A number of recent articles claim that, “You can’t reinforce fear.” What they mean is that you cannot use positive reinforcement to create or maintain fear. This would be true. I can pay you to fake an Oscar worthy performance of being afraid of spiders. However, I cannot pay you to BE afraid of spiders.
However, you can, absolutely enable fear via negative reinforcement. Negative reinforcement is a bit challenging to understand. You can read how it works in more detail here in a previous blog post.
The Reader’s Digest condensed definition of negative reinforcement is that you stop something unpleasant in order to increase behaviour. Presumably you have to start being aversive in order to stop it. For example, you can’t turn off a car unless someone started it first. The important point being that it is the cessation of something unpleasant that has all the power.
In dog training, negative reinforcement is used a fair bit. Turning off continuous shock when a dog obeys is one example. Nagging is a classic example of negative reinforcement. Fear also works. Dogs can learn if they act calm, they can move away from something that makes them nervous.
Negative reinforcement supports a wide range of fears and phobia. For example, some people stay in their home because they are afraid of social situations. Anxiety reduces when outings are successfully avoided. Staying at home increases.
In order to survive in this world, we develop coping strategies. Not all of them are healthy. Some people might brave the world if a close friend tags along. For lack of a better word, the friend acts as a crutch or buffer. Their presence prevents the phobic person from facing the root of their fear, on their own.
These are maladaptive coping strategies. Other publications call them safety behaviours and safety signals. Maladaptive strategies create a rapid drop in anxiety, but stand in the way of addressing the primary fear. Continued practice of these maladaptive strategies makes the patient better at avoiding the real problem. Fear festers in silence.
Similar things happen to our dogs. Fear can be maintained with negative reinforcement. In order to cope, our pets can create maladaptive strategies that deceptively look like cures.
For example, nervous dogs might take a back seat to another dog in the home. So long as “big brother” takes the lead in social situations, owners fail to see any sign of fear. Anxiety festers. The dog fails to develop social skills – until one day big brother is no longer around. The poor dog is ravaged with terror.
Owners can also become a crutch by repeatedly bailing (rather than occasionally helping) dogs out of difficult situations. Dogs become dependent on the owner in an unhealthy way. Instead of learning how to handle difficult situations, they learn how to escape when thing get uncomfortable. Owners offer an easy out.
If it makes you recoil at the possibility of being an enabler, understand that negative reinforcement is powerful and subtle. The curative appearance that safety behaviours create can fool expert eyes.
Our role, as owners or coaches is to help empower our dogs. We need to give them roots and wings. This means careful self-reflection – bitter, painful, self-reflection. Are we serving the dog, teaching it to face its fear? Or are we intentionally giving the dog an escape route?
By no means am I advocating that dogs sink or swim. There is no reason to work a dog to the point of discomfort. There are plenty of options such as desensitization and counter conditioning. New advances in their execution are improving upon already impressive results.
However, what I am pointing out is that you can enable fear through negative reinforcement. You can fool yourself into thinking that you have cured said fear if you intentionally or accidentally create a safety behaviour that allows or encourages escape.
Perhaps you are fine with that, so long as your dog stops being embarrassing in public. Remember that life isn’t always fair. Owners get sick, divorced – heck they go on holidays. Dogs become lost and end up in a shelter. Like the dog that lost its “big brother”, how will your dog feel and behave when abruptly forced to face a very scary world all alone? What will your dog do when they no longer have you to turn to as a routine escape plan? What will you do when your dog is cornered unexpectedly because they are charged by something or someone scary?
Regardless of your personal decision, understand that negative reinforcement plays a role in keeping phobias alive whether you like it or not. If you’re helping a nervous dog to overcome its fears, it pays to understand how negative reinforcement works, how it maintains fear, and the risks that tag along for the ride When a quadrant has this much power, it pays to know it inside out..
**Safety behaviours and safety signals link only works in some browsers like Chrome. If the link doesn’t work, please try changing browsers.
So then in theory in order to avoid negative R from creating a maladaptive coping strategy, you must keep dog under threshold and at a distance from triggers at all times (not just when you are using CC and desensitization) otherwise you run the risk of reinforcing the escape / avoidance behaviour when you leave the presence of the trigger?
Or am I not understanding what the blog post is trying to say?
The occasional mistake is going to happen. We don’t plan for mistakes.
You can create a positive association even if over threshold. It’s just that the dog is over threshold. And that isn’t exactly nice to set out to do. But it will “work.”
In classical conditioning, long breaks with fewer reps gets you faster results. So you’ll want to do those down time breaks just because it’s more efficient.
But the more important point of the blog is we can feel better or safe escaping from things. I might feel better avoiding an ex. or spiders or clowns or crowds of people. While relief feels good, and while you would seem better in the moment, that escape is actually standing in your way of overcoming. You can’t overcome fear by avoiding. That avoiding prevents you from learning about the thing that scares you. It prevents you from realizing that you have nothing to fear and that it will do you no harm. It prevents you from learning that you got this. Hence maladaptive. Feels better in the moment. Prevents you from growing for the future.
The initial assumption in your article is false: “Presumably you have to start being aversive in order to stop it” is inaccurate. The trainer does NOT need to present an aversive to use negative reinforcement. Simply allowing the animal to leave a stressful situation runs the risk of negative reinforcement, so what would you do, require the animal to stay there in order to avoid that quadrant?
Environmental stressors abound and teaching our dogs that they can control their distance to stressors is the most humane thing we can do for them. Giving animals control over significant events is the very foundation of humane training. Controllability significantly reduces fear and also has a protective effect for the next aversive stimulus that is encountered.
Here’s an example: http://opinionator.blogs.nytimes.com/2013/04/07/for-the-anxious-avoidance-can-have-an-upside/
Ledoux? Used shock and then told people to do desensitization for a long-term strategy?
He tells people that if they are too scared to go to work, they should go shopping. Which the NY times article is entertaining, it really isn’t very reflective of his research where he shocked rats and then taught them to escape at the sound of a tone.
The act of deciding to put a dog into a rehabilitation program is making a decision for the dog. The aversive is added whether I bring the aversive in, or if I put the dog into the aversive.
Or are you suggesting that dogs live in a perpetual state of anxiety? I know abused dogs that live like that. We should get them out. I don’t know why I would ever create a state of constant anxiety. To me, that would be cruel.
What would I do? Salvage and review how I possibly screwed up. Then vow to not do it again. Intent vs mistakes. I don’t push my kid in the road to teach them to escape.
Ledoux’s team calls it “EFF” by the way. ESCAPE FROM FEAR. Called controversial, difficult to reproduce, weaker than avoidance learning, response specific and adversely motivated.
Yes, let’s please talk about Ledoux’s work.
I can’t read the full text on that link, but “Successful EFF learning also resulted in better long-term elimination of a passive fear reaction (freezing)” struck me, because in BAT, for example, the practitioner is told to keep the animal (in this case, a dog) moving to prevent freezing. If an animal attempts to freeze, according to LeDoux, that seems to indicate the animal is already afraid and that we have ALREADY missed the point in “doing no harm” in our attempt at behavior modification.
In other words, should the animal *not* have been introduced so closely to the trigger stimulus in the first place? Do humans need a more flamboyant display, such as expression of anal sacs or defecation, to confirm that something is very aversive to the dog?
Sure, if you are in a scary place, you should leave, but why should anyone deliberately put a dog in a scary place so that they can manipulate his leaving and call it preference or freedom instead of avoidance?
To me, the leash is an issue in any behavior modification attempts, either when you walk toward or away from a stimulus. Is the dog really choosing to approach an aversive stimulus because he wants to, or because he’s used to going where the human is going? Research already indicates that dogs seem to interpret human gesture well (http://psycnet.apa.org/journals/com/115/2/122/), so indicators of body position that aren’t as obvious as hand signals may also prompt directional movement in dogs.
Sans leash, some dogs will approach an aversive stimulus, either to investigate, or as part of conditioned food-seeking, or because a human is looking at the subject stimulus as well. Regardless of the reason, making an assumption that a dog approaching, or even taking food, from a person that seems scary, is not necessarily indicative that the dog has formed a positive conditioned emotional response. In fact, there are many instances of dogs biting a person after having just taken a food offering. Perhaps the whole subject of whether to use methods involving R- in such situations should be revisited, in light of LeDoux and Miklosi et al.
I understand what you’re saying Grisha. I noticed a similarity to Ledoux’s findings in a piece I saw on 60 minutes recently for the treatment of PTSD. It’s so very interesting to me, and I know a friend and I do something similar to help us move forward when our heads are attempting to avoid a situation entirely. As for my dog, he does significantly better when he is able to engage in adaptive coping.
Ledoux is very interesting. He warns that in a clinical setting his findings might lead to maladaptive coping strategies.
Might I suggest that if you want to understand research, you read research. As interesting as the NY Times, 60 Minutes and Youtube might be – taking a few sentences out of context often presents a gross misrepresentation of the actual findings.
LeDoux’s specific warning:
In considering clinical implications, it will be important to pursue further what might underlie the retardation of re-acquisition induced by our behavioral procedure because it could, in principle, result in maladaptive behaviors in some cases. Future experiments will aim to determine whether we can successfully tailor our procedure to render a once-fear inducing stimulus simply neutral, without necessarily turning it into a safety signal.
P.S. I ask not because I am trying to only understand the theory but because I question what my dog is experiencing when he appears to be “indifferent” towards things that used to send him lunging/barking. I have been doing a mixture of CC/DS with treats, and also moving away.
It’s a very – very good question that I wish more people asked. I’ll be covering a bit more on it in the next blog. Understanding exactly how these things work I think helps. So look at it from a definitions point of view.
Counter conditioning – Make a new association. Spiders = candy. Enjoy the positive experiences during the process. Leave while relaxed and happy. Presumably, you really don’t care if you leave. You might want to stay. Gradually get closer, always ENJOYING the nice thing and leaving when having an okay time.
Desensitization: Learn to relax in baby steps getting closer to the scary thing. Leave while relaxed.
Negative reinforcement. Become uncomfortable enough that you’d rather leave. You feel happy to leave. You are rewarded by leaving. You leave when you’re uncomfortable.
You can’t actually do DS/CC at the same moment you do negative reinforcement. They are very different.
The question is “how does the dog feel when they leave?”
They are so similar that people get confused. HUGE difference between negative reinforcement and conditioning/desensitization. One you want to escape. The other you want to stay.
I can’t tell the difference. Sigh.
So we are walking along. A scary person appears and is approaching. Boogie and I both see the scary person. Boogie appears to be indifferent (his body language doesn’t display any tension) I give him a treat. As soon as the scary person walks past us and is behind us. Boogie does a whiplash head turn towards me. “Mom, he’s gone. Treat?”
To me it looks like Boogie is happy when the scary person leaves. So it’s -R? (I bet if I stopped and introduced him to the scary person, he would do a quick sniff and back away) But at the same time it’s also CC because he is associating the scary person with “I get a treat”?
I realize now (based on exchange of comments) that this argument is related to BAT. My understanding of BAT is that we don’t push the dog into scary territory in order to take him away from it. The dog chooses to go forward and we stop him before he goes too far, and then wait for him to disengage when he is ready. I like BAT because when Boogie disengages I know it’s because he is emotionally, mentally ready to move on (it’s an independent choice) and not because I have treats. The “moving on” part is -R but it doesn’t feel like a bad thing to me. It’s more like, “this is boring now, let’s move on” vs “I need to escape, let’s move on”.
Lili: I just posted video of a rat trained using negative reinforcement by shock on my facebook. Promise, not a video that would be upsetting to watch, except that you see that the animal looks very calm and thoughtful. The problem with negative reinforcement is you CANNOT tell if the animal is stressed or WAS stressed during the process. Avoiding feeds good. “If I get off a yucky facebook thread, it feels good.” or “If I leave work on Friday for the weekend, escaping my boring job – it feels good.”
That is a very big problem with negative reinforcement. It’s also why you can find videos from shock trainers that use pain to teach, their dog look happy. Ever wonder why they keep pointing to their dogs saying, “My dog is happy?” When you see video the dog might actually look happy? (Not all shock trainers, some.)
So it’s a good question … how do you know? That is not a comfortable question, especially if someone has ventured into the land of negative reinforcement. And then all that talk about under threshold.
Understand that threshold changes with technique.
Counterconditioning the dog should be able to ENJOY the pleasant thing you’re giving.
Desensiziation, awareness but the dog should be able to RELAX.
Negative Reinforcement, the dog should want to ESCAPE. It’s built into the definition of negative reinforcement. You are escaping from an aversive.
If you’re in negative reinforcement, the dog isnt’ working for food. Imagine the feeling of escaping something (spider, boredom, lousy facebook thread). That’s the “reward” the dog feels. The relief.
Good idea Lili. I`m not sure we can ever truly know what someone else feels. When I`m thinking through a process I start by imagining my worst fear. I think of something that would make me blow up and lose control in the right place and time.
I imagine that someone that does not speak English and comes from a culture with different gestures and customs is going to rehab me. I can`t talk. I can`t act it out.
I think of the the person doing the rehab is green – a novice.that will make mistakes. They will not be able to ask me how I`m doing.
Then I ask myself how I would be feeling in the moment through various techniques.
Ì can tell the thing is there, but I can relax and have no desire to leave.
I can tell the thing is there and I`m loving the nice things happening right now.
I`m wanting to go…really. Now. Not blowing up. But want to go.
That`s my process. As soon as I get to R-, the Ì want to go, it would really make me feel relieved to go,`I get this knot in my stomach.
But that visualization process is just how I do it.
Yvette, coincidentally I read something yesterday that says that repeatedly practiced “avoidance” can eventually extinguish fear. I don’t totally get how this works (it might be through some other rationale & not -R that makes it work?) …your blog post makes clearer sense.
However, could you please take a look at the comment in this post -http://www.scienceofconsequences.blogspot.com/2012/08/the-joy-of-escape.html – I’d like to know your thoughts.
The blog on going to Hawaii is escape from daily drudgery? Or surviving a bomb blast in WW2 resulted in happy feelings?
Escaping irritating and traumatic things does feel good.
Does it make irritating and traumatic things feel good? Or do we want to keep avoiding them?
If I escape abuse, I don’t start loving abuse.
If my kid escapes homework, I doubt he wants or loves homework more.
If my kid creates a maladaptive safety behaviour – a way out – that prevents him from dealing with his dislike of .. math…then he just gets better at avoiding it.
In dogs, if my dog fears other dogs, we should really be careful to ask what we are REALLY teaching.
Will escape make the dog LOVE other dogs? Or will they get a very happy feeling avoiding other dogs?
Yes, escape feels damn good. That’s why dogs can work for negative reinforcement and sometimes look very happy. They are happy. They are happy escaping. That is NOT the same as learning to like or appreciate the scary thing.
Negative reinforcement has this odd side effect where it supports escape. You can’t trust your eyes and say, “That dog looks defeated.” A dog can look happy and be figuring out the best way to get out of the situation.
If you look in the comments section of the link you provided, someone says almost the same thing, but about people.
Thanks, Yvette. It was actually the COMMENT in the blog I linked to, that threw me.
quote: ” Ironically, his success in avoiding these feared objects will negatively reinforce his avoidance behavior and also prevent him from being exposed to the object long enough to extinguish his fear and realize that it is, indeed, harmless.”
The “extinguish his fear” part is what got me confused because it contradicts what you have written about LOVING ESCAPE not being the same as NOT FEARING.
Let me re-word that sentence without jargon….
Because he is good at avoiding scary things, he is getting rewarded for running away and escaping. That stops him from ever getting close enough to the scary thing and he won’t be able to face his fear or get over it or find out that it really isn’t scary at all.
Reinforce means strengthen. Running away (or walking) is being reinforced.
See scary thing…move away…..that feels good.
If you don’t let the jargon throw you off, it means exactly the same as what I wrote in the blog. That you are rewarded for avoiding. It doesn’t mean you learn to like the scary thing. 🙂
But you still learn not to be afraid of it? Sorry for being so nitpicky. I get that you don’t learn to like the scary thing, but can you learn to be indifferent?
If you learn to escape, not likely. You learn to avoid, especially if you have a safety cue that lets you get out.
See below your other comment for more info.