Working Dogs Crossover to Companions

But brookcave, just because you have always used a lot of corrections in your training doesn't mean that's the only way you can do it, nor does it mean it's "the most effective way".   Look at schutzhund  and hunting dogs-- "everyone" used to believe you have to use corrections and/or ecollars when training those dogs, but that attitude is starting to change as people demonstrate it just isn't true.
 
Really, all the behaviors you talk about- waiting before charging through doors (ANY door, you don't have to train on each and every door), following off-leash, going in crate, not entering or leaving crate unless allowed to do so, basic doggy manners-- are quite easy to teach using purely positive methods. Go to an agility trial or class, and you'll see the proof with your own eyes.
 
I have no experience with herding, but if I wanted to roduce a herder I'd approach it just like training an agility dog. With an agility dog, you can use a clicker to train little essential pieces of behavior away from the equipment-- follow my body language, down, stay, any verbal cues you may wish to use. Then you're done with the clicker by the time you get onto a real course. You could do the same with herding, teach the basics of control from handler with the clicker then put it aside and move onto sheep. Same with biting dogs, train the basics, then put the clicker aside and move onto the decoy.
 

I have no experience with herding, but if I wanted to produce a herder I'd approach it just like training an agility dog.

Actually, to produce a herder you have to start with genetics. Thank you for an excellent response. This gets us to the question that ADogWhisperer had on the other thread - what goes into training a livestock working dog.[]

Dogs that are bred for herding, or have been bred in the past for herding, ideally come with three basic innate abilities:

• Prey drive
  
• The ability to read and react properly to the flight or fight zone - this is called "balance"
  
• The ability to stop the escape of stock without unduly panicking them - this is called "style" or "eye"
  

Herding dogs have these abilities in varying degrees - the first, prey drive, is a non-negotiable. The second seperates real herding dogs from the pretenders. The third can be trained to some extent, but dogs that have it naturally are most highly prized by those who need stockdogs for practical work.

Notice that these three things don't all obviously work together. Prey drive would tell a dog to chase the sheep, while "balance" tells the dog not to chase the sheep. And "eye" is the threat of chase, with no follow through, but the sheep has to believe the follow through is there, so it's not as simple as just training a dog to LOOK fierce. The dog's got to believe he has the power to follow through also, or the sheep (or other stock) will have his number.

With an agility dog, you can use a clicker to train little essential pieces of behavior away from the equipment-- follow my body language, down, stay, any verbal cues you may wish to use.

I should note here that I've used clicker training to shape agility, flyball, frisbee and obedience behaviors (of course), therapy dog training, rehabbing rescues, and goosedog training. I've also used it where you admit you have no experience - herding.

I'll share how the initial introduction of a dog to stock works, and let you decide where a clicker would fit in. I have a small number of quiet sheep in a fifty foot wide pen. I take the dog in on leash. If he is quiet and minds me, I let him off leash. If he is too excited to even walk on a loose leash, we'll work on that first by going in and out of the pen until he realizes that we are going nowhere until he settles.

Prey drive is the number one instinct - as it should be. Most dogs run straight at the sheep. Fine. I let the dog bust through, then the dog turns around and heads back to the sheep. Meanwhile I've walked away from the sheep and have started around to his side. The sheep follow me, opening up more space on his side of the pen.

Now different things may happen, depending on what kind of dog I've got. Some dogs are so excited that they want to keep heading straight into the sheep. These dogs will need to bust through and get stopped by me a few times before they start to think BEFORE the correction, and in that moment feel, on their own, the flight zone of the sheep.

If the dog feels the flight zone of the sheep on his own, enough to stop heading straight back into them, the dog will start to circle around instead. Now there's two forces at work - the dog's desire to run straight at the sheep, and my corrections pushing him away. Like gravity and escape velocity acting on a satellite, the dog is pulled "into orbit" where he'll experience the balance between the flight zone and loss of control of the sheep.

I'll walk the sheep away from him, then back up and turn the other way, so that he ends up heading into my space. Most dogs will then turn and go around the sheep the other way. In this way we'll make figure eights, with me encouraging the dog with verbal corrections to make arcs that are farther and farther from the sheep, until the dog realizes that he can simply stay at the tails of the sheep and control them from there, as long as he stays well back for "leverage".

This awakens instinct #3, the ability to control the stock with a minimum of movement, or "eye." This comes from the dog learning that while following his prey drive is welcome, he must "read" the sheep to adjust his approach, so as not to frighten them, whether it's circling them or following them straight. At the time that the dog settles in behind the sheep and is comfortable and able to hold them to me with no commands or corrections, and without pushing them past, as we walk all over the small pen, we are ready to go out in the big field.

You could do the same with herding, teach the basics of control from handler with the clicker then put it aside and move onto sheep.

As you see above, none of the basics we start out with, involve commands, control, or specific behaviors. The dog must come up with his own way to control the sheep, because it's different for every dog, and in fact it's different every day, and even from hour to hour. Sheep, all livestock, behave differently at different times, so teaching rote behaviors is useless.

Sheep are not agility obstacles. Teaching a dog to perform "dry" behaviors on inanimate objects, and then hoping it will carry over to the livestock, is like suggesting that you could learn something useful from teaching a robot dog to run an agility course. You won't learn the most important part of the equation - how to relate to another living being and adjust your actions accordingly. Dogs have to learn that about the stock they work.

When I ask a dog to do something, say move to the right to turn the sheep to the left, he not only has to do that, but also calculate, based on what he sees in the sheep, how much movement can be made, how far to stay away from the sheep while moving laterally, how close he needs to be to maintain control, and assess the mental state of every member of the flock to ensure that no rogue elements will be taking advantage of his releasing pressure to the left side. He also has to calculate whether the sheep will bolt to the side they will be going, and how to stop it potentially if it happens. Finally, he's still got a brain cell left to keep tuned into me, for my next signal, and if he gets no signal, he's got to decide what it was I intended ultimately and how to get the sheep there.

What I experimented with, using the clicker, was fixing some difficult problems I ran into with two of my dogs. It seemed to be successful with one - the other was problematic - it's hard to seperate the severity of the problem I was addressing from the limitations of the tool and method.

Ben, the first dog I ever trained, had a problem where he was avoiding "contact" - in other words, he'd avoid the point where the stock were held from escaping and came straight on to me. He wold flip back and forth on the other side of the sheep and if the sheep pushed back or challenged him, he'd crash into the back of them and scatter them everywhere. Not very useful.

One day I took some young ducks, Ben, and sat in the front yard and clicked Ben every time he turned the heads of the duckies back to me. Ducks are very, very sensitive, so Ben had to work pretty hard to get it right. However, I had to drop it almost immediately, as I noticed he was combining some other unacceptable habits the second he made contact. Everything happens almost simultaneously, so even a click is too long of a cue to ensure that discrete behaviors were being rewarded without reinforcing bad ones.

And that's one big fat problem. There's very little that happens out there that is discrete. It's far better to correct when the sheep/stock are going wrong, and leave everything else alone.

Which brings me to my adventure with Don. He came to me with a deathly fear of the "lie down" command. So I used the clicker to train it "dry", then went out and started working him again. For a while it seemed to be working like a charm. I used a verbal cue instead of the clicker, of course. Then he started diving under the sheep and slashing their bellies when asked to "walk up" after a "down."

Looking back on it now, I believe what was happening was that I was asking him to lie down when things were getting tense, to relieve the pressure on the sheep. When I reinforced him for physically lying down, in his mind I was actually reinforcing the prey instincts that were building up in reaction to the increasing chaos. I have experimented since then with deliberately instilling this in my goosedogs (who chase geese who can fly away, and need this kind of super-prey-drive to keep them motivated). It works like a charm.[]

The point of all this is not to claim that there's no other way to train a stockdog. Instead, I bring this training up as an example of a whole class of dogs that are trained without benefit of clickers and yet are happy, well-adjusted, and probably some of the best examples of independent problem solvers in the canine world.

Becca,

I guess my interpretation of this is that #1 and #3 are pretty instinctual behaviors that feed off of each other and that #2-the balance is really independent of the other two and is dictated by both the dog and the sheep. 

Like you said the dog has to think, has to learn how to read the particular flock as to how far he should move, when to use the eye or the chase and when to back off to acheive the desired result- that is where the shepherd wants the flock to go.

Isn't all of this easily learned by puppy BCs following and mimicking their mother, watching for cues, hearing the various commands?

Sheep are not agility obstacles. Teaching a dog to perform "dry" behaviors on inanimate objects, and then hoping it will carry over to the livestock, is like suggesting that you could learn something useful from teaching a robot dog to run an agility course. You won't learn the most important part of the equation - how to relate to another living being and adjust your actions accordingly. Dogs have to learn that about the stock they work.

Excellent insights! You've put down so many things on paper that I couldn't quite grasp, and you've given good examples.

I cannot see how that does NOT apply to our non-working compainion dogs. It really is relevant to any organism functioning in a environment it's learned to live in; us of course, too...

My dog comes from working lines, and although he is not "employed" as a hunting dog, it's interesting to see how adapt he is in putting the right amount of pressure on variety of objects he retrieves. We play this little silly game where he finds and brings objects we ask him to bring; he can lightly hold my tiny hair band or a little plastic bottle without ruining it, etc. The first time he brought me that bottle he squished it by biting too hard on it. I would not know how to teach him to put the right amount of pressure on an object using a clicker; 'pressure' is not something I can show him, it's not discrete. And then again, I am not a trainer, so what do I know... [] But, I could easily do it with a verbal, pronged correction when he was going at it too hard, and a happier relieving "yes, yes" when he was getting a feel for it - like: "cold, warmer, warmer, yes". My friends are happy that he can bring them their expansive shoes without punching a hole through them []. Just something I've noticed...

Also, I've never thought about giving your dog 'options' to chose from while not giving it discrete instructions on what to do. Your average Joe wants instructions, instructions on how to manipulate a washing machine: "What do I press and when to I stop." He keeps on hammering at it like an A-student. But good trainers do have to walk that fine line, looking at the dog they have in front of them, look at its slightest signals of frustration or delight, and act based upon that. I believe that training (as well as laws for that matter) should not be tailored towards the lowest common denominator - your average Joe. That doesn't make for the best trainers and best working dogs; or, companion dogs for that matter.

Another thing I've noticed, and you reassured me of it, is that the best reward to a dog is letting it do the behavior it wants to do at that *moment*.

I think there's two reasons a pup can't just imitate an older dog and learn this stuff (though some of it can be learned while working in tandem with another dog).

First, a pup doesn't have the full maturity necessary to work sheep the way a mature dog can. A dog has to come up with his own individual way of controling the stock, and that method is developed over many years of experience working with stock. In other words, what works for a dog's mother is like not going to work for the pup, and probably never will.

Second, one of the key things a dog has to learn is how to balance constantly watching and reading the stock, while still being ready to take direction from the human partner. The best way to help a dog learn to do this is to let the dog do what is natural first, then gradually introducing the human factor. Trying to do it the other way around makes the dog think that the natural part is undesireable and may tend to discourage independent behavior - a dog that will "step up" when needed.

There's a third reason, and that is that adding an additional dog complicates things, no matter whether the dog is one that has some kind of relationship with the young dog, apart from working. Some mature dogs get excited when the young dog gets excited and pack prey drive goes into action - wheee! Now I've got a rodeo to take care of. Others get locked up, intimidated by the presence of the other dog, who is "correcting" the free movement of the young dog. This doesn't always happen and sometimes it's actually useful to use another dog in these ways - either to calm things down or rile them up. But the other dog is withdrawn quickly once the instincts of the young dog are solidly in play without help.

Thank you so much for your recent posts regarding training! I am delighted that what you#%92ve offered is lifted some limitations in my thinking.

Interesting thing about CM is that he formed his ideas about dogs on a family farm - with working dogs. He maintains that model as “ideal“, and now addresses dogs who are made imbalanced by neurotic sub/urban human-dog relationships. Some of his techniques feel ”unnatural“ to me, but besides the fact that I'm a newbie, I chalk some of that up to the fact that he#%92s working with people who have chosen to live with some weird behavior from their dogs, sometimes for a long, long time. He#%92s giving tools, most often, to people who seem to have little dog sense, and are perhaps a little awkward in how they relate to the world (thus his focus on how owners project their emotional baggage onto their dogs).

I've been unable to reconcile the ”calm submissive“ state CM insisting that dog maintain. I guess, for owners who aren#%92t able to handle more energy in their dog, that#%92s a good policy. It taught me why my dog wasn't able to sit calmly for me while I was whooping her up with treats, voice and excitement.

But, a herder can have a dog revved up with high energy, and still get a down. So, what about when we want to focus drive rather than just burn that energy off in exercise?

In another thread, Angelique offered this distinction (http://forum.dog.com/asp/tm.asp?m=342093&mpage=4 )

ORIGINAL: Angelique
A dog with "drive engaged" is anticipatory, ready, and is functioning within a state of controlled excitement through the "energy" level of their leader who is encouraging the dog's drive to be engaged.

A dog with "drive off" understands what is expected in social situations, is calm, and looking for directions from their leader - who is indicating by their demeanor what is expected of the dog.

What do you think about the idea of managing drive for different contexts like the living room, the sidewalk, the agility ring, or in the fields?

ORIGINAL: Ixas_girl

.......I've been unable to reconcile the ”calm submissive“ state CM insisting that dog maintain. I guess, for owners who aren#%92t able to handle more energy in their dog, that#%92s a good policy. It taught me why my dog wasn't able to sit calmly for me while I was whooping her up with treats, voice and excitement.

But, a herder can have a dog revved up with high energy, and still get a down. So, what about when we want to focus drive rather than just burn that energy off in exercise?

.........
What do you think about the idea of managing drive for different contexts like the living room, the sidewalk, the agility ring, or in the fields?

I'm not sure I understand your question so I apologize in advance if I'm answering something different than what you are asking. But, I think one of the very very common misconceptions about herding is that herding dogs need to be "revved up" when working. This in turn leads to the very very very common misconception that a border collie should be manic/hyper/need tons of constant physical exercise. There are lots of reasons why a good border collie, or good herding dog of any breed is difficult to live with, but hyperactivity shouldn't be one of them. When managing livestock, dogs do need to be able to pour on speed on occasion, but the overwhelming majority of good stock management is done slowly and calmly. Ideally, the dog is mentally calm and focused, NOT revved up. At agility trials, you frequently see handlers "revving" their dog by playing tug games, speaking in an excited tone of voice, that sort of thing. You don't see handlers taking their dog to the post at a herding trial playing tug with the leash or a toy. The dog is walking, dare I say it, in a "calm submissive" manner, as the dog and handler approach the stock. And, even less would you see someone in an actual real working situation "revving" their stock dog up before sending it off on a job. You hear handlers instructing their dog to "steady" or "take time" far far far more often than you hear handlers exhorting their stock dog to "go, go, go".

I think a lot of people see herding as being all about speed and physical force, and often young inexperienced or poorly trained dogs do work mainly based on speed and force. Certain tasks do require a decent amount of speed, but often when you see a dog running hell bent for leather it's to compensate for mistakes. Similarly, certain livestock do require the dog to exert some real force, but often the dog that is biting is doing so to make up for a lack of real power, rather than displaying power. Good herding is boring to the untrained eye. The epitome of good herding is calm relaxed livestock being controlled by a calm and relaxed (yet intensely concentrating) dog being controlled by a calm and relaxed human.

Not sure if any of this answers your question, but this is a topic I love to discuss.

And, as an aside, I agree with the original post that clicker training can be an extremely powerful tool to train specific actions but that herding is only a tiny bit about the dog learning how to respond to specific commands. Training herding is far more about setting up situations that allow the dog to feel the good and bad consequences of his actions, and learning how to avoid the bad consequences. Herding instinct, in my view, is the ability of the dog to recognize and feel good when it is in calm control of the livestock, and to feel uncomfortable when it is not in calm control of the livestock.

I see the actual "working" part of working dogs, whether it be herding, hunting, tracking, biting the decoy, to all be instinctive-- the dog already knows how to do it, and with some practice he gets better at it. You don't actually train your BC to herd, or train your pointer to point, or your malinois to bite; what you train is your ability to control and focus the dog so it exhibits the behaviors when and how you want them to be exhibited.
I don't know what commands you actually issue to herding dogs-- an instant down, a go around? I don't see why you can't train those away from the sheep, and then introduce the sheep. The sheep are the ones that teach the dog, right? when the dog messes up, the sheep let the dog know.

I don't know what commands you actually issue to herding dogs-- an instant down, a go around? I don't see why you can't train those away from the sheep, and then introduce the sheep.

It's because the mechanical motion of a command is only a teeny bit of what's required from the dog when I ask for it - some of the commands actually don't mean literal mechanical motion of the dog. I'll illustrate:

"Lie down" - stop putting pressure on the stock. My youngest pup is learning this one right now - I get the sheep into a place where they won't move if he correctly doesn't push on them, and then ask for the "lie down." What he gives me doesn't matter - he's learned if I say "lie down" and the sheep aren't pressured, that's correct. He is then free to use whatever change in attitude or physical posture to effect that. If I teach him that lie down means go plop on the ground, it will take him a longer time to unlearn that you also have to let up the pressure - it's very easy for a dog to go plop on the ground AND lean into the sheep. And it's super hard to undo that habit.

"Come bye" usually means circle the sheep clockwise. It also means maintain contact with the sheep so they don't get away, but don't push on them so they move perpendicular to the arc of the dog's movement. Most of the time the dog will be asked to do this when the sheep and dog are already engaged in a lateral movement of some kind, so it's really important that the dog not understand this as blindly circling, but as a movement that incorporates rule #1 - don't get increase pressure unless asked. Again, if you teach this dry, you'll end up in a situation where the dog has to unlearn bad habits and that's more stressful by a long shot than learning the right way to begin with.

"Way to me" means circling the other way. I taught these to my oldest dog, dry, and I'm still living with the damage I caused. If you teach them dry, it's super easy to teach the dog that you mean for the dog to circle you - because without the sheep it's hard to see what the dog is interacting with. That's what Ben learned. On the other hand, if you manage to teach the dog to circle inanimate objects, you do two things: one, the dog will always prefer to interact with inanimate objects as a default behavior when the pressure gets too much (pressure from the livestock), and two, as I said before, the dog learns the easy part without the main point of it, and once there are rules to the game, you'll have a devil of a time getting the dog to enjoy it.

"Walk on" is so complex I'm not sure I could explain it adequately to someone who doesn't mess with livestock much. Basically it means I want to increase the dog's pressure on the stock.

I'll try to illustrate. Saturday I was catching up lambs for tail docking, and one was being a real beast about getting away from me. She was savvy as all get out - she'd slip up next to the ram (who weighs about 350 pounds) so I couldn't see her or get hands or crook on her.

We went round and round for a while, and finally I realized we needed to move her off that ram. I stepped so that I was sideways to both, facing the side that the ram was on. Cord balanced automatically and then froze, feeling the increased pressure.

Now I leaned forward a bit. The ram's ears twiched to the side. I leaned back and hissed twice (almost the very softest "Walk" cue I have). Cord flicked his eyes at the ram, who turned away from the lamb for just a second.

Now I took a whole step back and turned to the lamb - she had her head turned nervously looking after the ram, but for just a second her ears flicked back. I clicked my teeth once. Cord still didn't move but shifted his eyes to the lamb, who turned away from him - and me! I snaked my crook out and hooked her around the shoulders.

The sheep are the ones that teach the dog, right? when the dog messes up, the sheep let the dog know.

Again, it's hard to describe in words to someone who's not been around livestock much, but it's far, far better if your dog learns when it's messed up from you, than from the sheep.

My friend tells the story of two littermates he had. One he gave to a fellow to raise and train, because the man wanted to prove he could clicker train a dog to work. He's a top disc dog competitor so his clicker training skills are top notch. The other pup stayed with the sheepdog trainer and was raised his way.

When the dogs were a bit over a year old, the frisbee trainer paid a visit to the sheepdog trainer. He went out into the big field and demonstrated that he could put his pup anywhere on the field - over 400 yards away, even.

The sheepdog trainer, being quite open minded, took the dog back and proceeded to introduce him to livestock. The first time he went in with him and tried to use the commands with sheep in the picture, the dog fled in terror. He is a very patient man and used no harsh methods with the dog, but the dog never reconciled moving about like that, with the pressure he had to negotiate from the sheep. In the end he made a decent pet for someone, I hear.

Meanwhile the littermate learned his job rapidly and is today working on a ranch and doing well in trials.

When a person is involved, they can "make what's right easy" even as they are correcting something that is wrong. If my pup dives in amongst them instead of stopping when I asked, I can speak his name to get his attention, then walk into the space where he's NOT supposed to be, pull the sheep way back and then ask for that stop again. Without my interference, there'd be no clue where he was supposed to be as the chaos got worse and worse in the paddock. Sheep are not forgiving - they just want to survive and unless I demonstrate my ability to protect them, they will beat my pup to a pulp.

Buster, thanks for your thoughtful and informative reply. That makes lots of sense. I think I was being simplistic in lumping agility, protection, herding ... all kinds of "work" as one category, and domesticity as another.

You also helped explain why my ACD mix's default behavior is rather calm. []

I still wonder about the "crossover" part of this conversation. I'd love to hear more about:

ORIGINAL: brookcove
So part of "manners" training means learning the rule, "Always wait at open doors and gates." It takes a lifetime to proof every door and every gate that exists, but it takes a very short time for a dog to learn that its place is at your side, not charging through a door willy-nilly.

If I didn't work my dogs I'd still train them this way. I don't think everyone should, certainly.

ORIGINAL: mudpuppy


I don't know what commands you actually issue to herding dogs-- an instant down, a go around? I don't see why you can't train those away from the sheep, and then introduce the sheep. The sheep are the ones that teach the dog, right? when the dog messes up, the sheep let the dog know.

Becca already provided an excellent answer to why you can't just "dry" train responses to commands and then transfer that to herding. I do dry-train a recall, a sit, a stay, and a down before introducing the dog to stock, but again, those commands are just a minute part of what a dog has to be able to to when actually workng stock. As Becca says, a command like "down" on stock is way more nuanced than "roll over". It's true that if the dog messes up, the sheep will let the dog know, but that doesn't mean the dog automatically knows how to fix the mess, or that he even wants to fix the mess. Chasing stampeding sheep can be pretty fun for a dog, too. Plus, those messes can be life threatening to the stock, the dog, and perhaps even the handler. One really doesn't have the luxury of waiting for the dog to "offer" correct behavior when one is standing in a pen with some 200 lb hoofed animals running amok, and a dog that hasn't yet figured out how to get to the correct position or experienced the euphoria of being in the correct spot. If the handler isn't pretty darned proactive about helping the dog get to the correct place, sheep with broken necks, stomped dogs, and handlers with broken limbs or worse are all possible (and not rare) outcomes.
And yet, as Becca pointed out in her original post, these dogs that do sometimes have the scruff of their neck grabbed, that do sometimes have a stick slapped on the ground, that do sometimes have the handler step toward them in a manner that doesn't look all that inviting, somehow manage to be among the most creative thinkers in the dog world and are often intensely bonded with their trainer. They are pretty much the complete opposite of the dogs that you described in the "clicker training-??" thread that are afraid to do anything except stand there. I think the key is what Becca pointed out in her original post. One can both offer a positive punishment for undesireable behavior, AND follow that up immediately with the opportunity to earn a positive reward. Given the joy with which a well trained herding dog works, and the intense bond that is often developed between these dogs and their occasionally "punitive" handlers, I don't think that good and fair and humane dog training has to be religiously confined to only one of the four quadrats of the operant model.

I'm not sure I believe you, in that punitive training methods are necessary or even the best way. These sound like the same sort of "excuses" rotection dog trainers and hunting dog trainers make to justify using harsh corrections. The truth is, people train using these methods because that's what they used in the past, not because they are the "best" way.
 
It's not difficult at all to train a dog to never charge past the handler through doorways without using corrections, or to become an off-leash "velcro dog"; claiming you have to use corrections to get these behaviors undermines any claims you make about the validity of applying punishments to getting other behaviors.

There are definitely no "excuses" here. There have been many, many discussions in the herding world about methods - probably long before non-forceful methods were ever considered in obedience or other venues. It was in the 1940s that JM Wilson's quiet, calm handling of his dogs brought admiration and astonishment from dog trainers and enthusiasts right up to the Royal Family. His dogs never wore collars from birth to death and his motto was "You should never need to touch a dog except in affection."

ETA: To this day people come over from agility and other sports, always with the idea that they will be the ones to revolutionize herding training as hunting and protection work have been. I've been watching them come over for ten years now and nary a one has figured out the puzzle - or they all come to the conclusion that herding IS unique.

It's not difficult at all to train a dog to never charge past the handler through doorways without using corrections, or to become an off-leash "velcro dog"; claiming you have to use corrections to get these behaviors undermines any claims you make about the validity of applying punishments to getting other behaviors.

The point about using corrections in those contexts is that it's the time to build a relationship with my dog, where the dog knows that he'll be informed if he's incorrect, BUT that communication will be both clear and fair, AND the dog is free to find the correct choice - and ideally I'll make the boundaries of "grace" very, very clear. Building this give and take outside the working environment is vital - starting it in the training pen is trying to introduce too many concepts at the same time. I'm sure you are aware what a no-no that is.

I reiterate that the point is not that this is the only way to train a dog, nor one that I'd recommend to the casual pet owner, but to maintain that dogs trained this way aren't unhappy, shut down, mindless robots. I think it's odd that people continue to insist that all punishment automatically instills this in dogs.

One of these days I'll video one of my sessions with Cord, who's still being fine tuned on several major points. He was trained rather aggressively (not by me) and still has a fear of training sessions, so he often slinks out to the field if he catches on that it's training, not chores. I don't throw treats and my encouragement is only of the calmest, most boring kind - mostly it's verbal corrections. You'd hear - "Lie down. AH-AH! Lie down! That's it, walk on, come bye NO! come bye, down, there, HEY, walk on, HEY, walk on heeeeeeeeey you then . . .walk on (that's when he comes on a bit too strong while bringing me sheep). Lie down, LIE, come bye AH! come bye . . walk on . . ."

What you'd see is a dog that starts out extremely tense, moving jerkily, too fast even to the inexperienced eye - you'd no doubt notice the sheep running and my raised voice. At the end you'd see Cord, sheep, and me all very calm and relaxed. Cord has his ears forward confidently and is no longer waiting for commands to take control of the sheep, to keep them moving at a good pace and in a straight line.

That's the thing I'm looking for - the time I spend "talking" him through his initial mistakes at the beginning actually empowers him as he feels out the RIGHT thing to do on his own - I don't take that power away from him by MAKING him do the right thing. His confidence grows and when I call him off, he's grinning, perking his big ol' mule ears, and doing kangaroo jumps like a fool. The next time I put him in a similiar situation you can see the spark in his eye as he's saying to himself, "This is something I can do right."

ORIGINAL: mudpuppy

I'm not sure I believe you, in that punitive training methods are necessary or even the best way. These sound like the same sort of "excuses" protection dog trainers and hunting dog trainers make to justify using harsh corrections. The truth is, people train using these methods because that's what they used in the past, not because they are the "best" way.

I eagerly anticipate the video on You-Tube of you in the field with some livestock and a dog and a clicker, demonstrating the "best" way to train a herding dog. Or, I'd be happy to see anyone doing it.

It's not difficult at all to train a dog to never charge past the handler through doorways without using corrections, or to become an off-leash "velcro dog"; claiming you have to use corrections to get these behaviors undermines any claims you make about the validity of applying punishments to getting other behaviors.

I didn't claim that you have to use these corrections to teach a dog not to barge through a gate. Becca did say that she does use corrections, but even she didn't say she had to. If the only training I did was teaching basic manners, obedience, agility and tricks, I might very well not use any corrections. But I have a fair amount of experience training in a venue where failure to administer corrections can sometimes be life threatening. Since my self preservation instincts, as well as my desire to keep my dog and my livestock as safe as possible outweigh my religious convictions about a training method, I've administered corrections in the herding field. And, I've found that lo and behold, this didn't appear to do any terrible psychological damage to my dogs. Indeed, they are very eager to participate in herding, and have become more eager as they have been given the opportunity to do it correctly. And they become increasingly creative and independent thinkers as their experience expands, the opposite of the fearful repressed dog that is frequently portrayed as the inevitable result of ever straying out of the R+ only quadrat.

The "cross over to companions" part of this discussion is that having discovered that working dogs don't respond to corrections as abuse, I've found that certain behaviors in other venues can also be effectively and fairly taught with the judicious and well timed use of mild aversives. And if those corrections hasten the process of the dog learning to do something correctly, and hence hasten the dog's opportunities to be rewarded, then sometimes the use of corrections may be just as "dog friendly" as rigidly adhering to a philosophy that admits no nuance in any alternatiive approach.