Negative Reinforcement does not mean Punishment

[GlassOnion]

I wanted to share this blog entry with you guys, I found it quite informative-

http://bestinflock.wordpress.com/2011/0 ... ame-thing/

[Cage Cleaner]

That's true, actually. I had completely forgotten that.

Negative reinforcement is still "reinforcement" of a behavior. Opposed to positive reinforcement which is adding something to reinforce behavior. Negative reinforcement is taking away a stimulus, which then is perceived as a reward. E.g. when giving a riding aid to a horse (leg pressure, for example), and removing it when the horse responds correctly.

[evaleen]

but isnt it still harmful? i mean, wont the bird learn to associate the training sessions with negative stimuluses?

[Cage Cleaner]

No, b/c it's a removal of a stimulus, not a negative one. E.g. You ask a bird continuously to "come here". When he does it, you stop. If he realizes that you did because of what he did, then that's negative reinforcement.

[Shelby]

but isnt it still harmful? i mean, wont the bird learn to associate the training sessions with negative stimuluses?

No. Negative means that something is being taken away. Positive means something is being added.
Watch this video:
http://www.youtube.com/watch?v=nuERdlr2BZg
This is a video demonstrating a *negative reinforcement* type of training. These birds don't like being approached by hands. They act defensively when a hand/person comes near them. When they calm down, they are rewarded: the hand that they don't like is removed. These birds *like* having that hand go away! Eventually, they are ok with being touched, because they know the hand will go away and leave them alone after a little while. Remember that removal of a stimulus is "negative" in training terminology.

[Michael]

No, b/c it's a removal of a stimulus, not a negative one. E.g. You ask a bird continuously to "come here". When he does it, you stop. If he realizes that you did because of what he did, then that's negative reinforcement.

The problem is where the aversive condition is coming from. The reason why negative reinforcement could be bad for training is that if the aversive condition is being purposefully fabricated for the sake of taking away for reinforcement, it can reflect poorly on the trainer. For example, pulling a parrot's wings open and holding them there until the parrot stops resisting and opens them himself. The negative reinforcement is that you'll stop tugging on its wings if it open them itself. The problem with this is that the motivation for opening its wings is to avoid the discomfort of you doing it rather than something positive for it (like a treat). The parrot can learn to avoid the human all together rather than be subject to the discomfort which gets taken away for doing what it wants.

Negative reinforcement is further ineffective because of desensitization. Over time the parrot will get used to the aversive and taking it away won't be so effective. In most cases where I use/recommend negative reinforcement is in the case of desensitization. For example if a parrot is scared of a harness, I present the harness to the closest distance possible before it panics, hold it there, and then take it away as a reward for tolerating it. However, I virtually always couple positive reinforcement as well with it because it is more effective.

There are very few cases in which negative reinforcement is effective or advisable for parrot training but it is not inherently a bad thing.

On the other hand our parrots are very good at applying negative reinforcement to their owners. They will bite or scream nonstop until their owner stops what they are doing and gives them attention or toys. The parrot stops screaming/biting which negatively reinforces the owner to give attention at the onset of biting/screaming.

[GlassOnion]

Michael, are you sure the screaming/reward example is a valid example for negative reinforcement? It seems to fall into the 'reward' category, or the 'positive punishment' category from how I see it.

[Cage Cleaner]

Michael, are you sure the screaming/reward example is a valid example for negative reinforcement? It seems to fall into the 'reward' category, or the 'positive punishment' category from how I see it.

The screaming example is an example of negative reinforcement, yes. It is also a reward, because it's meant to encourage behavior. It's not positive punishment. I'll try to explain:

This is negative reinforcement: Bird screams, screams, screams, you run over, take it out of the cage, bird stops immediately.

This is positive punishment: you put your attention away from the bird to a friend in the same room, the bird takes a chunk out of your ear as soon as you do so.

In the first case, the stimulus (screaming) is applied first, and taken away as a reward. Negative -reinforcement- is indeed a reward category, because it is used to -reinforce- behavior. The reward for rushing over in response to the screaming and taking it out of its cage is the removal of the horrible stimulus (screaming).

In the second case, the punishment is applied after the unwanted behavior occurs. Punishment (biting) occurs after the unwanted behavior happens, as a result of it, and not before. It's a consequence of the unwanted behavior, which in this case is taking attention away from the bird.

I hope that explains it. If not, let me know.

Ignore the terms I use in my other, completely unrelated, tangent thread about horse training. I had forgotten about the techinical term of negative reinforcement when I was writing that. I meant positive punishment in that context.

[Michael]

Beat me to it. Good examples. Point is that it's much easier for our parrots to use negative reinforcement on us than on them. They are free to abuse us all they want and we can't do anything about it. But you just dare use something too aversive on them and they won't come to you for a week.

[Jamie Edie]

Negative reinforcement is further ineffective because of desensitization. Over time the parrot will get used to the aversive and taking it away won't be so effective.

Satiation works the same when using a positive reinforcement approach using food, praise, or physical affection to reinforce behavior. Does this mean positive reinforcement is not effective? Certainly not.

IMO, there is far too much baggage and confusion associated with any training approach that incorporates anything other than "positive reinforcement" to shape behavior, and this baggage stems from ignorance, or downright deceit, at all levels, from private owner to "professional" trainers.

Using the positive reinforcement approach popularized today, there comes a point in time real quick when the animal realizes there is food involved in the behavior equation, and it EXPECTS that food. When the animal performs a behavior that does not meet your criteria, that food is taken away, which is NEGATIVE PUNISHMENT in accordance with operant conditioning principals. You are, in essence, taking away (negative) a stimulus (food) to decrease behavior (punishment for wrong behavior). Animals are smart, especially parrots, they know what is at stake during training after you establish a precedence.

People are either mislead by "trainers" trying to sell a product, or ignorance, into believing that "negative" and "punishment" have a draconian context when it comes to shaping behaviors, and that is simply NOT the case. You either add or subtract something from the equation to either discourage or reinforce behaviors. Period.

When a pup bites too hard on momma's teet while feeding, momma nips at the pup, and the pup learns not to apply so much force. When a bird uses too much force preening another, it gets nipped as well. Both are examples of positive punishment in accordance with operant conditioning practices, and both approaches have been effective in shaping behaviors since.....well, forever.

There are FOUR quadrants in operant conditioning, and a good trainer needs to learn how, and when, to employ ALL of them.

Best to you and yours.