Positive Re-enforcement Only Techniques

[pgm]

"pgm: Why can't you train a dog with operant conditioning?"

I didn't say you can't. Many dogs and animals have been trained using operant conditioning. I just don't believe that in the case of dogs (in particular) such training develops the dog and its relationship to its fullest extent.

[Tralee]

pqm. ;)

You make an interesting point and one worthy of further investigation.

[pgm]

pewithers, I not currently in discussion with MattandBuddy. I am rather following a line of thought developed by Tess32 and myself. MattandBuddy may well find such discussion interesting or not. But as far as I know there is no rule that forbids a thread from developing into a different, albeit related, line of thought.

You have said: "You suggest that there should be a separation between practice and understanding."

I have suggested no such thing and fail to see how you have come to such a conclusion.

[Bonniescot]

pgm: Can I ask what method/training you believe develops the relationship to its fullest extent?

Cheers

[abzndbonnie]

what i do is most positive training and some firm training. by firm i mean using a deep voice. Not yelling, but letting her know who's boss. I think dog's need it. In the wild dogs get growled at if they do the wrong thing, so this is a way for me to let her know im in charge. Its also the way I was taught at obedience training. Works for me. still plenty of treaties but! ;)

That's still positive training - positive training doesn't mean sounding all nice and happy all the time and shoving treats into their mouth at every opportunity

Nat

O. lol ... Well i have tugged at the chocker chain, but only for a week or so because i found IT DOESN'T WORK. the dog doesn't get it. you do it often enough the dog gets used to it. So i started the watch me command. found it works much better. I am just discussing with myself lol im a bit lost in what's going on with tess32 and pgm and all those. ;)

[pgm]

Tess32: "What do you mean by the dog's understanding? Understanding of the point of the behaviour or?"

What I mean is something along the following: What is or is not reinforcing to the dog? It depends on the dog. Everytime you reinforce a behavior you are doing so within a particular context. Different dog equals different context - hence some dogs will work for food, some wont, some will work for toys, some wont. Equally, the same dog in one particular context (the backyard) will find a treat reinforcing, but in a different context (at the park) wont find the treat reinforcing.

Training a dog is all about your ability to read the context in which you are using reinforcement. Same as being successful on the stockmarket depends on your ability to know what is and what isn't low, and what is and isn't high. What will work for one dog, will not necessarily work for another, what will work in one particular context will not necessarily work in another.

So the question is: how does what the dog understands or doesn't understand change the context in which you are working? It can change the context significantly. It can mean the difference between applying a correction that the DOG INTERPRETS as positive punishment, to one that the DOG INTERPRETS as positive reinforcement.

But can you see the problem? How the dog interprets (what's going on inside the dogs head) is irrelevent to theory of operant conditioning. The internal states of animals and how that affects behavior is rendered null and void in operant conditioning.

But what is going on inside the dogs head - how the DOG ITSELF interpets the context - is the most important aspect of training. A good trainer knows this, whether implicitly or explicitly, and hence develops the right kind of understanding in the dog.

Edited February 8, 2005 by pgm

[pgm]

"pgm: Can I ask what method/training you believe develops the relationship to its fullest extent?"

I am not going to suggest any particular method as there is more than one method that can be used.

But I will say this, the difference between trainers that use operant conditioning is found in their vocabulary - they talk of shaping and modifying behavior.

The trainer that I like and admire on the other hand, talks in a vocabulary that includes talk of treating dogs with dignity, developing their character and sense of responsibility, of always being fair to the dog.

The difference in vocabulary is significant and signifies a great deal - but it a significance that is not always apparent to the naked eye. For instance, my dog may look up to my eyes because he loves and respects me, on the other hand he may look up to me because I have trained him to do so with food treats.

Such a difference is not easily discernable to a spectator (indeed, to the behaviorist who is soley concerned with the external manifestation of behavior there is no difference at all) - but there is a difference, and a significant difference at that.

[pgm]

"O. lol ... Well i have tugged at the chocker chain, but only for a week or so because i found IT DOESN'T WORK."

abzndbonnie, no training tool works or fails to work of its own accord. The person using it makes all the difference.

Believe me, if you were instructed in its correct use, you would it find it working within a couple of 10min sessions at the very longest.

[abzndbonnie]

i was taught by all different trainers on it, but bonnie just got used to it in the end and quite frankly sick of it. And i was sick of tugging on her little head. Making her look at me and watching what i was doing and not me pulling her head and her going "oo where is she tugging me next" worked much better. She concentrates on what i'm doing much more and we have gotten much more benefit out of it I understand some dogs may be different but she just got over it.

[pgm]

abzndbonnie, if what your doing works and your happy with the results then by all means stick to it.

But I can tell by your description - "but bonnie just got used to it in the end and quite frankly sick of it" - that you were not taught correctly. Taught correctly, the only thing that the dog gets 'used to' is staying in the heel position.

But as I said, if you are happy with the results you are now getting by all means stick to what you are doing. Whatever works...

[Rusky]

The difference in vocabulary is significant and signifies a great deal - but it a significance that is not always apparent to the naked eye. For instance, my dog may look up to my eyes because he loves and respects me, on the other hand he may look up to me because I have trained him to do so with food treats.

Such a difference is not easily discernable to a spectator (indeed, to the behaviorist who is soley concerned with the external manifestation of behavior there is no difference at all) - but there is a difference, and a significant difference at that.

yes there is a difference, one dog gets free food and "positive punishment"???

So looks at you with adoring eyes

the others have to work for their tucker

poor puppies :D

Life is tough for a dog

[pgm]

I am sorry Rusky, I have no idea what you are trying to say.

[Bonniescot]

pgm: mmmm INTERESTING!!

My knowledge on Dog Psychology is limited, so I will leave it up to others to comment.

When my pup comes running up to me at the park, without being called, she looks up to me for a reward (she is rewarded on a random ratio). Is she confused/does it effect our relationship.........who knows!

Cheers

[Guest Tess32]

But can you see the problem? How the dog interprets (what's going on inside the dogs head) is irrelevent to theory of operant conditioning. The internal states of animals and how that affects behavior is rendered null and void in operant conditioning.

PGM - I've been re-reading your posts but still not sure I truly understand what you mean.

As far as I can tell, there is nothing in OC that defines what the reward is, or what will be reinforcing to any particular animal. I have not read any primary texts, but all the secondary texts I have read make a point of saying that a trainer should choose a reinforcer that truly is....reinforcing. This does not specifically state, but implies, taking into account the individuality of the learner and their response to context and drives.

Then I was thinking that maybe what you mean is that OC doesn't really impart "knoweldge" or understanding - it merely exists to explain why and how behaviours increase or decrease. Eg...it doesn't refer to a dog learning what a "sit" really is, just that plonking its bum on the floor is reinforced more often than not, and so is worth doing.

Maybe it is possible for a dog to learn without understand click = correct, but I think most come to learn it and I think that IS truly understanding. My dog will offer me behaviours now - he knows that he won't get a treat until he can figure out what I want.

I believe, but am not sure, that the clicker is useless unless the dog understands clicker = correct. You could continue shaping just by reinforcing a behaviour with food etc and the dog will increase the behaviour. It is when the dog understands the click that true communication is going on - the click says - yep, that's what she wants, repeat it.

Sorry if I have misinterpreted you - it's a cloudy subject and hard to put into words.

Nat

[sidoney]

This topic is so perennial.

What it boils down to FOR ME, is this:

Dogs are capable of learning by many approaches. Any approach that is used must be consistent and, to be ethical, not harm the dog mentally or physically, rather it should be rewarding and enjoyable for the dog - and handler!

People are capable of training by many approaches and what suits one trainer will not necessarily suit another. Also, trainers may vary in their ability, no matter what approach they use.

If you have a training approach that is effective and ethical, that suits you, then great. It won't suit everyone. But then, nothing suits everyone.

If your training approach isn't effective, ethical, or suitable, then either you need to get better at what you are doing, or you need to consider switching to something else.

And if someone has a different training approach to you, that doesn't work for you, but it meets all the criteria (effective, ethical, suitable for them), then good luck to them.

[Rusky]

sidoney

[Guest Tess32]

sidoney

Ditto

Nat

[Vickie]

That'd make 3 ditto's

[pgm]

Tess32: “As far as I can tell, there is nothing in OC that defines what the reward is, or what will be reinforcing to any particular animal.”

PGM: True. But then please explain why so many (not necessarily yourself) ‘positive’ trainers, who rely of the theory of operant conditioning as an authoritive source for their methods, are so often downright hostile to traditional trainers who employ physical corrections as part of their methods?

In other words, as you have stated, there is nothing in OC that defines what will be reinforcing to any particular animal. Because what I am saying is that physical corrections can be, and indeed are, when properly administrated in the right CONTEXT, positively reinforcing.

But I find in general, that people under the influence of OC theory, cannot (will not?) understand this. And as a rule of thumb, the more infatuated they are with the theory of OC, the more hostile they are to the idea of physical corrections.

Now the short answer as to why they fail to understand how corrections can indeed be positively reinforcing, is because they fail to take into account (because it plays no part in the theory) how what is going on inside the dog’s head (the way the dog itself interprets) influences the context in which reinforcement is applied. Exactly the same reinforcement can change from being positive reinforcement to positive punishment depending on the change in context. What is going on inside the dog’s head (its level of understanding) is just as influential in shaping context as anything else. Failure to understand that typically results in the kind of hostility directed at trad trainers.

Tess32: “Maybe it is possible for a dog to learn without understand click = correct, but I think most come to learn it and I think that IS truly understanding.

PGM: No. The point at which a dog can truly be said to ‘understand’ is when ‘click’ is no longer necessary.

Tess32: “My dog will offer me behaviours now - he knows that he won't get a treat until he can figure out what I want.”

PGM: so long as your dog is motivated by reward, or even just the possibility of reward, it can be said that he has YET to actually understand.

This kind of thinking (not necessarily your own) is reflected in statements people make to the effect: I want my dog to come to me because he knows he will get a reward, rather than come to me because he is afraid of being corrected. In a dog that ACTUALLY understands, as opposed to one as YET to understand, reward and punishment are irrelevant. Understanding itself (as opposed to the reward/correction paradigm) is the motivating factor in behavior. In other words, motivation is internal, rather than external.

Tess32: “Then I was thinking that maybe what you mean is that OC doesn't really impart "knoweldge" or understanding - it merely exists to explain why and how behaviours increase or decrease.”

PGM: the problem with OC is that it does not suffice as an explanation for behavior, for it can only explain behavior in terms of external stimulus. Remember what I said about B.F. Skinner, he dismissed out of hand the INTERNAL states as having any bearing on what motivates (hence explains) behavior.

[pgm]

Sidoney: “Dogs are capable of learning by many approaches. Any approach that is used must be consistent and, to be ethical, not harm the dog mentally or physically, rather it should be rewarding and enjoyable for the dog - and handler!”

PGM: in one sense I agree with this, in another I disagree. But I appreciate your willingness to be open-hearted to different approaches.

But firstly, to put it frankly, the question of ethics is entirely dependent of where you place the emphasis. Personally, if one could successfully train a dog with an electrical cattle prod I would have no objection. Likewise, if one could successfully train a dog by continually banging it over the head with a crowbar, I would likewise have no objection. This because I put the emphasis of ethics on the end result, not the process of training. (when I say emphasis, I mean I put the emphasis, not the whole of the matter).

To anticipate a possible objection, what I mean by a trained dog is a confident, healthy, well balanced, obedient and reliable animal. That one could not produce such a result with a crowbar is not in dispute: I am simply saying if it were possible, I would not object.

Where I disagree with other approaches is their preoccupation with the process of training over that of the result. For instance, I could give you many examples of activities that were highly satisfying and brought enormous fulfilment and sense of achievement, yet were far from ‘enjoyable’. Running a marathon for instance, could hardly be said to be ‘enjoyable’, yet achieving the end result of finishing brings enormous satisfaction to those who do. Likewise, climbing Mt Everest is a PROCESS of physical and psychological deprivation, yet brings enormous fulfilment to those that achieve the top. So much so that many do it again, knowing full well the physical and emotional suffering the process entails.

Where this mentality impacts dog training is that people cannot stand even the thought of a dog suffering discomfort in the process of training. I disagree with this mentality not because suffering is good for its own sake, but because the refusal to ask anything of dog that might entail discomfort sells the dog short – sells the dog short of the inherent satisfaction that comes from achievement, even from achieving something that might be difficult.

The worst and most extreme form of this mentality comes from people who say, ‘I would rather euthanize Ben [a dog] than treat him in such an abusive manner’. This is an actual quote from a clicker trainer discussing the training methods of another trainer who uses a prong collar to rehabilate dogs with aggression problems.

For myself, this is unethical, to deny the dog its life simply because rehabilation might be difficult and entail some pain. This to put the process of training over and above that of the results – for me this is unethical.