I have long put off writing about this topic as it is so big and important I wanted to try and get a deeper personal understanding before committing my thoughts to the blog. I am not sure I am all the way there yet, but I'll give it a go and rely on comments to clear up any mistakes I make.
To start with what am I talking about ? When we train horses we mostly use a process of negative reinforcement to get the responses we want. Don't get worried about the word negative, it does not mean anything bad. In fact from the horses point of view negative in this context is good. It refers to the removal of applied pressure when we get the desired response, so the negative refers to the taking away of pressure (see my blog Reinforcement and Reinforcement 2 for further details).
There may be some people out there who think that the application of any pressure to a horse is a bad thing and they train using only positive reinforcement where you patiently wait for the horse to offer a desired response and then reward that response with a treat. Clicker training is an example of this method, and to an extent it can work, but I don't think it is appropriate to use it exclusively for horse training. Horses are just not that tuned into treats, as we know their hierarchy of needs is "Safety, comfort, play, water, food". Thus using food as a reward in positive reinforcement is rather a long way down the list, but the comfort of removing pressure is near the top of the list and thus a stronger reward for them. Even this statement needs some qualification as for left brain introverts food can be a good motivator, but then a left brain introvert already has the first three needs covered otherwise he wouldn't be a left brain introvert !
Pressure should be provided in phases and in the Parelli program this is simplified to four discrete phases although in reality each phase of pressure can be subdivided infinitely. Thus to get the horse to back up when on the ground phases one might just be a stern look, phases two a wiggle of a finger, phase three a light wiggle of the rope, phase four a strong wiggle on the rope.
Back to the issue of timing. When we first apply a new pressure to a horse it has no idea what its response should be, thus to find comfort and have the pressure removed it will attempt a variety of already known responses in the hope that this will get the reward of pressure removal. I would like to suggest that while the horse it attempting new responses then the current phase should be maintained, but if it stops new attempts that is the time to go up a phase. If you go up phases whilst the horse is still making new attempts then in effect you are punishing the horse for trying to solve the puzzle. Go too far down this road and you have achieved a state of learned helplessness in the horse.
In the learning period of a new skill as soon as the horse even thinks about the correct response then quit the pressure and reward the horse with comfort by going to neutral or even retreating away from the horse. What do I mean "thinks the response"? Well, for a backup this may mean pointing the ears backwards, or shifting the weight back, any small sign of thinking backwards.
As we refine the response we wish to get the desired result with lighter and lighter pressure so that eventually the horse appears to be following our thoughts. (They aren't of course, our thoughts just generate small changes in our bodies that maybe even we don't notice so it feels like mind control !). At this point the changes in our phases need to change. The horse knows now to go backwards at say phase 3, but not at phase 1. We also have to ask which bit of phase one we want the response from. The answer of course is the onset of phase 1, not thirty seconds later. Thus in our minds we need to relate the timing to the onset of a phase, not sometime into it. Now evidence from some "persistence of object" trials (McGreevy) suggest that horses short term memory is something less than 10 seconds. This suggests that if you keep phase one going for more than 10 seconds then they have forgotten all about its' onset and are responding to a continuing phase one, not its beginning. Note this is attempted ONCE the horse has the basic response sorted out and is to refine that response. I do not mean to imply that if the horse is still trialling responses at whatever phase you are currently using you should blindly increase that phase in 10 seconds time.
There is an even more subtle aspect to the application of applying pressure which if mastered will really pay dividends. When I started Parelli I was very guilty of asking for a response by applying pressure just at the point that the response was actually impossible for the horse. Thus there was not a chance that it would (could) respond to the onset of pressure which in hindsight was a bit unfair. Let me use the idea of an indirect rein to illustrate. When riding an indirect rein is where we ask the horse to pick up a hind leg and step it under the belly, in front of the other hind leg. That's the simplest way I can think of explaining it ! If we ask for this move by applying phase one pressure just as the desired hind leg has landed on the ground at the beginning of a stance phase of the gait it is a bit unreasonable to ask it to pick that leg up and move it. The application of the phase of pressure should happen just as the weight is coming off that leg as it enters the swing phase and the move is physically possible. An increase in the phase should also be timed to coincide with a point in time that the move is possible, not just at an arbitrary point where we think "he should have done it by now". This timing of the phases gives the horse the best opportunity to respond to the onset of a phase rather than at sometime later. After all if the horse cannot physically even attempt the movement we cannot reward the response close to the onset of a phase but only at some point where a) the horse can physically make the move, and b) he has actually trialled it as a response. I cringe now thinking of all the times I have asked for a physically impossible response and then upped the phase too soon as the response was not forthcoming !!
As we continue to refine a response it is important to know how to quit an ask. For example if we have gone all the way to phase 3 to get the response and then the response has occurred it is tempting to give a reward by instantly quitting all pressure and hoping that the horse has remembered that phase 2 and 1 came earlier. If having got the response we drop back to phase 1 and then quit if the response continues at phase 1 pressure then we have made a stronger connection of the response to phase 1 than if we quit at phase 3. If the response fails when we drop to phase 1 then up the phases again to the required level and drop back down when it occurs. This will rapidly get the horse to associate phase 1 with the response.
I'll go into one last aspect of quitting the phases. The quit should be done on a yield, not an escape. For example, if asking for the forequarters to yield away from a stick and bag it is important to stop the pressure as the horse yields away with confidence and relaxation, not as the head shies away in fear. If you do the latter you are well on the way to developing a head shy horse. Thus if I am teaching this response I will be asking just the head to move away to start with and in addition looking for the head to move in a relaxed way. This maybe just a fraction of an inch to start with (a thought), but done with confidence before I quit. This is preferable to a full spin done by the horse out of fear as an escape. It can therefore be seen that we need to quit not just on the response, but the quality of the response. In this example we may need to follow the head of the horse around as it "escapes" waiting for that golden half second where it is a yield from understanding and confidence. I guess this is an illustration of principle 3 "communication is two or more individuals sharing and understanding an idea". If the horse is escaping from the pressure it is not understanding the idea and is communicating to you that "this is too much pressure for me to be able to think through the problem, give me a chance". The snag is that to the heavy handed horseman they will feel that on the surface they got the desired response, whereas in reality they just got an escape.
There is much much more to say on the subject, as in effect it is the essence of horsemanship. We mainly communicate to our horses through the medium of pressure and release, so the correct application on release of that pressure is what underpins the majority of our interactions with them. I hope this blog will provoke thought and comments on where I have misunderstood concepts and thus improve my own horsemanship.
To start with what am I talking about ? When we train horses we mostly use a process of negative reinforcement to get the responses we want. Don't get worried about the word negative, it does not mean anything bad. In fact from the horses point of view negative in this context is good. It refers to the removal of applied pressure when we get the desired response, so the negative refers to the taking away of pressure (see my blog Reinforcement and Reinforcement 2 for further details).
There may be some people out there who think that the application of any pressure to a horse is a bad thing and they train using only positive reinforcement where you patiently wait for the horse to offer a desired response and then reward that response with a treat. Clicker training is an example of this method, and to an extent it can work, but I don't think it is appropriate to use it exclusively for horse training. Horses are just not that tuned into treats, as we know their hierarchy of needs is "Safety, comfort, play, water, food". Thus using food as a reward in positive reinforcement is rather a long way down the list, but the comfort of removing pressure is near the top of the list and thus a stronger reward for them. Even this statement needs some qualification as for left brain introverts food can be a good motivator, but then a left brain introvert already has the first three needs covered otherwise he wouldn't be a left brain introvert !
Pressure should be provided in phases and in the Parelli program this is simplified to four discrete phases although in reality each phase of pressure can be subdivided infinitely. Thus to get the horse to back up when on the ground phases one might just be a stern look, phases two a wiggle of a finger, phase three a light wiggle of the rope, phase four a strong wiggle on the rope.
Back to the issue of timing. When we first apply a new pressure to a horse it has no idea what its response should be, thus to find comfort and have the pressure removed it will attempt a variety of already known responses in the hope that this will get the reward of pressure removal. I would like to suggest that while the horse it attempting new responses then the current phase should be maintained, but if it stops new attempts that is the time to go up a phase. If you go up phases whilst the horse is still making new attempts then in effect you are punishing the horse for trying to solve the puzzle. Go too far down this road and you have achieved a state of learned helplessness in the horse.
In the learning period of a new skill as soon as the horse even thinks about the correct response then quit the pressure and reward the horse with comfort by going to neutral or even retreating away from the horse. What do I mean "thinks the response"? Well, for a backup this may mean pointing the ears backwards, or shifting the weight back, any small sign of thinking backwards.
As we refine the response we wish to get the desired result with lighter and lighter pressure so that eventually the horse appears to be following our thoughts. (They aren't of course, our thoughts just generate small changes in our bodies that maybe even we don't notice so it feels like mind control !). At this point the changes in our phases need to change. The horse knows now to go backwards at say phase 3, but not at phase 1. We also have to ask which bit of phase one we want the response from. The answer of course is the onset of phase 1, not thirty seconds later. Thus in our minds we need to relate the timing to the onset of a phase, not sometime into it. Now evidence from some "persistence of object" trials (McGreevy) suggest that horses short term memory is something less than 10 seconds. This suggests that if you keep phase one going for more than 10 seconds then they have forgotten all about its' onset and are responding to a continuing phase one, not its beginning. Note this is attempted ONCE the horse has the basic response sorted out and is to refine that response. I do not mean to imply that if the horse is still trialling responses at whatever phase you are currently using you should blindly increase that phase in 10 seconds time.
There is an even more subtle aspect to the application of applying pressure which if mastered will really pay dividends. When I started Parelli I was very guilty of asking for a response by applying pressure just at the point that the response was actually impossible for the horse. Thus there was not a chance that it would (could) respond to the onset of pressure which in hindsight was a bit unfair. Let me use the idea of an indirect rein to illustrate. When riding an indirect rein is where we ask the horse to pick up a hind leg and step it under the belly, in front of the other hind leg. That's the simplest way I can think of explaining it ! If we ask for this move by applying phase one pressure just as the desired hind leg has landed on the ground at the beginning of a stance phase of the gait it is a bit unreasonable to ask it to pick that leg up and move it. The application of the phase of pressure should happen just as the weight is coming off that leg as it enters the swing phase and the move is physically possible. An increase in the phase should also be timed to coincide with a point in time that the move is possible, not just at an arbitrary point where we think "he should have done it by now". This timing of the phases gives the horse the best opportunity to respond to the onset of a phase rather than at sometime later. After all if the horse cannot physically even attempt the movement we cannot reward the response close to the onset of a phase but only at some point where a) the horse can physically make the move, and b) he has actually trialled it as a response. I cringe now thinking of all the times I have asked for a physically impossible response and then upped the phase too soon as the response was not forthcoming !!
As we continue to refine a response it is important to know how to quit an ask. For example if we have gone all the way to phase 3 to get the response and then the response has occurred it is tempting to give a reward by instantly quitting all pressure and hoping that the horse has remembered that phase 2 and 1 came earlier. If having got the response we drop back to phase 1 and then quit if the response continues at phase 1 pressure then we have made a stronger connection of the response to phase 1 than if we quit at phase 3. If the response fails when we drop to phase 1 then up the phases again to the required level and drop back down when it occurs. This will rapidly get the horse to associate phase 1 with the response.
I'll go into one last aspect of quitting the phases. The quit should be done on a yield, not an escape. For example, if asking for the forequarters to yield away from a stick and bag it is important to stop the pressure as the horse yields away with confidence and relaxation, not as the head shies away in fear. If you do the latter you are well on the way to developing a head shy horse. Thus if I am teaching this response I will be asking just the head to move away to start with and in addition looking for the head to move in a relaxed way. This maybe just a fraction of an inch to start with (a thought), but done with confidence before I quit. This is preferable to a full spin done by the horse out of fear as an escape. It can therefore be seen that we need to quit not just on the response, but the quality of the response. In this example we may need to follow the head of the horse around as it "escapes" waiting for that golden half second where it is a yield from understanding and confidence. I guess this is an illustration of principle 3 "communication is two or more individuals sharing and understanding an idea". If the horse is escaping from the pressure it is not understanding the idea and is communicating to you that "this is too much pressure for me to be able to think through the problem, give me a chance". The snag is that to the heavy handed horseman they will feel that on the surface they got the desired response, whereas in reality they just got an escape.
There is much much more to say on the subject, as in effect it is the essence of horsemanship. We mainly communicate to our horses through the medium of pressure and release, so the correct application on release of that pressure is what underpins the majority of our interactions with them. I hope this blog will provoke thought and comments on where I have misunderstood concepts and thus improve my own horsemanship.
No comments:
Post a Comment