Topic For Discussion - Keep It Nice, Folks ;)

[Kavik]

...giving no treats or not rewarding the dog is already a form of positive punishment ...'positive reinforcement' works only in combination with 'refusing positive reinforcement' as the negative consequence after an undesired behavior. Once the dog is used to positive reinforcement every refusal of positive reinforcement is obviously felt as punishment respectively unpleasant scenario.

Not really. For that to be the case, you would see a reduction across the board of that behaviour that failed to be rewarded. This usually results in general behavioural suppression, which is a pain in the arse for training. Rather, the behaviours that are not rewarded are abandoned through extinction.

...I made a second attempt to digest this, but na, I don't get it...???...

I think she is saying that punishment/corrections can result in general behavioural suppression (dog unwilling to try new things/offer new behaviours) which makes training through positive reinforcement more difficult (especially if you like shaping). Withholding rewards leads to behaviours which are not rewarded not being repeated/going away through extinction.

[OSoSwift]

...and I never met a dog that didn't bark from time to time at other dogs...if this is classified as 'problematic behaviour'...oh dear...

You haven't met many dogs then.

if your dog never barked I would get him checked ...

I had a dog that was 4 before he barked which only happened because he was super excited - he was a tad confused as to where it had come from. He has barked a handful of times all only when super excited and I haven't heard him bark in ages.

There is nothing wrong with him actually.

[kayla1]

Willem - not quite. Positive punishment ADDS something to DECREASE the likelihood of a behaviour occurring again. If anything it's -P BUT "punishment" doesn't describe the dog's emotion but the outcome.

hm, IMO it depends on what becomes the 'normal' scenario. A dog used to treats (or a pat) if he makes the right choice get used to it...I recall my dog and she comes as she knows she gets the treat, and sometimes she even comes without having here called, sits in perfect heel position and her eyes asking me, 'hey where is my treat, that was a perfect recall...come on..'...so this behaviour becomes the 'normal scenario' after intensive training. If she does a lousy job I can 'add' the scenario where I refuse to give here the reward...IMO a positive punishment once the good behaviour pattern with the reward is established.

You add/remove a stimulus, not a 'scenario'. The stimulus in your example is treats. You are withholding the treats, which as others have already explained, is negative punishment.

[karen15]

I can't say I think that hard about training. I like my animals obedient, but I love them to be able to express themselves. The two aren't mutually exclusive. If you establish consistent rules and boundaries the animals know what these are and can then comfortably express their personality within those boundaries. This is immensely helpful when training as there is no wrong answer, only tries. If an animal stops trying, you've done something really wrong.

The most powerful thing I've learned in training is "no" is not a command or direction. "No" what?? Once you get your head around that it is very easy to train as you stop focussing on what the animal is doing and actually direct it to do what you want, which you can then reward.

I have lovely obedient animals, the dogs learn cool tricks as well. They're all happy, outgoing and confident. People really like them.

Edited January 1, 2016 by karen15

[corvus]

...giving no treats or not rewarding the dog is already a form of positive punishment ...'positive reinforcement' works only in combination with 'refusing positive reinforcement' as the negative consequence after an undesired behavior. Once the dog is used to positive reinforcement every refusal of positive reinforcement is obviously felt as punishment respectively unpleasant scenario.

Not really. For that to be the case, you would see a reduction across the board of that behaviour that failed to be rewarded. This usually results in general behavioural suppression, which is a pain in the arse for training. Rather, the behaviours that are not rewarded are abandoned through extinction.

...I made a second attempt to digest this, but na, I don't get it...???...

I think she is saying that punishment/corrections can result in general behavioural suppression (dog unwilling to try new things/offer new behaviours) which makes training through positive reinforcement more difficult (especially if you like shaping). Withholding rewards leads to behaviours which are not rewarded not being repeated/going away through extinction.

Pretty much.

Punishment suppresses behaviour. Whenever you use it, you create some avoidance, and unless the dog knows exactly what results in punishment, they could associate it with several things. For example, if I am in the middle of shaping a paw target, and I punish an attempt that doesn't meet criteria, how does my dog know that he's been punished for not lifting his paw high enough rather than, say, lifting his paw at all? How does he know it was about his paw lift and not about him turning his head slightly, or the appearance of another dog, or a slight shuffle before he lifted his paw?

In contrast, by reinforcing only the behaviour that is closer to what you want, the dog learns what is rewarding, and they naturally do more of it and therefore less of everything else. No punishment required, and this is better, because if you punished everything that you didn't reinforce, then it would be really hard to get any of those behaviours you didn't reinforce later on. I don't want my dog to sit when I am trying to get a down, but I still want him to sit at other times, so I should not punish sits.

[Simply Grand]

...giving no treats or not rewarding the dog is already a form of positive punishment ...'positive reinforcement' works only in combination with 'refusing positive reinforcement' as the negative consequence after an undesired behavior. Once the dog is used to positive reinforcement every refusal of positive reinforcement is obviously felt as punishment respectively unpleasant scenario.

Not really. For that to be the case, you would see a reduction across the board of that behaviour that failed to be rewarded. This usually results in general behavioural suppression, which is a pain in the arse for training. Rather, the behaviours that are not rewarded are abandoned through extinction.

...I made a second attempt to digest this, but na, I don't get it...???...

I think she is saying that if the dog saw not getting the reward every time they do the behaviour as a punishment it would stop attempting the behaviour at all. They don't generally though, as long as they get rewarded frequently enough to know the opportunity is there they keep attempting the behaviour - and that goes for both behaviours we purposely train and self-rewarding things like stealing off the counter.

And when behaviours stop under these circumstances it is because of extinction due to lack of reward, not because the behaviours are being punished.

ETA - Corvus answered for herself while I was posting :laugh:

Edited January 1, 2016 by Simply Grand

[Mrs Rusty Bucket]

http://denisefenzi.com/2015/12/28/have-to-vs-want-to/

Podengo - really good article - that says what Huski said pretty much - it's more about how good your training is.

From the article

If you’re looking for a guarantee, it should be obvious that neither corrections (have to!) vs. cookies (want to!) is going to get you there.

The solution is excellent training. The best trained dogs have the highest pass rates. Not the ones who got the most cookies or the hardest corrections – the ones that are well trained, regardless of the handler’s choice of method.

Willem - I think you're still a bit confused about the what the quadrants mean - it's easy to do which is why I avoid using the the terminology.

The people who claim to be "positive only" are also confused. Hitting a dog - is the adding (+) of an aversive (something the dog doesn't like) in the hope of reducing/stopping an undesirable behaviour.

Apply this to a human example - pinched from Susan Garrett...

You're driving to dog training class... you might have been speeding. The police officer pulls you over and instead of giving you a fine - the police officer shoots your dog.

Do you think you'd risk driving again? With your new dog? What about a child or friend? Is the behaviour the police officer was trying to reduce or stop - what you change? Shooting your dog is only +P if it successfully changes your behaviour. Actually shooting the dog might be -P - taking something away to reduce a behaviour or it could be +P - adding a bullet or an aversive - dead dog. I do really hate the jargon.

Now imagine the police officer who hands out rewards big enough that you're happy to be stopped - maybe $10,000 cash? if he catches you driving nicely in way that helps the traffic flow? This is +R - if you drive nicely more often. R stands for "re-inforcement". Would you be happy to see the next police car or would you still have a little startle in your seat?

So Re-inforcement is about anything that encourages more of a (desirable) behaviour. Eg pulling on lead and the dog getting closer to where it wants to go (reward) - results in more pulling on lead - self re-inforcing. Hence people with a pulling dog - really need to stop the reward of the dog getting to go where it wants when it's pulling. Three pops on the chain but the dog is still getting where it wants to go when pulling - isn't going to work.

I agree with Simply Grand? who said that Cesar has been changing his methods. And with TSD about the body language of dogs in photos with Cesar. I watched one early episode where he "trained" a boxer to be "calm submissive" loose lead walking and the dog at the end looked frightened and stressed. Not relaxed and comfortable. I haven't seen many of the more recent episodes. I hope he's been keeping up with the newer techniques. Tho they are not all that new.

The scariest thing about reward based training is it can be a form of "brain washing". The advertising industry use the techniques a lot and successfully. They're all about changing human behaviour and rewarding their product purchases with "feel good". Never mind if the purchases are at all beneficial to the buyer in the long term.

The techniques work quite well on humans but does require a lot of creativity. Like how do you get an old school trainer to try something new? To even want to try something new. Some of them don't even want to own a computer or a smart phone. Telling them they're doing it all wrong and this way is better - doesn't work. I learn that the hard way.

At the moment at our clubs - a lot of the top competitors who are regularly winning (at agility and obedience) are using the reward based training (reward / no reward and prevent/stop undesirable behaviour), but they're not instructing. The old school instructors won't let them.

At the beach this morning - I saw a lot of dogs pulling on lead. My dog did too occasionally - when we were too close to someone handing out treats for nothing - ok treats for barking and jumping. But most of the time she was good loose lead walking. Sometimes the pulling dog people ask how I got that. Most of them think it's something to do with my dog, not training and that their dog will "grow out of it" even if it's 6 yo already or then it's a breed thing. I've even heard labs are impossible to train. WTF?

And others just pull on by as if it is hopeless. But I can't tell them different unless they want to hear it. How do you get someone to open their mind to something different?

[Salukifan]

It happened when people started to equate "negative reinforcement" with abuse. Like the trainer who told a pet owner that saying "no" to a dog was abuse. Seriously

Same way some "balanced" trainers associate positive only trainers with dogs that had no boundaries. There used to be a standing joke at my club about how easy it was to spot the 'positive only' trainers' dogs - they were the ones that had to stay on lead. I think there is a lot of misunderstanding on both 'sides' of the discussion.

I used to give a damn about what other people thought of my training method and how other people train. Now I simply use what works for me. I'll reward what I want and no reward or verbally correct what I don't.

If you don't wish to make any real demands on your dog, have a fairly biddable dog and don't need to place strong boundaries on acceptable and unacceptable behaviour then positive only will probably work for you.

However a lot of trainers I"ve seen over the years are not what they say they are. They 'no reward' their dogs and consider that's 'positive'. No check chains but a halti is fine. If you think you have to hurt a dog to correct it then perhaps you need to learn a bit more about operant conditioning.

Dogs correct each other. Certainly most of the respected trainers I know will tell you a dog learns faster when it knows what and what is not desired behaviour. When you are training an animal that can chose to blow you off at any time then you'd better be rewarding. However some of us don't have hours and hours to shape behaviour and we don't have dogs that will keep offering behaviour. We need to achieve results more quickly - for both the sake of the dog and the owner.

Edited January 1, 2016 by Haredown Whippets

[Weasels]

Denise Fenzi's follow-up blog > http://denisefenzi.com/2015/12/31/controversy/

Edit: I tell my dogs no all the time. It means 'no, you aren't getting the rest of my lunch so you might as well sod off now'

Edited January 1, 2016 by Weasels

[The Spotted Devil]

Denise Fenzi's follow-up blog > http://denisefenzi.com/2015/12/31/controversy/

Edit: I tell my dogs no all the time. It means 'no, you aren't getting the rest of my lunch so you might as well sod off now'

Love it and so so true - this argument could and does go around in circles forever....

[OSoSwift]

Yep HD you have summed up what I would have posted.

[Simply Grand]

Good blog

When we would start to work with shelter dogs that were untrained and boisterous and had terrible lead and door and mealtime manners we never told them "no" (or stop or don't or whatever), we wouldn't say anything to them, they would just not be given access to the thing they wanted until they showed more desirable behaviour.

So when they got excited and jumped around at having contact with a human the human would stand still and straight and not look at or touch the dog until it stopped jumping around a had a moment of four feet on the floor, which would then be marked and rewarded.

If they pulled on the leash the person would stop, stand still and not look at or touch the dog until the dog moved to a position where the lead was loose, then the dog would gain the reward of moving on again.

If they went to barge out open doors the door would simply be closed again, until the dog showed a more desirable behaviour of sitting or standing still, then the dog would get what it wanted, to through the door.

We used clickers and food rewards as well as the "life rewards" like walking on and going through the door, so it was possible to do the whole thing without using any words at all, but it could certainly be done only ever using the word "yes", never telling the dog "no", and never using "force" or "correction" (bear with me on those terms, I can't think of better ones to use) (Oh, and we didn't do it without words or touch, we used lots of praise and pats, but technically it could have been done without them).

So is that the kind of training people refer to as "positive only"? Or are there really people out there who take the approach of not doing anything at all to inhibit undesirable (and dangerous) behaviours and literally just ignore them and reward the good?

Surely that's not the case, and as Denise says in that latest blog, people are getting confused around terminology and then digging their heels in to say whatever "the other side" is saying is completely wrong (not people in this thread, just in general).

Edited January 1, 2016 by Simply Grand

[The Spotted Devil]

I certainly don't spend hours training and shaping behaviours - my dogs are lucky to get 10 min a day apart from organised classes or workshops or days like today where they spent about 20 min training and 4 hours running and swimming - I train for 3 distinct sports, work long hours and am still studying. Most of my "manners" training is done as part of everyday life.

[Salukifan]

So is that the kind of training people refer to as "positive only"? Or are there really people out there who take the approach of not doing anything at all to inhibit undesirable (and dangerous) behaviours and literally just ignore them and reward the good?

No and yes.

[Simply Grand]

Sorry for my random, train of thought posts..!

I feel like there is also confusion around the term "correction". In a lot of the stuff I've read, including the OP article, it says that a "correction" is something like a collar pop, tap, hair pull (!?!), or in Cesar's case a kick, that is to interrupt the dog's train of thought and get its attention back on the handler.

That a "correction" is not a "punishment".

But "correction" also seems to be mean a consequence that makes the dog behave differently, specifically to stop doing an undesirable behaviour. Well, that is a punishment, isn't it? And different from just getting the dog's attention.

Edited January 1, 2016 by Simply Grand

[Simply Grand]

So is that the kind of training people refer to as "positive only"? Or are there really people out there who take the approach of not doing anything at all to inhibit undesirable (and dangerous) behaviours and literally just ignore them and reward the good?

No and yes.

Does anyone have links to people professing to train effectively in that no limits way? Are they REALLY doing or just saying they are because they conveniently don't consider things like using a collar and lead, shutting doors, teaching meal time manners etc as outside the "positive only" sphere?

ETA sorry, I shouldn't assume, does anyone here train without doing anything to inhibit undesirable behaviours, only rewarding desirable ones? And if so, what the results been?

Edited January 1, 2016 by Simply Grand

[The Spotted Devil]

So is that the kind of training people refer to as "positive only"? Or are there really people out there who take the approach of not doing anything at all to inhibit undesirable (and dangerous) behaviours and literally just ignore them and reward the good?

No and yes.

Does anyone have links to people professing to train effectively in that no limits way? Are they REALLY doing or just saying they are because they conveniently don't consider things like using a collar and lead, shutting doors, teaching meal time manners etc as outside the "positive only" sphere?

ETA sorry, I shouldn't assume, does anyone here train without doing anything to inhibit undesirable behaviours, only rewarding desirable ones? And if so, what the results been?

I do know of trainers that use a lot of luring and don't train for failure like I do - my opinion is that it does take longer because the dog can't focus on what they are doing as all they see is the treat. At the same time I know of plenty of trainers (who use varying methods - positive or otherwise) who "nag" the dogs constantly - repeated commands drive me nuts. I like to put the onus on the dog to make good choices. If I say "watch" to my retrieving dogs that means...sit straight at heel, look forward, watch the mark and remain rock steady until I tell you to retrieve it or leave it. Obviously I train it in stages but it's mostly left to the dog to figure out. Do this? Get this!

Edited January 1, 2016 by The Spotted Devil

[karen15]

Simply grand, why wouldn't you give the dog an instruction while doing the physical activity? It's the same as looking at a person and expecting them to know what you're thinking. Giving the command and doing the action to get the behaviour surely is a faster way to train?

When I say "no" is not a command, I mean when a dog is jumping, "no" gives it no direction. If instead you give the dog the command for the action you want eg sit, then you can reward the dog for sitting, this coincides with the dog ceasing to jump which is what you were wanting to stop. Being consistent will mean the dog will sit on most occasions instead of jumping. If your dog is running after something, "no" is meaningless. "Leave it" is much more useful and again can be rewarded when the dog does it. All of my animals have a "with me" command, that means stop what you're doing and assume your designated position - dogs at heel to the left, horse at heel to the right, cat infront. Currently new dogs are learning "with me".

Tell the dog what you want it to do, not "no".

[Simply Grand]

Karen, I'm assuming you are asking about in the case of the shelter dog training? It's because they don't know the cues yet. They haven't learnt than when I say "sit" I mean "put your bum on the ground and stay there". Saying it would just be noise to them. But without any prior training they can quickly understand that "when my bum is on the ground I get something yummy". You add the cue in afterwards, once they have the behaviour.

We also want the desirable behaviour to be the default, without them having to be given the cue - as in, if you aren't told to do anything else, please put your bum on thr ground and stay there, if we are walking and you haven't been told something specific please keep the lead loose, etc

Incidentally we taught exactly the same techniques in puppy classes, with the aim of avoiding the problem behaviours ever starting but with the same idea that an 8 week old puppy has no idea what human words mean yet but can understand "when I do this I get something good".

Edited January 1, 2016 by Simply Grand

[huski]

Denise Fenzi's follow-up blog > http://denisefenzi.com/2015/12/31/controversy/

Edit: I tell my dogs no all the time. It means 'no, you aren't getting the rest of my lunch so you might as well sod off now'

Love it and so so true - this argument could and does go around in circles forever....

100% agree, and I usually find the most vocal are the quietest when you ask them to show you their dogs. Funny that.