Why Dogs Pull on the Leash

Lee Charles Kelley

As for your explanation (and Corvus’), that dogs are interested in seeking out resources, and that the pulling behavior is reinforced, that tells me: a) you haven’t read my article very carefully, and b) you’re imputing a human-like thought process onto the behavior.

 

I read your article, and I thought it was far-fetched and unnecessarily complicated. I also thought I said pretty much exactly the same thing in one small paragraph that actually made sense to dog owners (I hope). Given you're the one talking about Freud, I kinda feel like you're the one putting a human-like thought process onto dog behaviour. I get that you think it's attraction drawing the dog on, but why can't it just be attraction for the inherently rewarding activity of collecting information? Throw in a few delicious surprises here and there and you've got dopamine spritzing all over the place (in Sapolsky's words) and an extremely motivated puller. Pre-sex, though...? What does that mean to anyone except that someone is quite likely a crackpot? If you want to talk to dog people and make a difference you've got to tell them things that make sense to them.

True, you’ve camouflaged it a bit by defining a resource as something the dog wants, but that still pre-supposes that the pulling – or, indeed, as Jaak Panksepp (who was initially an evolutionary psychiatrist, by the way), would call it, the “seeking” – behavior is part of a linear thought process rather than a simpler biological impulse: a drive to connect.

 

I disagree. A dog doesn't have to be conscious of their innate desire to seek information to want to seek information and find it rewarding to do so. The pulling is just a side-effect, as it was pointed out by Kim that you need two to pull. As far as I'm concerned it is an impulse, and it is an attraction, and it is rewarding. All under Panksepp's SEEK banner with a little Behaviorism thrown in there. How could that be?? Someone making sense of dog behaviour with Behaviorism?? Madness.

Going back to the “pre-sexual” aspect of my article, it’s clear to me that Freud was right in saying that “Even though it is certain that sexuality and the distinction between the sexes did not exist when life [on earth] began, the possibility remains that the instincts which were later to be described as sexual may have been in operation from the very first." As I put it: “Atoms need to connect to one another in order to form molecules. Molecules form connections so as to evolve into living organisms. Living organisms are vitally driven to connect to sources of energy: air, water, sunlight, food, etc. The human body and brain operate together through myriads of connections: we couldn't sustain life without them. It makes sense that the body's connectors and connectees would need to have some form of [energetic] attraction to one another in order to ‘hook up.’”

 

Okay, but when you reduce something to its most basic parts it doesn't mean you can automatically use information about those basic parts to infer things about the whole. Lots of things get in the way. Lots of barriers. Like skin, for example. You know the top few layers are dead, right?

I also make a very clear comparison between the “non-realness” of my hypothetical drive to connect and the “non-realness” of positive reinforcement (and +R is just a clinical outgrowth of Freud’s pleasure principle, by the way). If my argument is correct, it means that any argument for +R and against the idea of the pre-sexual drive to connect is more a matter of personal choice than of science.

 

Then why argue at all?

I wrote: “The truth is, positive reinforcements are not actual, physical objects any more than my hypothetical drive to connect is. They're more akin to a function of statistics, measured solely in terms of a behavior's response strength. We can only know if a tangible object, such as a toy or liver treat, might or might not have provided the mechanism for reinforcement by interpreting the resultant behavior after the fact, through a +R lens. Since it's also possible to interpret any behavior through the opposite lens6, behavioral science loses credibility in this regard. Then, once you add the necessity for determining what kind of reinforcement schedule was at play (and there are far too many to list here)7, it simply boggles the mind how anyone can say they know with any certainty, other than as a pure leap of faith, that any behavior of any kind has been reinforced, or what the mechanism of reinforcement actually was.”

 

I disagree that +R is "unreal". Is the behaviour increasing? Chances are it's +R. Measure it if you want. If you want to know what the reinforcer is, come up with some candidates and test them. Don't get me wrong, I'm doing my PhD on cognition and affective state in dogs. I'm not a huge fan of Behaviorism, but the thing is, I'm only interested in building on it to make it more useful, not throwing it out.

Meanwhile it's hard for a dog to form a cathexis with clicker, or with someone who's dominating him.”

 

Well, the clicker at least is arguable. Steven Lindsay feels that shaping with a clicker doesn't wholly belong in operant conditioning in the first place. He thinks it falls under the SEEK mode, if I interpreted him correctly. The click is both a pleasant surprise and information. I know that my little dog gets to a nice place with clicker training where he starts caring more about the click than the treat. I can drop the treat and he'll just ignore it and push for another click.

Liesje

I think the "tree" method works great with a lot of dogs and is very easy, no complex training or tools needed, but it doesn't always work when dealing with a dog with much higher drive.  For high drive dogs, being restrained builds frustration, and frustration actually increases the level of drive and agitation. 

I've had very little success with the tree method - I find it easier, smoother and more successful to change direction, and pair it with practising "heel" frequently off lead and on lead, so that Next To Me is a rewarding place to be.  I don't think that's because I've had a high drive dog as such.... just what I find easier personally.     The basic premise is still the same.... dog is pulling because he wants X, X is removed, dog stops pulling, dog is granted access to X if possible, or given some other reward on an intermittent basis.  I always thought that being a tree doesn't always work because it doesn't always remove X, so the dog keeps trying... the influence of drive wasn't something I'd thought of.

corvus

Well, the clicker at least is arguable. Steven Lindsay feels that shaping with a clicker doesn't wholly belong in operant conditioning in the first place. He thinks it falls under the SEEK mode, if I interpreted him correctly. The click is both a pleasant surprise and information. I know that my little dog gets to a nice place with clicker training where he starts caring more about the click than the treat. I can drop the treat and he'll just ignore it and push for another click.

Sooooo interesting to me.  I've often marvelled at how a dog can work so much harder and happier for a CT than just a.... well, T.  That would explain it, no?

Yep, another person who can contest that the "be a tree" method had limited uses in very active, driven dogs. I use restrained recalls, and build focus and drive control with restraint by using restraint by the collar and saying "Readddyyyyyyyyyy, readdyyyyyyyyyyyy, GO!".

I can use it with Zipper, and he learns quite well. But with Gaci it was the same old thing. Stop, slacken, leash, begin moving, right to the end of the leash. Gaci is so much quicker than I, and is so goal-driven, that it looked just like a new driver learning to drive a standard. There was no fluidity, and it never really helped her. I will use it to stop and wait for her to return attention to me if she diverts her attention during training, but not as a leash-walking exercise.

Chuffy

dog is pulling because he wants X, X is removed, dog stops pulling

In my situation I dont "remove x",  "x" can co exist with my dog, i just teach my dog that not because "x" is there that means he needs to get to it.

With other dogs you cant "remove x" simply because "x" is not there yet but they are looking for it (pulling on the leash in the process). Therefore if the dog knows that he wont get "x" anyway he wont look forward to find it.

Thanks for the detailed reply. It's very helpful in terms of seeing how I should have explained some of my ideas more fully. Also, I can see that my attempt to use pulling as the basis for an exploration of canine consciousness might've gone off track with some people. For example Kim made the comment that pulling is caused solely by the opposition reflex, and that if you drop the leash the dog stops pulling. That's fine if the article were simply about pulling, but it's about how dogs are pulled toward things that attract their instincts. Yes, dropping the leash will stop the pulling behavior, but it won't stop the underlying process of being pulled on by objects of attraction.

I can also see how bringing Freud into the discussion may have been much too off-putting for some tastes.

Yes, I anticipated that some people would feel that way, hence my statement: Now, some might complain: "Why complicate things with this pseudo-Freudian, unscientific gibberish about invisible sexual energies, and this imaginary drive to connect?"

To me, what I've proposed -- that dogs pull because of an unconscious drive to connect, energetically, with things that attract their instincts and emotions -- is actually less complicated, in terms of the levels of consciousness necessary to explain the behavior, than the idea that dogs are seeking information from the environment. I agree that, on first glance, the second idea (yours and Panksepp's) may seem simpler, but it's dependent on higher levels of thought than dogs could reasonably be capable of. The dog would have to have a sense of self and other (in order to see himself and the environment as two separate things), and have a linear, "if/then" thought process (if I do X then I'll get Y), which is the simplest form of logic.

Meanwhile, my idea may sound more complicated because it's a bit convoluted, i.e., one has to remove several layers of thought in order to see the process as just a more complex expression of the same evolutionary force found in chemical bonds, single-celled organisms, etc. I can well understand how, on first glance, that may sound far-fetched and unnecessarily complicated. And I'm sorry now that I took out a paragraph explaining Freud's view that the evolution of consciousness -- which can be clearly seen in the layers of the human brain, with newer forms being laid down on top of older ones -- must be governed by the laws of physics, particularly the first law of thermodynamics.

I know it may seem as if I'm making things even more complicated by bringing the conservation of energy into the equation, but the gist of Freud's view is that evolution can't invent new forms of consciousness any more than nature can invent new forms of energy. Both nature and evolution take the laws of physics, as given, and create different, and in most cases, more complex forms, from the same material. This is consonant with Darwin's idea that the different levels of consciousness come from a difference of degree, not of kind.

With that in mind, it makes more sense, and is more parsimonious, to suppose that dogs "pull" because of an older, less evolved form of consciousness than that they do so because they have a sense of self and other or the ability to form linear, logic reasons for their behavior.

Corvus: I disagree that +R is "unreal". Is the behaviour increasing? Chances are it's +R.

Positive reinforcement is not a thing, event, or marker. It's a way of determining, after the fact, what took place from a behavioral science standpoint. And while it may be as helpful as pi is when determining the circumference of a circle, it isn't any more real, physically-speaking, than saying that the pulling behavior was caused by 3.14159.

Corvus: I'm not a huge fan of Behaviorism, but the thing is, I'm only interested in building on it to make it more useful, not throwing it out.

Well, I think, at some point, it has to be thrown out. The more you build on a shaky foundation, the worse off you are.

I recently spent a great deal of time discussing the flaws in behavioral science with a behavioral scientist. And it was his view that in the minds of real academic behavioral scientists, the subject is far too complicated to be truly understood by most dog trainers, even those who've studied the subject. He said that even Karen Pryor gets things wrong.

I agree that saying a specific behavior was the result of positive reinforcement is a short-hand, user-friendly way of explaining the learning process in dogs. But it's only when you can describe the process in terms of the laws of energy, that you get closer to the truth.

"My dog pulls!" an owner complains.

"Well, how much energy is he expending on this behavior?"

"I don't know. He pulls really hard, especially when he sees a squirrel or another dog."

"Okay. Instead of seeing him as being bad or misbehaving because he's pulling toward toward those things, or that he's pulling away from you, what if you saw it as a form of energy, like a kind of emotional magnetism?"

"That's interesting. But what does that mean exactly?"

"What if he's trying to connect to squirrels and other dogs because they carry a force of attraction that pull on his instincts and emotions? And what if you could actually use that energy, turn it around, and get him to feel emotionally magnetized to you instead?" 

"How do I do that?"

"Play games, like tug and chase me, anything that will make you more interesting to the dog's instincts and emotions."

Is that more or less complicated, more or less helpful, more or less exact, than telling owners that their dog is pulling because he's seeking information from the environment? What information can the owner give her dog that will stop him from pulling? If the dog's goal is to garner information, vital to the dog's adaptive, survival needs, where does that leave the owner? But if the dog's goal is to connect to objects of attraction, then the owner's goal is much clearer: become more of an object of attraction, just as you were when the dog was a puppy.

Thanks again,

LCK

tiffy

Liesje

I think the "tree" method works great with a lot of dogs and is very easy, no complex training or tools needed, but it doesn't always work when dealing with a dog with much higher drive.  For high drive dogs, being restrained builds frustration, and frustration actually increases the level of drive and agitation.  Restraining dogs this way is used a lot in training and drive building.  So if you've got a dog with super prey drive and "be a tree" when he sees a squirrel and starts fixating, you are not going to achieve a whole lot.  For a dog like this (like Nikon), I think it's easier to combine management (avoiding or putting distance between you and whatever arouses the dog) with training incompatible behaviors.  For example, Nikon cannot "sit and watch me" while fixating on a squirrel, or he can't fixate on a squirrel if I simply do an about turn and take him out of sight.

 

Agreed Liesje. Tootsie has 0 prey drive.

Agreed, Liesje.  Sequoyah has overdrive lol.  Which is why Kim's oppositional reflex statement resonates with me.  For her, the leash was initially very frustrating.  So, unlike Sioux, who is more like Tootsie, Sequoyah learned her skills sans leash.  I just made it more rewarding to be next to me than not to be.  The combination of management and skill-building that you mention takes time and timing to instill.  But, in the end, I now have a dog that can sit still and watch me with distractions going on in the environment.  I think one aspect of training that is often overlooked, or misunderstood, is the building of duration, the gradual addition of distractions, and proofing.

LCK: Thanks for the detailed reply. It's very helpful in terms of seeing how I should have explained some of my ideas more fully. Also, I can see that my attempt to use pulling as the basis for an exploration of canine consciousness might've gone off track with some people. For example Kim made the comment that pulling is caused solely by the opposition reflex, and that if you drop the leash the dog stops pulling. That's fine if the article were simply about pulling, but it's about how dogs are pulled toward things that attract their instincts. Yes, dropping the leash will stop the pulling behavior, but it won't stop the underlying process of being pulled on by objects of attraction.

I can also see how bringing Freud into the discussion may have been much too off-putting for some tastes.

Lee, I think your intent might have been clearer had you used different terminology.  Dogs are attracted to things, but nothing is pulling them there, they're going of their own volition.  I'm attracted to chocolate, and I may choose to go find some, but it isn't literally pulling me anywhere.  When opposition reflex comes in to play the dog IS being pulled - by us.  So, he pulls back. 

As to Freud being off-putting,  I was a psych major, so maybe I *have* had enough of that guy to last me a lifetime.

spiritdogs

Lee, I think your intent might have been clearer had you used different terminology.  Dogs are attracted to things, but nothing is pulling them there, they're going of their own volition.  I'm attracted to chocolate, and I may choose to go find some, but it isn't literally pulling me anywhere.

Hi, thanks for the feedback. It's very helpful.

Actually, I chose that phrase -- that dogs are pulled on by objects of attraction -- for a very deliberate and specific reason.

Going back to my article, "Objects of Attraction: Why Dogs Chase Moving Objects" (which you might not have read or might not've thought relevant), I make the point that when we're at the movies (particularly suspense films), or engaged in watching a sporting event, or playing video games, we identify with the observable action not just through our mental thought processes ("Watch out! He's hiding in the closet!" or "Come on, you stupid ball, go over the fence!";) , but through actual, though minor, physiological responses, which are a kind of sympathetic vibration not only with the movements of the actors or athletes, but very often with the movements of the baseball, football, or hockey puck. And that I  believe this purely unconscious physiological response in humans has an equivalent in how and why dogs are attracted to moving objects such as Frisbees, squirrels, and tennis ball because "the objects create an emotional displacement in [the dog's body, in much] the same way that heavy objects create physical displacement in a body of water."

So in that sense, I would say that both the pulling and chasing behaviors may, in part, be volitional, yes. But that they're also largely the result of this kind of virtual (not literal) feeling of displacement a piece of wood feels when caught in the tide. In other words the dog is acting partly through his own volition, true enough. But he's also, at least partially, carried along on a wave of instinct and emotion, just as a surfer dude can only control the movements of his own body in relation to the motion of the wave, but he can't control the wave itself.

As for chocolate, I'm pretty sure some people would say that their attraction to it is less than volitional, too.

Thanks again,

LCK

Kim_MacMillan

Dogs pull on the leash, because something is pulling back on the dog. Simple oppositional reflex built into all mammals.  It takes two for pulling to occur. Drop the leash, and the dog is no longer pulling. 

Not always, while i agree that if you "pull back" that might trigger a "tug of war", the simple fact of dropping the leash wont always stop the dog (of course if we go to semantics we can argue that yes it stops the dog because now he is not pulling anything literally)

But once again it's all about  the "x factor" i mentioned in my last post. We will have to agree that if a prey driven dog sees a squirrel and starts pulling, dropping the leash wont make him stop, to the contrary it will be like giving him the green light to start chasing the squirrel. Now if we add that if this could happen on a busy street, dropping the leash thinking the dog will stop pulling will be actually a very dangerous thing to do.

Yes, some dogs might stop because if the human drops the leash then they "loose" the connection that is telling them that the human is behind and they will turn around to see if the human is still there. I think that would work more with not so independent breeds. I could name so many other breeds that could not care if the human is there or not, in their mind, the human better keep up with them once they drop the leash.

The worst enemy for the human and at the same time the best friend for a dog during the walk is: excitement. If you can control the dog's excitement you can control the dog.

espencer

Chuffy

dog is pulling because he wants X, X is removed, dog stops pulling

In my situation I dont "remove x",  "x" can co exist with my dog, i just teach my dog that not because "x" is there that means he needs to get to it.

Most people with a trained dog report the same - I was speaking of a hypothetical dog who hadn't learned this yet.

espencer

With other dogs you cant "remove x" simply because "x" is not there yet but they are looking for it (pulling on the leash in the process). Therefore if the dog knows that he wont get "x" anyway he wont look forward to find it.

Do you disagree then that dogs don't find the search ITSELF inherently rewarding?  Because that explanation makes a lots of sense to me.

espencer

But once again it's all about  the "x factor" i mentioned in my last post. We will have to agree that if a prey driven dog sees a squirrel and starts pulling, dropping the leash wont make him stop, to the contrary it will be like giving him the green light to start chasing the squirrel.

Of course.  No one has disputed that.  Sheesh, if my dog were pulling because he wanted something across the road and I dropped the lead, HOW stupid would that make me?!!!  No one is suggesting that it will stop the dog moving forward or that trying it is a good idea.  ALL Kim was saying was.... if you drop the lead, the dog is no longer pulling.  He might still move forward, but he isn't pulling.  Another way to phrase it is: it takes two to pull.

Yep, I'm not discussing what the dog is pulling "towards", I'm discussing the pulling "behaviour". Because it's the behaviour you change. *G*

Of course if you drop the leash the dog will continue doing what it was doing when it was pulling. I'm just classifying that the only reason dogs can *pull* is because something is pulling back on it, and the dog has learned that it is okay to continue that tension, and that it doesn't stop the goals of the dog.

espencer

The worst enemy for the human and at the same time the best friend for a dog during the walk is: excitement. If you can control the dog's excitement you can control the dog.

I want all my dog's interactions and time with me to be exciting, the more exciting the better.  As I explained earlier, for a high drive dog with an "edge", I just use an incompatible behavior.  No pulling AND I don't have to desensitize him to whatever makes him want to pull.  I like my dog's prey drive the way it is, but I can take him on a walk with two other dogs, a jog or bikeride, do a BH routine off leash in a field full of wildlife, etc without the dog pulling or darting off.  Why?  Because I've raised him to be equally excited about working with me and the rewards that I provide.  I use toys and play games with the dog that mimic chasing prey.  The dog does not know a difference between training and play.  So I guess I disagree with most people in this thread in that I don't use the "tree" method, I don't turn around and move away from what excites the dog, I don't seek to squash his excitement or prey drive, and I don't do tons of desensitization and proofing.  I think a lot of it has to do with bonding and respect.  I see so many training and behavior issues that just go back to a lack of bond and meaningful interaction between the dog an owner.  Living with a dog and training even to a high level should really not be that complicated, even with a dog that has high prey drive, is dominant, and can be reactive (all describe Nikon).  It should not be a constant struggle or need months of proofing.  I'd rather spend several months learning how to be silly and play with my dog, take him on a walk and see him ignore squirrels because he is automatically looking to me for fun than months trying to walk him without pulling.  I guess for me I like to keep it simplified.  The dog is either paying attention to the handler and willing to obey or he's not.  I like to set that threshold very high from the beginning and not waste a lot of time slowly pushing it up.

Lee Charles Kelley

To me, what I've proposed -- that dogs pull because of an unconscious drive to connect, energetically, with things that attract their instincts and emotions -- is actually less complicated, in terms of the levels of consciousness necessary to explain the behavior, than the idea that dogs are seeking information from the environment. I agree that, on first glance, the second idea (yours and Panksepp's) may seem simpler, but it's dependent on higher levels of thought than dogs could reasonably be capable of. The dog would have to have a sense of self and other (in order to see himself and the environment as two separate things), and have a linear, "if/then" thought process (if I do X then I'll get Y), which is the simplest form of logic.

Yes, you said that, and I fail to see why a SEEK mode necessitates consciousness is what it comes down to. Why does a dog need to have a sense of self and other to seek information? He doesn't need to know he's seeking information. He just needs to satisfy a deep-seated urge - that of seeking information.

Incidentally, there was a paper I came across that I think you should read, but I can't remember what it was called or where I saw it. I will try to find it. It's to do with the surprising ability of some animals to develop a linear sense of events in time, I think. I will say that quite to my shock and surprise, Erik when he was a pup started to get enormously excited when my partner called up to tell me he was on his way home. Erik's excitement would gradually increase until it peaked when my partner walked in the door 40 minutes later. WTF? I have no idea. But it stopped when I started sitting him on his mat with a pig's ear or Kong when the partner was due home.

With that in mind, it makes more sense, and is more parsimonious, to suppose that dogs "pull" because of an older, less evolved form of consciousness than that they do so because they have a sense of self and other or the ability to form linear, logic reasons for their behavior.

By my way of thinking there is no reason, consciousness or logic to it. Dogs naturally seek information about their environment the same way they naturally chase and bite things, and they find it inherently rewarding in of itself because they are programmed to. The same way many of us find doing internet searches on things inherently rewarding.

Corvus: I'm not a huge fan of Behaviorism, but the thing is, I'm only interested in building on it to make it more useful, not throwing it out.

Well, I think, at some point, it has to be thrown out. The more you build on a shaky foundation, the worse off you are.

We'll have to agree to disagree, then. It may be flawed, but it's the best thing we've got for now IMO. The laws of energy don't explain animal behaviour to me any better than an understanding of evolution and learning theory does. What helps me even more is an understanding of prediction-control expectancies, but just because I like prediction-control expectancies better than operant conditioning doesn't mean that if I reward a behaviour it won't keep happening.  

Is that more or less complicated, more or less helpful, more or less exact, than telling owners that their dog is pulling because he's seeking information from the environment? What information can the owner give her dog that will stop him from pulling? If the dog's goal is to garner information, vital to the dog's adaptive, survival needs, where does that leave the owner? But if the dog's goal is to connect to objects of attraction, then the owner's goal is much clearer: become more of an object of attraction, just as you were when the dog was a puppy.

Well, I like being the gatekeeper. When my dogs want to check something out I usually check it out with them. Rather than say "No, you need to stay close to me" I say "You reckon that will be interesting? Then let's go check it out together!" I rarely see them happier than when I go on an exploratory adventure with them. I've been idly thinking for a while about setting up exploratory adventures and rigging an area with "surprises" the dogs can discover with my help. *shrug* Maybe I'm mad, but I think the more you direct the fun activities your dogs get up to when out and about the more inclined they are to pay attention to you and the easier they are to control off leash. It's all just habit. For leashes, I consider pulling a bad habit, not an indication of some underlying relationship problem. It's easy enough to have pet theories for why they pull, but in the end it is the same old thing. We have a bad habit and we need to break it and form a new habit. Whether that be to pay more attention to the owner, to respond to a tight leash by returning to the owner, or whatever other habit you want to establish, it's the same process. Bit of Premack, bit of distracting, take what has been working out of the bad habit and provide an alternative to reward. This as opposed to, say, barking, which is something I would consider as communication rather than a bad habit. What is he saying? How do we change his life so that he no longer has a need to say that?