How Did The Aversive Get There? A Call for Honesty

I am mystified by one particular argument of those who use protocols for fearful or reactive dogs other than desensitization/counterconditioning (DS/CC). These other protocols often use negative reinforcement; if not that, then sometimes desensitization without counterconditioning; sometimes extinction; sometimes habituation.

People who practice these protocols intentionally expose their dogs to their triggers at an aversive level at times, as opposed to people who practice pure DS/CC, which is ideally practiced at a distance or intensity such that the trigger is not aversive to the animal.

The argument that bothers me is this:

It’s OK to expose the animal to a trigger at a potentially aversive level as long as we are not the ones who put the aversive there for them to be exposed to. We’re not adding an aversive; it’s already there.

I wrote a post a while back addressing this idea in part. I pointed out that for negative reinforcement protocols, the ethical and definitional difference is not about how the aversive got there. To say so is to invoke the naturalistic fallacy.  The ethical difference rides on whether the trainer chooses to put a contingency on the animal getting away from it, not whether the aversive is “natural.” Do they ask for or wait for a certain behavior before retreating? Because that is a choice. If the dog gets close enough to the trigger that she starts showing stress, there is always the option of getting her humanely out of there, with no requirements on her behavior from the handler.

Where the aversive came from is ethically irrelevant, since the trainer makes a choice whether or not to use it, however it got there. Most would agree that such a use is an ethical choice, to be carefully considered.

So the fact that people are still mentioning this irrelevancy about “who put it there” seems like a lot of hand waving to shoo away the real issue: choosing to use an aversive.

But wait–in case it matters–how did it get there?

How It Really Got There

My hand, my voice, my phone.

My hand, my voice, my phone.

I have a formerly feral dog with whom I have been working for a few years, gradually getting her socialized to people, and making lovely progress with DS/CC.

Even though my goal is to keep the triggers (people, in her case) under the threshold of aversiveness, I realize that I am dealing with potentially aversive situations when we go out into the world. And I arrange for and seek out those situations for her sessions. For instance, I make phone calls at times to arrange for a controlled session with a person unknown or partially known to her.

If I do this and blow it and let her get too close or stay too long, I have exposed her to an aversive. How’d it get there? Me! Entirely through my choices! I arranged it. I deliberately sought it out with her. I made the phone call, drove my dog to the meeting place, and exposed her to the trigger. I added it to her environment, or added her to an environment where I knew it to be.*

People following any protocol generally arrange for triggers to be present in this way, including people, dogs, specific things like people on bikes or scooters, or other animals. So if someone is doing any type of exposure treatment, how can they claim that they are not responsible for the aversive being there? Did the Tooth Fairy bring it? Can their dog pick up the phone and drive the car?

It is not logical to claim to have nothing to do with the aversive being in the environment if you planned it, arranged for it, or sought it out in the first place. And that includes stealth sessions. If you are out there looking for triggers to use without their knowledge, you are still the one choosing to expose your dog to them. Finding = adding.

Empathy

You can probably detect that I find this irritating, but I seek to look at it in an empathetic way.

I have been reading some posts by Behavior Adjustment Training (BAT) practitioners in particular who express that they feel attacked and beleaguered by questions about negative reinforcement and humane training attached to their protocol. I get that they feel pushed into a corner.

I can empathize with that. Here is something you believe in, and people are asking difficult, pointed questions about it. Sure, anybody would be defensive. As a blogger, I have to deal with all levels of criticism. Even the most reasonable of criticism hurts.

There are people who react to these questions with dignity, though. They say yes, they are using negative reinforcement at times if they use certain protocols. They have thought it out, see good results, usually use other protocols as well, and are ultra careful about side effects. They don’t play like the presence of the aversive has nothing to do with them. Although I may not agree about all methods these folks use, I can appreciate their transparency and honesty about the science.

But it really worries me that there are still people who claim not to be responsible for getting the aversive into the environment. If they are trying to elude responsibility for that, even though it’s completely a side issue, what else are they willing to overlook, justify, or push out of their minds?

Thank you to all the people who do their best not to adjust the science (or even basic logical thinking) to justify their own preferences.

Coming Up:

  • The Girl with the Paper Hat Part 2: The Matching Law
  • Punishment is not a Feeling
  • Why Counterconditioning Didn’t “Work”
  • How Skilled are You at Ignoring? (Extinction Part 2)
  • What if Respondent Learning Didn’t Work?

Eileenanddogs on YouTube

* I am not confusing positive punishment and negative reinforcement, here. To use negative reinforcement, there has to be an aversive in the environment to be removed or escaped. We’re talking about how the aversive got there in the first place.

5/25/14 Addendum

This post is an urge to be honest about one aspect of the use of aversives.  I believe that all trainers, regardless of method, should be honest about their training choices and philosophy. You do it: own it. That’s the message in a nutshell. And I directed it to an argument that I believe does the opposite of “owning it.”

However, one of the common responses I have gotten over the past week  is comparisons of the ranges and setups of DS/CC protocols and those using negative reinforcement, often in an apparent attempt to minimize the differences.

I have previously provided a webinar and a movie on the differences and similarities of the major protocols for addressing fear in animals, with particular emphasis on their ranges and setups.

To review a few relevant points: Debating who starts further from the stimulus is a moot point.  No matter how far away you start, you are required to go into the aversive range for a negative reinforcement protocol to work.  In desensitization and counterconditioning you have no need to cross into the range of stimulus aversiveness in order to get effective results. In R-, aversive exposure is necessary. The protocol depends on it. In DS/CC, aversive exposure is by accident and hopefully rare. That is an important distinction between DS/CC and negative reinforcement-based protocols.

The other important distinction is that you can get a positive conditioned emotional response from DS/CC. With DS/CC and negative reinforcement there are two very different types of learning going on.

 

 

Share Button

About eileenanddogs

Passionate amateur dog trainer, writer, and learning theory geek. Eileen Anderson on Google+
This entry was posted in Desensitization and Counterconditioning, Escape/Avoidance, Extinction, Fear, Negative Reinforcement, Terminology and tagged , , , , . Bookmark the permalink.

49 Responses to How Did The Aversive Get There? A Call for Honesty

  1. Gen bergeron says:

    I also found it unbelievable that some people actually argue that -R doesn’t always have to involve an aversive.

  2. clickerdog4 says:

    It sounds like you are describing the previous version of BAT (1.0), which is very different from the newer 2.0 version. It seems like beating a dead horse when BAT is still described as negative reinforcement, with the retreat to safety as being contingent on desired behaviors. It’s no longer that way. Your blog would be a good vehicle to help BAT critics get up to date. Will you be attending the upcoming BAT for geeks online seminar? It ought to be interesting- I’m looking forward to what Grisha has to say. 🙂

    Naomi

    • Hi Naomi! No, I’m not describing only the previous version of BAT, although all of it is still approved as part of BAT 2.0 per the webinar I attended in February and several other communications I have seen. (The webinar also specifically included that BAT 2.0 could include negative reinforcement, either solely through the actions of the dog, or mediated by the handler as before in BAT Stage 3. It may also include some of the other processes I mentioned at the beginning of the post.)

      This post is not focused on negative reinforcement, though. It’s about aversives, and we can have exposure to aversives with or without handler mediated negative reinforcement. The post is about taking responsibility for bringing aversives into the environment, and/or using them for training. I wouldn’t have written it if I weren’t still seeing the remark I quoted about how it’s OK to use them if we didn’t add them. I would love it if our whole community would get up to date on that. Thanks for the comment, Naomi.

      • Ingrid Bock says:

        You go, Eileen! This is another high value treat for us. I’ve already found several occasions to share it, and was pleased to see Reisner sharing it, too. You rock and roll.

        • Thanks Ingrid! I know it’s hard for you to get your comments on here, so big thanks from me.

          • A note to those with pending comments: I had quite the influx today and I have another writing project in the works. So Mel, Sarah, Paul, and Linda: your comments have been received. I will restart the discussion in a few days. In the meantime I may send through a few that are responses between other commenters, but the ones I listed above, and probably any new ones, will need to be on hold for a little while. Thanks for understanding.

  3. Reblogged this on pawsforpraise and commented:
    The always thought provoking Eileen does it again.

  4. Only when we, as a profession, hold ourselves to scrupulous standards of competency, ethics, and science, will we attain excellence. Thanks for holding our collective feet to the fire with your ever-examining and questioning blog posts.

  5. Paul McGee says:

    Given that both (say) BAT and CC/SD (may) require a setup situation with a helper to enact the protocol … what is the real-world difference between these approaches? If we have a “scary” dog held by a friend on an otherwise empty park … what do you envisage as the major differences in say distance approached, time in “exposure”, number of trials, breaks taken, etc? (for a hypothetical case)

    • Hi Paul, well you know what I’m going to say, don’t you? <>

      Distance: laid out in my webinar (and posts) about threshold. Distance/intensity from the trigger is the big difference between the protocols.

      Time in exposure, number of trials, breaks taken are all things that both need to be planned in a general way, but also need to be able to change on the fly in response to the dog in any protocol.

      Thanks for checking in, Paul.

    • awesomedogs says:

      Bit like comparing applies to oranges. The question might better be, “What is the expected end result?”
      With R-, you don’t get a +CER to the scary thing. It forms to the warning. The scary thing stays scary.
      With DS/CC, you’re working to create a +CER directly to the scary thing.

      Things such as the number of exposures or time are less important if you realize that the end results are very different between those two processes.

      I’m not going to comment on BAT because it seems like everyone has their own version that they have created. Would be difficult to even comment on which version someone is asking about.

  6. Very interesting thoughts Eileen– thanks for sharing. I suspect that some people that make unreasonable claims about their methods actually do not thoroughly understand the mechanics and theories that they are using,…and it’s hard to have a reasonable discussion with them because of that. It took me awhile to figure all of this out, with the help of some very smart teachers and an inquisitive nature. All we can do is keep learning and sharing more!

    • Hi Betsy. I suspect that is true. Goodness knows it’s a journey, and all of us have our learning curves. There are many times I’ve thought I had it all figured out and was pretty badly wrong. Like you, I have been very fortunate in my teachers and look forward to learning more. Thanks so much for your comment.

  7. Me says:

    Of course we put it there, and if we didn’t, we’re keeping it there by not taking our dogs away from it, which is pretty much the same thing as putting it there in the first place.

    And YES, it IS an aversive at an over threshold intensity, otherwise the dog wouldn’t feel the need to be making displacement and appeasement signals. There’s no and, if, or buts, about it. This is negative reinforcement.

    I appreciate the back and forth motion for it’s desensitization merits, but one doesn’t have to work the dog close enough to cause the dog to feel the need to do displacement/calming signals, and one can easily factor a counter conditioning element into the procedure. When correctly used, it shouldn’t detract from the procedure but be of benefit.

    Very brave for speaking up against this, Eileen!

  8. calkinsbetsy says:

    Hi Eileen, Thanks, as always, for your thought-provoking article. Do you think that some of people who react emotionally (ie without dignity) may be acting out because they don’t deeply understand the mechanics and theory behind the different methods? It’s hard to continue a high level discussion when one runs out of thoughtful points — then the emotional ammunition comes out. Everyone is at a different point in their journey of understanding. That has frequently been my experience, anyway. I find it easier to walk away from nonsensical arguments. I am also left frustrated later, so I do empathize with you! I try to remember when I knew less, which feels like only yesterday. (oh—wow—I guess it was just yesterday…!) 🙂

    • Thanks, Betsy. I don’t know if you intended both comments, but this one adds more, so I’m publishing it too. Thanks! Yeah, yesterday for me too. Or an hour ago…..

  9. Mel says:

    You are calling for trainers who use something other than CC/DS to treat fear reactivity to be honest about where the “aversive” comes from.

    1) The same place it comes from when you CC/DS. We’re starting at the same distances, sometimes farther away. In what world does this mean we are exposing dogs to stimuli at aversive levels and you are not? At least ours get to choose whether they want to approach or retreat.

    2) How do you know your dogs in a CC/DS protocol are below the “aversion threshold”? Just because they are not showing signs of displacement doesn’t mean they do not find the stimulus aversive.

    3) Who says we are not also doing CC/DS?

    4) The “sanctioned” negative reinforcement comes in at the same times someone doing a CC/DS protocol would simply retreat. The difference is we take the opportunity to attach desired behaviour to the retreat. It does not mean we stand there while our dog is distressed waiting for desired behaviour. The whole point is to move them when they are still doing something you like. E.g. dog stops and looks towards stimulus for a couple of seconds, you mark and retreat. You marked the lower intensity behaviour of a long look and rewarded with distance. Distance you would have given them anyway. You can give it non-contingently, or you can make it seem like the dog made it happen with some lower intensity reaction. It still comes, just as fast, at the same point in the procedure. What do you suppose the difference to the dog is? Do you think they don’t notice what happens before you pull them out non-contingently because you are doing a respondent protocol, not an operant protocol? They don’t know they are doing a respondent protocol, not an operant protocol.

    If you don’t catch them in time and they start reacting, there are no good opportunities for training, here, and they are simply given the space they need. Oh noes. The differences between us are unfathomable. I expect that “non CC/DS only” practitioners are trying to argue that they are not introducing an aversive that wasn’t already there when they first set the situation up. If it wasn’t a set up, they are not taking their dog directly to a situation that is aversive in order to remove them upon contingent behaviour. They are using what happens in spite of their best efforts to control the environment.

    5) I actually prefer to desensitise without counter-conditioning as a general rule, because the potential to introduce pressure through food is generally greater than the potential to introduce pressure through the secure base effect. I have seen professional trainers pushing dogs too close to aversive stimuli with the promise of food too many times to count. Yes, it is a CC/DS protocol done incorrectly. It happens. In no way is CC/DS an automatic moral high ground over other methods. If I were as overly generalised and negatively biased as this piece, I might argue that this in fact means that trainers who use CC and DS to treat fearful and reactive dogs are introducing unnecessary pressure into training and rehabilitation and I would ask them to be honest about whether they actually can claim to be using no negative reinforcement at all. They could be training an approach with negative reinforcement.

    If you want science on negative reinforcement, I have plenty of that right here: [from Eileen: link removed, but you can find Mel’s piece by searching on the title of the article: Negative Reinforcement: The good, the bad, and the ugly.]

    • Hi Mel,

      Thanks for being honest and being clear that you are using an aversive.

      You are correct that I am calling for trainers who use something other than DS/CC to treat fear reactivity to be honest about where the “aversive” comes from. And as you wrote, it’s the same place it comes from when I do DS/CC. I bring it or arrange it.

      Exactly! Maybe you missed the section of my piece (and the photo) where I acknowledge bringing in something potentially aversive. I think it is something everyone should acknowledge. Not sure why you think it’s a point of contention when its the main focus of the piece. I addressed my piece to the people whom I notice are trying actively not to acknowledge it.

      The rest of your comments appear to have little to do with the rather narrow premise of the article, but there is one point I’ll address.

      Regarding not being able to tell whether my dog is under the threshold of stimulus aversiveness: of course! It’s an area where errors and assumptions are easy to make. This very point—that we don’t know exactly what is going on inside the animal—is a good reason to err on the side of being conservative with one’s assumptions and aim for a non-aversive environment and learning situation, rather than aiming for some supposedly acceptable amount of stress from exposure at an aversive level. Because what is going on for the animal could be a lot more aversive than we think.

      You must also be aware that researchers have been investigating ways to test for aversion levels for many years. Some of these can come out of the lab and be useful in real life. One of the ways it’s tested is that the animal is taught an R+ behavior that is available to engage in. When a warning of the onset of an aversive or former aversive happens, the question is whether the animal continues unbothered with the R+ behaviors or do they disengage and heed the warning?

      The level of engagement/disengagement can be quantifiable, and one could take a ratio response to this, kind of similar to what Jean Donaldson demonstrates in this video about quantifying a dog’s success in a operant learning protocol.

      http://www.youtube.com/watch?v=0frgWXcRvTA

      Again, Ms. Donaldson is dealing with operant learning; we’d have a different metric in a classical conditioning scenario (and be aiming for 100%). I actually have a blog coming up on the various tested ways of knowing if an animal is over the threshold of stimulus aversiveness.

      I want to point out to my other readers that Mel makes several more claims that are common memes from those who are seeking to justify practices that center on the use of aversives. One is the false implication that dogs who are undergoing DS/CC are being micromanaged and can’t move or make choices. Another is the erroneous claim that exposures for the different protocols happen at the same distance or intensity. In my webinar on thresholds, I discuss the one scenario in which an animal might undergo a negative reinforcement protocol farther away than undergoing DS/CC. It is a rare one, and I don’t discuss it in my other resources on thresholds. But the important point is that it is not the raw distances that matter, it is simply whether the exposure to the trigger is at an aversive level or not. Negative reinforcement can’t take place at a non-aversive exposure, and DS/CC is not properly performed at an aversive one. While finding the “line” by watching the animal’s behavior takes skill and may not be exact, the difference between the necessary levels of exposures is definable and not blurry.

      One should always take heed when a writer attempts to to blur the differences between protocols that are centered on the use of aversives and those that are not. Again, I am not saying that accidents can’t happen and a dog undergoing DS/CC couldn’t end up in an aversive situation. But the bases of the protocols are different. This attempt at blurring appears to be an example of the continuum fallacy, which erroneously claims that because two possible actions are on a continuum of behavior, there is no meaningful difference between them. This is a link to my piece on the continuum fallacy and the attempts by some people to define away the beliefs or practices of others. If you read the piece, be sure and heed the part near the end where I discuss where one group tries to claim another doesn’t really exist. And which group it usually is.

      • Mel says:

        “I want to point out to my other readers that Mel makes several more claims that are common memes from those who are seeking to justify practices that center on the use of aversives. One is the false implication that dogs who are undergoing DS/CC are being micromanaged and can’t move or make choices.”

        If they have no way to signal that they are uncomfortable other than actually looking uncomfortable, then they do lose a level of control over the situation. If you’ve ever trained a safety behaviour, you would find you get some interesting surprises when it gets performed. It’s like “What? I always thought you were fine with that. Apparently I was wrong.”

        “Another is the erroneous claim that exposures for the different protocols happen at the same distance or intensity. In my webinar on thresholds, I discuss the one scenario in which an animal might undergo a negative reinforcement protocol farther away than undergoing DS/CC. It is a rare one, and I don’t discuss it in my other resources on thresholds. But the important point is that it is not the raw distances that matter, it is simply whether the exposure to the trigger is at an aversive level or not. Negative reinforcement can’t take place at a non-aversive exposure, and DS/CC is not properly performed at an aversive one. While finding the “line” by watching the animal’s behavior takes skill and may not be exact, the difference between the necessary levels of exposures is definable and not blurry.”

        I did a BAT session last week. It looked exactly the same as a desensitisation protocol. Because it was. I’m not blurring lines, here. It really does exist on a continuum (level of aversiveness), and the ‘distance’ on the continuum between the two protocols is small in some cases and non-existent in others. I don’t think that BAT is a protocol “centred on aversives” any more than CC/DS is. How can you call it a protocol centred on aversives if you could do the protocol, leave out the marker, and it would look the same as a protocol that is allegedly not centred on aversives? I said the distance was the same, but it’s the same because the aim is the same. No other reason.

        You presume that for negative reinforcement to occur, the stimulus must be present at an aversive level, but you also acknowledge that there are difficulties with assessing what is aversive in the first place. There is a grey area called “tolerance” between what is aversive and what provokes an observable response. If we provoke an observable response, we can assume the stimulus is aversive. But if we don’t, we can’t assume it is not. Equally, if the stimulus is aversive, but has not provoked a response, we can still remove it, in which case, we are still removing an aversive stimulus and negative reinforcement could occur. So the difference between these levels of aversion are observable, but not absolute. It is quite easy to CC/DS an animal to something when they are showing signs they find the stimulus aversive. If you miss the signals, or for some reason can’t get the reduced stimulus intensity you need, the protocol will still work. I’m not sure who has decided it must be done below some difficult to identify aversion threshold in order to be done “properly”. It’s being done “properly” if it successfully conditions a new emotional response to the trigger.

        • As I’ve stated before, thank you for stating in your first response that BAT involves aversives. That is the point of my post, that people—all of us—be honest about using them and bringing them into a situation. I appreciate that you have stepped up.

          The rest of the discussion is going very far afield. I’m publishing this comment from you as a courtesy but will not be replying point by point. Comparing the level of aversive exposure between DS/CC and BAT is not the topic of this post. I think it’s an interesting topic, but if I want to discuss it, I’ll write a post about it, rather than discussing it in the comments of another.

          Since you are not one of the people who denies the use of aversives in BAT, or splits hairs about where they come from, perhaps if you are in contact with people who do you could clarify that with them. Again, I appreciate your honesty.

          • A note to other readers: I suppose it’s tiresome for me to keep plugging my webinar, but so much of what Mel is talking about is spelled out in mine on threshold, including graphics for the visual learners. For instance, it covers in detail the range in which all of these protocols, separately and some together, can be successfully performed. There are some surprises there. I put as much of it as I could in the threshold movie, but there is much more detail and background in the webinar.

          • Mel says:

            Thank you for indulging me and publishing my somewhat combative comments. I confess I am more confused than ever about what the problem actually is and what is on topic, here. You start out saying you are mystified by trainers insisting they are not adding an aversive stimulus to their training and go on to point out that everyone is adding it, by orchestrating setups. But these claims about where the aversive is coming from are being aired in the context of ethical standards. There are subtleties here that are being ignored. These trainers are not being asked to account for exposure to aversive stimuli. They are being asked why they would “add” an aversive at all (which I have always thought peculiar, seeing as the consequence is something is removed), and they respond that they are not adding it. Certainly not in the sense that a trainer would “add” an electric stimulus so that they could then take it away. And now they get side-swiped for not acknowledging that they do actually have setups, just like any other trainer. So, your gripe is kind of out of context to begin with.

            Next, you say it is irrelevant anyway, because what is relevant is that the trainer decided to use an aversive at all, thereby setting the stage for the context these discussions ARE taking place in, where any aversive is apparently ethically questionable. But when I tried to point out why it does matter by contrasting with e-collar use, you said it was completely unrelated to your post. Okay, I guess I haven’t figured out what the post was about, yet.

            You then go on to make some comment about people adjusting “the science” to suit their training preferences, but earlier you claimied that CC/DS protocols if done properly are performed below an aversion reaction threshold, which I have never seen quantified in a scientific study and I would love it if you pointed me in the right direction. Same for negative reinforcement only occurring over this threshold. But this is off topic, although you were the one that brought it up. I merely argued it wasn’t so clearcut as all that and CC/DS practitioners were not automatically safe from doing a little negative reinforcement of their own.

            So… I dunno. You want people to admit they deliberately expose dogs to scary things? Why? Everyone does. It’s kinda obvious, isn’t it? Where we are talking about the differences between protocols, I would argue it’s an irrelevant point, because it is the same between protocols, and clearly that is not what other trainers are referring to when they say they didn’t add an aversive stimulus. And a lot of them are not doing setups. They are just going about their daily lives dealing with scary things when they bump into them.

            Ah well, never mind. I am sorry if I came out swinging because I misunderstood what you were getting at. I think I am still misunderstanding, but that’s okay. I’ll leave you to it.

            • As you said, everybody has setups (and also of course has real life exposures to assess and contend with). I made no swipe about having setups. But only in the context of a negative reinforcement protocol have I seen the (repeated) claim that everything is OK as long as you aren’t the one that put the aversive there. I argue that they did put the aversive there. It’s that simple. They can make another argument about why it is better than other protocols or delineate the differences, but they did arrange for the aversive to be there.

              For reference, for other readers, this earlier post of mine makes an argument that learning theory does make a distinction about whether a third party is _using_ an aversive, but not whether or not they added it.

              Anytime someone says, “This particular aspect of aversive use is OK because….” I pay close attention. Often the next thing I see is the statement spread like wildfire. I believe we need honest discussion about the ramifications of aversive use for both the dog and the trainer, not dismissive statements about it. Which by the way is the topic of my next post.

              My statement about “proper” use of DS/CC comes from a compilation of several sources, including ones for DS alone (human and animal, which are different because of some ways that the original DS protocols with humans can’t be replicated with animals) and CC alone. Also from a review of standard practices from non-scholarly training books on how to perform DS/CC most effectively. It’s in stuff I have already published, and I may compile it for a post in the future.

              You also asked for a source for negative reinforcement only occurring over the threshold of stimulus aversiveness. It is by definition. Negative reinforcement is an increase in behavior resulting from the removal of a stimulus. If there’s no stimulus at an aversive level there’s no R-.

    • “Just because they are not showing signs of displacement doesn’t mean they do not find the stimulus aversive.”

      Assuming one is schooled in how to read the communications of a domestic dog’s body language and vocalizations, and assuming the dog is not showing displacement, appeasement, or stress signals but instead, tension free body language that indicate the dog is at ease and a CER+ is in the works, what is to suggest that the dog finds the stimulus aversive? Is a dog is sub-threshold, as is supposed to be the case with CC/DS, then by definition, the dog should have no reason to find the stimulus aversive, and in fact, if they do, then the procedure isn’t correctly being executed. The fault therein, lies with the human variable, not the procedure.

      I don’t understand the merit of dismissing CC/DS based on the premise that you *could* do it incorrectly. The very same could be said for *any* procedure, including BAT.

      Just as you could err when doing CC/DS and lure the dog in closer than would be considered beneficial, you could very well do the same by placing pressure on the dog when utilizing negative reinforcement. One could create a response depression specific to this during the exercise alone and suffer the effects outside of controlled exercises.

      Also of note is that food is only the tip of the iceberg when it comes to training. One does not *have* to utilize food as their counter conditioning element. But food, like anything else, is a tool and needs to be correctly utilized.

      • Mel says:

        I would assume if the dog notices the aversive stimulus, their experience is probably at least a little bit unpleasant. Why wouldn’t it be? To make a positive association with something scary, you have to be noticing the something scary. I don’t know about anyone else, but if I am noticing something scary that is far enough away to be not very scary, I will forget about it when the candy comes out. It needs to be scary enough that when the candy comes out, I will still be keeping an eye on it. This is why I don’t think it’s ethical to just counter-condition willy-nilly. Think hard about whether the dog actually does need to be cool with this.

        And no, I’m not discounting CC/DS because it can be done improperly. I use it as much as anyone. I am arguing that it is not automatically an ethically superior training decision.

  10. For CC/DS the stimulus must be present. How exactly do we know, that the animal does not find it aversiv.
    For BAT the stimulus must be present. Is it in that case more aversive just because the handler choose to add distance as a (one part of the) reinforcere(s).

    For me the ethical question is neither how the aversive got there, nor which quadrant is better than the other, but how intense the aversive is.
    As both protocols start below threshold (in case of BAT WAAAAAAAAAAAY below threshold becaus there is noch food distraction…) I can’t see the ethical distiction between CC/DS vs BAT being very much favouring CC/DS.

    • Hi Martina,

      Please see my previous remarks to Mel about the question of whether something is aversive or not.

      No, I don’t say that the handler’s choice of adding distance means something is more aversive (I do understand why you mention that). But if the stimulus is at an aversive level, then and only then will adding distance act as a reinforcer. If one is actively trying to use distance as a reinforcer, one is actively trying for exposure to the stimulus at an aversive level. Sure, trying is not always the same as doing, but I am assuming a level of skill where the handler can detect whether behaviors are being reinforced or not.

      I appreciate you don’t center your arguments on how the aversive got there. That’s really the point of this post.

      About food affecting location of thresholds: sure, and many other things as well can operate as safety signals or lack thereof. But I contend that raw distances are immaterial. Let’s say that the threshold of stimulus aversiveness is 100 meters if the handler has no food and 5 meters feet with food (a deliberately extreme example). If you are at 101 meters without food you are 1 meter from the dog’s threshold. If you are at 6 meters with food you are also 1 meter from the threshold. If the longer distance is why you say the dog doing BAT is waaayy under threshold, I disagree. It’s still just 1 meter. Thresholds can change because of all sorts of factors and are not defined by raw distance. The big difference between the protocols is what happens next. For negative reinforcement to “work,” you have to cross that threshold.

      I see the ability of DS/CC to build a conditioned positive emotional response as an ethical distinction.

      Thanks for the comments.

      • Paul McGee says:

        This stuff warrants discussion, because there are subtleties upon subtleties in deciding how well any simple model fits the real world. George Box, statistician, said “all models are wrong, but some are useful” … but even better “all models are wrong; the practical question is how wrong do they have to be to not be useful”.

        You have mentioned about the presence of food (eg) modifying thresholds. I can think of two paths for thinking further along that track.

        Firstly, the food we hope helps creates a +ve association, which is great, but it also lets us be closer we suppose for the “same level of discomfort”. There must be a reason we want to be closer. We could stay back where we started from … it is still the same food.

        Being closer allows the animal to experience the “scary object” at a different amplitude of sight, sound, smell … and still (we assume) feel equally ok. So that is an instance of experience they couldn’t have had (in quite the same way) at the lesser exposure. I mention it just to say there must be more than just emotional learning “going on” in this process of building +ve association and “confidence”.

        The second point is kind of a mechanical one. “Presence of food” is a bit of a vague concept when you think about it some more. Say for arguments sake that that we could tell a dog is X amount comfortable with an object at 60 feet “naturally” (on a given day) but 40 feet “in the presence of food”. We wear a bumbag … it smells of roasted pumpkin perhaps.

        We could be rewarding each step we take, or every 10 seconds, or once in two minutes. We could be walking the dog in to whatever distance, patting and praising, then leaving. The rewarding rate can change. We may be rewarding operant behaviours. We might run out of treats and not let the dog know about it immediately.

        The thing I wonder is, assuming we could theoretically see and know these thresholds in real time, would they change from the “natural” on the sighting of the bumbag? on putting on the hip? what if it were left on the sidelines, on a chair to run back to? would there be a continuous high frequency waxing and waning of these lines as each treat was produced and given … or a smooth transition over time from the preparation for training, to treat delivery, to exhausting the supply, … and – the initial thought that came to me – what effect does this variability have on the animal?

        I wonder whether approaching these procedures in a standard, ritualised way (somewhat) has a additional advantage in possibly regularising the streaming emotions experienced in real time by our dog … or whether it is going to be highly individual to each animal. I know some dogs that seem to be processing 19 thoughts a second, where mine is chewing on a single thought for seconds at a time.

        • I mentioned food changing things because it seems central to people’s arguments about BAT starting farther away. There are countless other factors that could make a dog feel more or less safe in a particular setup and change their comfort range. I think (and I believe I have shown in the threshold work) that the raw distance doesn’t matter in the definition of what kind of exposure is going on, only 1) whether the dog can perceive the stimulus; 2) whether the stimulus exposure is at an aversive level, that is, whether the dog would like to leave; and 3) whether the dog is at the threshold of a fear response.

          See also an addendum I wrote in the body of the post.

          The only other thing I would respond to in your musings about food letting us be closer for the “same level of discomfort.” Just a note to say again that the goal for DS/CC is NO discomfort.

  11. Andreja says:

    First let me say that I am not very sciency about this 😉 While I can dissect situations according to R+, R-, P+ and P- I don’t usually think in those terms. But then I only use BAT with my own dog. I look at his emotional state, results next time around and in how this affects his trust in me and that’s how I decide whether I like a method or not.

    I never thought of BAT as requiring the dog to do something in order for us to get him out of situation, though technically this is true. But needing to get him out of situation means failure. What I want is for him to get out of the situation on his own and he will only do that if he is quite comfortable. Of course I act as his safety net – I will not let him get past the stage where he *begins* to show discomfort and if he decides he’s not going there that’s great! He is free to increase the distance at any point and I will praise and treat him for it if he does it on his own… so I don’t really see an ethical problem here. If this was a child and I would take him to the ZOO, but he would be afraid of tigers and I would NOT push him in any way to see the tigers (though he would know they are at the ZOO) but would instead follow him wherever he wanted to go (to see the seals on the other side of the ZOO perhaps) I would not feel like I’m maltreating the child.

    Of course most children are smart enough not to pull toward the scary thing. My dog pulls toward his triggers and BAT has been wonderful in teaching him that he has a choice. He doesn’t have to look at tigers. In fact, I will be happy if he doesn’t. We can go and watch the seals instead.

    • Hi Andreja,

      You sounds like a sensitive and thoughtful trainer.

      I think it is valuable to think through the science. Waiting for a dog to perform a particular behavior (“displacement behavior” or “cutoff signal” they are sometimes called) is putting a contingency on escape, and from what I understand from taking the webinar, that is still a part of BAT 2.0, though less prominent than in BAT 1.0. In Stage 3 BAT (1.0) that was the protocol. It was a successful repetition, not a failure.

      But actually, I like it that for you, needing to get him out of a situation means failure. It does for me too. It means I went too fast with DS/CC. Or sometimes just that something happened that was not under my control.

      Exposure to scary tigers in a zoo is something I would manage, and what I would do sounds like it might look very similar to what you do. But that would be a rare occurrence in my dog’s life. If and when my dog is bothered by something that is going to be a regular part of her life, then I want to take the best action I can for her.

      Having fear turn to joy through the process of DS/CC has been a wonderful experience for my dog. It has opened up her life to many more opportunities and reinforcers. They were there all along, and she would never have experienced them if she had been reinforced for avoidance instead. If you haven’t had that experience of building a positive conditioned emotional response with your dog, I encourage you to try. The ways it opens up the dog’s behavior and life experiences go far beyond just eradicating the one fear (which is cool enough itself). I would never want less than that for my dog.

      Thanks for the comment and good luck with your dog.

  12. Mel says:

    I think it is a fair distinction trainers are making, here. If we say it is okay to expose a dog to an aversive experience in order to teach them a favourable behavioural response, we are essentially condoning what a lot of trainers do with tools like e-collars and prongs, which we would not do ourselves. But there is a difference between doing that and pairing a behaviour with retreat and safety from a scary thing. Let me make it clear for you:

    BAT
    We don’t want our dogs to be afraid of this thing, which is why we are facilitating their retreat from it.

    E-collar/prong etc.
    Would quite like the dog to continue wanting to avoid the stimulus forever.

    BAT
    Not exposing dogs to scary things in order to get behaviour we want, or “proof” behaviour. We are exposing them to scary things to help them develop successful ways of coping with it so they do not have to be scared of it.

    E-collar/prong etc.
    Exposing dogs to noxious stimulus to provoke a behaviour regardless of whether it is ethologically sensible at the time.

    BAT
    Abort training and prioritise getting space if dog becomes reactive. Controlled retreat if dog becomes distressed.

    E-collar/prong etc.
    Up the ante if dog becomes reactive or distressed.

    If you boil it down to quadrants, which is what a lot of “force free” trainers are doing, you put us all in the same camp. This is not fair, because we might be using the same quadrant (on occasion), but we are using it very differently with very different goals, and very different ways of assessing effectiveness and welfare. Is it any wonder that BAT practitioners are feeling beleaguered and attacked? They are being lumped with trainers they have fundamental differences with by trainers they have very few differences with, and they are somehow expected not to try to defend themselves and explain why they are distinct. If you and yours can’t see the distinction, then I think that is your problem, not ours. I find it incredible that some force free trainers seem to think they can perform a perfect DS/CC protocol every time and transition smoothly to a DRI without ever drifting into “bad” quadrants. I call you to be honest with yourselves. Can you be sure you don’t use negative reinforcement with fearful or reactive dogs unintentionally? Better the devil you know, I always say.

    • **Boggle.**

      Mel, do you have me mixed up with some other blogger? One who is out there saying that BAT and prong/e-collar use are the same?

      I have written plenty about the breadth of the R- quadrant (and the others, as far as that goes). I have not made a blanket condemnation of it, but I have written against blanket acceptance of it. I have also cautioned about general arguments in favor of it because they can indeed become adopted by those who are using painful correction collars and the like.

      I almost didn’t publish your comment because it’s completely unrelated to my post, and I personally don’t need the education about BAT (1.0 or 2.0). I’m letting it stand but will not publish further posts comparing and contrasting BAT to shock collars. That’s not what we are talking about here.

      Likewise, you are barking up the wrong tree if you think I won’t “admit” to unintentionally using R-. I don’t know that it happened often enough to effect behavioral change (part of the definition) but of course I have accidentally used escape, pressure, the whole nine yards! And I probably will again. Much of it is delineated in this blog. I’ll take your word that there are people out there claiming absolute perfection, but I have never read that. In the context of my post and my own training and ethos, the perfection claim is a big hulk of a straw man.

      • awesomedogs says:

        Who has not at some time erred? I have slipped with a knife and cut myself. I don’t stop using a butcher knife. I do however try to use better knife skills.
        I’m not sure anyone should be choosing any method based on what happens when it’s done wrong. They should be looking at the process, the possible end results of various methods and side effects that may happen if done as directed.

      • Mel says:

        Nonono, you have misunderstood me, and probably rightly so as I jumped ahead. I was trying to say that the reason why BAT practitioners are using this argument about not adding the aversive is precisely to differentiate themselves from trainers using R- in very different ways they do not agree with. The whole point of saying we are not adding the aversive ourselves is certainly not to absolve ourselves of responsibility if our sessions go awry or we find ourselves in a situation where the dog is distressed. I could only interpret this blog post in terms of what happens when training sessions go off track, because I don’t think the protocols are very different at all, and the biggest difference is how those situations are handled by the trainer, and that’s what you seemed to be talking about with both protocols.

        I wasn’t comparing BAT to e-collars and prong collars. I was contrasting them to show why BAT practitioners are eager to differentiate what they are doing. There is an underlying assumption that if an aversive is being “used” by a trainer, that they have made a decision that is ethically questionable. You say yourself “So the fact that people are still mentioning this irrelevancy about “who put it there” seems like a lot of hand waving to shoo away the real issue: choosing to use an aversive.” Whether you make the direct comparison or not, you are putting them in the same camp, because you are not acknowledging that there are grey areas, here. Either you use an aversive or you don’t. Apparently it is irrelevant whether the aversive occurred in the exact same situation as it might in a DS/CC protocol (where, for whatever reason, you need to increase distance again), in exactly the same process. It’s not irrelevant just because it is an aversive. It’s perfectly relevant how you choose to use an aversive, otherwise we may as well all be split into “will use aversives” and “won’t use aversives”, which is a dichotomy based on intent, not reality, and makes no differentiation between something gentle that improves an animal’s emotional state and something heavy that makes it worse.

        • You said, “I was trying to say that the reason why BAT practitioners are using this argument about not adding the aversive is precisely to differentiate themselves from trainers using R- in very different ways they do not agree with.”

          I think anything that highlights differences between protocols is helpful, but only if the distinction is accurate. There are plenty of differences between BAT and other protocols. But “who adds the aversive” is not one of them. I see it used to justify aversives over and over again. My argument is still the same: if a trainer is arranging or seeking the presence of an aversive, whether it is a device, pressure from their own body, or a stooge dog or person, they are adding it. And if they are using it to train avoidance via negative reinforcement, they should say so.

          I’ve addressed the general arguments about DS/CC and R- protocols looking similar or being performed in the “exact same location” as an addendum in the body of the post.

  13. thejen123 says:

    In response to Mel’s comment above regarding where “sanctioned” R- comes in. Mel claims it is at the same point in a DS/CC protocol where the handler would get some distance. In this example, while you are still doing BAT and using R-, the competent DS/CC trainer simply switches to a DR+I.We would do Let’s Go! or Touch! and heavily reinforce the dog. We switch to R+, you switch to sanctioned R-. An effective DS/CC protocol includes DR+I. We don’t merely transition to DRI, as Mel has erroneously stated, it is an integral part of the protocol before, during and after DS/CC procedures. Mel says the dog doesn’t know whether they’re doing an operant or respondent protocol. She goes on to ask “What,do you suppose, the difference to the dog is?” The difference to the dog is in the R- and the R+. Which would you choose for your dog? You can argue all you want about how you’re teaching the dog to cope with BAT, but there is no CER+ possible with R-. The motivation, the brain chemistry is very different. I applaud Eileen for shining a light on the differences in these protocols.So, to answer your question, Mel, I suppose it matters to the dog very much, indeed.

  14. Kelsie says:

    You have such a skill for putting things into words! Wonderful as always.

  15. Thank you, Eileen, for all your work. I watched your webinar (with PPG) and have been following the discussions of BAT 2.0 in different forums. I am always interested in learning and I do learn a lot from your blog. I have not used BAT but from what I am learning, I do not think it is a protocol I would recommend.

  16. Sarah Owings says:

    Challenging, thoughtful post as always Eileen. Thank you. I want to comment on your two main points: taking responsibility for the presence of aversives, and the issue of contingency.

    First, I am hopefully one of those trainers you feel does take responsibility for the aversives I arrange–no matter what protocol I choose to use. But I am also someone that feels that what I have learned from CAT and BAT in terms of allowing room for choice and built-in patterns of relief for the dog are actually a great complement to CC/D–as long as I keep a sharp eye on thresholds (something I’m even better at now thanks to your great webinar), and strive to eliminate the elements of contingency and extinction (extinction being easier to fully jettison than contingency sometimes). On the flip side though, learning about CAT and BAT has also really forced me to look hard at my implementation of CC/D too, because as other commenters have noted, I have made the mistake as well over the years of using food as a crutch, or as a micro-managing tool, or even a tool to “get the job done,” inevitably resulting in pressuring the dog–mistakes I take full responsibility for as well.

    As for the issue of contingency and R-, this is an interesting one. The evolution away from 100% rigorous adherence to *only* reinforcing desired behaviors is an important new emphasis in dog training, a development which I think many people have not yet caught onto fully yet. It is hard to let go of the old mind set of “reinforce what you like, ignore what you don’t like.” Whether you strive to exclusively be a R+ trainer (always with the potential for frustration when the inevitable P- occurs), or one that intentionally utilizes R- too, the concept of non-contingency as a starting point for learning is a whole new frontier. When it was first developed, CAT was about total contingency and the use of extinction–a process which was very hard on the learners at times. Grisha’s BAT protocol sought to lesson that stress by eliminating the extinction part, but yes, the older version of BAT and many of its current practitioners still rely on contingency more than they often realize–and I absolutely agree with you that once contingency is a part of a R- protocol, you are using aversives to build behavior. There is no sugar coating that–and it shouldn’t be sugar coated either. But, Eileen, I would also encourage you to see the thinking behind CAT to BAT to BAT 2.0 as a gradual evolution *away* from the contingent use of aversives–and that is an evolution which should be celebrated, not criticized in a way that has the *potential* side effect of encouraging people to completely dismiss these protocols altogether. Behavior change is challenging, difficult, complex work, and I personally believe we need all hands on deck in terms of open-minded thinking in order to help dogs live happier, less stressful lives.

    As you know, I often work with dogs with challenging behavioral issues. Sometimes the complexities of their lives as well as their owner’s lives make 100% aversive-free learning almost impossible. Sometimes just by stepping into these dogs’ homes, or coming into view at all, or even the sound of my voice on the other side of a door, I become the aversive. Once I am already in a situation with a dog where I am the aversive, it is often more ethical I think to be sure the dog can escape, and /or to provide as much non-contingent relief from my presence as possible. Sometimes I do this by walking or turning away. Sometimes I do this by tossing treats so the dog moves away from me. Sometimes I do it by being sure the dog has plenty of room to move freely with minimal restraints–all principles highlighted in BAT 2.0. Non-contingency is an ideal to strive for, but it is not always achieved perfectly….But yes, you are absolutely right, it is only by recognizing that I have created (or become) the aversive in the first place, that I can be aware enough to relieve it as quickly and as often as possible as I work towards building a happier CER.

    However, IMO if we are going to really be honest with ourselves about these things, the one true alternative to *potentially* stressing a dog in a learning set up, is avoiding that set up altogether. As Susan Friedman’s Humane Hierarchy says, a preferred solution to behavior problems –whenever possible– is changing antecedents, not actively changing behavior. Level of intrusiveness increases once we take any step towards “fixing,” or “training” a behavior problem away. R+ can be intrusive too. So, to your point, when I hear BAT oriented trainers talk about the aversive “already being present in the environment,” what I have always took that to mean is: the aversive is something unavoidable in the animal’s life. Now we can debate on a case by case basis whether the trainer or owner’s perspective on this choice is one we agree with, but it there is always a choice being made for the dog. Every choice we make for our animals should be weighed against this standard: is the aversive avoidable? Is the chance of stressing the animal worth it? However, we also have to ask ourselves, is doing nothing at all ultimately going to create more stress? The two alternatives should be weighed very carefully and often we don’t get it right.

    Eileen, your work with Clara is beautiful, careful, thoughtful work, and lucky for her in the meantime, she gets to live in a perfect bubble at home with you where almost no aversives are present. My own dog Zoe also lives in a bubble where I strive each and every day to ensure she does not have to deal with other dogs rushing into her space. For years now I have been mostly successful with this–and her life is mostly aversive free too. Sadly, this is a reality most dogs don’t have. Sometimes, as a professional trainer, I really do have to “get the job done.” I have very high ethical standards for my work of course, and definite lines I will not cross, but sometimes the entire process is truly less than perfect.

    Just as an example, I was at a training workshop this weekend at a great training facility filled with kind, thoughtful, and skilled trainers, all there helping a variety of novice pet owners with their dogs’ issues. There were no corrections seen anywhere, lots of food reinforcement, tons of consideration for dog comfort, and only friendly, sociable dogs were invited…and yet, and yet, just by being in that space, just by being restrained on leashes in a crowded room full of other dogs and new people, many of those dogs–including the advanced, and highly trained dogs–all of them still showed signs of stress throughout the workshop, all of them. By the end of the day, I remember thinking that it would have been best maybe to send most of those dogs home hours earlier. Just go let them be dogs, you know?

    Humans are meddlers. We intrude. We control. We set agendas. We decide all sorts of things for our dogs that they would really rather not do much of the time. All we can be is as ethical as possible based on the knowledge and awareness we each have. I also feel we should all be as open as possible about all the training choices available, and of course, as conscientious and observant of our learners as possible. Ultimately, to me the issue then isn’t who’s responsible for the aversive being there, or which protocol is more intrusive than another, so much as, are we making the most ethical choice for the dog right in front of us right now?

    Anyway, this turned into a much longer comment than I anticipated. Sorry about that! Thanks for provoking a lot of thinking! With respect, Sarah

    • Yes Sarah, I had you in mind when I talked about those who take responsibility!

      It’s good to learn about choice, and I’m glad you have learned things that improve DS/CC (that sound condescending but I absolutely don’t mean it that way). I have never seen DS/CC as something that negates or limits choice on the dog’s part though.

      I’m going to leave the food discussion for another day, grin.

      I’m glad you are also critical of the sugar coating of the use of aversives. Since we live in a world where there are so many euphemisms and such fog about behavior, I think sugar coating can be very, very damaging, to individuals (their dogs!) and as a cultural trend.

      Your remark about evolution: although I see in part some of the changes you indicate, I can’t really answer this remark in full without launching into a full scale elaboration on why I am not willing to jump on that bandwagon or encourage others to. I will continue recommending caution and education and sticking to what we know works well, is unquestionably humane, and still has fewer risks of fallout.

      What I state in this blog is generally what on would find in a Learning Theory 101 textbook, and practices deduced from that information. And the little I might say beyond that comes from direct, explicit recommendations from experts such as Susan Friedman. I think it is most responsible of me to keep it that way.

      I understand your proposed explanation for talking about “the aversive already being in the environment” as making a distinction about something that is or is not typically present in the dog’s life. I appreciate that you addressed the focus of the post. Whether something is likely to be a regular occurrence in a dog’s life does certainly affect choices about whether and how we address it. FWIW that’s not the way I think I have seen the comment couched, though. Sure, we can’t interpret what every single individual means by it, but I have most often seen it in the literal sense of an arranged setup, when people are arguing about whether something is or is not aversive in the first place.

      You mentioned: “the concept of non-contingency as a starting point for learning is a whole new frontier.” I know you were heading in another direction, but I want to remind other readers: desensitization and counterconditioning is a non-contingent protocol, known about and used since the early 20th century. There is no contingency on the animals’s behavior. The trigger appears, and great things happen no matter what the animal is doing.

      What to do when one’s own self is the aversive is another great ethical discussion. It’s something that I have no doubt you handle well with your client dogs. A skilled and thoughtful person can turn themselves into a cookie, playmate, or friend remarkably quickly in most cases, it seems.

      One last comment: You said: “Ultimately, to me the issue then isn’t who’s responsible for the aversive being there, or which protocol is more intrusive than another, so much as, are we making the most ethical choice for the dog right in front of us right now?”

      I agree that the main issue isn’t how the aversive got there, as I state in my post. But to me assessing the protocol for intrusiveness is wedded to making the most ethical choice for the individual dog. Hopefully not in a head-in-the-clouds theoretical way, but as Dr. Friedman suggests, using the ethical signposts and the behavior we see in front of us.

      Thanks for posting. Always good to hear from you.

  17. Comments are now closed. I will post my replies to the pending ones later today. Thanks!

  18. Pingback: Why Scratching an Itch is Not the Same as Performing a Force Fetch - eileenanddogseileenanddogs

Comments are closed.