Renee and everyone else... what do you think?

[Kendall Aliza]

For the FULL article http://www.leecharleskelley .com/#tip
Is he essentially saying that instead of training with treats we should use tennis ball (or something or dog is obssessed with retrieving/having) to train our dog with? He is somewhat vague on how it works exactly but I guess that was his intention. Anyone know anything about what he is talking about and/or who he is?

"What this means in terms of dog training is very simple. There is absolutely no point in trying to make your dog be submissive to you as his "pack leader". No such animal exists in the mind or in the experience of any dog. All this does is create an unwilling and stressful obedience, based on fear. This is a completely unnatural way for a dog to learn and obey. Using food as the focus of training is also unnatural. No dog or wolf ever teaches another dog or wolf how to behave by offering food rewards to reinforce "good behavior". Of the four basic positive reinforcements-food, praise, play, and contact-food is, generally speaking, the least effective overall, and by far the most unnatural. Food is supposedly a "primary reinforcer", meaning that it satisfies a basic survival need. But the pack's prey drive is actually genetically engineered to override the survival instincts of its individual members. How else could canines deliberately go out and hunt something that could easily kill or maim them? (The hooves and antlers of large prey are lethatl weapons.) Then there's the fact that-as Coppinger puts it-"once the hunt is on, a wolf or a dog is not going to stop for liver treats." He only stops when the prey is down.

The fact is, dogs and wolves learn to adjust their behaviors naturally, through their prey drive. They want to learn and be part of a group activity, as long as it involves this need to hunt. It's as natural to them as breathing. In order to apply this to your dog, all you have to do is properly stimulate and satisfy his prey drive. (In fact, this works so well that when you build a dog's drive to a high enough level,, some commands can be taught once-just once, I'm not kidding-and the dog will never forget the command for the rest of his life.) Of course, some people are confused by this, thinking that stimulating the prey drive will make the dog harder to control. But in reality, properly stimulating then satisfying a dog's prey drive leaves him calm, relaxed, willing to do anything you ask of him, and totally happy to obey. This is how search-and-rescue dogs, drug-sniffing dogs, and bomb-detection dogs are all trained. And, with certain recent exceptions*, they're the best-trained dogs on the planet.

The same basic philosophy and techniques used for training these working dogs have been adapted by Kevin Behan (as well as by myself and others) for training pet dogs. It's the most effective way to train because it's the only one that feels totally natural to a dog's instincts and emotions. A study shows that most dog trainers are only successful 40% of the time. But, by using a dog's natural instincts for obedience, we're successful closer to 100% of the time; with all the dogs we train.

It's truly amazing.
"

[Kendall Aliza]

Yeah... Apparently I missed part of web page about the tennis ball business. So what do you guys think? How is this different from positive r? Isnt it the same since the ball, instead of a treat or a click, becomes the positive r?

[Renee]

Kendall-

I've never heard of this person......

Quote:
"This is a completely unnatural way for a dog to learn and obey. Using food as the focus of training is also unnatural. No dog or wolf ever teaches another dog or wolf how to behave by offering food rewards to reinforce "good behavior".

Since when does a wolf or dog "train" another dog to do anything - they don't ...(according to Beth Duman - Michigan's wolfpark rep)....especially in the way we train our dogs. Food is a primary reinforcer (food, air, sex, water to be exact) - it is a survival need. This is how initally teach behaviors, but we immediately phase out food and use other things like praise, petting, toys, access. We basically use the Premack principle when dog training ....I'm not sure what he means by saying it's unnatural to use food...I happen to think it is very natural and motivating to many dogs.

Quote:
"This is how search-and-rescue dogs, drug-sniffing dogs, and bomb-detection dogs are all trained. And, with certain recent exceptions*, they're the best-trained dogs on the planet."

I think he fails to mention that many of these dog are trained with severe punishments - especially the military's mine detection and bomb-detection dogs. This is why the military has dogs who "wash out" of the program - because they couldn't take all the harsh corrections (i think Bob Bailey has a book on this subject). So I'm not really sure how we would compare this to training our pet dogs.

Pack theory is a model we use to describe behavior - a very poor one and poorly understood one at that. We really don't need this pack theory/dominance paradigm to help us train or understand our dogs. Pack theory has been historically detrimental to the way people view their dogs - so I certainly agree with him on that point (as do the majority of dog and wolf professionals). The problem is when we see wolf or dog behavior, it isn't rapped up in tidy little answer or simple explaination. That is what makes multi-dog household problems so complex, dynamic and changing. Many times dog owners are the primary cause of these types of problems because of how they use the dominance paradigam and unknowingly create tension between their resident canines.

Many of the behavioiral studies we have done on wolves and pack theory have been done in captivity. The problem is that wolf behavior is different when studied in the wild (which is obviously much harder to study). That and dogs are not wolves, there are probably just as many differences between them as similarities. We can't even agree if dogs came from wolves, jackals, or coyotes. If dogs are closely related to wolves, did dogs "come from" wolves or did they just share a common ancestor (much like primates and humans). Dr. Ray Coppingers book Dogs is the theory that is most widely accepted by biologists (the garbage dump theory), but there are also many things in that book which biologists don't agree with either. So we can't quote it as fact. We still need to do some more advanced DNA studies to really verifiy the theories that we are using.

Wolf Park (http://www.wolfpark.org/) has stopped interacting with wolves based upon the dominance paradigm, which has resulted in people receiveing few bites from the captive wolves. Dr James O'Heare (my mentor at Cynology College) did his PhD dissertation on dominance theory and published a book on it (http://www.dogpsych.com/). The bottom line of his research and book was that we can't clearly define dominance theory and we are seeing some very problematic trends in applying it to companion dog ownership.

There is really not too many new theories as specifically related to dog training. We use the laws of learning - operant, classical conditioning etc. It is just how people decide to package it. You can use many different +R techniques to teach your dog not to jump on you etc., but they are all rooted in the same principles of learning. I think he is just trying to put a different spin on dog training to sell his product. I guess there is no harm in that..... I think that kind of vibe personally tends to push me away (even if I agree with the content). Dr. Patricia McConnell, Trish King, Pia Silvani, Pat Miller, Dr. Pamela Reid, Brenda Aloff, Karen Pryor, Jean Donaldson, Terry Ryan, Turid Rugaas, Chris Bach, Suzanne Clothier, Dr. Ian Dunbar, Bob Bailey, Dr. Myrna Milan, Dr Karen Overall, Dr. Suzanne Hetts, Dr. Ray Coppinger (I could go on...but I won't) tend not to frame their "products" in this way. These people are just a few of the book writers, researchers, behaviorists, vets, and dog trainers who have made this field what it is...


I will be heading to wolf park next weekend for a 3 day seminar on (http://www.wolfparkstore.co m/store/index.php3?cat=28293 7&item=8297) wolf/canine behavior. I will get to meet Suzanne Clothier, Pat Goodmann and Beth Duman. We will be discussing what is fact and what is anthopomorphic fallacy. I will get to observed and interact with the captive wolves at wolf park (this scares and excites me all at once). Perhaps when I return, I can organize my thoughts in a more scholarly manner...

Hope this makes sense,

[Kendall Aliza]

Yes, you are making a lot sense. I am going to try and do a little of my own research on some of the things you mentioned and see if I can find anymore info about this guy. The wolf park looks very cool! Let us know how it goes and which of the wolves and foxes that you met.

[Amber]

We had a wolf hybrid (98% wolf, 2 % shepard) brought to us at our shelter here in Louisiana. We also do wildlife rehabilitation, so animal control let us keep her, as wolves and wolf hybrids are illegal in our state. Anyway, if you haven't been around any wolves before, Renee, you are going to have so much fun! I know it was a wonderful experience for me...one that I will probably never experience again.

[Lee Charles Kelley]

I was gratified to see some material from my website showing up on a discussion group, and thought I ought to respond a little to what was being said. (Pardon me if I don't know how these dicussion groups work – there may be glitches in the way my comments are presented onscreen.)

Renee:
�I've never heard of this person......his website seems like an infomercial...�

Kelley-
My website is a promotional tool. And it's not an infomercial, obviously, since that's a generally derogatory term used for a specific type of television programming, not internet content. Your use of the term, besides being derogatory, is also irrelevant, sort of like saying you don't like the techniques used by the Dog Whisperer because he's on TV. (I don't like his techniques, either, but his being on television doesn't invalidate them; the fact that they're based on an outmoded, unscientific theory of behavior does that.)

The material on my website exists mainly to promote my training practice and, to some extent, my ideas. That is its main purpose. Education is only its second. I'm sorry if that offends you.

Renee:
�Since when does a wolf or dog "train" another dog to do anything - they don't ...(according to Beth Duman - Michigan's wolfpark rep)....especially in the way we train our dogs.�

Kelley:
Perhaps I should re-phrase that since the point is that dogs (and wolves) learn to do quite a number of phenomonal and amazing things naturally, on their own, without food ever being offered as a reward. In obedience training, food has a natural place in teaching only two, perhaps three behaviors: the sit, sometimes for a recall, and sometimes for the heel.

For instance, a dog will naturally either sit or try to jump up to grab a piece of food that is just out of its reach. This is a natural, predictable behavior and can be used successfully in training the sit. Natural Dog Training uses natural behaviors, which are already "hard-wired" into the dog’s nervous system to produce reliable learned behaviors on command. The truth is, everything we want from our pet dogs, from sitting to staying to coming when called, is already there as part of the prey drive. In fact, the prey drive is actually what makes dogs sociable and trainable. As for using food, it actually tends to get in the way of the natural learning process.

Renee:
"Food is a primary reinforcer (food, air, sex, water to be exact) - it is a survival need."

Kelley-
Okay, first of all, sex is not a survival need. You're mixing the first and second biological imperatives. (My dog has never had sex, he's thirteen, and he's survived!) And, yes, according to operant conditioning, or learning theory, food is a primary reinforcer. The problem is, the laws of learning theory break down where instincts are concerned. Keller Breland, who studied under B. F. Skinner and invented clicker training for dogs (and later taught Karen Pryor how to use it with dolphins), noted this in a 1961 article in American Psychology (http://psychclassics.yorku. ca/Breland/misbehavior.htm). He found that some animals that had been conditioned to produce certain simple behaviors drifted into choosing instinctive behaviors instead. He concluded, "There are definite weaknesses in the philosophy underlying these techniques."

Plus, the survival instinct is not always primary. That is a major flaw in the operant conditioning model of learning. Individual survival often becomes secondary when animals mate (they're heedless to danger). It's also secondary when canines hunt because in order for a pack to hunt large prey, the survival instincts of the individual members have to be blunted or short-circuited, or else they would never pursue anything as dangerous as deer and elk, and would stick to killing small birds and rodents.

Clearly, the canine prey drive has evolved in such a way as to override the survival instincts of the individual pack members. This explains why when a dog’s prey drive is aroused food is ineffective, and why the search-and-rescue dogs at ground zero put their jobs ahead of their own survival (see below*).

It's also why when you use the hunting instinct as the focal point of training, as both inducement and reinforcement, you'll find that a dog will not only ignore food but squirrels, pigeons, and cyclists, and will focus exclusively (and quite happily) on his owner's commands. And though it may sound odd, dogs are always calmer and more sociable when trained via their predatory instincts. Even Pavlov remarked: "Conditioned reflexes, in which positive emotions arising in connection with the perfection of a skill [such as playing a hunting game], irrespective of its pragmatic significance at a given moment, serve as the reinforcement."

(If some of THIS sounds like an “infomercial� it's because it's from another bit of my promotional material that hasn't been uploaded onto my website yet.)

Renee:
"We basically use the Premack principle when dog training ...."

Kelley-
That's good, except that Premack's Principle is also flawed, in that the training doesn't work because a "more probable" behavior can be used to condition a "less probable" one, it works because a more innately satisfying behavior can be used to train a dog to produce a behavior that is not as innately satisfying. The way the principle is worded shows again the underlying flaw in operant conditioning itself. The focus is on the resultant behavior, not on the underlying instincts or emotions that create that behavior.

For instance, with some dogs you might think that since they don't like to play fetch, that playing this game is therefore a "less probable" behavior, and you'd be right. You might even think that playing fetch is not a satisfying behavior for that dog. But if you did you'd be completely ignoring a terribly important training tool.

I recently worked with the owner of a bloodhound who said his dog wasn't interested in chasing a ball; that he'd only chase it for a few feet, then lose interest and start tracking. The dog also had a spotty recall and was just generally uninterested in learning ANY obedience commands.

Because I knew that anything connected to the prey drive will be the most satisfying behavior for any dog, I was able to teach the dog to fetch very quickly (in about two minutes) and then used his interest in the tennis ball to teach several other commands, particularly the recall. If I had believed in Premack's Principle, particularly in the way it is worded, I never would have thought to do this.

Renee:
"I think he fails to mention that many of these dog [detection dogs] are trained with severe punishments - especially the military's mine detection and bomb-detection dogs. This is why the military has dogs who "wash out" of the program - because they couldn't take all the harsh corrections (i think Bob Bailey has a book on this subject). So I'm not really sure how we would compare this to training our pet dogs."

Kelley-
Excellent point, and since Kevin Behan took this into account when he formulated his Natural Dog Training system, any concern for the “humane" nature of training pet dogs this way has already been taken care of.

*(And by the way, you have to be careful where you get your information on this topic from. There's a tendency still, though it's dying out, to favor political correctness over good training. This is why, in my opinion, and in the opinion of many others, many working dogs these days are poorly trained and will sometimes indicate a false positive just to get a treat. A dog trained via his prey drive would never do that. the only problems they had with the search-and-rescue dogs at Ground Zero was getting them to rest. They would have kept searching until they dropped dead themselves rather than stop to eat OR drink. And it's better for a dog to wash out of the program than to go out on an important job looking for liver treats, not bombs!)

[Lee Charles Kelley]

Renee:
"Pack theory is a model we use to describe behavior - a very poor one and poorly understood one at that. We really don't need this pack theory/dominance paradigm to help us train or understand our dogs. Pack theory has been historically detrimental to the way people view their dogs - so I certainly agree with him on that point (as do the majority of dog and wolf professionals)."

Kelley-
It’s good to know that people are starting to catch up with me and Kevin on this. Kevin first proposed the idea that dominance does not exist in the dog's mind in 1992. Ray Coppinger, who's studied wolves in the wild, has observed: that when wolves settle near a dump, and don’t need to hunt in order to survive they "form very loose social arrangements. They have them, but they're not a pack." And since coyotes and other canids also form packs, but only when they need to hunt large prey, it's clear that sociability in canines is dependent on predation, not on rank, status, and hierarchy.

Still, there certainly seems to be a pack pecking order. But that could just be a manifestation of individual variations in the prey drive, not of strata in a social echelon. These variations are a biological necessity for a chase-and-ambush style of hunting, while forming a dominance hierarchy serves no such biologic purpose.


Renee:
"There is really not too many new theories as specifically related to dog training."

Kelley-
Yes, there is one. It’s not just very well known yet. And it’s the only underlying theory that really works all the time with all dogs because it's based on the way dogs really think and operate.

Renee:
"We use the laws of learning - operant, classical conditioning etc."

Kelley-
Which are flawed and incomplete models of learning. For instance, dogs don't necessarily learn through trial and error, through association, or through repetition. When a dog's prey drive is hightly aroused he can learn certain commands once, just once and he will never forget them. No trial and error, no association, no repetition necessary.

Also, praise can be used to stop an unwanted behavior as it's happening. What I mean by this is that under certain conditions, if a dog is doing something you don’t want it to do and you simply praise the dog, very enthusiastically, while the dog is producing the unwanted behavior the dog will immediately stop. If you continue to do this under various circumstances, the behavior will be quickly be extinguished; the dog will simply stop doing the thing he's being praised for.

Example: a family with a shi-tzu puppy, living in an apartment with a dining room/kitchen area, with a large island in the middle of the kitchen. They wanted the dog to stay in the dining room and not wander past the island, into the food preparation area. So, I sat cross-legged on the floor and played fetch with the dog. Everytime she chased the toy and grabbed it in her mouth, I praised her. We set up a rhythm. After a while I threw the toy past the island, into the kitchen. The dog raced after it, and as soon as she got to the far edge of the island, I praised her. She stopped short, then turned around and came running back to me. We repeated this several times until I got her to the point that she would stop automatically at the edge of the island and not go into the kitchen, whether I praised her or not. By praising her at a specific moment, I stopped her from producing a behavior she was being "rewarded" (by being praised) for. How does that make sense using the operant conditioning model of learning?

Renee:
"I think he is just trying to put a different spin on dog training to sell his product. I guess there is no harm in that, but again it feels like an infomercial."

Kelley-
Again, the main purpose of the article was to promote my training system, just as the other parts of the website are there to advertise my mystery novels (which have dog training ideas woven into the storylines). And I'm not putting a spin on anything. I'm offering a completely new way of looking at training and behavior.

Renee:
"I'm a person who takes dog behavior and biology very seriously, so I think the infomercial vibe really tends to push me away (even if I agree with the content). That and I don't see Dr. Patricia McConnell, Trish King, Pia Silvani, Pat Miller, Dr. Pamela Reid, Brenda Aloff, Karen Pryor, Jean Donaldson, Terry Ryan, Turid Rugaas, Chris Bach, Suzanne Clothier, Dr. Ian Dunbar, Bob Bailey, Dr. Myrna Milan, Dr Karen Overall, Dr. Suzanne Hetts, Dr. Ray Coppinger (I could go on...but I won't) selling their "products" in this way. These people are just a few of the book writers, researchers, behaviorists, vets, and dog trainers who have made this field what it is and don't need gimicks to sell their products."

Kelley-
When they write their books on behavior and training they'e not writing websites. You seem to be very intelligent, experienced, and well-read. My suggestion would be to critique the content of my material, not the format.

[Grace Erick]

I've seen scent recognition dogs for the police trained with a toy or rope etc. At least that is what they seem to end up with as a reward for finding drugs etc.

Bye, Grace

[Lee Charles Kelley]

I've seen scent recognition dogs for the police trained with a toy or rope etc. At least that is what they seem to end up with as a reward for finding drugs etc.

Bye, Grace

Yes, and there's a trend nowadays with search-and-rescue dogs for the "victim", during the training process, to lie in wait armed with a tidbit of cheese, so that when the dog finds said victim in the "rubble" he gets rewarded with the cheese, when if he'd been trained properly he wouldn't need the extra reinforcement Just finding a live body (and maybe being given a tug rope from his handler) would be the proper reward.

If you train with cheese, you're gonna get cheesy results.

[Mario Niepel]

Interesting discussion. For starters, would like to say that I am not particularly set on using food as a reinforcer. In fact, I am currently attempting to use tug-of-war as a primary reinforcer. Unfortunately, my dog does not much care for playing tug, so I have to first teach him. I do have a few questions though.

Clearly, the canine prey drive has evolved in such a way as to override the survival instincts of the individual pack members. This explains why when a dog’s prey drive is aroused food is ineffective, and why the search-and-rescue dogs at ground zero put their jobs ahead of their own survival (see below*).

I am not sure that the statement is true that when a dogs prey drive is aroused that food is ineffective. I managed quite well to control my dogs squirrel prey drive using food as a primary reinforcer. In fact, he has gotten good enough to even break off a run at a squirrel tree.

In fact, in nature most predators know very well to interrupt their prey drive in order to eat, drink and rest. In wolf packs I think that roughly only a third of all hunts are successfull. Obviously, predators (unless 100% successfull) have to develop a mechanism to interrupt their prey drive for more pressing needs like eating, driking and resting. Otherwise they'd die out very rapidly.

Which are flawed and incomplete models of learning. For instance, dogs don't necessarily learn through trial and error, through association, or through repetition. When a dog's prey drive is hightly aroused he can learn certain commands once, just once and he will never forget them. No trial and error, no association, no repetition necessary.

The example you give does not invalidate learning theory or specifically distinguishes prey-drive-learning from other learning. A single successful trial falls into the class of trial and error and R+ learning. For example, my dog never reacted to me opening cans of food... until I gave him canned food for the first time. Ever since, he is in the kitchen every single time I open a can.

Plus, the survival instinct is not always primary. That is a major flaw in the operant conditioning model of learning. Individual survival often becomes secondary when animals mate (they're heedless to danger). It's also secondary when canines hunt because in order for a pack to hunt large prey, the survival instincts of the individual members have to be blunted or short-circuited, or else they would never pursue anything as dangerous as deer and elk, and would stick to killing small birds and rodents.

I don't really understand why this is a problem with operant conditioning. As far as I understood it, operant conditioning theory does not say that all behavior exhibited by an animal is elicited via mechanisms of operant conditioning. So, instinctive behavior, like giving chase, can live side by side with behavior learned through operant conditioning.

Also, why is it a flaw in the operant conditioning model of learning, that (as you say) survival instinct is not always primary? Where is the problem in reconciling the two statements.

There are many things to consider before this statement (survival instinct is not always primary) is made. Starving wolves, for example, are far more likely to attack prey even if they might get hurt in the hunt, while wolves that are not starving would rather break of a dangerous hunt and see if they can't find something that might be easier to kill. Also, wolves, when they are not in dire need for food, do not go out to chase large and dangerous animals for the fun of it. They do so only when their survival is on the line. So, it seems that survival instinct may in fact be always primary. The danger of starvation is just sometimes greater than the danger of hunting a large animal.

Even Pavlov remarked: "Conditioned reflexes, in which positive emotions arising in connection with the perfection of a skill [such as playing a hunting game], irrespective of its pragmatic significance at a given moment, serve as the reinforcement."

True, but that is based on regular learning theory. Pavlov also knew that food can serve the same function.

Yes, there is one. It’s not just very well known yet. And it’s the only underlying theory that really works all the time with all dogs because it's based on the way dogs really think and operate.

I guess, like Renee I failed to see the the novelty. But maybe if you can answer my questions, I will get closer to understanding why you deem this theory to be new and different from other training methods used. As far as I can tell, you use prey drive to elicit behaviors (in learning theory it is not really relevant how the behavior is elicited) and you use prey drive (chase or tug) as a primary reinforcer. But I may of course be missing something. Looking forward to your answers... mario

[Rebekah Hartman]

Wow, Mario - you've really done your homework. That's the second time tonight I've been really impressed to read your responses to the more scientific side of dog training and behavior. You've given me some great practical advice in the past on other threads, but I didn't realize there were so many on this forum who really dig in to the meat of the science, theory and philosophy of dogs.

Boy, I really wanted that to sound intelligent but I think it's just getting too late.... All to say it's fun to see the intellectual side of folks on this forum.

[Mario Niepel]

Lee, another question just occured to me: You mention a couple of times that drug/bomb sniffing dogs sometimes give false positives in order to get a treat. How was that determined? How did people tell that the dogs gave a false positive 'in order to get a treat'? Maybe they just made a mistake. In any case, doesn't that just signify that the behavior was just not trained well enough? Finally, since prey drive is such a powerful primary reinforcer, why don't these dogs give false positives in order to get the tug/chase reward?

In any case, that's it for tonight... mario

[Mario Niepel]

Wow, Mario - you've really done your homework. That's the second time tonight I've been really impressed to read your responses to the more scientific side of dog training and behavior. You've given me some great practical advice in the past on other threads, but I didn't realize there were so many on this forum who really dig in to the meat of the science, theory and philosophy of dogs.

Boy, I really wanted that to sound intelligent but I think it's just getting too late.... All to say it's fun to see the intellectual side of folks on this forum.


Rebekah, you give me way too much credit. I'm just a layperson who read a few books. I have virtually no knowledge on the primary literature of behavior and learning theory.

What really got me going was Karen Pryor's book 'Don't Shoot the Dog!'. I can warmly recommend to anybody even remotely interested in behavior (in any animal).

[Renee]

Hi Lee-

I'm sorry if I offened you by using the word infomercial....I went back and edited my post hopefully to your satisfaction...Perha ps it's the difference between my midwestern "Umwelt"(the looking glass through which each of us views the world) and your New York "Umwelt"...


I'm really excited that you could join us here at Doggie Bag. I would love for us to talk about learning theory, prey drive, wolves, canids etc. These are some of my favorite topics!


I have a few questions/comments for you....

Here is a quote from you website:
"A study shows that most dog trainers are only successful 40% of the time. But, by using a dog's natural instincts for obedience, we're successful closer to 100% of the time; with all the dogs we train."

-I was wondering what studies you are specifically referencing. I would be very interested to read them - that is if they are not to expensive to purchase (which most are!).

Quote:
"If all members had a so-called alpha temperament, or a gamma, or a beta, the hunt would surely fail."

-In Canids and primates there is a social heirarchy - alpha, beta, omega. I've never heard the term gamma used. Is this new/old terminology or are you refering to omega?

Quote:
"Perhaps I should re-phrase that since the point is that dogs (and wolves) learn to do quite a number of phenomonal and amazing things naturally, on their own, without food ever being offered as a reward."

- I agree...but I would like to point out that dogs and wolves will do an amazing amount of things for a food reward. The wolves at wolf park would do just about anything for a food reward....not to mention a piece of paper or a little bit a chap stick (I was very careful to empty my pockets before I entered the wolf pens at wolf park - believe it or not chap stick is a very dangerous thing to have when you are amongst 4 or 5 wolves).

Quote:
"The truth is, everything we want from our pet dogs, from sitting to staying to coming when called, is already there as part of the prey drive."

- Lee, could you elaborate what you mean by this....While wolves and dogs have distance increasing/decreasing signals, "comming when called" is not a part of their behaviorial repertoire (PhD. Zoologist Patricia McConnell has written and spoken about this topic)

Quote:
It’s good to know that people are starting to catch up with me and Kevin on this. Kevin first proposed the idea that dominance does not exist in the dog's mind in 1992.

-I'm not saying that dominance doesn't exist (because it does). I'm just saying the way in which many casual companion animal owners use "dominance theory" is incorrect (because of all the misconceptions surrounding it) and it is really not helpful in terms of training their beloved pets.

Quote:
" And since coyotes and other canids also form packs, but only when they need to hunt large prey..."

- Coyotes don't typically hunt large prey...they eat small animals and they scavenge. I learned at Wolf Park that Coyotes cannot bring down large prey (it would consume too much energy) and prefer small rodents. I think Wolf Park's point is that they are not hurting the farmer, they are helping....

-Also, when we are talking about prey drive in wolves as compared to prey drive in canines, we need to consider that canines have gone through many generations of selective breeding. (This is why people who own wolf dogs and wolf dog hybrids have to be very careful with how they house and manage these types of animals - they have low thresholds for predatiory behaviors.) In our canines, it is much harder to trigger predatory behaviors as opposed to their wolf relatives. Every behavior that is exhibited in wolves, is also seen in dogs. But it is the intensity, frequency, duration, context and threshold level is where the differences can be found. In many breeds of domesticated dog the hunting instinct/prey drive has been significantly lowered. So this is why we use many different types of motivatiors depending on the breed/personality of each canine.

Thank you in advance for you thoughts.....

[Renee]

Lee-

I have one more thing...

Quote:
"In fact, the prey drive is actually what makes dogs sociable and trainable. "

A wolf's prey drive is much stronger than that of the domesticated dog. My feeling is that dogs with a stronger prey drive aren't necessarily more trainable. This is why there are problems with wolf dogs and wolf dog hybrids biting and killing children. Here is a quote from Monty Sloan of Wolf Park "The reason you cannot ever trust a pure wolf with children is because of the aforementioned lower threshold for the trigger mechanism regarding predation, and the lack of any alteration of their predatory behavior once the tigger is released....These behaviors are genetically encoded - they cannot be eliminated by "proper socialization" or "training"; and at best can only be suppressed"

What do you think?

[Angel Siders]

Its seems this trainer is only looking at the prey drive, he is missing that there are other drives that dogs go for, ie food drive.

[Lee Charles Kelley]

I managed quite well to control my dogs squirrel prey drive using food as a primary reinforcer. In fact, he has gotten good enough to even break off a run at a squirrel tree.

In fact, in nature most predators know very well to interrupt their prey drive in order to eat, drink and rest. In wolf packs I think that roughly only a third of all hunts are successfull.

According to Coppinger, once the prey-making sequence starts, a wolf will not stop for anything, not food, not rest, not water, not even to go after a smaller animal instead of the larger, more dangerous one. Granted, if the hunt is not successful, and the animals are worn out from their efforts, they have to then look for alternative food source.

As for your dog and squirrels, I would question whether he truly has a strong prey drive or just shows an avid interest in squirrels. But you've seen his behavior, I haven't. The other part of it is that the dog's prey drive is an incomplete form of the one that exists in the wolf's psyche. It's been modified by Man for tens of thousands of years.

I DO know that many dogs who chase the cat, chase cyclists, chase cars, chase pigeons, can't be distracted by a simple offer of food. Perhaps in your conditioning of the dog there were other elements involved besides offering a food reward for giving up the chase.

my dog never reacted to me opening cans of food... until I gave him canned food for the first time. Ever since, he is in the kitchen every single time I open a can.

One of the hallmarks of conditioning behavior through food rewards is that it requires numerous repetitions, with a variable reinforcement schedule, etc. And your dog's learning to respond to the sound of the can opener, on his own, is quite different, IMO, from a dog learning to produce a behavior on command instantly, and that dog always obeying that command when it is given, from that moment on, no matter what else the dog may be feeling or doing at the time it is given.

Going back to the rules of variable reinforcement, one of the things said about this aspect of conditioning is that a behavior will be more reliable if the animal is NOT consistently reinforced. And this is usually true. However, when detection dogs are trained, the animals are always reinforced, every single time they perform correctly. When they AREN'T reinforced consistently, they're not as reliable.

One of the things that search-and-rescue handlers started doing after 9/11, is that when the dogs are at a sight where few survivors are found, and they become depressed or dispirited (at not being able to "complete the hunt"), some of the handlers will go out into the rubble and hide so that the poor animals can have some success at doing their jobs. So why are detection dogs always MORE reliable when they're consistently reinforced, and yet the rules of operant conditioning say that this shouldn't be so? Clearly there are gaps in the logic of learning theory.

operant conditioning theory does not say that all behavior exhibited by an animal is elicited via mechanisms of operant conditioning. So, instinctive behavior, like giving chase, can live side by side with behavior

Okay, according to most proponents of learning theory all behavior comes as a result of either operant conditioning or classical conditoning. There is no room in theory (at least for these proponents) for behaviors that are NOT rewarded and yet repeat themselves despite this lack of reinforcement. For instance, a male dog who seeks out a mate whenever he smells the pheremones emanating from a female in heat, and does so despite personal injury, and continues to do so every single time he detects that pheremone despite the fact that he never achieves his goal.

Also, why is it a flaw in the operant conditioning model of learning, that (as you say) survival instinct is not always primary?

wolves, when they are not in dire need for food, do not go out to chase large and dangerous animals for the fun of it. They do so only when their survival is on the line. So, it seems that survival instinct may in fact be always primary.

As I said before, once the sequence starts, they don't stop until they're successful or they can't physically continue. Food is not an issue once the hunt is on. The point is that the prey drive supercedes the survival instinct in a very real way even durig the hunt itself, though, as you say, the ultimate driving force is to find food.

Pavlov also knew that food can serve the same function.

No, to me dog training is like teaching a kid to play baseball. Certainly you offer the kid positive reinforcement through praise and encouragement, and successive approximation when he makes a good attempt but doesn't quite master the skill-set, but every time he makes a good catch you don't pop a piece of hot dog into his mouth to make he sure he's reinforced. Just successfully catching the ball is reinforcement enough for him. (By the same token, you don't need to give a treat to a dog who's just emptied his bladder -- he's already been reinforced by the reduction of physical tension and pressure on his bladder.) The pride and enjoyment he gets is integral to the learning process.

Like wise, when you train a dog through his prey drive you begin to see that excitement and happiness the dog gets at simply learning how to play the game and to succeed at it. I've worked with dogs who'd be confused and distracted if food were on the table (so to speak) during training. I've also worked with dogs who once the food is available can't concentrate on anything else. They don't learn how to obey, they learn how to get food. I've also seen this type of training increase a dog's anxiety level because the learning process doesn't satisfy a deeper instinctive need, while training via the prey drive DOES.

It sounds backwards to say that food doesn't satisfy a deep instinctive need, but again you have to go back to the natural way of learning. If everything in the learning process is about getting food, you're putting the learning process on a purely survival level ("how can I survive?"), which is an uncomfortable place to be. A nicer place to be comes through training via prey-oriented games ("how can I have fun?"). You might think that fear based training (which is another survival mode) is far worse than food training (and I think it is). If so dominance is the EVIL twin of survival mode training, but treat training, though not as evil, is still its twin.

I'm not saying food can't be used in training. (I'm not saying "dominance" can't be used, either -- it's sometimes necessary to calm down a certain type of dog.) I use both food AND dominance myself in certain circumstances. They're just not the most effective training tools overall.

I guess, like Renee I failed to see the the novelty. But maybe if you can answer my questions, I will get closer to understanding why you deem this theory to be new and different from other training methods used. As far as I can tell, you use prey drive to elicit behaviors (in learning theory it is not really relevant how the behavior is elicited) and you use prey drive (chase or tug) as a primary reinforcer. But I may of course be missing something. Looking forward to your answers... mario


Well, I object to the use of the word "novelty", as it makes the method sound like a trinket in a toy store, rather than the breakthrough (in the understanding of how dogs think and operate) that it is. The main component of Kevin's theory, the one that was new and unheard of back in 1992, is that the pack instinct only exists to serve the need to hunt large prey; that no such thing as alpha or dominance (or submission) exist in the dog's social repertoire. Coppinger said something similar last year, during an online discussion the day after the PBS show on dogs in February.

The other piece of it -- and since you seem to be grounded in o.c./c.c. in a way which might make you unable to see this I'll try to word it as carefully as I can -- is that learned behavior in dogs comes from emotion and from a rudimentary thought process involving finding a way to reduce emotional tension. (And I use emotion here to mean any underlying feeling state.)

In predators, the prey drive is designed to create tension, which creates behavior designed to find a way to reduce that tension. This is where the conflicts between pack members comes from (and which has been so widely misunderstood for so long as the need to rise in status, etc., as explained by the alpha fallacy). When the pack is not hunting, the prey drive has to find an outlet for its tension, and so you get friction among pack members. This friction is also designed to improve the hunting skills of the animals involved. But for the individual animals, most, if not all, behavior is simply about reducing thier own emotional tension.

Now, if you wanted to you could apply o.c. language to this and say that all learned behavior happens through negative reinforcement and you wouldn't be far wrong. (You'd still be a little bit wrong, but not much.)

[Lee Charles Kelley]

Its seems this trainer is only looking at the prey drive, he is missing that there are other drives that dogs go for, ie food drive.

I'm not missing anything. There simply is no such thing as "food drive". It's a popular expression among some trainers but it's not a real thing (unless you're using it as a euphemism for "hunger", which is not a drive but an underlying feeling state). Nor is there any such thing as ball drive, or tug drive, or pack drive. There are only two (maybe three) drives: prey drive, sex drive, and perhaps something akin to pack drive, which is the drive to reconnect when separated from the pack.

[Lee Charles Kelley]

Lee, another question just occured to me: You mention a couple of times that drug/bomb sniffing dogs sometimes give false positives in order to get a treat. How was that determined? How did people tell that the dogs gave a false positive 'in order to get a treat'? Maybe they just made a mistake. In any case, doesn't that just signify that the behavior was just not trained well enough? Finally, since prey drive is such a powerful primary reinforcer, why don't these dogs give false positives in order to get the tug/chase reward?

That's a good question! And I don't mean that in the usual sense, "boy, you got me there!" What I mean is they don't give false positives to get a toy because for the dogs the process of finding bomb scents or drug scents or cadaver scents is about completing the hunt.

Every hunt takes place in five stages: the search, the eye stalk, the chase, the grab bite, the kill bite. When detection dogs are put to work, they're in the first stage. However, during training, all five parts of the sequence (if the dogs are trained properly) are utilized. So when these dogs are trained properly, and they're out working, there would be no underlying emotional need for them to go straight from search to "gimme something to bite". It's not psychologically possible for the sequence not to play out in the order that nature designed it to. Introducing food rewards into the process interferes with that natural process.

(Sorry for the italics on my other response. I don't know how that happened.)

[Mario Niepel]

Lee, you have to bear with me. English is not my first language so sometimes I use words that might have the wrong connotation.

I tried to word my questions as careful as possible to not give the appearance that I am attacking or challenging your theory. I merely have the desire to understand what makes the theory you explain unique and what the flaws of classical learning theory are. I understand that since you have many long posts waiting for you, that it might feel a little like we are ganging up on you. All I can tell you is that this is not the case. And I really am not 'grounded' in oc/cc. I actually only discovered the application of oc/cc for dog training less than a year ago. If something makes sense to me and if it works, then I will try it.

The main component of Kevin's theory, the one that was new and unheard of back in 1992, is that the pack instinct only exists to serve the need to hunt large prey; that no such thing as alpha or dominance (or submission) exist in the dog's social repertoire. Coppinger said something similar last year, during an online discussion the day after the PBS show on dogs in February.

I don't really have a problem with this. But, as I see it, this is somewhat irrelevant to learning or differentiating your method form classical conditioning methods.

The other piece of it -- and since you seem to be grounded in o.c./c.c. in a way which might make you unable to see this I'll try to word it as carefully as I can -- is that learned behavior in dogs comes from emotion and from a rudimentary thought process involving finding a way to reduce emotional tension. (And I use emotion here to mean any underlying feeling state.)

It may be true that I am unable to see something, but I thank you for trying. I am still not quite sure I understand what you are actually saying. You state that 'learned behavior in dogs comes from an emotion'. What do you mean by that? That the emotion makes the dog learn something? That the emotional state of the dog triggers the actual acting out of a learned behavior?

Maybe it would help if you could state in exact language what you mean by 'comes from and emotion'. Also, it may help if you could answer my other questions. I am aware that I did fire a lot of questions at you... but curiosity does usually get the better of me. If there is something to learn I will attempt to do so.

[Renee]

Hi Lee

Because of selected breeding in our domesticated dogs, not all dogs have every step of the prey sequence...livestock guarding dogs would be an example

[Lee Charles Kelley]

[QUOTE=Renee B]
Here is a quote from you website:
"A study shows that most dog trainers are only successful 40% of the time. But, by using a dog's natural instincts for obedience, we're successful closer to 100% of the time; with all the dogs we train."

-I was wondering what studies you are specifically referencing. I would be very interested to read them - that is if they are not to expensive to purchase (which most are!).

It's a statistic I throw out from time to time. It comes from Stanley Coren. I have no idea where he got it from. (It also validates my own, personal, non-scientific observations of the dogs I've known.)

Quote:
"If all members had a so-called alpha temperament, or a gamma, or a beta, the hunt would surely fail."

-In Canids and primates there is a social heirarchy - alpha, beta, omega. I've never heard the term gamma used. Is this new/old terminology or are you refering to omega?

What you're referring to you (and you've apparently studied with wolf "experts") is first, second, and last. My adding "gamma" is partially a dig at the whole idea of social hierarchies in canids, since there isn't one. You've quoted part of my website but not the pertinent part: that the variations in hunting temperament that are a biological necessity for the chase-and-ambush style of hunting have been misinterpreted for years as variations in a nonexistent social echelon.

How does a dog or wolf form an awareness of rankings like alpha, etc., which require symbolic thinking? It can't be purely instinctual, because instincts originate in the hypothalamus while symbolic thinking either originates or is processed in the frontal lobes. A dog's brain is equipped with a healthy hypothalamus, but his frontal lobes are small and undeveloped, so he can't understand or process this kind of information.

In The Symbolic Species (1997) Terrence Deacon, a leading neuroscientist at Boston University, writes, "Species that have not acquired the ability to communicate symbolically cannot have acquired the ability to think this way either." It's quite apparent that dogs don't communicate by using symbols so according to Deacon they can't possibly make symbolic references. But wouldn't they have to in order to understand designations like alpha, beta, and omega? Perhaps there is no hierarchy, at least not from an individual dog or wolf’s point of view.

Quote:
"The truth is, everything we want from our pet dogs, from sitting to staying to coming when called, is already there as part of the prey drive."

- Lee, could you elaborate what you mean by this....While wolves and dogs have distance increasing/decreasing signals, "comming when called" is not a part of their behaviorial repertoire (PhD. Zoologist Patricia McConnell has written and spoken about this topic)

Obeying a trainer's commands isn't part of their normal, natural behavioral repertoire either. But returning to or rejoining the pack when there's an emotional need to do so, is. (Hint: play hide-and-seek with a five-month old puppy.)

Quote:
It’s good to know that people are starting to catch up with me and Kevin on this. Kevin first proposed the idea that dominance does not exist in the dog's mind in 1992.

-I'm not saying that dominance doesn't exist (because it does).

How could the urge to dominate others possibly exist in dogs? What is your definition of dominance? Mine is that the animal has the urge or desire to put another animal into a lesser position. I think you're confusing assertiveness or the need to control access to resources with dominance.

Quote:
" And since coyotes and other canids also form packs, but only when they need to hunt large prey..."

- Coyotes don't typically hunt large prey...

And they don't typically form packs. They do sometimes, though, and WHEN they do it's always about hunting large prey. The wild dogs of Africa, the most social of all canids, even hunt small prey as a pack. Lions are the only social cat in nature, and they also hunt the way wolves do. There is a definite connection between group predation and forming a social unit. Again, there is a clear biological need for variations in temperament, as it relates to the hunt. There is no such clear need for forming a social hierarchy.

-Also, when we are talking about prey drive in wolves as compared to prey drive in canines, we need to consider that canines have gone through many generations of selective breeding. (This is why people who own wolf dogs and wolf dog hybrids have to be very careful with how they house and manage these types of animals - they have low thresholds for predatiory behaviors.) In our canines, it is much harder to trigger predatory behaviors as opposed to their wolf relatives. Every behavior that is exhibited in wolves, is also seen in dogs. But it is the intensity, frequency, duration, context and threshold level is where the differences can be found. In many breeds of domesticated dog the hunting instinct/prey drive has been significantly lowered.

I agree.

So this is why we use many different types of motivatiors depending on the breed/personality of each canine.

Well, my story about the bloodhound reflects both the danger of the o.c./Premackian mindset, and the usefulness of understanding what you've said above. The problem this dog's owner had was that he didn't tease the dog enough before he threw the ball, and when he threw the ball, he threw it too far. That's why the dog's genetic predeliction for tracking took over. Before he ever got to the ball, his nose took over.

But when I teased the dog mercilessly with the ball (building his interest, or drive for it), then tossed only about fifteen feet away, he was energized, happy, totally committed to that part of the prey drive, which though somewhat muted in him due to his breed, was still there to be worked with. And by playing the game with me, his resistance to obedience disappeared. (There was a bit more to it: I also had the dog jump up on me and chase me around.)

The same is true for all dogs, not matter the breed or temperament type. Yes, the fullness of the prey drive is not there in dogs, but enough of it IS there that when you know how to do it, you can work with ANY dog. Or sometimes you use food as a prey object to get the dog started. There are a lot of variations, but it needs to always come back to the prey drive because (as I said in my long, italicized post to Mario), it's the most reliable way to reduce emotional tension, which is how learning takes place.

[Grace Erick]

Hi Lee,

I just woke up and was reading your post about giving the victim cheese for the dog, and I didn't focus enough yet, and thought you meant a victim would just be lying there and they would not help the person until they put cheese on him or her for the dog to find, lol!!!! I was thinking, that didn't sound right until I read it again:)

False/positive: I no nothing about dog behavior, but I would have thought the dog's drive was stronger to sniff out an object than to try and fool the handler for the reward. On tv, they seem to be very focused on their sniffing. I don't see customs agents at the Canadian border with a toy, unless they are using a snack that I can't see or nothing. I always thought rewards, if any were given for a correct behavior, the dogs were supposed to be weaned off of them and just perform the behavior or perhaps only given a reward sometimes during training so the dog does not need a reward after the behavior is learned, kind of like when I rewarded my dog for "doing her business" in the yard.

Bye, Grace

[Lee Charles Kelley]

Hi Lee

Because of selected breeding in our domesticated dogs, not all dogs have every step of the prey sequence...livestock guarding dogs would be an example

Not true. All dogs have the sequence, Renee. It hasn't disappeared entirely (it would take at least a few more millenia for that to happen, genetically speaking). What IS true is that certain aspects of the prey drive have been amplified and others have been muted, depending on what the breed has been used for. But all five steps are there in all dogs. You just have to know the exercises that will bring them out.

[Lee Charles Kelley]

[QUOTE=Grace Erick]I would have thought the dog's drive was stronger to sniff out an object than to try and fool the handler for the reward.

Hi Grace, You're right, it would be if the dogs were trained using their drive exclusively and the trainers hadn't unnecesarily introduced food into the equation.

On tv, they seem to be very focused on their sniffing. I don't see customs agents at the Canadian border with a toy

Drug detection dogs are typically given a toy of some sort when they indicate the presence of drugs. I've read of customs agents in Canada having the dog jump up on them. I know I sometimes use jumping up as a reward in training.

[Grace Erick]

Lee, I have to say that I have never seen if a reward was given or what it was since I have not seen anyone bagged for drugs at the border.

I do see police dogs attack on command, and when they return, they barely get a quick pet and a "good boy" from the police handler. It appears the act of attacking the person seems to be a reward in itself for good behavior, so it "appears" just wanting to please their owner/handler can be their reward in itself.

For baggage claims, I have seen them do a two parter. The dog finds the drugs, then he has to sit by them. I haven't always seen a reward given afterwards, but with tv, you can't always be sure where they edit the tape.

I know dogs train us with positive reinforcement. If we do something to make them happy, they wag their tales or show some type of joy. We are then happy which is our reward with no treats or rope toys:)

Bye, Grace

[Lee Charles Kelley]

Lee, another question just occured to me: You mention a couple of times that drug/bomb sniffing dogs sometimes give false positives in order to get a treat. How was that determined? How did people tell that the dogs gave a false positive 'in order to get a treat'? Maybe they just made a mistake. In any case, doesn't that just signify that the behavior was just not trained well enough? Finally, since prey drive is such a powerful primary reinforcer, why don't these dogs give false positives in order to get the tug/chase reward?

In any case, that's it for tonight... mario

Hi, Mario,

I've posted way too much on all this today but you're right, in a way, there is no way of knowing for sure that the dogs didn't just "make a mistake". It's very improbable that that's what happened, as dogs who are properly trained have a success rate somewhere in the upper 90% range (I believe).

In the Shandra Levy and Elizabeth Smart cases, it's a bit more clear that the dogs used to search the park in Washington DC and the area near the Smart home in Salt Lake City were not doing their jobs properly. Levy's body was found in a place the dogs had searched months earlier (and no, her body hadn't been moved), and Elizabeth Smart could see the search-and-rescue dogs from where she was being held captive in the woods near her home. I believe the dogs in both cases were trained using "o.c." (i.e., food), and not the traditional method, based solely on the prey drive.

And no one has yet to address the successful use of praise to eliminate an unwanted behavior as it's happening.

[Lee Charles Kelley]

Lee, I have to say that I have never seen if a reward was given or what it was since I have not seen anyone bagged for drugs at the border.

I do see police dogs attack on command, and when they return, they barely get a quick pet and a "good boy" from the police handler. It appears the act of attacking the person seems to be a reward in itself for good behavior, so it "appears" just wanting to please their owner/handler can be their reward in itself.

For baggage claims, I have seen them do a two parter. The dog finds the drugs, then he has to sit by them. I haven't always seen a reward given afterwards, but with tv, you can't always be sure where they edit the tape.

I know dogs train us with positive reinforcement. If we do something to make them happy, they wag their tales or show some type of joy. We are then happy which is our reward with no treats or rope toys:)

Bye, Grace

Hi, Grace,

Yes, when attack dogs finish their job they don't need a toy as a reward. If you look at the sequence for the prey drive it starts with the search and ends with the "grab bite" and "kill bite". Detection dogs are using the first part of the sequence so the toy is necessary (to grab with their teeth) to finish the process and leave them satisfied.

Drug detection dogs are routinely given a ball or other toy, even if they don't show it on TV. Dogs trained the new way are sometimes given a treat, which is a huge (HUUUUUUUUUUUGE) mistake. It doesn't complete the sequence and it's irrelevant, psychologically speaking, to what the dog has done.

You're right about dogs training us! Sometimes it's with a tongue on the face!

[Renee]

Quote
"My adding "gamma" is partially a dig at the whole idea of social hierarchies in canids, since there isn't one."

-Could you specifically explain this statement. I would also be interested in specific studies and scientific literature on this topic.

[Lee Charles Kelley]

Quote
"My adding "gamma" is partially a dig at the whole idea of social hierarchies in canids, since there isn't one."

-Could you specifically explain this statement. I would also be interested in specific studies and scientific literature on this topic.


As for explaining it, simple common sense dictates that there can't possibly be a social hierarchy in canids. Still, as far as I know most of the scientific community is still tied to this silly and outmoded idea. Masson and McCarthy mention that this attachment is changing within the scientific community in When Elephants Weep (the real title, I think, as opposed to the one I cite in the material quoted from my website, below).

The following is from Why Elephants Weep by Masson and McCarthy (words in brackets are mine):


In recent years the idea of the dominance hierarchy has become more controversial, with some scientists asking if such hierarchies are real or a product of human expectation ... Some ethologists now argue that while dominance relationships [Rex is more dominant than Spike] may be real, dominance ranks [Spike knows that Rex is alpha and he's beta] are not.

The authors go on to say that scientists have now found that "pecking orders" don't necessarily exist in all chicken groups (which is where the whole thing started in the first place), and that some social hierarchies, previously thought to be ruled by the alpha male are actually controlled by a middle-ranking female!

There are three fatal flaws in the alpha theory-three ideas that, when analyzed properly, don't make any sense.

Flaw #1 - You Can't Pee on a Concept

To a dog, a thing is what it is and that's all that it is. It never stands for something else. Alpha is only a designation; a way scientists have of representing or signifying an animal's rank or status within the social hierarchy. But rank, status, role, and hierarchy are all concepts, symbols, or designations. They are not tangible, sniffable, audible, or visible, which means that they can't possibly exist in a dog's mind. After all, you can't bite, sniff, chase, lick, or pee on a concept.

Some might say that when a dog chases a tennis ball it signifies (or represents) a squirrel or other prey animal. But is that really the case? If we look for the simplest explanation, we see that a dog's hunting instinct is hard-wired to respond to anything moving in a certain way. Think of a puppy on his first walk. Even if he's never seen a pigeon before, the moment he sees a leaf or a bit of paper caught in an updraft he starts to chase it. The leaf doesn't symbolize or represent a pigeon to the dog (especially if he's never seen one before). He chases it only because it's moving in a way that automatically stimulates an unconscious, genetic reflex. But (some might ask) couldn't the recognition of rank and status also be instinctive and genetic? Couldn't the dog's brain be hard-wired for that as well?

No, because there's a huge difference between a simple neuro-muscular reflex (originating in the hypothalamus), and the ability to think and act symbolically (originating in the frontal lobes). Dogs have a hypothalamus but they don't have frontal lobes (at least not very large or well developed frontal lobes) so they can't recognize concepts, symbols, or designations. Without this inherent cerebral ability to think symbolically, how can a dog relate to things like rank and status? He can't. He simply doesn't have the type of brain nor the accompanying cognitive architecture to process them.

So we have to ask ourselves this: when one dog acts submissively towards another is he doing it because a) he recognizes the other dog's rank and status? Or because b) he recognizes that the other dog is stronger physically or emotionally? The answer is probably b. We could go even further and say that the dog doesn't even recognize, cognitively, the other dog's superior emotional and physical strength, he only senses it or feels it. This makes more sense, and yet we could go even further than that-and in so doing be much closer to the truth-and say the dog isn't even able to feel or sense the other dog's superiority. All he can really do is feel the changes in his own temperament when the two come into contact. Still, no matter how specifically we want to look at this, we have to realize, once and for all, that there can never be any recognition or awareness in a dog's mind of his own or of anyone else's rank or status in the pack.

Flaw #2 - There Are No Cocktail Bars in Nature

Alpha theorists seem to think that dominance is the defining characteristic of the pack instinct when it's really just a secondary aspect of the sex drive (the primary one being the actual, physical act of mating). It also may have an influence on two other survival behaviors-eating and sleeping-but let's look at the reproductive aspect first:

The main manifestation of the dominant/submissive polarity in animal behavior comes when two sexual rivals vie for the right to breed with an available partner. This rivalry can only take place between two males or between two females, but never between a male and a female. And never, ever, between a dog and a human. (This behavior occurs in all species, by the way, not just canines. Think of two rams butting heads, for example, or two guys in a bar fighting over a cocktail waitress.)

A second manifestation occurs when a dam steals pups from her less dominant counterpart to raise with her own litter. In some cases-such as when food is scarce-a dominant female may even kill a rival's newborn pups.

There are two other situations where dominance may rightly be said to occur. One is a rivalry over food. The other relates to the best place for sleeping. Still, these are both survival, not social behaviors, because food and sleep are necessary for survival.

As to who gets the best place to sleep, when the space is particularly desirable-a stronger wolf may force a weaker wolf to move so he can lie down in the best spot. But he doesn't do this because he's alpha and the other wolf is beta or omega. He does it simply because he can, because his "will" and desire for the spot are stronger.

Keep in mind however, that whenever dominant behavior does occur (if it does) it has absolutely nothing to do with the pack instinct. Sex is not a social activity for animals-its only purpose is to insure the survival of the species; or more correctly, to insure the survival of the genetic code. Dogs and wolves-no matter how socially developed-are still just dogs and wolves. To them, sex is a completely asocial experience. There are no mixers, dating services, or cocktail bars in Nature.

(Still, people often tell me, "My dog is alpha," or, "My dog is very dominant." This is simply not the case. The language needs to be more exact: the dog is simply "assertive", not dominant.)

Flaw #3 - "Let's Get Together and Kill Us a Moose"

What really sets dogs and wolves apart from other social animals is not the pack hierarchy but how they hunt. The fact is, the pack instinct only exists to enable canines to hunt large prey by working together as a cooperative social unit.

According to Ray Coppinger, in Dogs: A Startling New Understanding of Canine Origin, Behavior, and Evolution, when wolves settle near a garbage dump, and are able to scavenge for a living, rather than having to hunt large prey, the pack's "social structure" becomes much less clearly defined. Other wild canids, such as coyotes and jackals, only form packs when the conditions in their environment make it necessary for them to hunt large prey. When they don't need to hunt large prey, they don't form packs. It's also notable that lions are the only social cats in nature, and they hunt in a similar manner to the way wolves chase and ambush large prey. Meanwhile, the wild dogs of Africa, who are so distantly related to dogs, genetically speaking, that they're practically not a member of the same family, not only hunt large prey as a pack, they also hunt small prey this way as well. And they're the most social mammals on the planet.

The question becomes obvious: is there a direct correlation between sociability and the canine prey drive? The answer is just as obvious-yes there is.

When you look at the alpha fallacy with these three flaws in mind, it makes no sense. No wonder some ethologists are starting to question it. Now, some alpha theorists are suggesting that there isn't just one alpha wolf, there may be as many as five of or six! How much sense does this make to you? However, if you begin to look at the pack from the point of view of a new scientific discipline called Emergence Theory, which began to develop in the late 1950s, you may begin to understand that the pack is not a top-down hierarchy, but a bottom-up heterarchy. Knowing this totally changes how you relate to and train your dog.

[Lee Charles Kelley]

This is another article from my website. As far as I know, there is no one else besides me working on the idea that the pack is a self-emergent system. I hope that will change soon.

Emergence Theory and the Pack Instinct

After reading "Natural Dog Training", I began looking at pack behavior with an open mind and realized there's a simpler explanation than the alpha theory: Kevin Behan was right, the pack's social structure exists solely to enable canines to hunt large prey. But how does the pack organize itself, if not around a "strong leader"? Then I came across a new scientific discipline called emergence theory. Originally used to describe the behaviors of slime molds, emergence now includes ant colonies, the immune system, the growth of urban neighborhoods, voice recognition software, artificial intelligence, and certain aspects of the Internet. It also explains some of the contradictions inherent in the alpha fallacy. According to emergence the pack is not necessarily a top-down, dominance hierarchy in which the behaviors of the individual members are controlled from above, but may instead be a bottom-up, self-emergent heterarchy, in which every member is different yet equal.

On the face of it, this seems absurd. All members are equal? We all know that the alpha wolf rules the pack, right? But when you stop to consider that the pack's only biological function-it's only reason for existing-is to facilitate the hunting of large prey, it makes sense that every member is equal, or at least equally important to the success of the hunt, since it's these differences that enable them to work well as a unit. If all members had a so-called alpha temperament, or a gamma, or a beta [sorry, there's that quote again], the hunt would surely fail. There is a basic need for some members to drive the prey, others to circle and harass, and still others to lie in wait and ambush. It's only when you have this combination of emotional types that the pack (and the hunt) succeeds. And it's this mix of temperaments-necessary only for hunting large prey-that gives people the mistaken impression that the pack is a hierarchy, even though canines are never concerned with abstractions like their rank in a nonexistent pecking order, or their status in an imaginary social echelon, only with the very real and concrete details of the hunt.

Are their abilities amazing? Absolutely. Do they seem super-intelligent at times? Yes. But one of the key features of a self-emergent system is that the individuals within it always display at least one scale of intelligence higher than what they actually possess. So canines are not super-intelligent at all, they just act synergistically as if they were. This explains why so many alpha theorists believe that dogs have the capacity for logic, language, and abstract thought when no such abilities are demonstrable outside the parameters of the alpha theory, and certainly not outside the dynamics of the hunt. Emergence not only explains these so-called abilities, it is actually dependent on them because emergent systems are always smarter than the sum of their parts; in fact, in most cases, the dumber the parts, the smarter the system.

As I see it, each member of the pack is responsible for making his own choices, based on the type of temperament he has. If he has an assertive temperament, his behavior will reflect that. The same is true for animals with cooperative and tentative tendencies. All social interactions between pack members result from this. As the number of such interactions increases over time, clusters of behaviors form and a hierarchical structure seems to emerge. However, the individual members of the pack aren't consciously aware of it, they just behave at times as if they do. So, it seems entirely logical that the pack really is a self-organizing heterarchy (ruled by differences and individual behaviors) not a hierarchy (ruled from above).

The latest trend in training-using food and clickers-continues the alpha fallacy, though food trainers don't talk about an alpha dog, just that dogs need a "strong leader." That's not what they need, though, nor do they understand what a leader is. In an emergent system there are no leaders because the system itself is the leader; a hard concept to grasp, I know; since, according to Steven Johnson in his book, Emergence, even artificial intelligence guru Marvin Minsky-who knows more about emergent systems than anyone-had trouble seeing this at first. What dogs really need and want is to align their individual needs with the needs of the pack as a whole.

This idea of emergence in the canine pack is, as far as I know, mine alone at this point. But new research being done at Yale University may soon lead others in this direction, too. Called the "Feral Robotic Dog Project", cheap robotic dogs have been outfitted with rudimentary computer programs and sensors, designed to detect environmental toxins. Then they're then released into "the wild". The purpose of the project has been to locate hazardous materials without posing any danger to humans (or real dogs). But a side-effect has been the creation of a model of behavior, created by simple computer programs, that mimics an actual pack: The robotic dogs will stop whatever they're doing and move toward, then follow any dog whose sensors are the first to pick up toxins in the environment. [If you look at the computer model of the robotic dog's behaviors when this happens, you'll be amazed to see how, in rudimentary fashion, it resembles a wolf pack, searching for prey.]

If we took this a step further and programmed one of the dogs for other behaviors, such as having the most sensitive sensor, plus the most direct method of approaching the target material, and others in the group were programmed to gravitate to him, based solely on pattern recognition, this would be closer to imitating real pack behavior. You could add as many variables as necessary, and a complete model of pack behavior would emerge, all without giving the robotic dogs a conscious awareness of anybody's rank or status in a mythical dominance hierarchy. Their only motivation would be to find the most direct access to "prey". The upshot is that canines do not form [hierarchical] packs based on who's the most dominant. They do it in order to hunt together.

[Renee]

Lee

I think we are going to have to agree to disagree on the many points that were discussed. I do thank you for taking the time to answer our questions and also for sharing your perspective on canine behavior and learning.

Best wishes,

[Lee Charles Kelley]

Lee

I think we are going to have to agree to disagree on the many points that were discussed. I do thank you for taking the time to answer our questions and also for sharing your perspective on canine behavior and learning.

Best wishes,

No, we certainly are NOT going to agree to disagree! So far you haven't successfully refuted any of my ideas. You haven't even tried! I could certainly agree to disagree with a credible counter-argument against my positions. And if your logic were strong enough I would even agree to agree with you on a given point. But you haven't said anything in this forum that I haven't already taken into account and already dismissed as illogical.

Agree to disagree? What a cop out! For example, you haven't made one comment on the idea that praise can be used to correct a behavior as its happening! I suppose if you feel threatened by the idea that learning theory has holes in it, I can't fault you for that. But I don't agree to disagree.

Granted, you have much of the dog world AND the "scientific" community on your side of the debate. What I have is logic and common sense. And it's my belief that logic and common sense will always win out over hidebound, academic "intellectualism".

sorry if I seem disagreeable in this post,

LCK

[Renee]

Lee-

Quote:
No, we certainly are NOT going to agree to disagree! You haven't made a single argument of any kind refuting my positions. I could certainly agree with a credible counter-argument against any of my positions, and would love to hear them (but haven't).

Granted, you have much of the dog world AND the "scientific" community on your side of the argument. What I have is logic and common sense.

However, it's my belief that logic and common sense will always win over hidebound, academic "intellectualism".


Thank you for discussing your position. Good luck to you...

[Mario Niepel]

Lee, I don't really know if I disagree with you or not-- and I certainly don't have any personal belief system at stake. I am not a trainer, I am not a dog behaviorist and I am not a writer. The reason why I don't know if I agree with you or not is that you are making it really hard to understand what you are actually advocating.

As an example, I really don't understand what 'learned behavior in dogs comes from emotion' means.

Next, you make fairly sweeping statements, like that learning theory has holes in it. Can you explain to me what these holes are? In addition, just because a theory has holes in it (i.e. can not explain everything) that does not mean that the theory is invalid. For example, Newtons laws are valid theories even though they break down when relativistic speeds are approached.

You gave an apparent example where you say learning theory is incomplete. I suggested that this example is not at all with regular conditioning. So far I have not seen another example that would show that classical learning theory are flawed and incomplete models of learning.

In fact, when I asked you to specify how your way of teaching differs from classical and operant conditioning, you actually did not answer. Maybe it was an omission on your part since there were many questions to answer, but this one seems particularly pertinent. So, can you explain to me what kind of learning theory you base your training on that is different from conditioning? Bear in mind though, that learning theory does not make any statements on how and why actions are performed. I.e. whether a dog performs an action to please an owner, a dog performs an action in order not to get whipped, or a dog performs an action to satisfy an inner emotion is irrelevant.

When I browsed through your website where you explain a couple of examples (sit, redirected agression, ...) you are actually applying classical lerning theory. The dog soon learns that it will not be rewarded with a treat by snapping at it, but by sitting calmly. Likewise, the dog realizes that barking and lunging at a dog does not lead to any reward, but staying close to the handler and not lunging and barking leads to a reward (chomping down on a toy). Likewise, the dog learns that lying down leads to continued playing (reward).

So, you might see why I am confused about why you think that learning theory is flawed and how your approach differs.

I did have a few other questions about what you wrote which can be found in my first post in this thread. While I think they might be interesting to you, because they show how some of your statements may need further clarification or further support by facts, the crucial question is the one stated above: What are the flaws in learning theory? Do these flaws invalidate learning theory? What learning theory are your training methods based on?

I hope you will find the time to answer these questions. Cheers... mario

By the way, while interesting in itself, I think discussions about Alphas or Emergence Theory might be interesting, but currently they only confuse the issue (at least for me). Before I can apply your theories to other concepts (like pack emergence) I have to clearly understand your theory of learning in the first place.

[Kendall Aliza]

Glad you are picking up the ball Mario! I too am not sure I understand some things and am looking forward to the rest of this discussion .

[Stefie C]

No, we certainly are NOT going to agree to disagree! So far you haven't successfully refuted any of my ideas. You haven't even tried! I could certainly agree to disagree with a credible counter-argument against my positions. And if your logic were strong enough I would even agree to agree with you on a given point. But you haven't said anything in this forum that I haven't already taken into account and already dismissed as illogical.

Agree to disagree? What a cop out! For example, you haven't made one comment on the idea that praise can be used to correct a behavior as its happening! I suppose if you feel threatened by the idea that learning theory has holes in it, I can't fault you for that. But I don't agree to disagree.

Granted, you have much of the dog world AND the "scientific" community on your side of the debate. What I have is logic and common sense. And it's my belief that logic and common sense will always win out over hidebound, academic "intellectualism".

sorry if I seem disagreeable in this post,

LCK

LCK, I freely admit most of this discussion is way over my head but I've been looking forward to the week-end when I'd have to time to really study your theory's. But the tone of this post is so off-putting I don't think I'll bother. Renee is an trusted & respected member of this community. Your thinking that she, the whole world & scientific community is wrong & you are the only one who's right goes against my sense of logic & common sense.

[Kendall Aliza]

I have to say in his defense he went back and edited it after he realized Renee was not going continue discussing everything with him. It was a lot more respectful the first time before he edited it. I think both sides remained a healthy amount of respect for the other views while explaining their belfief systems. But LCK didnt expect Renee to just up and stop discussing the issues and then giving no real reason, in his, eyes for doing so. I would be quite frustrated to after spending a good few hours of my time trying to talk theory to just have someone walk away without some sort of explaination.

[Kendall Aliza]

His original post looked something like this...

No, we certainly are NOT going to agree to disagree! So far you haven't successfully refuted any of my ideas. I could certainly agree to disagree with a credible counter-argument against my positions. And if your logic were strong enough I would even agree to agree with you on a given point. I could certainly agree to disagree with a credible counter-argument against my positions. And if your logic were strong enough I would even agree to agree with you on a given point. Granted, you have much of the dog world AND the "scientific" community on your side of the debate. What I have is logic and common sense.
sorry if I seem disagreeable in this post,

LCK

[Lee Charles Kelley]

[QUOTE=Mario Niepel]Lee, you are making it really hard to understand what you are actually advocating.

I'm not trying to make it hard; it just is hard. It's a whole new way of looking at things that goes against almost everything we've been taught.

As an example, I really don't understand what 'learned behavior in dogs comes from emotion' means.

First of all, it's not just dogs, but all animals. And I explained it already: behavior comes from a need to eliminate some kind of underlying emotional tension (and I use "emotion" to mean any feeling state). When a dog learns that a certain behavior will reduce his emotional tension, he tends to choose that behavior in similar situations. On the surface this sounds a lot like learning theory, and it is. But there's a big difference in where you put your focus: on the underlying emotion, not on the behavior itself. The Skinnerian model is that behaviors tend to repeat themselves. The way I see it is that the animal learns to choose certain behaviors because of a desire to reduce emotional tension. (And please don't say I'm just using semantics; semantics means how you use words, and how you use words is important here.)

Next, you make fairly sweeping statements, like that learning theory has holes in it. Can you explain to me what these holes are?

I think Skinner's mistake, which many people still buy into, is (again) that behavior comes simply from anything which is reinforcing in some way to the individual organism producing it. It's not that simple. A behavior may provide a positive outcome to a group, but not to the individual member of the group who produces the specific behavior; it may even be detrimental, harmful, or lethal to the animal producing the behavior, yet the animal will predictiably produce that behavior anyway. That's one instance, I think, where Skinner's philosophy breaks down.

If the canine pack is a self-emergent system (and even if it's not), then the needs or pleasures and positive outcomes of the individuals within it are sometimes subservient to the needs, comforts, and ultimate success of the system as a whole. Let's take two examples: mating and protecting the young. Both can be extremely dangerous for the individual organism. In terms of mating, think of two rams butting heads over a potential mate, or two guys in a bar, fighting over a cocktail waitress. The process is filled with aggression, danger, and possible death (if the other guy's got a gun, etc.) When a mother protects her young from a predator she puts her own life in jeopardy and will continue to do so, over and over, even if she's seriously injured each time.

So what I'm saying is that the need of the individual dog is simply to reduce his own tension, and in doing that simple act (attempting to reduce his emotional tension) he's unwittingly serving the needs of the DNA of the species. And if the dog never achieves his own positive outcome, yet he persists in producing the behavior, this would explain why he's driven to do it over and over, even though it doesn't provide him with a positive outcome. It's designed to provide a positive outcome for the DNA of the species, not for the individual organism.

Now, admittedly, these are strong instinctive behaviors, and you said previously that there's supposedly room within the context of learning theory for instinctive behaviors and learned behaviors to co-exist. I say it just ain't so. According to learning theory if a behavior persists it's being positively reinforced in some way. But in both cases mentioned the behavior is not being positively reinforced; just the opposite.

In addition, just because a theory has holes in it (i.e. can not explain everything) that does not mean that the theory is invalid. For example, Newtons laws are valid theories even though they break down when relativistic speeds are approached.

Well, first of all, most people I come across in the dog world don't think Learning Theory has holes in it. To them all behaviors come from either classical or operant conditioning, when the truth is they don't. So what you have is a situation where a lot of people keep expecting the Newtonian/Einsteinan model (or a Pavlovian/Skinnerian one) to hold up at a quantum level (a Behanian/Kelleyesque one).

Secondly, by not thinking this through and examining it carefully, a lot of dog trainers, and by extension, dog owners, make the mistake of not praising a dog for doing something wrong (as one example that goes against the precepts of Learning Theory). Just yesterday I had a session with a coton de Tuelar named Boomer, who'd been biting people. As soon as I came in the door he snarled and snapped at me. And I immediately praised him, enthusiastically. "Good boy! You want to bite me? What a good boy! What a good bitey boy, you are!"

Cut to twenty minutes later and he's lying next to me on the couch with his legs in the air, asking for a belly rub. Why? Because by praising him I wasn't reinforcing his aggression, I was reducing his emotional tension. (Admittedly, this is an encapsulated version of what went on -- there was a bit more to it -- but the very first experience that dog had of me was me praising him while he was trying to attack me!)

You gave an apparent example where you say learning theory is incomplete. I suggested that this example is not at all with regular conditioning. So far I have not seen another example that would show that classical learning theory are flawed and incomplete models of learning.

In fact, when I asked you to specify how your way of teaching differs from classical and operant conditioning, you actually did not answer. Maybe it was an omission on your part since there were many questions to answer, but this one seems particularly pertinent. So, can you explain to me what kind of learning theory you base your training on that is different from conditioning?

I've actually done that already. (And not just in Boomer's story, above.) I may not have answered the question directly to your post, but it's in the material I've posted elsewhere in this thread.

I've got to go now, and will try to answer your other questions later.

LCK

[Lee Charles Kelley]

LCK, I freely admit most of this discussion is way over my head but I've been looking forward to the week-end when I'd have to time to really study your theory's. But the tone of this post is so off-putting I don't think I'll bother. Renee is an trusted & respected member of this community. Your thinking that she, the whole world & scientific community is wrong & you are the only one who's right goes against my sense of logic & common sense.

Dear Stefie,

I apologized in the text of that post itself for my disagreeable tone. It was late when I wrote it and I was in a sour mood.

I apologize again.

LCK

[Mario Niepel]

I'm not trying to make it hard; it just is hard. It's a whole new way of looking at things that goes against almost everything we've been taught.

Touchee. But I am sure that if you are patient, you will be able to explain it to me. I may not be fast, but, like dogs, I do have the ability to learn.

First of all, it's not just dogs, but all animals. And I explained it already: behavior comes from a need to eliminate some kind of underlying emotional tension (and I use "emotion" to mean any feeling state). When a dog learns that a certain behavior will reduce his emotional tension, he tends to choose that behavior in similar situations. On the surface this sounds a lot like learning theory, and it is. But there's a big difference in where you put your focus: on the underlying emotion, not on the behavior itself. The Skinnerian model is that behaviors tend to repeat themselves. The way I see it is that the animal learns to choose certain behaviors because of a desire to reduce emotional tension. (And please don't say I'm just using semantics; semantics means how you use words, and how you use words is important here.)

Lee, the reason why all of this is confusing is because you are bringing two separate concepts together. Why do animal do things (emotional tension relief) and How do animals learn things (...?). Classical learning theory does not make any statements or prediction on the first question. And so far, I have yet to understand what your position is on the second question. Specifically, in your statement the question is 'How does the animal learn to choose certain behaviors that reduce emotional tension?'


I think Skinner's mistake, which many people still buy into, is (again) that behavior comes simply from anything which is reinforcing in some way to the individual organism producing it. It's not that simple. A behavior may provide a positive outcome to a group, but not to the individual member of the group who produces the specific behavior; it may even be detrimental, harmful, or lethal to the animal producing the behavior, yet the animal will predictiably produce that behavior anyway. That's one instance, I think, where Skinner's philosophy breaks down.

Skinner did never make such a statement. Skinner never talked about behavior in general, but about LEARNED behavior. So Skinner's philosophy does not break down. In addition, learning theory can be entirely consistent with an animal learning and repeating behavior that is detrimental to it... namely, if overall the behavior is still positively reinforced. Looked at smoking... it is certainly detrimental to the organism, yet, each act of smoking is positively reinforced. Likewise, I can teach my dog to run across a busy highway by conditioning. Even though you and I know that this can be extremely detrimental to my dogs health, the dog will faithfully repeat this behavior until the dog gets killed.


If the canine pack is a self-emergent system (and even if it's not), then the needs or pleasures and positive outcomes of the individuals within it are sometimes subservient to the needs, comforts, and ultimate success of the system as a whole. Let's take two examples: mating and protecting the young. Both can be extremely dangerous for the individual organism. In terms of mating, think of two rams butting heads over a potential mate, or two guys in a bar, fighting over a cocktail waitress. The process is filled with aggression, danger, and possible death (if the other guy's got a gun, etc.) When a mother protects her young from a predator she puts her own life in jeopardy and will continue to do so, over and over, even if she's seriously injured each time.

Again, this does not really have to do with learning theory. The same statements from my previous paragraph apply.

So what I'm saying is that the need of the individual dog is simply to reduce his own tension, and in doing that simple act (attempting to reduce his emotional tension) he's unwittingly serving the needs of the DNA of the species. And if the dog never achieves his own positive outcome, yet he persists in producing the behavior, this would explain why he's driven to do it over and over, even though it doesn't provide him with a positive outcome. It's designed to provide a positive outcome for the DNA of the species, not for the individual organism.

Again, this is not about learning theory.

Now, admittedly, these are strong instinctive behaviors, and you said previously that there's supposedly room within the context of learning theory for instinctive behaviors and learned behaviors to co-exist. I say it just ain't so. According to learning theory if a behavior persists it's being positively reinforced in some way. But in both cases mentioned the behavior is not being positively reinforced; just the opposite.

And I think this is really where the confusion comes from, so I will repeat the relevant phrase.

Now, admittedly, these are strong instinctive behaviors

An instinct is an INBORN pattern of behavior. That means, instinctive behavior is NOT LEARNED. Nobody needs to show salmon on how to swim upstream for spawning in order to do it. Spawning does not have to be positively reinforced to occur. Salmons know how to do it. Why they do it, be it to relief emotional tension or some other reason is irrelevant. Likewise, chase behavior in dogs does not need to be learned. It is an instinct. Succesful hunting does need to be learned (positive reinforcement of the kill or progression in the hunt) but the chase behavior itself is inborn. So, it appears that you are using not-learned behavior to make statements about learning theory. That makes it really difficult to understand.

Well, first of all, most people I come across in the dog world don't think Learning Theory has holes in it. To them all behaviors come from either classical or operant conditioning, when the truth is they don't. So what you have is a situation where a lot of people keep expecting the Newtonian/Einsteinan model (or a Pavlovian/Skinnerian one) to hold up at a quantum level (a Behanian/Kelleyesque one).

Again, no person with a sound mind thinks that all behavior in animals comes from conditioning. Some people think that all LEARNED behavior can be explained by conditioning. That is a huge difference.

Secondly, by not thinking this through and examining it carefully, a lot of dog trainers, and by extension, dog owners, make the mistake of not praising a dog for doing something wrong (as one example that goes against the precepts of Learning Theory). Just yesterday I had a session with a coton de Tuelar named Boomer, who'd been biting people. As soon as I came in the door he snarled and snapped at me. And I immediately praised him, enthusiastically. "Good boy! You want to bite me? What a good boy! What a good bitey boy, you are!"

Cut to twenty minutes later and he's lying next to me on the couch with his legs in the air, asking for a belly rub. Why? Because by praising him I wasn't reinforcing his aggression, I was reducing his emotional tension. (Admittedly, this is an encapsulated version of what went on -- there was a bit more to it -- but the very first experience that dog had of me was me praising him while he was trying to attack me!)

This is a nice example. However, it is not at all incosistent with learning theory. The dog did not cease his aggression to you because you were actively praising him (i.e. positively reinforcing the aggression). By approaching the dog relaxed and praising, you may just not be a trigger for his aggression. In fact, the dog may have already learned that only fearful people or aggressive people are likely to make his life miserable while condfident praising people do not.

I've actually done that already. (And not just in Boomer's story, above.) I may not have answered the question directly to your post, but it's in the material I've posted elsewhere in this thread.

So, as you see, my questions are not really answered. In fact, some of the answer have made things even more confusing. But, I spelled out clearly where my confusion stems from. Maybe you can explain things to me so I do understand.

Just as a side note: You seem to be a talented writer. You wrote multiple books and you write a lot of stuff on your website. But to make things easier to understand it is really helpful to NOT quote complete articles from webpages, but to try and answer in a more direct and pointed manner. Often the articles on the website are only partially related to the question asked, which makes it hard to determine what part of the information is in direct answer to the question, and what part of the information can be discarded in this particular context.

I've got to go now, and will try to answer your other questions later.

LCK

Can't wait. CHeers... mario

[Rebekah Hartman]

Lee-

It seems to me that part of the confusion many of us are having is that you seem to imply that what we use to describe our human observations of animal behavior (e.g., alpha, pack) is understood by the animal himself. Dog: "I am the 'alpha', therefore I am supposed to respond thus..." Clearly the terminology that humans have coined to describe their observations of animal behavior and the dynamics of interaction between animals is not something that animals walk around thinking about to determine how they'll respond to a given situation or interact with one another.

I'm also interested in learning more about where your research has been published. I don't mean this to sound snotty in any way, either (just in case it does) - I work in healthcare with humans and am always cognizant of whether the studies I'm reading are from peer-reviewed sources. I'm always looking for this type of information related to animals (dogs specifically) as well. I've found it can be somewhat more difficult to determine whether the information I read about my dog is, indeed, accurate as these sources seem to be harder to locate. I've specifically looked for a lot of information about vaccinations and adverse reactions, but I'd like to find research related to learning and behavior as well.

Thanks for taking the time to respond to all of these posts, just please remember that we are a positive group made up of mostly just dog lovers who want the best for our animals. We come here for encouragement, support, assistance and fun - so keep it light!

[Lee Charles Kelley]

[QUOTE=Rebekah Hartman]Lee-

It seems to me that part of the confusion many of us are having is that you seem to imply that what we use to describe our human observations of animal behavior (e.g., alpha, pack) is understood by the animal himself. Dog: "I am the 'alpha', therefore I am supposed to respond thus..." Clearly the terminology that humans have coined to describe their observations of animal behavior and the dynamics of interaction between animals is not something that animals walk around thinking about to determine how they'll respond to a given situation or interact with one another.

You would think that most people would understand this, but the way the theory has been presented makes it difficult for the average dog owner to make the distinction. "If a dog does (such-and-such) it's because he thinks he's alpha," "Dogs from social hierarchies based on dominance," "Dogs don't like learning the down command because it signifies submission and all dogs want to be alpha," etc. The impression given is that dogs have the capacity to think along lines that are actually far beyond them.

I'm also interested in learning more about where your research has been published.

I wouldn't call it "research", I'd call it theoretical thinking. Nevertheless, you can find more in the following:

Kevin Behan, Natural Dog Training; Lee Charles Kelley, A Nose for Murder, Murder Unleashed, To Collar a Killer, and Twas the Bite Before Christmas. Also Coppinger's book, Dogs, A New Understanding is good, though incomplete.

My books are a combination murder mystery, romantic comedy, and dog training manual, rolled into one. But there are a lot of interesting ideas interwoven into the stories.

[Lee Charles Kelley]

Lee, the reason why all of this is confusing is because you are bringing two separate concepts together. Why do animal do things (emotional tension relief) and How do animals learn things (...?). Classical learning theory does not make any statements or prediction on the first question.

Well, that depends on who you talk to. Many pr-type trainers that I've talked to seem to be under the impression that all behaviors, even instinctive ones, will eventually be extinguished if they’re not positively reinforced in some way. That ain't the case.

And so far, I have yet to understand what your position is on the second question. Specifically, in your statement the question is 'How does the animal learn to choose certain behaviors that reduce emotional tension?'

Say an animal is feeling emotional tension and acts in such a way as to reduce that tension. When that same type of tension arises again, he’ll probably try doing the same thing he did before to alleviate those uncomfortable feelings.

There are times when an animal repeats a behavior over and over even though it doesn't reduce his emotional tension or get positively reinforced. Then the owner cries to the vet, a Behavioral Scientist comes in and prescribes drugs to try and solve a problem that could've been solved quite easily by simply showing the dog exactly how to relieve his tension. (Hint: through properly stimulating and satisfying the prey drive.)

Skinner never talked about behavior in general, but about LEARNED behavior. So Skinner's philosophy does not break down.

Skinner said that any behavior that is positively reinforced and that isn't positively punished will tend to repeat itself, and any behavior that isn't positively reinforced, or is positively punished will tend to be extinguished. So you're right. He wasn't talking about just any old behavior that might arise willy-nilly and then immediately disappear. Puppies produce a lot of absurd behaviors that seem to have no meaning or purpose (which make us laugh at their antics). Most of those behaviors quickly disappear. So Skinner was talking about predictable, repeated behaviors. And there are some such behaviors that repeat themselves despite not being positively reinforced and despite the fact they are severely positively punished.

In addition, learning theory can be entirely consistent with an animal learning and repeating behavior that is detrimental to it... namely, if overall the behavior is still positively reinforced. Looked at smoking... it is certainly detrimental to the organism, yet, each act of smoking is positively reinforced.

You’re talking about a behavior that artificially stimulates dopamine (pleasure) levels in the brain. That’s a no-brainer; there is a positive reinforcement built into the experience of absorbing nicotene. Same with heroin, cocaine, alchohol, pot, etc. I’m talking about a dog going after a female in heat. I have observed this behavior on numerous occasions and can see no evidence that the search for the source of the smell is at all pleasurable for the dog. It seems intensely frustrating, if anything.

An instinct is an INBORN pattern of behavior. That means, instinctive behavior is NOT LEARNED.

Yes, but according to Learning Theory, at least as I understand it, any instinctive behavior, once produced, if not positively reinforced, or if positively punished with enough force, will not persist. The animal will learn not to produce that behavior again, instinct or no instinct. Now, that position may have changed after Keller Breland's article on "Instinctive Drift" in American Psychology. I wouldn't know. All the Skinner I've read came before 1961.

The dog did not cease his aggression to you because you were actively praising him (i.e. positively reinforcing the aggression). By approaching the dog relaxed and praising, you may just not be a trigger for his aggression. In fact, the dog may have already learned that only fearful people or aggressive people are likely to make his life miserable while condfident praising people do not.

Well, that is not what Boomer has learned, not at all. He wants to bite anybody he doesn't know, whether they're fearful or not. And you are right, in that my praise didn't trigger his aggression. Why? Because it made him feel safer, calmer, and more secure. In other words it reduced his emotional tension. But how many pr-type dog trainers would understand this? To them the idea of praising an aggressive dog while he’s “aggressing� would be unthinkable because they don’t understand what praise is or what the real process of learning is made of. To them Praise = Positive Reinforcement, and therefore you should never praise a dog for doing something you don't want it to do, when the opposite is the case.

I discovered this aspect of praise one day when my dog picked up a chicken breast in Central Park. He looked at me, immediately dug in, ready to run away. For some strange reason (probably pure pigheadedness more than intuitive brilliance) I praised him very enthusiastically. "Good boy! What a good boy you are!" Surprised, he dropped the chicken breast and ran back to me, where we danced a happy dance together. I repeated this exercise for several days until his scavenging behavior was entirely extinguished. The only thing I did was praise him every time he became interested in a forbidden food object.

I experimented with this for a while and found that, under the proper circumstances, praising a dog while he's producing an unwanted behavior will eventually extinguish that behavior entirely. Why?

Just as a side note: You seem to be a talented writer. You wrote multiple books and you write a lot of stuff on your website. But to make things easier to understand it is really helpful to NOT quote complete articles from webpages, but to try and answer in a more direct and pointed manner. Often the articles on the website are only partially related to the question asked, which makes it hard to determine what part of the information is in direct answer to the question, and what part of the information can be discarded in this particular context.

Sorry, but as you’ve noticed I tend to write very loooooong posts. I want to answer as many questions as I can, but I really have a lot of other things to do. Cutting and pasting ideas from my website is so much easier. I try to keep them in context but I don't always succeed . . .

[Vanessa Lee]

Well, that is not what Boomer has learned, not at all. He wants to bite anybody he doesn't know, whether they're fearful or not. And you are right, in that my praise didn't trigger his aggression. Why? Because it made him feel safer, calmer, and more secure. In other words it reduced his emotional tension. But how many pr-type dog trainers would understand this? To them the idea of praising an aggressive dog while he’s “aggressing� would be unthinkable because they don’t understand what praise is or what the real process of learning is made of. To them Praise = Positive Reinforcement, and therefore you should never praise a dog for doing something you don't want it to do, when the opposite is the case.

In Jean Donaldson's book "Dogs are from Neptune" she suggests, for a dog that is reactive to unfamiliar men, classical conditioning using treats. In the presence of men the owner is advised to supply a consistant flow of the dogs favorite treats (regardless of the dogs behavior) only ending the treat flow when the man disappears and repeating as another man appears. Doing this repeatedly in the presence of a variety of unfamiliar men in a variety of locations is supposed to change the dogs emotional state from tension to relaxed at the sight of unfamiliar men.

Is this similiar to what you are doing using praise or is this completely different?
Could praise be used successfully in the above situation if the owner were to heavily praise the dog for lunging/reacting at unfamiliar men?
Also, could praise stop a dog from alert barking at every little noise?



I discovered this aspect of praise one day when my dog picked up a chicken breast in Central Park. He looked at me, immediately dug in, ready to run away. For some strange reason (probably pure pigheadedness more than intuitive brilliance) I praised him very enthusiastically. "Good boy! What a good boy you are!" Surprised, he dropped the chicken breast and ran back to me, where we danced a happy dance together. I repeated this exercise for several days until his scavenging behavior was entirely extinguished. The only thing I did was praise him every time he became interested in a forbidden food object.

I experimented with this for a while and found that, under the proper circumstances, praising a dog while he's producing an unwanted behavior will eventually extinguish that behavior entirely. Why?

But this only works if the dog is experiencing emotional tension and is engaging in a behavior (undesirable to us) in order to red