How You Get Behavior Really Does Matter
The following was originally printed as a three-part article in Teaching Dogs magazine.
A sit is a sit is a sit. When you get to a final, finished behavior, you won’t be able to tell whether the initial behavior was molded, lured, or captured, so as long as it doesn’t cause fear, pain, or distress to your dog, it really doesn’t matter what method you use to get behavior. Or does it?
All methods of getting behavior are not created equal, and none is right for every trainer, every dog, and every situation. In this three part series, we’re going to examine the most common dog-friendly methods of getting behavior: social facilitation, molding, luring, targeting, capturing, and shaping. In part one, we’ll discuss the consequences of choosing a particular method and overview the pros and cons of each. In parts two and three, we’ll explore each method in detail and look at some practical applications.
Solving the jigsaw puzzle
If you were going to teach a small child to solve a jigsaw puzzle, how would you do it? Would you put the child’s hands on each piece, guiding it into position? This would get this puzzle worked very quickly, but it wouldn’t help her learn how to solve the next one. Would you talk her through it, verbally helping her through the process? Again, that would get this puzzle worked fairly quickly. In fact, she would probably learn to repeat those behaviors and soon be able to work this puzzle herself. She might even learn a tip or two to help her with the next one.
But what if your goal weren’t just this puzzle, but all jigsaw puzzles? Instead of helping her, physically or verbally, you could let her work it out, acknowledging correct choices and helping her work out strategies like “Straight edges go on the outside.” It would probably take her longer to work this first puzzle, but as she learned the concepts she could apply to other jigsaw puzzles, her speed with subsequent puzzles would increase.
Examine your goals
Training your dog is similar to teaching a child to work a jigsaw puzzle. The method you choose affects more than the behavior at hand. It affects your dog’s mind, teaching him how to learn, how to approach problems in the future.
When choosing your method, consider not only the immediate behavior but your overall training goals. Is this a pet animal, who will learn only a small handful of behaviors, owned by someone who desires a steady, predictable companion? Or is this a performance dog, who will ultimately have a large repertoire of behaviors, who may need to think on his feet in a working situation, or who might benefit from problem-solving skills?
Some people want a thinking dog. Others very definitely don’t. Neither is right or wrong, just desirable or undesirable for the tasks the dog will be asked to do. Before you choose a training method, ask yourself the following questions:
- How many behaviors will this dog need to learn? The more behaviors you need to teach, the more important it is to teach concepts that can be generally applied to different behaviors.
- How quickly do I need to train the behaviors? Some methods are faster for beginners than others, but those methods tend to get the same results forever. Other methods start slow but increase in efficiency—bypassing the initially faster methods—as the trainer and dog get more experienced.
- How precise are the behaviors? Some methods are better-suited to larger, less precise behaviors, and others excel with precise behavior.
- How skilled is the primary trainer? Some methods are easier for beginners. Some require both physical and mental skill.
- How interested in training is the trainer? Not everyone who trains a dog aspires to be a “dog trainer.” Some methods require more knowledge and skill than others.
- How experienced is the dog in each method? The more familiar a dog is with a particular method, the more rapidly he will learn new behaviors taught with that method.
- Will the dog need to “think on his feet,” or will he primarily respond to well-known, well-rehearsed cues? Some methods teach the dog to wait to be shown what to do. Others require the dog to figure out what’s needed. These habits carry over into their working lives.
Weigh your options
Once you have your answers, compare them against the training methods. All have pros and cons—none is a magic pill perfect for every dog and every trainer. The six most common techniques for getting behavior are described below.
Instinct, allelomimetic behavior, mimicry, social facilitation. While these are each subtly different, they all fall into the category of naturally occurring, “non-training” methods of learning. With this method you’re taking advantage of the dog’s instincts to get the behaviors you want. For example, many dogs will naturally follow another dog. You could, then, pair that dog with a dog who has a terrific off-leash recall to help build a habit of returning when called.
- Pros: These natural methods of learning are ideal for the canine brain. They are an excellent choice when working on simple behaviors with a group of dogs.
- Cons: Although there are numerous anecdotes of dogs learning trainer-defined skills using these methods, depending on them to teach behaviors that aren’t natural canine behavior is an iffy proposition. Bottom line: if it works, terrific! But don’t count it as a sure thing.
Molding. Molding is physically guiding or otherwise compelling a dog to do a behavior. Pulling up on the dog’s collar while pushing down on his rear is a method of molding a sit. Molding also includes the use of physical props, such as working against a wall to force a straight heel or putting tape on the dog’s face to elicit a paw over the nose.
- Pros: Molding is easily understood by humans, and thus it’s very easy for beginners. It’s a quick, easy way to teach large behaviors.
- Cons: Though good for large behaviors, molding is limiting for trainers who want more precise or advanced behaviors, and it requires a great deal of trainer participation, which then has to be faded from the picture. The dog has to do very, very little thinking—his body is set up to perform the desired behavior.
Luring. Luring is a hands-off method of guiding the dog through a behavior. Lures are usually food but may be target sticks or anything else the dog will follow. A common method of luring the sit is to hold food in front of the dog’s nose, and then move the food up and back. As the dog’s head follows the food, generally the back end will drop to the floor.
- Pros: Luring is fast and flexible, and it’s easy for beginners.
- Cons: Lures must be faded early or they become part of the behavior, and properly fading a lure is not easy for beginners. Luring, like molding, requires little mental effort by the dog. You’re telling him everything he needs to know, and helping the dog becomes habitual—for both of you.
Targeting. Targeting, at its most basic, is the behavior of touching a specified surface with a particular body part. In practice, targeting is much more flexible. Targets can be used to position an animal, to manipulate its body position, or transferred to a different surface—or used in combination to get incredibly complex behaviors.
- Pros: Basic targeting is a simple, easy to teach behavior that can be generalized to different body parts fairly easily. Targeting is fast.
- Cons: After the initial behavior is taught, this method requires little mental effort by the dog—the trainer gives the dog all the information he needs. Trainer participation is heavy and must be faded.
Capturing. In capturing, the trainer waits for the dog to offer the behavior, then marks and rewards it. Simple!
- Pros: Capturing is easy for beginners if the desired behavior occurs frequently. Even better, it requires mental effort from the dog to figure out why it’s being rewarded.
- Cons: Unfortunately, capturing is limited to naturally-occurring behaviors—it’s not likely you can capture a competition-perfect drop on recall. The trainer has to be ready to capture the behavior when it’s offered.
Shaping. Shaping is a technique of training a complex behavior by teaching, and gradually building upon, the behavior’s individual responses. To shape a spin, a trainer might start with just a glance to the left. Then a glance and a weight shift. Then a glance, a weight shift, and movement of a front paw, continuing until the dog is performing a complete spin.
- Pros: The clicker makes shaping a powerful technique, enabling incredibly precise behaviors. Its flexibility is unmatched. Once the trainer and dog are skilled with the method, shaping is extremely fast. Best of all, shaping requires significant mental effort, creativity, and problem-solving ability by the dog.
- Cons: Shaping requires good observational skills, and until those are developed, shaping can be frustrating to the trainer. It also requires the ability to break behavior into small enough increments that your dog remains consistently successful. If the trainer can’t do this, the dog can get frustrated. Shaping can be frustrating to dogs and trainers who aren’t method-savvy, aren’t comfortable experimenting, or aren’t good problem-solvers. Until the dog and trainer are experienced, especially if they lack a mentor to help them learn the method, progress can be slow.
No perfect solutions
In parts two and three of this series, we’ll look at each of these methods in detail. As you compare your goals with the pros and cons of the different methods, you may find that there are conflicts. Perhaps you have a working dog who needs to problem-solve, but you’re under a strict, tight time-table. Maybe you aspire to a sport like canine freestyle which emphasizes both precision and creativity, but both you and your dog have a learned reliance on lures. Don’t lose faith. As we delve into these techniques, we’ll give you specific tips and tricks for solving those problems.
In the last issue, we discussed how the method used to get behavior influences how dogs approach that and other problem-solving situations. In short, how you get behavior affects how dogs learn and think. No one method for getting behavior is perfect for every dog, every behavior, or every trainer, however. Each has pros and cons. In this issue we’re going to look in depth at three dog-friendly methods of getting behavior: social facilitation, molding, and luring.
Instinct, allelomimetic behavior, mimicry, social facilitation
In these closely-related methods of getting behavior, you’re taking advantage of an instinctive desire to do what others are doing to get what you want. Dogs are not known as mimickers. You cannot, for example, have one dog demonstrate a perfect finish, and then expect the other dogs in the class to copy it. Yet, they still have a tendency to repeat some natural—“untrained”—behaviors that other dogs do.
For example, if one dog in a group begins to bark, the others frequently bark. If one gets up and goes outside, the others often follow. If you signal a recall for one, the others may come in as well. Observations at a doggy daycare show this tendency quite clearly… groups of dogs tend to run, bark, and even lie down to nap at the same time.
Trainers of working dogs have taken advantage of this tendency for generations. It’s common to pair new dogs with older, experienced dogs. Sometimes this pairing is quite literal—young fox hounds are braced with older hounds so they have no choice but to follow the older dog’s lead. In some kennels, the training of the young hounds is done exclusively by the older dogs.
Trainers with multiple dogs can take advantage of this method for simple, natural behaviors if they have one dog who is already fluent. Cue a recall—or sit or down or other behavior—and reward the dogs who respond. Many dogs will respond just because the other dog did it. And those who don’t may be motivated to figure out why the others are getting reinforced.
Less obviously, allelomimetic behavior may be used to modify the emotional state of the dog. Allelomimetic behavior is “mutual mimicry.” It includes not only “acting like,” but also “feeling like.” And it isn’t limited to dog-dog interactions. Tense handlers have tense dogs. Calm handlers have calm dogs. Households experiencing unexpected acute stress may have dogs exhibiting unprovoked aggression or other stress displacement behavior. Trainers can use this tendency by modeling the calm, focused emotional state they want their dogs to exhibit.
Some real-life examples of social facilitation and other instinctive methods of learning include…
- A group of daycare dogs sitting quietly when the gate is opened… and new dogs following suit the very first day!
- Young puppies and adolescent dogs performing—and later repeating—recalls at the dog park because their older playmates respond to the cue.
- A dog learning behaviors as varied as pushing a box along the floor and walking backwards by copying the behavior of another dog being rewarded for those behaviors.
That final example is the most controversial. Repeated laboratory experiments have concluded that dogs are not capable of learning by observation. However the training world is filled with anecdotal evidence disputing that. The bottom line is this: Try it. If it works, you’ve found an easy solution ideal for the canine brain. If it doesn’t, simply try something else.
Molding is a method of setting the dog up so that he must perform the behavior. It includes physically compelling behavior or using props. Trainers sometimes use the term “modeling” interchangeably with molding. Technically this is incorrect. Modeling more accurately refers to demonstrating, rather than compelling, a behavior. Examples of molding include standing on a dog’s leash to force him into a down or putting a scrunchie on a dog’s paw to elicit a wave or a “limp.”
Molding has a bad reputation among clicker trainers. Although it includes techniques as innocent as heeling next to a wall to encourage the dog to move in a straight line, trainers, associating the method with force, often dismiss molding entirely. Unfortunately, by doing so they risk ignoring a method that could potentially make learning easier for their pet-owner students.
Some pet owners are not interested in dog training. They don’t want to become good trainers. They aren’t interested in theory. They simply want their dog to obey simple cues and live harmoniously in their homes. For some of these people, molding is both clearer and easier than other methods. If they are successful, and the specific technique they use doesn’t hurt or frighten the dog, why not take advantage of the method?
Of course, the method does have drawbacks. To be used successfully, the compulsion shouldn’t trigger the dog’s opposition reflex. The common techniques of pushing on a dog’s hips to get a sit or stepping on a leash to get a down are poor training choices because the dog’s natural response is to resist. Instead, use the opposition reflex to your advantage. Got a dog who shifts his weight backwards in a show stance? Tug lightly on his tail—and watch him resist the backwards pull by shifting his weight to the fore.
Another common and more accepted use of molding is the use of props—wire guides through the weave poles, a piece of tape on the nose to encourage “hide your eyes,” a channel to force a straight approach on a recall. The problem with these and most other examples of molding is that the props and trainer participation has to be faded from the picture, and that’s not always easy, especially for beginners.
Molding is similar to allelomimetic behavior in that both can be used to effect a change of emotional state. Emotions and physical position—body language—are closely linked. So closely linked, in fact, that scientists have an ongoing argument about which comes first! Regardless, it is possible to change a dog’s emotional state by having him perform behaviors consistent with a different emotion. For example, a dog has trouble maintaining aggressive behavior when he’s wagging his tail.
Amazingly, the change in emotional state occurs even if the dog isn’t volunteering the behavior. In other words, molding the mannerisms associated with a particular emotional state can create that state. If your dog is growling at the end of his leash, try smoothing his hackles and lowering his tail to calm him.
Luring is a hands-off method of guiding the dog through a behavior. Lures are usually food but may be target sticks or anything else the dog will follow. The method is incredibly simple. Using the food or other lure to control the movement of the dog’s head, the dog is maneuvered so that his body performs the desired behavior.
Luring is a very, very popular method of positive training. It’s fast, it’s flexible, and it’s generally easy for beginners. Because it’s hands-off, it’s considered superior to molding. After all, the dog performs the behavior of his own volition!
However, if the dog is focusing on the food, he’s not focusing on what his body is doing. Fading the lure, and making the jump from guided performance to offered performance, is a feat that makes this seemingly simple method far more complex. Some trainers mitigate this problem by turning the lure into a hand signal cueing the behavior. Others keep the dog’s brain in the game by using a non-food lure, like a target stick. Still, despite these drawbacks, skilled and unskilled trainers are able to use luring to produce a huge variety of behaviors.
If luring is both dog- and trainer-friendly, why do some clicker trainers complain that luring isn’t “true” clicker training? When using the lure, the click is essentially superfluous. The reinforcer can be delivered at the moment the behavior is complete, the conditioned reinforcer imparts no additional information.
But the real problem with luring is that the trainer is doing all of the work. She tells the dog everything he needs to know, helps him through every step—and helping the dog becomes habitual, for both of them. For a hobby trainer, this may not be a problem, but for a serious trainer, a dog who cannot problem-solve, who cannot work his way through a puzzle, is ultimately limiting.
When I hear a trainer complaining that he really can’t see what the big deal about clicker trainer is or that he isn’t getting the mind-blowing results that other claims, almost invariably I find that he is luring almost exclusively.
Each of the three methods for getting behavior that we discussed is a valid way of getting behavior, each with significant pros—and significant cons. Social facilitation and other related methods are excellent when working with a group of dogs, but are limited primarily to naturally occurring behaviors. Molding and luring are easy and beginner-friendly, and luring is extremely flexible, but neither requires much mental effort by the dog, and both require significant fading of trainer participation.
In the final installment of this series we’ll look at three more dog-friendly methods of getting behavior: targeting, capturing, and shaping. Then, once we’ve got a comprehensive view, we’ll tackle once and for all the question of “Which method should I choose?”
In the first two installments of this series we discussed how training method affects how dogs think and learn, and we looked in depth at social facilitation, molding, and luring. In this final installment, we’ll look at the pros and cons of three more dog-friendly methods of getting behavior—targeting, capturing, and shaping—and we’ll formulate some realistic strategies for choosing the best method for achieving your training goals.
Targeting is frequently lumped in with luring, and, indeed, the two methods have much in common. But targeting has some unique characteristics and deserves to be considered on its own merits.
Basic targeting—touching a specific spot with a nose or paw—is frequently one of the first behaviors budding clicker trainers ever learn. Some teach their dog to touch the end of a target stick, and then use the stick to lure the dog through future behaviors. Others teach their dogs to touch their finger, their palm, or something like a margarine lid.
For many, this simple touch is all the targeting they do. Others go a step further, generalizing the touch cue to get their dogs to interact with other objects or using targets placed in specific places to move their dogs around. Common uses of targeting include:
- Transferring the touch to a door or a light switch to teach common service dog behaviors
- Teaching a dog to pick up an indicated object
- Touching specific areas of the contact obstacles in agility
- Teach the dog to “go out” to the ring fence in competition obedience
Some trainers are going even further with their targeting. First, they teach their animals to target with different body parts. They might have separate cues for targeting with each front paw, each back paw, shoulders, hips, nose, chin, and chest! Then they use a series of targets and combinations of these cues to create chains of behaviors.
For example, to teach a dog to “pray,” a trainer could ask the dog to sit, cue him to target his left front paw to the trainer’s knees, cue him to target his right front paw on top of his left paw, and then cue him to target his chin to his chest. Chain it together, and you’ve got a fast route to complex, precise trick!
This method has been used with great success in zoos and aquaria for many years. With exotic animals, it’s often difficult to use a traditional lure because the trainer has limited (if any) contact. Targets give those trainers a limited set of powerful, well-known behaviors that they can call on for a variety of uses.
Uses I have seen include teaching each member of a group of chimpanzees to go to and remain at his own target while the trainer works another member of the group; teaching an elephant to continuously target a spot with his forehead and then, one at a time, targeting a hole in a fence with each foot so his nails and foot pads can be cared for; and teaching a gorilla to grip and hold a bar, which is moved further down a sleeve, so eventually he is presenting his arm for blood draws.
As positives, targeting is both fast and flexible—creative trainer can use targeting to teach most any behavior. But the method does have some drawbacks as well. Unlike luring, the target and trainer participation can’t always be fashioned into a hand signal used to cue the behavior. Trainer participation is high, and it’s not always easy to fade the target and the trainer from the picture, especially for beginners.
Additionally, the level of problem-solving performed by the dog is limited. After the initial targeting is taught, other behaviors are simply combinations of the initial behaviors chained or back-chained together. The trainer shows the animal exactly what is required and continues to show him until he learns the pattern. Since all the information is provided to him, there’s just no need for creative problem solving.
Capturing is usually the first “hands off” training technique tried by most new clicker trainers. The concept is incredibly simple: When the dog does what you want, click and reinforce it!
Capturing is necessarily limited to behaviors that occur naturally in their finished form. It’s rather unlikely that the average dog is going to offer a full set of weaves the first (or second or third or…) time he sees the poles, but it’s a pretty good bet that he’s going to sit, lie down, or bark at some point.
It’s also limited to behaviors that occur with enough frequency that the dog can figure out a pattern to the click. It seems obvious to us what we’re clicking, but the dog may not be focused on that particular aspect—or any aspect—of his behavior at that moment. It’s only with consistent capturing of that behavior that he can figure out the common denominator of each clicked situation.
Like the Boy Scouts of America, the motto of any trainer who wants to capture behavior should be “Be Prepared!” Behavior happens quickly, and if you aren’t ready, an opportunity to catch it can be missed. This doesn’t mean you have to follow your dog around, clicker in hand, twenty four hours a day. Instead, identify the times that the behavior most commonly occurs or the events that generally precede the behavior and be ready to capture the behavior then. For example, to capture a bow, catch your dog when he is stretching after waking from a nap.
Although capturing is limited to the frequently-occurring behaviors included in a dog’s personal repertoire, it ranks fairly high on problem-solving ability because the click is the only information given. The dog must experiment to work out what he can do to earn a reinforcer. Happily, nearly all of the behaviors desired by pet owners occur frequently enough to be captured. It’s quite easy to teach a complete beginner’s class—even a class of pet owners with no desire to become trainers—using only capturing.
Another pro to capturing is that it teaches new trainers to anticipate behavior and to see the smaller responses that occur just before the desired behavior. This, of course, is the first step on the road to shaping.
Shaping is sometimes portrayed as the Holy Grail of clicker trainers. Some go so far as to say any method other than shaping isn’t real clicker training. Its perceived complexities terrify neophytes and some experts. Some are too intimidated to try. Others, discouraged by their first fumbling attempts, decide it isn’t worth the effort. Others follow the instructions but find their dogs didn’t read the same manual. What, then, is so great about shaping? In a word, power. Unlimited power.
Before we cover the pros and cons of shaping, let’s talk more about what it is. Shaping is another term for successive approximation. Complex tasks are broken into achievable chunks, and then taught bit by bit. Shaping comes in two “flavors.”
In the first, commonly called free-shaping, the trainer creates a new behavior “from scratch.” For example, to shape a competition-quality sit, the trainer might start by clicking any sit. Then he would increase his criteria and click only tucked sits. Then only tucked sits on the haunches. Then tucked sits on the haunches with even front feet. And so on.
The second type of shaping builds on an existing behavior. This is how elements such as distance, duration, and distractions are incrementally added. For example, once that competition sit is on cue, duration is added gradually. . .half a second, one second two seconds, four seconds, seven seconds, ten seconds, and so on.
Although I speak of two flavors of shaping, they are really applications of the same method. Shaping is shaping. However, beginners often find using shaping to add elements to an existing behavior is easier than free-shaping a brand new behavior. Using successive approximation to add distance or duration is a great way for trainers to experiment and get comfortable with the technique and to begin to hone their observation skills. However, the real power of shaping, and its major benefits, come from free-shaping.
Shaping is, by far, the most flexible training method available. It is limited solely by the animal’s ability and the trainer’s skill. You can’t teach a dog to fly, but if you had the skill, you could teach him to do a solo freestyle routine cued by the music. For an experienced dog and trainer team, shaping is as fast as targeting, and its precision absolutely can’t be matched.
The key, however, is experienced. Shaping isn’t fast for new shapers or inexperienced animals. It takes skill to break behavior into responses so tiny that your dog can be consistently correct, even when you increase criteria. It takes skill to see subtle nuances of behavior and to anticipate them well enough to time your click perfectly. For the dog, it takes a lot of brain power to work out what a click means (especially if the click is less-than-perfectly timed). It takes creativity and a willingness to experiment and make mistakes. If the dog or the trainer doesn’t understand the method, isn’t comfortable experimenting, or isn’t a good problem-solver, free-shaping can be very frustrating.
Despite these significant challenges, do I think it’s worth the effort to learn to shape? Yes, for most anyone interested in truly learning to train, I do. So often, I see clicker trainers stick to luring and cheat themselves out of possibilities they truly can’t imagine.
Clicker trainer Sue Ailsby has a Portuguese Water Dog named Scuba who has been extensively free-shaped. Watching Sue shape a new behavior is a jaw-droppingly amazing experience. It takes just one or two clicks at each criterion for Scuba to catch on, so she learns at seemingly light-speed. It’s as close to pure communication as training can possibly be. Until you see it, however, you simply can’t conceive of it. Not really.
Shaping is unlimited possibilities.
So which should I choose?
Each of the six dog-friendly methods of getting behavior is a legitimate technique. All have pros and cons. None are suitable for every trainer, every animal, and every situation. I firmly believe that a good trainer should understand and be able to use all of the methods.
Choosing the right method for a particular situation depends on several factors which were covered in part one of this article. Briefly, they are:
- How many behaviors will this dog need to learn?
- How quickly do I need to train the behaviors?
- How precise are the behaviors?
- How skilled is the primary trainer?
- How interested in training is the trainer?
- How experienced is the dog in each method?
- Will the dog need to “think on his feet,” or will he primarily respond to well-rehearsed cues?
Sometimes the best solution is to use a combination of methods. Use a lure to get the dog focused on the general type of behavior you want to work on, and then free-shape. Set up your environment in a way that your dog can’t help but succeed—a type of molding—and then capture what you want. Stretch your imagination and your boundaries!
No matter which method you choose, remember that training is just a tool we use to deepen our communication and relationship with our dogs. Don’t get so focused on the result that you forget to enjoy the journey. Clicker training is the road less traveled, and that makes all the difference!