Training A Horse To Steer On A Verbal Cue - Day 1
Why train a horse to steer of a verbal cue? There are many benefits, the first of which being that it creates another mode of communication between rider and the horse. Where the traditional pressure cuing system vaguely translates to "more pressure = more response", a verbal cue allows us to precisely communicate what it is that we are looking for. A verbal cue is clear, and easy to attach to the polished behavior. Although verbal cues are not permitted in competition, we can use this as a bridge to creating a clearer (and less forceful) physical cue. One that does not promote adaptation and habituation as the traditional gradient pressure cuing system does. So how is it done?
For this process I used a technique known as shaping, which involves rewarding successive approximations (or small steps) towards the finished behavior. At first I stand waiting for Cherokee to stumble upon something that slightly resembles what I am looking for. I don't add a cue right off of the bat, because I don't want her to associate the cue with some unfinished version of the behavior. So, I will work on that later. Right now my only job is to wait, watch, and capture with a click any moments in which she gets close to what I am looking for. Not knowing what I am looking for, Cherokee immediately resorts to behaviors that have a history of reinforcement, things that she has already been trained to do including her personal favorites; leg lifts, and "kiss". After a short while she realizes that those behaviors are not getting rewarded, so she tries with more effort, sometimes effort is the criteria to earn reinforcements, but still no click comes, that must not be it. So she looks to the left looking for anything that may provide a clue as to what I am looking for, *CLICK*. The click markes the moment in time that Cherokee has performed something correctly, and earned a reward. Every click muct be followed by a reward to maintain the efficacy of the tool. I feed cherokee on the left, to give her some more information about what she did correctly. I take a step back, giving her space and again. After a few more guesses, her head turn to the left becomes more deliberate. She has figured out what I am clicking for. Notice that I take frequent breaks, this allows her to process the information she is getting without getting frustrated or stressed.
I ask her to move within our training space, to ensure that she does not learn to turn towards an object, but rather to the left no matter which direction she is facing. Left is relative to her, not the environment. Proofing is the process by which we ensure that the learning is not only performed in one contextual environment, but rather becomes generalized to a variety of situations.
As her behavior becomes more deliberate I can raise my criteria, I want her to turn a certain distance before I click and reward now, so I no longer click for any small glances to the left. After she understands the new criteria through repetition and practice, I begin clicking for looks to the left the come after she "resets" by straightening out again. Once I am sure that her movements are deliberate, and that she understands what earns reinforcement and what does not, I begin to attach the verbal cue. However, I must be careful not to add it too early, I don't want to attach it to standing still, I want to ensure that she understands the correlation between the turn and the word. Otherwise the cue won't be effective. Cherokee doesn't understand what "turn" means by definition, that is a human construct, but with practice and consistency she can learn to pair the word to the behavior. To begin, I watch her closely, and say "Turn" when I see her begin to offer the behavior, moving to the left. I chose turn as my cue, because "Left" sounds too similar to "Lift" her cue for the leg lift behavior she was offering before. Over and Over we practice. In time I can test her response to the cue, does she understand the relationship?
To test the cue, I say it before she is performing the behavior, does she immediately perform the behavior? If yes, I can raise my criteria to rewarding correct responses to the cue, if not I must stay where I am.
After she passes the test a few times, I can raise my criteria once more. Now I will only reward behaviors that fit my performance criteria, and follow the cue. I will no longer provide reinforcement when the behavior is offered without first being cue'd. This is the first step to getting the behavior under stimulus control. You will likely notice that she offers the leg lifts and "kiss" behavior even when she is not signaled to perform these, they are defaults, we have not put those behaviors under stimulus control yet.
At this point, I could fade the use of the clicker, as we are past the initial acquisition phase of learning, but I must continue the reinforcement. I want to have a strong history of reinforcement before switching to a V.R.R.V. schedule of reinforcement. (Variable Ratio schedule of Reinforcement with Reinforcement Variety). That will come much later after further proofing and working in the bit-less bridle, and with a rider mounted. This will happen in days to come.
When I have a solid behavior moving to the left, I can begin to work on the right side in the same fashion. However it is important to remember that I have just heavily reinforced turns to the left side, this will make shaping to the right more difficult. You may have to lower your standards significantly from your initial starting point with the left side. You will see in the video I began clicking for approximations in which she was almost straight rather than turning to the right at the beginning. The trick to successful shaping is to start at a point where the animal can succeed. If my criteria were high, expecting a full head turn to the right for instance, she would fail, and I would become frustrated. To be successful I have to be mindful of the information that I am providing her with. Starting small allows me to give her clues as to what it is that I would like. I click for a straight head (this lets her know that I am no longer working on the left turns, we have moved on to something new). She gets that pretty quickly, so I start clicking for a slight glance to the right, now more of a glance. Onwards and upwards until we have reached our goal.
Good luck, and remember to have fun!
Contact Simply Animal Training LLC:
(847) 363 2775