Man, in my dog training course and putting together curriculum to hopefully be teaching my own beginners and puppy classes in the new year, the hardest thing going has been this: figuring out which 'label' of dog-training I fit under.
I have been doing lots of reading on the scientific positive-only based models that reward for correct responses and ignore incorrect responses. Their belief is that lack of attention in and of itself is a punishment/consequence to a dog. I have been watching lots of videos with really good information and technique to creating bomb-proof behaviours out of puppies. However, coming from a world of rescue where very, very few are blank-slate puppies of 8-10 weeks key behaviours have already been rewarded in an act that they are self-fufilling. IE: The dog just likes to do the behaviour, and doesn't really mind if there's a lack of attention from their handler because of it. IE: The behaviour has been practiced and habitualized too much already to (in my opinion) take away based on a pure 'ignoring' approach. Of course they go on to say to interrupt behaviours, etc, and suffice to say in short form that I agree with almost all positive-based training.
We must mould, shape and mark behaviours in order to create a dog that learns and thinks. At the same time, I'm truly struggling with where respect comes in within a purely-positive model. How exactly can you tell a dog what they SHOULD NOT be doing if you only ever tell them how good they are?
I found this good analogy:
"And the problem with (positive-only training) is:… Imagine somebody trying to teach you to drive from Los Angeles to St. Louis … but they ONLY tell you when you’re going in the right direction, and they NEVER tell you when you’re going in the wrong direction.
Either you’re going to get completely lost, or it’s going to take you a VERY, VERY LONG TIME to get to St. Louis. Of course… you might have a lot of fun taking a long, long time… but if you’re trying to get home on time for Christmas dinner… forgettaboutit!~ Plus, when you train exclusively with positive training techniques… you will never end up with a dog that is 100% reliable in a “street smart/around town” environment.
I recommend using BOTH positive and negative motivation. The wise trainer will always adapt to the dog and respond with more or less positive or negative motivation in response to what the dog is giving him."
In my training I have been exposed to the use of pressure, a 'hands-less' form of correction that deals with the brain/hierachy of a pack. Dogs use pressure on each other all of the time, it is something that is ingrained in them. If you think of horsewhispering, it is very similiar. It is using body language to own space -it teaches dogs to be soft and does not involve physical force like leash corrections, etc. To a traditional 'positive' trainer, pressure is intimidation and therefore is punitive.
How though then, if you do not use any form of saying 'no, that's not what I want', do you gain reliability and respect. Should a dog be a machine when it comes to commands -no-, but I want to know that even under a tense environment, that my dog will still look for my guidance as their leader.
Only saying yes also creates a great possibility that the environment of giving only positive lets the dog do the bear minimum to earn their reward. Smart dogs that don't naturally want to be a follower catch on to this very fast -I saw first hand a little mini aussie who had gone through 12 intensive weeks of clicker-training who still had no respect for her handler-. The handler was doing everything the trainer asked, but being a good leader -in my opinion- is knowing when to give praise, when to withhold it, and when to disagree with behaviour via a pressure correction.
It also goes back to each individual dog in their motivation to 'be good', and what they will work for. Even in my own house it varies greatly.
For example, Beckett almost solely works for 'freedom': IE: off-leash priviledges. He couldn't care about food 9/10 (though will take anything if he doesn't have to do anything for it). His motivation then is negative motivation to begin with. In his mind, the leash is punitive. Freedom is his positive. He is not a sensitive dog, he has to know who is boss. If we had only ever applied positive with him he'd be running us. It goes back to being an aloof dog. If the only consequence was not having my attention -he wouldn't care at all.
For a consequence to be a consequence, it has to mean something to the dog shouldn't it?
Lemon will work for food, and love and she is overall quite a soft dog. She needs much more positive than negative to excell and to be honest it's very rare that she would need a correction -hands off or otherwise-. She is a dog who could be get away with a purely positive approach. It stems back to the 'consequence' of absolutely no attention being a true consequence for her. Stems back to motivation.
I go back to the same line:
For a consequence to be a consequence, it has to mean something to the INDIVIDUAL dog shouldn't it?
In a pack of dogs, a true alpha type dog applies their leadership when and where it is needed, and if you watch them they apply different force and timing to each situation.
When I start training, I want to train people to have their dogs work FOR THEM and be rewarded for working FOR THEM, and not just rewarded on correct 'behaviour' alone. To me, it isn't fair on any person or dog to ONLY give them 'the way' without showing them where they've gone wrong and getting them back on track.
But, whenever you are teaching something new, it should be POSITIVE and FUN! It is unfair to give correction if the dog doesn't know to perform the correct response to begin with.
My two cents. I love my dogs to pieces, but they are my dogs and I am responsible for their actions in our world. They are a gift to me, but they are not god's gift to the world and aren't perfect all the time. I will praise them when they succeed, I will set them up to succeed, but I will also steer them back on the path to success if they happen to decide to stray from it. Whether that makes me a bad trainer in the eyes of some, so be it.
If there's one thing two trainers can agree on, it's that the third trainer is doing something wrong.
In foster terms, Urchin is headed over to Vancouver tonight. He is going to visit a potential family but if they aren't a match for him he's being fostered in Vancouver while I'm away anyway. He's a cute little stinker and sure has come a long way in a few weeks.
This video is too long to upload directly to the blog, but if you want take a peek: http://www.facebook.com/#!/photo.php?v=10151175681038671&set=vb.729918670&type=3&theater
In it I play a game with Urchin so that he allows his collar to be touched, and goes towards an outstretched hand. This is a great game to play with sensitive, shyer dogs as if they got lost somehow, they are more likely to come to a strange hand. It can also help with hand-shy dogs who have been been hurt in the past. Counter-conditioning for the win!
Fostering dogs in Victoria, BC. Giving them lifeskills, cushy pillows, structure, confidence and copious snuggles. From chihuahuas to Rottis, to huskies and anything in between, amusing stories, mostly useless dog tips, and cute dog pictures abound.
3 comments:
Emily, our trainer (a certified Karen Pryor Academy trainer) explained the positive model this way:
If you tell your dog "NO!" when he does something wrong, you are not telling your dog what he SHOULD be doing. He (or she) may understand from your tone of voice that you are not happy (or more likely, will understand you are angry) but without the information of what behaviour you do want to see, the options are limited, and in all likelihood the bad behaviour will be repeated another time. Also, if what the dog wanted was attention (by jumping up on you, for example), you have rewarded the behaviour - because you gave the dog attention even if it was 'angry voice' attention.
The idea is to teach the dog the behaviour you do want, so the behaviours you don't want are extinguished. If you want the dog to sit when greeting someone, reward him for sitting (spontaneously or with a voice or hand command) when someone approaches, let him learn that the correct way to greet someone. Then the jumping on people becomes a non-issue because it disappears.
I found this explanation very helpful in understanding why the word 'NO' is forbidden in our training program.
I totally agree with teaching a dog what to do not 'what not to do'-it is unfair to any dog to use anything negative on behavior they haven't been taught to perform. At the same time, from what I gather positive trainers are using an interrupter noise, like a 'leave it and come to me' noise. In a way, that is saying 'No'. Does it matter if you use the word 'no' which you train to use the same context: stop what you are doing and COME to me.
My concern doesn't come in initially training a new skill -as the learning stage should all be positive, or else why would the dog want to learn to begin with.
I'm thinking more for a dog who has recalled 1000+ times beautifully and KNOWS what his or her handler is asking, but is just blowing them off (maybe I am too familiar with the husky breed). In two examples from my guinea pig class. One husky mix knew her commands, but would test her people -in typical husky fashion- to see whether they would do anything about NOT doing the command. The reward of pushing them was more rewarding to her than the food (great food, and for weeks she'd only been fed by hand). How would a solely positive approach be able to get respect from this dog when her initial motivation is getting away with stuff?
Same with Beckett. He would never be able to go off-leash reliably without the once in a blue moon 'mother means business, you down and stay or your punishment comes in fast'. For him, the leash in many senses is punitive in and of itself. I don't have to use it often, but it is a tool I employ or he just won't take me seriously.
The other example in class is a dog-reactive dog. Of course you put counter-conditioning to work in getting them comfortable around the other dogs, and getting them focused on their handlers, but do you just ignore if your dog outbursts at another dog without disagreeing with the action (to me a pressure, hands-off correction)? What if your interrupter fails?
I understand putting all the pieces into place in hard training to hopefully have a remarkable dog who is wanting and willing to work hard for their handler. However, I think some onus has to be placed on the dog once they have learned the behavior you've asked. You've done your part in giving them all the tools they need to play their role in the relationship.
Am I saying you should beat your dog or use harsh physical corrections. NO!!!!!!! I'm just saying it's unrealistic to have a relationship built on a participant never being able to be 'wrong' EVER.
I've also seen a few dogs while I was doing my certification who had been through classes and classes of Karen Pyror based classes, and the dogs didn't have respect for their handlers in the slightest (I know that doesn't speak for all dogs though).
It comes down to the obvious though: not all training methods work for all dogs. I go positive until the dog's actions require that I implement and supplement with correction in terms of what works for the individual.
I love this article (directed mostly to children, but positive reinforcement is positive reinforcement regardless who it is used on): http://www.drmaglio.com/articles/parenting/negatives.php
Especially this quote: " A parent using only positive reinforcement unwittingly encourages the child to develop manipulative techniques to avoid responsibility for inappropriate actions."
And also this quick article too, more in line with dogs: http://articles.petoskeynews.com/2012-09-28/obedience-classes_34153534
Sorry, the second article was this one, I had to search the computer's history to find it: http://www.dogproblems.com/dog-behavior/the-problem-with-using-only-positive-dog-training-techniques/
Post a Comment