I have been thinking long and hard about this subject trying to fully understand it and now, with the help of Equitation Science, I think I have got it ! I am going to use this blog to try and explain my understanding of it and maybe if I still have it wrong someone can use the comments section to correct me.
There are two types of reinforcement that are used in training animals (or humans), "piddle" and "niddle" (bear with me on these strange words). I would like to use the analogy of a banker working in a bank where he cannot directly speak the language of the bosses to get the idea across. Again bear with me on this, all MAY become clear !
So this mythical English speaking banker has arrived at his new job in a foreign land. Snag was no one told him they didn't speak a word of English, but he knows his job and so figures he can work out what the bosses want. He is a little perturbed when his dismissal notice appears on a noticeboard in the office, but relieved to note that the signature of the boss is not attached. This has placed him under some psychological discomfort and he realises that this must mean that he is not performing the job as expected. He tentatively goes about his business and tries to appease the boss by buying a few Greek debt bonds and is alarmed to see that the first two letters of the signature has been added to the notice, so rapidly sells them. One letter it tipexed out. Phew I must be headed in the right direction. He has just had his first taste of Niddle reinforcement, and whilst the notice is still there, feels a little more comfortable. The next day he scans his investment strategy possibilities and buys some subprime mortgages from the USA. Five letters of the signature appear. That can't be right and so now feeling very uncomfortable he sells them to a suitable sucker and four of the letters are suddenly removed. That feels much more comfortable and he has had his second taste of Niddle reinforcement. He also now has a good idea of what the boss does not want, but the notice is still there. He now tries, tentatively, buying a few ounces of gold for which he pays a fair price. All the signature is now removed and the signature space is blank. Ah ha, feeling much more comfortable now but relalising that more must be needed to get rid of the notice altogether he buys more gold and some platinum for good measure. The notice is removed altogether and replaced buy a nice fat bonus cheque, which not only removes all his psychological pressure but makes him feel valued and secure at the same time. He has undergone his third niddle reinforcement and received his first piddle reinforcement. He relaxes and puts his feet up. After a few days the notice returns causing a little concern, but he also realises there is a way to make the notice disappear. There must be more work to do. This time he scans the market, sells his platinum and buys some copper. Notice is gone. So this is a metals trading exchange he is in. Now he has a good notion of how to behave and it becomes easier and easier to get that notice removed from the wall (niddle reinforcement), he also gets more piddle reinforcemts in the form of bonus cheques, dinner vouchers etc.
After a while he has been doing so well that he is continually getting piddle reinforcements and comes to regard them as part of his basic salary, feeling undervalued and annoyed at his employer when he fails to receive them for doing nothing.
This causes his employer a problem. To get to the point at which niddle reinforcement again works he has to give the banker a shock to his system, remove all the piddle rewards and again place the notice on the board. He realises that he is now risking having the banker resign altogether as the banker now perceives the removal of the piddle reinforcement as a pay cut, but has to take the risk. On this occasion all is well and the banker gets back to work, but the employer realises he must not take this risk too often and gets better at balancing the niddle and piddle reinforcements.
From this it can be seen that niddle and piddle reinforcements all lie on a continuum. The middle point of the line is no reinforcement at all (the few days when he did nothing in the analogy). On either side of this "neutral point" lie two different regimes. If niddle reinforcement is on the left side of neutral and piddle on the right it should be apparent that to move along the line from left to right the banker must make and attempt at doing the correct thing and is rewarded by being made to feel more comfortable. If they are to the left of neutral this may involve the removal of some of the discomfort that it being applied. Removal of discomfort is by definition going to make him feel somewhat more comfortable.
Lets get back to reality. I've purposely used the words Piddle and Niddle to overcome the layman's disquiet of the words positive and negative. To the average person something that is "negative" has the unfortunate connotation of being bad. In the animal behaviorists dictionary this is not the case and the terms have a more defined meaning. I hope it is obvious that I am using the word Niddle for negative and Piddle for positive. So niddle reinforcement is negative reinforcement.
Niddle reinforcement occurs when an applied discomfort, such as pressure on the bit, is removed, hence the term "negative". It is used in it's mathematical sense of "to take something away" (in this case discomfort) rather than in its popular sense of "being bad".
In Parelli terms, then, let us consider the phrase "Pressure motivates, release teaches". Applying the pressure is NOT reinforcement, it is asking for the response. Thus applying discomfort is NOT negative reinforcement, subtracting the pressure when the correct response occurs IS negative reinforcement. The release of the pressure is the niddle reinforcement that teaches the correct response to the original applied pressure.
Positive reinforcement lies on the other side of neutral and involves adding something nice, like a treat, to reinforce the offered behavior. The trick, as the employer above discovered, is to carefully balance the use of piddle reinforcement and niddle reinforcement so that the neutral point stays at neutral.
If too much piddle reinforcement is given then the employers idea of a bonus can quickly be seen by the banker as a right. Thus unbeknownst to the employer the banker has a new neutral point in the reinforcement continuum and the withholding of what are bonuses, in the employers eyes, will be seen as application of pressure in the bankers eyes. The neutral point from the point of view of the banker is now in what was originally the piddle region of the scale. The snag with horse training, of course, is that it may be impossible to give the bonuses continually at the right time to keep the horse feeling that it is in the neutral spot. It could be at the end of a 45 foot line after all. (The banking community is having similar problems as we speak of course ! How to readjust the lack of a bonus to be the neutral state).
With this post I have hoped to convey that positive and negative reinforcement are in effect the same thing, but they lie on different sides of a neutral point. One of our responsibilities is to ensure that neutral point occurs in an appropriate and consistent place. I have also tried to explain that "negative reinforcement" is nothing bad, it is just a scientific term. I had someone explain the other day that they think negative reinforcement is awful and they never use it. I asked if they ever put pressure on a bit, with the leg, even with the eyes. Of course they did. This showed that they did use negative reinforcement, but just did not understand the term and it's application.
I believe that it is the popular use of the word "negative" and it's connotation of something bad, that leads people to think that it is the application of pressure that is negative reinforcement, whereas in fact it is the removal of that pressure that is negative reinforcement. To overcome this inbuilt prejudice you may find it easier to use the words Niddle and Piddle, but I suggest you do it under your breath in the presence of others
There are two types of reinforcement that are used in training animals (or humans), "piddle" and "niddle" (bear with me on these strange words). I would like to use the analogy of a banker working in a bank where he cannot directly speak the language of the bosses to get the idea across. Again bear with me on this, all MAY become clear !
So this mythical English speaking banker has arrived at his new job in a foreign land. Snag was no one told him they didn't speak a word of English, but he knows his job and so figures he can work out what the bosses want. He is a little perturbed when his dismissal notice appears on a noticeboard in the office, but relieved to note that the signature of the boss is not attached. This has placed him under some psychological discomfort and he realises that this must mean that he is not performing the job as expected. He tentatively goes about his business and tries to appease the boss by buying a few Greek debt bonds and is alarmed to see that the first two letters of the signature has been added to the notice, so rapidly sells them. One letter it tipexed out. Phew I must be headed in the right direction. He has just had his first taste of Niddle reinforcement, and whilst the notice is still there, feels a little more comfortable. The next day he scans his investment strategy possibilities and buys some subprime mortgages from the USA. Five letters of the signature appear. That can't be right and so now feeling very uncomfortable he sells them to a suitable sucker and four of the letters are suddenly removed. That feels much more comfortable and he has had his second taste of Niddle reinforcement. He also now has a good idea of what the boss does not want, but the notice is still there. He now tries, tentatively, buying a few ounces of gold for which he pays a fair price. All the signature is now removed and the signature space is blank. Ah ha, feeling much more comfortable now but relalising that more must be needed to get rid of the notice altogether he buys more gold and some platinum for good measure. The notice is removed altogether and replaced buy a nice fat bonus cheque, which not only removes all his psychological pressure but makes him feel valued and secure at the same time. He has undergone his third niddle reinforcement and received his first piddle reinforcement. He relaxes and puts his feet up. After a few days the notice returns causing a little concern, but he also realises there is a way to make the notice disappear. There must be more work to do. This time he scans the market, sells his platinum and buys some copper. Notice is gone. So this is a metals trading exchange he is in. Now he has a good notion of how to behave and it becomes easier and easier to get that notice removed from the wall (niddle reinforcement), he also gets more piddle reinforcemts in the form of bonus cheques, dinner vouchers etc.
After a while he has been doing so well that he is continually getting piddle reinforcements and comes to regard them as part of his basic salary, feeling undervalued and annoyed at his employer when he fails to receive them for doing nothing.
This causes his employer a problem. To get to the point at which niddle reinforcement again works he has to give the banker a shock to his system, remove all the piddle rewards and again place the notice on the board. He realises that he is now risking having the banker resign altogether as the banker now perceives the removal of the piddle reinforcement as a pay cut, but has to take the risk. On this occasion all is well and the banker gets back to work, but the employer realises he must not take this risk too often and gets better at balancing the niddle and piddle reinforcements.
From this it can be seen that niddle and piddle reinforcements all lie on a continuum. The middle point of the line is no reinforcement at all (the few days when he did nothing in the analogy). On either side of this "neutral point" lie two different regimes. If niddle reinforcement is on the left side of neutral and piddle on the right it should be apparent that to move along the line from left to right the banker must make and attempt at doing the correct thing and is rewarded by being made to feel more comfortable. If they are to the left of neutral this may involve the removal of some of the discomfort that it being applied. Removal of discomfort is by definition going to make him feel somewhat more comfortable.
Lets get back to reality. I've purposely used the words Piddle and Niddle to overcome the layman's disquiet of the words positive and negative. To the average person something that is "negative" has the unfortunate connotation of being bad. In the animal behaviorists dictionary this is not the case and the terms have a more defined meaning. I hope it is obvious that I am using the word Niddle for negative and Piddle for positive. So niddle reinforcement is negative reinforcement.
Niddle reinforcement occurs when an applied discomfort, such as pressure on the bit, is removed, hence the term "negative". It is used in it's mathematical sense of "to take something away" (in this case discomfort) rather than in its popular sense of "being bad".
In Parelli terms, then, let us consider the phrase "Pressure motivates, release teaches". Applying the pressure is NOT reinforcement, it is asking for the response. Thus applying discomfort is NOT negative reinforcement, subtracting the pressure when the correct response occurs IS negative reinforcement. The release of the pressure is the niddle reinforcement that teaches the correct response to the original applied pressure.
Positive reinforcement lies on the other side of neutral and involves adding something nice, like a treat, to reinforce the offered behavior. The trick, as the employer above discovered, is to carefully balance the use of piddle reinforcement and niddle reinforcement so that the neutral point stays at neutral.
If too much piddle reinforcement is given then the employers idea of a bonus can quickly be seen by the banker as a right. Thus unbeknownst to the employer the banker has a new neutral point in the reinforcement continuum and the withholding of what are bonuses, in the employers eyes, will be seen as application of pressure in the bankers eyes. The neutral point from the point of view of the banker is now in what was originally the piddle region of the scale. The snag with horse training, of course, is that it may be impossible to give the bonuses continually at the right time to keep the horse feeling that it is in the neutral spot. It could be at the end of a 45 foot line after all. (The banking community is having similar problems as we speak of course ! How to readjust the lack of a bonus to be the neutral state).
With this post I have hoped to convey that positive and negative reinforcement are in effect the same thing, but they lie on different sides of a neutral point. One of our responsibilities is to ensure that neutral point occurs in an appropriate and consistent place. I have also tried to explain that "negative reinforcement" is nothing bad, it is just a scientific term. I had someone explain the other day that they think negative reinforcement is awful and they never use it. I asked if they ever put pressure on a bit, with the leg, even with the eyes. Of course they did. This showed that they did use negative reinforcement, but just did not understand the term and it's application.
I believe that it is the popular use of the word "negative" and it's connotation of something bad, that leads people to think that it is the application of pressure that is negative reinforcement, whereas in fact it is the removal of that pressure that is negative reinforcement. To overcome this inbuilt prejudice you may find it easier to use the words Niddle and Piddle, but I suggest you do it under your breath in the presence of others
No comments:
Post a Comment