Cues: Knowing Your ABCs
You have taught your dog that certain behaviours will earn him good consequences (Reinforcement or food and fun). You have taught him the following:
B (Behaviour) > C (Consequence)
The next logical step is to teach your dog which behaviour you would like him to do and when you would like him to do it. This is a cue: a word or signal that lets your dog know when you want him to offer the associated behaviour. The use of cues or commands is where clicker training is very different from traditional training methods. Traditional training methods are command based; they operate on the assumption that commands drive behaviours. They say a word such as Sit, then manipulate the dog in such as way as to get him to perform the behaviour associated with the command. This is absolutely backwards and although dogs can be taught with these methods, they are inefficient and often stressful for the dog.
An Antecedent is something that comes before a behaviour. Cues are a form of Antecedent called discriminative stimuli: an antecedent that lets your dog know that if he performs a certain behaviour now, he is likely to get reinforced for it. When you put behaviours on cue, you now have the following:
A (Antecedent) > B (Behaviour) > C (Consequence)
It is useless to teach the dog a cue before he understands that the behaviour will earn him reinforcement. This is why you must reinforce the behaviour first through shaping; you must first focus on the B > C. Remember, consequences drive behaviour, not cues. Cues merely suggest to the dog which behaviour will earn him reinforcement at that particular moment.
Let's say you are out in your car and stopped at a red light. You wait until the green light flashes, and then you start driving. It may appear that the green light is what prompted you to start driving but it did not. You start driving because you need to get home from work. Arriving home is the happy consequence to the behaviour of driving. Going home reinforces driving. The green light merely suggests that it is an opportune time to resume driving but it does not compel you to drive. The green light is mere an antecedent or a cue.
In order for a cue to function as an antecedent, it must be predictive which is to say it must come before the behaviour. To add a cue then, you would give the dog the cue before he engages in the behaviour and the cue should not be attached to any other behaviour. For example, if you want to teach your dog to sit, you would begin by shaping it. Once you click the dog for sitting, the dog will begin to offer you sits. He must be offering sits in a predictable manner so that you can add the cue just a moment before he sits. If you say Sit while he is already sitting, the cue will not work because it will not be an antecedent. It must come before the behaviour to be effective. On the other hand, if you say Sit and your dog is not ready to sit, he will not learn the cue because the cue will again not predict the behaviour.
In order to add cues, your dog must first be proficient at the desired behavioiur. If you want to put Sit on cue, your dog should be offering sit regularly and predictably. If your dog is offering you a sit every two to five seconds, you can easily add the cue and know that the dog will offer the behaviour after the cue. Putting sit on cue will look a little like this:
You cue the dog to Sit (antecedent) > Dog sits (behaviour) > Dog gets cookie (consequence)
You must be careful when adding cues that you are actually getting the behaviour you want. If you want your dog to sit straight up and not flopped over on one hip, do not add the sit cue until your dog is sitting straight with predictability. If you add the Sit cue and your dog flops over onto his hip, you are attaching the Sit cue to the wrong behaviour. When it comes to cues, the quality goes in before the name goes on!
Now here is the best part about clicker training. Cues will become rewarding for your dog. He will love it when you ask him to do something for you. If you are skeptical of this, you just need to learn a bit more about the science behind behaviour.
You might remember Ivan Pavlov from high school or college. Ivan Pavlov was a physiologist doing research on digestive function and he used dogs as a part of his experiment. He would feed the dogs a meat paste and measure the amount of saliva and digestive juices produced by the dogs in response. He started to run into difficulty however, when the dogs started salivating before he was ready to measure them. As the dogs gained experience, they quickly learned that white coats proceeded meat powder and they would begin salivating as soon as somebody entered the room.
Despite earning a Nobel prize for his work in physiology and digestion, Pavlov was fascinated by this response. He changed gears completely and began the experiments that most people are familiar with today. He would ring a bell before feeding the dog and the dog would soon salivate every time they heard the bell, whether food was present or not.
Pavlov called this Classical Conditioning or relational learning. The dogs learned that because the bell always preceeded food, the bell was a predictor of food. Since food was a stimulus for salivating, the bell that predicted food become a 'conditioned' stimulus and the dog reacted to the bell the same way he did food.
Within the context of Classical Conditioning, cues become predictors of food. This is why it is important to train the behaviour before adding cues: when we add the cue fully knowing that the dog will comply, he will get a reward in response. In this way, cues such as Come or Sit are conditioned stimuli and our dogs look at them as money in the bank! Your cues will trigger an emotional response in your dog that will motivate him to want to comply.
Now that you understand how to add cues, you just need a few more details on develop skill with the clicker.
Clicker Skills