Fox Trotting With Dogs – Attaching commands to behaviors.

Imagine that you are having a bad dream. You are dressed in formal evening clothes and it is obvious that you are competing in a ballroom dance contest. When your number is called, you grasp your partner and glide out onto the floor — only to realize that you don’t know how to dance. As the music starts, your partner hisses at you through clinched teeth, “Fox-trot, you idiot!” and begins to dance. As the command to fox-trot fails to cause you to do the dance, your partner starts chanting the word “fox-trot” and tries to pull and tug you around the dance floor. Soon the crowd has realized your ignorance and wishes to help you by yelling, “FOX-TROT! FOX-TROT! FOX-TROT!”. This is obviously a vain hope that you will instantly be able to do what you don’t know how to do, merely because they are chanting the name of the dance. As you awaken from this nightmare you breathe a sigh of relief. Only in a nightmare is someone expected to learn and perform a complex behavior by being pulled and tugged while people are yelling unintelligible words at them – unless you are a dog.

One of the most obvious and generally unknown aspects of training is that teaching a behavior and attaching a command are two completely separate tasks. There is a logical sequence to this process. A command without a behavior is as useless as yelling fox-trot to someone who doesn’t know how to fox-trot. A behavior without a name is equally dysfunctional – the person may know how to fox-trot but can’t connect the knowledge to a setting where it would be the correct dance to perform. That means the animal must know how to do a behavior before you can possibly “command” the behavior to happen. If Fluffy does not know how to retrieve a ball, it will do no good to chant “fetch” or “ball” while she is learning to fetch. Waving one’s hands around won’t help either. Screaming and jumping up and down will only make Fluffy convinced that you have some kind of nervous condition. The only way that Fluffy can learn that “fetch” means “go get the ball” is if she already knows how to retrieve — and has the desire to do it. A part of your brain is nodding yes to this concept. The rest of your brain is either saying “huh?” or “wait, that doesn’t make sense. How can you get a behavior to happen if you don’t tell the animal what to do?” If you go back to our fox-trot metaphor, you’d answer your own question with “how can a dog do anything merely by throwing words at it?” More to the point, do the words you chant help or hinder your dog’s learning? Vast evidence indicates the words either do nothing to aid learning or actually distract from effectively connecting the command to the behavior. Here’s why.

The traditional and most common way to train is to chant commands at an animal while attempting to manipulate or lure it into position. Because the animal is forced to chose between paying attention to the commands, or learning the behavior, one or both will suffer. Either the animal understands the behavior but must be told several times to do it, or the command is obeyed instantly, but in a sloppy fashion. If the animal decides to focus on the behavior, it must ignore the chanted command while it learns the new skill. This will cause the dog to intentionally ignore the command as it is learning and after the behavior is mastered. Specifically, if you have been chanting fetch while Fluffy learns the behavior, she must actively ignore the word “Fetch” to focus on the flight of the ball and where it lands. Once the behavior is learned, the command is effectively dead in the water. You will have to go through extra lengths to take the now meaningless command and connect it with the behavior.

While all this may seem perfectly logical, the implications may still feel alien. If one does not talk to the animal first, how can the behavior be taught? If one trains silently, how does the command become connected to the behavior? Before the strangeness of the concept causes you to reject it automatically, consider how you could use this method to teach your pet to sit. (You will notice that I said “pet”, and not “dog– this method works on almost all mammals and many birds.)

* Lure a dog into a seated position. Don’t say anything as you do it. (If you aren’t familiar with luring, this following is a description of the process.) Take a favored piece of food and hold it in front of the animal’s nose. As the critter starts nibbling, slowly move the treat over its forehead. If the animal relaxes its rear end a little, give it the treat. Over a series of repetitions, try to get the animal to relax a little more each time until its rear end sinks to the ground and the animal “sits”. Continue with this process at least 20 more times so that the animal learns that it can always get the treat by sitting.

* Now that Fido, Fluffy or Flicka is consistently sitting in order to get the treat, touch a treat to his/her nose and then quickly put it behind your back. If the animal sits, give it the treat. If it stands there and looks confused for several seconds, be patient. Wait at least 30 seconds to get the animal to offer the behavior. If the animal does not offer a “sit” during the 30 seconds, lead the animal by the nose as you did in the beginning to make the behavior happen again. Try it ten more times.

* Fido, Fluffy or Flicka will now sit whenever a treat is presented. To attach a command to the behavior begin saying “sit” just before you think the animal is about to do it. How do you know when it is going to happen? The repeated reinforcement for the behavior has gradually put it on a cycle. Technically, for the first several attempts to attach a command, you are merely guessing that the weight of your reinforcements will cause the behavior to happen. If you have done the procedure correctly, it’s a pretty simple guess. By this time, you already know approximately how long it takes from the moment you tease the animal with your treat and the onset of the behavior. (Note: Though you may not observe it with a behavior so easily taught as “sit”, in many cases, the first time you attempt to attach a word to a behavior the animal will instantly forget the behavior. As I said earlier, the animal is either going to pay attention to the behavior or the cue as it learns the pattern. That very first time you give the signal, you are going to draw your animal’s attention away from the behavior. That goes along with what I said about the dog paying attention to either the behavior or the signal, but not both at the same time. Don’t worry about it. Reinforce the behavior a few more times and the word/hand signal will no longer be novel and the behavior will return.)

The pattern is unfolding like this…

  1. You touch a treat to the animal’s nose
  2. The animal pauses for a few seconds
  3. The animal sits
  4. You deliver the treat
  5. There is a small delay as the animal eats the treat, realizes that you are at the end of the sequence but has not yet offered another behavior.

What I am suggesting is that you can now do this pattern…

  1. You touch a treat to the animal’s nose
  2. The animal pauses for a few seconds
  3. The animal sits
  4. You deliver the treat
  5. There is a small delay as the animal eats the treat
  6. You say “Sit” or any other word you choose.
  7. Repeat

If you do this 20-50 times, the animal will start to connect the originally meaningless word to the behavior. Note that I said “start to connect.” Ivan Pavlov, the great Russian physiologist, said it took him about this many repetitions to connect the sound of a bell to food. That simple association is nowhere near as complex as attaching a cue to a learned behavior. In practical terms, the real number of times you have to repeat the sequence for the animal depends on a number of factors, but will probably be between 200 and 300 repetitions over several training sessions. Don’t let that number scare you. I recently taught a 12 week Husky puppy to “down at a distance.” I did 65 repetitions of the behavior in 15 minutes. By that standard, I can connect the word “Down” or a hand signal to the behavior in about 5 sessions to have a rock-solid connection between the command and the behavior. By contrast, many trainers do a half-baked job of attaching the cue to the behavior and end up with literally thousands of attempts to connect the two, while never really getting solid performance. Taking the process one step at a time in the correct sequence reduces the number of repetitions and greatly increases the dependability of the cue.

Methods that attempt to teach the behavior while simultaneously chanting the command often fail to develop consistent performance. Teaching the behavior first is an efficient alternative that can help your pet be more obedient. Avoiding your pet’s worst nightmare is as simple as teaching him the step before you ask him to dance. Here are some other thoughts that may help you see this perspective on the process.

  • I have a simple rule of thumb about when it’s time to attach a command to a behavior. If you aren’t willing to bet $5.00 that the behavior will occur, naturally, within a given 30 second period, it’s too soon to add the cue.
  • Some dogs connect signals to behaviors faster than others. Stay in this for the long haul and you will create a rock-solid association between a specific behavior and a specific cue. If you bail and starting throwing words around, I recommend that you use “fox trot” for all commands. If you are going to be impatient and use sloppy technique, you might as well advertise it.
  • If you wish to have multiple signals for a behavior, put the signals in sequence. Example: The dog will lie down when you tease it with a treat. In between the repetitions, make sure the dog is looking at you. Offer your hand signal of choice. Wait a few seconds. Say the spoken word. The dog then does the behavior, not because you waved your hand or spoke, but because it’s on a cycle. Continue with the pattern repetitions and soon the signals will start triggering the behavior. After you see that the first signal is triggering the behavior, switch the order and move to using the signals interchangeably. Note: some dogs are more visual than auditory – they will learn hand signals faster than spoken commands and vice versa. You may have to work on one signal more than the other to keep them working at the same level of dependability.
  • Never give any two signals at the same time while attaching them to a behavior. Do them in sequence, like this…
  1. hand signal
  2. pause
  3. spoken command
  4. behavior happens
  5. click/praise
  6. treat

The reason to put them in sequence is to guard against creating a compound signal. If that happens, the dog will require you to give the hand signal and the spoken command before it does the behavior. When this occurs, if the dog can’t see your hand signal, the spoken command isn’t going to trigger the behavior by itself and vice versa. Your goal is to make each work independently.

  • A command can be anything the dog can perceive – a sound, a sight, a touch or a scent.
  • Never offer a hand-signal or spoken word twice-twice. Only do it once-once. Otherwise, your dog will require you to always do it twice-twice. Don’t believe me? Try going into a toy store and asking for a “yo.” You’ll get blank stares unless you say it twice-twice. This phenomenon is not a matter of intelligence. Your dog is a very literal creature. If the command is consistently sit-sit, you’ll always have to say sit-sit and the single word, sit, will draw the same blank stare you got when you asked for a “yo” at the toy store. This is very similar to the reason you don’t want to give spoken commands and hand signals at the same time. Your dog will swallow whatever you consistently use as the signal, even if it doesn’t make sense.
  • Never teach or give a hand signal that is in front of your body. Hand signals should include a change to your silhouette and movement. Dogs are shorter than we are an often you are backlit. Your silhouette looks like a big black blob to the dog. Your hand signal will be virtually invisible. If you need ideas on what would make a good hand signal, look up the word semaphore on the internet. Semaphore is an old form of signaling used on land and sea that uses flags to represent letters of the alphabet. You can see a good graphic of this at http://www.infovisual.info/05/070_en.html
  • My last piece of advice is simple. Never be in a hurry to attach a command to a behavior. The longer you wait, the better the behavior will be and the more reliable the command.

Leave a Reply

Your email address will not be published. Required fields are marked *