Teaching Commands

[Vickie]

The recall thread got me thinking about something. I am a little hesistant to post because I think this idea/method could be misused or I could be misinterpreted...but...

Are there things/commands you teach your dog effectively without external rewards or punishment?

What & why? Are there certain dogs you would/would not do this with?

There are so many things we teach our dogs without even realising or trying to, simply b/c dogs are creatures of habit. Often we suddenly notice one day that when X happens our dogs do Y. Often these behaviours are more dependable than the ones we set out to teach.

As an example, the way I taught Shine to go to her crate was simply to say a word & pop her in her crate, right from day one. In a very short time, she learnt the word & volunteered to go in. I've never really rewarded her for it, nor have I punished her for not doing it, it has always been very matter of fact & I generally say good girl as she goes in. It's not something I want done in any specific manner, just so long as she does it.

I've used a similar method for her recall, I'm sure I gave her treats as a tiny puppy, but not for very long. The day she tested me to find out if she really needed to do it or not, she found out that she did. I've never been harsh with her & if she has tested me I have just helped her make the right decision for a couple of seconds before allowing her to continue on her own.

In all honesty I think recall is too important for it to become optional. I think sometimes by varying reinforcment rates & values, the average trainer can make behaviours less solid than if they are just an everyday thing that we make a habit without ever giving them another choice.

Does this make sense? What do you think?

Please don't think that I don't reward dogs because I use rewards all the time & think it is one of the most important things we can do in training, I am just curious as to other people's thoughts.

Edited October 21, 2007 by Vickie

[sidoney]

IMO there ARE external rewards and punishment (or negative reinforcement or whatever) for the things that you are describing, it's just that they are not as obvious as when you think about what you are using.

E.g. when you take away the choice, how do you do that, and what kind of effect does that have, in terms of learning theory?

Perhaps the question could be rephrased.

[Erny]

As per what Sidoney said ..... the perception of *reward* doesn't always need to be as obvious as a game of tug, or a piece of food. Especially when pack drive is strong - the greatest reward is your acceptance of the dog to your *pack* and your attentions/approval (yep .... "good dog" for going in the crate can do it - especially if the crate did not represent an aversive in the first place). These are the subtleties that sometimes we don't appreciate are working 'behind the scenes'. Dogs are experts at detecting "subtleties".

Edited October 21, 2007 by Erny

[Rom]

I think that basically when you take away choice but are insisting on a behaviour, you are using compulsion or force. I have learned that this doesn't necessarily need to be harsh or cruel to be effective. The way it works from my understanding is that being forced can be stressful = +P, even if it is just a slight amount of stress and when the force is taken away you get -R. So yes it can work inside the terms of operant theory.

Not sure if I'm on the same page or not :D

[Erny]

Not sure if I'm on the same page or not

:D S'ok Rom. 'Cause when I posted, I did so wondering if I was on the right page or not.

[sidoney]

It might be reasonable to suggest that if something is taught in a simple way with repetition, without juggling rewards and so on, that the repetition and predictability may enhance learning, while inconsistency from, say, trying out several training ideas, might inhibit it.

[percyk]

some dogs learn some things really well and quickly even without the reward

one dog of mine has learned to do lots of things...like lie down at my feet when im working or go to sleep on the end of a bed without a fuss..simply cos it meets his needs too! hes a sooky boy that thrives on company

so i guess u dont have to reward those behaviours that are self-serving

the real challenge is to get them to do stuff that they dont wanna do

like not bark to come in (when the same dog has quickly learned to settle at ur feet)

lots is related to the breed and then to the individual...

[Vickie]

You're all on the right page :D . I know I explained it badly . I guess I am really talking consistency when I say habit.

Sometimes I think the problem with theoretical models of learning is that they are based on a consistent approach. Hard as we try, most of us are not consistent. So while the theories of course work perfectly, often we do not, despite our best efforts.

Sidoney I think your rephrasing says it nicely. And yes there are very subtle rewards & punishment I guess to both the examples I gave.

The crate one was an interesting one. I simply placed her (picked her up & put her) in there every day with a single word. Every morning at 9 I put her in the crate before I take the kids to school. I was surprised to see her waiting at the crate one morning at 9, so I opened the door & in she went . She had learned my habit every morning (maybe triggered by keys?) & knew what came next. From then on, if she wasn't waiting I just opened the door & said the word.

So now you have me wondering...was she there waiting for me to hear "good girl"? or to avoid being picked? or simply because it is what we did every day?

I guess you could assume that she sees me picking her up as a negative? (unless you have seen her laying spreadeagled in my lap :D ) As far as positives, I've never really said "Good girl" in a motivating tone for her going in the crate, b/c I am about to leave her, so it has been very matter of fact, but I agree pack drive could come in here. She has on occasion been fed while in the crate but not for going in. I don't think I've used any aversives with the crate, because I can't remember her ever questioning/challenging me to go in. In answer to your question Sidoney, if she did challenge this, I would take her collar, walk her forward a couple of paces & repeat the command as she went in. She also slept in the crate at night as a small pup & did the normal amount of puppy winging initially. In general terms I would say neither positives or negatives have been a strong influence in this process.

In any case. I guess there are a number of ways to train a dog to go into a crate. To me this way just seemed so easy, not very time consuming & has met with no resistance.

[sidoney]

So now you have me wondering...was she there waiting for me to hear "good girl"? or to avoid being picked? or simply because it is what we did every day?

You know her better than anyone. What do you think? Also, is it really possible to sort out a tangle of possibly ambiguous cues/motivations/reinforcers? And is it worthwhile?

That is possibly where it's good to have a clear training situation with clearly understood motivators - because the ambiguity in motivation is less. And so the training situation may be more repeatable, in terms of taking it to other contexts and situations. In your case, it was the same behaviour and context each time, from your description, and so possibly it becomes less important to know exactly what were the motivations, reinforcers and so on.

Maybe I'm rambling. :D

[Mushaka]

it just means you are training with a different method. Alot of dogs are trained as you have described. I also train my dogs this way. Never harsh.they are praised with "good dog",and corrected verbally (not yelling)

And occasionally when they "come",there will be a treat for them.

[Erny]

With the crate .... like an earlier poster suggested, it is likely that it became a comfortable/nice place to be. Consequently going there becomes a self-serving reward. Not to mention any additionals from you such as "good dog" or whatever. Try doing the same thing but having your dog go to an area that does not provide these "self-rewards" and I'd expect you'll find that you'd need more than simply a "good dog" here and there for your dog to volunteer the behaviour.

[Vickie]

So now you have me wondering...was she there waiting for me to hear "good girl"? or to avoid being picked? or simply because it is what we did every day?

You know her better than anyone. What do you think? Also, is it really possible to sort out a tangle of possibly ambiguous cues/motivations/reinforcers? And is it worthwhile?

I think the latter, & yes I think it's worthwhile, that's why I started the thread . What I am trying to investigate is actually the lack of "tangle of possibly ambiguous cues" etc. The basic, don't make a big deal either way with lots of consistent repetition. There is no fading of anything which I think can sometimes make a behaviour less dependable.

That is possibly where it's good to have a clear training situation with clearly understood motivators - because the ambiguity in motivation is less. And so the training situation may be more repeatable, in terms of taking it to other contexts and situations. In your case, it was the same behaviour and context each time, from your description, and so possibly it becomes less important to know exactly what were the motivations, reinforcers and so on.

The reason this has interested me is because her responses are so ingrained that I see no deterioration of them when taking them out & about. I'm not sure that going into the crate is a self rewarding thing...in fact sometiomes it's the opposite, quite often it is so I can train another dog...nothing rewarding about that. When I ask Trim to go into her crate in this situation, she does it, but I can also see her thought process & a split second hesitation. Shine just goes matter of factly, no questions.

Anyway I'm not trying to negate proven theories, 95% of my training will still be based on reward, but it's something that has been going around in my head for a few years & so far I'm happy with my experiments.

[sidoney]

I think that there is a tangle of cues, motivations and reinforcers, but you may not be seeing it because the situation seems simple and straightforward.

Cues - could be a physical movement, could be voice, could be context (getting ready to leave). Could be a mix.

Reinforcers - there is a physical placing - could be P+ and R-? There is the approval from you (even if mild) P+.

There are varying motivations - Shine has different motivations from dogs that are not so keen to have a relationship with a human, for instance. I remember you saying how Bryte just wants to work for the person, too. For a different dog, one that is not so motivated by human-dog relationship, for example, what you are doing may not achieve the same results.

Something may seem straight forward but may not be. Some dogs may respond to what seems straight forward (possibly most) but others won't.

[stumped]

l think the trend to over complicate the reasons , motivations and drives used for training dogs has clouded the fact that a dog will naturally do most basic activities we require for a well mannered dog. l f you are observant and have good timing you can form an association between a cue and an action with no need for reward or punishment. lt is just that, an association, and if a word or signal is related to a specific action enough times, they become linked.

[Vickie]

l think the trend to over complicate the reasons , motivations and drives used for training dogs has clouded the fact that a dog will naturally do most basic activities we require for a well mannered dog. l f you are observant and have good timing you can form an association between a cue and an action with no need for reward or punishment. lt is just that, an association, and if a word or signal is related to a specific action enough times, they become linked.

Thanks Tony, you said very succinctly in 4 lines what I was trying to say (badly) in about 200 lines

My point was that I feel these "associations" can be more dependable simply b/c they are not over complicated & therefore have less possibility of being corrupted by the "average" trainer.

Edited October 22, 2007 by Vickie

[sidoney]

How does one form an association between a cue and an action without obtaining the action? That is a fundamental that is missing in this description.

Yes, there can be a tendency to over complicate things - part of that is people not understanding what can be simple.

But just saying "you can form an association between a cue and an action" is lacking. To form the association is simple classical conditioning. But, how does one get the action?

How do you get an action without any incentive for the dog (i.e. reward, punishment, whatever you care to name it)?

In what way will a dog "naturally" do "most basic activities" (and what are those basic activities)? Will it do it with no input from the person at all?

If we don't give any thought to why a dog does things - the simple idea that there is some motivation there - then we may start to walk down the path of "my dog should do it because he loves me" - and then the average dog owner may punish the dog because he's not doing it. Or something of that ilk.

Edited October 22, 2007 by sidoney

[stumped]

Does a bear cr*p in the woods if no one is there to watch it ?

Dogs go through life every day without someone having to provide an "input" for every action.

All dogs will sit, stand, lie down many times during any given day, they will do many things that can easily be associated with a word.

This is just a component of training like any other.

[Erny]

TMC ... Are you suggesting this is the case if "there's nothing in it for them"? No foreseen reward of any description? Even no 'self-serving reward' ?

[sidoney]

Dogs' self directed behaviour is done for some motivation - to smell a smell, investigate a sound, move into the sun to get warm, sit or lie to rest their legs, and so on. They associate certain cues with actions - tired legs mean sit, fridge opening means go and check it out in case of food, interesting sound means go and investigate the possibility of a chase or bark. All satisfy some desire in the dog. They don't "just do it". Cues from people can also be associated with behaviours - but they don't "just do them" any more than they do their self directed behaviour without some motivation.

[stumped]

The question was regarding, providing a reward or correction.

l suppose if you examine anything there is a self serving reward, they breathe to stay alive, they walk to get somewhere, they lie down to rest, and so on, but indeed to form an association with an action they are carrying out , you do not need to reward or manipulate.