Jump to content
BC Boards
Alaska

Before You Get Your Puppy

Recommended Posts

Problem is, I've come across some dogs who actaully won't jump up, or put their paws on anyone or anything because they were punished for it.

 

That doesn't mean that every single dog who is punished, or corrected, for putting paws up (on a counter, on the person, etc.) will end up like this. But I know it can happen and it's not a risk I'm willing to take.

 

Okay, certainly won't argue with you, it's your choice.

 

But you do see what I mean, don't you, about food being in the forefront of the dog's mind when he is forming the intent to countersurf? Sometimes it can be tricky to know exactly what connection the dog will make with a correction. If you correct him for putting his paws up on a fence, for example, he *might* think he shouldn't put his paws up on something else he might approach with the same thought in mind as the thought he had in mind when he put his paws up on the fence. (That's why I like to make the corrections mild, almost like part of a conversation, so that it's easy to clear up any misconceptions you may later find the dog has formed, and refine his understanding of exactly what's off limits.) But in the countersurfing case, you KNOW the dog is doing this to get food. He is thinking "food!" His nose almost looks prehensile in its fixation on the food. Food is his sole motivation; it's filling his mind at that moment. Therefore, you know he is going to connect the correction with the food, and not with non-food-connected places. Or so it seems to me.

Share this post


Link to post
Share on other sites
Okay, certainly won't argue with you, it's your choice.

 

But you do see what I mean, don't you, about food being in the forefront of the dog's mind when he is forming the intent to countersurf? Sometimes it can be tricky to know exactly what connection the dog will make with a correction.

 

True. I see misplaced understanding of what the correction was actually for happen fairly regularly while out and about in training classes with people who use them in training.

 

Granted, misplaced reinforcement can foster unwanted behaviors, so there is some risk involved in training, no matter how it's done. Poor timing, misapplication, etc. can create problems no matter how one is going about doing the training.

 

While I would agree that the food is probably in the forefront of the dog's mind when he is intending to countersurf, I still consider the risks of correction to be much higher than the potential benefits, and I consider the risks of reinforcement to be much lower than the potential benefits. Again, this is my own choice. I am not saying that your conclusion would be the same.

 

The paw thing was certainly not my only reason for choosing to use a reinforcement based approach. It was highly efficient to incorporate other lessons into the counter training. Since this was one of the earliest pieces of training that I worked on with my former surfer, it was the perfect time to teach some foundation behaviors that have come in very handy in both sport training and real life things.

 

And then there's my plain and simple personal preference.

Share this post


Link to post
Share on other sites
That was my point. He has learned what corrections are (usually just words or a particular tone), suggestions are, directives are, and so on, through working with me both on and off stock. It's about communication above all else and as Eileen, I think, noted, [ETA: looking back on the thread, I see that it was Denise who made this comment, and not Eileen, just for clarification] communication that ignores part of the richness of what's available *to communicate* (good tones and bad tones) is a bit stilted. I don't have to tell him necessarily to do a particular thing (as in the case of the lamb) but just give him a quick reminder to have his "thinking cap" on and then let him do what he thinks is right. Is he always right? No, but that's where training (more communication) comes in. For example, I also use him to hold the main flock off the feeders in the mornings (it's good for building his confidence for driving since he has to push sheep away from someplace they very much want to be). I don't say anything to him as we're walking out into the pasture because he knows what the job is. When I'm done pouring feed, I give him a flank around the sheep so he's no longer blocking them and let him bring them to me. Lately he hasn't wanted to take that flank to release the sheep. I can't very well take the sheep from him (they're not nearby because he's pushed them well away--necessary for him to be able to cover them adequately without anyone breaking past on the sides, and I can't physically correct him, so I have really nothing to use but my *tone of voice* to tell him that he's not doing what I wish--this is why I cringe at all the punishment comments--usually the dog isn't anywhere near enough to me for me to punish it if I wanted to, and I really don't bother with punishment after the fact as I think it's a waste of time. When he does comply (is obedient), the sheep are released and I will give him a command in a happier tone (walk up, or perhaps that'll do) to convey to him that I am pleased with him.

 

ETA: I don't think the bulk of learning comes from watching me work with the other dogs. The one exception is that I think pups very quickly learn a recall command or whistle by joining the pack when it complies with those commands. I've never had to specifically train a recall whistle on any youngster--they just get it from doing what the "big dogs" do in response to such a whistle.

 

J.

 

Thanks fr your detailed reply, but I feel like we are talking about 2 different things.

 

I understood your question to be: if I don't train my dog with corrections, how would he know my tone meant I was unhappy?

 

Based on that, my suggestion is your dog knows when you are unhappy by your tone just from living with you, not necessarily because it was preceded or accompanied by a correction.

Share this post


Link to post
Share on other sites

Bwahahahahahahahaah! You said "prehensile nose"... I'm picturing that thing that horses do with their upper lip when they're begging. Too funny.

 

This is a fascinating thread, BTW.

Share this post


Link to post
Share on other sites

WOW...

 

First off, I train my dogs on stock, trialed in USBCHA trials, and we also do some contract work for local farms.

 

I also train and trial my dogs in agility, where they do quite well.

 

I think people on EITHER end of extremes are to be avoided. Just my opinion of course. Extremists are just that...extreme.

 

My dogs understand corrections as a form of communication. Think of it as a game of hot and cold. A positive reward is "you're getting hot! Smoking hot!" and a correction is "you're FREEZING. Nothing more.

 

I have to admit I was more on the positive everything side of the training debate until I witnessed the amazing communication between a very talented stockdog clinician and her dogs. They were happy! They understood what was expected of them! They LOVED their owner! And they were well behaved, well adjusted dogs. And yes, they were corrected.

 

And that changed all things for me. I think we don't give dogs credit for their extreme intelligence, their resourcefulness and their ability to understand humans.

 

I had a dog that was definetly a very hard dog to train on stock...but after learning how people train stockdogs, she turned into a completely different dog, on stock and in life. Not one time was she abused, hit, or in anyway harmed. But she was communicated with. In tones she understood, with pressures she understood and using her instinct to let her figure things out.

 

I never understood how the word "lie down" could mean 10 different things...but now, I do.

 

I tell everyone that comes to me for lessons, if you are not into any kind of correction, go watch a stockdog clinic or trial. Watch the amazing communication between the handler and dog. I never had that with my dogs until I learned how to train a dog on stock. We communicate with each other using corrections...why do we think a dog is too sensitive or stupid to do the same?

 

Just my 2 cents...which is about all it is probably worth :rolleyes:

 

Loretta

Share this post


Link to post
Share on other sites
I have to admit I was more on the positive everything side of the training debate until I witnessed the amazing communication between a very talented stockdog clinician and her dogs. They were happy! They understood what was expected of them! They LOVED their owner! And they were well behaved, well adjusted dogs. And yes, they were corrected.

 

And that changed all things for me. I think we don't give dogs credit for their extreme intelligence, their resourcefulness and their ability to understand humans.

 

I'd say this is roughly my experience too. It took me a long time to understand the actual consequences of accepting that behaviorism is a theory of mind (or non-mind)--rather than just a theory of training and/or learning-- and that as such, it misses a lot of opportunities to harness the cognitive processing that is involved in training (of dogs or other species).

 

I think there is plenty you can teach using principles linked to behaviorism, but for me there came a point (and, not to overly romanticize it, but that point came by watching dogs trained on stock--and I have since also seen it in a couple of agility trainers and their dogs), where what I take to be fundamentally incorrect assumptions about the actual minds of those being trained were too glaring for me to overlook.

Share this post


Link to post
Share on other sites
I understood your question to be: if I don't train my dog with corrections, how would he know my tone meant I was unhappy?

 

My question was a bit more subtle than that, I think. This morning I wasn't telling Pip I was *unhappy* because I wasn't unhappy. I was telling him to take care because if he didn't take care, then I could potentially be unhappy. Do you see the difference? Calling his name in a certain tone meant to him to think and be careful; it wasn't a corrective tone that said "don't do that" because he hadn't actually done anything wrong. It was a cautionary tone, and that's why I labeled it a "sort of" correction, for lack of a better thing to call it.

 

Based on that, my suggestion is your dog knows when you are unhappy by your tone just from living with you, not necessarily because it was preceded or accompanied by a correction.

Sure my dogs understand what my tones mean just from living with me, because I deliberately use those sorts of tones when training and in everyday living so that they *will* understand them. BUT, what I think you're missing in this discussion is that if the tone is meant to be a correction then the tone itself *is* the correction. There is no preceding or accompanying *other* sort of correction. I get the feeling your envisioning something where I might say "Hey!" in a "don't do that voice" and then follow up or accompany that with a physical correction, but that's not the case at all. So if I want Pip to be more thoughtful in his actions, I might say "Piiiip" in a drawn-out ascending tone, whereas if I want to correct him I might say "Pip!" in a short, harsh tone. In both cases I used his name (not saying this is always how I do it, but just using it as an example). In the former, he should give me a sort of "What?" response, which actually involves checking his actions (whatever they may be at the time) and becoming more thoughtful in those actions, and in the latter he should stop entirely what he's doing and come up with an alternate response (a Plan B, if you will). So in the case of the lamb if I see him jump up in a hard flank meant to catch the lamb on the head, my "Piiiip" will convey the message to him to slow down and be careful and not bully the lamb. If in a similar situation a ewe were trying to barrel past him, her sheer size alone would mean that Pip would have to be more forceful and I wouldn't say anything at all (unless I thought he was being unfair to the ewe). When I'm using my voice like this, the livestock are also doing their own thing, so there has to be leeway for him to read and respond to the actual behavior of the livestock with my voice just adding to his information gathering, so to speak. I hope that makes sense.

 

J.

Share this post


Link to post
Share on other sites

Time to reply with a book:

 

Julie,

 

Excellent post. You've stated very eloquently what I have been trying to say.

That was all I was looking for. Not avoiding the situation altogether. Not manipulation the environment around the dog so the dog never has to face adversity. Not waiting for the dog to offer the correct behavior. Just clear communication. Period. Call it negative. Call it whatever you want.

 

Can I clarify a few things? Its like people are speaking 2 completely different languages here.

I hate the term “purely positive trainers” because its meaningless and almost never true. I train with a conditioned reinforce (a clicker) and attempt to train by rewarding behaviors (R+ as much as I can; P- as well) I like as opposed to using R- and P+. I have trained using R- and P+ in the past and it does work, but for a multitude of reasons I will not address here (a different topic for a different day). This doesn’t make me “positive,” it means I train using R+ and P- more than R- and P+. I train with as much positive reinforcement as makes sense.

“Not manipulation the environment around the dog so the dog never has to face adversity.”

This is not a component of training with positive reinforcement at ALL.

Yes, when a dog is* first* learning a new skill, you manage the environment so the dog doesn't have an opportunity to be reinforced for a behavior you don't want him to be reinforced for.

Ex: you don't leave an untrained dog in a room with food on the counter if you don’t want him to counter-surf: you clean the counters if you can't be there. If he gets on the counter in your absence and gets the, there is a history of being positively reinforced for counter-surfing (woooo! Beef!) and any behavior that is reinforced is more inclined to be repeated. You can then work with the dog to train him to keep feet and noses off , and even do “set ups” to tempt him and then show him a different way to get what he wants. In addition, when training this you would start with a dog who has just been fed (therefore not even more tempted because of hunger.

Even if you don’t train with positive reinforcement, a dog who learns he can get stuff that really tasty off the counters in your absence will likely surf when you are not there. Believe me…I owned the worst counter surfer in the world back when I was more likely to use compulsion, and I never could keep her from checking out the counters because she had gotten good stuff too often early on. It was fairly easy to use corrections to keep her off of the counters while I was there, but as soon as I left the house she was straight up on the counter (I spied on her with a video camera and by looking in a window). I used booby traps, red pepper and even a shock collar (while peeking at her through the window). She still stole off of counters while I was gone.

Once you have had a chance to actually insert some training, you can then step back and test the dog, and in fact you might even test the dog further by deliberately leaving things so that he can be rewarded for not taking them.

Manipulating the environment is only a small part of the process.

Although it can be more, a correction is usually just a verbal noise that communicates to a dog that it is wrong and better stop what it's doing and try something else. When you tell your dog "No!" you are training your dog using a correction.

Semantics even more: It depends on the dog. To some dogs a NO is devastating and a severe correction (very soft dogs who have difficulty failing), to some it’s a very mild aversive and to some its not aversive at all, only a no reward marker.

Dear Doggers,

…Many themes in this discussion are familiar, not just to me but probably to all of us. I was genuinely startled to read this. For research purposes I have visited pet dog trainers of different persuasions and attended national dog trainers conferences and though all these trainers asserted a goal of "off-lead reliability" of the several thousand dogs I saw both in training and accompanying their masters few were off lead and only then in dog parks or otherwise controlled circumstances. The gap between the goal and its achievement was striking. I don't view this as wickedness or human failure but merely that one's training method becomes one's default and nobody wants harm to come to their dog. Treat trainers always carry treats, ecollar trainers keep the collar on fully trained dogs, I never go anywhere without my shepherd's whistle.

Donald McCaig

Dear Mr. McCaig;

Good dog training is good dog training and bad dog training is bad dog training regardless of the method used. A good trainer knows how to apply reinforcement (regardless of type) to suppress or encourage behavior. A good trainer employs splitting (breaking behavior into pieces small enough to be taught), great timing (applying reinforcement, be it positive or negative, at the exact right time to communicate that this specific behavior is one I want you to do or no do), observation (being able to read the dog and see if he understood, if his brain is full for the day, if he is scared, if he is not engaged), and consistency.

If these things are used, the dog will indeed learn what is expected and once that learned behavior is fluent (meaning it has been trained well enough that the dog can do it in any context) you no longer need the reinforcement (be it positive or negative) to have fluency.

I know many, many positive reinforcement trained dogs who are off lead all the time, because they clearly understand what is expected. I hike off lead (illegally) on a city trail with 3 dogs who understand their job is to stay on the trail, stay within sight of me (only going far enough ahead than I can see them, and that if I call them they are to return IMMEDIATELY (so we don’t get fined or eaten by someone elses off lead dog).

OT: I am tickled about addressing you personally as Nop’s Trials is a favorite. I have read it 6 times. Thank you for that.

 

See next post (too much quoting for IPB). :rolleyes:

Share this post


Link to post
Share on other sites
Just a quick, irrelevant question: Would most operant conditioning trainers agree that this is an example of -P ?

 

No its not.

And I guess your dog's behavior of attacking other dogs can be "extinguished" by just never letting him see another dog?

You extinguish a fire, you use extinction to decrease the frequency of a behavior.

Removing the food from the counter is not extinction. It is management.

No. Extinction would be if you had him attack a dog and it had no effect (say he attacked a fake dog): the definition of extinction is the lack of any consequence following a behavior. When a behavior is inconsequential, producing neither favorable nor unfavorable consequences, it will occur with less frequency.

Extinction would be a horrible choice to address a dog attacking another dog, because the other dog is going to react by fighting back or running away and therefore there is no outcome that will result in a lack of consequences.

You can use extinction on a behavior like barking at night in the crate. Ex. I baby sat my neighbors Lab when they went on vacation. He was 10 months old and needed to sleep in his crate (my 12 year old dogs needed the peace!). He barked as soon as the lights went out. I could have used negative reinforcement (yelling at him or hitting the crate when he barked) or positive reinforcement (waiting for him to stop for a second and rewarding him with a piece of food. I elected to ignore him so crate barking had no consequence. After 30 minutes (we watched TV in the bedroom with the close caption on) he slowed down, and after 2 hours it was over completely. He didn’t bark the rest of the week. Barking received no consequence so he stopped doing it.

 

OK, so you've got a dog who jumps up on people all the time. The dog does it once to me, I let it know that I don't like that behavior (an "acchht!!"), and the dog puts all four on the ground. The dog may still jump up on other people, but as it approaches me, you can tell it's thinking about jumping up, and decides not to. So, I did not have to direct the dog to do something else (counter-conditioning),

That would not be counter conditioning, that would be training an incompatible behavior.

I just extinguished a behavior I don't like.

If you mean you made the behavior stop, yes you did. If you meant you used extinction, see above.

Now, will the dog continue to jump up on other people? Probably, because they allow it.
and because it is intrinsically reinforcing enough that he wants to do it.

I don't think you give dogs enough credit (I think Eileen was just saying this as I was beginning to post). Like Eileen, I have brought adult dogs into my house who have never been indoors before, and have taught them in one easy step that that behavior is not allowed;

 

And if he got it in one lesson with only a noise you in fact used a negative reinforce very well, with excellent timing (so he knew what he did that you didn’t want) and you probably also had already developed a bt of a relationship in that short time that speaks wonderfully about how well you rwad and like dogs. I’d quit for you too!

I also took in a dog who *did* have a habit of counter-surfing in her own home, and, again, one easy lesson--poof! No more counter-surfing... to never leave anything on your counter seems mere avoidance to me, and if the dog has a proclivity to counter-surf, you never address the problem at all, but merely avoid it.

100% agreed…however, do you think the dog would have stayed off the counter if you left the room and left a bit of raw meat on the counters edge? (I’m not trying to infer he wouldn’t, btw) How many times had he got something off of the counter? See above: I had an intractable counter surfer who never touched anything until she figured I was gone. She even learned to wait to hear the car drive away (I learned by spying).

I see this in all types of behavior--the owner just avoids the problematic behavior, rather than addressing it, fixing the problem, and moving on,

And if that happens there is not TRAINING, positive or negative, but management…and as M. Shirley Chong says (one of my favorite trainers) Management Always Fails.

I have to admit I was more on the positive everything side of the training debate until I witnessed the amazing communication between a very talented stockdog clinician and her dogs. They were happy! They understood what was expected of them! They LOVED their owner! And they were well behaved, well adjusted dogs. And yes, they were corrected.

I would guess shes is a skilled an excellent trainer whose dogs are happy because they understand what she wants due to her excellent training skills. In addition, stockdog work has the advantage of the best reward available for getting it right: the stock. No cookie or tug toy in the world can compete with that! I wish I could bottle it… :rolleyes:

…I tell everyone that comes to me for lessons, if you are not into any kind of correction, go watch a stockdog clinic or trial. Watch the amazing communication between the handler and dog. I never had that with my dogs until I learned how to train a dog on stock. We communicate with each other using corrections...why do we think a dog is too sensitive or stupid to do the same?

The reason I choose to use MOSTLY positive reinforcement is this: I get quicker results for the things I train when I use a motivator (I have not yet trained stock dog work) than I did when I used negative reinforcement, because of the psychological effect rewards have on a dogs brain. In the situation above (the dog jumping up) I might have used an “annnck” to get the dog off…its clear and not too aversive and get the job done. In addition, I would never reward the jumping up by petting or acknowledging at any other time. However, if I had a scared dog who had no clue about house rules and was soft as a feather emotionally, I might choose to ignore the jumping up and marking the behavior I liked (4 on the floor) with a really really reinforcing treat.

Share this post


Link to post
Share on other sites
It took me a long time to understand the actual consequences of accepting that behaviorism is a theory of mind (or non-mind)--rather than just a theory of training and/or learning-- and that as such, it misses a lot of opportunities to harness the cognitive processing that is involved in training (of dogs or other species).

 

Since I'm wordy and argumentative today, can I define behaviorism?

 

Definition:

 

The term behaviorism refers to the school of psychology founded by John B. Watson based on the belief that behaviors can be measured, trained, and changed. Behaviorism was established with the publication of Watson's classic paper Psychology as the Behaviorist Views It (1913).

 

According to behaviorist theory, our responses to environmental stimuli shapes our behaviors. Important concepts such as classical conditioning, operant conditioning, and reinforcement have arisen from behaviorism.

This is not incompatible with any kind of training.

 

Behaviorism holds that only observable behaviors should be studied, as cognition and mood are too subjective.

 

This is where you might part ways with behaviorism, understandable. I have issues with the 'black box' theories too.

 

However, when we talk about training regardless of how we do it, we are engaging in reinforcement which is a part of behaviorism.

Share this post


Link to post
Share on other sites
(Eileen Stein @ Mar 17 2010, 01:56 PM)]

Just a quick, irrelevant question: Would most operant conditioning trainers agree that this is an example of -P ?

No its not.

 

(Eileen Stein @ Mar 17 2010, 03:03 PM)]

And I guess your dog's behavior of attacking other dogs can be "extinguished" by just never letting him see another dog?

You extinguish a fire, you use extinction to decrease the frequency of a behavior.

Removing the food from the counter is not extinction. It is management.

 

I trust you realize I was challenging the statement that removing the food from the counter was either -P or extinction, not asserting that it was. I agree that it can only be considered management (although I do find it interesting to see the confusion these terms can generate, even among their adherents). And I used "extinguished" in quotation marks because I was unwilling to say "extincted." "Extinguish" is often, quite logically, used as the verb for extinction; see, e.g., http://en.wikipedia.org/wiki/Operant_conditioning; http://online.sfsu.edu/~psych200/unit3/32.htm. ETA: See also, Merriam-Webster: "extinguish: to cause extinction of (a conditioned response)"; Free Online Dictionary: "Psychology To bring about the extinction of (a conditioned response)."

 

I'm interested in this, however:

 

I don't think you give dogs enough credit (I think Eileen was just saying this as I was beginning to post). Like Eileen, I have brought adult dogs into my house who have never been indoors before, and have taught them in one easy step that that behavior is not allowed;

And if he got it in one lesson with only a noise you in fact used a negative reinforce very well, with excellent timing (so he knew what he did that you didn’t want) and you probably also had already developed a bt of a relationship in that short time that speaks wonderfully about how well you rwad and like dogs. I’d quit for you too!

 

Why do you characterize Anna's verbal correction (what you term "a noise") as "negative reinforce" or "negative reinforcement"?

Share this post


Link to post
Share on other sites
No its not.

You extinguish a fire, you use extinction to decrease the frequency of a behavior.

Removing the food from the counter is not extinction. It is management.

 

I trust you realize I was challenging the statement that removing the food from the counter was either -P or extinction, not asserting that it was.

 

Yes, and I apologize for the weird formatting that did not make it clear...IPB got mad at me for posting too much. :rolleyes: There should have been a space in there.

 

I agree that it can only be considered management (although I do find it interesting to see the confusion these terms can generate, even among their adherents). And I used "extinguished" in quotation marks because I was unwilling to say "extincted." "Extinguish" is often, quite logically, used as the verb for extinction; see, e.g., http://en.wikipedia.org/wiki/Operant_conditioning; http://online.sfsu.edu/~psych200/unit3/32.htm.

 

I think the difference for me is we are having a discussion about specific terms in operant conditioning, and extinguish can have multiple meanings: a plural of extinction but it can also mean to stop a behavior by other means so it can be confusing, plus the two words are similar. I apologize if it seems like I was questioning your intelligence, that was not my intention. I was just trying to be specific.

 

I'm interested in this, however:

And if he got it in one lesson with only a noise you in fact used a negative reinforce very well, with excellent timing (so he knew what he did that you didn’t want) and you probably also had already developed a bit of a relationship in that short time that speaks wonderfully about how well you read and like dogs. I’d quit for you too!

 

Why do you characterize Anna's verbal correction (what you term "a noise") as "negative reinforce" or "negative reinforcement"?

 

The way I read her post she had a dog who jumped up, she applied an aversive (a very very mild aversive, but an aversive nontheless :D ) and he put his feet down, so she stopped applying the aversive. The removal of the aversive was the consequence of him putting his feet on the floor.

 

He got the message "oh its this she doesn't like!" and stopped immediately.

Share this post


Link to post
Share on other sites

Dear Doggers,

Ms. Rushdoggie wrote(in part):

 

"Good dog training is good dog training and bad dog training is bad dog training regardless of the method used. A good trainer knows how to apply reinforcement (regardless of type) to suppress or encourage behavior. A good trainer employs splitting (breaking behavior into pieces small enough to be taught), great timing (applying reinforcement, be it positive or negative, at the exact right time to communicate that this specific behavior is one I want you to do or no do), observation (being able to read the dog and see if he understood, if his brain is full for the day, if he is scared, if he is not engaged), and consistency."

 

This is unexceptionable: yeah, sure. But . . .

 

Splitting. In sheepdog training one very rarely can teach only one thing. Early days:

Your young dog is blurring around you so rapidly you are getting too dizzy to stand and the sheep are pressing against your legs and threatening to trip you. Your dog more or less wants to split those sheep and savage one. You'd like him to down (submit) on command. You want to deepen your working relationship with him and you wouldn't mind it if the dog reversed its flanks from time to time nor if it started to associate your flank commands with the proper directions; taking into account the dog's personality: hard, soft, sulky, keen, indifferent, grippy. Oh, and you do (mostly) want to protect the sheep.

 

And your instructor has instructed you to: "Push that dog off you. It needs space to think."

 

Sheepdog (and handler) training is, initially, a dance learned by two inexperienced but more-or-less-willing partners (and at least three very experienced unwilling partners). It is nothing like "splitting: and the language and theory of behaviorism aren't so much wrong as utterly beside the point. Yes, one can concoct a description of what's happening using Skinner's four quadrants but one might as profitably produce a description in Urdu.

 

A second point. We often speak as if timing and reading the dog is something one learns and then possesses. These critical elements of man/dog communication vary with circumstance. When a snowbound handler complains "I haven't worked my dog in weeks" the dog is likely to have retained a sharper working edge than his handler has.

 

In the broadest sense the experienced dogger has so much better timing and reading ability he/she seems miraculous to the unskilled. You and I both can see a dog fight building in the dog park more readily than any civilian (incliuding the owners of pre-combatative dogs) and we can prevent the fight from erupting (often accompanied by an owner's "Why you talkin' so mean to Fluffy! He was only playin'").

 

But sometimes good timing and good reading fades in the presence of a gift. Last year about this time I was taking a lesson from Kevin Evans, the affable young welshman who won the 08 International. My 4 year old Danny, has a serious psycho-motor problem. From time to time something clicks in his mind and locks in and goes after a sheep. He's been that way since he first started working stock and otherwise he's an excellent sheepdog. He's won pronovice trials but I don't dare run him on a bigger course.

 

Anyway, there we were in Georgia with Kevin Evans who wouldn't have put one more hour or one more dollar into Danny (though he was too polite to say so). Thinking afterwards, what I realized was: when we both saw the problem abuilding and gave Danny a verbal correction, Kevin's correction was a mini- second quicker than mine. I overlapped his. Ot was my dog I'd worked for two years and this young welshman, first time he ever saw Dan, was faster than I was. Kevin Evans has a gift.

 

Donald McCaig

Share this post


Link to post
Share on other sites
Since I'm wordy and argumentative today, can I define behaviorism?

 

I don't think there was anything in my post (sorry, I don't know how to double embed the quotes) that was incompatible with the definition you provided (which, I think, came from about.com:psychology--is that right?), but if there was, please let me know.

 

This is not incompatible with any kind of training.

 

Nor did I claim that it is

 

This is where you might part ways with behaviorism, understandable. I have issues with the 'black box' theories too.

 

However, when we talk about training regardless of how we do it, we are engaging in reinforcement which is a part of behaviorism.

 

This is actually getting at the essence of my post. I came to the point where I understood that it wasn't really possible (for me) to accept a theory for training purposes while rejecting its fundamental, underlying premise--namely that there is no cognitive processing that intervenes in the "stimulus-response" chain.

 

For at least 40 years, cognitive science has shown that the behaviorist theory of mind can't be correct (which doesn't speak one way or another to the issue of measuring and observing behavior--it might have more to do with issues regarding changing it). Beyond that, the ability to observe cognitive processing directly as it is happening (ETA: in humans) has also advanced significantly and continues to do so at a rapid clip. These observations also show that cognition matters for behavior.

 

In having read through this entire thread (and many of the others like it), I think the fundamental difference underlying the various positions is what you believe (whether you articulate it or not) about the cognitive processing going on--I think that's why the people arguing against behaviorism keep bringing up very complex situations with many environmental factors involved.

 

In any event, one thing I've observed that I've often wondered about is something about the relative timing of positive and negative reinforcement--namely that dogs seem much more able to process poorly timed positive reinforcement than they are poorly timed negative reinforcement. Thus, for new dog trainers (I'm thinking here esp. of people learning to train pet dogs), it's a lot easier to get results using positive reinforcement (because your timing needn't be quite as precise--which it simply can't be as you're learning the skill). I have no idea if this is an accurate observation or not, though. Has anyone else noticed the difference in issues of timing with new trainers with positive and negative reinforcements?

Share this post


Link to post
Share on other sites
Semantics even more: It depends on the dog. To some dogs a NO is devastating and a severe correction (very soft dogs who have difficulty failing), to some it’s a very mild aversive and to some its not aversive at all, only a no reward marker.

And of course in this discussion, neither Anna nor anyone else has said that these verbal corrections are absolutes. We all read *the dog* and tailor corrections to what the dog *can* take. A very soft dog can certainly take a correction, if it is timed correctly and of appropriate intensity for the dog's personality and learning style. The fact is that even the softest dog on the planet, if it's going to work stock, is going to make a mistake while doing so that requires a correction. So that dog will need to learn to take a correction. A good trainer will certainly know how to temper the correction to suit the dog's temperament. I have had soft dogs, and have one who comes for lessons now. Yes, it may take a little more skill or ingenuity on my part, but I can most certainly give the dog a correction without devastating it for life.

 

As for the "splitting" topic, even though Donald already covered that adequately, I want to add my two cents. I may go out thinking I will start teaching the dog to drive, for example, but what training actually goes on out there will also depend on the "unwilling" participants and I've never yet gone out and taught just one little "split off" behavior related to working livestock without of necessity covering a whole bunch of other things too. Just one quickie example: I may want to work on the dog not going into chase mode when the sheep break back to the barn, but the dog does need to learn to cover--that is, get out around the sheep and far enough ahead of them to have enough influence to turn them back from where they want to go. During this sort of training, I can't just work on "don't chase;" I also have to work on "stop," "get out" (flank out wider, which will have the dual purpose of making "chase" less likely and also put the dog into a position where the lead sheep can see it and be influenced by it), and "cover." Also "don't grab," BUT if you successfully get in front and head the sheep and one tries to run over you, then it's okay to bite her face. If the sheep are running hell bent for leather, I have to encourage the dog to be fast, but not to chase. If the dog goes into chase mode, I have to work on a stop, because a dog in chase made isn't likely to give me that wider flank I need to cover once it's locked into the chase. And of course chasing can lead to gripping, and I have to get the dog to understand that gripping while chasing is inappropriate but could be okay if a sheep is running him over. I can't just correct the gripping willy nilly, though, because the last thing anyone wants to do is take the grip completely out of the dog. Yes, all those things are individual items the dog can learn (stop, widen, hurry up, cover, grip, don't grip), but they really have to be taught *together* in a fluid and changing environment so that the dog gets that it's okay to go after sheep that are running away but that it needs to do it with purpose and thoughtfulness--and control. And also that there are times when it's okay to let the sheep get away. During all this, I'm also "teaching" the dog that it needs to be obedient to me but also needs to think for itself and perhaps ignore my command if it knows I'm making a mistake. It truly is like a dance where neither partner knows the steps beforehand and so they have to adjust to their partner and also to the other dancers (the stock). And of course while I'm trying to train all these things in the dog, the sheep are doing their own thing (they don't know my training agenda and don't care). Maybe once the dog has correctly headed the sheep, one is still determined to break back to the barn, while the rest are either stopped or coming back toward me, or perhaps even moving to another part of the pasture. Maybe I've got two bad actors, who have split off in different directions, determined to beat the dog. There's simply no way I can work on just one specific item of training and what actually gets worked on depends on the reaction of the stock to what we might be working on/doing in any given second. In fact, you'll often hear stockdog trainers, when asked "When do you start to teach X?" say that they do so as the opportunity arises. This is because the stock will present training opportunities that you may not have actually planned on or been specifically ready for, but when it happens you take advantage of it. Again, I just don't see how that sort of training fits into the discrete splitting you refer to.

 

As Eileen said earlier, I'm sure a behaviorist could look at what I'm doing and try to fit discrete moments of training into little boxes, but from my POV it's so fluid and ever-changing and in the moment depending on all sorts of outside factors (livestock, weather, terrain, time of day, mental acuity of the dog--or trainer :rolleyes: --at that moment, etc.) that I wouldn't even try to split it into little boxes.

 

J.

Share this post


Link to post
Share on other sites
This is actually getting at the essence of my post. I came to the point where I understood that it wasn't really possible (for me) to accept a theory for training purposes while rejecting its fundamental, underlying premise--namely that there is no cognitive processing that intervenes in the "stimulus-response" chain.

 

I am very interested in a specific citation from a reinforcement based dog trainer who claims this. Name, book, article, etc.

 

In having read through this entire thread (and many of the others like it), I think the fundamental difference underlying the various positions is what you believe (whether you articulate it or not) about the cognitive processing going on--I think that's why the people arguing against behaviorism keep bringing up very complex situations with many environmental factors involved.

 

OK. What do you think I believe about cognitive processing, based on the fact that I choose to train through reinforcement, without correction?

 

Again, specifics.

 

In any event, one thing I've observed that I've often wondered about is something about the relative timing of positive and negative reinforcement--namely that dogs are much more able to process poorly timed positive reinforcement than they are poorly timed negative reinforcement.

 

I've found this to be true - just from my observation, not any statistics or anything. A big part of what the students are there to learn in a reinforcement based class is how to reinforce most effectively. But poorly timed reinforcement can still give the dog a ballpark idea of what is desired. There's that cognitive process at work - the dogs are often intuitive enough to figure it out even when the handler is not perfect with timing, especially with very basic things. As the handler's timing improves, precision will improve, so will speed in training. But perfect precision is not required right from the beginning. Dogs that I see in training classes who receive poorly timed correction tend to either tune it out altogether (Aht! Aht! becomes the constant elevator music in the background), or they associate the correction with a behavior that the owner did, in fact, want the dog to continue, and then there is some sorting out to do. Just my observation.

 

Thus, for new dog trainers (I'm thinking here esp. of people learning to train pet dogs), it's a lot easier to get results using positive reinforcement (because your timing needn't be quite as precise--which it simply can't be as you're learning the skill). I have no idea if this is an accurate observation or not, though. Has anyone else noticed the difference in issues of timing with new trainers with positive and negative reinforcements?

 

I can't give a comparison, but I will say that I give pet owners a lot more credit than people typically do when it comes to the ability to learn how to reinforce effectively. Yes, it's a skill. Yes, they have to put a little thought and practice into it. Those who care to do so, even a little, tend to get it very quickly. Those who don't care, or don't want to use reinforcement - they don't learn because they have chosen not to. I'm not implying judgment on that choice in making that statement. It's just an objective observation. What, after all, will you learn to do well if you either don't care, or don't want to?

 

One thing that I see undertaught in reinforcement based classes is how to effectively transition what the dog has learned with food to a point where the dog is completely fluent in real life situations and food no longer need be part of the picture. That's a flaw in the instruction process, though, not in the approach itself. It's not rocket science - it's just part of the process that seems to get forgotten in pet dog classes.

Share this post


Link to post
Share on other sites

Why do you characterize Anna's verbal correction (what you term "a noise") as "negative reinforce" or "negative reinforcement"?

The way I read her post she had a dog who jumped up, she applied an aversive (a very very mild aversive, but an aversive nontheless :rolleyes: ) and he put his feet down, so she stopped applying the aversive. The removal of the aversive was the consequence of him putting his feet on the floor.

 

He got the message "oh its this she doesn't like!" and stopped immediately.

 

So you view Anna as reinforcing her dog's getting down from the counter, rather than punishing her dog's getting up on (surfing) the counter? Hmm. From that perspective I guess her training was a failure, since the behavior of getting down from the counter must have decreased -- if he never again got up on the counter he couldn't be doing much getting down from the counter. Or do you just share my dislike for the use of the term "punishment" in this context?

 

Gosh, this focusing on terminology seems boring and beside the point compared to the discussion going on about actual training, doesn't it?

Share this post


Link to post
Share on other sites
And of course in this discussion, neither Anna nor anyone else has said that these verbal corrections are absolutes. We all read *the dog* and tailor corrections to what the dog *can* take.

 

Interesting, we read *the dog* and tailor reinforcement to what the dog needs.

 

I get the impression that some of you (not necessarily you, Julie) have the idea that we live our lives in a box or something where we never interact with our dogs beyond training. Like we treat the dog as a subject that does not have a personality, preferences, input into training and his or her life in general, and that there is not a regular old give and take relationship between dog and handler that is not only in place in regular everyday life, but in the training process itself. Like the training is something done *to* the dog and not *with* the dog as a partner with just as much, if not more, to offer than the handler.

 

If I'm wrong on this, I'd welcome clarification. :rolleyes: And if I'm not wrong, I'd love to know where this idea is coming from. It is completely contradictory to my own experience with reinforcement based training.

Share this post


Link to post
Share on other sites
The way I read her post she had a dog who jumped up, she applied an aversive (a very very mild aversive, but an aversive nontheless :rolleyes: ) and he put his feet down, so she stopped applying the aversive. The removal of the aversive was the consequence of him putting his feet on the floor.

 

He got the message "oh its this she doesn't like!" and stopped immediately.

So you view Anna as reinforcing her dog's getting down from the counter, rather than punishing her dog's getting up on (surfing) the counter? Hmm. From that perspective I guess her training was a failure, since the behavior of getting down from the counter must have decreased -- if he never again got up on the counter he couldn't be doing much getting down from the counter. Or do you just share my dislike for the use of the term "punishment" in this context?

 

Gosh, this focusing on terminology seems boring and beside the point compared to the discussion going on about actual training, doesn't it?

 

I feel kind of like I am being baited here...but I will bite.

 

If she made the aversive noise while he was jumping up (and now I have replied to multiple posts so I am confused but I thought the situation was he jumped on her? It doesn't matter I suppose) and had his feet on the counter, the aversive happened, and as soon as his feet were off the aversive stopped, then she made having his feet on the floor a good thing.

 

Not having seen the situation, its possible the noise was a punishment (by definition), added after the feet hit the counter that caused the dog to jump off, but when I read the post what I saw in my minds eye was the former. Ether way, I'm not sure why you would call it a failure? It clearly worked, worked well enough she said the behavior stopped and such a mild aversive did not carry any negative consequences with the dog, right?

 

I don;t find the terminology boring or besides the point, myself. Knowing how any why something works in my training is very interesting to me!

 

And to Mr. McCaig and Julie:

 

I don;t know much about stock dog training, but I was confused by your posts:

 

Do you mean to say that when you train a dog to work sheep, that you expect the entire set of behaviors to happen all fully formed at once? You can take an untrained dog out on a course and just start expecting the full picture? Don't you start in an area and help the dog understand one thing, then move on to increasingly more complex situation? I honestly don't know, I am asking.

Share this post


Link to post
Share on other sites
And of course in this discussion, neither Anna nor anyone else has said that these verbal corrections are absolutes. We all read *the dog* and tailor corrections to what the dog *can* take.

 

I get the impression that some of you (not necessarily you, Julie) have the idea that we live our lives in a box or something where we never interact with our dogs beyond training. Like we treat the dog as a subject that does not have a personality, preferences, input into training and his or her life in general, and that there is not a regular old give and take relationship between dog and handler that is not only in place in regular everyday life, but in the training process itself. Like the training is something done *to* the dog and not *with* the dog as a partner with just as much, if not more, to offer than the handler.

 

If I'm wrong on this, I'd welcome clarification. :rolleyes: And if I'm not wrong, I'd love to know where this idea is coming from. It is completely contradictory to my own experience with reinforcement based training.

 

Well, I'd guess that Julie wrote what you quote at least partly in response to the positive-reinforcement-based trainers' repeated statements that the dangers of correcting a dog are too great to risk -- dog may shut down altogether, dog may think he daren't do things other than what he was corrected for, etc. Julie is pointing out that a good trainer modulates the corrections to the dog.

 

I don't think she or anyone else has suggested that you never interact with your dogs beyond training.

Share this post


Link to post
Share on other sites
Not having seen the situation, its possible the noise was a punishment (by definition), added after the feet hit the counter that caused the dog to jump off, but when I read the post what I saw in my minds eye was the former. Ether way, I'm not sure why you would call it a failure? It clearly worked, worked well enough she said the behavior stopped and such a mild aversive did not carry any negative consequences with the dog, right?

 

You termed it negative reinforcement. Reinforcement is that which increases a behavior, right? What behavior did Anna's correction increase? What behavior was it intended to increase?

 

And to Mr. McCaig and Julie:

 

I don;t know much about stock dog training, but I was confused by your posts:

 

Do you mean to say that when you train a dog to work sheep, that you expect the entire set of behaviors to happen all fully formed at once? You can take an untrained dog out on a course and just start expecting the full picture? Don't you start in an area and help the dog understand one thing, then move on to increasingly more complex situation? I honestly don't know, I am asking.

 

You don't expect "the entire set of behaviors" to happen all fully formed at once, but it is impossible to work on them one at a time. I trust Julie and/or Donald will answer more fully.

Share this post


Link to post
Share on other sites
Guest echoica
Why do you characterize Anna's verbal correction (what you term "a noise") as "negative reinforce" or "negative reinforcement"?
The way I read her post she had a dog who jumped up, she applied an aversive (a very very mild aversive, but an aversive nontheless :D ) and he put his feet down, so she stopped applying the aversive. The removal of the aversive was the consequence of him putting his feet on the floor. He got the message "oh its this she doesn't like!" and stopped immediately.
So you view Anna as reinforcing her dog's getting down from the counter, rather than punishing her dog's getting up on (surfing) the counter? Hmm. From that perspective I guess her training was a failure, since the behavior of getting down from the counter must have decreased -- if he never again got up on the counter he couldn't be doing much getting down from the counter. Or do you just share my dislike for the use of the term "punishment" in this context?

 

Gosh, this focusing on terminology seems boring and beside the point compared to the discussion going on about actual training, doesn't it?

 

It is NOT negative reinforcement. The target behaviour is putting feet up. The aversive is applied (positive)...and the consequence is feet removal (decrease target behavior - punishment). If you consider the removal of aversive to be a desired 'consequence' rather than the behaviour itself, I say you are not going to be a very effective trainer in this particular circumstance.

 

...oh and I still hold that my earlier example of removing food to reduce counter surfing is negative punishment. Regardless of if you want to put a modern name on it (which is fine...theory changes all the time) but under the operant conditioning model it IS absolutely negative punishment (technically) or at the very least an attempt at extinction...

 

*yawn* it is getting boring :rolleyes:

Share this post


Link to post
Share on other sites
I used booby traps, red pepper and even a shock collar (while peeking at her through the window).

 

:rolleyes: Isn't that a little abusive? Wouldn't it have been better to prevent it from happening at all?

 

She still stole off of counters while I was gone.

 

My dogs find that a bit hard to do from their crates.

 

you might even test the dog further by deliberately leaving things so that he can be rewarded for not taking them.

 

Cool! What a great idea! Maybe I should go buy a fire hydrant and put it in my living room and reward my dog for not pissing on it! Or I could walk around with my leg stuck out to the side and throw him treats for not humping it! He'll love that!

Share this post


Link to post
Share on other sites

Regarding training stock/sheep work, IMO:

 

It is not possible to train a dog to work sheep in a proper relationship with the sheep and the trainer without using what are technically aversives, however mild.

 

The very essence of what a good working border collie is bred to naturally do is the gather, where the dog goes out and around sheep, in a way that doesn't disturb them until it is behind them and in a position to bring them in a fairly straight line back to the trainer.

 

The first thing most trainers do is try to get the dog to go away from them and around to the opposite side of the sheep. This means the dog will need to react to the presence or pressure of the trainer to be cued to move away from him/her in order to go around to the proper position to balance or hold the sheep to the trainer.

 

Depending on the dog and trainer, this aversive stimulus could be as mild as a strong presence of the trainer or as strong as the trainer needing to put pressure on the dog by moving toward it, or using any number of devices from training sticks to rakes or what have you.

 

Once the dog is behind the sheep, and starting to balance or hold the sheep to the trainer, it should start to click into its genetics that it is in the proper position with regard to the trainer and sheep. Then it begins a three way relationship that makes sense to it because of its very specific breeding to understand this relationship.

 

Once this genetic potential starts to be unlocked, and if the dog has the talent, much or even most of what follows depends on the attitude the dog develops. There is very little that actually needs to be trained if the dog is talented and has proper respect for the sheep and trainer. Developed, yes, but trained, no. The development of this respectful attitude will by its very nature require the application of pressure (which is considered aversive) and the release of pressure by both the trainer and the sheep.

 

The dog is getting information it needs and this information is not only, yes, that's right, but, no, that's not right, that's close, etc. The reason the dog is happy to get all kinds of information is because this information helps it get to this proper relationship between the sheep and the handler, which is genetically at the dog's core. It is seeking this relationship.

 

So many people say the sheep are the reward - you give the sheep and take the sheep away like a treat depending on the behavior of the dog. I'm not going to say it's more complex than that because that statement even undermines what is really happening. The intrinsic reward is not just the sheep. The intrinsic reward is the proper relationship that the dog understands in its genes. The dog doesn't need to learn specific behaviors. The talented dog is learning a proper attitude and once it has that, its actions are correct, or it will seek correct behaviors on its own or with reminders from the sheep or trainer.

 

You just can't break that down into specific behaviors.

Share this post


Link to post
Share on other sites
Interesting, we read *the dog* and tailor reinforcement to what the dog needs.

 

I get the impression that some of you (not necessarily you, Julie) have the idea that we live our lives in a box or something where we never interact with our dogs beyond training. Like we treat the dog as a subject that does not have a personality, preferences, input into training and his or her life in general, and that there is not a regular old give and take relationship between dog and handler that is not only in place in regular everyday life, but in the training process itself. Like the training is something done *to* the dog and not *with* the dog as a partner with just as much, if not more, to offer than the handler.

 

If I'm wrong on this, I'd welcome clarification. :rolleyes: And if I'm not wrong, I'd love to know where this idea is coming from. It is completely contradictory to my own experience with reinforcement based training.

Kristine,

My post was a *direct response* to Rushdoggie's comment that some dogs are too soft to take corrections and that a simple "No" would devastate such dogs. My response was simply that even soft dogs can indeed take corrections, IME, as long as the trainer tailors the correction to the individual dog. That's not an indictment of anyone else's training methods or interactions with their dogs, but is simply my response disagreeing with RD's belief that some dogs can't take a correction. My experience leads me to believe otherwise. Nothing more, nothing less.

 

Maybe you're getting this impression simply because the discussion is about training and not about how we spend the rest of our lives with our dogs? (Honestly, not being facetious here.) I just assume that people who live with their dogs do indeed interact with them on some sort of "non-training" level but that they also use opportunities that present themselves to apply training, even if no training was planned. I think that's what any good owner or trainer would do, and I expect that the other folks to whom you refer (the other stockdog trainers here) would believe the same.

 

But when the discussion turns to behavioral terminology and putting things in boxes, then I think it also necessarily turns to talking about specific, discrete instances of training and not the whole, rich world of interaction that we all presumably have with our dogs.

 

So in my post above, I wasn't trying to imply that someone like you, who trains differently than me, is not having the same rich relationship with your dogs that I have with mine, but when behavior theory is discussed and some parts of that *theory* imply certain things, including dogs as "robotic" subjects, then naturally the discussion is going to take a turn in that direction. I think where you may be getting caught up is that people like me are basically objecting to the idea that the theory explains all and people like you are perhaps taking our objections to be objections to the specific methods you use, rather than to some of the theory *behind* those methods. At least I think that's how we're managing to talk at cross purposes.

 

J.

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.

×
×
  • Create New...