Jump to content
BC Boards

My reasons for "positive" training


Recommended Posts

That is a key difference, then, in how we view training and learning.

 

When I'm teaching sit, I'm not approaching what the dog is learning as "don't stand, don't lie down, don't . . ." It is "sit". So, I am not teaching by communicating "don't, don't, don't". "Sit = rump on the floor and stay still" is the goal at hand when I teach sit. It's about what the dog is to do, not what the dog is not to do.

 

I think I can clear some of this up…

Here is my understanding was that in the instance of the dog taking the tunnel instead of the jump.

 

I was under the impression that you indicated to the dog that you wanted it to take the jump. But the dog headed for the tunnel instead. (for the sake of this aspect of the discussion it doesn’t matter why the dog didn’t take the jump) So instead of saying “No, Jump.” you simply say “Jump.” (I’m also assuming that the command “Jump” means approach the thing that looks like a cavaletti and jump over it, as opposed to “Over!” which might mean jump over whatever is in front of you now.)

In that instance, saying “Jump!” would presumably have the effect of letting the dog know that the tunnel was not what was wanted, as he seems to think, but that the jump is what’s wanted. So although you are not verbally saying “no,” the “no” is implied, since he is in fact barreling toward the tunnel. It also lets him know that what you do want is the jump.

 

When I teach "jump" I'm not teaching the dog not to take a tunnel or A Frame or weaves or whatever. I'm teaching the dog to jump over an obstacle. When I'm teaching left heel I'm not teaching the dog not to walk on my right or in front of me or behind me, but to walk on my left side in a certain sustained position. When I'm teaching a dog to wait politely at the door, I'm not teaching "don't jump on the door", but sit and wait at the door.

 

Ok, understood, but again, unless the dog is doing something other than waiting politely at the door, you have no need to teach it to wait. You can reinforce the sitting and waiting with a treat, but there is no guarantee that the dog will not think he received the treat because his tail happened to be draped over his right forefoot as he sat there – especially if that tail happened to move at the moment you clicked or popped the treat into his mouth.

On the other hand, if he is jumping up on the door and you say “Wait!” and he responds by sitting, and you click/treat, he will very likely make the connection that jumping was not what was wanted, and sitting was what was wanted.

 

Think of the dog who grabs treats from your hand roughly. I use the word “Easy!” for this behavior. At the end of the process the dog understands that what is wanted is for him not to be rough. He can, and usually does interpret this to mean that it’s OK to lick the treat up from your hand or to take it gently – but the main lesson is what he should not do – that is, be rough.

In many, many training situations it is not necessary to wait for the dog to get the behavior wrong. You can simply lead the dog by the method of your choice into doing what you want him to and rewarding the correct behavior. But in the case where the dog is exhibiting an unwanted/ incorrect behavior, you can clarify by issuing a correction, and then illustrate or ask for an alternate behavior. In that case the dog is learning “this, not that.” The type/severity of correction is dictated by the situation. If the dog fails to sit, a soft “No.” is probably sufficient. No pain, no anxiety. If he’s manhandling a sheep, the stock stick may need to come into play, either being banged on the ground, or in the case of a bloodletting, a bop on the bean.

 

Sure, everything that a dog does on cue eliminates other possible actions. I'm surprised to find that there are people who consider every single cue (or "command", as some say, or "directive") as a correction, no matter what the dog is doing when the cue is given. Knowing that does help me understand why we are not always able to understand each other's training choices or point of view.

I guess this strikes me as oddly as the idea of not utilizing corrections in training strikes many of you. I really don't consider everything I ever ask my dog to do to be a "correction" because I am, in essence, telling the dog to stop doing something else (whether it is right, wrong, neutral). To me that honestly doesn't make any sense.

 

Here I think that it may be safely said that our definition of "correction" is somewhat different - at least in some situations. (Sorry if this is confusing to read - I haven't quite mastered the art of using alternating quotes/ responses)

Link to comment
Share on other sites

  • Replies 210
  • Created
  • Last Reply

Top Posters In This Topic

Can you explain this? Do you just keep removing the dog from the situation or do you actually teach the dog acceptable behavior (by whatever method)? I think the latter would be the goal.

 

J.

 

You could do either. If every time the JRT goes after the kitten, it gets removed from the situation, it should learn, a. whenever I chase the kitten, I get removed from where I want to be (with my family) so b. If I want to stay with my family, I shouldn't chase the kitten.

 

I would say if you do that, that is training using punishment or a "time out." However, if you let the dog get into that situation, it is your fault, not the dog's or kitten's, but you don't have many options at that point.

 

The latter is the goal, so if that situation comes up, you should know to not put the dog in that situation again, until the dog is properly trained, or can be controlled enough to train.

 

I was mainly using this example, because I was confused why someone in that situation wouldn't just remove the dog? Why put the kitten through trauma, and let the dog be reinforced for inappropriate behavior? Just because I use positive training methods doesn't mean I don't make mistakes, and it definitely doesn't mean because I made those mistakes, I'm going to just let the dog/kitten/whatever suffer because I'm too stubborn to admit it. When I make mistakes, I do whatever I have to do to correct that at the time, then figure out what I can do better next time.

 

Autumn

Link to comment
Share on other sites

Personally, in that situation, I think the kitten's safety should trump the dog's feelings. A verbal correction AND removal from the area is what I would do. Kind of like a little kid drawing with crayons on the wall. Simply removing the child from the room would not stop him from coming back and doing it again. Expressing your displeasure AND stopping the activity seems more effective to me. Just my 2 cents :rolleyes:

Link to comment
Share on other sites

In that instance, saying “Jump!” would presumably have the effect of letting the dog know that the tunnel was not what was wanted, as he seems to think, but that the jump is what’s wanted. So although you are not verbally saying “no,” the “no” is implied, since he is in fact barreling toward the tunnel. It also lets him know that what you do want is the jump.

 

I definitely still have a completely different view of the situation.

 

Maybe the dog understands the unspoken implication and maybe he doesn't. In any case, I don't see any type of correction whatsoever in this scenario. Even if he understands, "I'm not taking the tunnel now, I'm taking the jump", it's more of a change of game plan than a correction.

 

Ok, understood, but again, unless the dog is doing something other than waiting politely at the door, you have no need to teach it to wait. You can reinforce the sitting and waiting with a treat, but there is no guarantee that the dog will not think he received the treat because his tail happened to be draped over his right forefoot as he sat there – especially if that tail happened to move at the moment you clicked or popped the treat into his mouth.

 

In practical terms, it is very easy to convey that sitting is the behavior that is being reinforced. If the click is happening as the rump hits the floor after the dog hears the word "sit" (of course, I teach sit before I teach sit at the door), it doesn't take a dog who understands the meaning of the click long to know what is being reinforced here.

 

And, honestly, if my dog learns to sit at the door and wag his tail while he's waiting, that's perfectly fine. Shoot - if I could get a tail wag on cue, I'd consider it more valuable than a sit and wait at the door! In practice I've found that capturing a tail wag can be quite challenging. If I got one by mistake like that, it would be like training gold!!

 

If I am really concerned that the dog does not understand that sitting is the correct behavior, I can play several games to make it clear. I could cue different behaviors and reward each with food, but the door only opens when the dog holds a sit stay, for instance. Another one might be to teach the dog to back away from the door and sit.

 

On the other hand, if he is jumping up on the door and you say “Wait!” and he responds by sitting, and you click/treat, he will very likely make the connection that jumping was not what was wanted, and sitting was what was wanted.

 

In this example, I'm not seeing any correction. You give an unknown cue "wait" and wait to see what the dog offers. He offers jumping. That doesn't work. He offers a sit - jackpot! So, he learns after a few reps that "wait" means "sit at the door". (Or, he might learn that "wait" means start to offer behaviors!! - But let's just say that doesn't happen in this case!)

 

Where you're seeing "not jumping", I'm seeing "polite sit and wait". Where you are seeing "wait" as a correction that is telling him not to jump, I'm seeing "wait" as an unknown cue that the dog must offer behaviors to figure out the meaning of.

 

But in the case where the dog is exhibiting an unwanted/ incorrect behavior, you can clarify by issuing a correction, and then illustrate or ask for an alternate behavior. In that case the dog is learning “this, not that.” The type/severity of correction is dictated by the situation. If the dog fails to sit, a soft “No.” is probably sufficient. No pain, no anxiety.

 

That's one option.

 

Another is to clarify by either showing the dog what you want, or giving the dog to figure out on his own what you want (through shaping, perhaps), and reinforce it to teach the dog what is desired.

 

Whether or not that is done in one quick session, or if it requires additional training (usually in the case of a dog that does not have a solid foundation in reinforcement yet) is dictated by the situation and the needs of the dog.

 

If the dog needs more work on sit, the criteria is lowered and the rate of reinforcement is raised until the dog knows it well enough to transfer it to a more difficult situation.

 

And for some, this just makes more sense, just as your choice makes more sense to you.

 

Here I think that it may be safely said that our definition of "correction" is somewhat different - at least in some situations.

 

Yes, I would agree. I must say that I find your definition surprising. It is not one that I'd heard before.

Link to comment
Share on other sites

Hopefully this does not confuse more people, but these are my definitions for the words being used to describe communication to dogs while training.

 

Praise/Click: Positive, “YES! Do that.”

 

Correction: Negative, “No! Don’t do that.” Can be followed by praise when the dog does do what you want, or the hopes that the dog will just learn you don’t want them to do THAT behavior. (For example, giving a leash correction when the dog is not in proper position, and not giving a leash correction when they are.)

 

No reward marker: Neutral, “That’s not what I want, but please try something else. Nothing bad will happen if you try.” Followed by praise or another NRM until the dog figures out exactly what you want.

 

Positive interrupter: Positive, “When I do this, come here and good things happen.”

 

Redirective: Neutral, “Do this command, no matter what you were doing before.”

 

Doing nothing when the dog does something “wrong”: Neutral. A lack of communication/directive. Can extinguish unwanted behavior by not rewarding the dog in any way. (Negative attention is still attention)

 

Hopefully I didn't leave anything out.

 

Also, in case I did confuse anyone. I consider myself a positive trainer meaning positive reinforcement is my ultimate goal and first choice. If I can't figure out how to work that into a particular scenario, I will use other things, such as verbal correction/pressure, time-outs and no reward markers, until I can figure out (if possible) how to do it a positive way. I will not let bad behavior/habits continue just because I can't figure out how to communicate in a way that I want. Maybe that makes me a hypocrite or not a "true" positive trainer, but I would disagree. No one is perfect in any form of training they choose.

 

Autumn

Link to comment
Share on other sites

Where you're seeing "not jumping", I'm seeing "polite sit and wait". Where you are seeing "wait" as a correction that is telling him not to jump, I'm seeing "wait" as an unknown cue that the dog must offer behaviors to figure out the meaning of.

 

Actually, I'm seeing both of those things happening at once. The dog is politely sitting and waiting, which also means not jumping in this case.

 

That's one option.

 

Another is to clarify by either showing the dog what you want, or giving the dog to figure out on his own what you want (through shaping, perhaps), and reinforce it to teach the dog what is desired.

 

Which is what I said - " you can clarify/reset by issuing a correction, and then illustrate or ask for an alternate behavior." Naturally, when the dog gets it right he gets rewarded.

 

 

Whether or not that is done in one quick session, or if it requires additional training (usually in the case of a dog that does not have a solid foundation in reinforcement yet) is dictated by the situation and the needs of the dog.

 

agreed

 

If the dog needs more work on sit, the criteria is lowered and the rate of reinforcement is raised until the dog knows it well enough to transfer it to a more difficult situation.

 

I'm fine with this too.

 

And for some, this just makes more sense, just as your choice makes more sense to you.

 

Well, we each do what works for us, and falls within our definition of fair & humane.

 

Yes, I would agree. I must say that I find your definition surprising. It is not one that I'd heard before.

 

It doesn't seem like anything new to me. It was how I was taught, modified by my experience with various dogs. It's all evolving all the time - learn new stuff - plug it in where it's indicated. I'm sure it's the same for you.

Link to comment
Share on other sites

our kitty moment this morning....

 

I let the sheep out to graze with Lilly the lgd. DH came out of the house, Lilly runs up to greet DH and with that she's within "playing" range of the chickens. She wants to play but it ends up playing dead chicken.

So, I holler for Lil, not in a neg. way just in a come here way. But she is not formally trained in anything but willing to do anything to please. She honestly didn't hear or registar my call while in chicken mode.

The CHicken game is way to intense for her to even hear me. I run up, grab an old piece of pvc pipe on my way up. I get there in time to stop the game from going to dead chicken. I could of clicked treated or anything positive till the cows came home or the chicken died and Lil wouldn't of heard me. She was not disobeying me, she didn't hear me as her brain was in dead chicken mode.

I wasn't mad, but I had to make the choice to either correct her or let her kill the chicken. There was no manhandling (or woman handling) the chicken out of her mouth in a positive manner. I didn't "hurt" her more than her feelings but it took physical connection for me to hurt her feelings and I really mean hurt her feelings not her. She is a huge tough dog, would take more than me to really hurt her unless I had something quite hard or a gun. But once she is listening she is the most willing to please dog I know.

We didn't end our day mad or upset. She ended her game by choosing to come with me and got rewarded with a treat (believe it or not, it was an egg) once I had her in the sheep pen and the chickens were safe.

 

It ended well, chicken lived, Lily realized at that moment I was not a happy camper with her game and decided it was time to do what I asked.

 

What would the positive people suggest I do if it happens again. Of course the ideal would be for the situation not to present it self but it wasn't intended in the first place so likely it will happen again as stuff happens on the farm.

 

Just wondering

Link to comment
Share on other sites

Even if he understands, "I'm not taking the tunnel now, I'm taking the jump", it's more of a change of game plan than a correction.

 

So, now the "directive" is a "change of game plan"?

 

And to nitpick further because I think that the distinction is important, it is a change of the dog's game plan, not yours. Your game plan has always been for the dog to take the jump.

 

Before I dismount from this merry-go-round, the dictionary definations of "correction" may be useful:

 

http://www.thefreedictionary.com/correction

cor·rec·tion (k-rkshn)

n.

1. The act or process of correcting.

2. Something offered or substituted for a mistake or fault: made corrections in the report.

3.

a. Punishment intended to rehabilitate or improve.

b. corrections The treatment of offenders through a system of penal incarceration, rehabilitation, probation, and parole, or the administrative system by which these are effectuated.

4. An amount or quantity added or subtracted in order to correct.

5. A decline in stock-market activity or prices following a period of increases.

 

And the word, "correct": http://www.thefreedictionary.com/correct

 

cor·rect (k-rkt)

v. cor·rect·ed, cor·rect·ing, cor·rects

v.tr.

1.

a. To remove the errors or mistakes from.

b. To indicate or mark the errors in.

2. To punish for the purpose of improving or reforming.

3. To remove, remedy, or counteract (a malfunction, for example).

4. To adjust so as to meet a required standard or condition: correct the wheel alignment on a car.

v.intr.

1. To make corrections.

2. To make adjustments; compensate: correcting for the effects of air resistance.

adj.

1. Free from error or fault; true or accurate.

2. Conforming to standards; proper: correct behavior.

Link to comment
Share on other sites

You really don't know what Geonni's approach would be with a scared dog. It's just as likely that she would have recognized that the dog was scared and done something to ease the dog's fear as that she would have just plunked it into the tub despite its fear.

J.

 

Thank you! I was describing a dog with the "Ewww! A bath!" attitude. Of course if the dog is frightened of the bath we do things very differently!

Link to comment
Share on other sites

Thank you! I was describing a dog with the "Ewww! A bath!" attitude. Of course if the dog is frightened of the bath we do things very differently!

 

And see that's the problem with a thread like this, everyone is so defensive that no matter what is typed you can read it as accusatory. I said that Geonni may have used a different approach based on something she posted earlier, but that was simply an example. Of course I don't know her, and I can't read her mind. My point was many people would feel that shaping the dog to jump in the was a waste of time because they could have just dropped Kobe in and that would be that, and that would have been an acceptable answer and certainly not abusive, just a different plan. See, even when I go *out of my way* to be nice and agree with people I get jumped on!

 

I just don't think we can meet at an agreement here, I think there are too many of us speaking different languages. Its so very frustrating. No one want st be told that what they are doing is wrong, so its too easy to be if you disagree with any post of point to feel defensive.

 

My point in starting the thread was to answer the question: if using corrections work for an OTCh trainer, why would I choose to do something different? The answer is because for me, it works better and results in a better relationship with my dogs. This does not mean that other people may have a different experience.

 

And with that, I am done here..except I would be interested in hearing what kind of a scientist Blackdawgs is and what kind of a lab he or she worked in based on the comments made about learning theory.

Link to comment
Share on other sites

So, now the "directive" is a "change of game plan"?

 

And to nitpick further because I think that the distinction is important, it is a change of the dog's game plan, not yours. Your game plan has always been for the dog to take the jump.

 

Actually, once the dog is heading toward the tunnel, my game plan has now changed. Now it is "direct dog from where he is at this moment to where he needs to be next." If the dog's game plan changes, mine has changed. I have to make a new decision based on the dog's previous choice.

 

Now to nitpick a bit myself, based on the definition that you provided, do you consider any and every cue (directive) given to a dog in any and every circumstance to be a correction?

 

Example: I have walked into the ring with my dog and he is standing next to me. I tell him to sit and he does. Do you consider that "sit" cue to be a correction?

Link to comment
Share on other sites

Geez Louise! :rolleyes: Why the quotation marks?

 

Do you believe there are mild corrections, or only (so-called, but not really) "mild" corrections?

 

The quotation marks were intended to emphasize the word. Perhaps putting mild in italics would have been more appropriate.

 

Of course there are mild corrections. There are harsh corrections, mild corrections, and everything in between. Conversely, there is high value reinforcement, pretty darn skimpy reinforcement, and everything in between.

Link to comment
Share on other sites

Well, we each do what works for us, and falls within our definition of fair & humane.

 

Absolutely! On that we can definitely agree.

 

It doesn't seem like anything new to me. It was how I was taught, modified by my experience with various dogs. It's all evolving all the time - learn new stuff - plug it in where it's indicated. I'm sure it's the same for you.

 

True. I just learned recently about how to make cues into very strong reinforcers and I've barely begun to scratch the surface of using this in my training.

 

I suppose that if cues can become powerful reinforcers, they can become corrections, too, for those who choose to use them in that way. I'm definitely going to go the cues-as-reinforcers route. I'm pretty excited to see where that will lead.

Link to comment
Share on other sites

I just don't think we can meet at an agreement here, I think there are too many of us speaking different languages. Its so very frustrating. No one want st be told that what they are doing is wrong, so its too easy to be if you disagree with any post of point to feel defensive.

Yep, and so you probably understand the frustration some of us feel when we say we use corrections and it's implied over and over again, ad nauseum, that we're abusive and causing pain and fear in our dogs when there's a better way (or ways) that we are apparently too obtuse to understand or just plain unwilling to understand or attempt. The funny thing to me is that most of us who use corrections do so in very specific contexts and actually use reward-based training in other contexts. No one on the "other side" in these discussions ever seems to acknowledge that at least some of us do understand reward-based training and do use it where it's appropriate to do so. For some of you, it's always appropriate. For some of us, the reality of farm life might mean that it's not always appropriate.

 

In the example of the JRT and the kitten, I would first take into consideration the high prey drive of a terrier. I would also take into account the prey drive and personality of *that particular terrier.* And I would take into account whether in my household I could always be sure that I would be there to control interactions between the dog and the kitten. I would certainly have picked up the dog and removed it from the situation, but before doing that I would have let the dog know in no uncertain terms (and those terms don't have to cause pain) that I was *very unhappy* with dog's behavior. For me, it makes sense to be very explicit in situations where a life is at stake. Sure the dog might eventually figure out "if I bother the cat, I get put somewhere else." For me, I would want the dog to make that connection very quickly, so my correction would be loud and instantaneous. Then the dog could go somewhere else and think about it.

 

I had a similar discussion with someone who came out here with her young dog. Every time the dog showed interest in the chickens, she gave the dog a redirect command so that the dog would watch her instead of the chickens. My question to her was "What happens if the dog is in the yard with the chickens and you're not there to give the dog that command?" The reality, of course, is that in that dog's normal life, it would never likely be in a yard with chickens without the owner present. But I expect my dogs to share the yard with my free-ranging chickens without my supervision. So I need to come up with a training method where the dog understands that "don't bother the chickens" means just that whether I am present or not. I think this is where directives (correction) is useful. If the dog starts to stare at, stalk, or chase the chickens I can say "no!" or "ah, ah!" and the dog understands that the behavior it was engaged in at the time of the correction is unacceptable to me. The dog should be able to extrapolate that information to apply even when I am not standing right there watching because the information I gave it wasn't "do something else that involves me" but simply "don't do that, period."

 

Blackdawgs posted the dictionary definitions of correct and correction and I don't think any of those accurately described what is meant by a correction, at least in the sense that I use for stockdog training. As someone else said, a correction can be a simple verbal directive to a dog, or it can be body pressure placed on the dog (that is, using my body to block the dog's intended route), or it could be a bop on the nose for a dog that's grabbing a sheep. Note that in the latter case, a correction might indeed cause pain or discomfort to the dog, and that's why such corrections are reserved for the most egregious mistakes. In fact, if the dog grabs a sheep in the act of trying to catch and turn said sheep, I'm not likely to correct the dog at all, but if the dog is just in there grabbing and slashing for no good reason, then the dog most certainly will get a physical correction. Of course the one variable in these scenarios that doesn't occur with agility training, pet manners training, and the like is the fact of the stock--another sentient being that also feels pain and experiences fear. They have as much right not to have undue pain and fear caused by a dog as the dog does to not have those things caused by a human. But of course the dog needs to get trained in order to be useful, so there has to be compromise in that regard.

 

J.

Link to comment
Share on other sites

Kristine,

What Blackdawgs said in the "definitions" is what I was referring to earlier.

 

In my mind, redirection is a correction in that it alters one course of action and replaces it with another.

 

Example...

Bernie likes to jump up on people and I do not want him to do that.

When he meets someone new, in his usual over-exhuberance, I know that he wants to go over to that person and jump up on them.

Knowing that, I will tell Bernie to sit. Not "Bernie, No, Sit" just "Sit".

To me, that is a correction and if I am reading what you have written correctly, you would say that is a redirection when for either or both of us, it is a distinction without a difference and what we are really discussing are simple semantics.

 

It is the same as if driving a car and you start to veer to the left out of your lane. I may call it a (course) correction and you may call it redirection, but the efforts taken to return the car to its proper lane are the same, you move the wheel to the right until you are where you want the car to be. Now if we were looking at "no corrections" until the desired result was achieved, we could both be part of an on-coming semi.

 

Certainly I am not questioning, knocking or attempting to trivialize your methods. It works for you and it causes no harm to you or the dog. On that front, I say more power to you.

 

ETA,

While I was posting, you posted in part...

"Example: I have walked into the ring with my dog and he is standing next to me. I tell him to sit and he does. Do you consider that "sit" cue to be a correction?"

 

If the dog is standing next to you, no. If the dog is straying off or perhaps because you know your dog and it is staring at that Aussie you know it hates and would just love to rip its throat out and could be seriously contemplating that action, then yes.

Link to comment
Share on other sites

I just learned recently about how to make cues into very strong reinforcers and I've barely begun to scratch the surface of using this in my training.

 

I suppose that if cues can become powerful reinforcers, they can become corrections, too, for those who choose to use them in that way.

 

Sheepdoggers do this all the time. For example, one well-known and highly respected trainer advocates giving a command (or cue, as you would say) -- "Lie down," for example -- in a commanding voice, and then when the dog does it, repeating it in a sweet, praising voice, rather than saying something different like "good dog," "good job," "thank you," "yay," etc. as I usually hear sport and pet trainers do. Conversely, he would repeat the lie down in quite a harsh voice if the dog ignored the command (which of course would be anathema to those trainers who believe you should never repeat a command), again followed by the same phrase in the praising voice when the dog obeys. The phrase thus performs three functions -- command, correction, reward -- depending on how it is said.

Link to comment
Share on other sites

Sheepdoggers do this all the time. For example, one well-known and highly respected trainer advocates giving a command (or cue, as you would say) -- "Lie down," for example -- in a commanding voice, and then when the dog does it, repeating it in a sweet, praising voice, rather than saying something different like "good dog," "good job," "thank you," "yay," etc. as I usually hear sport and pet trainers do. Conversely, he would repeat the lie down in quite a harsh voice if the dog ignored the command (which of course would be anathema to those trainers who believe you should never repeat a command), again followed by the same phrase in the praising voice when the dog obeys. The phrase thus performs three functions -- command, correction, reward -- depending on how it is said.

 

It's a little bit different, though. It's not the tone of voice that makes the cue reinforcing, but the actual cue itself. It could be said in a neutral voice, an excited voice, encouraging voice, etc., but the cue itself, and the meaning behind it becomes something that is highly valuable to the dog. A cue that had become a reinforcer would not be said in a harsh or commanding voice. That would be like saying "good dog" in a harsh or commanding voice.

 

Tone of voice has a place, but it's not exactly the same thing.

Link to comment
Share on other sites

It's a little bit different, though. It's not the tone of voice that makes the cue reinforcing, but the actual cue itself. It could be said in a neutral voice, an excited voice, encouraging voice, etc., but the cue itself, and the meaning behind it becomes something that is highly valuable to the dog. A cue that had become a reinforcer would not be said in a harsh or commanding voice. That would be like saying "good dog" in a harsh or commanding voice.

 

Tone of voice has a place, but it's not exactly the same thing.

 

wait... Huh? I don't get it. It seems to me that the tone of voice is the whole deal. It refers back to the original command, yes, but the tone makes a different statement about the command each time.

 

Like if you tried to share a bowl of pork rinds with some folks and said in your best, look what a great treat I have to share with you! voice, "Pork rinds!"

Then your vegetarian says in a voice tinged with horror, "Pork rinds?"

And your Manwich-eating junk-food junkie, with covetous eyes aglow said, PORK RINDS!

 

It's the same word all three times, but the emotion conveyed with each is totally different.

 

One is an announcement, the command. One is a condemnation, the rebuke. And one is an expression of delight, the praise.

Link to comment
Share on other sites

I see a lot of "them" and "me" in this thread. To make gross generalizations, many individuals belong to the "agility/obedience" group, another group belongs to the "stockdog" group. I think there may be middle ground to be found. At least I hope so. Perhaps if I offer an example that neither involves stockwork, nor involves "agility/obedience" training, it'll give some of you who view yourselves on one side a chance to offer me suggestions?

 

Me, I view myself as a "positive" trainer (I use "treats" when I feel it's appropriate) who also uses (mild) "corrections" (again, when I view them as appropriate) for my 16-month-old adolescent dog. (I classify myself as a "pear", if you will). To wit: when training certain behaviors in my pet Border collie (like "go to your bed", "check in with me while on off-leash walks"), I use treats. For "expected behaviors" (lie down, sit, stay, wait), that are pretty entrenched, I use nothing - at most a "good dog!" if I recognize that it required some self-control to obey. If there's some wavering - such as a dog starting to shove its way to the door in front of me, then I use a "hey, get back!" to remind them of what I expect as good manners.

 

So here's the scenario I recently found myself in. It's neither "agility/obedience", nor is it "stockwork". Nor does it involve "dog instinct" (as in a terrier vs a kitten). If someone can suggest a way of training this - without using any (verbal or otherwise) "corrections", I'd love to hear it. Especially if I'm likely to find it of use within the sort of time constraints I'm facing at present.

 

I'm spending 4 weeks on a 250-acre island in NH. Island is wooded. Only 12 houses, so sparsely settled; each property is unfenced. (Idyllic? definitely). I'm taking care of a FIL with Parkinsons and Alzheimers. Also trying to write a critically important proposal (this is NOT my "vacation time"; thank goodness for the Internet). All this to explain why I don't have unlimited time at my disposal to wander outside to train the dog.

 

Dog is doing well at hanging around the boundaries of the house. He thinks he's in heaven because the water is at our doorstep; there are squirrels and chipmunks scolding him right and left; he gets to go for walks each day in the woods; and his human brother, whom he adores, will play with him when his human "mom" is busy. Everything is going great until a "neighbor" cuts through the property with her dog. My dog wants to tag along after them. He suddenly gets it into his mind that following them along the trail is a Good Idea.

 

I don't want said dog to Get Lost. Nor do I want him to get Run Over by the few (but non-negligible) vehicles on the island. Nor do I want him to cross the bridge onto the mainland. He's still an adolescent, remember. No "car sense". And this is not his regular home, so I can't expect "homing sense".

 

Options: (1) put him on a tie-out; (2) keep him indoors all day (he LOVES being outdoors; I can keep a semi-close eye on him by working on the screened porch - but only so long as he stays at the front of the house); (3) train him that there are "boundaries" he should respect.

 

I opted for (3). Took him out and said "let's PLAY!" He ran off behind the house in the direction of the dog he'd seen. I waited until he crossed behind the house and said "Hey, GIT out of there!!!". He came back to me immediately. We did this a few more times, in slightly different circumstances, until it seemed to me that he had learned that he shouldn't cross this invisible line in my absence. (He's smart; it didn't take more than three corrections for him to recognize an invisible line that he Would Not Cross). For several days thereafter I've been able to work on the porch and he hasn't wandered an inch.

 

I don't believe I've crushed his spirit by letting him know that there was a line he shouldn't cross, simply by using my voice. He flattened his ears a bit and wriggled coming back to me, but as soon as I praised him for heading back in my direction, he knew he was once again treading on the path of righteousness, and his body language showed that to me.

 

I'd welcome specific suggestions for how to train him NOT to cross that invisible line (and wander), while not using some sort of correction. Or would all of the "never use corrections" crowd opt for tying him out? Personally I'd rather have him exerting impulse control - once he knew what impulses I wanted him to control.

Link to comment
Share on other sites

Forget purely positive or correction-based methods... there's always this tried and true way of training.

 

:rolleyes::D

 

...I kid :-P

 

Hee! That's funny...

 

But you know, sometimes it does kinda work... My no-longer-poop-eating dog used to get this kind of treatment from me when she was caught "scarfing a loaf." I would say "That's dis-GUS-ting!" and when she came to placate me I'd draw away and say "Eeeeeew! Don't touch me!"

When I started doing this regularly she started looking all shame-faced. Later she would glance longingly back at the treasured objects... and forgo them. Eventually even the longing glances stopped. Now, to be fair, she may have simply grown out of it, or when her condition started to improve, (she was underfed, ill, and full of parasites when I adopted her.) she might simply have lost her taste for her own stools.

Link to comment
Share on other sites

Sheepdoggers do this all the time. For example, one well-known and highly respected trainer advocates giving a command (or cue, as you would say) -- "Lie down," for example -- in a commanding voice, and then when the dog does it, repeating it in a sweet, praising voice, rather than saying something different like "good dog," "good job," "thank you," "yay," etc. as I usually hear sport and pet trainers do. Conversely, he would repeat the lie down in quite a harsh voice if the dog ignored the command (which of course would be anathema to those trainers who believe you should never repeat a command), again followed by the same phrase in the praising voice when the dog obeys. The phrase thus performs three functions -- command, correction, reward -- depending on how it is said.

 

Good stockdog people always use positive training techniques. It's the foundation for teaching young dogs.

 

If the dog exhibits the correct behavior, it is rewarded by being given control of the stock. If the dog does not exhibit the correct behavior, it is not given control of the stock (neutral).

 

The same trainer Eileen mentioned (if he is the one I am thinking about) also modified the old method of pushing a dog who is tight on its flanks out by putting pressure on the dog. He advocates directing the "pressure" at the point on the ground where the dog is turning in prematurely. The concept is easier for beginners not to screw up.

 

<Heresy> It's like clicker training </Heresy>

 

The problem a lot of beginners get into with training stock dogs is that the "pressure" or the "correction" comes too late and is applied at the wrong point. Putting the pressure on a non-moving target (the ground) is less like a "correction" and more like a "reinforcement". The "marker" is allowing the dog to have the sheep. Another great stockdog trainer is always saying; "make the wrong behavior difficult and the correct behavior easy, and the dog will almost always choose the correct behavior". Sounds pretty positive to me.

 

Following this discussion, I don't see that there is no common ground. I see four things that are clouding the discussion:

 

1) People on either side are defensive because they take a question about the effectiveness of a given technique in a given circumstance as a personal attack. It's not. And, if you can't defend your position with reason, clarity, and specificity, then you either don't understand it fully or it's untenable. We can disagree and challenge one another's position without being disagreeable. Defending your position is the quickest way of discovering how much of it is defensible.

 

2) Semantics. The jargon is different. I believe that the concepts are more similar than one would think

 

3) Context. Positive-only methods (which, from what I've read here, aren't strictly positive only. The corrections are mild and are called something else) may work for some dogs, in some instances. They may be the best methods for some dogs in some situations. They may be ineffective and inappropriate for other dogs in other situations.

 

4) Bad trainers. Lots of people on the "positive only" side equate "correction" as physical correction or abuse because they've seen corrections applied inappropriately. Similarly, lots of people on the other side view the positive only approach as "voodoo" because they've seen it misapplied by people who don't really understand it and don't get the desired result from it.

 

Good practicioners on both sides can achieve the desired results and healthy happy dogs. Charlatans on both sides can ruin perfectly good dogs.

 

The gulf is not as wide as one might suppose from reading this thread, at least in my opinion, and both sides can learn something by listening to the best of the other approach with an open mind.

 

Pearse

Link to comment
Share on other sites

Good stockdog people always use positive training techniques. It's the foundation for teaching young dogs. . . .

 

Another great stockdog trainer is always saying; "make the wrong behavior difficult and the correct behavior easy, and the dog will almost always choose the correct behavior". Sounds pretty positive to me.

 

Following this discussion, I don't see that there is no common ground.

 

It's quite true that both sides use positive methods. I guess that could be said to be "common ground." But I think everybody recognizes that these are things we have in common.

 

The dispute is not about positive reinforcement, which both sides speak in favor of. The dispute is about corrections, which one side favors in addition to positive reinforcement and the other categorically opposes. I don't see much common ground on that point. But then, I don't expect much. :rolleyes:

 

The same trainer Eileen mentioned (if he is the one I am thinking about) also modified the old method of pushing a dog who is tight on its flanks out by putting pressure on the dog. He advocates directing the "pressure" at the point on the ground where the dog is turning in prematurely. The concept is easier for beginners not to screw up.

 

Yes, that's the trainer I was referring to, and I agree that the "dangerous ground" concept is a very valuable one, but I can't agree that it's "less like a 'correction' and more like a 'reinforcement'." To call it that seems to me a distortion of the terminology, if a well-meaning one. It is a correction -- that's just what it is.

 

[in stockdog training, I]f the dog exhibits the correct behavior, it is rewarded by being given control of the stock. If the dog does not exhibit the correct behavior, it is not given control of the stock (neutral). . . .

 

It's like clicker training

 

I really hate to see this formulation used, because clicker trainers who don't train in stockwork are all too ready to believe and assert that stockdog training is just like clicker training with sheep as the reward, and as you know it's just plain not as simple as that.

 

Yes, sheepdog trainers do talk in terms of "letting him have the sheep" and "not letting him have the sheep," and yes, it is sensible in many instances to let the dog continue to work the sheep when he is right and to stop him when he's wrong. (Technically, BTW, since the dog is already working with the sheep when the incorrect behavior occurs, under the rubric of behavioral learning theory what you're doing when you take away the sheep is actually negative punishment. Letting the dog continue on with the sheep is not positive reinforcement under that rubric; it's neutral, because nothing is being added and nothing is being taken away.) But unless you're working with holographic sheep you can't withhold that "reward" with the kind of meaningful timing you can a treat even in a small area, much less in a big field. And there are times when not letting the dog have the sheep when he's wrong would cause undesirable consequences. Hence the routine observation that beginners rely too much on the down -- if it were simply a matter of withholding "the reward" when the dog is wrong, then it would always be right to down him. Another example -- a dog who starts out wrong on his outrun can be stopped and called back, and thereby not allowed to have the sheep, but if you do that more than a very few times you risk creating other outrun problems. In short, I can think of quite a few incorrect things a dog in training may do (and I'd be surprised if you couldn't think of many also) that cannot be dealt with effectively, or most effectively, by simply giving or withholding control of the stock.

 

I would never deny that reinforcement (with intrinsic rewards, not extraneous ones) is a big part of stockdog training, but I myself would not be willing to say or imply that good stockdog training is "all positive," or that corrections are not a necessary part of it. Or that they are not very valuable in dog training quite apart from stockdog training, as Ooky's receiving-blanket example (post #47) indicates. And that's where the dispute centers.

 

But I'm all for disagreeing and challenging one another's opinions without being disagreeable. :D

Link to comment
Share on other sites

wait... Huh? I don't get it. It seems to me that the tone of voice is the whole deal. It refers back to the original command, yes, but the tone makes a different statement about the command each time.

 

So this is something different. I'm not making a statement about the cue. It is what it is.

 

Like if you tried to share a bowl of pork rinds with some folks and said in your best, look what a great treat I have to share with you! voice, "Pork rinds!"

Then your vegetarian says in a voice tinged with horror, "Pork rinds?"

And your Manwich-eating junk-food junkie, with covetous eyes aglow said, PORK RINDS!

 

This is actually a good example. The emotion that has changed is not that of the person offering the pork rinds, but the person who is potentially receiving the pork rinds.

 

When the cue becomes a reinforcer, the response that you get from the dog when the cue is given is like the Manwich-eating food person. I could offer him pork rinds in a neutral voice, in an encouraging voice, in an offering voice, or maybe even a slightly confused voice, but the response of the Manwich person is going to be more or less the same. He does not love pork rinds because I make them sound good, but because he knows that they are good.

 

It's not the tone of voice that gives the pork rinds value to the Manwich person, but the pork rinds themselves.

 

And it's not the tone of voice that the cue is given in that makes it reinforcing, it is the cue itself.

 

Very different.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...