Part 2 of Ian Dunbar Seminar!

Since apparently Crystal is going to take awhile to write about he Ian Dunbar experience, I might as well finish up on the other things that really caught my attention from Saturday.

One question I had about his RRNR (basically repeating the cue and walking closer continuously until the dog does the behavior) was if it could inadvertently reward the behavior you are trying to punish.  What first made me ask this question was when Ian was talking about teaching dogs to heel.  A lot of trainers only focus on the negative and are constantly leash popping their dogs for being out of position.  Ian was arguing that not only do we need to reward the dogs when they do the right thing, but that we need to give more constructive feedback such as teaching the dog a word such as slow, or hurry up and then use those cues as information to the dog.

My question to Ian, was that can't a cue that was positively trained actually reward a behavior?  In the same way that a finish can be a reward for a straight front, couldn't a dog learn to forge and then be rewarded by the cue to "slow."  Or as a better example, I think a lot of dogs out there learn to jump up first before sitting.  They jump up to get the owners attention, are asked to to sit, and then nicely fold back down where they receive lots of praise.  They learn not only that sitting gets petting, but as part of a behavior chain they learn to jump first, then sit.  I did try and ask this all to Ian, but either I didn't ask it correctly and he misunderstood me, or he just blew it off because he doesn't think it occurs if you use RRNR correctly.

Another aspect I found interesting on Saturday was that he actually teaches his dogs different levels of performance based on their name.  Ian formalizes disobedience :)  90% of the time Ian argues that we don't actually care if the dog obeys or not.  If he really wants the dog to listen, then Ian uses the dogs full name.  So Rover Sit would be optional.  But Rover Doggy Sit would always be enforced.  And if he wanted pizzaz, a nice tucked it sit, that Ian would use yet a third name to go be before the cue sit.

I thought about it and I do currently have a little system for my dogs. But instead of using a cue to go before the real command, I just use different commands.  If I say sit, down, stay, come, all done, etc then I ALWAYS enforce it.  But if I tell Vito to go relax, just lie down, come cuddle, come over here, no more, then they are basically optional.  I would like it if he would down, or come, but I don't really care if he does it or not.  And if I want precision, then I use his formal obedience cues such as front and heel.  Having more cues makes it a bit more complicated than just putting a different name before the cue, but I would think that it would be easier for the dog to learn what you expect.

And finally, my last observation about Saturday was the use of the command in the praise as in Good Sit.  Now I know that dogs are really great at learning what we want so it probably doesn't matter one bit.  Traditional trainers have always said the command after the dog did it so it clearly can be done that way.  But I still find it extremely annoying.  I believe that dogs see cues as actions.  Sit means to lower the butt to the floor.  And that is why a lot of dogs have zero clue what you mean when suddenly the dog is lying down and then you tell the to sit.  They can't lower their butt any further, so it is like an entirely new command!  By saying Good Sit, I would imagine the dog would hear it as good....Sit rather than as a quality of the position.  Especially in the learning stages when you are naming the behavior, I always give it right BEFORE I think the dog is going to do it, not after.  Just a pet peeve of mine!

Does anyone else get what I'm trying to say about cues rewarding behaviors?
Do you guys have different levels of obedience, and do you think a different name is an easier idea than different cues?

Robin Sallie  – ( November 4, 2010 at 10:15 PM )  

I use different cues.

The "Good Sit!" is a pet peeve of mine, too.

Robin Sallie  – ( November 4, 2010 at 10:43 PM )  

A cue can become itself a good and rewarding event or a conditioned reinforcer. And, it properly timed, can be used to reward reinforce the behavior before it and cue the next behavior. Heck, that is why behavior chains are so easily (and often unintentionally) built. Using the next cue as a reinforcer creates reliable behavior. An cues can come from the environment.

Look at your example. Dog sees owner and jump up for attention. Owner says sit and rewards the sit. Owners legs are a heavily reinforced cue to jump. It is self rewarding to the dog. And the owner spent a lot of time touching the dog, looking at the dog and or talking to the dog for jumping before a trainer told the the dog needs to sit for attention.

So now the owner asks for sit as soon as the the dogs paws touches the owners legs.

Jump is rewarded with the verbal cue to sit which has been rewarded with lots of food and praise and petting for a long time in a lot of different contexts.

The owner pets, praises, feeds the dog for sitting - most of the time.
Day after day, the owner walks in. The dog jumps up and even the before the owner can, say sit, the dog does. And gets paid for it.

A behavior chain is born.

Kristen  – ( November 4, 2010 at 11:23 PM )  

The key to cues as behavior chains may be that this applies mostly to positively trained behaviors? And his behaviors are not strictly so?

Maybe?

I do agree that we have to be careful about unintentional chains. Though there's also that fine line between "training" and "using" behaviors. Students LOVE to learn a formal leave it. And it's useful. Until your dog starts seeking out things to leave. Do I still teach it? Yes. Is it our primary solution to distracting things? NO!

Amy / Layla the Malamute  – ( November 5, 2010 at 12:25 AM )  

I completely understand what you mean about cues rewarding behavior. The jumping up example was a perfect explanation (I thought) of what you were trying to ask. In trying to teach Layla to howl/talk on command, she's inadvertently picked up the behavior chain of laying down beforehand. She talks when she gets annoyed, so when she was figuring out what I wanted her to do, she must've tried laying down and then got rewarded as well. So now when I give the command, she won't do it in whatever position she was in at the time; she has to lay down. We're working on fixing that.

Crystal  – ( November 5, 2010 at 10:05 AM )  

hahahah, I'll post another tonight. I think I'm going to have increase my posting schedule for awhile...

On topic: I think Kristen might be on to something. The way I understand it, cues can be reinforcing because they've been classically conditioned to be awesome, right? This cue means I might get a treat, therefore I love responding to this cue!

On Sunday, Ian talked a lot about "losing the lure." But more than just phasing out the lure, he likes to phase out the food reward entirely much earlier than I do... He said that he starts using differential reinforcement from the second time a dog does something. Which means that, in theory, the dog only gets rewarded half the time.

It seems that as a result, there just isn't as much value built in, so the cue isn't as rewarding as it might be otherwise. When you add in the RRNR stuff later, it seems like the cue becomes more of an annoyance than an exciting opportunity.

Ninso  – ( November 5, 2010 at 11:05 AM )  

I totally get what you're saying! It's like the dog who know what "off" means when he's already on the couch, but doesn't know not to get on the couch in the first place! Why have a "slow" command rather than just teaching the dog what heel means in the first place. Heel means a particular position. If the dog is out of position, "heel" should be able to get him back into it.

Now that I think about it, I also have "commands" and "suggestions" but I can't say I always enforce commands 100% of the time or that the dogs even necessarily know what the "suggestions" mean vs. responding to body language or tone of voice.

And the whole "good sit" thing annoys me too. It's like it assumes dogs understand English syntax.

Robin Sallie  – ( November 5, 2010 at 3:25 PM )  

Yes, cues must be "classically conditioned to be awesome" to be used as rewards.

Honey the Great Dane  – ( November 16, 2010 at 4:44 PM )  

Ooh - you've done such a great job of summarising his seminars, I'm not going to bother ever getting round to posting about them - I'll just give a link to your posts!! :-)

No, seriously - I totally agree with you about your last point = that's one of my pet peeves too coz a lot of training classes here do that. The instructors are always telling people to say "Good Sit" or similar after the dog has done it adn I always think that it must really confuse the poor dog because he's thinking, "What? But I'm doing it already!" - as one trainer who was against this said to me very aptly, "Dogs don't understand grammar" - we understand "good sit" as an adjective describing the quality of ths Sit, ie. a form of a compliment - but the dog only understand each cue as an action. If you wanted to let the dog know he's doing well, I thikn you're better to just use a praise word - like Good boy or Well done. But as you say, it still somehow seems to come out OK althoguh I suspect that's more to do with dogs being flexible and forgiving than to do with good training skills!!

I also thought it was very interesting what Ian said about the different levels of performance. I do agree with it actually - I think dogs can perform better if they can 'relax' a lot of the time. And I realised that I actually unconsciously do this already with Honey in that I have different expectations of her at different times and communicate this to her clearly - but like you, I use different words rather than the same word with different names. It seems to make more sense to me to use different words, even if they sound alike. So my basic commands (eg. Down, Come) are non-negotiable but I have casual versions of those commands ('LieDown', 'Come on') which Honey understands as meaning that she has the choice to obey or can obey more slowly...I find this works very well for us.

Hsin-Yi

Post a Comment

Thank you for reading! If the link to Post a Comment is not working, click where it lists "X Comments" at the bottom of the post, right after the date field.

Thanks for reading my blog! Please Subscribe by Email!

Contact Me!

Dogs are not our whole life, but they make our lives whole.- Roger Caras

Email: lkwaudby (at) gmail.com

  © Blogger template Shush by Ourblogtemplates.com 2009

Back to TOP   

href="http://laurawaudby.blogspot.com"/blog/feed/" onclick="pageTracker._trackPageview('/feed/');"