Corrections Are Punishment (If They Work)

Correction is a term used in certain segments of the dog training world. It commonly applies to jerking the dog’s leash (also called a “leash correction). Sometimes “correction” refers to other physical things people might do to a dog.

Trainers who use corrections do such things when a dog is performing an undesirable behavior. For example, they will perform a “leash correction” when a dog is pulling on the leash, is in the wrong position, or is not focused on the handler. The magnitude of a leash correction can range from a twitch of the leash to jerking hard enough to lift the dog partially off the ground or knock him off balance.

stuffed dog wearing prong collar getting leash corrections

Feisty receiving a leash correction

Corrections are intended to decrease an undesirable behavior. You never hear anyone say, “My dog was doing a gorgeous job of heeling so I gave him a correction.” You won’t hear “My dog behaved perfectly when the guests were here so I gave her a correction.”

You might hear a trainer say they gave a correction “to get the dog’s attention.” That implies the dog was not paying attention. The trainer wants to decrease sniffing, pulling, fixating on squirrels, or whatever the dog was doing instead of paying attention.

Punishment

In behavior science, what do we call a learning process in which a behavior decreases? Punishment.

There are two types of punishment. One is the removal of something appetitive (desirable), as a consequence of a behavior. An example would be holding out a treat near the dog while he is supposed to be maintaining a position. If the dog moved out of position to get it, the handler would pull the treat back out of the dog’s reach. This would constitute punishment if the behavior of moving out of position decreased in the future.

The other type of punishment is the addition of an aversive stimulus as a consequence of  the dog’s behavior.  An aversive stimulus is something the dog will work to avoid if they can. An example of this type of punishment would be stepping on the dog’s back foot whenever he tried to jump on you, if in the future the behavior of jumping on you decreased. (I am not recommending this technique. There are humane and effective ways to train dogs not to jump on people.)

Positive punishment, as this latter type is called, need not be obviously harsh. It need only be effective. I used to have a dog who would leave the room when I coughed. She hated that sound. (Poor dog; I have asthma.) I could probably have used coughing as a punishing stimulus.

Another typical example of positive punishment is…jerking on the leash. A.k.a a leash correction. If your dog moves out of position (undesirable behavior) and you jerk on the leash (added stimulus) and moving out of position decreases, that behavior has been punished.

Defining Words to Suit One’s Purpose

Why then, does the segment of the dog training world that employs corrections often deny that they constitute punishment? You can read this myth in a thousand articles online. You can hear it from ten thousand trainers. And it will be stated in countless beginner obedience classes. Here is a quote from a well known and successful trainer:

The purpose of a correction is to get a behavior change. It’s not to punish a dog.

Wait, what? In behavior science, the definition of punishment is about behavior change.

The reasons for avoiding the use of the word “punishment” are pretty obvious. In a world where positive reinforcement-based training is becoming more well known, many potential clients would not like the idea that a trainer may hurt their dog or instruct them to do so. So trainers who do use positive punishment often call it something else.

Many also use the word “punishment,” but they use the lay definition of the word instead of the behavior science definition. That way they can claim not to use it.  The lay definition is associated with retribution and cruelty.  Many such trainers use the word “punishment” to mean doing something harsh to the dog in a fit of emotion, and state that they don’t believe in doing that. They reserve the word correction to mean something planned, deliberate, and by implication less severe. And they often call it “communication” or “getting the dog’s attention” or some other benign-sounding phrase.

Defining “Correction” This Way Is a Red Flag

People are free to use different definitions of words. Heck, the retribution definition of punishment is probably the first in the dictionary. But if you choose a trainer who uses corrections and claims they aren’t punishment, you can know that one of two things is true.

  1. This person is ignorant of the correct terminology used in the science of behavior, even though they are claiming professional expertise and taking money for changing behavior. Or:
  2. This person is using an idiosyncratic definition on purpose. Whatever their explanation is for this, the effect is to mask what they are actually doing to dogs. They are minimizing the fact that they are startling, hurting, or causing some kind of discomfort to the dog. They are avoiding transparency.
A brown and white stuffed dog iw being held forcefully on her back. You can see a woman's arms coming down and her hands are on the dog's belly and the underside of her neck, pushing hard.

Feisty being “alpha rolled.”

I realize there is a wide spectrum of aversive techniques for training dogs. Some are harsher than others. If I used positive punishment, I, too, would want to distinguish myself from those who used more violent methods than I did. But there’s a more honest solution to that problem. The solution is to be specific and transparent about what one does. Such a trainer could say, “I use punishment in the form of jerking on the collar or scruffing dogs, but I don’t ‘helicopter’ dogs, hit them with plastic bats, or kick them.”

But that last sentence demonstrates a good reason why people who use any aversives may not want to be specific. Even the mention of more violent ways to train dogs is going to be off-putting to much of the general public. How much easier and benign sounding it is to say, “I use corrections, but not punishment.”

When Corrections Aren’t Punishment

There are situations where “corrections” aren’t punishment. And that is when they don’t work to decrease behavior. And that does happen quite often.  So I suppose we could add a #3 above. If someone says that corrections aren’t punishment, it could be true if their methods don’t work to reduce behavior over time. That would also be a good reason not to let that person work with your dog.

I Use Punishment

I use negative punishment. This is the type where you remove something desirable when the dog performs an unwanted behavior. You do this with the intent of decreasing (punishing) the behavior. This method can be very effective in clarifying to a dog what behavior works and what behavior doesn’t work. But I try not to overuse it. There are some situations where it is unpleasant for the dog, even though it doesn’t employ an aversive stimulus. I’d rather train a strong behavior to begin with, using positive reinforcement, than be pulling away cookies, toys, or attention frequently. Also, with several common applications of negative punishment,  positive punishment can easily creep in. I do not knowingly perform positive punishment, nor do I ever design a training plan to include it.

I just applied the word punishment to my own training and explained the ways I might use it. So if I can use the word…how about you folks out there who use “corrections”?

Copyright 2018 Eileen Anderson

 

Share Button
This entry was posted in Operant conditioning, Punishment, Punishment culture and tagged , , . Bookmark the permalink.

15 Responses to Corrections Are Punishment (If They Work)

  1. peggieditmars says:

    Perfectly true, perfectly clear, perfectly helpful in that it’s written for laymen and ready to share to clients. There used to be a clothing store which had the great slogan: ‘An educated consumer is our best customer’, and that’s how I feel about my training business. Thank you for this; sharing it on my training page.

    • Eileen Anderson says:

      Thanks, Peggie! I usually write more for trainers, but I was aiming for the public this time. Glad you feel it’s shareable!

  2. BK says:

    Yeah, I think I’ll keep using the word “correction” because it’s less confusing for the “layman.” The word “punishment” had its vernacular sense for hundreds of years before the behaviorists decided to co-opt it, and it retains that traditional sense in the popular vernacular today (i.e. retribution for a perceived slight). I know how to use the behaviorist terminology, and if another dog trainer or animal expert wants to talk in that language I can do so, but I choose not to when I’m with clients because, first of all, I’m a dog trainer, not a scientist, and I don’t pretend to be anything but that, but secondly, I find that trying to navigate the dual meanings with someone who knows nothing about behavioral psych is pointless, so I use the word “correction” when I’m talking to clients because its vernacular sense is much more in line with what I’m trying to express. Training dogs is as much about communicating with canines as it is about communicating with humans, and since the dog doesn’t care what symbol I use to refer to my techniques, I use the symbols that are most likely to produce understanding in the humans.

    By the way, you’re wrong that all corrections are punishments in the behaviorist sense. In my heel work, any correction the dog receives is actually the aversive component of a negative reinforcement schedule. What do I mean? When I’m teaching a dog to heel, I literally just walk at a normal pace, exercising my God-given right to turn in any direction I desire. Based on the way that I hold the leash, my footwork when I’m making turns and where I have the dog when we start walking, there is one, and only one, possible location that a dog on a 6-foot lead can be and not either be walked-into or receive pressure on his collar when I make a turn, and that’s on my left side with his head parallel with and about a foot to the left of to my left knee. If he’s walking in this location, which is called “pleasant heaven,” the direction of the feeling of pressure on his collar caused by the weight of the slack lead is directly downward between his front paws. However, if he starts to forge, lag, or swing wide, the pressure-vector shifts, and he quickly learns to associate an imbalance in the pressure-vector with the corrections that inevitably result from my turns. He’s not being punished for moving away from me, he’s being reinforced for keeping that pressure-vector directly downward between his toes, the primary reinforcer being a slack lead in this case. The goal of our early heel work isn’t to teach him the behavior “watch your handler,” he’s merely learning to keep the weight of the lead attached to his collar directly between his legs, and because of my handling, there’s only one possible way he can do that, by remaining in “pleasant heaven.”

    • Eileen Anderson says:

      It’s interesting that you cite the “coopting” of the word punishment, when such coopting happens all the time as science develops. It would be hard to discuss electricity without the words “charge” or “circuit” or “current.” The electronics hobbyists I know all use them, so as to have a common vocabulary. (They aren’t all scientists.) All of those words predate the scientific discipline into which they were incorporated. Although granted, “punishment” may create a more unusual situation since the two meanings we are discussing can overlap.

      You didn’t mention whether you use the term “imbalance of pressure vector” with your clients, but sounds like you do. It doesn’t seem like a particularly simple term to use, nor an accurate one. “Application of pressure vector” would be more scientifically accurate, or simply, “application of pressure.” Using the word “imbalance” takes away the agent, like the pressure just “happens.” And before you say it–yes, the agent could be the dog, but you, the trainer have created the setup to make that happen.

      I don’t object so much to the word “correction” except when it is coupled with the denial of punishment, which is the point of the piece.

      I agree that with leash work, we are usually dealing with both negative reinforcement and punishment, a dual contingency process. The aversive stimulus of pressure on the leash can indeed be part of a negative reinforcement process. And in the avoidance mode of negative reinforcement particularly, the application of such an aversive can be very quick. But you yourself have distinguished between the change in the pressure and the “correction that results from your turn.” The process of negative reinforcement (application of pressure) will allow the dog to work to diminish the magnitude of the correction and eventually avoid it entirely. But the correction itself is still a noxious stimulus that comes after the behavior. I.e., it works as punishment or attempted punishment.

      Finally, the “pleasant heaven” you refer to is the kind of phrase I warn people against. Escape from an aversive stimulus is indeed a primary reinforcer. It’s a primary negative reinforcer, something that positive reinforcement based trainers don’t always learn about and would do well to keep in mind. But the feelings associated with escape from an aversive stimulus are more accurately described as relief or safety. You have created a little box using aversive stimuli and taught the dog that nothing bad happens as long as he stays in the box. Then you have named it “pleasant heaven.” Pleasure is associated with positive reinforcement, not negative. You have been polite in your comments here and argued in apparent good faith, but I have to point out that “pleasant heaven” is a disingenuous phrase.

      Thanks for your comment.

  3. David Todtman says:

    Thank your for the article and I agree with Peggie that the article clear. It lays matters out simply, and is helpful for me–a layperson.

    Some trainers–Donaldson, and maybe others–employ the verbal cue “unh unh” or “too bad” for unwanted behaviours. The cue is referred to as a “no reward marker” and is described, I think, as a communication to the dog that a reward will not be forthcoming for the behaviour that preceded the no reward marker.

    Technically, the cue “unh unh” is punishment since it is designed to reduce the probability of the unwanted behaviour. Right? Maybe, if the “unh unh” is delivered in a mild and pleasant way I am splitting hairs?

    • Eileen Anderson says:

      Hey David,

      Thanks for the kind words. You have stumbled onto one of the few topics I don’t discuss in my blog: so-called NRMs or no-reward markers. My own teacher just says to wait until they are listed by that name in a learning theory textbook (OK, someone is now going to come up with a mention, but that’s OK! It would help!) Her point is that people mean different things by them and that they are not officially codified.

      I don’t mean to discourage you from looking into them. Anything people do as part of training is worth analyzing. I just don’t write about them here because I don’t find it worth the ensuing fights that tend to happen.

      But I can give you a couple of resources. In the April 2009 issue of Clean Run agility magazine, available as a digital download, Pam Reid has an article about them. In the article, she describes the process of conditioning a negative punishment marker, just as some people condition a clicker sound. The conditioning she describes consists of literally having an array of treats before the dog, making the noise, and taking the treats away. Then using that sound later as part of training. It just makes you think. She also refers to the respondent version of “good thing isn’t coming,” Pavlov’s “conditioned inhibition.” That’s also worth looking into.

      I hope that gives you something to take a look at. As far as I know, her article is the only one so far that approaches it in a scholarly way, but I may be wrong. Susan Friedman and Ken Ramirez have been discussing the “LRS” or least reinforcing scenario as well. It is a different process, but since I mentioned scholarly approach I thought I’d better bring it in. It is slightly related.

      Good luck! And just a heads up to other commenters: I won’t publish further comments on NRMs or LRSs here. Thanks for understanding. (I know some of you are giggling out there. I just don’t have time!)

  4. Alison says:

    Yes! Absolutely.

  5. Iben Fals says:

    😁👍👍👍👍👍💕

  6. Shannon says:

    Hi Eileen!

    Great post, as usual.

    What are some examples of negative reinforcement which would occur without having positive punishment applied 1st?

    Thanks,
    Shannon

    • Eileen Anderson says:

      Good question! I have a whole post coming out about that someday, but am still doing some research. But I think my shower mold example is a pretty good one. The threat of shower mold motivates me to clean my shower. Cleaning the shower is negatively reinforced. I can think of some things that **could** be punished by shower mold, but does any behavior HAVE to be punished? I still got in the shower just as often. I still looked at the shower walls, in fact, I looked at them MORE. So as one example, when you have a gradual onset, duration aversive (for which the duration is a large part of the unpleasantness), I think you have the possibility of no P+. Think about what I might have been doing when I first perceived the mold. Can we show that it must have been punished?

      One other thing to keep in mind. It takes a stimulus of a certain magnitude to get positive punishment. Lots of times the behavior momentarily decreases, then increases back to its previous level. Positive punishment is not some automatic thing that happens because there is a noxious stimulus around. Whereas a very very mild stimulus, such as a strand of hair brushing into your face, can get behavior via negative reinforcement.

      Just some thoughts. I still have some reading to do.

  7. Suzanne says:

    As usual, your article clarified some key concepts for me. Thanks, Eileen! Always a pleasure to read and then runinate on your posts…

  8. Sarah says:

    I never use the term “punishment” to describe my actions in the context of dog training, or anything else for that matter. I do use words and phrases such as correction, feedback, communication, and getting attention.

    1. Dog training is not behaviourism. Behaviourism is an academic field that has relevance to dog training. I do not teach behaviourism. I teach dog training and my clients are not behaviourists, so I use everyday language that accurately represents what I am doing. One should use language appropriate to the audience.

    2. My use of the word “correction” accords with accepted usage in everyday language as well as the definition provided by the Oxford English Dictionary:

    “The action of correcting or setting right; substitution of what is right for what is erroneous in (a book, etc.); amendment. Hence, loosely, pointing out or marking of errors (in order to their removal)…”

    That last bit sounds like an NRM to me, so without commenting on NRMs themselves, I would point out that they meet the definition of “correction” above, there is disagreement as to whether they are positive punishers, and unless you can prove that their use is positive punishment, you cannot prove your point that those who say that a correction is not always “punishment” are wrong if they define NRMs as corrections.

    3. I use language to describe accurately what I am doing. I do not reject behaviourism – it is a useful tool among many – but I have issues with behaviourism and reject the term “punishment” as it is loaded, not objective.

    A. Behaviourism lays claim to being a science and to using the rigourous, objective standards demanded by science. If behaviourism were truly scientific and objective, it would not use such a loaded term as “punishment” when there are plenty of neutral terms that could have been chosen.

    B. The loaded term “punishment” maximizes the negative outcomes of an action. The seriously negative connotations of the word in everyday usage have been conflated with the meaning of the technical term, significantly skewing perception of the actions described by the term.

    C. Behaviorism lends itself to binary thinking and absolutism. Everything is positive or negative. Everything is a reward or a punishment. Everything is good or bad. This model, and the worldview it engenders, has no place for neutrality or ambiguity. Its misuse encourages extremism.

    D. Binary thinking and absolutism lead to lumping. All of the following fall into the P+ quadrant: a gentle tug on the leash; a hard tug on the leash; a stimulation from an e-collar that is mildly irritating; a stimulation from an e-collar that causes pain, saying “hey” loudly, prolonged yelling, stepping onto a dog’s feet; stepping into a dog’s space (social pressure, not physical contact), a tap on the shoulder or some part of the body, kicking a dog, hanging a dog, helicoptering a dog, beating a dog. Lumping real abuse with mild stressors is problematic. Lumping pain, discomfort, stress and fear together leads to any sign of stress being seen as an “animal welfare concern”. However, the research below suggests that exposure to mild stress and adversity has quite different effects to those of abuse.

    E. This last is not so much about behaviourism as the misuse of research. It has become popular among dog trainers generally to use peer-reviewed research to prove a point. First, that is not the purpose of a study. The purpose of a study is to further knowledge and the reason it is published is to open it to critique. Second, most of these dog trainers have not even read the studies they cite. At best, they may have read the abstract. There is no acknowledgement of the caveats. There is no understanding of biases such as confirmation bias, observer bias and selection bias. There is no understanding of the politics of academia and how they affect applications and funding for research and the peer-review system. There is no understanding of publication bias and issues such as politics and null results that create it. I am not accusing amateurs of deliberate misuse or deception – most dog trainers aren’t researchers or academics and therefore do not go looking for studies that would contradict the research they cite and have no reason to be cautious of the system that produces it. The resultant skewing may not be deliberate, but it is still problematic.

    You say, “The solution [to creating transparency] is to be specific and transparent about what one does.” Well, that’s exactly what I’m doing by specifying that I am using a correction, telling the dog he’s made a mistake, getting attention etc. as opposed to using loaded language that twists and exaggerates what I do.

    You end with this: “I do not knowingly perform positive punishment, nor do I ever design a training plan to include it.” That’s fine, and you are entitled to train your dogs how you feel best based on your knowledge and experience. But please note that there is a fairly large body of research that supports an alternative approach: the occasional judicious addition of stressors in a controlled environment and teaching the dog (or child) to work through them so that when the encounter stressors in the real world, the reaction is not anxiety, loss of impulse control and aggression, growing problems in the canine and human population, concomitant with the rise in the force-free approach to upbringing for both. I agree there are risks in this approach, but there are also risks to the force-free approach – see Seery below and what he has found about exposure to little or no trauma as opposed to some trauma. You also have every right to take the research I list with a pinch of salt. You should, given Point E above. But to be a truly critical thinker, you must give it due consideration and apply an equal pinch of salt to the research that reinforces your beliefs. Acknowledging that your opponent has valid points is standard practice in academia and does not imply a requirement to accept the totality of their position or change your own views, as your opponent will also happily acknowledge that you have a point too when you make an intelligent attack their argument as opposed to calling into questions their professionalism or intelligence because they disagree with you. Happy reading.

    Becirevic, A., Critchfield, T. S., & Reed, D. D. (2016). On the social acceptability of behavior-analytic terms: Crowdsourced comparisons of lay and technical language. The Behavior Analyst, 39(2), 305-317. doi:10.1007/s40614-016-0067-4

    Critchfield, T. S. (2014). Skeptic’s corner: Punishment — destructive force or valuable social “Adhesive”? Behavior Analysis in Practice, 7(1), 36-44. doi:10.1007/s40617-014-0005-4

    Dooley, L. N., Slavich, G. M., Moreno, P. I., & Bower, J. E. (2017). Strength through adversity: Moderate lifetime stress exposure is associated with psychological resilience in breast cancer survivors. Stress and Health, 33(5), 549-557. doi:10.1002/smi.2739

    Hanley, G. P., Piazza, C. C., Fisher, W. W., & Maglieri, K. A. (2005). On The Effectiveness Of And Preference For Punishment And Extinction Components Of Function-Based Interventions. Journal of Applied Behavior Analysis, 38(1), 51–65. http://doi.org/10.1901/jaba.2005.6-04
    Joober, Ridha, MD, PhD, Schmitz, N., PhD, Annable, L., Dipstat, & Boksa, P., PhD. (2012). Publication bias: What are the challenges and can they be overcome? Journal of Psychiatry and Neuroscience, 37(3), 149-152. doi:10.1503/jpn.120065
    Marsh, D. M., & Hanlon, T. J. (2004). Observer gender and observation bias in animal behaviour research: Experimental tests with red-backed salamanders. Animal Behaviour, 68(6), 1425-1433. doi:10.1016/j.anbehav.2004.02.017

    Sarkar, M., Fletcher, D., & Brown, D. J. (2015). What doesn’t kill me…: Adversity-related experiences are vital in the development of superior olympic performance. Journal of Science and Medicine in Sport, 18(4), 475.

    Schalke, E., Stichnoth, J., Ott, S., & Jones-Baade, R. (2007). Clinical signs caused by the use of electric training collars on dogs in everyday life situations. Applied Animal Behaviour Science, 105(4), 369-380. doi:10.1016/j.applanim.2006.11.002

    Seery, M. D., Holman, E. A., & Silver, R. C. (2010). Whatever does not kill us: Cumulative lifetime adversity, vulnerability, and resilience. Journal of Personality and Social Psychology, 99(6), 1025-1041. doi:10.1037/a0021344

    Shapero, B. G., Hamilton, J. L., Stange, J. P., Liu, R. T., Abramson, L. Y., & Alloy, L. B. (2015). Moderate childhood stress buffers against depressive response to proximal stressors: A multi-wave prospective study of early adolescents. Journal of Abnormal Child Psychology, 43(8), 1403-1413. doi:10.1007/s10802-015-0021-z

    Tuyttens, F. A. M., de Graaf, S., Heerkens, J. L. T., Jacobs, L., Nalon, E., Ott, S., . . . Ampe, B. (2014). Observer bias in animal behaviour research: Can we believe what we score, if we score what we believe? Animal Behaviour, 90, 273-280. doi:10.1016/j.anbehav.2014.02.007

    Tuyttens, F. A. M., Stadig, L., Heerkens, J. L. T., Van laer, E., Buijs, S., & Ampe, B. (2016). Opinion of applied ethologists on expectation bias, blinding observers and other debiasing techniques. Applied Animal Behaviour Science, 181, 27-33. doi:10.1016/j.applanim.2016.04.019

    van der Schot, Agnes A, & Phillips, C. (2013). Publication bias in animal welfare scientific literature. Journal of Agricultural and Environmental Ethics, 26(5), 945-958. doi:10.1007/s10806-012-9433-8

    • Eileen Anderson says:

      Part 1 of 2 (Eileen’s response split over two comments)

      Thanks for your thoughtful response.

      I have no problem with the word “correction” on its own. My specific beef is obfuscation.

      In case people don’t want to read my entire response, I’m stating an important point up front. This commenter’s reference list does not support her claims. It’s always easy to be impressed by a list of journal articles. And I applaud anyone who attempts to go to original sources. But in this case, the articles fail in basic ways to support her point of view. Those reasons are not subtle or complex. I’ll explain further after I address some of her other points.

      1. The good trainers I know use vocabulary appropriate to their clients and take the clients as far as they want to go with terminology. Barraging a layperson with jargon is not good teaching in any field, but teaching them terminology on a need-to-know basis is. The definitions of punishment and reinforcement in behavior science have meanings that are both little known the and important to the understanding of how organisms learn. It’s not a case of assigning a known word to another concept we are familiar with. It’s a case of teaching a concept that most people are unfamiliar with. (The commenter’s section 3D may demonstrate that lack of complete understanding of punishment.) I defined punishment in this blog post and others. Teaching the concept necessitates using the words that, however unfortunately, have come to represent them.

      2. Again, no problem with “correction.” I think perhaps I didn’t make it clear enough in my piece that it’s the combination of “correction-not-punishment” that virtually always demonstrates a lack of transparency. But if a correction is intended to decrease behavior, it is being applied as punishment. And it needs to be intrinsically unpleasant enough to do so. I wouldn’t particularly mind a trainer who didn’t use the word punishment—but bringing out the word “punishment” in order to deny it and lessen the impact of what they are doing is dishonest. That usage is rampant. If that shoe doesn’t fit this commenter—great. I can respect transparency even if I disagree strongly about methods. But since the commenter mentions the words correction, feedback, communication, and getting attention, I still don’t know what she says to clients who ask if any of these actions are painful or unpleasant for the dog. I also don’t know why a person who is transparent about their methods would want to ally themselves, through a defense, with people who make these glib denials, and whose claims are the target of my post.

      3A. The commenter says that if behaviorism “were really a science it wouldn’t use such a loaded term…” I would submit that that is not a good criterion for deciding whether a decades-old field of study is a science or not. Not that it’s even up to any of us to “decide.” She’s welcome to her opinion, of course. As for terminology: lots of sciences have confusing or paradoxical terminology. I could say that if chemistry were really a science they wouldn’t call sodium NA, when most other elements have handy, memorable abbreviations that are in accord with their words in English. I agree that the term punishment in behavior science is a particular problem. It is a much-discussed problem. But there are ways to straightforward ways to manage the problem without obfuscation, again, the topic of my post. And the terminology may change. It has changed before.

      3B. Interesting claim about the use of the word making people’s actions more negative, and not one I’ve heard before. But note: the problem of people responding to the term “punishment” by being harsh would only be a possibility if one is using positive punishment, the type involving aversive stimuli. Trainers who avoid that procedure don’t have to worry too much about “negative outcomes being maximized” when their students remove an appetitive in the case of negative punishment. One of the goals of positive reinforcement-based trainers, and I would hope all modern trainers, is to teach humans that rage responses have no part in training an animal. Even though there is a sad tradition of them in dog training.
      3C. The claim about behavior science being binary: good/bad, positive/negative, etc. is simply not true. It is a common belief of those who have not delved deeply into the science. (I’m not saying that is necessarily the case with this commenter, but it is a common correlation.) I refer readers to this post.
      D. I like it that this list is so inclusive, but it is not a list of positive punishers. This is a list of noxious stimuli people do to dogs, some of which could possibly decrease behavior contingently. There is no such thing as a list of positive punishers. This, too, is often a mistake of someone who has not delved deeply into the science. Nothing fits into a “quadrant” unless we know there has been behavior change. The contingent processes of operant learning are observable, and behavior change is built into their definitions. Without observing outcomes we stay in fantasyland.

      Certainly we sometimes use the shorthand that a certain thing falls into a “quadrant.” But this commenter is making claims about the science of behavior and its extent and attempting to do so in a scholarly way. Is leaving out any mentions of outcomes in a list of “punishment” shorthand, or is this a lack of knowledge of the science? It’s a dangerous shortcut when having a serious conversation. Also, there’s another problem with the list. Positive punishment is not all that easy to achieve. The weaker the stimulus, the less likely it will even work to punish, so including the weaker stimuli to make some kind of point about the wide spectrum of punishment is misleading. Many of them will not act as punishers at all. If we are discussing the technical meaning of punishment, we need to use it correctly.

      As I said in my piece, if I used P+, I would want to make these distinctions. I understand the desire to do so. But drawing a line between “real abuse” and “mild stressors” is impossible because it will always depend on the dog. It makes more sense to group them by possible function, and whether they work to achieve behavior change.

      Note the last sentence about the “research below” is not true. Again, the research cited doesn’t support the statement. See 3E and also what I wrote about the references below.

      3E. Regarding misuse of research. I agree that “citing a study” or cherry picking articles is a misuse and misunderstanding of research. I have written about that myself. Science is ongoing. That’s why I cite textbooks and review articles. The bulk of the research over decades supports any claim I make in this blog. The claims I make are basic. If my error or further evidence shows them to be wrong, I publish a retraction or an amendment. More on this below.

      Regarding the References

      A note to all readers: it’s impressive and can be very helpful when someone posts a lot of sources. But you need to assess them to be able to determine whether they actually support the person’s position. I took the time to do that because this commenter took time with her own comments, which were polite, considered, and substantial.

      I read everything seriously. I was familiar with some of those, but not all. I looked at them all. Had to order a couple. They do not support the commenter’s position, with one possible exception, a small point about terminology.

      On Bias: The articles about observer and publication bias in studies are great, in and of themselves. We need to know these things. I am all for improving our experimental techniques and fighting bias in ourselves in both professional and personal settings. And it’s also great that those are being published. It’s an example of the self-correcting nature of science. I read a ton about biases. Studying behavior science helps me fight biases.

      But these articles about bias do not discount the overwhelming bulk of the classic literature about the negative effects of using aversive stimuli. Those are not modern studies. They have been going on (and been replicated) since at least the 1940s. Over time, the effects of publication bias or factionalism have been ironed out. Plus, in these studies, nobody was trying to sell or demonize a training technique. Positive punishment and negative reinforcement have been studied for 70 years. And the literature consistently tells us that the risks of ill effects are high. These effects are not usually subtle. This is why I cite textbooks and review articles in which an expert has already combed and reviewed the literature. And my claims about the effects of aversives are from those textbooks and review articles. They are from the bulk of the literature. Not individual studies here and there.

      < >

      • Eileen Anderson says:

        Part 2 of 2 (Eileen’s response to Sarah split over two comments)

        On Terminology: The Becirevic article about terminology is excellent and I’m glad the commenter provided it. I don’t know any good trainer who just uses jargon on people cold turkey and expects them to understand. They teach the concepts and use terminology gradually, not snowing people under. They shape the usage. The people’s responses as recorded in this study are good to know about, but I don’t see them as presenting insurmountable problems.

        On Stress: Perhaps not all readers know that the claim about stress is a fairly common argument used by trainers who use aversives. The argument has three parts:

        1. Stress is good for dogs in that it makes them more resilient.
        2. Using aversives in training serves that function and will create that resilience.
        3. Dogs trained without aversives or with exceedingly rare ones don’t develop coping skills in life.

        To attempt to support this argument, the commenter has provided us with 6 studies showing positive effects of stress (or in one case, of punishment). But they are all studies of humans. So first, we have to consider whether this is an area where we can assume humans and dogs have similar responses. Given that dogs are believed, at best, to have the cognitive abilities of human toddlers, the answer is no. Studies about Olympic athletes and breast cancer survivors (to cite the most extreme examples) benefiting from previous stressful life experiences depend on years of experiences, living their consequences, and reflecting on them. They depend on human cognition, as do all of the included articles about stress.

        These articles on stress don’t offer any evidence about dogs and stress or dogs and punishment.

        But there’s more. It’s not true that we R+-centered trainers believe in protecting our dogs from all stress. Life presents plenty of stressors, and we use evidence-based techniques to help the dogs learn to deal with them. And we prepare them for the kinds of stressors they may actually encounter: relevant ones. Hence, we expose them to as much of the human world as we can so they can build positive associations and coping skills. Vet visits. Medical procedures. I help my dogs get used to sudden noises or me dropping things. Noisy, chaotic environments.

        Building coping skills by controlled exposures has good science behind it. So I dispute utterly the commenter’s claim to have presented evidence of “risks to the force free approach.” There are evidence-based ways to help dogs build resilience. Using aversives in training is a red herring.

        The sweet irony of all this is that building resilience in many ways consists of habituation to slightly unpleasant stimuli. So if, for example, I desensitize my dog to having a chain thrown on the ground next to her, I have removed its power as a punisher, should my dog have the misfortune to fall into the hands of someone who tried that. Building true resilience against a noxious stimulus makes that stimulus ineffective. So using aversives in training with a goal of resilience could easily build habituation and render one’s own techniques ineffective. There’s strong evidence that habituation is the common response to mildly noxious stimuli.

        I wrote a post recently about my dog who died during surgery a year ago. I wrote of my regret that I didn’t do more conditioning to vet visits, because her last hours were spent waiting in a cage at the vet’s. I did some conditioning, but probably not enough. At no time did I think, “Whew, I’m sure glad I trained her with a prong collar and with jerks on her collar before I crossed over, because that prepared her for the adversity of this stressor.” I’m not trying for pathos, here. I’m pointing out the obvious: to prepare a dog for stressors you practice with those stressors. What would have helped would have been more controlled exposures and vet visits with positive or neutral associations.

        I appreciate the time this commenter took to write. I learned something and the comment allowed me to make some really important points to my readers.

        I hope that when she and other trainers discuss corrections, etc with their clients they tell them honestly how they are intended to work and what effects they may have on the dog, whether or not they use the word “punishment.”

Comments are closed.