Lures, Rewards and Bribes

iStock_000005304730XSmall.jpg

"I don't want to bribe my dog" is a frequently cited criticism of using food in training. Neither would I. I would no more want to bribe a dog with food, than I would want to bribe it with toys, games, attention, or affection. In fact, I wouldn't want to bribe a dog at all. Bribing seldom works. And when it does, it seldom works for long.

Lures, rewards and bribes have all been used to attempt to modify the behavior of animals and humans. Whereas bribes are ineffective, lures and rewards have many highly specific uses for teaching any animal almost any exercise. Moreover, lure/reward training is simply the fastest way to train any animal.

Lures

A lure is presented before a desired behavior, specifically to entice the trainee to perform the required response voluntarily and on cue. When luring dogs, it is taken for granted that the dog would gladly do what we wanted, if only he knew what we wanted him to do. Luring specifically teaches the dog which response is required for each request. In a sense, luring teaches the meaning of the commands — English as a Second Language.

Without the use of lures, when asking an eager-beaver novice dog to sit for example, even though the dog may be perfectly willing and eager to please, he has no idea what we want him to do. In fact, the dog may have no idea that we want him to do anything at all. Once he eventually realizes we want him to do something but not knowing exactly what, the dog may rapidly run through his entire behavior repertoire in a frantic quest to find at least one response to please his owner. Characteristically, the average border collie will zoom up to his owner, sit, lie down, roll over, sit up, bark three times, spin right, spin left, circle the owner — come by, 'way to me and jump up and down, all in the space of three seconds. The owner’s timing has to be impeccable to reward the desired response. Obviously, the dog is only trying to please his owner but he just doesn't know what the owner wants. The dog's enthusiastic responses tend to irritate less patient and less understanding owners and sadly, some dogs are reprimanded for their enthusiasm.

Some dogs fail to fathom that the request to sit means action. After all, we witter away to our dogs so much of the time that the dog may have learned to interpret all human dialogue as Larsonian verbal diarrhoea — Blah! Blah! Blah! Blah! Blah! If the dog doesn't sit but remains immobile however, his inactivity is often interpreted as stubbornness and/or dominance by a less patient and less understanding owner and sadly, the dog is physically corrected or punished.

On the other hand, by simply requesting the dog to sit and moving a lure appropriately, not only does the dog understand that action is both required and desirable but also, which specific action is required ... and the dog sits. It's as simple as that. Without a doubt, luring is by far the quickest way to teach a dog the meaning of words that we use as requests, directions and commands and eventually as instructive reprimands. Lure/reward training is a user-friendly and time-efficient technique for all trainers, especially those with limited patience, i.e., many novice owners.

For most exercises, luring is an all-or-none response. Either the dog sits quickly and completely, or he doesn't. However, even when luring proves to be tricky, once the dog has been lured successfully on just one occasion, subsequent trials rapidly become progressively easier, such that in no time at all, the dog will lure quickly every time. When teaching some body positions (e.g., down and rollover) or when teaching more complicated responses (e.g., heeling and retrieval), luring may be used effectively to facilitate a shaping procedure, whereby the dog is rewarded for successive and progressive approximations to the ultimate response.

Now, some trainers might respond, "Why not physically guide the pup into the required body position?" Well, firstly, whereas an experienced trainer could easily guide a dog into almost any body position, a novice owner might not have the necessary skill or patience, especially when working a novice dog. Moreover, when training the dog at home, frustration at being unable to copy a simple exercise demonstrated by the trainer in class, frequently manifests itself as manhandling and the novice dog is pushed and pulled around.

Secondly, physically prompting the dog to assume each desired position delays the ultimate learning process of establishing verbal control. By using a lure in training, the dog immediately starts to learn the meaning of hand signals (hand lure-movements) and verbal requests from the very first trial, whereas any physical contact introduces an additional, unnecessary learning stage. Virtually all untrained dogs selectively attend to physical contact rather than the words we use. Certainly, dogs quickly learn collar and rump contact means sit, but it is still necessary to teach the meaning of hand signals and verbal command, which is, after all, the whole point of the exercise.

Thirdly, using physical prompts delays mastery of off-leash distance control. Since dogs are lure/reward trained off-leash from the outset, the trainer does not come to rely on proximity and physical contact and the dog learns distance commands quickly. This is beneficial in itself but also solid distance control is essential for proofing distance stays. (When a dog is over thirty yards away for example, a distal instructive reprimand is the only way to correct the dog instantaneously.) However, distance training is a much more time-consuming and complicated endeavor if the trainer has become reliant on physical prompts (and maybe physical corrections) by hand or leash. Yes, the dog may respond if he’s on-leash or within arm’s reach but it’s a different matter when he’s off-leash and just four feet away. Physical contact in training quickly becomes a crutch for the owner, often fostering an over-dependency on various other physical training tools. Basically, the leash is overused simply because it's there, i.e., dogs come to class on leash. However, it does not necessarily follow that we have to overuse the leash during training, even when teaching walking-on-leash or precision heeling. After all, most good shepherds and gundog trainers achieve formidable distance control by teaching the dogs off-leash using sheep and bumpers as effective lures.

And of course, there’s a fourth reason why we train puppydogs off-leash; off-lead control is what the owners want. Dogs spend the majority of their lives off-leash at home. Thus, much leash training is irrelevant for most pet dogs. Instead, the owner requires off-leash control and they require it urgently.

Rewards

Rewards are given to the dog after the desired response, specifically to reinforce the immediately preceding appropriate behavior so that it is more likely to occur again in the future. In a sense, rewarding a dog following desired behavior teaches the dog the relevance of the exercise, so the dog learns why it should do what we request.

Praise is certainly the best reward since it may be used at almost anytime and in any scenario, especially even when dog is some distance away. Additionally, the dog’s favorite toys, games and activities and of course affection may all be used effectively.

When initially training the pup in the home and for the first couple of classes, most novice owners seem to get on better using small, discrete, palpable rewards, such as toys and treats. Few novice owners have the requisite, exquisite timing of an experienced trainer and usually, it is easier for the owner to learn appropriate timing using toys and treats rather than praise and affection. Also, not all novice owners have the inclination to praise their pup, especially in public. Indeed, many owners have to be taught how to praise, just as they have to learn how to reprimand.

Bribes

A bribe is offered, or promised, before a required behavior in an attempt to coerce the trainee to perform the specific task, usually against its will. Unlike luring, which is a precise educational process used to teach a willing puppy and/or novice dog the meaning of the Request, we assume bribed individuals know what we want, but simply just don't want to do it. Thus, bribery is a coercive attempt to corrupt the will of the trainee. And as such it is bound to failure. Additionally, when a dog doesn’t want to do what his owner asks, the training program is already off-track and the dog/owner relationship is out of whack. Now, that is not to say, all dogs will always comply. On the contrary, the integral purpose of reward training is to teach the dog to want to comply. Teaching willing and eager compliance is easily accomplished using rewards, but not with bribes.

Unlike reward training, whereby specific and desired behaviors are reinforced by rewarding the dog after he has responded, bribes are offered or promised before the dog has done anything. In a sense, a bribe is an attempt to reinforce a response before it has happened, which of course is theoretically and practically impossible!

Not only is bribery ineffective, it creates all sorts of training problems. Some trainees may gladly accept the bribe but then still refuse to do what the trainer wants. Other trainees may comply if a bribe is in the offing but otherwise refuse. Indeed, bribe-contingent reliability is the most common problem created by dog trainers, parents and politicians misusing (otherwise extremely effective) lures and rewards as bribes.

A Potential Problem

A combination of luring and rewarding — luring the dog to do what you want and then reward him for doing so — is the fastest way to put behaviors on cue. However, when the same item (food or toy) is used both as a lure and as a reward for any appreciable length of time, lure/reward training approximates bribing. The dog’s response becomes unreliable and contingent on the trainer having food in their hand or pocket. If the owner has a toy, a treat, or a happy frame of mind, the dog may do it, when he’s good and ready, but … if the owner doesn't, the dog probably won't! These potential problems may be easily avoided by using different items as lures and rewards and especially by beginning to phase out the use of food lures and food rewards during the very first training session.

Use a wide variety of lures and rewards in training and specifically, use different items as lures from those used as rewards. For example, lure the dog to sit with a squeaky toy or chewtoy but offer a food treat from your pocket as a reward. Or, use a food treat to lure the dog to sit, down, sit, stand, down and rollover but then praise the dog and invite him up on the couch as a reward. Or, lure the dog to sit with a hand-signal but throw his tennis ball and say "Fetch" as a reward.

If you use food as both lures and rewards, then during the dog’s very first training session, the prime directive is to phase out the use of food lures and to begin to phase out food rewards, or any other training tool for that matter. Lure/reward training comprises five stages, the first stage involves phasing out training lures and the second stage involves phasing out training rewards (usually food).

There are occasions when it is essential for the dog’s safety that the dog respond quickly and reliably whether or not we have food or toys available. Consequently, we want to be absolutely certain that the dog’s compliance and motivation is not contingent on the trainer having food in their hand or food pouch. However, once response-reliability is assured, food and toy training rewards may be brought back to refine precision and panache. Also of course, we would never phase out the use of food for classical conditioning — to teach the dog to thoroughly enjoy the proximity and actions of other dogs and people, especially, children, men and strangers.

Phasing Out Lures

Food lures should be phased out once the dog has learned the meaning of the hand-signal (hand lure movement) after just 6-12 repetitions. From then on the hand-signals may be used as lures to teach dog the meaning of verbal commands — a process that requires precise timing. The verbal request must precede the hand-signal by one third to one half of a second. Additionally, the trainer must not move (or even flinch) when delivering the verbal request, otherwise the dog will respond to the movement rather than the verbal instruction.

Phasing out food lures is easy; just go cold-turkey. After successfully food-luring the dog to come and sit, put all of your food treats into your pocket or food pouch and show your empty hands to the dog, then back up and repeat the same routine as if still holding a food lure in your hand. After the dog has come, sat and laid down, praise the pup profusely and offer a food reward from your pocket. Thus, the dog learns, if he follows directions (verbal requests or hand signals), he may still receive a food reward, even though one was not in your hand at the outset. Well done. You have now learned that you do not need food in your hand to get your dog to comply.

Certainly, verbal praise is the most useful and easiest-to-use reward. However, do not fall into the trap of using a loving tone and sweet expression as a lure to entice your dog to respond, otherwise you’ll end up with a dog that may respond (when convenient) if you speak in a sweet tone but will probably not respond if you shout. The prospect of praise should not be apparent before the dog responds appropriately. Always give requests/commands in a neutral voice with a neutral facial expression. Call the dog in a neutral voice but then giggle, and praise the dog happily as soon as he arrives.

However, it is also a smart idea to teach the dog to respond quickly to a shouted command. Remember, emergency commands are invariably shouted, or screamed in an ugly tone. It would be a disaster if the dog interpreted the tone and volume to mean that you were angry and so failed to respond, or even ran away. Instead, we want the dog to learn that a requested recall usually brings praise and reward and a shouted recall signifies urgency and always brings better rewards. The process is simple — gradually and progressively change the tone and/or increase the volume with successive trials and give the dog better rewards as the volume increases and your tone gets gruffer.

Phasing out Rewards

Implement a differential reinforcement schedule from the getgo and at the very most, only reward the dog’s above-average responses, i.e., no more than 50% of responses. (And of course, better responses get better rewards and the best responses get the best rewards.)

The number of food rewards may be further reduced by progressively asking the dog to do more for less. The first night in puppy class, the average Golden Retriever will happily complete ten doggy push-ups (sit-downs) in a row, i.e., twenty consecutive responses for the prospect of a single food reward! Thus, 19 out of 20 food rewards are unnecessary and may easily and quickly be dispensed with right away. So, by using verbal requests and hand signals only, instruct the pup to come for a food reward, then to come and sit for the second reward and then come, sit, down, sit, stand, down and stand for a third reward and so on. Gradually and progressively increase the number of responses required for a single reward.

Next, empty your pockets and take off your food pouch. Either put the food on a table to use as a distraction as well as a reward, or give the food to another person to reward the dog whenever you say, ”Good dog!” Thus, the dog learns, if he follows directions (verbal requests or hand signals), he may still receive a food reward, even though one was not on your person at the outset. Well done. You have now learned that you don’t need food in your pocket or food pouch to get your dog to comply.

Now, replace food rewards with much more valuable life rewards, e.g., sit-play, sit-play, sit-sniff, sit-sniff, sit-walk, sit-walk, etc. Walking, sniffing and playing are the three most effective rewards in training. Integrate short training preludes and interludes into the dog's lifestyle. Ask the dog to sit, for example before every enjoyable activity — before you put on the leash, before letting him off leash, before telling him to say hello to other dogs, before throwing his tennis ball, before allowing him upon the couch and of course also, have the dog sit for his supper. The dog will soon enjoy sitting when requested. Similarly, frequently instruct the dog to sit for numerous short training interludes during each enjoyable activity. "Rover, Sit. Good dog. Go play." "Rover sit. Goooood dog. Go sniff." Wonderful! All potential distractions that otherwise would have competed with training, have now become life-rewards that facilitate training.

Eventually, all external rewards become unnecessary as the dog becomes internally motivated and the responses become self-reinforcing — Natural Motivation. Now the dog does it, because it wants to. In a sense, each cued response has become a reward for the dog.

This article was based on Dr. Dunbar's monthly Behavior column in the January 1992 issue of The American Kennel Gazette. Reprinted with the permission of the author and The American Kennel Club.