RSS Feed

Behavior Chains Part 8

Posted on

This blog will explore a second circumstance under which I’m likely to reward a dog who is making errors within a behavior chain. In this circumstance, If a normally un-engaged dog exhibits a significantly more enthusiastic attitude than what is typical for that dog, then a reward could be in order.  In effect, I’m rewarding an attitude at the expense of a behavior.

To be clear.  I’m not rewarding the dog to CAUSE the improvement in attitude, but if the dog brings it to the table then I want to acknowledge it.  The dog’s improved attitude drives my behavior; I am not using the cookie to drive the dog’s behavior.

Here’s a human example to give you a a better idea of my reasoning.

Imagine that you’re a relatively reserved person by temperament, but you find yourself telling a story with great excitement.  Just as you’re getting completely wound up, your listener interrupts to ask you to speak more slowly and clearly.  While you might finish your story, the interruption will likely deflate you, and now it’s just not so much fun to tell your story.  If you’re normally confident and excited, then an interruption might be just what you need to help you improve your storytelling skills, but if you’re less confident or more reserved by nature, then this could kill your storytelling career altogether.

Your excitement level is directly related to your belief that your audience is also captivated by your story; that you’re on the same page.  While the interruption will probably slow you down,  it will also change how you feel about telling your story and quite likely,  how you feel about the person who interrupted your enthusiastic start.

Let’s consider our dogs.  The question to ask yourself is this:   “Is breaking the  flow of training on this occasion worth the cost?”  Keep in mind that you’re actually interrupting both the behavior chain AND the dog’s enthusiasm at that moment. If your dog tends to struggle with attitude then it might make more sense to focus on that factor than on the precision chain.

Both speed and precision can be trained but dogs (like people) have natural tendencies.  I find it much easier to train precision into a “high” dog than to add speed to a “low” dog. (The same is true of children.  Getting a shy child to perform with confidence is much harder than teaching self control to an exuberant child)   As a result, I will reward a dog showing unusual amounts of  energy and confidence to get a better overall picture and more long term toughness, even if that means I’m allowing precision errors.  I ignore the errors to maintain the “flow” of a fast paced training session, because restarting chains always breaks that flow.

A great dog training session is a conversation.    A dog who comes to a training session especially engaged offers a valuable opportunity to build an exciting and worthwhile relationship.  Your dog’s attitude needs to be nurtured if you want to keep it!  You may not agree with every word in the conversation, but don’t interrupt.  Over time, as your storyteller’s confidence grows, you’ll have a chance to get in your opinion. In the meantime, let your dog be the storyteller; the center of attention!

Let’s use my young dog Brito as an example.  Brito is a happy worker, but he is not a high energy or  intense working dog.  I have noticed, however, that he does have some sessions that are clearly “much better than average” in the attitude department.  When those sessions show up, I run with them!  Some people prefer to get the precision first and then bring the speed.  Others work for speed and allow a bit of slippage in accuracy on occasion.  I’m in the second camp.

When Brito is in one of his really good moods, then I reward that intense CER above all else, even though his actual work might be something of a mess.  Super happy dogs tend to forge in heeling.  They forget how to sit straight.  They retrieve extra fast but sometimes drop the object because they never had a good hold on it in the first place.  The examples continue; speed and enthusiasm tend to come with some specific challenges for accuracy.

To avoid ruining our accuracy, I focus on work which allows Brito to express his energy, such as heeling, retrieves and jumping, while avoiding work which is more technical or slower paced overall, like scent articles or long stays. I’m not using the rewards to create the attitude but I will respond if the dog himself brings it to training.  If I select work which is too difficult when Brito is in a particularly excited frame of mind, then I’ll consider rewarding the occasional wrong attempt and then we’ll move away from that exercise and towards something more appropriate.  Since I’ll actively allow errors of enthusiasm within heeling, I may see some forging or bumping.  That’s ok; I can work with that as long as his attitude is strong.

I want Brito  to feel like a star and sure that he can do no wrong!  Even more excited about the next session!  As long as we’re still in the training phases and not proofing for competition, this is not going to hurt his career. Experience tells me that dogs who believe that they are superstars in training have much better resilience when I start demanding more precision. These dogs will stay in the game longer and will work harder, even when the going gets tough and the behavior chains get longer or more complex.

Dogs that are more passive in their outlook; accurate but uninspired, are much harder to keep in the game when the dog begins to wonder if the value of the motivators is sufficient for what you are asking in return. Eventually, I’ll pull attitude and accuracy back together again, but in the meantime it’s possible that I’ve created some problematic behavior chains. If you only reward a behavior chain if it is absolutely correct then you risk a slower and more careful worker, unless you have taken extreme care to train speed and energy into each behavior in the chain before you pull them together.

For some temperaments of dogs this careful training approach works quite well, because they are happy to “bounce” into perfect heel position or quickly touch a target with their nose by virtue of their temperament.  But other dogs simply cannot generate speed and enthusiasm until chains are formed and movement or games are involved.  In effect, they require “flow” to create their love of work and without movement, they’re dull workers.  Practicing uninspired work over a long period of time will give you more of the same.

The trick here is to only reward dogs that make technical errors if the dog is offering a SIGNIFICANTLY better attitude than what is normal for that dog.

This approach does cause some problems over time as I tighten up my criteria, because dogs often enjoy those “errors of enthusiasm.”.  But the alternative, always working the dog in a frame of mind where they can be very precise and correct, just doesn’t work for me,  so I sacrifice some behavior chains over the short term.  When my dog is stronger and more willing to stay in the game even under adversity is when I will work to bring more consistent criteria back into my chain.

Decisions to ignore behavior chains are made to prioritize classical over operant conditioning. The more fragile your dog by temperament, the more you need to consider your dog’s CER towards you and training.  The “stronger” your dog’s temperament the less important classical conditioning is, because your dog is already driven to really want to work with you!

If you’re a truly stellar trainer who is extremely good at creating and maintaining each behavior to absolute excellence before you create a behavior chain, and if you have a dog who is amenable, then you’ll rarely run into any of these issues.  But if you’re a more average trainer, or if you have a softer dog,  then it’s worth considering the ‘less than ideal’ situations where prioritizing your dog’s attitude over a specific behavior probably makes good training sense.

In the next blog, we’ll consider the related topic of the “challenging” dog; either by virtue of genetics or an unfortunate training history.

Behavior Chains, part 7

Posted on


The next few blogs in this series will consider the various occasions under which I’ll ignore excellence in behavior chains in order to support a dog’s Conditioned Emotional Response (CER) or self confidence within the training relationship.

Today we’ll consider managing the dog when the handler makes an error.

My general rule is to provide a reward, even if this choice will reward a fault within a behavior chain.   Keep in mind that when I say “reward” I mean either a classic one (toys and food), praise, or the chance to continue on in the chain.  Let’s look at this topic more closely:


I set up for a retrieve on flat.  I say “stay” and then I send my dog to fetch with the WRONG command.  In this example, I cheerfully ask her to “take it,” but her cue is “bring!”  As a result, she looks at me with a ready expression but she does not move.

At that point I would hand her a cookie and act like she was the most clever dog on the planet, and if she had chosen to fetch I would have treated that as super clever as well. In essence, whatever choice she makes is going to be right.

If you ask your dog to do something that she is not trained to do (in this case follow an unfamiliar cue),  then she cannot win because there is no “right” answer.  Fetching a dumbbell on the wrong cue is just as wrong as doing absolutely nothing.  As a handler, your options are to either start over (effectively punishing the dog), or offer a reward.  I choose to reward so that my dog will continue to feel absolutely confident.

Hopefully I’ll do better next time.  Fortunately, single incident learning within behaviors chains is relatively rare, and it is extremely unlikely that one cookie will cause massive confusion the next time you throw that dumbbell.

Scenario #2:

I am directing a handler through a heeling pattern.  As I call a right turn, the handler is unsure of what I asked for.  They “bobble” their handling which throws the dog off, so the right turn is wide and unsure.

Your dog has no idea what “bobble” handling means so there is no way for your dog to be correct.  If you simply start over then you risk a loss of confidence in a dog that has been working correctly.  For your dog to maintain faith in your leadership then you have to lead so that the dog can follow.  Keep the dog in the game!

Hand over a reward or simply continue the chain as if the error never happened.

Scenario #3:

You’re running an agility course and you pull your dog off of a correct jump with an inadvertent twist of your shoulders.  Your dog “guesses” about what you want.  It doesn’t matter if your dog guesses correctly or not; reward your dog with a cookie or by continuing the run.  In effect, the error never happened. 

In all of the above examples, your handling made it impossible for the dog to meet criteria.  Since re-starting a chain means “you’re wrong”, then any other choice effectively punishes your dog for a circumstance over which they have no control.  That choice may not matter for dogs with stronger temperaments, but it can be death on more fragile dogs.  If you’re not sure about your dog’s tolerance for errors, watch your dog’s behavior carefully after you make a mistake.  You might be surprised at how often your dog’s next attempt ends with classic avoidance behaviors like sniffing or zooming, or with more confident dogs just a slight decrease in speed and forward momentum.  Go ahead and reward your dog and try to prevent those sorts of errors from happening too often.

In the agility culture, this sort of thinking is better understood than in obedience, because dogs with less confidence in their handler’s abilities are often slow and methodical, and a slow dog is the death knell of a serious agility competitor.  In obedience, on the other hand, slow and careful can be rewarded by high scores, so there is less incentive to  consider WHY the dog is slow and careful.

Sadly, I screw up more than I’d like to admit, because screwing up covers a lot of territory in the world of dog training.  I may have asked for a behavior that my dog cannot manage in a challenging situation.  I may have miscued my dog.  I may have put her in a frame of mind that is conducive for routine work but not for learning new skills.  I may have gotten distracted and disconnected.  There are a lot of ways to screw up, and professional trainers are not immune.

Your dog should not pay the price for your learning curve; if you make a mistake or “bobble” in training then reward your dog.   If you follow this rule while you are learning to be a better trainer, then your dog’s attitude will remain intact, even if you’re making a bit of a mess of the process.  Teaching behaviors is relatively easy once you master the mechanical skills, but recovering a dog with a bad attitude is actually rather difficult, and since dogs often revert to early learning under stress it can rears its ugly head when we get to competition. 

Rewarding a dog each time we screw up helps us get around that.

My next blog in this series will consider behavior chains and dogs working in an unusually “high” frame of mind.  In other words, should you reward errors of enthusiasm?


Behavior Chains Part 6

Posted on

This blog will consider the issue of behavior chains, Classical and Operant Conditioning and Conditioned Emotional Responses (CER’s) with a dog who runs off in the middle of training for agility competition.

Here’s the scenario:

Your dog has been in agility classes for one year.  He started at 6 months and worked his way up through all of the classes.  You did a puppy class, a foundation class, a class on equipment and a class on basic sequences that pulls together handling skills and basic sequencing.  You’re now in a class that does longer and more complex sequences.  Many of your classes mates are beginning to compete.

Your dog, however, runs off.  A lot.  Maybe 25% of the time, your dog is heading for a tunnel and….keeps on going.  Past the tunnel.  Past the jumps and straight out to the dogs waiting outside the ring.  Fortunately your dog is friendly, so after wiggling around for awhile outside the ring, he is easily caught and returned to you.

Now what?  Let’s start with the issue of behavior chains, classical conditioning and CER’s.

For a change of scenery,  I want to talk about you for a bit.  Yes, you.  The handler.

Let’s talk about YOUR behavior chains, YOUR operant and classical learning, and YOUR Conditioned Emotional Response.

When you set up for your run, you are beginning a behavior chain; a way of handling and interacting with your dog.  What does it look like? Has your dog trained you to stay close?  To feed constantly every time he looks around or shifts?  To monotone “stay stay stay” with an outstretched hand on the start line?  And…if your dog breaks that start, has your dog trained you to simply turn and start running the course?  Does your behavior chain include feeding your dog every time you stop for a contact to prevent him from looking around?  Do you start running towards your dog, frantic when he looks away, even if the course goes the other direction?  And if your dog runs off, what do you do then?  Do you take your dog back to the course and continue on?  Do you crate your dog?  Take out another dog?  Go home?  Yell?  Throw up your hands in despair?

What has your dog trained you to do?  Have you been operantly conditioned to hover, feed excessively, stay in his line of vision and handle to maximize your presence rather than for the flow of the course?  If yes, then you’ve got a classic behavior chain; you’ve learned to handle to compensate for your dog’s behavior rather than adjusting your training to change your dog’s behavior.  And my guess is that it’s not pretty.

Now let’s consider your CER; your Conditioned Emotional Response.  How do you FEEL when it’s your turn to run?  Are you nervous about what your dog might do?  Are you concentrating on your dog or on evaluating the environment for the places you are most likely to lose your dog? Is your heart pounding?  Are you embarrassed about what your class mates are thinking about you?  Depressed  about your lack of progress? Irritated with your dog, who doesn’t seem to do what the other dogs do?

That’s classical conditioning at work, and you haven’t even left the start line.

How does that make you feel about your dog?  It’s likely that all of those negative feelings have also affected your relationship with your dog.  It’s easy to become angry and resentful, which of course will make your dog want to avoid you.  Uh oh.

Do you see how your behavior chain begins to intersect with your dog’s behavior chain?  You’ve trained your dog and your dog has trained you.  You have feelings about your dog and your dog has feelings about you, and those feelings impact your training. Both your conscious choices (operant) and your unconscious (classical) ones are involved, because nervous handlers are in no better a position to train than nervous dogs are in a position to learn.  Of course, learning is taking place, but it’s likely the exact learning that got you into this mess in the first place.

If this is you, then there is a problem and the problem may, or may not, involve the dog.

What has happened is that you and your dog are over faced.  You are in a situation that neither of  you can handle, and the longer it’s been going on without intervention, the harder it will be to fix it.  All teams make errors, but when you no longer make training decisions to further your sport interests but instead to prevent disaster, then you’ve crossed the line from an error to a problem.

Well trained and prepared dogs will mess up and may run off on occasion.  That’s ok because it’s part of the learning process, but it should be relatively unusual, and it should not impact your basic handling decisions.  If it does happen then you’ll want to take a moment to figure out what triggers may have occurred in the environment to cause your dog’s behavior.  Was there a dog unusually close to the ring?  Did you ask for work that was harder than usual?  Did you change your reinforcer?  What factors may have existed that caused this result?  Once you know that, you can either use it as an excuse (and change nothing), or you can set up scenarios that allow for a “miniature” version of the same situation within a more controlled environment, so that over time your dog can learn to work in spite of his triggers.

But once you’re at the point where it’s a routine happening, then the dog is probably training you rather than you training your dog.  The dog has set your handling and your reinforcement schedule.  And it’s likely that your CER towards training in general and your dog in particular is miserable.

You have to regain control of your training.  That means you need to forget about what the rest of class is doing and create a plan for how you are going to get your training back on track.

Where did you go wrong?  Is it a foundation issue?  Is your dog clear on each single behavior, and properly proofed for various distractions when only ONE obstacle is being considered?  How about your start line?  If  you cannot leave your dog, lead out and know that your dog will still be there, then you should NOT be running the course unless you had already planned on a running start. Get that foundation first.

Or maybe it’s a problem with your dog’s emotional state.  Is your dog nervous?  Scared of the other dogs?  Unsure of your expectations?  Avoiding you or the pressure that you represent?  These are not “agility” issues but they will sure cause you grief in agility.  Study dog behavior and learn how to make your dog emotionally secure.

Is your dog more excited about the world than whatever you might have to offer for motivation?  If your dog isn’t able to focus and stay engaged before you take the leash off, it’s over before you’ve even started. Focus and impulse control are trained behaviors!  Work hard on impulse control exercises in well managed environments until both you and your dog can handle just about any temptation thrown in your direction.

Or maybe your dog’s foundation is excellent, impulse control is perfect and your dog’s emotional state is also fine.  Maybe your dog is simply multitasking.  He likes to run agility and he also likes to visit other dogs, people, etc, so he leaves and then comes back to continue with work.  If that is the case then you have a classic behavior chain issue; your dog has learned that there is no real consequence to adding his own personal twist on the courses.

To be honest, I find this to be the least common scenario, but if that is what is happening then your run must end.  Every time.  Have the person who catches your dog put him in the crate while you continue the run without him.  Yes, you heard me right.  You will continue the run – having as much fun as you and your imaginary dog can muster.  While another person crates your dog.  No reason for you to be the bad guy.  And anyway, you’re busy on the course having fun with your imaginary dog.

With my young dog Brito, I have a pretty good chance of keeping him engaged under some, but not all, circumstances.  I know what motivators work best (high value food) and what surfaces he can function best on (concrete or indoors).  I know his best training times of day (morning or evening) and I know when he’d prefer to sniff and wander (afternoon).  I know which environmental triggers are deadly to our engagement (squirrels).  I know which exercises we can practice in public (easy ones where he is mostly facing me) and which ones need to be worked at home (working at a distance or independently.)   I know how long he can work before he gets tired (about fifteen minutes split into two sessions – shorter outdoors) and I know how being tired affects his behavior (sniffing and wandering)

I can manipulate these factors to allow him to succeed and if that is not possible then I can choose not to work him at all.  Every management decision I make is designed to further our long term goals; none are designed to “keep up” with a class.  I don’t care if the rest of the class can do scent articles facing a field of cows.  We’ll practice our articles in the bathroom so we can have lots of success.  We’ll join the class when we’re ready.

If you recognize yourself in this blog, then here’s what I think you should do at this point.

Go talk to your instructor and start an honest conversation.  You need to know exactly what has to happen at this point to get you back on track.  You need to know if your dog’s behavior is disruptive to others in class or is being perceived as dangerous in any way.  You need a plan.  And then you need to listen to what your instructor has to say. If she says that you’ve got a great big mess, then hear that.  If she says that you need to take your 18 month old dog back to a puppy class, then do it.  If she says that the other students in class do not feel comfortable with you and your dog, then hear that too, even if it makes you mad.  If your instructor says that she can create a plan for you, but it will require private lessons, remedial training or working on foundation skills on the sidelines, then give it some thought.  Most of us professional trainers are loathe to bring up these hard conversations with our students, so if you don’t ask outright then you’ll find that the conversation is not likely to happen.  What you do with that information is up to you.

If you are the instructor, then this is the time to be both honest and kind.  Create a plan to address their issues.  If you cannot or will not do that, then refer them to someone else.  And if you don’t think the dog’s circumstances can be helped, either due to the seriousness or complexity of the problem, then now is the time to put it out there.  There is no shame in admitting that you’re in over your head too.

If I could, I would lay out a plan right here, but there are thousands of reasons why dogs and handlers find themselves in this situation.  You need to figure out the root issue and then work to address that.

In the next blog in this series, we’ll look the main reasons why I’ll risk an undesirable behavior chain in order to further another interest.

Behavior Chains, Part 5

Posted on

In this series on behavior chains, we’ve considered what they are, how to use them to your advantage, how to adjust your expectations in new environments, and in the last blog, we’ve looked at classical and operant conditioning and how a Conditioned Emotional Response (CER) forms during both operant and classical conditioning towards both the trainer and towards training in general.

This part will consider the relationship between operant conditioning, classical conditioning, CER’s, and how each might be considered in training decisions that potentially lead to the formation of  behavior chains.

Let’s consider a scenario:

You take your four month old puppy to a new park for a little training.  You start a session and get a few seconds of work, which you reward generously. Your puppy wanted that cookie and earned it by offering behaviors that you wanted to see. That is operant conditioning.  You even managed to string together three perfect steps of heeling before reinforcing – a mini behavior chain!  All is well!

And then your puppy sees a butterfly and runs off to chase it before you have a chance to intervene.

One minute later, your puppy remembers that you exist and looks over at you.   You believe that if you call with enough enthusiasm, your (now tired) puppy will probably come back. What should you do?  Should you call?  And if the puppy comes, should you hand over a cookie? If you hand over a cookie, are you rewarding the recall, or the running off followed by a recall? Or should you go back to work and forget the recall cookie?  Or just pack it up and go home?

Remember that in the earlier blogs on this topic of behavior chains, we talked about the fact that whatever happens within work becomes part of the chain.  So if you call your puppy and reward, are you creating a behavior chain called, “Work, run off and chase butterfly, come back, and get a cookie?”

Maybe.  But in this scenario, the possibility of creating a behavior chain is probably not the most relevant factor under the circumstances.

In this example, did the puppy consciously choose to run off after the butterfly, or did he just go?  If the puppy made a choice (looked at the butterfly, looked at your cookie, and then took off), then I’d be a little worried about the creation of a very undesirable behavior chain since the dog is potentially learning to run away and then run back; talk about having your cake and eating it too!

But with a young puppy, it’s really quite likely that the puppy did not CHOOSE to go after the butterfly at all.  Its likely that the puppy saw the butterfly and the feet followed – prey drive kicked in and off he went!   And if the puppy was not conscious of making a choice, then there’s not much concern about creating an operant behavior chain.  As far as classical behavior chains, you dog does not CHOOSE to make those chains; they simply happen, and they become stronger with practice.  That’s why it’s up to the handler to manage training environments so that running off is fairly rare.  Over time, dogs learn focus, impulse control and a love of work.  The more these are developed, the less likely it is that a dog will run off in the middle of work.  In the meantime try not to let it happen (leashes, controlled training environments, constant contact, appropriate work for dog’s stage of training and emotional readiness etc. are all strategies to consider)

So back to the above scenario.  What would I do?

I’d give the puppy a cookie for coming back.  And I’d work hard not to let it happen again.

If you’re staring at your computer screen in utter disbelief then consider this:  There are tradeoffs with every choice you make.  It’s not all about behavior chains; don’t forget those CER’s that we talked about in the last blog!

What CER do we want your puppy to feel when they are with you in public?  Safe!  Happy!  Eating!  Approval!  Remember, the running off happened – it’s in the past.  The puppy is no longer thinking about what happened a minute ago, he’s thinking about how he feels right now as he’s interacting with you.  Make that CER towards you positive; feed and play, even if you’re less than thrilled with what he did a minute ago.

If you don’t give your puppy the cookie, is he more or less likely to come back next time?  Obviously, he will be less likely to come back.  And next time, as your puppy stands looking at you and thinking about whether to return, he is definitely in operant mode because he is making a decision based on prior experience.  Hand over a cookie this time and the next time he gets away your odds of getting him back go up quite a lot.

Indeed, the ideal situation in public with a very young puppy is a whole lot of cookies, toys and and play in a short period of time, for pretty much no work at all.  Just look in my general direction and I’ll do the rest.  The only thing the puppy needs to associate with you and being in public is how great it is to be there, eating cookies, basking in your approval, and developing a classically conditioned response to public places – keep an eye on you because food keeps coming.  Once the puppy has made that association and has developed a tendency to look towards you to get cookies, for no reason at all, then it’s a very small step to switch over to operant conditioning.

You just add some criteria – now you have an if/then statement.  “If you follow my cue then I’ll give you a cookie.”  At that point there’s no need to continue with free cookies in that environment.   The new question is, “what can you do to earn that nice string of cookies?”  At that point, if your dog decides they cannot or will not work that hard for the available reinforcement, then you can stop.  You don’t need to feed any more free cookies.  Just make sure that whatever criteria you are asking for is reasonable, so your dog is winning most of the time.  And reasonable, for a typical young dog in public, is probably one or two seconds of work.

You’ll make lots of mistakes in training, and that’s fine; even the most experienced trainers make less than ideal decisions.  Just make a point of evaluating your session when you’re back home.  What did you do very well?  What can you do better next time?  On balance, are you heading in the right direction?

If yes then all is well.  If, on the other hand, you find that you’re dealing with the same issues month after month, and long after the puppy phase, then it’s time to consider what you have.  What is your dog’s CER towards you and training? Is your dog actively avoiding  you?  Is your dog making conscious choices or simply responding to the environment?  If your dog is continuing to show you behaviors that you don’t like (running off, for example), how have you altered your training choices so that it is either impossible (long line or small area) or less likely (less interesting environment or better foundation skills)?

In the next blog in this series, we’ll consider an adult dog who runs off in the middle of agility training.  What are the relevant behavior chains, training considerations and CER’s in this scenario?




Behavior chains Part 4

Posted on

Now let’s veer away from Behavior Chains for a moment to introduce Operant and Classical Conditioning; we’ll combine these issues soon when we consider even more complicated scenarios.

Operant Conditioning is one way animals learn; it’s a simple “If/then” statement.  The animal consciously realizes that “if I do X then you do Y”.  So, if the dog sits, then the dog gets a cookie.  Consciousness is very important here; the animal must recognize how his behavior does (or does not) affect the outcome to be operant conditioning. Until now, this series has assumed operant conditioning when discussing behavior chains.

Classical conditioning is another way animals learn.  It is a simple association between two things.  It can involve emotions (happy, sad, nervous), physical reactions (increased or decreased movement) or involuntary reactions (drooling, hormone release, etc.).  From the dog’s point of view, a classically conditioned behavior chain might go something like this; “Mom is preparing my dog food.  When food is being prepared I get to eat.  I love to eat!  Drool is forming in my mouth because food preparation happens before I eat. Eating makes me happy.  Feeling happy makes me excited and that makes my body want to move more.  When I see food I feel so much excitement that I leap around!”

Now you have a classically conditioned behavior chain – the dog never had to “think” anything at all to cause this chain to happen; simply experiencing food preparation over time caused a chain to form.  In both operant and classical conditioning, learned has occurred but one was a conscious event and the other was not.

Did learning occur?  Yes.  Did that learning cause predictable changes in the dog’s behavior?  Yes.  Was it a conscious choice to drool, feel excited, and leap around?  No.  It’s a conditioned response to the process of eating and food preparation as opposed to a conscious choice, but the end result is the same – a behavior chain has formed.  In fairness, many examples of learning have both classical and operant components, but for the sake of simplicity…work with me here.

People also experience operant and classical conditioning.

I happened to experience a classically conditioned response recently.  A few friends joined me and we set up a practice ring for our dogs.  When one of my friends took on the role of the judge and called me into the ring, I found myself feeling nervous!  Even though I was well aware that this wasn’t a dog show, I still had the emotions that I experience at a real dog show when a judge calls me into the ring.  I became hyper aware of my dog’s behavior  and my heart rate increased!  Logical?  No; I knew that this was not a dog show.  Classical conditioning doesn’t have to be logical; it just happens.

So now that we understand that classical conditioning “just happens”, what is the relationship between the trainer and classical conditioning?  When you prepare your dog’s food and they become excited, or when  you train your dog to heel, are you relevant to the process of classical conditioning?

Yes, because the dog begins to associate you with the entire process of learning.  Whatever emotions your dog experiences in training will now also become attached to your presence as the trainer.  This is called a Conditioned Emotional Response (CER) and it takes place through classical conditioning.  Your dog does not choose to be happy when the food preparer shows up; it just happens.  The dog does not choose to be happy at the start of a training session with a trainer who uses lots of desirable reinforcers – it just happens.  That’s very good news for those of us who compete in events where our reinforcers are severely limited.  We may not be able to bring our food and toys into the ring with us, but we always bring our dogs dog’s CER into the ring.  If that CER is positive, then our presence provides a good deal of emotional support to the dog because they willl find themselves feeling “happy” in our presence in a training context.  Much the way I found myself feeling “nervous” in the presence of a judge, except hopefully….happy!

Both operant conditioning and classical conditioning cause learning, and both need to be understood when you make a choice about how to handle your dog’s behavior at any given time.

In order for an OPERANT behavior chain to form the dog must be aware that his behavior creates consequences.  “Every time I look up at mom in heel position I get a cookie”.  The dog tries out variations (sniffing, looking away,etc.) and finds that they don’t work and soon, the dog understands that the way to the cookie is to look up at mom continuously.

That is very different than the dog that leaps and drools when food is being prepared.

But…what if the owner does not appreciate the dog’s leaping around at food preparation time?  Then the trainer can make a point of causing the food to disappear every time the dog starts the leaping.  This will cause your dog to transition into an operant mode; they will work to understand why their food has disappeared.   They will begin to connect their behavior (which until this time was probably unconscious) of leaping with the food disappearing.  Then they can try out alternatives to see if they work better.  For example, “when I stay calm the food continues to approach me and when I leap it disappears.”

In the above example, you’ve transitioned from classical to operant conditioning in order to affect a behavior chain.  Timing is critical here. If the very first second the dog does a big leap then the food disappears,  the dog is likely to make the connection quickly.  But if the owner is either inconsistent  (sometimes the leaping causes the food to disappear, other times it is ignored), or if the owner is simply slow (the leaping is ignored for five seconds before the trainer responds by removing the food) then the dog will struggle to figure out what causes food to come or go.  Further, the longer that the dog’s classically trained response is ignored, the harder it will be for the dog to become conscious of the change desired by the trainer.  When the rules of the game are changed, especially after a behavior chain has a strong history of continuous or inconsistent (partial) reinforcement, then the dog is likely to be very frustrated because he won’t know how to “win”.  Frustration leads to all kinds of bad things, including, unfortunately, an association between the owner and negative emotions.  The owner becomes associated with negative emotions such as frustration rather than positive emotions built on mutual enjoyment.  To understand this, re-read the paragraph above on CER’s.  If the dog associates you with the removal of what they want (food in this example), then you’ve got a problem if you cannot also find a way to either communicate to your dog the desired behavior to bring the food back, or find another way to reduce your dog’s frustration.

To summarize.  Classical and Operant conditioning are different, yet connected.  ALL operant conditioning also involves classical conditioning, because the dog is unconsciously developing a CER when being trained – let’s hope it’s a happy, positive, and enthusiastic state, because that classical response to training is also being connected to how the dog feels about his trainer.  While all operant conditioning involves classical conditioning, the reverse isn not true; not all classical conditioning is operant, because operant conditioning is only in play when the dog is conscious of how their choices affect the outcome.

So now that I put you through reading all of that…why do you care?  Because understanding classical conditioning is the only way to understand why there are many times that I will give a dog a cookie or a toy even though I do not like the behavior the dog is showing me.  In those instances, I am choosing to prioritize the dog’s emotional state over conscious learning.  That is the subject of the next blog in this series.

Behavior Chains, Part 3

Posted on

In Part One of this series, we looked at Behavior Chains.  What they are, how they are formed, and why you care if you are training for competition.

In Part Two, we looked at Behavior Chains in heeling. A video example of Lyra showed both success and failure within a chain, and how to pull out a weak piece of the chain and work it separately before returning it to the chain.

Today we’ll look at behavior chains and how they are integrated into the process of generalization.  Specifically, how do we handle reinforcement when a chain is excellent in one environment but falls apart in another?

Let’s use my nine month old puppy as an example.

At home, Brito can be brilliant for 15 or 20 steps of heeling with all sorts of turns and challenges thrown in for good measure.  That chain earns a cookie.  Great!  That cookie reinforced the entire chain of excellence.

But in the local park?  That same 25 steps would surely include sniffing, sightseeing, and in Brito’s case, almost certainly running off to find other interesting things to do.  Considering he’s a curious puppy with many interests, his behavior is entirely normal, and there is no reason to “squash” his interests or try to force him to work before he is ready.  What I need to do is allow him the freedom to choose between his options; specifically,  limited exploration on a short leash or my fun games.  My job is to make sure that I ask for such a short chain that it becomes a no-brainer for him.  It has to be a much better deal to work with mom for one second, followed by a fabulous party, than to self entertain in a limited space.

That means that I have to rethink what is an appropriate behavior chain under these more challenging circumstances.

In the local park, I have to simplify my requests so that success is very likely.  Instead of 25 steps of heeling, I’m thrilled with just looking at me for something to do.  He can look at me from front position or side position. And if he appears to be inspired or unusually attentive then I can even see about one or two or three steps of heeling.  My job is to reward before the chain fails!  If he makes a mistake once a chain has started, then I simply step out of position and when he re-engages, we try again.  Give me five steps of excellence over 50 steps of mediocre any day.

Here’s a video of Brito training at the local park.  Note that all he really has to do is look at me with a bright attitude (the very first step in the chain).  If he looks at me AND gets into position I will give him cookies almost continuously as long as he stays there.  You’ll see that he tends to look away after each click/cookie.  That doesn’t bother me because the treat ends the behavior chain.  If after the cookie he does not look back quickly, then I step out of position, effectively ending the possible start of a chain called “ignore mom in heel position”. When I feel like Brito is settling in then I ask for a little more.  Brito isn’t really learning a heeling chain as much as a focus chain; he’s learning how to focus and enjoy his work in the face of environmental distractions.  To be honest, focus is much harder to get (and keep) than heeling, so it’s my priority.

Progress is slow and steady here but over the time, he becomes faster and faster to ask to work.  Faster and faster to maintain attention.  More determined when he gets to heel position.  And that’s all good!  By the five minute mark, I really have a dog.

If he is over faced (heels for a few steps and then leaves), then that’s fine.  He can go.  Without a cookie.  When he comes back, we’ll start that focus/heeling chain over.

If after a couple of failures I realize that I’ve asked too much, then I’m happy to scale back on my expectations, but if failure happens too often, you need to re-evaluate your training plan. Frequent “scaling back” suggests that you are consistently asking too much of your dog.

If your dog’s inability to perform is a result of emotional distress (anxiety, worry, hectic behavior), then that’s different from curiosity or attraction to the environment.  You need to get your dog’s head in the game before you even think about training.  The fastest route I know to creating a dog with a lifetime habit of shutting down or becoming frantic in new places is asking for more work than they can give whenever you enter a new environment.  Don’t go there.


Behavior Chains, Part 2

Posted on


In my last blog, I explained a bit about behavior chains.  I talked about what they are and how understanding them can help trainers make better training decisions.

Now I want to talk about Heeling, since it gives people more grief than any other exercise.  Why is something so simple in concept (walk next to my side looking up) so difficult in practice?

Heeling is a behavior chain made up of the same behavior repeated over and over and over…each step of heeling is either performed correctly (according to the criteria set up by the handler), or it is not.  The basic criteria never changes – hold your body next to me in exactly the same manner no matter I do, and look at a focal point.

If we look back at the last blog on behavior chains, then we understand that if a dog is allowed to move to the next step in the chain (the next step of heeling) then the last step must have been performed in a manner that was acceptable to the handler.

For example, the first three steps are perfect.  On the fourth step the dog looked around briefly but was back in good position at the fifth step.  By the sixth step the dog was lagging slightly until the ninth step and at the tenth through thirteenth steps the dog was brilliant. The handler was pleased with the end of the chain, so the dog was given a reward.

Remember that if the dog is allowed to move to the next step, then the previous step is reinforced by the next step, because any step can lead to the final reward.  The chain includes ALL OF THE BEHAVIORS seen from the start until the reward.  Therefore, allowing a dog to continue an exercise communicates that the behavior is being performed correctly and that the reward will come.  And it does!

And if it is not correct?  What if the dog takes a quick look around or lags on a few steps of heeling, but continues “well enough,” to eventually earn a reward?  What has happened then?

If that deviation was positive for the dog, then you have allowed it to become a possibility within the chain called heeling.  Next time you ask for heeling, you may well find that you’ll see those deviations again.

So what SHOULD you do?

End the behavior chain and try again.  We discussed this in the last blog.

If you aren’t pleased with your dog’s behavior and if those deviations could be self reinforcing for the dog, then you must eliminate them from the chain.

Of course, ending the chain and restarting only works if you dog knows how to win.  For example, if you dog sits crooked within the chain and you start over, that is only a sensible decision if dog knows how to nail the straight sit at heel when it is cued outside of the chain.  Adding expectations that are not trained to fluency outside of the chain won’t have any positive effect at all. Indeed, it is quite harmful and will probably cause your dog to shut down.

Here is an unedited video of Lyra working on heeling in a new location.  This is hard for her.

The chain fro 6 sec to 21 sec is as good as I’ve trained for – she is rewarded.

The chain from 31 to 46 s also as good as I’ve trained for, so even though she fails to perform the”spin”cue,  she is rewarded.  Logically I should either pull the spin out or give her more support when asking for it, since she’s not strong enough to do it independently within a chain at this time.

1:05 – she looks away and I correctly end the chain.

1:11 – she looks away on the next attempt and I correctly end the chain.

1:17 – again she fails (by now I should be worried that she is not capable of doing better).

1:24 Here she succeeds.  This reward covers the period from 1:21 to 1:24.

Now I attempt to try the chain from the beginning

1:49 – 1:54 – she fails at the end – the same point she has failed several times before.

It would make no sense to continue on this path of failure.  If a dog repeatedly struggles at the same point in a chain, then you need to stop what you are doing and pull it out to work on it separately.  How you do that depends on the dog and why you think they are having trouble.

In this case, what I do is simply go right up to gate where she is failing and only ask for that very last step.  I walk slower than normal to give her more support.  We do this for the next minute – with this level of support she is able to succeed several times in a row.

After a minute of isolating this variable, I add it back into the chain.

From 2:57 to the end we have our final chain.  She is successful and I end that session.

In the next blog (Behavior Chains, Part 3) we’ll continue to consider the issue of behavior chains and how they relate to competition training.


Get every new post delivered to your Inbox.

Join 982 other followers