Clicker Training Discussion Thread: Capturing, Shaping, Cue's and Stimulus Control

I guess being a less seasoned aversive trainer I was under the assumption (but I really should know better because we all know what happens in those circumstances) that the point of employing aversives was to stop a particular "incorrect" response/ behaviour. Dog breaks a stay- a correction is employed to stop the breaking (of course only used after the dog understands the behaviour). Dog is cued to come, dog doesn't respond, e-collar let's off shock which stops when dog returns- in this case the behaviour of not coming is being corrected and will be turned off when the dog reaches it's handler.

I know it all plays into the four quadrants of op conditioning, adding something positive, taking something positive away, adding something negative, taking something negative away. My core understand was however that aversives are generally used to either suppress undesired responses or train desired responses by taking the aversive away when the desired behaviour is performed (like the example of using an e-collar to teach an offlead recall).

I'm sure there are many sport leader's/ service dog/ SAR/ Police dog trainers who still use aversives in some form, but there are still many trainer's in leading fields who stick to the removing something positive ( access to reinforcement) as the only consequence for an incorrect response.

Some that I can think of off the top of my head include:

Agility trainer Susan Garret
Police Dog Trainer Steve White who was the head trainer for a long time of the Seattle Police Department ( First police department to switch to clicker training- their success rates sky rocketed post switch)
Cecile and Mortan- Two Top Shutzhund and Obidience competitors from Sweden ( I attended a lecture in Rhode Island by them and it was SO neat to see their Mal who was exclusevly trained via clicker training and withholding reinforcement for mistakes work- I have NEVER witnessed a mal be so concious and controlled in drive- WAY TOOO COOL!
Freedom Dogs- US based service dog trainer's for war vetrans
Michelle Pouilet of the San Rafeal Dog Guides for the blind ( one of the oldest service dog schools to switch to clicker training- they now have over five years of hard data showing how their success rates have sky rocketed)

Sorry Carrie I kept meaning to address this but kept forgetting. I 100% agree that there is a growing number of individuals in the scientific community and trainers from every walk of life in every single field who share my particular view.

Equally so though, there are just as many, if not more trainer's from every walk of life, in every field that are just as happy to stick to every combination of methods, with aversives mixed in one way or another.

It has been my general observation that training is akin to religion for most seasoned trainers thus the rapid rate of downward spiral that occurs in conversation's between people or person's from different training backgrounds. To be honest it's all very unproductive and it is my hope that everyone on Chaz feels comfortable enough to join in on this disscussion regardless of their personal idea's and methods without being attacked, or made to feel inferior/ be littled or abusive.

So what I ment to say, was everything I have posted is simply my take and opinion of marker based, aversive free training. However like Carrie pointed out I am not alone in this particular view point that however does not make me right and other's wrong- because until we, in the training community can get past the your wrong I'm right mentality open minded conversations on methods and the underlying principles which govern their effectiveness, and their drawbacks will never occur.

To me that would be a horrible thing, because we all have something to learn from each other, even if we choose not to dabble in another's methods we should all remember that most of us, at some time or another have started out using aversives and before lashing out at other's for it, one should take a good look at their training history.

The discussion is turning out to be great,

Thanks guys
Kayla

:hail: Well said.

So here's a question: What's an aversive?

To me, aversives are positive punishment techniques; the term implies physical discomfort. I understand that negative punishment (like removing attention, time-outs, etc.) can be very uncomfortable for dogs, but are they aversive? Do they cause physical discomfort to the dogs? Not really.... It's more of an emotional discomfort (which, of course, could have physical side-effects - they get upset stomachs, possibly headaches, etc.).

Do I train with NO aversives, even by my definition? Can't say that I do. But I use very MINIMAL aversives. Just today, I was working with a dog who was obnoxiously jumping up on me.... This is a dog who's been in training long enough to know that jumping is not going to get my attention, sitting is, and still she repeatedly jumped on me. I will have no trouble sleeping at night with the knowledge that I did in fact knee her in the chest to punish the jumping.... and, given the training history I have with this dog, I'm also confident it did not hurt our relationship nor will make her more sluggish or shut-down in training.

But I would NOT use aversives with dogs I have just begun to train; I would NOT use aversives unless the dog was doing a behavior that I would like to see completely extinguished (like, if he sat when I cued down, I would not use an aversive to punish the sit); and I always strive to use Karen Pryor's 10 (?) rules of using effective punishers.

Thoughts??

I consider P- to still be an aversive. Reading through some other articles on Karen Pryor's website, it sounds as if she also considers P- to be aversive and that she doesn't use any punishment ~ positive or negative? Kayla, maybe you would have more insight into that, having gone to her school?

The Myth of "Purely Positive" | Karen Pryor Clickertraining

I then read some articles from other sources that define aversive as P+ only - as being a physical unpleasant stimuli.

So, maybe I don't use the term correctly? Maybe others use it incorrectly or I've been misunderstanding them all these years? :dunno: I'm here to learn.

Great question Lizzy.

I know we all probably refer to it differently so maybe for clarity sake we really should just refer to the four quadrants of operant conditioning.

(I always manage to get these mixed up but bare with me and let me know If I've screwed up)

-You can add something positive ( Positive Reinforcement R+)
-You can take something positive away (Negative Punishment P-)
-You can add something negative ( Positive Punishment P+)
-You can take something negative away (Negative Reinforcement (R-)

For me an aversive is either adding something negative (P+) or taking something negative away (R-), now like anything else the dog will truely determine what is an aversive event however I think as I see these two quadrants applied in general within the dog training community I would classify these as aversives because they are usually applied in a way to cause pain/ stop a behaviour/ shape a new behaviour by ending the pain when a behaviour occurs.

Now do these things need to be physical in nature to be aversive- I would again say it would depend on the animal. I know horse whisper's, among other techniques use a marker signal combined with removing something negative (themselfes) to shape wild horses to accept their presence and then a halter and saddle, and in "Don't Shoot the dog" there is a story about using Negative Reinforcement to shape a llama to accept handling.

Again though, taking something positive away (P-) can be equally aversive to an animal, like CP mentioned it is to Tyr. So again the animal should be obsereved to determine it's particular definition of an aversive event.

What we were taught in the Karen Pryor Academy course was the basic and advanced and elegant (though I haven't even had a chance to try out some of these yet- things like using stimulus control to extinguish nusiance annoying behaviours or using a limited hold in which an animal must perform x behaviour within a certain time frame to recieve reinforcement- again only after its been shaped to fluency and on cue) uses of clicker training.

While adding something negative and taking something negative away are obviously not a part of the course, for good reason IMO, there is more leeway on things like NRM, and taking away something positive. The bottom line being that these things are not needed in the first and middle chunk of shaping, getting the behaviour on cue, and then shaping to fluency before switching to a performance cue to seal in the more advanced things you've shaped in (your distractions, the precision of the behaviour, the speed in which it is performed and the distance you can be from the dog when cueing,etc).

Poisoned cues are a big part of the course as well, and it really only takes one or two aversive events in association with a certain behaviour/ cue to have it poisoned and at which point, especially for those of us who train for competition, you do end up needing to re-shape and add a new cue.

The last thing is that since the course is geared towards teaching pet dog owners, the use of things like taking something positive away is not recommended because of the frequency and hasheness that novice pet trainers tend to use any sort of technique. (I'm sure many trainers who use physical corrections or just verbal correction's in their classes see beginner owners abuse and over use these at the inappropriate time).


We all are, the day we close our minds off to the possibility of new things in the training community is the day the field stops moving forward period.


Ouu lastly I wanted to invite everyone to post some of their favourite training quotes. I have tons but I think my favourite for crossover clients who are new to clicker training is "Positive does not mean permissive" by Susan Garret.

Kayla

That is one of my favourites too Kayla!!!

Whoa, Kayla, I'm having a REALLY hard time understanding that last post.... Let me clarify some of the language you are using:

In our vernacualr (regular, every-day speech) "positive" means good, "negative" means bad.

BUT, when you're talking about operant conditioning, these are the definitions:
"positive" = "giving,"
"negative" = "taking away,"
"reinforcement" = "to increase the likelyhood that the behavior will happen again"
"punishment" = "to decrease the likelyhood that the behavior will happen again"

So, Positive Reinforcement means giving something to increase the likelyhood that the behavior will happen again. Negative reinforcement means taking something away that will increase the likelyhood that the behavior will happen again.

I know that you know what they all mean, but it just gets very confusing when discussing this with other people, if you use "positive" to mean giving AND good. That also confuses the public.... I talk to people a lot about the dangers of positive punishment, and they have a hard time wrapping their minds around the terms - how is punishment good?? And why is that so bad??

One of my old bosses repeatedly told me about how he talked to potential clients about my training methods, telling them "She uses positive reinforcement, not negative punishment!" when technically, this is not true.

It's NEVER recommended? What about when a dog jumps? When he pulls on the leash? When he's demand barking? When he bites your hand while playing tug?


ETA: One of my favorite quotes. When talking about the dangers of teaching people punishment methods: "When you give a man a hammer, everything he sees is a nail."

It's not that it's never recomended, I should have been more specific, it's more that their is a stronger focus on using R+ to teach incompatiable behaviours (for the example of jumping/ pulling on leash). Demand barking and nipping during tug, yes I too would withdraw attention.

Love this quote Lizzy.

I would never use the four quadrants in day to day conversation with a client I just thought for the purposes of clarifying what an aversive is it would be useful in this conversation. I still prefer to just stick with, either adding or taking away something negative or positive as I agree it gets way too confusing otherwise.

Hopefully that's slightly more clear?

Why don't you just say "good or bad" instead of "negative or positive"??

Most of my clients have heard of "positive reinforcement" training, though they don't understand exactly what those terms mean. If I tell them that we're going to be using "positive reinforcement training," I'm not going to turn around and tell them that negative = bad. Of course I don't get into a four-quadrants discussion with them, but if I want to talk about what we don't do, I'll just say that we're not going to use corrections or aversives and leave it at that.

I've had several clients, too, who have degrees in psychology; they do appreciate it when I don't underestimate them and do use the terms correctly.

Personally, I would just leave "good" or "bad" out of it all together. It doesn't really matter what it is. Don't know if it is helpful or not, but here's how I explain it:

Positive means adding something.

Negative means subracting something.

Reinforcement is for increasing the behavior.

Punishment is for decreasing a behavior.

Soooo.....when you add something in order to increase a behavior (or cause the dog to tend to repeat it in the future) it is positive reinforcement.


If you take something away in order to increase behavior, it's negative reinforcement.

If you add something to stop a behavior, it's positive punishment.

If you subtract something to stop a behavior it's negative punishment.

If the thing you add or subtract does not alter the behavior, it has missed the mark and can not be called a reinforcer or a punisher.

Thanks for the clarification Carrie and Lizzy, I always find it difficult to remember the technical terms but thinking of it the way you've described is helpful.

Kayla

Glad it helped.