Peer-Reviewed Papers on Dog Ethology

One of my cognition lecturers described a simple means-end experiment once- put a dog on one side of a long (but not closed off) wire-mesh fence, with food/toy/something it wants on the other side. Does the dog try to go around the fence? or does it try to go through it?

 
Or over...or under... That experiment seems better suited to dogs' abilities and would show a means-end understanding, no?

Hmm, good point Anne. Do you think breed-specific characteristics might have influenced the string pulling experiment too?

They did use different breeds, I think, but I don't remember seeing a list of the breeds. I do think that if they used different breeds they would have checked back to make sure it didn't significantly affect the results, like they did for the age of the dog and that kind of thing right at the start.

Good point about the smelly string, Vinia. Maybe you could control the smell well enough by soaking the string in meat broth or cheese or something that smelt very strong and then positioning it with gloved hands. But it would still be dicey and some scientists would probably have a problem with it.

Talking about evolutionary basis for means-end understanding, I remember seeing a doco with a jumping spider that was trying to lure another spider out of its web by mimicking the mating strum of that species on the silk lines. It didn't work that time, so the spider wandered off and reappeared above the web a few minutes later to try another strategy from above the spider. I was so astounded that a little jumping spider could make such a leap in comprehension that it could move out of sight of the prey in order to position itself for a different strategy in catching the prey. Would that be means-end comprehension in a jumping spider? Because that's pretty awesome. []

I can't think of where a wild dog might need to understand means-end, but as opportunists and hunters, I would assume that understanding means-end would be sure to come in handy at some point. Something like a mouse inside a small hollow log would be much easier to tackle with some very formidable problem solving skills.

ORIGINAL: Vinia

Hmm, good point Anne. Do you think breed-specific characteristics might have influenced the string pulling experiment too?

I don't know - possibly.  After all, it matters, in terms of scent ability, whether you are a Bloodhound or a Saluki.  But, all dogs have the ability for scent discrimination sufficient that they would notice even subtle differences.

I found that part of the discussion interesting that centered on the fact that dogs may simply have developed a sensitivity to reading the humans' wishes.  Thus, they are not as adept at problem-solving, since they never have to do it.  But, as I have seen in my own pack, the dog that was bred for her ability to work stock on her own with minimal human interference also has better problem solving ability than my other dogs. Sounds like there's something to that conclusion that needs further investigation.
Also, I don't know if any of you saw the news this morning, about elephants realizing it's them in the mirror.  I think Sequoyah "gets" that it's her (and me) in the mirror, although she certainly did not when she was a small puppy.  I'm not sure about the others, since I've never put them to the test.  But, I think that, as with all species, there are brighter bulbs in some lamps than others.

I am of the opinion that alot of the behavior these dogs were exhibiting were natural problem solving displays. Dog will "naturally" paw at surfaces to get to food or try to get their tongues into small spaces to get to the last scrap of a tasty morsel.
The string is hook to a larger solid object that just so happens to be "in the way" of the dogs pawing motions, that in turn "pull" the string. So, I don't think they really KNOW if they pull the string, that the food will come out. They are just doing what come natural for them.
Also, I believe that some of the results have been adultered and skewed by using "treats" for pulling (pawing) the string and by conducting portions of these experiments outside of a controlled environment. The results could have quite possibly be alot different had they been more tightly controlled.
The use of reinforcement for pulling the wrong strings was nothing more than operant conditioning. Get the dog pulling ANY string for a treat, and he will eventually get it right!
 
While I believe that dogs have extreme problem solving skills, I don't think they really make the connection to the means-end relationship.

many cases their food is hidden inside something and they need to use a tool or otherwise make the connection in order to get the food.

 
I have watched Shadow find a pecan on the ground, crack the shell, and get at the nut inside to eat it.
 

Agreed.

And it sounds like we are still seeing problems with this test and how those problems may limit the viability of a conclusion from it as regards a "means-end" solving ability. I mentioned earlier how I have watched Shadow crack a pecan shell and get the nut out of it to eat. True, he did not go to the drawer in the kitchen and pull out a nutcracker. He simply cracked it in his jaws a few times until the whole thing fell apart and ate only the nut, leaving the hard shell alone.

I think motivation goes towards problem solving. Some dogs may not bother to problem solve so much. They know they will get a treat, regardless. Is that problem solving? That is, they are smart enough to know that they'll get a treat as soon as they do enough to make the human smile and give forth? And not have to expend a lot of effort in the meantime?

I agree with Anne, too, that there should be more investigation in the role of human in a dogs solving ability while in the company of humans. Dogs live in groups, cats do not. Cats may have more readily apparent problem solving skills. Jade can reach a doorknob now and paw on it to try and turn it like she sees humans do. She is teaching herself how to open doors. So, I better quit leaving the car keys laying around.
 
Shadow doesn't bother trying to turn a knob. Why bother? He behaves and knows he gets access when we are ready, so he is more patient and more willing to wait for inclusion on our activities.
 
Again, though, perhaps tighter control groups and criteria may or may not help.

Here's my quick take on the subject and then I'm off to work (two aggressive RR's this morning). I think most would agree that dogs have some form of reasoning skills. But so far I've seen no proof as to what they...actually are. First off since dogs raising dogs do not use speech (beyond calls or correction, and play) to teach their young, one that has studied dogs can conclude from their studies that a dogs (be they wild or domesticated) most natural from of learning is through...observation. They watch, observe, and learn. It is also...IMPO...that it is difficult if not almost impossible to actually learn much about dogs in a controlled environment in contrast to their natural one, the wild for wild dogs, and the human home for domesticated dogs.

Wow everyone! This is awesome and exactly the kind of discourse I, for one, have been craving. Life (er, the near-death stupidity of myself and my dog) kind of got in the way for me to really effectively participate in the current paper discussion, but I promise I'll be ready to go for the next.
 
Given one of the directions the current discussion is going, I thought it would be neat to take a look at [linkhttp://www.psy.cmu.edu:16080/~tedtutor/cues_in_dogs.pdf]this article[/link]. Let's begin our discussion Monday (or over the weekend if you feel excited about it).
 
Miklosi, A., Polgardi, R., Topal, J., & Csanyi, V. (199. Use of Experimenter Given Cues in Dogs. Animal Cognition, 1:113-121.
 
Also, I just heard about a study done by Dr. Esther Schalke about the use of shock collars and increase in saliva cortisol levels in beagles, and try as I might I couldn't find it anywhere. It's possible that it's written in German (she's a DVM and PhD at the University of Hannover in Germany), but if anyone has any info on how I can get my hands on this in English, please let me know.
 
The summary that I heard is that she had three groups of beagles--all from the University's breeding stock, all living together at the same kennel. They spent a few weeks getting baseline saliva cortisol levels for each individual and habituating the dogs to having their cheeks swabbed (cortisol levels found in saliva are an indicator of stress level). For the experiment, there were three conditions meant to simulate the use of shock collars as a tool to get dogs from chasing unwanted prey (apparently Germany has a major wild boar problem). A fake lure rabbit was used to simulate the prey and the beagles were allowed time to play with it unmolested and get to see it as a desirable thing to chase. In the first condition, the dog recieved a shock the second their mouth touched the rabbit. In the second, the dog was given a pre-trained recall command and if the command was ignored, recieved a shock. The third condition was meant to simulate the typical use of shock collars by John Q Public, which is sort of random, with bad timing and changing criteria (sometimes it is used as in the first condition, as soon as the dog touches the unwanted prey, sometimes as in the second, to enforce a recall, and sometimes just kind of randomly as the dog is beginning to chase the prey). Cortisol levels were checked at the end of each session. The results at that point were that the dogs in the first condition showed elevated cortisol levels, by about 30% on average. In the second, cortisol levels were greatly increased, by over 100%. In the third, predicatbly, cortisol levels were through the roof, up by over 300%.
 
The beagles were then given a few weeks off, and then brought back to the room in which the experiments had initially taken place. The prey animal was not present, nor were the electric collars. Cortisol levels were checked again. The dogs in the first condition showed little to no increase. The dogs in the second and third conditions, however, displayed a marked increase in cortisol levels, approaching the initial increases from right after the aversives were applied during the experiment.
 
The conclusion that the experimenters drew is that shock collars greatly increase stress levels in dogs on a lasting basis, unless they are utilized by a trained professional with exquisit timing and even then only when the cause for the shock is abundently clear (touching the unwanted prey animal with the mouth).
 
She delivered this paper at the recent APDT conference and I'd love to read it and study it first hand if possible.

That's a really interesting article. I wonder, if this experiement was to look at all species, how many species we would find that also use their eyes to communicate directionally as we do.
 
I am also curious what the results would be if wolves could be tested similarily. Would they have the same pre-disposition to learn & understand our cues?

That's a really interesting study, Cressida.

It would indeed be interesting to know how a wolf would fare in the means-end string-pulling experiment. I was wondering about that.

Judging by the results of the string-pull test and the social cue test, I think the wolf might do a little better at the string pull, even with similar limitations of vision, because they are not depending on a human to lead them but they would probably not due as well understanding a human cue. In fact pointing might be taken by a wolf as something entirely different than "The food is there."

An interesting bit on the neoteny, the trait of carrying juvenile behavior into chronological and biological adulthood. It is apparent in both dogs and humans. And it, according to this second article, is what makes us able to learn, i.e., we never quite grow up and can still see things as "new." This also supports what I had read many years ago and mentioned here once or twice before. Man is not as evolved as, say, the mountain gorilla. We are "unfinished." Zoologically, we are great apes.

It was also interesting to note how man and dog have evolved somewhat together and this may affect social behavior. Which, superficially, supports how I have stated observing dog behavior in humans and have used dog behavior on a human, with startlingly effective results. A guy was trying to take over my project last year and I literally barked at him, loudly. Then I stated, "This is my fire hydrant." He had flinched, which was my intention. He backed down and did the task he was originally sent there to do.

Did the first article say that the owners told the dog it was okay to get the treat?
 
I know that with my dogs I can leave a plate of food on the side table long enough to go to the other room and get something, return and none of the dogs will have touched my food.  (Certainly I wouldn't push it and take a shower and expect a steak to still be there.  Two of them are, after all, basenjis.)  So, back to the topic at hand (paw?), I wonder if this had an effect on the results. 
 
Also, I don't think the study specified what type of treat was used.  I know the boyz will work for meat, but may lay next to a hard dog biscut all day.  The malinois is happy to work for tossing a ball.

Good point. In Shadow's case, if you use a carrot, you'll have a long wait because he's not interested.