I've heard that one way to teach a dog a complicated command is to teach him a simple command then use it to teach the dog the more complicated command.

For instance: If you want to teach a dog to ring a reception bell, you teach the dog "touch" first, then use "touch" to teach the dog to ring the bell.

I've never actually seen anyone do this, and I don't understand all of the steps that would be involved. If you train your dog to touch your hand with the "touch" command, do you hold your hand close to the bell and say "touch"? Do you only reward when he actually touches the bell? How do you introduce the new verbal command? ("Ring the Bell") Is this even a real way to teach your dog new commands?

It seems to me that this would be very confusing to the dog. As a matter of fact I'm confused right now.

Maybe you are referring to chaining, or breaking the wanted action down into the smallest possible parts, only stringing (chaining) them together when each part is solid.

For example, if I were teaching the dog to fetch something, I'd use back-chaining, and if it was for play (as opposed to formal retrieve) I'd start with training the dog to "give," or to place the item in my hand, first, rather than tossing the object and "hoping for the best," as another forum member (Bob Scott) phrased it once.

Part of the reason for the back-chaining rather than forward-chaining (for me) is that since the dog learned the last step first, it's the one with the most practice and the strongest action of the chain, and it's the solid behavior that the dog is working toward when the complete command is strung together.

But either way, it's always easier to teach the little parts one at a time rather than a long complicated behavior all at once.

Was that what you were asking?

A better definition:
http://www.k9web.com/dog-faqs/activities/obedience.html#backchaining

Thanks for the response. Backchaining seems to be a great way to teach a dog a complicated command. I will use that when we get to that point.

I don't think it is the answer to my question, though. Maybe the word complicated was not the right way to describe the harder trick.

Using my example of ringing the bell and the command "touch": Teaching "touch" is easy, because the dog will likely touch your hand when you present him with it. He likely will not ring the bell, because it takes a specific action (pushing the button, with force).

I was thinking that if you teach the dog the touch command, you might somehow use that command to make it more likely that he rings the bell. I just don't know if this is the right thing to do or how to go about doing it. It seems like saying touch, but only rewarding him only for ringing the bell will confuse him.

Ah. I couldn't have gotten the question wronger if I'd tried, then!

If I have it right now, then my answer is no, I don't use the same (or even a similar) command for two different actions.

I also tend to use nouns. One noun. In your example, I'd probably use "bell."

OK, I may be off as well, but using your specific example of ringing the bell, I taught the “target” command first, then later interchanged it with bell. Let me explain.

First: “Target” was touch the end of the stick with paw. This was done over and over and with so much praise and treats that it was fun and given enthusiastically. (Please forgive the over explanation)

Second: The stick is moved to what ever it is I wanted her to touch. For the bell, obviously right next to the little depressor, or whatever form of bell you use. The command (for now “target/BELL” with the emphasis on “bell&rdquo was given and praise was given for ringing the bell. No bell ring, no praise. Not a correction just no praise. Lots of praise when it does.

Third: After many reps (dozens and dozens over many play times) the word “target” was dropped and now is just “bell.” I threw in “target/bell” randomly and, less and less, until it just faded away. To test it, after just “bell” was giving good results, I would omit the target stick once out of every five or six tries. If that worked, yeah!!! Tons of praise!!! If not, with a very easy “no” I would reintroduce the stick and give the command again. Praise when that does work.

Fourth: After “bell” was reliable with no use of the stick, I would reintroduce the “target” on the stick. The first few times I got a look like “Wait a sec, what?” but very quickly that behavior came back. After using the target stick to train multiple behaviors, it became easier and easier to wean off of the stick and still retain the “target” command.

I also taught, “shut the door/cabinet” “flush the toilet” “turn on/off the lights” etc…

Good luck, and I hope it works for you too.

I would use a clicker to isolate the behaviors you want.

http://www.leerburg.com/books.htm Scroll down to the clicker section.

We offer 3 really good books that can help you with training just about any behavior. Once your dog get the idea that you are looking for him to initiate behaviors you may be surprised what he offers you.