Normally I don't comment on the clinics I give, but I am just back from
the MOST OUTSTANDING weekend, I just had to write something to the list
about it.
Dolores Arste hosted the clinic at her farm near Galway NY which is only
about an hour from where I live. January clinics are always risky. You
never know what the weather will bring. Last year I did a January
clinic over in MA. I don't think the temperature ever climbed above 5
degrees, and that's Fahrenheit, which means it was bitter cold. We
spent most of that clinic huddled together in the tack room. We had a
great time, and actually worked horses, but we never ventured out of the
barn.
This past weekend we faired much better with the weather, but it was
still too cold to spend the entire day outside. I knew at the outset
that I needed to structure the clinic so we weren't working horses the
entire time. And that's what made it such a great clinic. The cold
gave us an excuse to stay inside and talk. Imagine the clickryder list
coming to life in your living room. That's what the clinic was like. We
had the most amazing, enthusiastic discussions! Now we know what
happens when you put eight horse-crazy clicker enthusiasts together in a
room! You get an incredible weekend!
We had a mixed bag in terms of clicker experience. Most of the people
had been clicker training for a long time, and were either clients or
had been to my clinics before. But we also had a couple of beginners.
They had seen my tapes, so we could skip past some of the basics without
leaving anyone out. On the questionnaire I sent out before the clinic
one of the common concerns people had was what to do with pushy,
in-your-face horses. That's one of the recurring questions on this list
as well, so I had decided in advance of the clinic to make emotional
control the organizing theme for the weekend.
My focus in general is on helping people to get started with clicker
training. Once you've got the basic idea and your horse is working well
on the clicker, clicker training is a blast. But if your horse is
getting frantic about the treats or mugging you with behavior, it can be
anything but fun. One of the things that I so appreciate about this
list is the willingness everyone has to help each new person who joins.
That's really why I am writing this post. It's something we talked
about at the end of the day on Sunday when we were all sitting in
Dolores' living room reluctant to have the day end. It doesn't matter
how many times here on the list we've answered questions about horses
that grab their treats, or are afraid of the clicker, or get overly
excited. Instead of the older members screaming GO READ THE ARCHIVES,
they answer the questions.
There was a movie that came out this summer called "Pay It Forward". I
didn't see it. I'd rather ride horses than watch movies, but I did love
the premise of passing on good deeds. That's what this list is all
about. As each person makes the successful transition into clicker
training, they in their turn help the next wave of newcomers. That's
something I love about this list, the willingness everyone has to help
people make a smooth transition into clicker training.
There are issues that come up with clicker training. When I talked
about this at the clinic, someone asked the obvious question: what are
those issues? The main one is enthusiasm. What do we do with horses
who truly love their work? People get distracted by the food issue, but
the real issue is what do you do with a horse who is mugging you with
behavior, who is offering you every trick he can think of to get you to
click him? We aren't used to our horses being so intent on performing
for us, or sticking to us like glue. What do we do with our
enthusiastic over-achievers? How do we help people understand this
stage where the mantra is "get behavior, get behavior, get behavior".
How do we stabilize these behaviors without resorting to punishment, so
the end result is a polite, settled, incredible horse?
This question was the organizing theme of the clinic. We had the
perfect horses for it, and the perfect distraction: snow sliding off the
arena roof. We were working on emotional control. What developed was a
wonderful format for going beyond basics. The weekend created a study
circle for the development of clicker training. Very exciting!!
Let me give you a small taste of what was covered over the weekend.
Sunday morning I talked about an exercise I did recently with Robin.
Robin has been suffering from cabin fever the last couple of weeks. Our
paddocks are closed for the winter, so the horses get very limited play
time. Robin is a horse who requires a lot of mental stimulation, and
when he doesn't get it, he invents his own games. To short circuit this
I stepped in with a game of my own. I've been turning him into a "three
hundred peck pigeon."
This refers to an experiment in variable reinforcement schedules I heard
about a couple of years ago. A scientist was looking at how you build
long duration variable schedules. She taught her pigeons to peck a bar,
and by gradually extending the number of pecks they had to give her, she
got her pigeons on a VR schedule of 300. That means that on average they
had to peck the bar 300 times before they would get reinforced. That
sounds like a tremendous amount of behavior, and it is, but it is in
fact what we want from our horses. Think about dressage riders. They
want their horses on the bit throughout an entire test. In other words
they want a "three-hundred-peck pigeon". What this researcher discovered
was the way to get there was to build the schedule slowly in very small
increments. Most riders build their schedules too fast, and in too few
steps. They jump from being delighted if their horse goes on the bit for
a step or two, to expecting it as a matter of course. The result:
resistance sets in.
In my training I generally emphasize quality of movement over duration.
That's not to say I don't build duration, but my focus is more on HOW
something is done rather than HOW LONG it is done. In this experiment I
shifted my emphasis to duration. I picked a simple, very basic behavior
to work with: walking next to me. That was my only expectation. I just
wanted Robin to take a step forward when I did. When he took a step
forward, click, I reinforced him. The way I played the game was I
increased the number of steps he had to take by one after each click. I
kept count, and I walked marching band style in a consistent, metronome
beat. If Robin fell below a certain standard, the count automatically
reset back to zero.
If you've seen my videos, you know Robin is a very sophisticated in-hand
horse. Basic leading is baby stuff for him, except that right now he's
so full of energy he can hardly contain himself. He wants to play, and
he feels very frustrated by the lack of turn-out. Plus he hates being
wrong. If he makes a mistake, he tends to get mad. He expresses both his
frustration and his anger by grabbing at the lead. That was the impetus
behind this experiment. I wanted to explore some non-traditional ways of
eliminating this problem.
Since I can't change the turn-out schedules, my goal was to be able to
have Robin come directly out of his stall and without any play time go
to work in a settled, focused, relaxed manner. That's a lot to ask of a
young horse, and especially a high-energy horse, but that's what I wanted.
Robin is usually the "shoemaker's kid who has no shoes". He gets very
little focused training time, but over Christmas some of my clients were
away, and I actually had time to work my own horses. That's when I
turned my attention to duration. It's been an interesting experiment,
and I'm continuing to reap the benefits from it.
At first it was an easy game for Robin. All he had to do was walk
forward one step, and click! I reinforced him. The next time he had to
walk forward two steps, then three, etc., baby stuff for him, except
that Robin is a very sophisticated clicker-wise horse. He was looking
for the standard of behavior I was reinforcing. So as we built duration,
he started experimenting. What was I reinforcing? He went into his
mainstay, his "pose". He arched his neck and carried himself in a
beautiful dressage-horse equilibrium. I didn't click him right away, but
kept to my count. Well, that was all right. He was used to some duration
in the behavior. So we'd go maybe ten steps, and I'd click him.
The next time I'd go eleven steps. The next time twelve. clicking him on
that 12th step, no matter what he was doing, so long as he wasn't
grabbing at his lead. If he grabbed at his lead, the count automatically
went back to 0, and I'd start again. Robin did pretty well in the early
stages of the game, but as the count started to lengthen out, he got
frustrated. He was posing. He was doing great. I should have clicked him
by now! Robin hates being wrong. He's very bright, and he's very eager.
He gets mad if he gets the wrong answer. So as my count stretched out,
Robin would get frustrated and grab at the lead. Zap! We might be one
step away from a click, but the count would go back to 0.
As I explained this exercise to the people at the clinic one of the
questions that came up was the whole issue of randomness. A variable
reinforcement schedule should be just that, variable, and yet here I was
describing a very patterned exercise, the count automatically increased
by one after each click. My pattern was set from MY perspective, but
from ROBIN'S I was being the most variable and unpredictable that I had
ever been.
I'm a sucker for quality. Normally, if Robin shows an extra bit of
brilliance, I click him. But in this game, at least at this stage, it
absolutely did not matter how gorgeous he was. If we were not yet at my
count, I did not click him.
I had begun the clinic by showing a tape of an experiment that was done
with orangutans. The tape begins with a study conducted on children
where preschoolers were given a timed test. They were told that most
children of their age could complete the test in the allotted time. Then
the researcher manipulated the time to ensure that the child either
passed or failed. What they were looking at was the body language, the
postural changes, the child exhibited when he failed.
A similar test was then presented to an orangatang. The orang was using
a language board, matching symbols to pictures for a treat. She was very
good at this, and almost always got the right answer. But then the
researcher made the test deliberately so much harder that she couldn't
help but fail. When she got the wrong answer, she exhibited a similar
posture to that of the preschooler who failed his test. She clearly had
an awareness that she had made a mistake, and it upset her.
I believe our horses show similar emotions. Our clicker-trained horses
understand the game, and they are eager to please, eager to get the
right answer. That's especially true of horses like Robin who are very
bright, and very confident. Robin hates being wrong.
This is a very important dynamic to understand. Robin expresses his
frustration by grabbing at his lead. That's the postural equivalent to
the body language the pre-schooler and the orangatang showed. The ape
flung her arm over her head. Robin structurally can't do that. Instead
he uses his mouth to express frustration.
This is NOT a question of respect, and if I addressed it as such, I
would create some major training issues. Imagine how you'd feel if
someone reprimanded you every time you got a wrong answer. Think how
willing you'd be the next time to try anything. You might shut down, or
you might get angry and act out more violently the next time.
Robin and I have had respect issues. When he was younger and his
position in his horse herd was shifting, he tried to reverse our
relative positions, as well. I dealt with that as the dominance issue
that it was. But that's not what this lead-rope-grabbing behavior
represented. For our horse's sake, it's important to understand the
difference.
Clicker training opens us up to seeing the differences and it offers us
new strategies for teaching emotional control. Robin was trying to
figure out what I wanted. When he tried his hardest to offer me things
that previously would have earned reinforcement, he became frustrated.
That's very understandable. He expressed his frustration by grabbing at
the lead. I didn't reprimand him for that. I simply regrouped, got him
settled again, and marched off as promptly as I could. I made a point of
responding as little as possible to the unwanted behavior. If we got to
our count, click, he got a treat. If he grabbed at the lead again, the
count reset. So, to go to, for example, a count of twenty steps, we
might actually walk fifty or sixty steps. The next time we might go
straight away to 21 steps and a click. But to get to 22, we might walk
80 steps or more, depending upon how many times he grabbed at the lead.
As the count lengthened we went through several interesting stages.
First, he coupled his "pose" into the game. He'd hold it longer and
longer, but when I didn't react in my usual manner to his gorgeous
posturing, he tried other things. At first grabbing the lead was one of
the things he tried, mainly because he was frustrated that something
that always worked for him, wasn't working now.
One of the advantages of this game was I was counting. I could see when
things happened. For example, around twenty-five he went through an
extinction burst centered around the "pose". That's when he decided that
the carriage I normally reinforced was not what I wanted. He gave up on
looking outstandingly gorgeous and tried other things. At forty-one it
was clear he had figured out the underlying criterion: that whatever he
was doing, he had to keep doing it a little longer each time.
Robin experimented with several different alternatives: posing, walking
with his head at chest height, dropping his head to the dirt. I kept to
my count so that all of these were at times getting reinforced. But it
was clear Robin was making choices. He started walking for longer and
longer stretches with his nose in the dirt. I had not made what he did
as we headed to our count one of the criteria. As long as he didn't grab
the lead, he could walk any way he wanted. I clicked him when he made
our count. But Robin was becoming more consistent in offering head
lowering, so I decided to incorporate that in.
Again, at the clinic one of the questions was how did I know I could add
in the head lowering at this point, and didn't that alter the structure
of the process? By and large, you know you are ready to add the next
layer to your criterion when you see it already occuring at least 70% of
the time. That's what was happening with Robin. More and more, he was
keeping his head down. If it popped up for any reason, within a couple
of steps he had it back down again. Once I saw that happening I knew I
could target it directly with the clicker.
I hadn't started out looking for head lowering. My criterion was much
more general. Robin was simply to walk next to me. But now I made nose
to the dirt also part of the game. If his nose was up when we made our
count, I would keep walking until it dropped back down. So we might go
58 steps towards a goal of 59. If on the 59th step his nose popped up, I
kept walking until it dropped back down on the 64th step, and then I'd
go a step or two beyond that before clicking him.
Our game had become even more interesting. I was not asking him to keep
his nose down the entire time, but I would not click him until the count
had been met, and his nose was to the dirt. Again, if he grabbed at the
lead, the count reset to 0. The first night we worked for a little over
an hour. We ended at a count of 125 steps with Robin keeping his nose to
the dirt the entire time. Pretty neat.
The next night I made my goal of 300. That's 300 consecutive steps where
Robin kept his nose to the dirt. Very neat.
I found in my count that after about hundred and twenty I could start
jumping up in larger increments. Instead of increasing one step at a
time, I jumped up in increments of 10 to 15 steps. Robin had certain
sticky points. 65 was one of them. At 65 the count kept having to be
reset. It was as though Robin was saying: "I get it. I know what I'm
supposed to do, but I don't want to do it that long." We got over that
hurdle and had smooth sailing until we hit 100. That was another
emotional barrier for him, as was 150, but after 150, I could build
towards 300 very quickly. He had the behavior, and had accepted the idea
of doing it for longer and longer periods. I think if I wanted a "1,000
peck-pigeon", I could get there with ease.
In general most of us don't stay with an exercise long enough to see
what it can do for us. That's certainly true for an exercise like this.
Without the goal of 300 I would have been satisfied if he'd just gone
forty or fifty steps with his nose to the dirt, but somehow I knew
getting all the way to 300 was important for Robin. I could not have
been more right. I've said many times before you always get more good
things than you expected out of a lesson when you use positive
reinforcement, and that's certainly been true of this exercise.
We met our goal and then some. Robin has been coming out of his stall
and, without any turn-out or play time, going right into work. I don't
have to lunge him, or let him run to get his energy out. Even after this
past weekend when he had even less turn-out than usual, he could come
right out and go to work. He's no longer biting at the lead rope. What I
see now is he has an automatic default behavior, dropping his nose to
the dirt. When he feels frustrated, he drops his head. That's a safety
valve we didn't have before, at least not as strongly as he know has it.
I also have a much stronger, more consistent cue for head lowering. I
wasn't working directly on stimulus control. In fact when he began
offering head lowering, I was not asking for it. But out of the
consistency of the behavior evolved a much clearer, more consistent cue.
My previous cue had become linked to some other behaviors. Now I could
get head lowering by itself without triggering the rest of the chain.
Since head lowering leads to calmness, this is an important development.
Robin gets lots of reinforcement for brilliance. He needs the balance
the head-lowering creates.
The head-lowering has made something else much better - respect of
space. Again, this isn't something we worked on directly, but he's
become much more adept at shifting out of my space when I ask him to
change sides, and that's resolved another whole layer of issues and
questions he's had. He's become an even more amazing dance partner than
he was before. Prior to this exercise I always had the feeling that he
was a little clutzy with the changes. It was like dancing with somebody
who steps on your toes. He wasn't quite managing to manuvre his big body
out of my space, and that always made him a little resentful. Now it's
just really pretty the way he can switch from one side of me to the
other without either of us ever feeling crowded. Pretty neat.
When I started with this game, I would never have thought I'd be
resolving these other issues. That's, of course, the fun of training.
You nver know all the good things you're going to get. But I did know
stretching us both to 300 would be good for his emotional development.
Robin is an athletic horse who learns things easily. Sometimes that
means too easily. He's the bright kid in class who doesn't always want
to do his homework. He knows the principle, why should he bother with
the details? Working towards 300 helped with his emotional maturity and
mental discipline.
So, how does this relate to the clinic? What I like to teach is process.
That's fundamentally more important than presenting a nuts and bolt
cookbook. When I find an exercise I like, I experiment with it with
different horses. I ask them what they think of it. That's what I did
when I first stumbled across clicker training. I liked what it did for
Peregrine, and then I just kept experimenting, and look what's evolved!!
After I worked with Robin, I experimented with Peregrine who is ten
years Robin's senior and has already worked through many of the
emotional issues that Robin is still struggling with. Duration is not an
issue with Peregrine. If I wanted to work 40 minutes between clicks I
could. But I still played the game with him, and discovered that he
loved it. With him I picked another simple criterion, keeping your ears
forward while we walked together after our ride. He thought it was a
highly entertaining game.
So now the question was what would it do for other horses, and at what
stage in a horse's training could you head for 300? That's the question
I threw out to the group. As I shared Robin's story, the wheels began to
turn. Several people saw immediately that this would indeed be a useful
exercise for their horses. One person with a barn-sour horse felt it
would help with her horse's issues. She could see walking away from the
barn, and then turning and walking directly back. She'd play the 300
peck game by increasing the number of steps she went by one after each
successful cycle. She said this would keep her from over-facing her
horse by asking him to leave too far, too fast. And that's exactly the
point. This exercise helps the handler stay focused and consistent.
Dolores saw applications for her young horse, and that's when we decided
it was time to go out to the barn. After all, you go to people for
opinions and horses for answers.
One of Dolores' goals was to have her young horse, Cadberry, stand
quietly for saddling. Cadberry is a very energetic 4 year old Morgan she
bred, and whom she intends to use for endurance riding. Energy is good,
but it needs to be channeled and connected to her. The connection part
has always been the challenge with Cadberry. When she describes him,
Dolores tells the story of the day he was born. She had to bring him in
from the field to get him out of a rainstorm. Most foals will follow
their mother, but not Cadberry. He was less than a day old, and already
he was off exploring, doing his own thing. That's his basic personality.
He's a very confident, independent horse. In many ways, he's very like
Robin.
Dolores went on to say that Cadberry is very good under saddle. Her
issues with him are all on the ground where he tends to be a very pushy,
impatient horse. Standing still would be something she'd love to see
stretched into a 300 peck-pigeon exercise. That was our goal.
The rules of shaping say: never start with your goal. Instead put as
many steps between where your horse is now and where you want to be as
you can think of. That was the lesson Cadberry illustrated for us.
In the arena we decided we were going to teach him to stand on a piece
of plywood. In a sense we were using the plywood as a target. He was to
place his left front foot on the plywood. We began by teaching him to
step on many different objects, the lid off a supplement container, a
fly mask, the plywood, etc. Once he had the idea down, "I step on these
things on the ground," we focused just on the plywood, and began to
build duration. At first standing still for just a second or two was
hard for him. All of his "respect of space issues" came into play. He
could keep his foot still, but he'd be snaking forward with his nose,
telling Dolores to clear out of his space. As his nose pushed forward
into Dolores' space, it took his feet with it.
At the heart of clicker training is learning to be a good observer and
chunking behaviors down into smaller and smaller pieces. By noticing
that the precurser to his moving his feet was his snaking his nose
forward, Dolores became more adept at checking his shift in weight. She
could ask him to reset his foot on the plywood even before he had begun
to lift it. The result was he began to truly stay back out of her space.
All the juvenile pushiness, and impatient body language, including some
ugly faces disappeared.
He became solidly anchored on the plywood. Dolores could step back from
him first to the end of the lead, and then she could drop the lead and
step progressively further and further away. Even with snow sliding off
the roof, he stayed on his plywood target. She could then pick and
choose her moment to reinforce him. Ears forward became part of the
criteria. It was neat to watch, but it was also very clear that he was
not yet at a point where we could stretch him out to 300. Cadberry
needed a high rate of reinforcement. Going prematurely to 300 would have
confused and frustrated him. Before we could build that much duration
into the base behavior, we needed to reinforce many different aspects of
the behavior.
For example, Dolores could work on stepping away from him, stepping
further away, stepping off in different directions, having both front
feet on the plywood, having both front feet square on the plywood, ears
forward, etc., etc. As she worked on each element individually, she was
keeping Cadberry on the high rate of reinforcement he needed. At the
same time, she was building duration in the base behavior. Cadberry's
feet became glued to the plywood. He was getting clicked frequently,
but since Dolores was going to him for the treats, his feet stayed
planted. He was learning that he could stay still AND keep his ears
forward to earn goodies. Snaking his nose at her, or making ugly,
"get-out-of-my-space" faces just delayed the click.
Cadberry showed everybody how you stabilize behaviors and develop the
emotional control that is such a necessary part of clicker training.
It's all well and good to have a horse joyously flinging behaviors at
you. In the initial stages of clicker training that's part of the
attraction, but it can quickly get out of control. What we focused on in
the clinic was how to stabilize behavior. How, for example, to get a
horse to step back out of your space and TO STAY back. Prior to this
Cadberry was yo yo-ing back into Dolores' space demanding attention.
Now he was learning a more appropriate, people-pleasing behavior. This
is all part of the emotional control that evolves as horses learn how to
learn via clicker training.
Because of the cold weather at the clinic we were able to explore this
concept in much greater depth than we might have done otherwise. In our
discussions we went well beyond what the horses at the clinic were able
to show us. But the horses helped us to ask many questions about VR
schedules and the building of duration. And it gave me a wonderful
opportunity to share with everyone how I approach training questions and
the whole process of exploring an idea to see what it can do for our
horses. I expect at our next gathering we'll have much more data on "300
peck pigeons".
As Dolores has posted to the list, we are going to have another of these
clinics, Feb 17-18. The format will be much the same, with time for
discussion, and time for training. One of the points that came up this
past weekend is that you can head out to the barn with the intent of
working on a particular lesson only to have the horses send you off in a
totally different direction. Our organizing theme for the next gathering
will be lesson plans. How do you find an appropriate starting place for
your horse? And if you find yourself starting at a much more basic step
than you had anticipated, how do you stay on track so that you are
building towards your ultimate goal? We'll have Dolores' horses to work
with again, and I am also planning on bringing Robin, Crackers, and our
two Icelandics. Come and join us.
