Purely Positive

[Mattweiser]

New question:

Do those of you who adherer to the basic clicker training concepts feel that a No Reward Marker (NRM) is an aversive? is it P+?

[DNeff]

Originally posted by Mattweiser
Do those of you who adherer to the basic clicker training concepts feel that a No Reward Marker (NRM) is an aversive? is it P+?

I think at some level you must consider the NRM as P+ in the technical sense. P+ means the addition of something to the situation that causes behavior to decrease.

The objective of NRM is to cause a behavior to decrease through extinction. Whether your dog views that cue as neutral information (leading to extinction) or slightly punishing (P+) would depend on the way it is said, the owners demeanor, and the dog's temperament.

IMO in order for something to lead to extinction, nothing would be added or taken from the situation. I feel that a NRM adds something to the situation and is there for "technically" P+.

Therefore a "true" NRM in the techincal sense, would be to NOT click (no verbal cue used) an incorrect behavior. Nothing is added or taken away and the dog is given information that is valuable (if they are clicker trained), which helps them understand that behavior is incorrect and causes extinction.

Dawn
U-CD Cammcastle's Mystic Riven CDX OAC OJC OGC OAJ NA TT CGC
http://members.aol.com/dplantier
http://members.aol.com/rottweilerworld

[Rott-Wiley]

I agree. This also will depend on the dog. To some dogs, a NRM means to offer another behavior or "that's not what my owner wants so I'll offer something else for the click/treat." Whereas to a dog that has had more training, the NRM could be inturpreted better as, "don't do that again cause it gets me nothing." Which is not a result the dog wants.

[Judi W]

The weakness in simply withholding the reward, is that many actions are self-rewarding above and beyond the offered alternative.

[Mattweiser]

Originally posted by Judi W
The weakness in simply withholding the reward, is that many actions are self-rewarding above and beyond the offered alternative.

I agree.. like barking, jumping, digging, dumpster diving... pretty much all of the unwanted things!! ;)

[Mattweiser]

Originally posted by DNeff
Therefore a "true" NRM in the techincal sense, would be to NOT click (no verbal cue used) an incorrect behavior. Nothing is added or taken away and the dog is given information that is valuable (if they are clicker trained), which helps them understand that behavior is incorrect and causes extinction.

That would not be a NRM, but rather P-, wouldn't it?

[DNeff]

Originally posted by Mattweiser
That would not be a NRM, but rather P-, wouldn't it?

Extinction and P- are very similar and have a similar effect on behavior. Extinction usually relies on a lack of action to change behavior and P- requires an action to take place (the removal of an item) to change behavior. You can argue not clicking is a lack of action (since technically nothing is being added or removed in the physical sense) and therefore causes behavior to extinguish or you can argue that the lack of clicking removes the opportunity for reinforcement and is therefore P-.

I think you can make a case for either theory.

Dawn
U-CD Cammcastle's Mystic Riven CDX OAC OJC OGC OAJ NA TT CGC
http://members.aol.com/dplantier
http://members.aol.com/rottweilerworld

[damp]

What does P+ and P- means???

[DNeff]

Here's a quick summary of the definitions from a previous post of mine:

Positive Reinforcement (R+): Adding something to the situation that will increase the chances of that behavior reoccurring. Giving a treat to a dog for sitting.

Negative Reinforcement (R-): Removing something from a situation that will increase the chances of that behavior reoccurring. Pinching a dog's ear until they take the dumbbell, then removing your fingers when they have the DB.

Positive Punishment (P+): Adding something to a situation that will decrease the chances of a behavior reoccurring. A collar pop.

Negative Punishment (P-): Removing something from a situation that will decrease the chances of a behavior reoccurring. Walking away from (shunning) a jumping dog.

Dawn
U-CD Cammcastle's Mystic Riven CDX OAC OJC OGC OAJ NA TT CGC
http://members.aol.com/dplantier
http://members.aol.com/rottweilerworld

[damp]

Thanks!

[Mattweiser]

Originally posted by DNeff
Extinction and P- are very similar and have a similar effect on behavior. Extinction usually relies on a lack of action to change behavior and P- requires an action to take place (the removal of an item) to change behavior. You can argue not clicking is a lack of action (since technically nothing is being added or removed in the physical sense) and therefore causes behavior to extinguish or you can argue that the lack of clicking removes the opportunity for reinforcement and is therefore P-.

I think you can make a case for either theory.

True, I think that is one of the biggest problems when discussing theoretical concepts such as these.. there is a bit of room for "point of view".... Two different trainers could go round and round about whether something is or is not "X"....

but I do want to agree with the other thread, that it is more important how a trainer makes use of their theoretical information. One of my associates uses a quote, no idea where she got it so if anyone here recognizes it let me know:

When Information is Combined With Content and Experience it Becomes Knowledge

[Beckysmom]

So, there is classical conditioning.

And there there is operant conditioning with a 4 quadrant box... positive reinforcement, negative reinforcement, positive punishment, negative punishment.

And while all of that seems to fit in the toolbox of training, as it were, there also seem to be things beyond.

Where for example would "yawning" as a calming signal fit into the behaviourists concepts? Wait I mean Behaviourists with a big B, I think. Because it certainly is a tool that also works in some situations. In certain situations with my dog it is working better than any rewards or punishments.

[LynnS]

Hi Kate

Can you elaborate on that (in what situation giving a calming signal works better that P+,P- etc..? I am guessing this is more related to a behavioral issue as opposed to an obedience issue?

[Beckysmom]

Originally posted by LynnS
Hi Kate

Can you elaborate on that (in what situation giving a calming signal works better that P+,P- etc..? I am guessing this is more related to a behavioral issue as opposed to an obedience issue?

Okay begging forgiveness in advance if I have my letters and plus minuses etc. mixed up...

First off, I think behavioural issues and obedience issues are very closely tied. It is not much help to say my dog is extremely obedient --but only so long as we are in our own living room with no cats, squirrells, other dogs, people, or trucks around. etc.

Yawning etc. seems to work better for him to get him to ignore certain distractions at a level for him that neither things like treats and toys, nor corrections will touch, so that he can focus on and carry out whatever the obedience matter at hand is.

More specific example, neither high level treats nor punishment (collar corrections) would get him to hold a stay around our three new kittens when they were carrying on which included leapfrogging (short direction) the dog's back etc. Yawning--whatever it may be or not be on the whole operant conditioning conceptual framework-- did work.

I suppose it is possible to construe yawning as a calming signal to be both a lure and reinforcer for the specific behaviour of ignoring distraction that in certain particular circumumstances is more meaningful to the particular dog than bits of roast beef...just as in certain specific circumstances tug toy is a better +R than roast beef. Perhaps specifically it is a reinforcer that works where calm is needed and building yet more drive as with a toy is not desirable. But I think calling it a reinforcer is something of a stretch, and that it really is in a realm that is outside of the operant conditioning world.

Having found to work so well with cats (also chickens--similar but not as tough because they are not as interesting as kittens)--a comparatively lower distraction for this dog, I am now also starting to experiment with yawn+"leave it" + other command as fits circumstances (eg come, stop, sit, heel etc.) with regard to truck chasing. (If it works on that I will consider it a miracle and will let you know if you like.)

I cannot yet report on the outcome. We've got a good result at 100 feet from trucks--which is already an improvement--, but the real test will not be till we are on a one lane road with minimal shoulders and we can get a calm sit as far off side as we can get while a big fast noisy truck passes by.

Positive rewards were not working at all--the last dog I had from puupyhood and from puppyhood she learned using +R that a truck means to get to edge of road and sit still. But this one was a rescue with ingrained truck chasing and truck confronting habit (he was found chasing trucks on the interstate). --probably already classically conditioned and reinforced with adrenalin rush which was stronger --apparently--than operative conditioning to overcome in lasting reliable way. The adrenalin rush made the behaviours self reinforcing beyond any positive reinforcer I had to offer.

Punishments +P (ecollar, pinch collar) could keep us from having a fatality--which in the circumstances puts it well above +R as a dead dog and handler are precluded from further training progress, as well as other activities. But so far as I could tell were only working as a control device, not actually training the behaviour out of existence.

Once classically conditioned that a particular thing brings on a particular strong visceral response... be that bell to salivating, or truck to adrenalin rush, I think the OC toolbox becomes lame to deal with it. So that at least for us dealing with trying to rehabilitate dogs that did not start out on a good footing, OC can take us far, but not always all the way we need to get.

I think Probably because once again, for my dog in the truck situation while managed by P+ and/or R+, the dog still gets his adrenalin rush every time truck went by even if he did not actually chase it that time. For some probably not great analogies but ones that come to me in this moment, I think he may have even been building up desire to chase a soon as opportunity permitted, almost like a person on diet might be building up a craving for chocolate cake. Or even like a person who has gotten sick from a bad food of a type they love, will, notwithstanding the P+, go back to it when they are better. Or like, despite the pain of childbirth a P+ for some women though not all, other aspects of the situation of having children (R+es) will not make the one child the last.) We discussed that some behaviours may be so self reinforcing that they are hard to match /override with a R+ treat. I think that is also true for some behaviours being so self reinforcing that they are hard to touch /override in a lasting way with a P+. (though by definition a P+ would be something that would make the behaviour be offered less, such that technically if the behaviour isn't modified, offered less, eventually to to the point of extinction, then the conditioner was not a strong enough stimulus to be a P+--but if the only thing strong enough to be a real P+ would be that the truck hits the dog and renders it unable to chase ever again, then there is a basic problem with that definition in its absolute sense as opposed to real world usefulness. Sorry I didn't manage to express that more clearly because I think it is a really important point and basic defect with the OC terminology. I thought this might be the sort of thing that Mattweiser was actually considering in bringing the OC article up for discussion in the first place ) Anyway, in terms of real life, I think there may be situations beyond the OC four quadrant toolbox ... and also tools beyond that toolbox.


Kate