The Training/Behavior "Chatter" Thread

I haven't lost a dog yet... I do not look forward to going through what you guys have... I can barely stand to read about it, thinking of my own loved ones here...

spiritdogs

To me, they are never less than - they are just different. 

Yeah... 

Merry Christmas everyone.  

FourIsCompany

The question is not COULD you clicker train a child or use positive reinforcement on a child, but WOULD you clicker-train YOUR child the same as you clicker-train a dog (if you do), by withholding food and having them try different behaviors until they got it right? If you're going to make a direct correlation from  flooding a dog to flooding a child, then make the direct correlation to clicker training a child. 

I'm a bit behind, but I wanted to comment. I would most certainly clicker train my child in that fashion if it so happened that we did not speak the same language somehow. This is the important distinction to me. You can explain to a child in a language they comprehend what you want of them and why. You can't do that with a dog. If for some reason I couldn't do that with a kid, I would consider using marker training. I would be unlikely to withold food because a child's natural inclination is to communicate verbally with other humans. They are driven more by the desire to communicate than by wanting food. I may well withhold something like sweets or chocolate, though, much in the same way I see parents hold back the sweet until the child says "please".

I don't know about other people, but I don't withhold things an animal needs to survive to train them. I withhold treats and favours and things like attention that the dog doesn't have to have, but would very much like.

FourIsCompany

The reason flooding doesn't work with thunderstorms is that the dog isn't IN the fear long enough for it to fade away. During the storm, the thunder comes and goes and each time the dog goes INTO the fear and then calms down somewhat between lightening strikes. The only way to flood the dog successfully would be to have nearly constant thunder that lasted however long was necessary for the dog's fear to fade and to realize that the thunder wasn't actually causing harm. The dog (or person) cannot stay "in fear" forever and so the body moves out of fear even though the stimulus is still present.

I've been reading about flooding and what actually happens psychologically and it's very interesting.  

Well, maybe to a point. However, my dog is equally terrified of thunder and fireworks. A few years back we lived across the road from a public park where they put on extensive fireworks at New Year. After 20 minutes straight of constant bangs, cracks and pops, my dog was more scared than she'd ever been of fireworks or thunder. I had to sit down with her and physically prevent her from running around the house in hysterics. I've never seen her that bad before or since. I got her calm, but she's still terrified. Maybe she could have worked through it if the fireworks had continued for several days, but as it was it was constant bangs right on top of each other for about 30 minutes, and if anything it made her worse.

I think there's a time and place for flooding, and that's when an otherwise confident dog suddenly gets vaguely spooked by something simple. If Penny balks at something, I hustle her back and get her to face it before she can decide she's afraid of it. However, if she gets as far as running and hiding and won't respond to my calls, it's time to be more sensitive and desensitise her. I personally know a dog that was tossed into the water to 'get her used to it'. I could hear her yelping from my house next door and she never ever went in the water past her ankles after that. On the other hand, it took me several months to convince Penny that water was harmless by wading in it whenever I got the chance and presenting her with a small river crossing. For a long time she chose the log to cross rather than paddle, but in the end, she realised it wasn't so bad after all and now you can't keep her out of the water. I like my way better, but for some dogs flooding might work equally well, especially if you don't have time on your side.

kind of bored with the "fears" concept-  let's discuss this idea that some people have brought up that one is "withholding" stuff if you're trying to use +R (clicker, marker, whatever) during training and the dog didn't manage to earn the reward and this "withholding" may be very frustrating for the dog.

Has anyone else played the training game? where humans actually volunteer to get clicker-trained? we did it at one class, the only "reward" we got was the click (unlike dogs, the instructor could explain the meaning of the click without conditioning us with food to enjoy hearing it). I first drew an "instructor" who had horrible timing. The depth of frustration I felt was truly amazing- I had no idea what that idiot wanted me to do despite earning many clicks. If I'd been a dog I would have either bitten her or walked away. Next "instructor" had ok timing, and the depth of joy both of us, trainee and trainer, felt when I rapidly learned "the trick" was also very surprising. Not once did I feel "frustrated" about not earning clicks, it was getting poorly timed clicks that were just confusing that caused frustration. In fact, once I'd mastered my "trick" I earned fewer clicks per minute and had to work harder to get that one click, but it felt better to get the trick right and get clicked for it than to earn many clicks during the "training" process.

mudpuppy

kind of bored with the "fears" concept

 

Me, too. 

mudpuppy

let's discuss this idea that some people have brought up that one is "withholding" stuff if you're trying to use +R

I just wanted to say that I know not everyone withholds food to use for training. But I know that even Karen Pryor suggests to use the dogs meal for clicker treats.

I don't have anything to add about humans getting clicker-trained at this time.  

I just wanted to say that I know not everyone withholds food to use for training. But I know that even Karen Pryor suggests to use the dogs meal for clicker treats.

 

I've used parts of meals to charge clickers, not really to train though.  When I do it, it's a two-fold purpose, charging a clicker and either doing it with a dog that eats way too fast and will choke (Coke at first) or a dog that I'm trying to build a trusting relationship with (Kenya at first).  Now that both dogs trust me and neither eats too fast (most of the time), I no longer use meals for that purpose.  They also know the click = treat so I only "charge" with half a dozen pieces or so, a marginal amount that won't make them fat if they get it but won't starve them if they don't.  Their actual training treats are so small and usually healthy that it doesn't matter if they get them or not.  Right now I'm actually using cat food for Kenya (really tiny kibbles) b/c I'm trying to put on a pound or two, but otherwise I use pieces of cheese or cooked chicken breast the side of the zero on your keyboard.  In an hour long training class or practice group, one of my dogs would eat less than a string cheese even earning treats constantly.  Usually if I have bits left at the end, I make one final trick be the "jackpot!" and they earn whatever is left. 

Each trainer I've taken obedience with has suggested using kibbles, but not for the purpose of withholding food or forcing a dog to earn its meals.  In the first obedience class I took, *everyone* was either over-treating, or didn't think to bring treats at all.  Half the dogs were getting whole milk bones for every command and the other half didn't have any treats.  The trainer suggested using kibble mainly as a reference for the size and value of a basic training treat.  Also, that was a beginner/puppy class so most of the dogs didn't know their names.  With the "Name Game" you want to be naming like 50 times in a row.  For a puppy, 50 treats per session is a lot of extra treats and would get unhealthy pretty fast.  The trainers suggested just using a meal.  The point is not that the dog is earning its meal - they will get the meal regardless.  If you skip the Name Game for a day, you still feed your dog.  We had to use kibble for Coke's Name Game because he gets massive diarrhea when he eats most things besides what his stomach is used to.  Using his kibble, he learned his name solidly in two days.  When we got sick of proofing, we set down the bowl and let him eat the rest of the food (or hand-fed handfuls at a time to slow him down). 

I'm not real sure I see the problem with using a meal for training if the dog will actually work for kibble. The dog will get fed regardless but why not kill two birds with one stone if you have some training to do? I don't free feed, so even to get their bowls of food there is a procedure the dogs have to follow in order for me to put their bowls down. That's working for food. Every day when we leave to go to work the dogs are left with their breakfasts packed in to Kongs and similar toys that they have to work to unpack before eating. Working for food. So if it's around dinner time and I've got something I'm working on with Marlowe, I have no problem using part of his dinner in training (he finds a piece of kibble just as motivating as a piece of liver--it's all in the same category of "OMG FOOD!" for him). I'll maybe use half to train, then fill the bowls with the rest and it's dinner as per usual. If you're a pretty good trainer with pretty good timing, the dog is getting fed for doing some pretty easy stuff and I don't see what the problem is with that.

mudpuppy

...let's discuss this idea that some people have brought up that one is "withholding" stuff if you're trying to use +R (clicker, marker, whatever) during training and the dog didn't manage to earn the reward and this "withholding" may be very frustrating for the dog.

Has anyone else played the training game? where humans actually volunteer to get clicker-trained?

Hello, "some people" here.

I played the training game in class.  The point of the game was to show the human how the dog learns using C/T.  In our game, there was no food treats but the reinforcement was the association of the click with the body language of the trainer, a smile.  The tasks were to go and retrieve a specific object in the room and then do something with it.  There were no timing issues.  E.g., if you turned right and got no click, then you turned left and if you still got no click, then you went straight ahead and then you got a click and a smile.  No timing issues at all didn't mean there was no frustration.  Like any game, sometimes you win and sometimes you lose.  With the human game, when you got a click and smile, it was a good feeling.  When you offered something and got a blank stare, it was a downer feeling.  Its a game and it is always good to win.  Humans have the awareness that it is a game with an element of competition and tend to be able to properly process the emotional up and downs.  Do dogs know competition in this same setting?  Dogs shape their behavior to get pleasure and avoid displeasure.  What a dilemma for the dog. 

DPU

Humans have the awareness that it is a game with an element of competition and tend to be able to properly process the emotional up and downs.  Do dogs know competition in this same setting?  Dogs shape their behavior to get pleasure and avoid displeasure.  What a dilemma for the dog

Or is? At one time we say a dog is not a human and will not feel frustration or betrayal or fear during flooding like a human will, and in the next, we assume the dog will have human emotions of frustration without the human's abstract analysis of their own frustration. BTW, DPU, knowing that you are a smart guy, you would be one of the few I could think of that could actively analyze your own reaction to something, in real time. There are many who could not or would not.

So, which is it? Does the dog feel the same frustration as we do? In which case, could they also feel about flooding like we do? And physical punishment or correction, like we do? I'm specifically not comparing to Koehler or Leerburg.

I think the example of the human games also show the limitations of the humans themselves. As MP pointed out, the first person she worked with had no sense of timing and that caused to frustration. Properly timed and executed, clicking and treating is about leading to the reward. And if one has to wait too long to click, the goals are too long and should be shortened into smaller segments. Also, in the human version, it is assumed that a smile from the trainer is rewarding. With me, I might work better for chocolate.

But back to the frustration? Is it not also possibly frustrating to a dog that is receiving mostly corrections to not ever know what direction the human wants?

DPU

Humans have the awareness that it is a game with an element of competition and tend to be able to properly process the emotional up and downs. 

 

This is true. Humans have a huge advantage in being able to process the information in such a game. It's a totally different situation for human and dog. For one thing, humans are told they're playing a game. They're told the rules. They have so much more information available to them than a dog does. And they enter the game willingly.

I can see doing it to observe a feeling of getting a reward, but I think that's just about where the similarities between dog and human would end. I got the same thing by having a report card or having my mom tell me I did something well or that she was proud of me.   

This "game" can be played different ways, always with the outcome contingent upon the "priming" of the human participants before hand.

Another way to play, is to try and get a particular human to sit in a specific chair by saying "hot, cold, warm, cold, or burning hot", all with the same tone of voice and with a smile on your face. The benefit of this game is that you are also providing information as to the incorrect direction in a cooperative, rather than a negative (or a positive punishment" sense...(if this sounds like a "no reward marker", yay).

One more game, is to use facial, body, and eye gestures (or even pantomime) along with touch, rather than words  for "come to me", "go here", " stop", or "sit"...again, with a smile on your face.

With humans, the results of this game often rests on the influential power and skill of the Dungeon Master. 

 

(Carla, we keep posting at the same time...lol)

ron2

At one time we say a dog is not a human and will not feel frustration or betrayal or fear during flooding like a human will,

I never said this.  One would characterize a fear I have as 'fear of heights'.  That is on a ladder, on a roof, waaay above the ground.  But I don't have a fear of flying and looking out the window.  My fear of heights is really a fear of falling, or more generally something that will make me die because I don't have trust.  The emotions that overcome me comes directly from my survival instinct.  Dogs have survival instinct so I believe they share the same emotions.

DPU

I never said this.

I apologize, as I did not mean that you said this. Certainly, our breadth of opinion includes these various opinions that sometimes seem at odds.

DPU

My fear of heights is really a fear of falling

Same here. I don't fear death, however. I only fear serious and disabling injury. To me, being a vegetable for life is worse than death.

DPU

Dogs have survival instinct so I believe they share the same emotions.

You might have something there.

Again, I wasn't necessarily disagreeing with you but I happen to notice what I thought was a discrepancy in what we expect of dog emotions and thought processes and whether or not they relate to our own.

 

ron2

But back to the frustration? Is it not also possibly frustrating to a dog that is receiving mostly corrections to not ever know what direction the human wants?

*I* think so. Knowing what you did wrong doesn't necessarily clue you in to what's right. That's why I like PR training, using my voice or a clicker to mark the CORRECT behavior and following up with a food or play reward, rather than correcting for the wrong behavior. As long as your timing is good it's clear and unambiguous to the dog, and since there's no penalty for guessing wrong, there's always an incentive to keep trying until they get it right. It's not about withholding stuff for doing the wrong thing, it's about giving the dog the opportunity to GET stuff for getting it right. I'd think that would be empowering, not frustrating. My dogs know what they need to do to get us to throw the ball for them. They seem to think that's pretty darned awesome! Are they frustrated that we won't throw the ball if they break a sit or down before being released? Maybe, sometimes. But they know EXACTLY what the rules are and how to get what they want, so to an extent they are in control.

Cassidys Mom

 It's not about withholding stuff for doing the wrong thing, it's about giving the dog the opportunity to GET stuff for getting it right.

 

Yes. That right there. Withholding stuff is the realm of negative punishment, which I do use sometimes (door doesn't open until the dogs can maintain their sits until I release them for instance) but in a clicker training session, I'm looking for things to reward, not withholding things for being wrong. If the dog is not getting reinforced a whole lot, that is my fault as a trainer and I need to back up and come up with some way to make it easier for the dog. One thing you'll notice with really good trainers that use positive reinforcement is that their level of reinforcement for new behaviors is really really high. Way higher than your average novice. Novices tend to stand around and wait for the dog to out of the blue perform the complete correct behavior, which really just ain't gonna happen most of the time. That can lead to frustration, but that's because that's a really clumsy use of the technique. It's key to set the dog up to be successful--you want to give them lots of reinforcement, so you find things they can easily do that are baby steps towards the end behavior, and then you can reinforce to your heart's content. My dogs think this is the most awesome thing ever. I think they think they're training me, not the other way around.

I was at a seminar back in October watching people working with their dogs with the help of a professional trainer. I was sitting with my trainer and another professional trainer and they were just about at the edge of their seats going "REINFORCE!!! NOW! NOW!!!!" because the people working were just too slow with the reinforcements, just kind of standing around waiting for the dog to be perfect, meanwhile the dog was offering about a million acceptable behaviors that would have been excellent baby steps towards the end behavior if their handler had just noticed and reinforced them.