There are two words in the dog training world that may seem very similar at first glance, but have vastly different meanings. One is a term that has been around since the beginning of animal training. The other is just starting to become more commonplace. This change in terminology is not just a matter of semantic shift over time, but is representative of an important change that is occurring in the industry. If a person calling themself a dog trainer uses one term instead of the other, it’s an automatic red flag that they are not up to date on the science of training and likely not a trainer whose advice I would follow.
The words I’m talking about are “cue” and “command.” Both terms are used to refer to information given to the dog, usually coming from the trainer, that is intended to cause the dog to do some behavior. I can “command” my dog to sit by saying the word “sit,” or I can “cue” my dog to sit by saying the word “sit.” Same thing, right? Not exactly. Not only is there a different semantic connotation, the word command conjuring up images of a brutal war general ordering troops into battle while cue brings to mind a lovely rehearsed stage production with actors hitting their marks, but the terms also imply vastly different approaches to dog training.
I believe it was the pioneering clicker trainer Karen Pryor who first referred to a command as a veiled threat. If I give my dog a command, then I am saying to them “do this or else.” In order for that to be even the least bit effective, I have to teach the dog what the “or else” is. If I want to command “sit” and my dog’s butt doesn’t hit the ground, then it follows that I have to punish them. Assuming that the intended punishment actually acts as a punisher, then the dog will put their butt on the ground faster to avoid the scolding or the leash jerking. Because a command is built by punishment, it comes along with all kinds of yucky feelings for the learner.
Many trainers who teach commands believe that punishment is a necessary part of the learning process to inform the dog when they are wrong. Some will even critique clicker training for being permissive of “bad” behaviors. This critique reveals a fundamental misunderstanding of clicker training. My dog knows very clearly when he is wrong - he doesn't hear a click. I can inform my dog that he made an error without all of the ugly fallout that comes along with using punishment.
The fallout of punishment-based training is extensive and well-documented. (Like in this study, this study, this study, this study, this study, and this study to name a few.) Use of punishment in training is associated with increased aggressive behavior and avoidance behavior. It is likely to create fear or anxiety and damage the relationship between the animal and the trainer. Even the occasional use of punishment reduces the effectiveness of a positive reinforcement training strategy. If you mix punishment with reinforcement, the learner stops wanting to learn. The unpleasant events shift the dog’s emotional state from enthusiastic interest to avoidance, stress, and caution. This creates a learner who is acting only to avoid unpleasant experiences, and therefore acts a lot less. Unfortunately, this is a dog that many in our society may see from the outside and deem “good.” It’s a dog who doesn’t try new things, who is shut down and reserved, and goes through their day just anxiously trying not to get hurt.
Not only is punishment ethically problematic, it is not the most effective. (Another great study just came out directly comparing the effectiveness of shock collars and positive reinforcement. Guess who won?) Punishment does not teach the correct behavior. It may suppress a “wrong” behavior temporarily, but that behavior will resurface in the absence of the punisher. Learners habituate to punishment, so handlers will have to ratchet up the intensity of the punisher over time. Punishment does not lead to predictable behavior in the future, and even if I’ve thoroughly suppressed the animal’s first choice in a given situation, I may not like their second choice any better. Behaviorist B. F. Skinner summed all this up best, writing, "Punishment does not actually eliminate behavior from a repertoire, and its temporary achievement is obtained at tremendous cost in reducing the over-all efficiency and happiness."
A cue on the other hand has no threat implied. It is attached to a behavior that is built through reinforcement. It is not enforced, but the animal performs the associated behavior because they want to. A cue informs the learner that there is an opportunity for reinforcement. Because a cue is built with positive reinforcement, it comes attached with all kinds of good feelings and causes the dog to want to do the behavior.
If I’m in the process of teaching a new cue and my dog gets it wrong, I simply don’t click and treat. I don’t blame the dog for getting it wrong, but I take the mistake as information for me, the trainer. The dog is simply telling me that they do not know what I am asking of them in this situation. I then ask myself a series of questions. Did I increase my criteria too quickly? Is the level of distraction too much at this point of the training? Is my dog tired? Thirsty? Is the reinforcer I’m using of appropriate value for the difficulty of the task? This level of critical thinking requires admittedly more skill, which is why less knowledgeable or experienced trainers so often default to correcting the dog.
Another reason that cues are so wonderful is because there are so many options. Very few things can be a command because of the heavy reliance on my ability to enforce the threat. Commands limit the ways that I can communicate with my dog. Cues, on the other hand, can be anything that the animal can perceive. Sounds, words, objects, scents, environments, movements, and hand signals can all be cues for certain behaviors. The smell of drugs, money, or weapons can cue a scent dog to alert their handler. The sight of a curb can cue a guide dog to stop so their person doesn’t walk into traffic. The sound of the refrigerator opening can cue a dog to run to the kitchen and start drooling. Thinking in cue terms, how so many different stimuli and events elicit action, gives us a much better understanding of our dogs’ behavior than thinking in command terms.
When I was a student at Karen Pryor Academy, one of our assignments was to use four different types of cues for four different behaviors. I was able to teach Jack to lie on a blanket on the word cue “mat,” bark on the sound of a harmonica, walk through a hula hoop on the visual cue of seeing the hoop, and spin in a circle with a hand signal. Watch the video below and see if you can identify the cue for each behavior you see.
The difference between a cue and a command really comes down to the relationship that I want with the animal in front of me. Do I want my dog to do what I ask because he is afraid, or because he is an enthusiastic participant in the training process? Do I want to have to follow through on my threat, or do I want to better set my dog up to be successful the next time I ask? Do I want my dog to wither and give up, or do I want him to be confident and feel that he has control over his outcomes? Those questions are easy for me to answer. I teach cues, not commands. If you want to learn more about how you can make your training more cooperative and less aversive, I am happy to help.