Imagine that you are locked in a museum. Evil people have hidden bombs throughout the building. You cannot run away. There is no way you can call for help. If you smash the bombs with a hammer, they explode. Your only hope is to defuse the bombs.
Rushing about, you seek to find the explosives. As you find each bomb, your nimble fingers gently open the casing. You pry apart the mass of tangled wires and you deftly clip the wire and disable it. Then you rush off to find the next one and the next one.
Which quadrant did the evil villain use?
I love bizarre questions like this because sometimes the insanely exaggerated helps us understand the mundane. Running around looking for bombs is creepy and twisted. Thus, it is…interesting.
Plenty of behaviours are increasing: searching, prying apart wires, clipping them. We know that this is some form of reinforcement. There is an obvious aversive (something unpleasant that we would rather avoid). Bombs are clearly an aversive in the vast majority of situations. Thus, this is clearly negative reinforcement. Without the negative reinforcement, none of the behaviours, the seeking or the defusing would take place.
Here is the twisted part. Generally, we flee FROM things that are nasty, scary, painful, uncomfortable or just plain yucky. If a tiger is trying to eat you, you probably will run away.
In the bomb scenario, we would run TOWARD the aversive.
While it is natural to flee from aversives, the reason we run toward a bomb is because the behaviours that lead to escape are near or part of the aversive. In order to escape the aversive, you must search for the bomb, touch the bomb, examine the bomb and cut the wires.
Placing the abort sequence on an aversive creates an odd scenario that causes people to run toward things that are scary, nasty and even potentially dangerous because the behaviour of approaching is part of the escape route.
While our hypothetical example might conjure up images of blind panic, there are other examples that do not. People who search for land mines do the same thing. However, it’s with a slow and steady purpose. As J.M. Lohr points out in Clinical Psychology Review, “”there is no expectation that the predictability of an aversive event will reduce the aversiveness of the event.” Having the ability to safely find and defuse bombs does not make bombs any less aversive.
When it comes to dogs, this sort of “put the abort button on the aversive” scenario is interesting for a couple of reasons. Two things change when the location of the abort sequence is near the aversive.
- The direction of escape changes. Fleeing turns into approach. Do not assume that approach equals lack of aversive control.
- Lack of overt fear does not mean lack of aversives. Calm, cautious behaviour may indicate practice and predictability, not lack of aversiveness.
Trusting common sense and rules of thumb can only get you so far. Thankfully, quadrants do not lie. Some might claim that quadrants are fuzzy. As Leahy and Leahy pointed out, “Just because the boundaries between night and day are fuzzy, it does not mean they cannot be meaningfully differentiated.”
The quadrants are four little boxes – it’s not rocket science. Changing small details such as the location of the desired behaviour can change everything. Fleeing takes the form of approach. These details are interesting and important elements that require our attention.
Exceptions such as these are especially important for trainers and rescue workers who evaluate dogs for placement into family homes. Approaching children is not necessarily the same as loving contact from children. It’s an important distinction if one is placing a dog in a home with a child.
Trainers can place abort sequences on or near scary things. Dogs will learn to approach them. Anything is possible if you know how to wield the sword of behaviour modification. The question in dog training isn’t whether you can but rather whether you should. Sometimes the quadrants get a little sneaky and you have to look very carefully to see if there’s an aversive lurking about.
I’d like to give Jean Donaldson a thank you for her assistance – for letting me ask her questions. I have never met someone who is so open to questions and helping others.
awesomedogs 
You’re right, Yvette, such questions are both very important and very timely. Owners and trainers these days try to keep five balls in the air as they juggle effectiveness, fairness, safety, clarity, and relationship.
That leads me to ask a couple of questions. First, how do you define an “abort sequence”, both in general terms and specifically in this scenario? The reason I ask is that as I discussed this article with other trainers we realized we had different notions of what an “abort sequence” is . . . and its impact on the dog, performance, and our relationship with the dog.
Second, I love both your Blurred Lines title and the night/day distinction analogy. Drawing upon them and my four decades as both a cop and dog trainer I can tell you that the one hour a day that twilight and dawn cumulatively represent have a disproportionate share of traffic accidents. There are many factors for this but my point in bringing this up is that although there are meaningful distinctions between night and day, operating where the differences are less apparent entails substantial added risk. The question is dies that analogy carry over to training and living with dogs?
Hi Steve – Love the police work you do!
Unfortunately, I can’t take credit for the day/night quote. Blurred lines – well can’t take credit for that either! But it seemed to fit. The basis for what I wrote about came from a research report called, “A functional analysis of danger and safety signals in anxiety disorders.” by Lohr. Lots of things in there that sent me on a research treasure hunt and lots more that I think is interesting.
The paper splits things up into danger signals (Things that warn of danger) and safety signals (Things that cue you to safety). Also relief and respite. They look at predictability and control.
There were so many variables and examples, I’m not sure it’s an easy answer. But a few examples from the report.
“Danger signals may increase the immediate experience of fear while simultaneously contributing to the reduction of pathological fear. Safety signals, on the other hand, appear to reduce the immediate experience of fear while simultaneously contributing to the maintenance of pathological fear.”
They tried changing the location of safety and danger signals in one experiment they listed. A light warned of spider and it was near a door. Light being danger signal. The door being the safety. If the light was near the door, people responded more quickly. Maybe because their attention only had to be in one place.
However, “unreliable onset of danger impedes the reliable establishment of safe places or behaviors. The organism will eventually become exhausted from its unsuccessful and futile search, and likely retire to a state of learned helplessness.”
This idea of having a safety seems problematic on some levels based on the research. The report goes on to say, “inconsistency maintains fear and may lead to a continue an erratic search for safety.”….”over-reliance on safety signals may have the effect of maintaining catastrophic beliefs…..Some safety signals may also interfere with the habituation process during exposure to feared objects or situations….avoidance behavior can persist in the absence of fear once safety signals have been established…..may also interfere with or prevent the learning of corrective information thus functioning to maintain anxiety.”
What type of behaviours? “behaviors carried out within a situation with the intention of actively preventing the onset of anxious symptoms.”
The behaviour used to give relief or respite can be any number of things. Hand washing for people who compulsively wash their hands. Grabbing at a friend when scared in a social situation and getting help to get out. Searching for exists. Gripping a cell phone to ensure you have a way to call a helper. Asking someone else to get rid of a spider. From what I read, there really isn’t a limit to the behaviours that can be used to offer relief and respite.
Do they ever use the placement of a safety signal to encourage exposure? Yes. For example, if someone has a friend that acts as a “safety” they might be placed in a shopping center and the scared person has to seek them out. Report says there is very little research to support these types of treatments…bear that in mind.
For this blog, I just focused on the placement of the behaviour. “Where is the behaviour that you ask of the dog?” And if you ask the dog to do a behaviour near an aversive and then offer escape, you can create approach. Kind of like the person sitting in the shopping mall. I just thought the bomb offered a really clear example that approach does not mean you like something. I found that really an interesting point.
Do I think the lines are blurred? No. A dog is either getting relief through escape or reinforcement from approach. To the dog, I don’t think the lines are blurred at all. Do I think there is risk in those foggy areas? Sure. I think the lines are blurred because we can’t read our dog’s minds. And we can’t read their intent based on approach and retreat. I think interpretation is tricky and fraught with errors. If you’re going to use a strategy, my feeling is you better be just as informed about the risks. To me, it’s something I really feel is important to know.
Steve:
This post gives more examples of maladative coping strategies…also known as safety signals. Some examples in there. I always link to at least the abstracts. So feel free to read through the research links if you enjoy that sort of thing.
https://awesomedogs.wordpress.com/2013/12/01/yes-you-can-reinforce-fear-or-at-least-enable-it/
Nice article, and analogy.
A trivial (and largely moot) question: how does Lohr define level of aversiveness? If I was forced to choose between an unpredictable aversive and a predictable–especially a controllable–one, there’s no doubt I’d pick the latter. Which would mean I find it less aversive. Has this situation not been tested? (It’s also interesting to consider positive reinforcement, where unpredictability–i.e. variable schedules–has different effect than predictability.)
In any case, thanks for the interesting article. R- still seems like the trickiest quadrant to me; one more reason to avoid it.
(PS. You still have one occurrence of “diffuse”. Not that anyone’s counting.)
Lohr spends a considerable time explaining the problems and issues associated with it. He also discusses the difference between prediction and control. Very interesting actually.
Will go back and look again. LOL
Great post! My wolf dog is a bomb defuser.
We are still looking for the abort button. We look really hard, but I realise it is me who is not looking in the right places.
Excellent article and very clever metaphor. But please – “diffuse” is not the same word as “defuse”.
Thanks- I’ll fix. Shows how much I know about bombs. Like nothing.