Two training questions

[Root Beer]

Oh. I understood you to be saying that the person was describing their dog as doing something out of a desire to please, when in fact the dog was being forced to do it. From what you say here, I guess you're just saying that some people want a dog to do things out of a desire to please them, and that people who feel that way sometimes use force in training. Is that it? I dare say that's true.

 

Not quite. What I am saying is that those people consider what the dog does as a result of making the dog do what they want to be the dog doing those things out of a desire to please.

 

Some other examples:

 

Dog sits on the table because owner pushes on the dog's rump = dog sits out of desire to please.

 

Dog stops on a contact because the handler is holding the dog there by the collar = dog holds contact position out of desire to please.

 

Dog goes through tunnel because he has been pushed though = dog goes through tunnel out of desire to please.

 

"Force" isn't so much the issue, although all of those things certainly can be done forcefully. But these things can be very neutral as far as that goes for many dogs. So, I'm not really talking about force here.

 

What I am saying is that really, in those examples, and instances like them, the dog is really not doing anything to please the handler. Desire to please is not a motivation at all in those cases. The dog is simply doing what he or she is being physically led to do.

 

As I said above, whether or not such an approach is good Agility training practice is debatable. But really, the dogs in these cases are not working to please their handlers and to say that they are (as some people actually do) is misleading. And while you would not use the term in that way, in sport contexts there are people who do.

[Root Beer]

Well, I assume you're at least educating your students about the difference. I honestly find it hard to believe that most people think that forcing a dog to do something somehow equates with the dog working out of a desire to please.

 

I make the best effort I can. Most people are open to a different approach, but some are not. You know the old "lead a horse to water" saying.

 

I do want to clarify one thing, though. Most people do not think that making a dog do something equates desire to please. Out of a class of, say, eight people, there are typically one or two who come in with this kind of mindset. And of those, many of them are not particularly committed to thinking along those lines and are willing to try out a different approach. Most people actually do know that some kind of reinforcement is needed, at least to some extent. They may need to learn how to use it better, but most know that it is to be used.

 

However, while most people do not think this way, it is common enough that it sends up a red flag when I hear it. I actually have yet to meet a pet or sport dog IRL whose owner is unwilling to use treats or some other reinforcer in the early stages of learning who has actually demonstrated any willingness to learn for the sake of "desire to please" - 99.9% of the time those dogs are being dragged around. I am not saying it can't be done or doesn't exist. You and Eileen have given examples. I've given examples. Airbear gave an example. But I don't see it at all in training class situations.

[Eileen Stein]

Not quite. What I am saying is that those people consider what the dog does as a result of making the dog do what they want to be the dog doing those things out of a desire to please.

 

Some other examples:

 

Dog sits on the table because owner pushes on the dog's rump = dog sits out of desire to please.

 

Dog stops on a contact because the handler is holding the dog there by the collar = dog walks near handler out of desire to please.

 

Dog goes through tunnel because he has been pushed though = dog goes through tunnel out of desire to please.

 

"Force" isn't so much the issue, although all of those things certainly can be done forcefully. But these things can be very neutral as far as that goes for many dogs. So, I'm not really talking about force here.

 

What I am saying is that really, in those examples, and instances like them, the dog is really not doing anything to please the handler. Desire to please is not a motivation at all in those cases. The dog is simply doing what he or she is being physically led to do.

 

That's what I thought you were saying in the first place, when I doubted anyone could be so totally out of touch with reality as to say/think what you said they were saying/thinking.

 

Hmm. Maybe this will clarify it for me:

 

Are you saying that, when the owner pushes the dog's rump down into a sitting position, the owner believes that the dog at that moment is sitting out of a desire to please? Or are you saying that, when a dog whose owner has pushed his rump down into a sitting position to teach him to sit on the table later sits on the table spontaneously without having his rump pushed down, the owner believes that the dog is then sitting out of a desire to please?

 

Likewise, are you saying that when a dog is being pushed through the tunnel, the owner believes that the dog's transit through the tunnel is a result of the dog's desire to please? Or are you saying that, when a dog who has learned what he's expected to do upon hearing the command "Tunnel" by being pushed through the tunnel later goes through the tunnel on his own in response to the command, the owner believes those later trips through the tunnel on command and without pushing are a result of the dog's desire to please?

 

It seems to me that in each of these examples, the belief expressed in the second sentence may or may not be true, but is not particularly remarkable (or deplorable). The belief expressed in the first sentence, however, is a sign of mental disease or defect, and I have to think it must occur very, very rarely.

[Eileen Stein]

However, while most people do not think this way, it is common enough that it sends up a red flag when I hear it. I actually have yet to meet a pet or sport dog IRL whose owner is unwilling to use treats or some other reinforcer in the early stages of learning who has actually demonstrated any willingness to learn for the sake of "desire to please" - 99.9% of the time those dogs are being dragged around. I am not saying it can't be done or doesn't exist. You and Eileen have given examples. I've given examples. Airbear gave an example. But I don't see it at all in training class situations.

 

This may be a quibble, but in the example I gave of my first border collie learning the tunnel I would not characterize that as a "desire to please," exactly, although I suppose I wouldn't argue if someone wanted to broadly describe it that way. To me it's a desire to work with her person, to learn new things and perform them, to figure out what her person wanted and do it. She delighted in it, for its own sake, although perhaps she would not have enjoyed it so much if not for the delight she could tell I felt in response.

[Root Beer]

Hmm. Maybe this will clarify it for me:

 

Are you saying that, when the owner pushes the dog's rump down into a sitting position, the owner believes that the dog at that moment is sitting out of a desire to please? Or are you saying that, when a dog whose owner has pushed his rump down into a sitting position to teach him to sit on the table later sits on the table spontaneously without having his rump pushed down, the owner believes that the dog is then sitting out of a desire to please?

 

I would actually say that the owner would consider both instances to be some manifestation of "desire to please".

 

Has the person really thought it through and looked at it from this perspective? I can't say for sure, but my guess is probably not. Somehow, no reinforcer, dog does what is desired (regardless of how that happened) = dog working to please.

 

The belief expressed in the first sentence, however, is a sign of mental disease or defect, and I have to think it must occur very, very rarely.

 

I would disagree with that. I tend to think that many of the people who hold this point of view really do see "the handler is pleased" and the dog acting out of a "desire to please" as one in the same thing. Few of the people that I have worked with who do not want to use reinforcers because the dog should learn everything out of a "desire to please" have been mentally deficient. There are reasons why they hold that mindset - even if the conclusion being drawn is not exactly logical.

 

Whether it is remarkable or deplorable isn't really the issue from where I stand. I think there is a lot to be learned from our dogs and if a particular dog really is not going to learn out of a "desire to please" and needs reinforcement/motivation to learn and develop an authentic desire to please, the handler coming to recognize that fact will open the door to more effective training and a better training relationship between dog and handler.

[Eileen Stein]

I would disagree with that. I tend to think that many of the people who hold this point of view really do see "the handler is pleased" and the dog acting out of a "desire to please" as one in the same thing. Few of the people that I have worked with who do not want to use reinforcers because the dog should learn everything out of a "desire to please" have been mentally deficient. There are reasons why they hold that mindset - even if the conclusion being drawn is not exactly logical.

 

Kristine, I'm not talking about a mindset toward training without reinforcers being a sign of mental deficiency. Obviously I don't think that. I'm saying that someone's perceiving a dog being held in a sitting position as indicating a desire to please is a sign of a mental defect. It's almost like a rapist perceiving a bound and gagged kidnap victim as being there out of love for him. Not that placing a dog in a sitting position is anything like a crime of violence, but the bizarre inability to perceive reality is similar. And hopefully very uncommon.

[Root Beer]

Kristine, I'm not talking about a mindset toward training without reinforcers being a sign of mental deficiency. Obviously I don't think that. I'm saying that someone's perceiving a dog being held in a sitting position as indicating a desire to please is a sign of a mental defect. It's almost like a rapist perceiving a bound and gagged kidnap victim as being there out of love for him. Not that placing a dog in a sitting position is anything like a crime of violence, but the bizarre inability to perceive reality is similar. And hopefully very uncommon.

 

One thing I've found is that when it comes to their own dogs, some owners don't see these things all that clearly.

 

I was working with someone just a few days ago who did not want to use a clicker. The reason? She kept insisting that her dog does not like the clicker. What both my assistant and I saw was a dog whose eyes were alight, her Aussie rump was wagging a mile a minute, and she was picking up on the exercises in record time in just a few clicks and she very obviously wanted more, more, more! This dog clearly loved the clicker.

 

But somehow the owner saw a dog who did not like the clicker. She was sure if it. The clicker could not possibly work because the dog doesn't like it. It was very plain that it was not the dog who did not like the clicker - it was the person. But she really didn't see that - she saw her dog disliking it.

 

By and large I am someone who believes that the dog's owner knows the dog best. But sometimes these things happen.

 

And when it comes to seeing dragging a dog through the tunnel as the dog having a desire to please, I think much the same thing happens. The handler doesn't want the dog to do it "for a treat" and any means of getting the dog through "without a treat" becomes "desire to please" to that person's way of thinking. And when that is the case, I think the handler is better off coming to an understanding of what is really going on with the dog. He is being dragged through a tunnel - not wanting to please.

 

Again, I am not saying that there are not cases when desire to please doesn't come into play. I would say that can actually be manifested through training with treats and all of the available reinforcers. But the way the term is used most typically is misleading.

[Eileen Stein]

But somehow the owner saw a dog who did not like the clicker. She was sure if it. The clicker could not possibly work because the dog doesn't like it. It was very plain that it was not the dog who did not like the clicker - it was the person. But she really didn't see that - she saw her dog disliking it.

 

But there's quite a difference between not reading a dog's emotions accurately, and not seeing that a dog who's being dragged is not moving voluntarily.

 

I don't want to belabor this, and this will be my last post on this particular point, but is there any chance you are misinterpreting the owner's statements that s/he's training in a way that will utilize the dog's desire to please, as a perception on the owner's part that the dog's being dragged through the tunnel reflects the dog's desire to please? I only ask because you don't seem to see much difference between these two things. Have you ever asked a person dragging a dog through a tunnel if the dog is going through the tunnel *right now* because of a desire to please? If s/he replies, "No, but once I get across to the dog that this is what I want he'll do it out of a desire to please," you have a very different situation from the one you have if s/he replies, "Yes." You may consider both to be problematic because the owner is not recognizing the need for reinforcers in training, but in fact they are two very different situations.

[airbear]

Now let's suppose Rex could not care less about "Yay". Suppose he had looked at it and you said "Yay" and he turned and walked away from the tunnel, wanting nothing more to do with it.How would you then tap into desire to please? It would be a lot tougher. My guess from your description is that you would use another reinforcer if that desire to please were not really there.

Well, yes, I would rethink my game plan, including, but not limited to, the reinforcement piece. Why did he walk away? Was there something more interesting (maybe one of the other dogs was distracting him)? Is he exhibiting fear of the tunnel, since he has a history of interpreting my "yes!" as "you're on the right track, keep showing me stuff". I wouldn't dismiss immediately the reinforcer without analyzing the whole picture. I think training is more complex than that. If the only thing that keeps my dog engaged with me is the relative strength of the reinforcer, then we have a problem with our working relationship.

 

You have an established reinforcement history with several different reinforcers, including the "Yay".

What seems to have been glossed over here is Rex's background as a sheep dog. He has, since a very early age, received only a softly murmured "good dog" or perhaps a brief pat on the side for his work (this is aside from the intrinsic reward of working sheep, of course). In his sheep dog training, he has learned how to fail and to keep trying, how to be wrong and to move on. In my experience, learning to be a stock dog is a LOT more stressful mentally than learning to do agility/flyball/freestyle/rally. Before he turned 2, he had won 3 nursery trials, involving courses which took upwards of 8 minutes a piece to complete. Do you feel that this early life experience is contributing to his ability to work for praise, to thrive on praise (which I do interpret as a desire to please)?

 

When I think about it, I don't see that so much as "desire to please" but as a good solid training relationship on both of your parts. Desire to please is part of that and really that goes both ways.

WHY is our training relationship solid? Is it just because we've done a lot of stuff together, and he thinks he's going to get a treat? Or is it because I accept and respect who he is as a dog, I have set expectations of him, and I try to be fair when rewarding or sanctioning behaviour? It kinda reminds me of a comment that an agility friend made when I loosely tied Rex's leash to my chair and went out to the car to get something (the only reason I tied him to the chair was to meet the "no loose dogs" requirement of the seminar). She said "only a herding dog would stay put tied to a flimsy chair". I thought this was quite sad, that she felt that there was something remarkable about a dog behaving decently when tied out, that somehow it must be part of that scary "herding training".

 

One thing that I think is important to consider here is that in order for a dog to work for praise, praise must be a reinforcer/motivator.

Well, yes, I would think that is sort of the point of my post. Rex has only mild interest in balls. If I took a tennis ball out and tried to use that as a reward in the intro to tunnel scenario, I would be setting us up for failure. I have a dog that will work for praise, because he has learned that it feels good when he is praised for doing something. It is a very strong motivator for him.

 

Just as some dogs will spit out food when they are learning, some of them, figuratively, "spit out praise". It means nothing in that context, even if they love it at home (It happens with roast beef and it happens with praise!). They don't care about it. In that case, it is not a reinforcer and it won't motivate. And really, I wouldn't say that a dog like that does not want to please the handler in a larger sense. It really just says that the dog is not motivated by praise.

 

Now most agility instructors that I know insist that you can and should teach your dog to work for a toy. Why doesn't anyone suggest that you can teach working for praise? When I first threw a water bottle for Wick, she ducked. But then she learned that she could retrieve it, strip the label off of it, crunch it. Now she will work very happily for a water bottle. So in the beginning, no, it wasn't a reinforcer, it didn't motivate, so I showed her how much fun it could be to work for a water bottle. Why do you think that a dog can't learn to take immense pleasure not from a bit of food or a game of tug, but from a sincere "good dog" and a pat from his owner?

[topnotchdog]

The only difference I can see is that this time he saw that I was stuck, and needed help, and he was "motivated" to help me. I wouldn't say he did it because he "loved" me in any emotional sense -- not at all. (Nor did he do it for the sheer pleasure of working with me, the way my first border collie learned things other than sheep work.) But I would say he did it for a motivation that is different in kind from a liver treat.

 

I know that in behavioral theory something like this doesn't happen. But I was there, and it did happen. Spot was the best farm dog I ever had, right up to the week he died -- he learned to read my mind to a great extent -- and he never would work for anyone else.

 

 

It seems to me we humans have barely scratched the surface of understanding learning and behavior. Sure, there is merit to learning theory/behaviorism. But there are also other ways to understand behavior like observational learning, model-rival method, gene-environment interactions, temperament, social pressure or lack thereof, physical state/wellness and others, including those we just don’t have words for yet. I think we humans can get in our own way in trying to understand things; we only know so many ways to ask questions, so we only understand a certain slice of what animals do and why. In my view, our own opinions, personalities, experiences, rules (eg academia), desires, social alliances and traditions help us see some things, but also prevent us from seeing others.

 

I think Eileen’s story about Spot is an example of just because there isn’t a way (yet) to explain what happened with him that day within an established framework doesn’t mean it wasn’t learning, understanding, and partnership taking place. And (not that Eileen was implying this, it’s where my mind goes next) just because other sheepdog work hasn’t been described in learning theory terms before (for eg), does not mean a traditional method or training situation is working outside the laws of, say, operant conditioning. These divisions seem to come up often between more family dog training or sports-minded folks on the list and those more working sheepdog inclined. I guess it might be a little ground-shaking if things turned out to be not quite as each had imagined, or not quite all the time. Some people might see that as interesting, while others would possibly be unnerved by that or not see the point in fiddling with what works for them. While I have core values that guide my exploration, and I have more confidence in some methods than others, the day I sort it all out I guess I’ll need to take up stamp collecting or some other new challenge, but I doubt that will happen. Not just because I can learn from the people, but because I continually learn from the dogs.

 

On another note, I can’t help wonder if the notion of a dog doing something out of a “desire to please” is one side of a conceptual coin that has as its flip side the dog refraining from doing things out of “fear of reprimand.” While I can easily think of family dog or sports training scenarios in which that comes up, part of my more recent dog education has been in trying to understand it in the world of working dogs. For example, it has been mentioned here that sheepdogs are generally well behaved in public off leash and settle down easily even in a new or very distracting situation. This sometimes goes along with phrases like “the dog knows what’s expected of him,” as though the dog senses what is needed.

 

From what I have seen this kind of explanation does not do justice to the amount of work put into breeding the dog, socializing the dog, and correcting the dog repeatedly and consistently for undesirable behavior, and letting him go about his business or speaking in a nicer voice for desirable behavior. It also gives a bit of a mythological air about the working sheepdog’s behavior, which strikes me as not quite on the money, either. If a dog bred and trained to be alert to verbal reprimands or corrections gets in the habit of being quick to “behave less” in a given situation after a look or a word from the person whose every nuance he has learned to understand, then perhaps those factors play a large part in that dog behaving in a well-mannered way. (Maybe “knows what is expected of him” is shorthand for that, I don’t know.)

 

So, to get back to the coin, the dog may well do things out of desire to please his human partner. But might the dog not also be keeping himself in check or avoiding doing things that will feel unpleasant to him (could be physically, but really I mean even mentally)?

 

Whether or not the human or dog in question really minds this is another matter. As is whether there is problematic fallout or not, which in my experience varies according not just to the dog and the person, but also the context. A dog or a person can put up with a lot of strong feedback when they are determined to stay involved in the activity or behavior in question. And reward based training can be effective and not coddling, cumbersome, time-consuming, overly complex and the like. But that’s a different thread. Maybe all this boils down to is the old adage of “the only thing two dog trainers/handlers can agree on is that the third is wrong.”

 

As to the kibble v.s. a treat humans perceive as being better, why not just listen to what the dog has to say about it? And everything else being equal (i.e. the person is listening to the dog and doing no harm), if it makes a person feel good to feed a jackpot, teach the lie down on or off stock, use a clicker instead of a bridge word, etc., my feeling is, a person should do what they prefer (why shouldn’t they enjoy the training as well?). Dogs are constantly figuring out what we want, sometimes in spite of what we do and not because of it.

 

B.

[bcnewe2]

Now let's suppose Rex could not care less about "Yay". Suppose he had looked at it and you said "Yay" and he turned and walked away from the tunnel, wanting nothing more to do with it.

 

And that would be my Mick. hence the reason we don't do agility!

 

As to the kibble v.s. a treat humans perceive as being better, why not just listen to what the dog has to say about it?

 

Again...that is why Mick doesn't do agility. But put him on sheep and he's motivated till the cows or sheep come home.

you have to work with what you have!

 

Really, I tried agility with Dew, she did it cause I said let's try it and she did it cause she wanted to spend time with me. Those are her motivators. Treats made things easy to "show" her how to do, but it was her desire to please me and work with me that made her do them.

 

I thought agility was supposed to be fun. Yes it takes a few times learning how to do particular thing so there might be some bribing to "try" something, but after that I was under the impression it was a team sport that both human and dog enjoyed. I didn't think it's cause they liked treats!

 

Whatever works for you and the dog, no black and white rules of how you or the dog are supposed to enjoy the sport...or I guess I could be totally off base!

[juliepoudrier]

Great post, Barbara!

 

From what I have seen this kind of explanation does not do justice to the amount of work put into breeding the dog, socializing the dog, and correcting the dog repeatedly and consistently for undesirable behavior, and letting him go about his business or speaking in a nicer voice for desirable behavior. It also gives a bit of a mythological air about the working sheepdog’s behavior, which strikes me as not quite on the money, either. If a dog bred and trained to be alert to verbal reprimands or corrections gets in the habit of being quick to “behave less” in a given situation after a look or a word from the person whose every nuance he has learned to understand, then perhaps those factors play a large part in that dog behaving in a well-mannered way. (Maybe “knows what is expected of him” is shorthand for that, I don’t know.)

I think what you describe is exactly what is meant by "knows what is expected of him." I doubt anyone uses that term, at least from a stockdog perspective, to mean that the dog springs from the womb knowing what's expected of him (the mythological aspect), but rather that through the training process (on and off stock) the dog learns what is expected of him, such that it doesn't need to be constantly reminded or reinforced--hence a single look might be all that's needed to stop an unwanted behavior. On the flip side, how many of us tend to apologize for our dogs' leash manners (which may be nonexistent) because it's not something that's important in the normal context of their lives--walking at heel on a leash is not generally expected of them, and so they don't necessarily do it well.

 

It shouldn't be a big mystery to folks who train dogs using any method (that is, this isn't the sole provenence of stockdog trainers). Once your dog becomes your partner and learns your expectations in various situations, then it knows what's expected of it, right?

 

The key, I think, to this discussion, though, is that stockdog trainers don't necessarily feel a need to use high-value rewards to get their expectations across to their dogs (on or off stock). This is the point Kristi was making I believe, and it ties in with the idea of the dog working out of a desire to please--because a simple pat on the side isn't anywhere near as "thrilling" (maybe that's a distinction that's important only to the human mind though) as, say, a liver treat or a tug toy or water bottle. And yet for many of us, that simple praise is our paradigm. Granted, perhaps the partnership is already there that makes simple praise (or a look) meaningful for such dogs, but as Kristi's example illustrates, a dog doesn't have to be indoctrinated in a "big prize reward system" from the start in order to want to work with its human partner (which is how I personally define a desire to please), including offering behaviors in an attempt to figure out what the human wants.

 

J.

[Donald McCaig]

Dear Doggers,

 

I have enjoyed this conversation, particularly because I knew Eileen’s Spot and have attended Kristine’s pet dog training class. I thank them both.

 

Apples and oranges.

 

The honest pet dog trainer (and most are) has a difficult job, not one I’d want. That first day of a class we’ve a man with the body language of Richard Nixon and a nervous Dobermann Pincher, a woman whose snappy Pomeranian is her “baby”. Another’s bulldog is twenty pounds overweight and can’t breathe. Then there's the college professor with the badly socialized, frightened “shepherd-mix”. Prof Smith watches Cesar Milan every week and has learned all the wrong lessons.

 

They all believe their dog offers “unconditional love” and cannot understand why, if the dog loves them, he pees on the rug.

 

What fun.

 

Against this, imagine a beginning agility class comprised entirely of experienced sheepdog handlers with their young Border Collies. Do you think they’d need the introductory lecture”Behaviorism for Dummies” or the complementary treat bag?

 

When I meet an urban/suburban new Border Collie owner, I am likely to recommend they find an agility instructor. For most, agility and obedience are more readily available than stockwork and working with your dog is more vital to human/dog happiness than what work the two of you are doing.

 

In the unlikely event that owner asked:” should I seek a treat trainer or a traditional (corrections) trainer, I’d advise them to seek the more accomplished trainer as measured by ribbons and wins and if said trainer was compatible, could teach and provided the owner liked the relationship the trainer had with his own dogs he/she should sign on.

 

Like most sheepdoggers, I am a pragmatist.

 

Perhaps because of that, I am no behaviorist. My dogs get the gleanings when I butcher (we share “our” kill) but they don’t do anything to get their “treat” except make me trip over them.

 

Treats are dog money; an abstraction which adds a layer to the dog/human relationship. Before you teach the dog “heel”, you must teach the dog “This is a treat. Trust me, you want it. I treat when you do what I ask” . Many add a metalayer:“I click which stands for the treat which stands for your doing what I want.” If the treat is money, the click is the treasury bond.

 

Does it work, sure. But its the long way around the barn.

 

Badly nurtured, mishandled, and unsound dogs can test an experienced trainer’s mettle - whatever his method - but millions and millions of pet dogs are trained by Joe/Jane Dokes more-or-less satisfactorily, not because owners are gifted trainers but because dogs are gifted trainees.

 

 

Hence, I don’t think there is a “correct” method for training pet dogs to the usual minimum required :housebroken, don’t bite the kids or UPS man, don’t counter surf, come when called.

 

I have come to think that our disagreement is more profound than quibbles over training methods: we don’t see the same dogs.

 

From Watson on, the Behaviorist goal was human manipulation. A benign autocracy of disinterested psychologists could reform criminals, addicts and other social undesirables after enough research had been done on rats and pigeons to understand the levers that produced needed results. In the nurture/nature argument they were 100% nurturists.

 

By the early sixties, almost all their human experiments had been discontinued but they had nearly a hundred years of experimental research on laboratory animals and BF Skinner, modified by Karen Pryor, Peter Singer and a sentimentality surge produced a new dog training school promising to be scientific, sensitive to the dog and, above all, kind.

 

Like Watson, “positive” theory saw animals as pure nurture and perfectly manipulable by detached trainers, once the proper motivations were applied.

 

The theory has evolved and I don’t know any “positive” trainer who doesn’t take ethology and genetics into account. But the dog they see is fundamentally manipulable. Those that aren’t? Rather than try other “cruel” training methods, some “positive” trainers would have the dog put down.

 

This is a very different understanding than the shepherd’s. I won’t lengthen my post with examples; there are plenty but Eileen’s experience with Spot will do.

 

The shepherd knows that as the dog depends on him, he depends on the dog. The sheepdog can do many things he cannot and has sometimes done things the shepherd had thought impossible. The sheepdog is somewhat manipulable but what is most valuable to the shepherd is the dog’s intuition of its work, its master, its livestock; the constantly changing gestalt of its world.

 

The sheep dog is mysterious, unexplained by learning theory and beyond the shepherd’s comprehension.

 

Ralph Pulfer trained and trialed sheepdogs for a half century and was for some of that time the finest sheepdog handler in North America. When asked how much of it he uinderstood, he invariably replied, “About fifteen percent.”

 

Donald McCaig

[Eileen Stein]

Now most agility instructors that I know insist that you can and should teach your dog to work for a toy. Why doesn't anyone suggest that you can teach working for praise? When I first threw a water bottle for Wick, she ducked. But then she learned that she could retrieve it, strip the label off of it, crunch it. Now she will work very happily for a water bottle. So in the beginning, no, it wasn't a reinforcer, it didn't motivate, so I showed her how much fun it could be to work for a water bottle. Why do you think that a dog can't learn to take immense pleasure not from a bit of food or a game of tug, but from a sincere "good dog" and a pat from his owner?

 

Yes. Yay!

[topnotchdog]

Thanks, Julie. Just a few more thoughts...

 

It shouldn't be a big mystery to folks who train dogs using any method (that is, this isn't the sole provenence of stockdog trainers). Once your dog becomes your partner and learns your expectations in various situations, then it knows what's expected of it, right?

 

I think that's the case. One trainer may describe the consistent demonstration of desired behavior in a partnership it as 'learns what's expected,' the other may describe it as 'sufficient reinforcement history that the behavior is fluent,' and a third may describe it in another way still.

 

 

The key, I think, to this discussion, though, is that stockdog trainers don't necessarily feel a need to use high-value rewards to get their expectations across to their dogs (on or off stock).

 

This statement brings three things to mind. First, it reminds me that the value of the activity or the partnership in the dog's eyes plays a huge role (and perhaps there are instances in which access to sheep aren't necessarily that great a thing to a dog, hopefully the exception, just as agility in itself may not hold inherent value to a dog).

 

Second, (and this is not to remind you personally, but I think it bears repeating), not all trainers or dogs are going to be at ease with using the flip side of the coin, the reminding the dog of potential reprimand (i.e. the look). A really effective way to teach all kinds of even very challenging things is through rewards. But those rewards are going to have to be adequate to create value for the activity or behaviors if it is not there inherently (better known as classical conditioning, always at work as well). This all may be another way of saying exactly what you were saying. Anyway, the trainer commitment has to be there, the dog has to be trained in "everyday manners" like being tied loosely to a chair, staying calm in the face of excitement, being quiet when crated and the like. Maybe a lot of people don't train their dogs to do that who also happen to teach using reward-based methods. If that's so, it's kind of a shame, as it gives the impression perhaps those methods don't work for those sorts of behaviors. Or that their dogs can't do them because they are not sheepdogs.

 

And finally, I am not sure how much room there is in stockdog work for solely reward-based training off the stock. I can't yet tell how badly it clashes with the general use of corrections on the stock. If reward based training is your cup of tea, as it is mine, it's not that rewards have to be used constantly or indefinitely in general (just like you don't have to tell a kid who learned to ride his bike at age 5 "Attaboy!" as he pedals down the driveway when he's 14). It's more that it seems the handler and the dog both have to be comfortable with the kinds of corrections that will be needed on stock (which of course will vary from dog to dog and situation to situation). And it's not that sheepdoggers here have not told me that before, it's that I didn't really get it as well as I do now (actually, I am still mulling it over).

 

Haven't figured this out yet. Generalization alert: I do think that some reward based trainers tend to throw the baby out with the bathwater, and don't take advantage of other ways of listening to dogs and working with dogs. A dog is an amazing creature, not an input-output machine. And I do think a number of sheepdogs I've seen might have had things clearly & efficiently explained to them without the use of a correction (stuff off-stock), which to my mind is more pleasant, respectful and often fairer. Maybe people just don't question what they're doing, or stop questioning as much at some point.

 

I think the following short bit I found online provides one example (if you love or hate the clicker does not matter, she could have gotten the same results without it for the purposes of this discussion). Did this work because sheep are generally not as important to a schnauzer as to another kind of sheepdog? Did it work because she had the one big chunk of time and access to work through this? DId it work because she had a partnership already established with the dog, such that the dog relatively quickly began to play a game of, "what does she want of me so I can get what I want?" Would anything have been lost had she reprimanded, yanked, or waved a stick in front of the dog a few times and made her behave on the way to the pen? How important are the results she got once she got inside the pen as a reflection of what happened outside the pen? (These can be rhetorical or not, they are just the questions I am asking myself. Cuz, you know, I enjoy that sort of thing. ) Song and sheep

 

B.

[juliepoudrier]

Maybe a lot of people don't train their dogs to do that who also happen to teach using reward-based methods. If that's so, it's kind of a shame, as it gives the impression perhaps those methods don't work for those sorts of behaviors. Or that their dogs can't do them because they are not sheepdogs.

I'm actually going to go work a dog so I'll respond to the rest later, but I think what you're getting at here is a matter of numbers (although I could be misinterpreting!). If the general public is mainly introduced to reward-based training, then the numbers of people doing that (or calling it that) is overwhelming compared to the number of people who train stockdogs. The sheer numbers alone will skew the perception, because of course *some* (perhaps a lot) of those reward-based trainers will do it consistentely and correctly and get the expected good result, BUT a good number will approach it half-heartedly, apply it inconsistently, misinterpret what they and their dog(s) are doing, and so on. I think we've had discussions in the past here regarding reward-based training that is applied ineffectively and inconsistently and whether that might be as "harmful" to a dog as any sort of (misapplied) correction-based (or mixed) training. No matter what training method a person uses, if it's inconsistent and ineffectual, then the result is a confused (and likely unhappy) dog. But anyway, I don't think most of the sheepdog trainers here would argue that reward-based methods don't work or are completely unnecessary (for the general public)--to the contrary, I think we would argue that for the general public,reward-based training is most likely to be effective (if applied correctly) and least likely to do damage (when applied incorrectly). As far as the general public is concerned, though, I think that misapplication, half-hearted training attempts, etc., using reward-based training might well create a perception that such methods don't work well--because of course it's most often those ineffective and misapplied applications that we *see.*

 

And finally, I am not sure how much room there is in stockdog work for solely reward-based training off the stock. I can't yet tell how badly it clashes with the general use of corrections on the stock. If reward based training is your cup of tea, as it is mine, it's not that rewards have to be used constantly or indefinitely in general (just like you don't have to tell a kid who learned to ride his bike at age 5 "Attaboy!" as he pedals down the driveway when he's 14). It's more that it seems the handler and the dog both have to be comfortable with the kinds of corrections that will be needed on stock (which of course will vary from dog to dog and situation to situation). [emphasis added] And it's not that sheepdoggers here have not told me that before, it's that I didn't really get it as well as I do now (actually, I am still mulling it over).

 

 

I am working with a few dogs now who don't necessarily see sheep as the best thing on the planet (or maybe they see sheep as great, but only when it's on *their* terms), and it's interesting trying to figure out a way to get through to them what's expected and make some training progress without having them go off and sulk. Praise is involved, but at some level the dog does have to understand correction, because if what really inspires the dog is diving in and biting, then I'd be hard pressed to allow the dog what it likes in order to make progress in other areas of training. And I think this is what you're getting at here. I have no idea if any of these dogs have been consistently corrected in their lives before sheepdog training, but it can become a source of frustration when every time you utter an "ah, ah" (for the diving) the dog turns away and sulks, because progress is going to be very slow. (This is a very simplistic explanation, but since you've been involved with stockdog training, I assume you know what I mean.)

 

J.

[topnotchdog]

^^ Yes, I understand what you're saying.

 

The sheer numbers alone will skew the perception, because of course *some* (perhaps a lot) of those reward-based trainers will do it consistentely and correctly and get the expected good result, BUT a good number will approach it half-heartedly, apply it inconsistently, misinterpret what they and their dog(s) are doing, and so on.

 

Actually I hadn't though of it like that, good point. I wonder if a similar percentage of sheepdoggers do the same? Good point about the numbers.

 

B.

[juliepoudrier]

I imagine the sheepdoggers fall across a similar spectrum. Probably the biggest difference is those who can't apply the training consistently and effectively stop doing stockdog stuff because they aren't likely to make much progress. If you're applying reward-based training poorly, you may still get acceptable results (i.e., the dog eventually learns to sit), but that might not be the case with the sheepdog trainers who can't apply the techniques well. So the numbers of sheepdog trainers who can successfully train is smaller than even the numbers who try it, and both of those numbers are far smaller than the numbers who try reward-based training in some form.

 

J.

[Root Beer]

But there's quite a difference between not reading a dog's emotions accurately, and not seeing that a dog who's being dragged is not moving voluntarily.

 

Of course there is a big difference. But the similarity lies in the handler stating that one thing is happening when there is actually something quite different going on.

 

I don't want to belabor this, and this will be my last post on this particular point, but is there any chance you are misinterpreting the owner's statements that s/he's training in a way that will utilize the dog's desire to please, as a perception on the owner's part that the dog's being dragged through the tunnel reflects the dog's desire to please?

 

I don't see how. The statement, "I want my dog to do "xyz" out of desire to please me" is pretty straightforward. And seeing the handler turn around and do something like drag the dog through a tunnel is pretty obvious.

 

I don't really see what is there to misinterpret.

 

I only ask because you don't seem to see much difference between these two things.

 

Actually, I see a very big difference. That is the point that I was making in the first place about the phrase "desire to please". I don't consider a dog being dragged through a tunnel or held at the contact point or pulled over a piece of contact equipment by the collar to be the dog learning because of "desire to please".

 

Have you ever asked a person dragging a dog through a tunnel if the dog is going through the tunnel *right now* because of a desire to please?

 

I don't tend to get into discussions like that with my students for many reasons. Instead, I show them how to break things down, use a reinforcer, and allow the dog to make the choice to go through. Since the classes that I am teaching are all about self control (on the dog's part), students learn to allow their dogs to make these types of choices and we work on that from the first day of class. Usually by the time we get to the point where we are working with tunnels, the handler has reinforcement skills and is able to resist the temptation to drag the dog through if he or she does not go right away. If someone is on the verge of doing that, I take the leash so the handler has to figure out how to help the dog choose to go through the tunnel with hands off. I will suggest techniques that the handler has learned in class to help the dog make the choice to go through. We might use the dog's mat, we might use a target, we might use a shaping process, we might use the Give Me a Break game.

 

If s/he replies, "No, but once I get across to the dog that this is what I want he'll do it out of a desire to please," you have a very different situation from the one you have if s/he replies, "Yes." You may consider both to be problematic because the owner is not recognizing the need for reinforcers in training, but in fact they are two very different situations.

 

If the dog really will go through the tunnel without the need for reinforcers, I don't consider it problematic. Most dogs learn some things very quickly and without much help and need more assistance with learning other concepts and behaviors.

 

If someone really can point to a tunnel, tell the dog to go through, and the dog does, that's great. If the handler were to call that "desire to please", I wouldn't disagree.

 

I disagree specifically to citing wanting the dog to do something out of "desire to please" and then proceeding to drag the dog through the motions in some way. The dog is doing nothing to please anyone in that situation.

[juliepoudrier]

I think the following short bit I found online provides one example (if you love or hate the clicker does not matter, she could have gotten the same results without it for the purposes of this discussion). Did this work because sheep are generally not as important to a schnauzer as to another kind of sheepdog? Did it work because she had the one big chunk of time and access to work through this? DId it work because she had a partnership already established with the dog, such that the dog relatively quickly began to play a game of, "what does she want of me so I can get what I want?" Would anything have been lost had she reprimanded, yanked, or waved a stick in front of the dog a few times and made her behave on the way to the pen? How important are the results she got once she got inside the pen as a reflection of what happened outside the pen? (These can be rhetorical or not, they are just the questions I am asking myself. Cuz, you know, I enjoy that sort of thing. ) Song and sheep

I didn't realize this wasn't a video. I don't see what this person did as any different than taking an overexcited border collie and sitting down in a packed stall with sheep and a book, or sitting in a chair in the round pen with the flock right there and simply waiting for the dog to relax (except that she kept her dog on the other side of the fence from the sheep). You're right that she probaly didn't need the clicker. I think this is simply an example of someone who had the time, access, and patience to wait the dog out, and perhaps a dog who was either partnered well enough with her or smart enough to realize that not screaming got the desired result (that is, the dog's personality was such that waiting and ignoring it didn't cause it to ramp up even more but instead to settle down). I doubt one could attribute the seeming ease with which she fixed the screaming to a relative lack of interest in stock (since the screaming would imply that the dog was very excited about the prospect of working).

 

(repeating this part of the previous quote)How important are the results she got once she got inside the pen as a reflection of what happened outside the pen?

 

As you know I think the dog's mindset as it approaches work (i.e., approaches the sheep or pen) plays a HUGE part in how that work will go. Does that mean the dog will be perfect on sheep if it figures out to behave itself when approaching the enclosure? Of course not. But you have a much greater chance of not having those initial yeehaw moments if you can encourage your dog to settle in its mind before it gets in with the stock. A settled mind is a mind that's capable of learning and to me that's the key of the dog's approach to the stock from outside the fence.

 

Would anything have been lost had she reprimanded, yanked, or waved a stick in front of the dog a few times and made her behave on the way to the pen?

 

I think that depends on the dog and the handler. If the handler is uncomfortable with such an approach then likely the dog will sense the human's ambivalence and so the handler won't be as effective (or the dog might end up confused). But if the dog understands, say, a verbal correction like "ah, ah!" then I don't think that using such a correction appropriately in this situation would cause any loss of partnership or learning on the part of the dog. A lot depends on the dog and the person. The key, IMO, is to get the dog's attention (however that works for the handler) and get the dog to settle its mind before you let it work.

 

J.

[Root Beer]

Now most agility instructors that I know insist that you can and should teach your dog to work for a toy. Why doesn't anyone suggest that you can teach working for praise?

 

Some do. And you can. I find it to be a lot more straightforward, though, to use what the dog is naturally inclined to work for.

 

My Agility dog does not care a thing for toys. I actually tried very hard to try to teach her to work for toys and it really never took. I finally decided that if she is food motivated, I should just use food and that worked extremely well.

 

Later, when I got into using environmental reinforcers, I also began to use permission to sniff as a motivator for her. That really got her focused and motivated.

 

I know some instructors are very set on use of toys, or praise, with all dogs, but I really have had more success with using what the dog is naturally inclined to work for. I've found that really is where the learning power lies. If your dog likes praise - use it! If your dog likes toys - use those. If your dog likes food rewards - use that. If your dog likes environmental rewards - use those. If you really want to use praise and the dog is indifferent, of course it's an option to try to teach the dog to work for it. Seems like a lot of extra work to me, though, when a whole world of reinforcers are at my disposal, and there are going to be at least a few that my dog will work for naturally.

[topnotchdog]

I think this is simply an example of someone who had the time, access, and patience to wait the dog out, and perhaps a dog who was either partnered well enough with her or smart enough to realize that not screaming got the desired result (that is, the dog's personality was such that waiting and ignoring it didn't cause it to ramp up even more but instead to settle down). I doubt one could attribute the seeming ease with which she fixed the screaming to a relative lack of interest in stock (since the screaming would imply that the dog was very excited about the prospect of working).

 

That is how I interpreted it as well. If anything it might even be more straightforward to fix if the dog was that interested, since the dog could focus on the one goal of "how can I get closer to the sheep." I think it is pretty neat that she communicated to the dog in just a few sessions "choose to calm down and you will gain access to the reward." (If it was a dog who ramped up, I suppose one could add in distance or block the dog's view if those would help settle her enough to make progress.)

 

The key, IMO, is to get the dog's attention (however that works for the handler) and get the dog to settle its mind before you let it work.

 

So if I'm understanding your point, it's not necessarily the case that corrections would be ideal to use to get the dog's mind calm before approaching the sheep given that corrections are going to be used in part to communicate with the dog on the sheep. 'Course you may also be saying it's moot to generalize, given that so much depends on the individual dog and the handler (with which I couldn't agree more).

 

Come to think of it, I think that holds true for agility as well. Some handlers 'correct' a dog for, say, knocked bars, and some dogs deal well with that and others don't. Some handlers would not do that (or need to to improve the dog's performance) regardless of the dog's temperament. I like how Kristen said it's not black and white, no rules on how anyone should enjoy the sport (as she said, you work with what you have, which I would extend to the handler as well).

 

B.

[juliepoudrier]

So if I'm understanding your point, it's not necessarily the case that corrections would be ideal to use to get the dog's mind calm before approaching the sheep given that corrections are going to be used in part to communicate with the dog on the sheep. 'Course you may also be saying it's moot to generalize, given that so much depends on the individual dog and the handler (with which I couldn't agree more).

Well I think it's both. For some dogs corrections would be the most effective means of getting the dog to settle and start to think, especially if that's the training paradigm the dog is used to. The point is that you're not only trying to get the dog to settle down about the sheep, but also trying to get the dog to remember that a human is in the picture too. It does no real good to have a super-keen dog if the dog ignores the existence of the human. But are corrections absolutely necessary? Probably not. The woman in your example waited her dog out. But if that hadn't worked for her dog, she would have had to come up with a plan B or C, one of which might have included corrections. The flip side is that the correction itself could serve to ramp a dog up further--that is, a dog who is determined to get its way could just choose to "fight back" against corrections. So, yes, it does depend on the individual dog.

 

In the case of your example, the woman clearly did have to time to sit down and wait her dog out. That might not always be the case, nor might it be suitable for every dog. And if you are using corrections to communicate to the dog while working, then it's something of a natural extension to use them to communicate to the dog about its attitude when heading to work. JMO.

 

J.

[bcnewe2]

I think it's not really the corrections that get the mind to think, it's more snapping the dog out of not thinking that is the real issue. For some that might be corrections, others time, but it's getting the brain to go somewhere other than the excited or fear place it's in.

 

Example:

Mick has some dog aggression issues, being observant I can usually tell exactly what he's thinking, but if I try and correct him or raise my voice to intercept it sends him right off the deep end. Friends have teased me that correcting him in this issue is his kill command. It's not, but it obvioulsy doesn't snap him out the the frame of mind he's in. so for him a redirect works almost 100%.

 

Same as the reactive dog or excited not thinking correctly dog, a redirect of brain thoughts is what IMO is the key, it does matters as to what works but that's an individual dog thing.

[juliepoudrier]

I think it's not really the corrections that get the mind to think, it's more snapping the dog out of not thinking that is the real issue. For some that might be corrections, others time, but it's getting the brain to go somewhere other than the excited or fear place it's in.

I think we're saying the same thing just two different ways. IME, if a dog is lunging at the end of its leash just dying to get to the stock to the point that it's tuned me out completely, then an "ah, ah!" certainly does draw the dog's mind back to me and thus put me back in the picture. It also serves to stop the escalation that's going on with the desire to GET TO SHEEP, simply because the dog must acknowledge the correction, which naturally takes its mind off the MUST GET TO SHEEP wavelength, a resetting if you will.

 

I *did* note that for some dogs, corrections just escalate things. Ranger's littermate is one such dog. So her handler has to come up with a different method to get in her head and get her focused while still including the human in the picture and not letting her run wild among the stock.

 

Here I'm talking about a dog's behavior as it approaches work, though, and not aberrant behaviors toward other things (people or dogs).

 

J.