Interesting article

Is nobody else biting then? I am going to make a straw dog argument as pointed out at the start of the article 'cos this particular bit is the only thing wot I has any experience in. :p

"Science-style training has repeatedly failed to give dependable performance in the same settings. The examples of “positive” based failures are endless. Watch any agility competition and you will see dog after dog that will not reliably hold a “stay” at the start line..." etc.

Every single week I see examples of the lack of dependable performance mentioned in the article but I don't see this as a failing of positive training, it's a failure to understand and maintain the principles of dog training full stop. The people concerned think they are training in a positive way but what they are so often doing is not maintaining criteria, rewarding haphazardly and being too wary or ignorant of using non-reward markers. Of course, they're not alpha-rolling the dog or smacking it in the head with rolled up towels so it's positive training...

Most dogs will pootle along just fine with these kind of inconsistencies but some, often the border collies, malinois etc. I see being driven almost bonkers with it - and the handler also, wondering why they can't get a consistent contact from a dog whose contact criteria are changed every time they go on the equipment. Most of the time the handler doesn't even realise they are doing it (I have been guilty of this!)

My beef is that this slightly misunderstood use of typical positive-reward methods simply results in a dog that will do what you want it to most, but not all of the time. Aversive training done in a misunderstood fashion, though - hasn't that the potential to do far more harm? It will only take an eejit copying CM's alpha roll, neck punch etc. on the wrong dog to wind up bitten. I have also recently seen with my own eyes the result of using an electric collar on a dog and it was a bloody awful thing to witness. The former may result in a confused, pain in the arse sort of dog but not usually one at risk of nailing you or being PTS in a frothing frenzy.
 
Black Cob I think you pretty much said what I was going to put in that last paragraph - been saying it for years!

I predominantly use positive-based training, and it is always my first port of call, but I think there is a line.

Two examples that truly baffled me...
The first: to get a dog off the sofa he shouldn't have been on, go into another room and do something interesting so the dog gets off and reinforce him for being off the sofa (WHAT!!! turf the dog off!)

The second: recently went to the clicker exp and in the program it said that if you bring your dog along with you, you can't raise your voice or correct it, as it may upset the other clicker trained dogs. I was pretty amazed, if a dog can't cope with a raised voice (that isn't even directed at them) then we've all gone a bit mad.

I've side tracked slightly, my eyes are sore and I'm wafflin', but I think the article has some good points, especially with instinctive drift, that even the college courses are claiming that you can get round if you train correctly. You can manage it but I don't think you can ever truly over-ride it!!!
 
I often think with certain "training" models we get a bit bogged down in the language and actually start to misunderstand the message, I believe in using popositive training methods because I love my dogs and wouldn't want to hurt them or heaven forbid have them fall out of love with me (I am guilty of humanising them to an extent ) but my idea of positive training isn't to constantly give them treats for just not being naughty and being quite good. Mostly good behaviour gets a head scratch and us continuing what we were doing it our walk etc. As stated above I certainly wouldn't reward one of mine for getting on the sofa when they aren't allowed by encouraging them off it with a toy, I was helping a friend with a dog who had decided the sofa was theirs and became aggressive if you tried to get him off it so the sofa got tipped over to turf him off and he was promptly booted outside for a doggy time out - I am a big fan of the time out and have used it with great success. So if I believe in rewarding good behaviour and removing a dog from the group for bad behaviour does this mean I use positive or aversive training methods?
 
blackcob said:
My beef is that this slightly misunderstood use of typical positive-reward methods simply results in a dog that will do what you want it to most, but not all of the time.
Click to expand...

Which is fine for a pet or a sports dog, but for a working dog, a service dog or assistance dog, could be fatal for the dog, handler or both.
Re the comment from Stargirl about the clicker expo, that boggles my mind a bit. If you never expose your dog to stress, and then one day stress happens, or for some reason you are the cause of the stress...that is hugely unfair on the dog. I've seen it myself where everything has been hearts and flowers the whole dog's life and then BOOM, their world has changed - especially when they've been led to believe they are free to choose their own behaviours. At 12 months or 24...I think that is much crueller mentally than a correcting a dog that has learned how to cope with stress.
 
Interesting article, thanks for posting.

I think it's odd that the term 'aversive' is so often interpreted to mean a physical act such as an 'alpha roll' or e-collar shock. If a positive method can be as innocuous as a click from a clicker, or praise in a high-pitched tone of voice, then surely an aversive method can be as simple as a low 'no' or a growl? And I fail to see how such a response to a wayward dog can ever be a bad/damaging/dangerous thing.

I am waaaaaay off being any sort of expert, but I do find the notion that dogs should never be 'corrected' bizarre. If I raised my son that way I'm pretty sure he'd turn out to be a delinquent. I don't need to hit him to teach him the error of his ways: much more subtle methods serve to make him realise that he has let me down :D. The same applies to my dogs, who need to know that there are unpleasant consequences (at a level commensurate with the scale of the misdemeanor) to choosing to follow their instincts rather than my instructions :p Of course, it'd be nicer if they never screwed up in the first place, but life's not like that is it.
 
Positive training = ADDING a stimulus to help achieve/maximise DESIRED results.

Negative training = REMOVING a stimulus help achieve/maximise DESIRED results.
 
Last edited:
I think the lack of reliability in reinforcement based training is not the training itself, but once trained, most people don't remove the element of choice from the dog. You can condition to a point, but to remove the choice you have to enforce - so I can train a really good down stay, but at some point the dog will get bored and break the stay - unless I enforce it. Which I do by putting in an aversive of some kind against him breaking the stay until I ask him to. An aversive (or punishment if you want to call it that) can be anything the animal doesn't like - being pressed back down, water pistol, can of pebbles, citrus collar etc etc - NOT an electric collar, far too open to abuse - but you tailor it to the situation. The trick with aversives is it has to be immediate (not two seconds late, the animal may have presented a different behaviour in that 2 seconds) and it also ideally has to be unconnected to the handler in the animals mind, otherwise the animal just develops a fear of the handler.
Years ago I was taught this by a very good trainer to stop a dog ignoring the "down" and wandering off out of reach. A second person had a rattly can of pebbles and when the dog was within reach signalled to me to tell her "down". If she didn't, the can was dropped on her. Within a couple of times she was dropping like a stone and she never ever ignored a "down" again - and she never ran off out of reach either. I had taken away her option to ignore. Sometimes with a down stay I will just go and press the dog down again, but I have enforced the stay.
 
My dog will stay where I put her for as long as I want her to, she does TV and film work so she has to do what I ask. She is trained using positive reinforcement only .....

Scaring a dog a dog into doing what you want is not training, it's abuse.
 
california dreaming said:
Positive training = ADDING a stimulus to help achieve/maximise DESIRED results.

Negative training = REMOVING a stimulus help achieve/maximise DESIRED results.
Click to expand...

Maybe to you, but not to me. For me:

Positive training = giving a stimulus that the dog enjoys

Negative training = giving a stimulus that the dog doesn't enjoy.

Simples :p

I still don't get how, for example, not praising/ignoring my dog while it pees off after a small furry will ever help it to learn that's not the behaviour I want. I'm sure someone will have an answer :)
 
Broodle said:
I still don't get how, for example, not praising/ignoring my dog while it pees off after a small furry will ever help it to learn that's not the behaviour I want. I'm sure someone will have an answer :)
Click to expand...

Training isn't the same as emergency measures - if you have a trained dog it won't pee off in the first place. My dogs get treats every time they come to me at the yard. Result? They are too close most of the time, to the extent of being under my feet, and so focussed on watching for the next treat they totally ignore anything tempting.
Emergency measures are a swift "down" which should have been trained (and enforced) which can then get reinforced.
 
This is a thread of real interest, and an apparent though slight change of emphasis for this section.

From the rights and the wrongs of what we do with dogs, so we learn, and there is no one beyond that!! From the rights and wrongs, so we make our mistakes, all of us, and providing that they aren't too catastrophic, our dogs are mostly forgiving and offer us an eraser, and that's Time and further attempts! I find that with care, our dogs can be fairly forgiving of our inadequacies!

Alec.
 
Broodle said:
Maybe to you, but not to me. For me:

Positive training = giving a stimulus that the dog enjoys

Negative training = giving a stimulus that the dog doesn't enjoy.

Simples :p

I still don't get how, for example, not praising/ignoring my dog while it pees off after a small furry will ever help it to learn that's not the behaviour I want. I'm sure someone will have an answer :)
Click to expand...

To me giving a stimulus that the dog doesn't enjoy is called punishment.

Giving a stimulus that the dog enjoys is called reward. Simples.
 
Negative reinforcement is subjecting the animal to something it doesn't like and then withdrawing it when the desired behaviour is offered - like a check chain for example. It is a very powerful tool for example when subjecting a horse to clippers. The second it stops reacting and stands still the clippers are stopped. Think of it as the animal learning how to remove the unwanted stimulus by offering the behaviour you want.
The two ways to remove unwanted behaviour are by ignoring it or subjecting it to aversion therapy (you can call it punishment if you like but that implies physical contact, whereas it can be other unwanted actions such as turning and taking home etc). Ignoring takes time and you also need to remember that some behaviours such as chasing are self reinforcing, if not by catching the prey then by producing endorphins.
 
Would the belief that simply ignoring unwanted behaviour not very effectively reinforce, for the dog, that it's OK to continue? If I have a young Gundog puppy who takes off after a hare, then ignoring him will simply encourage further occurrences. Once engrained, stopping the dog will be impossible.

Alec.
 
Yes^^^ you may get away with ignoring some very basic behaviours by ignoring, but you certainly wont with those of a more serious manner, and indeed once ingrained the harder to fix. Some behaviours I woud never ignore.
 
Ignoring would be the scrounging at the table, or yapping at the door, and for attention seeking you can do what I call actively ignore and turn your back and walk away. Behaviour does take longer to die out if it is ignored, and you can get an "extinction burst" (trying harder before giving up) if the behaviour has been rewarded at some time in the past. So it isn't a suitable choice when there is behaviour you need to nip in the bud - that is where the experience comes in.
Like I said Alec, behaviours such as chasing are self reinforcing so ignoring would be to allow that to happen.
The really clever way to stop that behaviour is to train another which is inconsistent with it - so a down instead of a chase, or a recall instead of s scrap - and make the stimulus for the original behaviour the cue for the replacement. So if a dog gets aggressive when other dogs come in sight, make another dog appearing the cue for the recall/down whatever. THAT is training lol.
 
Last edited:
CorvusCorax said:
……..

http://clickandtreat.com/wordpress/?p=884
Click to expand...

I've spent an hour thinking and typing and correcting and more thinking and then more typing and I've now cancelled the bulk of what I've written.

There are so many complexities within the quoted article that it would be impossible to unravel, in but a few posts. The canine mind is not a complex organ, indeed the reverse, it's a remarkably simple mechanism by which a dog reaches decisions. One might think that we need behaviourists with degrees to understand! The simple fact is that we don't.

This stuck in my mind; **"She’s still doing it, more than 20 years later – and still hasn’t clicker trained a dog to any standard. Contemplate that for a moment – an expert who has never actually done what she claims is her specialty"**, and this is the problem.

Can anyone explain to me why the human voice needed to be replaced with a contraption which 'Clicks'? Sorry, but it's laughable!

I'm off on one again, so I'm off!! :D

Alec.
 
This has just appeared on my FB newsfeed...

n9b34.jpg
 
Alec Swan said:
Can anyone explain to me why the human voice needed to be replaced with a contraption which 'Clicks'? Sorry, but it's laughable!

Alec.
Click to expand...

Yes. I can. It is conditioning a la Pavlovs dogs. The dog learns to associate the sound of the click with a reward which is going to follow it, the click is a click for two very good reasons. The first is it is an entirely consistent sound - you can use your voice, or a word but intonations vary from minute to minute. Any consistent sound will do. The other is that it is a clear and precise marker of the behaviour you want - if you are just reinforcing without the conditioned sound by the time you have got the reinforce to the dog the moment may very well have passed. Not hugely important when you are training something simple like a down, but if you are shaping small increments in a more complex behaviour you will struggle to mark the precise action you want. The click does exactly that and is followed by and linked to the treat.
It isn't laughable, it is scientifically based in the same way the all operant and classical conditioning is. HTH Alec
 
JillA said:
Yes. I can. It is conditioning a la Pavlovs dogs. The dog learns to associate the sound of the click with a reward which is going to follow it, the click is a click for two very good reasons. The first is it is an entirely consistent sound - you can use your voice, or a word but intonations vary from minute to minute. Any consistent sound will do. The other is that it is a clear and precise marker of the behaviour you want - if you are just reinforcing without the conditioned sound by the time you have got the reinforce to the dog the moment may very well have passed. Not hugely important when you are training something simple like a down, but if you are shaping small increments in a more complex behaviour you will struggle to mark the precise action you want. The click does exactly that and is followed by and linked to the treat.
It isn't laughable, it is scientifically based in the same way the all operant and classical conditioning is. HTH Alec
Click to expand...

I also allows more than one person to work with any given dog and be able to give the exact same marker as any other trainer
 
I stand (well, ok, I'm actually sitting with Zak on my lap) amazed at the change in training. Used to be that a rolled up newspaper slapped on a leg was the best thing to make the dog behave and now, one should apparently click, treat, pamper, never raise a voice. It's the same with kids! I wonder if ever the old methods will return?

A group of lads asked my OH how long it had taken to get the youngsters to work separately and the other to stay in his spot and thinking about it, he said six months-all without a treat or a clicker in sight! Their treat is the ball then the ball being thrown/hidden again.
 
A clicker just acts as a marker (I use a verbal marker to let the dog know he is doing the right thing, personally I don't use a clicker as I don't have enough hands but they are handy when you have multiple trainers or handlers (I'll click someone else's dog in the exercises when the handler cannot look down or back at the dog - if the dog has sat, stood or downed quickly or is in the correct heel position or done a nice about turn) and as mentioned, they are used for guide dogs, service dogs, search dogs etc which may be passed around multiple handlers.

To me, the article isn't about the pros and cons of the clicker itself, it's about the belief that an animal that is prepared to withstand injury from a prey animal, cannot or should not have to cope with any negative consequence in training, and also the fact that a lot of the science of dog training is based on rats in boxes (including the one with the electrified floor....) , not dogs in the real world with all its various distractions.

Does nobody else think it is unfair as a dog owner to NEVER expose a dog to stress?

Where I live, if a dog chases lifestock, it's dead. I would rather a dog is wary of the consequences of showing strong interest in sheep, than be full of lead because of my moral or ethical stance.
 
People are making the mistake of comparing clicker training and aversives such as electric collars with nothing in between, being strict isnt about knocking the stuffing out of your dog its having a dog that respects its owner. for me if my dogs disobey me they know there are consequences, it might be as simple as sending them out of the room into their bed or harsher a strong check on the lead with umpteen others in between. My dogs are not obedient robots and will test the boundaries, they are not an easy breed as they are very stubborn and Ive never had one from a puppy but from 18months on as rescues. Im not making excuses for myself either as Ive explored all training methods and been kicked out of positive training classes for having a dog aggressive dog, they couldnt help me and didnt want to once they saw what this dog was like but luckily for us both we found an ex police dog trainer who didnt do clicker training and help turn the dog around.

I do use a clicker, its great for training tricks, shaping a behaviour but for very ingrained behaviours it cant work as this article has explained.

CC I think if people never expose a dog to stress they are not doing their dog any favours, lots of people are wrapping their dogs in cotton wool, treating them as their babies, they are dogs and think totally different to us. Its very interesting sitting in a vets waiting room and watching how people interact with their dogs, Ive spent a lot of time lately at the vets with Diesel so have had ample opportunity to observe, its amazing how often people sit down with their dogs beside them and start reassuring their dog the minute it starts to winge, what is that teaching a dog. I just tell mine to be quiet and he is.
 
I don't have a problem with the use of some aversives but I do have an issue with the paucity of human education. You would think that before deriding others on their lack of scientific understanding the author would do some basic reading himself. This would quickly allow him to discover an enormous number of studies on various aspects of dog behaviour (Lindsay's three volume summary is an excellent resource for non-specialists who want a starting point to understanding this research, McPhail's "The evolution of consciousness" is a brilliant, critical account of the emergence of various psychological and philosophical theories on animal behaviour), that Skinner's behaviourism is grossly outdated and not really the basis of clicker training other than in terms of a historical root or even practical facts such that John Fisher, one of the first trainers to develop clicker training and founder of the APDT, used aversives such as startling sounds in his training.
 
Why is it the theorists (and especially the American theorists!) on dog training always have to make things complicated? I suspect they spend so long arguing about the definitions that they never actually get around to training anything! I stopped exporting my dogs to the USA because they are such appalling dog trainers and probably about 500 years behind Europe. But when they come up with another bit of scientific nonsense, we all bow down and worship! <wanders off mumbling&#8230;>
 
You must log in or register to reply here.
Top