Tuesday, January 19, 2010

Nose v. Paw Target - Part Three

Last time, when I left off, I had observed that, when switching back and forth between nose and paw touches in a single session, Elo generally offered the behavior that was previously being rewarded, regardless of the command given. I had noticed though, that he seemed to always get the very first one right, and had decided to move to single-rep sessions. Well, my initial observations seemed to be a fluke. When we moved to single-rep sessions, Elo generally remembered what command I had asked for the last time we worked the target and just gave me the same one. At this point, I was ready to call it quits, for a couple reasons.

First, I was getting bored with the whole thing. Second, I've noticed that Elo does not seem to be a very verbally-inclined dog in general. He is HIGHLY sensitive to context, changes in my body position, tone of voice, position the room, etc. Numerous factors seem to affect his ability to respond to verbal commands in general, such that even after almost four months with me, he still sometimes has a hard time with "sit" and "lie down," and he often mixes up other commands that result in similar behavior (for example, rolling over v. putting his chin down on the floor). By contrast, the border collies (well, the hearing border collie anyway, though the deaf one does the same with hand signals) easily pick up verbal commands and associate them with behaviors, most of the time filtering out other contextual information, and are much better at generalizing of the signal and context are not exactly the same each time. How much more difficult, then must the nose/paw distinction be for Elo, when the target is not always in exactly the same place relative to my body or his body. When sometimes I ask for a target behavior when he is sitting, sometimes when he lying down, and sometimes when he is standing. When he sometimes looking and me and sometimes at the target when I give the command. When my tone varies. And on top of all of that, the occasional mis-timed or erroneous click when I think he's going for it with his paw, but ends up using his nose instead. Poor confused little dog!

I decided to keep working it, at least now and then. Laura posted some great tips on teaching distinguishing behaviors that I'm going to try to take into account. I've gone back to multi-command sessions, switching back and forth. Could have been luck, but last night I asked for step/touch/step touch and Elo scored 100%. In the mean time, I've learned something really important about Elo's method of processing information and his learning style that will hopefully help me improve his training and the training of future dogs.

1 comment:

  1. ugh i agree that training it was very frustrating and very boring. I felt it was a good project though to really see the differences in the way my dogs learned and train something completely. usually i free shape a trick and then never bother getting the dog really good at it on cue and in all locations/distances/positions.

    i think the nose/paw thing is extra hard for dogs too since it completely takes context and body language out of the picture and is even harder then a left vs right verbal cues since it has the added component of one object to think about.

    I hope Elo and you make some progress!