If you're so smart, why ain't you fat?
August 2, 2011 10:59 PM   Subscribe

My dog Duffy is very smart and even though he's pretty old he learns new things fast, so I am doing an experiment with him and I am finding his failure to pick up the trigger puzzling. He discerns the most subtle cues from others when he's trying to decide what to do next, so I wonder how this isn't working.

Duffy is a border collie. He learns things by a process of observation that is extremely subtle - for example, the way my partner puts down her wine glass (not her fork, not her napkin, her wine glass - we have checked) is his cue to think we are done at the table and he gets ready to be told "OK" when he can come off his down-stay.

Lately I have been trying to see if he will pick up a simple action of his own as the cue. When I have a snack I am not at the table, and he will try to get a bit of it by looking at me, the snack, me, the snack.... Therefore I started delivering a bit of the snack the instant he sat. And he does not seem to be aware that -his- action is what's governing snack-delivery.

Is the behavior idiosyncratic? If I -tell- him to sit, he does it, but if I deliver snack only -after- he sits of his own accord, he doesn't seem to get the linkage. He has even sat, had bits of snack, then stood up and got no more and his behavior isn't changing.

posted by jet_silver to Pets & Animals (11 answers total) 2 users marked this as a favorite
Try making a special sound when he sits. Then stop making it after you're sure he's caught on.
posted by vorfeed at 11:19 PM on August 2, 2011 [1 favorite]

I agree dogs, especially border collies, are really sound responsive, are you sure it's not the sound of the glass? Dogs are very observant but sounds are so much more intense to them than they are to us, which is why using different whistles to signal directions is so effective. I have no idea what sound could be associated with snack though.
posted by boobjob at 12:02 AM on August 3, 2011

Best answer: I can't answer why, but I can answer that his behavior is idiosyncratic if my dog is any indicator. She is a black lab and if she wants a treat that I have, she will run through her repertoire to try to get it, without any prompt from me. She initially tries sitting, and then giving me her paw (or pawing in my general direction). If I go so far as to say "sit", she'll get so excited she sits, paws, lies down, plays dead, and rolls over - all in a row, without being asked. She's a little bit clever... but not very. Also, I've known many dogs who beg by doing things spontaneously that they think their target will respond to, like giving a paw, or putting their face on your leg and giving you mournful eyes.
posted by treehorn+bunny at 12:35 AM on August 3, 2011 [7 favorites]

Best answer: When I have a snack I am not at the table, and he will try to get a bit of it by looking at me, the snack, me, the snack....

Does this ever work? Do you ever give him a treat just because he's looking alternately at you and your food? If you do, then his behavior (looking at you and the food) is already cuing a treat -- just not the behavior (sitting) that you're focused on. If you don't ever treat him for looking at you and the food, maybe part of the "problem" is that he expects to work for treats (i.e. do something you've told him to do), and you working for him, i.e. giving in to demand behaviors by treating him when he's done something unprompted, feels especially foreign.

Beyond that, what you're asking him to learn is that, when you are eating but not sitting at the dining table, he can cue you to give him some of your food by sitting. That's a fairly subtle pattern for him to pick up on. If he sits while you're eating at the table, no treat. If he sits while you're not at the table and not eating, no treat. If he sits while you're at the table but not eating, no treat.

I'm sure Duffy can learn this if you really want him to. You could have a snack away from the table, tell him to sit and then treat him for sitting, several times in rapid succession. He'll soon start to anticipate having to sit, and do so unprompted, which you can then reinforce. That said, I don't think it's a great idea to train a behavior that has you responding to your dog's cues.
posted by jon1270 at 3:49 AM on August 3, 2011

Best answer: I recently read Inside of a Dog which is an interesting, if fairly dry, read about how dogs' minds work. In the book the author really emphasizes how tuned into us domesticated dogs are - it's essentially what separates them from wild canines. Dogs learn at a very young age to use people as "tools" to get what they want. I'm sure I'm not doing a very good job explaining it, but I'm no dog scientist.

Anyway, if I were to venture a guess why your treat-sitting trick isn't working, it's because Duffy is looking to YOU for cues on what to do and not observing his own behavior. It's the same reason he picks up on your partner's wine glass cue.

If your ultimate goal is to have him sit when he wants a treat, that's easy. I know it's easy because I taught my dumb beagle to do it. :) When I was training her to sit (a long, arduous process), as soon as her butt hit the floor I would give her a treat. Once she figured out what "sit" meant, I made her sit before getting any of her treats. Now, whenever she wants some food item I have she sits (unprompted) to beg (very cutely).
posted by geeky at 10:48 AM on August 3, 2011 [1 favorite]

Look up 'superstitious pigeons'. It was an experiment in which pigeons in a skinner box were delivered food randomly without any prerequisite behaviors or cues. Many of the pigeons came to believe that whatever they had been doing the first couple times the food was delivered was the cue, so they would repeatedly turn in a circle, or peck their toe, or whatever. Their behavior was reinforced because they kept getting food, bolstering their suspicion. Your dog may believe it's the looking back and forth that's the desired behavior, and since that is eventually being reinforced he is unaware of the impact of sitting.
posted by whalebreath at 11:30 AM on August 3, 2011 [2 favorites]

I agree with geeky about the dog's mental frame being attuned to your cues rather than his own. That's just how Border Collies roll. They have to be focused on exterior cues (handler, sheep movement,) and their response comes from the stimulus of the cue.

That said, why do you think Duffy should pick up on such a subtle idea when you haven't even figured out AskMe's cues for posting pictures of your dog?

Animal AskMe = goggie pictures!
posted by BlueHorse at 11:32 AM on August 3, 2011

How long have you been at it? If he has a long history of doing the look-snack|look-you communication and has been rewarded with snacky goodness as a result for years, it's pretty ingrained in his mind that this is the way we do these things. Just because for some reason you are now not responding reliably doesn't signify to him that the old way doesn't work, but just that it doesn't work as well as it used to because you're maybe getting a little old and absentminded. :)

The problem is probably that you aren't modeling a new behavior as much as trying to re-model an old behavior (without explicitly telling him what you want, and while the old behavior is still in play), and that's likely to take a good bit longer. As far as he's concerned, he's still communicating in the same fashion, and still getting tidbits, even if fewer. I'd say if you want to test his ability to connect an action from you resulting from his own behavior, it would be better to start with something totally new.
posted by taz at 1:05 PM on August 3, 2011

If your dog knows the Sit command, just use it when you want him to sit, then reward him. He'll learn pretty quickly that sitting is a prerequisite to a treat and you'll find him sitting hopefully without any command, if my own border collie is anything to go by ... :-)
posted by Susurration at 9:38 PM on August 3, 2011

I think jet_silver's goal, though, was to determine if Duffy could figure out on his own that the treat was dependent on the sit. (Which, again, I'm betting he certainly can, but this particular test may be misleading due to the existing dynamic.)
posted by taz at 10:48 PM on August 3, 2011

Response by poster: BlueHorse, goggie pics here.

taz has it: I thought animals would do random shit to influence their environments until something worked. It was pretty clear Duffy wanted a cue from me about what it took to get $_TREAT but this was an experiment, if I'd told him "sit" he'd have done it. Changing $_TREAT to blueberries changed the conditions in his mind and he figured it out quickly with that value.

It's not clear Duffy liked this experiment, but it was attention and that works for him.

Thanks, folks.
posted by jet_silver at 6:51 PM on August 5, 2011

« Older What female bbc correspondant wrote an awesome eco...   |   "My Kindle can't play Angry Birds, but here's a... Newer »
This thread is closed to new comments.