Inspired by Maire and Ben, I decided to have another go at some clicker "training." I put the word in quotes as it's a rather fancy-sounding word, implying all kinds of fancy results. In my case, it would be more accurate to describe it as clicker-messing-about.
Anyway, I loaded up with some chopped apple and headed out to find George. We've done a little of this in the past, but for one reason or another, I haven't pursued it. Today, I practiced having him be polite-while-knowing-I-have-treats - and then clicking and treating.
After a while, I started working with backing up. George will back up up when you lean forward, point at him, and whisper, "Back." So I just added the click-and-treat. But George doesn't like backing up very much, and after a treat or two, he'd get a bit sad and say, "Well, it's just not really worth it" and wander off.
This was ok, because 1) he learned that he could leave when he wanted, and 2) when George walked off, Bridget would present herself, saying, "I don't know what's the matter with him - I'll walk backwards all day for you if you keep the apple coming." But of course Bridget is a princess and can do silly things for the humans without losing self-esteem. Noblesse oblige.
Then I thought - why should we work on a behavior which goes against the grain? Why not work toward establishing an understanding of clickering by working on an enjoyable behavior?
So I worked on kissing to George and having him walk towards me. This went much better, as he didn't get sick of it. Sometimes he came over and parked me near his hip (presumably hoping for a scratch), but I wouldn't give him the treat until his head was somewhere near me. He stayed relaxed and engaged.
I wouldn't normally want to keep asking him to do the same thing over and over just to practice. It's not so much that the treats function as bribery - more like they make it into a game, so that it's not boring or demeaning to keep on repeating the same exercise. And I think repetition could be helpful in order to establish a certain reliability of response to cues, rather than always relying on the moment-by-moment mutual responsiveness which we have at present. I am grateful for this - but it might be nice to be able to say "ho" and have the horse reliably stop.