DunbarAcademy.com

Within our legal system and dictionaries, punishment is defined is defined by its nature (incarceration, corporal, or capital) and reasons. In society, punishments usually relate to our sense of moral and ethical justice; they are ways to right a wrong, and so many people feel strongly that punishment should be unpleasant or painful.

However, in the behavioral sciences, punishment is defined by its effect on behavior, for example: A stimulus that inhibits an immediately preceding behavior, causing it to be less likely to occur in the future., in that scenario. This definition makes no mention of pain or discomfort. The stimulus could be aversive, painful, or scary, or it could be non-aversive, pleasant, and welcomed. A leash-correction or instructing "Steady" would each be defined as punishment, if they resulted in a dog pulling-less while walking on-leash.

When learning theory made its transition to animal training, most trainers assumed that the terms 'aversive' and 'punishment' were synonymous, which led to two huge misconceptions that continue to this day:
    1. All aversive stimuli are punishing, and
    2. Punishment must be aversive to be effective.

Aversive Punishments are a Rare Beast
The first clue that these two assumptions were not true, came to me in a flash when I heard someone say, "I hate seeing all these aversive punishments," I thought, but that can't be what you see. Only very few stimuli or procedures qualify as aversive punishment because they have such a fleeting existence. When they do their job and eliminate undesirable behavior, they are no longer required and therefore, no longer used. Therefore, what we must be seeing are lots of aversive stimuli intended as punishment, but they are not working, i.e., aversive non-punishment.

When dog training was formalized and went on-leash in the early 1900s, leash corrections became the default, de rigeur catch-all correction for most problems. When corrections and punishments do their job and reduce and eliminate undesirable behavior, they are no longer required and so, are no longer used. Essentially, effective punishment eliminates the need for more punishment. But is this what we see? Even from casual observation, we see people repeatedly jerking the leash in class after class, and on walk after walk. Continued and excessive use of corrections advertises their ineffectiveness; they are not working as intended, i.e., they are not corrections.

When aversive stimuli do not rapidly reduce the frequency of undesirable behavior, leading to its eventual elimination, then they cannot reasonably be defined as 'punishment'. Instead, they are merely aversive.

Once we realize that 'aversive' and 'punishment' are not synonymous, if we define stimuli by their nature (aversive or not) and by their effect on behavior (effective as a punishment or not), we have:

Aversive Punishment: an aversive stimulus that works as punishment — a rare beast.

Aversive Non-Punishment: an aversive stimulus that doesn't work as punishment — an invasive species.

Non-Aversive Non-Punishment: Nagging without compliance.
...which of course, leaves us with the BIG question: Is there such a thing as:

Non-Aversive 'Punishment': A non-aversive stimulus or procedure than acts like a punishment by reducing and eliminating undesirable behavior?

Yup! In fact, the Jewel in the Crown of Barking Up the Right Tree is the sheer wealth of non-aversive techniques for reducing and eliminating undesirable behavior, including of course, teaching a dog ESL via lure-reward training from the outset, regularly testing comprehension and motivation, and monitoring the response-reliability percentages for every verbal cue. Providing clear verbal instruction prior to task and specific verbal guidance when dogs err is the very best way to prevent or resolve misbehavior and noncompliance, and thus, the perceived need for aversive punishment.

By teaching dogs our language, training transcends to astonishing heights. What was so complicated, cumbersome, and time-consuming, trying to communicate with whistles, clicks, and buzzes and with feedback limited to food, jerks, and shocks, becomes so simple, quick, effective, natural, and enjoyable. We talk to our dogs: give clear instructions and provide ongoing, analogue, instructive, verbal feedback — commonly known as PRAISE!

When developing more effective reward-training techniques, I think it is prudent to look at why most aversive punishments fail to work. To be effective and work as a punishment, aversive stimuli must meet six stringent criteria: Punishment must be 1. Immediate, 2. Consistent, and 3. Fit the 'crime'. Trainers must 4. Teach the meaning of Requests and Rules beforehand, 5. Warn prior to punishment, and 6. (The Biggie) A punishment must be INSTRUCTIVE. These six criteria are quite an ask and require an extremely high skill-set. Often, immediate and consistent administration are deal-breakers for so many owners, and for dogs, the lack of instructiveness of aversive stimuli is the doozy, especially since any instructiveness depends entirely on immediate and consistent application.

Prompt Reinforcement?
With reward-training immediacy is not such a big deal. For short-duration behaviors, I usually reward a puppydog as quickly as possible for its first few cued-Sits, Downs, Stands, Heel-Sits, Come-Sits, Recalls, and Rollovers but then I would purposefully delay any possible reward while I praise in 'good dogs': "Good dog one... good dog two... good dog three..." etc. and then maybe offer a piece of kibble. That's how we teach dogs that "Sit" means 'Sit-Stay-Watch'. Obviously with long-duration behaviors, praise and occasionally reward whenever you like.

Be Inconsistent!
The sheer beauty of all reward-training techniques is that we need not reward consistently, in fact, we should not reward consistently. Consistently delivering too many food rewards decreases their reinforcing power. Inconsistency, uncertain expectation, and hopeful anticipation all increase reinforcing power. I love it. Reward-Training: Consistency Not Required. Just count out the number of food rewards beforehand and reward at random with a bias towards rewarding the dog only for the better and best responses.

Praise is Rich
Rather than the same old leash-jerk or shock for isolated behaviors, with ongoing praise, we can instantly 'dial it up' or 'dial it down', several times a second to reflect the constant state of flux of behavior change. Praise is analogue; not only does praise effectively inform a dog that they got it 'right', praise also precisely informs a dog how well they are doing from moment to moment.  

Yes, I Allow My Dogs to Ignore Me Most of The Time
Teach verbal cues and rules from the outset, and differentially cue responses to signal their urgency and importance, i.e., whether your instruction is a 'suggestion' or a must-do 'command'. Use a nickname prior to suggestions, which the dog may follow or not, but use a formal name in those few instances when immediate compliance is expected and required. Works great! Dogs and their people can be off-duty, relax, and enjoy life together 99% of the time, and they (both dog and owner) only need to pay attention and temporarily be consistent for those very short periods when formal names are used.

Let's Just Get the Dog Back on Track!
The crippling failure of most aversive stimuli is that they fail to clearly communicate what we would like our dog to do instead. Any 'instructiveness' depends entirely on precise timing and absolute consistency. Moreover, even when aversive punishment effectively inhibits and eliminates undesirable behavior as intended, two points:

1. Do we really want to inhibit normal and natural doggy behaviors without first teaching dogs how to express their basic doggy urges acceptably and appropriately when living with people? For example, where to pee, what to chew, when and on whom to jump-up (hug), when and for how long to bark or 'let off steam' (a barkathon in the car, or a game of Jazz-up & Settle Down in the living room on a rainy day). When we see training from the dog's point of view and provide timely and appropriate outlets for their natural behaviors, most misbehavior vanishes. In fact, I just love it when we use cued behavior 'problems' as rewards to reinforce their cessation, e.g., by yo-yoing Sit/Hug, Shush/Woof, Come/Tag, Watch/Go Sniff, Settle Down/Jazz-up, etc.

2. Inhibiting behavior is only part of the puzzle: Equally as important: we want to get the dog back on track as quickly as possible; we want the dog to do it 'right'. So, we routinely ask the question If this is 'Wrong', what is 'Right'? and then inform the dog of our answer: Outside! Toilet! Chewtoy! Sit! Shush! Settle! Bed! etc.