In January 2022, the dog trainer Ivan Balabanov emailed me to invite me on his podcast. I knew little about him at the time except that he was world famous in protection sports.
I declined. I’m a writer, not a trainer. I don’t think well on my feet in conversation. I wouldn’t be a good representative for the positive reinforcement training community, and that’s what I would be there for.
I had no idea of the bullet I dodged.
I observed Mr. Balabanov’s outreach to the positive reinforcement-based training community after that. And in February 2023, he published a podcast episode titled, “The Real Facts about Science Based Dog Training.”
I have thought hard, for more than a year, about whether to give this podcast any oxygen by responding to it. But now it’s pertinent to current events in the dog world. It’s important to pull back the curtain.
The “Real Facts” Podcast Episode
In this podcast episode, Mr. Balabanov employed many rhetorical fallacies. Primary among them, he did what is called a Gish Gallop. Here’s a definition:
The Gish Gallop is the fallacious debate tactic of drowning your opponent in a flood of individually weak arguments in order to prevent rebuttal of the whole argument collection without great effort.
A Gish Galloper spews out rapid-fire arguments of varying quality, from false, to unverifiable, to half-truths, and usually some legitimate points thrown in. The problem is that their opponent would have to take far more time and labor to untangle the mess than it takes for it to be thrown out there.
Between verbal mentions and citations flashed on screen, Mr. Balabanov cited about 50 books or studies by my count in a 65-minute podcast.
Some of the opinions Mr. Balabanov tried to persuade listeners of were:
- The AVSAB position statement on humane training is terribly wrong;
- Positive punishment (specifically shock) is necessary sometimes and not only not harmful, but has benefits;
- There is a ton of science to support his stance; and
- “Force-free” trainers and veterinarians are dogmatic, ill-informed, and cherry-pick the science.
In addition to the Gish Gallop, he employed straw men, the naturalistic fallacy, and ad hominem attacks on groups and one named individual.
I seek to adhere to the rules of fair debate in this post. So there won’t be any colorful language or even what most people think of as passionate writing. But this is a passion project for me. Gish Gallops can be very persuasive. The speaker sounds super knowledgeable to people who aren’t familiar with the technique or don’t know the subject. All those references!
Over 100,000 people have viewed the YouTube video, and thousands more on other platforms, I’m sure. I can’t reach those people directly, but I want an evidence-based response to the podcast episode to exist and be accessible.
How to Respond to a Gish Gallop
When a debater Gallops, it puts the person on the other side in the position of having far too much material to refute. This is why some points can be and often are total bullshit. You won’t have time to get to them all.
When confronted with a Gish Gallop in debate, the standard advice is to do two things:
- Point out your opponent’s use of the technique.
- Pick one claim and address it thoroughly, pointing out the flaws in the argument.
I am going to do a variant of this response, since I have a little more time than a debater. I’ll address a short selection of the fallacious points.
Here we go.
Arguments and Citations
There is no list of references in the notes for the episode, as should be included for a talk citing research. (Another trainer made one and posted it on their own website.)
The Episode Title
The title of the episode itself indicates we are not about to hear a scientific approach. Science is about evidence. No one can claim knowledge of the “real facts” of science-based dog training, much less cover them in an hour. Given the content, an expert in the field might have titled such a lecture “Some Evidence to Support the Use of Aversives in Dog Training.” But they also would have picked one or two references and presented them in context. They wouldn’t have packed dozens of studies, names, and opinions into an hour. It takes a lot of time and words to cover the results of even one study properly, because it needs to be in the context of the whole literature. This includes preceding studies, any later replications, and those with opposing findings.
Punished by Rewards
The very first reference presented set the tone. The Gallop was on. Mr. Balabanov said, after offering it as a reference: “There is a very well-written book, Punished by Rewards. It discusses some of the problems with positive reinforcement.”
That’s all he said about it.
I’ve read this book (Kohn, 2018) and it’s on my shelf. But it’s far from relevant to the claims in the episode. The title has the effect, though, of getting those words—punished by rewards—coupled in our heads.
The author, Alfie Kohn, despises behaviorism. He is an odd person for Mr. Balabanov to cite. Mr. Balabanov uses operant conditioning, and in his own words from the same episode is “a big advocate of positive reinforcement.” He also cites many articles by behavior analysts in the episode.
Punished by Rewards is about using rewards with children. A major focus is that Kohn claims extrinsic rewards destroy intrinsic motivation. The evidence has moved on from this stance; the topic is much more nuanced. But training dogs is much simpler. Extrinsic vs. intrinsic motivation is a minor issue, when it’s an issue at all. We understand that many of the things we ask pet dogs to do are not intrinsically motivating, so we make it worth their while. The book is irrelevant to dog training.
Mr. Balabanov spoke 18 words about the book in about five seconds, including nothing about its content or relevance. I wrote several paragraphs and barely scratched the surface. I didn’t even make a synopsis of the book; I only pointed out reasons the book doesn’t support Mr. Balabanov’s arguments. That’s the burden a Gish Gallop puts on its recipient. And neither of us did the subject justice.
The next two items are on the topic of comparing negative and positive punishment.
The “Just Think” Study
Mr. Balabanov quoted a study called “Just think: The Challenges of the Disengaged Mind” (Wilson et al., 2014). This was to support his claim that negative punishment can be “just as harsh or abusive [an] approach” as positive punishment. But there were neither negative nor positive punishment contingencies in the study. The study found that humans who were put into a room for a set time period with nothing to do but think or shock themselves often did the latter, even though they said before the experiment that they would pay to avoid the shock. That humans would choose to try a shock generator under their control when asked to be alone with their thoughts does not provide a comparison of negative punishment and positive punishment. There was no contingency on the shock, and the “timeout” was not a consequence for anything except signing up for the study. And leaving the room was likely an option considering the standard requirements for human study. I recommend reading the study, and particularly the next studies in that line of research, but just because they are interesting. Just realize that they have little to nothing to do with dog training.
Had I been in the study, I’m sure I would have explored the shock. I did that with our livestock electric fence as a kid, seeing how short a weed stem I could use to touch the fence and still tolerate the shock. I wasn’t trapped with nothing else to do. Humans are curious. A human shocking themselves a few times in a quiet empty room has no comparison with a dog being shocked contingent on their behavior, by a human, via an inescapable collar. Nor does a person joining a research study where they will be in a boring room for a few minutes have much in common with being put in a timeout contingent on a behavior (and controlled by a trainer).
Timeouts bear careful consideration. It’s not news that they can be aversive, so Mr. Balabanov’s remarks lean heavily on a straw man. Many force free trainers don’t use timeouts. Methods that rely on them are being replaced by better ones.
The “Quitting Signal” Study
This odd study is a favorite of defenders of shock and prong collars. Mr. Balabanov presents it to support a very general statement: “This suggests that negative punishment may be more stressful for dogs than other forms of punishment.”
I read the dissertation related to this study soon after it came out and got translated (Salgirli, 2008). I read the spinoff study when it was published in a journal (Salgirli et al., 2012). I’ve had a blog post about it in the works for years. In the latter study, it was found that dogs had higher cortisol levels after training that involved “negative punishment” (more on those scare quotes coming up) than positive punishment via shock or prong. A big problem with how the study is presented is that positive punishment wasn’t compared with negative punishment, but with a negative punishment marker, a conditioned punisher.
From the study:
Corrections made by pinch collar and electronic training collar were considered as representatives of the positive punishment while correction made by the quitting signal was considered as the application of the negative punishment.
Salgirli et al., 2012, p. 531
There was no consequence paired with the quitting signal, no withdrawal of the appetitive during the actual experiment. A negative punishment marker (encountered in an environment where it wasn’t trained and with a novel stimulus) shouldn’t be equated with negative punishment.
There are also problems with the training methodology, assuming it was what was described in the dissertation. There is insufficient detail in the published paper itself to allow replication, and oddly, the dissertation is not in the references.
But let’s zoom out a little. Set aside my remarks about the quality of the study. It’s not news to positive reinforcement-based trainers that negative punishment can be frustrating and stressful. Could there be a study that validly found that in a certain situation, negative punishment caused more stress than collar corrections to some dogs, most of whom were accustomed to them? It’s possible. Individual dogs react differently. But even if that study existed, it wouldn’t prove Mr. Balabanov’s general claim.
That’s because you can’t hang your hat on one study to “prove” an argument, or two if we count the previous one that had no contingencies. This isn’t a scientific approach. No matter how much we want studies that give firm evidence for our beliefs, what we need to pay attention to is the bulk of the amassed literature, the consensus of the experts.
That is what’s missing from the podcast episode.
Jack Michael’s 1975 Study
Mr. Balabanov mentions in passing, in an argument about the AVSAB statement, “…the 1975 study done by Michael, which says that every reinforcement includes both positive and negative form…”
No. That isn’t what that study says (Michael, 1975). It is a favorite for defenders of aversives to trot out. And I don’t have to explain what’s wrong with their argument in this post, because I wrote a whole post about it.
Positive and Negative Reinforcement by Jack Michael: A Misconstrued Article
At the end of the article, Michael concludes his exploration of the nomenclature by saying that we need a better way to describe the differences between positive and negative reinforcement, not that there are no differences. After asking whether we need the distinction, he says, “We need to make the distinction in order to have a name for the bad things in our world” (Michael, 1975, p. 43).
During the time Mr. Balabanov speaks of the Michael study, he shows on screen instead the Baron and Galizio study (2005). This paper does discuss a possible overlap between positive and negative reinforcement, and there were a few more papers in this vein that followed. But while these papers are mentioned in some textbooks, they still comprise a minority opinion. The familiar nomenclature and separation of positive and negative reinforcement are still the standard.
Benefits of Positive Punishment
Mr. Balabanov said:
“…studies show that the effectiveness of positive punishment in reducing problem behavior tends to be associated with a wealth of positive side effects. The positive side effects tend to outnumber any negative side effects associated with positive punishment.”
He cited seven studies on screen during the 15 seconds it took for him to make these statements. Most were from the 1990s; the most recent was from 2013.
I chose one claim to investigate, the one about the positive side effects outnumbering the negative side effects. It’s true that the studies he cited listed positive side effects of positive punishment or stated that there were more positive side effects than negative. One was a review study, although from clear back in 1989 (Matson & Taras).
I consulted more contemporary sources. I looked in six behavior analysis textbooks, all of which were at least a decade more recent than the review study. Behavior Analysis for Lasting Change had the most material on this topic (Mayer et al., 2019, p. 691–3). There were three pages on benefits of punishment, although they had caveats. Seven pages of undesirable effects followed (Mayer et al., 2019, p. 693–700). In the “benefits” section, the authors cited several of the same studies about the benefits of punishment (including the review) that Mr. Balabanov referenced. But the textbook included many other studies with opposite findings and didn’t come to the same conclusions. The authors opened the “Disadvantages of Punishment” section with, “If punishment works rapidly to reduce the rate of a behavior, why not use it as the first line of defense against unwanted behavior?” After describing corporal punishment statistics in the United States, they continue: “As you read about punishment’s disadvantages, though, you will begin to understand the information that has been causing those numbers to diminish slowly and steadily since the early 1980s” (Mayer et al., 2019, p. 693). Then they thoroughly describe 12 categories of disadvantages.
You might think I cherry-picked the textbook. But no. Except for a brief mention in Chance (2003, p. 205) at the beginning of the section on problems of punishment, the five others didn’t have sections on benefits of positive punishment at all.
We need to assess the bulk of the literature, and most of us, me included, are not equipped to do that. Textbooks are written by discipline experts and distill a vast mass of knowledge into one book. These experts, along with other behavior analysts, applied animal behaviorists, veterinary behaviorists, and people with graduate degrees in ethology and animal behavior are the subject experts.
They are in consensus about punishment. They consider the entirety of the literature, and disagree with Mr. Balabanov.
Assessing Research
I do my research, a lot of it. I’ve done a formal literature review for a master’s thesis. I distilled hundreds of papers into the handful pertinent to our experiment, critiqued them, and wrote about their relevance to my research. I’ve taken a course in assessing research in behavior. But my graduate degrees are in music and engineering, not behavior science. As much as I study, I will not have the in-depth understanding of the behavior science or ethology literature as people with advanced formal study in these disciplines. When I write about research, such as in my piece about the Jack Michael article, I run it by experts.
If you want examples of responsible reporting about research from people with better credentials than I have, Linda Case of The Science Dog and Zazie Todd of Companion Animal Psychology both do a great job. (Please do not assume they have anything to do with this post, which is entirely my creation.)
And read textbooks. Read the pages and pages of warnings, cautions, and caveats about using positive punishment that result from decades of research, collected by experts in the field.
And here’s an article of mine on how not to get caught in the “a study says” embarrassment.
Final Words: Stepping Away from Debate Guidelines and onto a Soapbox
Building bridges and helping trainers cross over have been hot topics on social media lately. I benefitted from people extending a hand to me, and I have extended a hand to others. This is best done one-on-one. I’ve observed that it’s usually most effective via a personal relationship, or it may (I hope) sometimes be via someone writing and communicating to readers. It seems unlikely that a panel discussion of people with mixed ideologies (as is scheduled soon and includes Mr. Balabanov) would cause an epiphany in someone’s thinking. Letting go of our cultural punishment mindset is hard.
I haven’t been invited to any such panel and I don’t expect to be. But hearing this Gish Gallop, hearing Mr. Balabanov’s savage ad hominem attacks and other bad faith arguments, and his low regard for his imagined debate opponents (in this case force free trainers, veterinarians, and veterinary behaviorists), made it absolutely clear to me that this is not someone who will argue in good faith. I don’t call myself a force free trainer, but they are my people (if they’ll have me)! I check all the boxes, and then some, in terms of how I train and live with my dogs. I see no benefit and lots of problems attendant to sitting down with someone who is so willing to use unsavory debate tactics and speaks of my colleagues with disdain. It would be a betrayal. There is no bridge there.
I made my decision in 2022 not to join Mr. Balabanov on instinct and a little luck. But now I get the complete picture. In the unlikely event I am ever invited again to a discussion including Mr. Balabanov, I will again decline. And that’s what I recommend to others in my community.
References
Baron, A., & Galizio, M. (2006). The distinction between positive and negative reinforcement: Use with care. The Behavior Analyst, 29, 141-151.
Bouton, M. E. (2018). Learning and behavior: A contemporary synthesis. Second edition. Oxford University Press.
Chance, P., & Krause, M. A. (2003). Learning and behavior. Thomson/Wadsworth.
Kohn, A. (2018). Punished by rewards: The trouble with gold stars, incentive plans, A’s, praise, and other bribes.
Matson, J. L., & Taras, M. E. (1989). A 20 year review of punishment and alternative methods to treat problem behaviors in developmentally delayed persons. Research in developmental disabilities, 10(1), 85-104.
Mayer, G. R., Sulzer-Azaroff, B., & Wallace, M. (2019). Behavior analysis for lasting change. Sloan Pub..
Michael, J. (1975). Positive and negative reinforcement, a distinction that is no longer necessary; or a better way to talk about bad things. Behaviorism, 3(1), 33-44.
Miltenberger, R. G. (2008). Behavior modification: Principles and procedures. Fourth edition. Wadsworth.
Pierce, W. D., & Cheney, C. D. (2008). Behavior analysis and learning. Psychology Press.
Salgirli, Y. (2008). Comparison of stress and learning effects of three different training methods: Electronic training collar, pinch collar and quitting signal (Doctoral dissertation, Hannover, Tierärztliche Hochsch., Diss., 2008).
Salgirli, Y., Schalke, E., Boehm, I., & Hackbarth, H. (2012). Comparison of learning effects and stress between 3 different training methods (electronic training collar, pinch collar and quitting signal) in Belgian Malinois Police Dogs. Rev Méd Vét, 163(11), 530-535.
Schwartz, B., Wasserman, E. A., Robbins S. J. (2002). Psychology of learning and behavior. WW Norton & Co.
Wilson, T. D., Reinhard, D. A., Westgate, E. C., Gilbert, D. T., Ellerbeck, N., Hahn, C., Brown, C., & Shaked, A. (2014). Just think: The challenges of the disengaged mind. Science, 345(6192), 75-77.
Related Posts
