|
Location: UK
Joined: Oct 2010
Posts: 1,096
|
|
Originally Posted by
smokeybear
There are good and bad trainers out there, just as there are good and bad dog owners.
Before I set foot in a class with a dog, I go and observe ALL the trainers etc, if I don't like I walk.
Like any "new" thing (although of course OC is 80 years old) people will jump on the bandwagon.
I have observed people who claim to be clicker trainers who have no understanding of the difference between CC and OC (Joe Inglis is case in point of a vet who is stepping outside his field of knowledge in his latest book and saying clicker training is the same as Pavlov and his dogs)!
They USE a clicker, ineffectively and are not even cross over trainers, they may have seen someone use one, or even been on a course, but they are unable to transfer any skills they may have learned into reality.
There are plenty of purely positive trainers out there, you just need to know where to look!
While you certainly can and should have "mainly positive", or "ideally positive", you cannot have"
purely positive" in effective training.
As soon as you withholding a reward, because the dog has not offered the required level of performance/compliance be it food,toy or attention etc, is not positive. Even having poor timing and being slow with the reward will involve a degree of potential frustration for the dog.
However rewarding everything indiscriminately, the only way to avoid this, will not train/teach anything.
Dogs who run through a gamit of behaviours, or gain speed and intensity (which may well be the intended training goal) until rewarded are acting through negative reinforcement. If the reward was
never given again this would have been an extinction burst, however if it is given it then becomes a performance increasing intermittent reinforcement schedule.
Over time expectation can cause this to be eager anticipation but in the early stages the motivation is likely to be frustration.
For the same reasons, those less motivated individuals will consider the prize not worth the price and decrease their performance
If there is a behavioural problem with the dog lunging on it's lead towards something the dog is getting a negative lead check, irrespective of whether that is initiated by the dog itself, or deliberately by the handler.
Training is essentially about handler-generated and external (i.e the rest of the physical enviroment) motivators being balanced against the dog's self-generated and internal (biochemical/emotional/genetically predisposed) motivators.
When a self-reward factor for the dog outweighs the reward for alternative desirable behaviour offered by the handler, in situations where the environment cannot be totally controlled, some aversives/negative inputs are likely to result, if only e.g. the the dog's head being turned from the distraction via a head collar or pressure on the front strap of a harness, etc.
I would also agree with another poster that a few of the smarter dogs who are never corrected in any way, including an Ah ah etc, will learn to repeat the behaviour that was substitued as a two step procedure to get a reward. the removal from the sofa was a good example as it e.g sit when greeting, or 4 on the floor, to correct jump when some will jump again to repeat the sequence, to gain more treats, attention, petting etc.
Some "positive only" claims are because people don't always recognise or accept when something is not positive but others are about cynical marketing.
E.g. a "signal of non reward" is not neutral, as that would signal nothing, and it is not rewarding/positive but rather it is an aversive as it signals disappointment. In my book that is what I would term a correction, so imo not positive only, although the correction is described in words that suggest a more positive outlook.
As it is relevant here I would also repeat part of my post from another thread:
Professor Daniel Mills (Professor of Veterinary Behavioural Medicine at Lincoln university) gave a talk on Putting learning theory into practice, in which he stated that claiming not to use punishment in training reflects a fundamental misunderstanding of the application of learning theory in practise and that, even without any physical punishment, the absence of an expected reward is a form of punishment, and to ignore that punishment is occurring under these circumstances is to ignore an important part of learning that is occurring during training.
He also said instinctive behaviour patterns may response more effectively to punishment that to positive reinforcement, but than this punishment should be used in skilled hands. He also said there can be problems arising from poor timing of rewards
So, according to Professor Mills, you cannot train by reward/positives alone, and even use of rewards/positives have to be carefully managed to avoid shaping the wrong behaviour or bringing about undesirable behavioural tendencies.
|
|