The "Eyes" Exercise - a Bit Different from "Watch Me"

double post!

This is very very similar to what we have done, although I never would have dreamed of making the "fake eye" with my hand.  It seemed to work on a similar level though.  HOWEVER, I have always done it, not because of the predator thing, but because direct eye contact is often seen as a challenge between dogs and part of it was, I wanted to counteract that.

However, in addition to that, I want to encourage the dog, in general terms, to "check in" with me often.  I want him to keep half an eye on me so that he will be more responsive if I need to cue him to do something.

I also find it is a good "control" exercise.  It's a good way of, how shall I put it, breaking his chain of thought?  If he is about to chase a bike, well he can't chase that bike if he is busy giving me good eye contact can he?  I don't expect him to hold taht for a long period of time, but to be honest, even if it only works for one second taht can be VERY useful if your timing is good and you use that second wisely. 

I often ask for "eyes" at meal times and at "door" times and during games.   As with most training I do, I like to shape desirable behaviours by linking them with things that are desirable to the dog, which is what I think Ron does.  And I call it "watch", also as I think Ron does.  I am failing to see a GIGANTIC difference between "eyes" and "watch", but then Subtle has never been my middle name.

 

I don't see the difference yet, either. I use "look" and the dog LOCKS onto my eyes. I taught it slightly differently, without the "finger eye", but I guess I'm not that good at subtle either.  

I was never a good clicker trainer probably because I am not patient but "look" was one command I did teach with the clicker, moved into queue using the word and subsequently I have always been able to have Rivers attention on me since.  Now I just say his name and he locks eyes.  I think any exercise that teaches a dog how to look at you and make you the main attraction is a must in dog training.  I believe Ron was right and they are the same.  They are taught differently but the end result and the reasons to want them to do it are the same.

Good Blog LCK thanks for sharing.

Lee Charles Kelley

I'm just interested in some of your reactions.

It's well-written, of course. But writing is one of your specialties. There are a few things I consider to be inaccurate or contradictory to what I understand about dogs and wolves. There is still much debate, even at the genetic level of nDNA. There is not enough genetic evidence to prove that dogs descended from wolves, Wayne's work on a singular locus of mtDNA notwithstanding. In terms of structure and physiology, dogs more resemble coyotes, which does not mean they descended from coyotes. And Wayne's supposition that dogs descended from wolves is a bit foolhardy, in that he did work on the canid genome but had no way to account for the effect of nearly a century or more of specific cross-breeding with wolves. You could take any of the hybrids around here and if the momma was a wolf, the mtDNA will show wolf, as mtDNA always comes from the female. It is more likely, though generalized, to say that dogs descended from an earlier canid that was similar to canids that became coyotes. But the anthro, I mean canipology (?) does not show a direct link between dogs and wolves. So, basing dog behavior on wolf behavior is no more accurate than basing it on the behavior of other canids.

Also, even though I don't firmly believe that dogs came from wolves, I also don't think the wolf views a steadfast moose as a predator. A predator is an animal that hunts, kills, and eats the prey. I think the wolf stops and retreats because the moose has precessed into an angle that is all wrong for a good kill. Which allows the other wolves of the pack to flank and attack anew. I don't think the wolf fears the moose and the moose will certainly not eat the wolf. So, I don't see it as a turning of the tables. True, the horns are possible defensive weapons but a moose is more likely to do damage by kicking a wolf while being pursued. Wolves are considered apex predators because no animal eats a wolf, not even Man, mostly. However, there are nations that eat dogs. China wants to breed St. Bernards as a food item. So, I might have to hedge my statement a little bit. Man and wolf are both apex predators, sometimes in cooperation, sometimes not. Man has killed way more wolves than any wolves killing men.

I feel it was important to address these things as I think you presented them as the underlying theory of why the exercise works. And we can always get into another cognition debate, later.

But kudos for showing exercises that involve rewards for the desired behavior. I hesitate to call it marketing but with your stance against clickers, you are able to reach people who have a thing about clickers and still teach them some value of reward training. So, your method has some value, imo, even if it wasn't your intention.

In one aspect, we might almost agree, with different words and different angles of approach to the end product, so to speak. You view the dog as viewing you as prey because you produce the food. Food they would otherwise have to get by predatory means. I view the dog as an opportunistic scavenger and excellent at guarding resources. And they will do what it takes to get resources, whether that means hunting or looking you square in the eye. An eye lock for dogs that don't know each other is a threat. But, between humans and dogs that know each other, it might be different. So, you say eyes, I say watch me.

It also seems that you are teaching a leave it by means of moving the treat away, even if you have to restrain the dog. You are getting him to leave and look at you before getting it. A valuable skill, even if I don't agree totally with your theory as to what causes the dog to specifically do things against his nature, namely, prolonged eye contact and not going for the food item. The dog does it because he has learned the new pattern to get that treat is to leave it alone and look at you. Sorry, but methinks I smell OC. Bad Ron, bad Ron.

And there is some difference between the methods. Yours involves restraining the dog, which I'm not saying is bad. Mine involves having rewards that totally outclass whatever it was he thought he wanted and that listening to me or looking at me is always the best thing. No leash involved. In fact, when I worked on drop it and leave it, there was nothing to keep him from going for whatever, save his decision to forego that thing for what I have. Because I don't play fair. I will have drippy chicken or roast beef or hamburger and for him, that's the mother lode. It could easily be a frisbee, if he was so inclined. And I'm not saying my way is better, just different.

And bravo for presenting your theories, even in the face of skepticism or questions or opposing views.

Ron, I think the predator idea is to explain that animals see-saw between wanting to approach and wanting to retreat. The way I see it, it's a constant process of risk assessment using input from the world around them, including things they intend to kill and eat if possible, to help them. If a large prey animal suddenly stops running and turns to face its attackers, there is a fair bit of danger in proceeding for a wolf. Now the prey can see exactly where the wolf is and they are ready to defend themselves in ways that can be deadly. I think wolves (and other predators) don't CHOOSE not to attack because of the danger so much as they instincitively pull up when the balance has changed and the prey animal is at least standing still, if not approaching. I often think any two animals have this see-saw going between them. A retreating wolf might get chased by a large deer or several of them. One time I watched Penny, who is quite confident around other dogs, approaching a playing GSD and Aussie. Suddenly they turned towards her and started approaching her at a dead gallop. She turned and fled, which was probably a bad thing to do in the situation, but her instincts told her that's how it goes. If you like, using the see-saw analogy, she had been the one that kicked off the ground at a polite speed, but then they kicked off the ground with a great deal of vigour and that changed her rise into approach into a plummet of retreat. I think it's only when two animals know each other well that this see-saw starts taking a backseat role in interactions. Just my opinion.
 

I think I get what LCK is saying about the prey/predator thing, although dogs and wolves are now a world apart, dogs are still animals and they still react at a bone-deep level to most situations, training notwithstanding.  The whole pull/push, desire/apprehension thing does make sense to me.  But I don't believe it has a huge impact on eyes/watch.  The method is still similar and the outcome is still the same.  Dogs STILL, when all has been said and done and picked ober and argued about up and downa nd round and round.... they STILL "do what works".  Sometimes that can override those bone deep reactions - you can get your dog to ENJOY making eye contact wth you and do so regularly and reliably on cue, just as Ron has successfully  taught a Sibe mix a beautiful "heel" cue despite generations of "pull pull pull!!!!!!" in his bones.

Isn't it amazing what the man-dog partnership can achieve? 

Chuffy

  Sometimes that can override those bone deep reactions - you can get your dog to ENJOY making eye contact wth you and do so regularly and reliably on cue, just as Ron has successfully  taught a Sibe mix a beautiful "heel" cue despite generations of "pull pull pull!!!!!!" in his bones.

Exactly. With however we train, we are able to have dogs do what is against their natural urge. Although I think LCK is trying to use the natural urge to achieve training success, hence natural dog training. But the outcome appears the same. In his theory, the dog approaches you because you are prey. In mine, the dog approaches you because it has been proven to be worth his while. "Hi, I'm Shadow. I'm in Acquisitions, mainly for myself. Here's my portfolio."

Now, the interesting thing I have noticed is that dogs do not necessarily transfer what they know of humans to other dogs. Your dog can lock eyes with you but will not necessarily do so with another dog, even though the human has rewarded the staring behavior. Because with that other dog, it is still a challenge to stare. I know we're skating close to another round of cognition but isn't it interesting that the dog differentiates, without any further cue from us, that it's okay to stare at a human for a treat but not stare at another dog?

Dogs do use eyes to communicate, I think. Like the video I have of Shadow poking around the yard and he spots Jade in the window and alerts me to her presence. It's subtle, fast, and unmistakable.

http://s216.photobucket.com/albums/cc122/ronws_photos/?action=view¤t=mov2.flv

For some reason, the video is shorter than it was originally but near the end of this clip, you can see his alert.

While editing this message to link the video, I noticed something I said which has another implication. Dogs don't always transfer what they know about humans to how they treat other dogs. In which case, does having them view us as prey or an obstacle to overcome have an relevance or bearing on their interactions with other dogs? Or is another process at work, despite whatever we think?

 

 

ron2

I don't don't think the wolf views a steadfast moose as a predator.

I'm sorry if I seem to have given that impression. I think it would be better to put in the reverse terms, which is that the wolf feels that he's no longer the predator, but is now prey. But again, you're right in that the moose isn't going to kill and eat the wolf. However, the wolf doesn't know that. All he knows (or I should say "feels";) is that his energy, which was flowing so pleasurably toward an objective -- involving a deep-rooted urge to bite what's moving away from him -- has now been reversed. The moose is no longer an attractor but a resistor. That's all I'm saying. And to my way of seeing things it's the moose's height and the level of his gaze that are the key ingredients in the wolf's sudden reversal from attraction to resistance. And it doesn't matter in the slightest if you care to replace the word wolf with coyote, or proto-dog, or pre-doglike-canid, or whatever. The energetic essence of the situation even works with cat and mice (and apparently with tigers and rubber plantation workers wearing masks on the backs of their heads!).

ron2

But kudos for showing exercises that involve rewards for the desired behavior. I hesitate to call it marketing but with your stance against clickers, you are able to reach people who have a thing about clickers and still teach them some value of reward training. So, your method has some value, imo, even if it wasn't your intention.

Cute.

Just to be clear, I have nothing against using rewards in training. I think you'd have to be either a fool or an idiot not to. And I have nothing against the basic bedrock stuff that Skinner and others came up with back in the early part of the last century. That stuff is pretty undeniably accurate most of the time. It's just that I see a dog's behavior, whether learned or instinctive, as being about reducing his internal stress, not about gaining an external, extrinsic reward, which is the way Skinner has come to be grossly misinterpreted in recent years.

Also, if you ask me, the Skinnerian model, no matter how far removed it may seem to be from one involving any sort of mental thought processes, still has that bit of mental sand in the machine. A trainer I butt heads with once in a while here in Manhattan is fond of telling people to let the dog know that when he behaves properly he gets access to the "good stuff," and that when he does something he wrong, "the good stuff disappears." On a certain level I sort of agree with her. But her phraseology belies a belief that there's a mental thought process going on. I think it's better to think of the dog as an energy system, one that naturally seeks a kind of emotional homeostasis, than a thinking machine that knows what "good stuff" is, and can learn how to manipulate the environment to make it "appear" and "disappear." Yes, on a certain level that appears to be what's going on, but I don't think it is. In a very real sense dogs don't manipulate the environment, nor does the environment control them. The dog and the environment are all part of the same energy system. And within that system the dog is simply attracted to things, behaviors, events, people, smells, etc., that reduce his own internal stress. And whether his attraction is a result of instinct or learning, or both, is irrelevant. It's all about attraction and resistance, not about making mental associations.

The objection I have to the clicker is expanded upon somewhat by Bob Bailey, in the same interview I pulled my signature quote from. He said that he didn't think clickers were a good idea for training pet dogs, that they can get in the way of real learning. Obviously you believe they work for you, and I would have no way of knowing personally if that's true or not. The only thing I can say in that regard is that in most dog training situations, roughly 85% of the credit goes to the dog and only 15% to the trainer or training method. Most dogs are born already knowing how to obey most of the commands we give them. The trick is making the cues relevant to their emotions at any given moment, particularly in critical situations.

But this is way off topic...

ron2

It also seems that you are teaching a leave it by means of moving the treat away, even if you have to restrain the dog. You are getting him to leave and look at you before getting it. A valuable skill, even if I don't agree totally with your theory as to what causes the dog to specifically do things against his nature, namely, prolonged eye contact and not going for the food item..

We disagree here, because I think the way I was taught the "eyes" as opposed to the way I've seen the "watch me" described (which is apparently done by free shaping) is that the NTD way of doing it doesn't go against a dog's nature. In fact it works with the dog's nature. That's the whole point.

It seemed strange to me when I was first told about making the make-believe "eye" with my thumb and forefinger and the nonsense about the tigers in Malaysia (which even though it was reported by the BBC, I'm not even sure it's 100% true). But the thing is when you actually do it that way there's a subtle, yet substantial difference in how the dog responds. Almost all dogs will initially go after the treat as a matter of course, whether you're using the stupid make-believe eye or not. They'll jump up, they'll try to grab it, they'll come around from the other side. But at some point, whenever that "eye" is part of the process, they just seem to stop going for it much quicker. They'll just sit and stare. I won't say they're mesmerized exactly, but they do just sit and stare at it. None of the dogs I've worked with have exhibited that same kind of response as much or as often when that "eye" isn't there. In many cases they've never exhibited it at all. I can't explain why it this is so, but I've seen it happen over and over.

ron2

Yours involves restraining the dog...

Not really. The leash is just mentioned so that the novice will have some ability to control the dog if necessary. Nowhere in my description do I mention or suggest using the leash and collar to restrain the dog or prevent him from getting to the treat. (I may have to rewrite the article to make that clear.) In fact I never use a leash myself. Or do so only rarely, and then not for restraint, per se, but just to give the dog a feeling of being under control (funny how simply having a leash on does that for most dogs).

And I think everyone here still seems to be missing the fundamental difference between this and "watch me," which is that this is NOT done by free shaping. It's done by deliberately frustrating the dog's desire for the treat, not by having it appear out of nowhere the instant the dog looks at you. The simple, undeniable fact is the more you frustrate him, without getting to the point that he shuts down or loses interest, the bigger the payoff when he finally figures out how to get what he wants. In the oc model the dog doesn't particularly want anything, he's just given this sudden "jackpot," for doing nothing more than looking at you. That's nice, and you can build on it from there. But I think my way of doing it goes much deeper because again, it's not so much about giving him an external reward for looking me in the eyes, but showing him that looking me in the eyes is a way of reducing his own internal stress. Stress reduction is probably a nifty, accidental side-effect of the "watch me" exercise. But here it's built into the way the exercise is done.

Anyway, that's how I see it.

LCK 

A trainer I butt heads with once in a while here in Manhattan is fond of telling people to let the dog know that when he behaves properly he gets access to the "good stuff," and that when he does something he wrong, "the good stuff disappears." On a certain level I sort of agree with her. But her phraseology belies a belief that there's a mental thought process going on.

Her phraseology is perfectly wonderful for one thing - getting her students to do what she is asking them to do.  And, frankly, if their dogs then learn to sit, come, or whatever the object is at the moment, then that's great.  The public doesn't usually give a rat's butt about our debates on Skinner or the cognitive capabilities of their dogs.  They just want the dog to stay or stop jumping up.  And, to be honest, explaining your current posit as "not the same as "watch me" implies that there is only one exercise that trainers call "watch me", which is not exactly correct.  Some trainers do what you do, others stick a treat in their hand at arms length and just wait for the dog to make eye contact instead of watching the treat.  The phrase is unimportant when you talk about behavior, it's the behavior that requires description.  I doubt that my dog feels that her internal stress is lowered by looking me in the eye - gee, wouldn't she have to have some advanced cognitive function, that you feel she lacks, in order to think of it that way?
Sometimes, there is no hidden meaning in all this, save that the dogs think our direct eye contact is pretty impolite - perhaps because we give them a lot of "good stuff", they forgive us our poor canine manners.

 

I don't get the difference either. It's basically teaching your dog to make eye contact. Who cares what command you use - eyes, watch, look - it's all eye contact. BTW, freeshaping isn't the only way to teach watch, it's just one technique of many. Personally, I like to teach my dogs to make eye contact on cue, as well as to offer it up frequently (without having to ask for it) as a default behavior. I started teaching watch by holding out a treat pinched between my fingers, (such as you describe), right in front of their face, and then marking and releasing the treat the second they looked away from the food and up at me. I also 'captured' eye contact by marking with a clicker and tossing a treat whenever they voluntarily checked in with me. I want them to choose to make eye contact with me and actively seek it, not just obey a command when given.

By working on both - teaching watch on cue, and marking and rewarding spontaneous eye contact - from a very young age, my dogs seek eye contact with me often, and view it as a pleasurable experience. They also don't mind eye contact with strangers, which some dogs can interpret as a challenge, and find threatening.

Just wanted to add that I also do almost all of my training at home off leash. I prefer not to rely on a collar and leash to keep the dog interested and engaged in training.
 

Lee Charles Kelley

roughly 85% of the credit goes to the dog and only 15% to the trainer or training method.

I could agree to these percentages and might phrase it as most of the struggle is for the human to learn to communicate, as the dog already can communicate. Or, to paraphrase, the training is mostly for the people.

Lee Charles Kelley

He said that he didn't think clickers were a good idea for training pet dogs, that they can get in the way of real learning

Save a few stark exceptions to the rule, just to include all statistical possibilities, I don't see how clickers get in the way of training. For me, and many others, they have become the primary way to train, often getting results in one session, 5 to 10 minutes long. For some important behaviors, it might take the same length of time in various circumstances or scenarios, bringing a grand total of an hour, if one totalled all the different scenes in which one might have had to approach generalization of the behavior. But I'm not going to debate that. And it would detract from my cute observation.

Lee Charles Kelley

And I think everyone here still seems to be missing the fundamental difference between this and "watch me," which is that this is NOT done by free shaping. It's done by deliberately frustrating the dog's desire for the treat, not by having it appear out of nowhere the instant the dog looks at you. The simple, undeniable fact is the more you frustrate him, without getting to the point that he shuts down or loses interest, the bigger the payoff when he finally figures out how to get what he wants

One of the points I am making is that what you are doing is OC, whether it seems that way to you, or not. But, specifically, I think in OC, I am removing the frustration that can lead to alternate undesirable behavior by exactly showing the way to get what the dog wants. That is, there is no frustration or confusion. If there is, I have done something wrong. And some dogs, like mine, who started out with lure/reward and corrections, will shut down by means of laying in a down and doing nothing. He will quit playing your game.

Lee Charles Kelley

But I think my way of doing it goes much deeper because again, it's not so much about giving him an external reward for looking me in the eyes, but showing him that looking me in the eyes is a way of reducing his own internal stress. Stress reduction is probably a nifty, accidental side-effect of the "watch me" exercise

Is not stress reduction a reward, part of the dog's desire for personal equilibrium, ala the non-linear dog theory? The dog is stressed because he wants the food. When he gets the food, stress is reduced. Just as working for a reward, ala DPU, might involve some induced stress. The dog was just fine until we started waving food in the air, so to speak. It seems to me that the reception of the reward, whether it is food, catching a frisbee, whatever, may result in some temporary stress relief and is therefore rewarding. Or gaining rewards can result in stress reduction.

And looking in your eyes as a way to reduce stress, by means of using the treat reward means you are relying on the work of Skinner and Pavlov. That is, one hopes the dog looks at you to resolve stress, once the behavior connection has been made, ala the ringing of the dinner bell bringing on salivation. It may or may not mean cognition at this point but it certainly is returning full circle to behaviorism.

Check and possibly check-mate.

 

ron2

One of the points I am making is that what you are doing is OC, whether it seems that way to you, or not.

No, technically it's Natural Dog Training, which uses some principles of operant conditioning. I already alluded to this in my previous post, to wit:

"I have nothing against using rewards in training. I think you'd have to be either a fool or an idiot not to. And I have nothing against the basic bedrock stuff that Skinner and others came up with back in the early part of the last century. That stuff is pretty undeniably accurate most of the time. It's just that I see a dog's behavior, whether learned or instinctive, as being about reducing his internal stress, not about gaining an external, extrinsic reward, which is the way Skinner has come to be grossly misinterpreted in recent years."

ron2

I think in OC, I am removing the frustration that can lead to alternate undesirable behavior by exactly showing the way to get what the dog wants. That is, there is no frustration or confusion. If there is, I have done something wrong. And some dogs, like mine, who started out with lure/reward and corrections, will shut down by means of laying in a down and doing nothing. He will quit playing your game.

Yes, Ron. I already covered that too: "The simple, undeniable fact is the more you frustrate him, without getting to the point that he shuts down or loses interest, the bigger the payoff when he finally figures out how to get what he wants." By setting up a training situation where you avoid frustrating the dog you also end up avoiding a deeper, longer-lasting form of learning, that's all.

Lee Charles Kelley

But I think my way of doing it goes much deeper because again, it's not so much about giving him an external reward for looking me in the eyes, but showing him that looking me in the eyes is a way of reducing his own internal stress. Stress reduction is probably a nifty, accidental side-effect of the "watch me" exercise

ron2

Is not stress reduction a reward?

Maybe you need a nap (I know I do). Look again at what I said: "it's not so much about giving him an external reward," as in he gets a treat, or "the good stuff" as SpiritDogs would apparently put it, when he looks you in the eyes. It's that feeling I described in my article, where the dog has a dilemma, a problem to solve. And when he looks to you for the solution and gets the treat at the same time: "You’ve not only given him the answer to his problem, as far as he’s concerned you are the answer." That's pretty strong stuff. A lot stronger than, "Oh, this is nice! I got a treat for something, I'm not sure what..."

ron2

The dog is stressed because he wants the food.

Yes. But also because his normal behavioral attempts at getting it are being frustrated.

ron2

When he gets the food, stress is reduced.

Yes, but you've left out the most important part, which I re-described just above. In other words we're not just rewarding him for just looking us in the eyes, but for looking to us for a solution to his dilemma. Doesn't that at all strike you as being something of a substantial difference?

The deeper the tension and stress, the more fulfilling the reward, the more fulfilling the reward, the stronger the learning. Correct? So by setting up the "watch me" exercise without building any stress or frustration into the equation, you're limiting its effectiveness to some degree. I don't see how you can get around that fact.

ron2

The dog was just fine until we started waving food in the air, so to speak. It seems to me that the reception of the reward, whether it is food, catching a frisbee, whatever, may result in some temporary stress relief and is therefore rewarding. Or gaining rewards can result in stress reduction.

Yes, but external rewards don't always reduce stress; in fact they sometimes increase it. In oc terms, food can sometimes, perhaps even often, be a punisher; there is research showing that extrinsic reinforcers actually undermine the learning process entirely. In this case (the "eyes" exercise) food is intrinsic to the way the it's performed because for the dog the real reward is the feeling of solving the problem, not the treat, per se. You could set it up with a tennis ball or the gate at the dog run and get a similar result (except you wouldn't have that strange make-believe "eye" for the dog to look at). In free shaping whether the treat is an intrinsic or extrinsic reinforcer would very well depend on how the dog is feeling prior to the moment he's "rewarded" with the treat, and what kind of treat is used, and how much he might be in the mood for that type of treat instead of another, etc. Here there's no question that the food is intrinsically part of the dog's reward, because it means he's found the solution to his problem. But that's the real reward: finding the solution.

ron2

And looking in your eyes as a way to reduce stress, by means of using the treat reward means you are relying on the work of Skinner and Pavlov. That is, one hopes the dog looks at you to resolve stress, once the behavior connection has been made, ala the ringing of the dinner bell bringing on salivation. It may or may not mean cognition at this point but it certainly is returning full circle to behaviorism.

Check and possibly check-mate.

 

I didn't know we were playing chess. And I already made my point about the correlation between the bedrock principles of Skinnerism and what I do. What I find troubling in current oc dog training circles is the way it's become conflated beyond recognition or repair.

LCK 

Well, if clickers get in the way of training, tell that to Sequoyah who learned to circle around me as I walk within about a two minute session.  She's a clicker savvy dog and all it took was an ex pen, some elementary targeting, and a verbal cue added as soon as she did the behavior. 

I think that it might be the use of the word "reward" that is a problem in the context of this conversation.  In reality, ron is quite correct that OC is taking place.  If we substitute the word "reinforcer" for reward, we are more accurately describing what takes place.  A reinforcer is ANYTHING THE DOG WANTS, and does not necessarily have to be something we think of as positive, such as roast beef or a tennis ball.  The reduction of stress is a reinforcer for many dogs.  But, again, that isn't what I want my dog working for.  I don't like the idea of purposely stressing a dog to get them to learn by figuring out how to de-stress if I can adopt a more positive framework that works equally as well and the dog remains as happy as I.