Mein liebsten fünf Dinge.
Top 5 Authors (no order)
- Ray Bradbury
- C. S. Lewis
- Piers Anthony
- Ursula Vernon
- Mark Rashid
Top 5 TV Shows (no order)
- Chuck
- Community
- The West Wing
- Smallville
- Neon Genesis Evangelion
Top 5 Movies (no order)
- The Last Unicorn
- Night and Day
- Hidalgo
- Unbreakable
- Brother Bear
Top 5 Bands (no order)
- Sister Hazel
- They Might be Giants
- Pink Floyd
- Trans-Siberian Orchestra
- The Patient Wild
Top 5 Solo Artists (no order)
- Nena
- Peter Schilling
- Peter Gabriel
- Alanis Morissette
- Christina Stürmer
And some comments from Facebook. Someone on a bitless grouped asked what Parelli was and what people thought of it.
One Lady's Comment: Parelli is dubbed as training by means of "negative reinforcement" ie. removing an aversive stimulus, but, in the case of Parelli type training, the aversive stimulus has to be introduced before it can be removed. Therefore you are initially applying positive punishment, in order to be in the position to give the release (negative reinforcement) when the horse offers the desired behaviour.
Some horses comply, others become very confused and anxious that they may be subjected to the introduction of the aversive stimulus at any time. Be very careful experimenting with this approach. It could seriously damage your horse's trust in you.
My response: This is not correct description of negative reinforcement versus positive punishment.
OLC: "Positive punishment is a concept used in B. F. Skinner's theory of operant conditioning. ... The goal of any type of punishment is to decrease the behavior that it follows. In the case of positive punishment, it involves presenting an unfavorable outcome or event following an undesirable behavior" e.g. the horse remains standing when cued to step back - the "standing" behaviour is "undesirable" therefore the trainer presents the "unfavourable outcome" of the rope shaking.
Me: If your goal is for the horse to spend less and less time standing still AND the horse was doing something desireable before standing still, then yes, that is an example of positive punishment.
Positive punishment does not work as a training technique if the horse is standing still and you suddenly decide you want it to do something else and start whacking it. That is abuse.
Negative reinforcement is the removal of an aversive (which does NOT have to be painful or mentally damaging) when the desired behavior is produced. Ideally the aversive stimulus starts mild and increases pressure *if ignored.*
Now, there are the definitions of the types of operant condition, but what people do and try to call them are five different things. You can't call something a punishment if it is happening seemingly at random (decide your horse shouldn't stand still after standing still for five minutes and whack him). Punishment is defined by occuring immediately after the undesired behavior presents itself (horse tries to bite you, you yell and whack it).
Simply put: punishment extinguishes behaviors, reinforcement encourages them
Second Lady Chimes In: But that's exactly what she said?
It IS positive Punishment before you release, because you add an aversive stimulus. It becomes negative Reinforcement as soon as the horse does what you want and you take away the stimulus.
Conditioning happens all the time, nonstop. So the world is based on the four operants.
Me: No, positive punishment is used to extinquish an undesired behavior, negative reinforcement is used to encourage a desired behavior. There is a difference.
Positive punishment is often at a higher intensity than you would initiate negative reinforcement. As I stated before, aversive does not equate painful. Negative reinforcement is not punishing the previous behavior, it is encouraging the next one, big difference.
SLCI: You are right, but for example if your horse bites you and you give it a slap and make a tantrum, it's positive Punishment. So when it moved away, you stop the aversive. Next time you lift your Hand, it will most likely back up. So your positive Punishment was also a negative Reinforcement.
We must always keep in mind that we are not acting in "laboratory conditions" - it's not as plain as simple as it is described in theory.
You are absolutely right that negative Reinforcement doesn't have to be painful to work - but it has to be *aversive* to work. And the horse decides about aversives and appetatives. A sensitive horse might find a stare of you aversive and a touch as HIGHLY aversive. So one could say "it's nothing bad, it's not painful" but for the sensitive horse it IS. Also, it must be "aversive enough" to make the horse move. And, the point where I personally had to Quit NHS was - the horse must know that the aversive increases.
So a green horse must learn the "cues" at first and often you need to use a lot aversive pressure. If you want your horse to go backwards, a horse that doesn't know the cues. (Referring to Parellis YoYo Game) You must most likely shake the rope very hard and immediately stop when it finally takes a step back. The horse will LEARN that this consequence will follow if it doesn't go backwards when only the finger is lifted. That's the reason why it goes backwards with a "light" cue.
I wasn't willing to go beyond phase 2 and my horse is pretty smart. So he knew that there wouldn't be consequences and started to react worse and worse to light pressure. ;)
Me: Yes it is situational to an extend. In your biting example, however, the first instance was positive punishment for sure, however that does not translate into a raised hand meaning back up. Backing up and no biting are different behaviors. Was the horse threatening to bite, the hand raised to administer punishment and the horse backed off? Still positive punishment as the action came after the initiation of the undesired behavior, the adversive was much less.
Yes, it has to be aversive. It has to be something that one (horsen human, whatever) would want to avoid. I don't like rap music, it doesn't cause me physical pain and I'm not emotionally distraught listening to it, but given the space I'd move out of hearing. Looking at a horse, not harming it, horse doesn't like the way you look at it, it moves away. This isn't negative reinforcement until you are using it to achieve another behavior.
And this is where the water gets muddied and the Parelli program often misses the mark. The aversive *IDEALLY* should not increase if the horse is responding. The horse is giving you an answer, don't yell at them for trying! Your horse reacted worse and worse because the aversive disappeared when he did so. It isn't because he found it more and more terrible, it's because acting worse caused the stimulus to go away. That is the fault in human timing.
My mother-in-law inadvertently "taught" her horse to "count" (tap the ground with a forefoot). She tried to ask him to change directions lungeing and he tried "counting" instead. She stopped swinging the stick, that must be the answer! I almost died laughing when she asked me to help her with getting him to lunge and he starts "counting" when I signal to change directions.
The other thing we have in working with our horses is it evolves to a point where we help them move better and feel better in their own bodies. The aid is there to encourage the horse into a better position that then becomes self-rewarding.
And I added: And what I was trying to say at the end there is that the aversive ultimately becomes appetitive because the horse then wants to go with that feel and seeks it out.
I just wanted to record it. This conversation is indicative of the problem I see in a lot of the bitless and positive reinforcement groups. Negative reinforcement is viewed as terrible because the scientific community uses the words "negative" and "aversive" to describe it and therefore it is terrible and evil because something you want to avoid must be emotionally destructive or painful.
Of course it isn't just the clicker training people who get the quadrants of operative conditioning wrong, but they are more prone to use emotionally charged words. The admins closed the conversation so I won't know if the people had any response to my last comments.
I got Kitt a new boot. Can you tell which one it is? She would have had two had the Scoot Boots not been too large. Tru-D had her first ride in the Scoots, which went pretty well. Anyway, couldn't afford two sets of boots so Kitt got one, Kash has one of Kitt's old ones, Chewy took Kash's smaller ones and now I have a spare. I guess I could have just gotten one boot for Kitt in the first place and saved me some money, but now I don't have to get Tru-D boots before taking her out on trail.
For snowhawk, click for full size.
- Ray Bradbury
- C. S. Lewis
- Piers Anthony
- Ursula Vernon
- Mark Rashid
Top 5 TV Shows (no order)
- Chuck
- Community
- The West Wing
- Smallville
- Neon Genesis Evangelion
Top 5 Movies (no order)
- The Last Unicorn
- Night and Day
- Hidalgo
- Unbreakable
- Brother Bear
Top 5 Bands (no order)
- Sister Hazel
- They Might be Giants
- Pink Floyd
- Trans-Siberian Orchestra
- The Patient Wild
Top 5 Solo Artists (no order)
- Nena
- Peter Schilling
- Peter Gabriel
- Alanis Morissette
- Christina Stürmer
And some comments from Facebook. Someone on a bitless grouped asked what Parelli was and what people thought of it.
One Lady's Comment: Parelli is dubbed as training by means of "negative reinforcement" ie. removing an aversive stimulus, but, in the case of Parelli type training, the aversive stimulus has to be introduced before it can be removed. Therefore you are initially applying positive punishment, in order to be in the position to give the release (negative reinforcement) when the horse offers the desired behaviour.
Some horses comply, others become very confused and anxious that they may be subjected to the introduction of the aversive stimulus at any time. Be very careful experimenting with this approach. It could seriously damage your horse's trust in you.
My response: This is not correct description of negative reinforcement versus positive punishment.
OLC: "Positive punishment is a concept used in B. F. Skinner's theory of operant conditioning. ... The goal of any type of punishment is to decrease the behavior that it follows. In the case of positive punishment, it involves presenting an unfavorable outcome or event following an undesirable behavior" e.g. the horse remains standing when cued to step back - the "standing" behaviour is "undesirable" therefore the trainer presents the "unfavourable outcome" of the rope shaking.
Me: If your goal is for the horse to spend less and less time standing still AND the horse was doing something desireable before standing still, then yes, that is an example of positive punishment.
Positive punishment does not work as a training technique if the horse is standing still and you suddenly decide you want it to do something else and start whacking it. That is abuse.
Negative reinforcement is the removal of an aversive (which does NOT have to be painful or mentally damaging) when the desired behavior is produced. Ideally the aversive stimulus starts mild and increases pressure *if ignored.*
Now, there are the definitions of the types of operant condition, but what people do and try to call them are five different things. You can't call something a punishment if it is happening seemingly at random (decide your horse shouldn't stand still after standing still for five minutes and whack him). Punishment is defined by occuring immediately after the undesired behavior presents itself (horse tries to bite you, you yell and whack it).
Simply put: punishment extinguishes behaviors, reinforcement encourages them
Second Lady Chimes In: But that's exactly what she said?
It IS positive Punishment before you release, because you add an aversive stimulus. It becomes negative Reinforcement as soon as the horse does what you want and you take away the stimulus.
Conditioning happens all the time, nonstop. So the world is based on the four operants.
Me: No, positive punishment is used to extinquish an undesired behavior, negative reinforcement is used to encourage a desired behavior. There is a difference.
Positive punishment is often at a higher intensity than you would initiate negative reinforcement. As I stated before, aversive does not equate painful. Negative reinforcement is not punishing the previous behavior, it is encouraging the next one, big difference.
SLCI: You are right, but for example if your horse bites you and you give it a slap and make a tantrum, it's positive Punishment. So when it moved away, you stop the aversive. Next time you lift your Hand, it will most likely back up. So your positive Punishment was also a negative Reinforcement.
We must always keep in mind that we are not acting in "laboratory conditions" - it's not as plain as simple as it is described in theory.
You are absolutely right that negative Reinforcement doesn't have to be painful to work - but it has to be *aversive* to work. And the horse decides about aversives and appetatives. A sensitive horse might find a stare of you aversive and a touch as HIGHLY aversive. So one could say "it's nothing bad, it's not painful" but for the sensitive horse it IS. Also, it must be "aversive enough" to make the horse move. And, the point where I personally had to Quit NHS was - the horse must know that the aversive increases.
So a green horse must learn the "cues" at first and often you need to use a lot aversive pressure. If you want your horse to go backwards, a horse that doesn't know the cues. (Referring to Parellis YoYo Game) You must most likely shake the rope very hard and immediately stop when it finally takes a step back. The horse will LEARN that this consequence will follow if it doesn't go backwards when only the finger is lifted. That's the reason why it goes backwards with a "light" cue.
I wasn't willing to go beyond phase 2 and my horse is pretty smart. So he knew that there wouldn't be consequences and started to react worse and worse to light pressure. ;)
Me: Yes it is situational to an extend. In your biting example, however, the first instance was positive punishment for sure, however that does not translate into a raised hand meaning back up. Backing up and no biting are different behaviors. Was the horse threatening to bite, the hand raised to administer punishment and the horse backed off? Still positive punishment as the action came after the initiation of the undesired behavior, the adversive was much less.
Yes, it has to be aversive. It has to be something that one (horsen human, whatever) would want to avoid. I don't like rap music, it doesn't cause me physical pain and I'm not emotionally distraught listening to it, but given the space I'd move out of hearing. Looking at a horse, not harming it, horse doesn't like the way you look at it, it moves away. This isn't negative reinforcement until you are using it to achieve another behavior.
And this is where the water gets muddied and the Parelli program often misses the mark. The aversive *IDEALLY* should not increase if the horse is responding. The horse is giving you an answer, don't yell at them for trying! Your horse reacted worse and worse because the aversive disappeared when he did so. It isn't because he found it more and more terrible, it's because acting worse caused the stimulus to go away. That is the fault in human timing.
My mother-in-law inadvertently "taught" her horse to "count" (tap the ground with a forefoot). She tried to ask him to change directions lungeing and he tried "counting" instead. She stopped swinging the stick, that must be the answer! I almost died laughing when she asked me to help her with getting him to lunge and he starts "counting" when I signal to change directions.
The other thing we have in working with our horses is it evolves to a point where we help them move better and feel better in their own bodies. The aid is there to encourage the horse into a better position that then becomes self-rewarding.
And I added: And what I was trying to say at the end there is that the aversive ultimately becomes appetitive because the horse then wants to go with that feel and seeks it out.
I just wanted to record it. This conversation is indicative of the problem I see in a lot of the bitless and positive reinforcement groups. Negative reinforcement is viewed as terrible because the scientific community uses the words "negative" and "aversive" to describe it and therefore it is terrible and evil because something you want to avoid must be emotionally destructive or painful.
Of course it isn't just the clicker training people who get the quadrants of operative conditioning wrong, but they are more prone to use emotionally charged words. The admins closed the conversation so I won't know if the people had any response to my last comments.
I got Kitt a new boot. Can you tell which one it is? She would have had two had the Scoot Boots not been too large. Tru-D had her first ride in the Scoots, which went pretty well. Anyway, couldn't afford two sets of boots so Kitt got one, Kash has one of Kitt's old ones, Chewy took Kash's smaller ones and now I have a spare. I guess I could have just gotten one boot for Kitt in the first place and saved me some money, but now I don't have to get Tru-D boots before taking her out on trail.
For snowhawk, click for full size.
