likelihood of a behaviour occuring; reinforcement, both + and -, works to increase the likelihood of a behaviour occuring. So if I want to instill a behaviour, rather than shut one down, I'll go the reinforcement route. Works for me too, I respond very well to reinforcement

Yes, but in the process of instilling the wanted behaviour are you not also weakening every other behaviour that is not wanted in that context?......for example, you ask your dog to sit but he stands, so you withhold the reward. You ask him to sit again and he lies down, so you withhold the reward. Finally you ask him to sit and he sits and you deliver the reward. You are effectively weakening the dogs response of standing or laying down when you ask him to sit by withholding the reward...that is a negative punisher. Even if he sits on the first command, you don't give him the reward until he sits.....so the behaviour of not sitting when you ask him to sit is weakened by the withholding of the reward....a negative punisher.

I don't exclude + and - reinforcement in my training....far from it, infact I'd go as far as to say that my dog is one of the most heavily rewarded dogs at my club and is one of the most enthusiastic and reliable workers there. There is however, still punishment in her training and if I don't recognise that, how can I use it to its best advantage?

Thats why I'm exploring this concept......and I really do appreciate everyones input here cuz you guys are really stretching my mind and making me think hard and really explore the material that I'm studying.

Another thought....the ABC of behaviour....Antecedent Stimulus>Behaviour>Consequence.

We know that dogs can learn without us (even though they're often learning things that we don't want them to ). When we put ourselves in the equation though, we think of our commands becoming the antecedent stimulus. But take us out of the picture and what is the antecedent stimulus? I don't think that it ceases to exist........and I can't think of an example where the antecedent stimulus is not a negative or positive punisher....can anyone else?

Edited February 25, 2008 by Rom

[m-j]

Hi

All learning/behavioural change begins with punishment, positive or negative.

I don't believe so in this case

Yesterday I was getting 3 mth greyhounds to chase a bit of sheepskin for the first time. When I put the skin on the ground they weren't to fussed, they just wanted pats, put a bit of movement into it and that was a whole new ballgame. I believe the motivator that changed their behaviour was an instinct, not a punisher. I can't think of why the movement/start of the chase game would be a punisher, the punisher would be not being able to catch it and that would stop them from bothering to try to get skins moved around by me. There really isn't too much chance of that I'm too slow

cheers

M-J

[Akitaowner]

All learning/behavioural change begins with punishment, positive or negative.

I don't believe so in this case

Yesterday I was getting 3 mth greyhounds to chase a bit of sheepskin for the first time. When I put the skin on the ground they weren't to fussed, they just wanted pats, put a bit of movement into it and that was a whole new ballgame. I believe the motivator that changed their behaviour was an instinct, not a punisher. I can't think of why the movement/start of the chase game would be a punisher, the punisher would be not being able to catch it and that would stop them from bothering to try to get skins moved around by me. There really isn't too much chance of that I'm too slow

cheers

M-J

this is a new behaviour not a change....?

this is the same basis of anything the dog will do when alone - for exampel not many people teach their dogs to counter surf but the dog does it! to stop the dog you have to punish

(JMO - i dont like the phrase 'negatively punish' - that infers a double negative - i prefer 'actively punish' or 'actively reward')

this is a good thread :D - i strongly believe the fastest and most efficient way to a modified behaviour is to increase the pressure of 'discomfort'

and then reward the desired. once you increase the pressure of discomfort the dog will automatically offer something new/different.

i cant think of anything that would change without pressure initially - if the dog is counter surfing then they are intrinsically rewarding themselves...same bottom line

[Akitaowner]

just on a side note can you define:

positive punish

and negative punish

are they different? - are they not both decrease in comfort level = punish?

[poodlefan]

I'd prefer to express the concept in terms of the result for the dog. A zoo trainer who's seminar I attended put it thus:

Behaviours an animal finds rewarding will increase.

Behaviours an animal finds unrewarding will decrease.

What a dog finds rewarding and unrewarding may or not be related to punishment. Dogs also self reward.

[Akitaowner]

yep so by removing the potential to intrinsically (self) reward (get the food off the kitchen bench by not leaving it there) you are exerting pressure to do something different (dog will still seek same reward) at which point you can reward a different behaviour offered

[poodlefan]

yep so by removing the potential to intrinsically (self) reward (get the food off the kitchen bench by not leaving it there) you are exerting pressure to do something different (dog will still seek same reward) at which point you can reward a different behaviour offered

Or by removing the food, you deny the dog the reward for the behaviour and it decreases.

Have you noticed how many undesireable dog behaviours are self rewarding?

Barking, digging, fence jumping, fighting, stock chasing. My theory is that these kinds of self rewarding behaviours are extremely difficult to extinguish without the use of aversives.

I have yet to hear of anyone who has managed to cure stock chasing with positive reinforcement.

Edited February 26, 2008 by poodlefan

[Akitaowner]

thats exactly what Rom has said in her initial post - every change in behaviour has to begin with a punishment (aversive)... i agree!

all these people who say they only ever use positive rewards either have a stupid dog that doesnt have a very high drive or they do (maybe subconsciously) use aversives!

i think ob trainers which claim to be 'all positive' are a waste of time and they (the individual trainer) rarely get success 100% of the time with any given problem.

[Ah Ngau]

Intuitively, we associate something bad as punishment but in operant conditioning, a punisher is something that decreases a behavior.

A positive punishment is something bad happens. A negative punishment is something good gets taken away.

Driving in heavy traffic is a bad thing. One morning, I left a bit earlier than usual and didn't run into traffic jam. The next morning, I left earlier again to avoid heavy traffic. My behavior of leaving earlier is strengthen by the consequence of avoiding heavy traffic.

In this example, there is no punishment in the sense of operant conditioning. My behavior was negatively reinforced.

Edited for spelling.

Edited February 26, 2008 by Ah Ngau

[Diva]

You are effectively weakening the dogs response of standing or laying down when you ask him to sit by withholding the reward...that is a negative punisher. Even if he sits on the first command, you don't give him the reward until he sits.....so the behaviour of not sitting when you ask him to sit is weakened by the withholding of the reward....a negative punisher.

No I don't think I agree that it's a negative punisher. The 'unwanted' behaviour simply elicits no reponse, nothing is changed in the dog's environment, there is no consequence, the stimulis conditions are unaltered. If the reinforcer was in sight and removed at the wrong response I'd agree it is a negative punisher, also maybe if I used a no reward maker (I need to think about that one), but if simply nothing happens? It's not then negative punishment in my lexicon, maybe we are using different frameworks.

Edited to be a bit clearer, hopefully

Edited February 26, 2008 by Diva

[Cosmolo]

Dog is lying down, rabbit runs by, dog gives chase. That sequence starts with instinct not a punisher. I also agree with Poodlefan

[m-j]

this is a new behaviour not a change....?

Yes but isn't making a dog sit when it was standing also? The pups changed their behaviour from soliciting pats, to running after a white fluffy thing. Judging by the reaction I got from a couple this morning they learned that the fluffy thing is fun. So with the right motivator when walking into their pen, I'm in the process of changing of a mindset and increasing the likelihood of a behaviour and extinguishing the pat soliciting and it all started by an instinct that the dogs find intrinsically rewarding.

cheers

M-J