Dog Learning New Behavior

[Eileen Stein]

An article in this morning's Washington Post describes an intriguing recent study on dog learning.

 

A group of dogs were trained to pull on a trapeze-like wooden rod to get a treat. They were free to do it any way they wished. Most (85%) spontaneously did it with their mouths, showing a natural canine preference to manipulate with their mouths.

 

A demonstrator dog (a border collie, natch) was then trained to pull the rod with her paw, either with or without a ball in her mouth.

 

When a second group of dogs watched her pull the rod with her paw while holding a ball in her mouth, 80% of them imitated her action using their mouths to pull the rod, rather than their paw. But when a third group of dogs watched her pull the rod with her paw while she had nothing in her mouth, 83% of them used their paw to pull the rod, just as she did.

 

This was similar to the results of research with human infants. It has been found that 14-month-old children will imitate an adult turning on a light with her forehead only if they see her doing it with her hands free. If the adult is clutching a blanket, infants will use their hands, presumably because they can reason that the adult resorted to using her forehead only because she had no choice.

 

The results of the dog experiment were surprising because they suggest that the dogs go through the same reasoning process as children do. Interesting, eh?

[Jim2000]

it is interesting, but is an extra 3 percent that big of a difference?. I am not a scientist so I dont know if that is more significnt than it seems to me.

[Eileen Stein]

I may not have written it clearly, but it's not just a 3 percent difference, it's a total reversal.

 

After watching a dog pulling rod with paw while holding a ball in her mouth, 80% of them pulled the rod with their MOUTH. It appears they thought she wasn't using her mouth only because her mouth wasn't free, rather than because it wouldn't work if she used her mouth.

 

After watching a dog pulling rod with paw while her mouth was free, 83% of them pulled the rod with their PAW. Since they could see her mouth was free, it appears they thought she wasn't using her mouth because there was some reason a mouth wouldn't work and a paw had to be used.

 

Reading the article may make it clearer.

[Dixie_Girl]

I read the article on MSN. I found it more interesting that the scientist were just now admitting they don't know all there is to know about animal intelligence! Reminds me of the head of the U.S. patent office who stated that everything that could be invented, has been invented..........in 1899!!!

 

The part of the article that had them excited was that the dogs were making a distinct difference in the dog using a paw with his mouth occupied as opposed to using it when he could obviously have chose his mouth. THAT is what made the difference. The normal way is to move it with their mouth, and they "understood" that when his mouth had a ball in it, he couldn't use his mouth. So, they didn't use their paw. But when the dog had nothing in his mouth, they understood that moving it with the paw was what was wanted. But heck, I coulda told them animals are smarter than scientist think! LOL

[Alaska]

Reading this article left me...uncertain, I guess. Is this objective science? I read the methods in the appendix and I still can't decide. Even with all the controls, and the fact that the owners/handlers supposedly did not know the purpose of the experiment, I'm uncomfortable with the various active roles played by the humans.

 

Did you watch the movies of Tracy and Duke? In the movies of the two dogs' first "independent" trial, the main thing I notice is that both dogs just start nosing around on the ground for treats. They haven't learned yet that there is a connection between doing something with the rod and the appearance of a treat. The owner has to repeatedly point at the rod to get the dog to do something with it (and some dogs never do figure it out).

 

Tracy doesn't even pay attention when the experimenter "loads" the treat dispenser, even though showing the treat to the dog first and having the dog watch the dispenser being loaded is an explicit part of the experimental design.

 

But when they do finally interact with the rod, the first thing each dog tries (nearly all of them, not just Duke and Tracy) is as reported: "mouth" if the demo dog's mouth had been full, and "paw" if the demo dog's mouth had been empty. That's too unlikely to be a random pattern, so they must have retained some information from watching the demo dog.

 

But the humans were watching the demo dog too. And the humans were pointing out the rod to their subject dogs.

 

All we really know is what the dogs did. We still don't know why. The interpretation that they "reasoned" is just an interpretation -- there's no competing hypothesis for why they did what they did.

 

Some dogs tried only one method and stuck with it. Some dogs tried one method, then switched to the other. Some dogs tried both methods, then switched back and forth for the rest of their trials (the two border collies, Duke and Tracy, were among the "experimenters"). What's the significance of that?

 

I'm left with more questions than answers. But then I'm not a psychologist, so I'm not familiar with these sorts of experiments.

 

Here is the link to the page with methods and movies.

[Eileen Stein]

Did you watch the movies of Tracy and Duke? In the movies of the two dogs' first "independent" trial, the main thing I notice is that both dogs just start nosing around on the ground for treats. They haven't learned yet that there is a connection between doing something with the rod and the appearance of a treat.

 

Well, apparently the subject dogs had been allowed to scarf up the treats that the demonstrator dog had elicited from the mechanism after each demo, so it seems to me only natural that they would look on the ground first even if they had grasped the connection between the rod and the treat, since they'd found treats there repeatedly during the demo phase.

 

But the humans were watching the demo dog too. And the humans were pointing out the rod to their subject dogs.

 

Is your thought that the humans may have pointed the rod out in a way that would suggest "mouth" to some and "paw" to others? What is it about the way they pointed that would suggest one or the other? Or is your concern something different? The experimenters do seem to have tested some dogs whose humans had their eyes covered during the demos, and found no difference in the dogs' responses.

 

All we really know is what the dogs did. We still don't know why. The interpretation that they "reasoned" is just an interpretation -- there's no competing hypothesis for why they did what they did.

 

That's very true. I think one of the challenges in designing studies like this is to set it up in such a way as to exclude other possible reasons for what happened. Obviously that's what they tried to do here. Is there a competing hypothesis you can think of?

 

Some dogs tried only one method and stuck with it. Some dogs tried one method, then switched to the other. Some dogs tried both methods, then switched back and forth for the rest of their trials (the two border collies, Duke and Tracy, were among the "experimenters"). What's the significance of that?

 

Personality differences, maybe? In humans, it seems to me some folks are more inclined to "experiment" than others, while some are content with the "tried and true."

 

If the percentages were not so dramatic it would be easier for me to view it as chance plus some "noise" that we can't identify. When they're as striking as this, and evidently parallel similar studies with human infants, it seems to me they're hard to discount.

 

BTW, five of my dogs and I once were subjects in a study designed to determine whether there are differences between breeds in the extent to which dogs understand the meaning of a pointing gesture by their person. The way that study was conducted, the involvement of the human (me), and the sort of impression of "messiness" you would get when watching it, were all totally similar to this one. I bet it's much the same with the infants.

[Bo Peep]

ooopgs