How dogs became man's best friend

February 21, 2004 7:50 AM

How dogs became man's best friend: Dr Hare's hypothesis is that dogs are superbly sensitive to social cues from people...
His experiment was simple. He presented his animal subjects with two inverted cups. Then he hid the cups behind a screen, put a small piece of food under one of them, and took the screen away. The animal had to choose which cup to look under. If the experimenter gave no cue, both species got it right 50% of the time, as would be expected. However, if he signalled in some way which was the right cup, by pointing at it, tapping it, or even just gazing at it, a dog would choose correctly every time, while a chimpanzee would still do only slightly better than chance. [More at Harvard Gazette]
-- My question: are you able to reproduce his results?
posted by MzB (21 comments total)

Main study:

The third idea was that sensitivity to human social cues is a recent genetic adaptation that has evolved specifically to allow dogs to enter a new ecological niche—that of being symbiotic with people. Testing this one was tough. But Dr Hare managed it by going to New Guinea.

This island has a population of dogs known as singing dogs, because they cannot bark, but instead yowl in a way some compare to human song. Singing dogs' body shapes are those of other dogs, rather than of wolves, indicating that they may once have been domesticated. Research on foxes in fur farms suggests that such body shapes are a side-effect of breeding for docility. But New Guinea's singing dogs are now completely wild, and the archaeological record suggests that they have been so for thousands of years. That, Dr Hare reasoned, would be long enough for natural selection to eliminate a trait that was no longer valuable to the animal.

And so it proved. Singing dogs, even if raised from puppyhood by people, were no better than wolves or chimps at find-the-morsel. The conclusion is that what natural selection had taken away from the singing dogs was something that it had first given to ancestral dogs: the ability to understand people's intentions almost as well as they understand each other's.
posted by MzB at 7:56 AM on February 21, 2004

Soon we'll be able to "domesticate" any living thing by gene therapy. "Paging Doctor Twostrand..."
posted by wobh at 10:10 AM on February 21, 2004

Interesting. Of course, most pack animals are expert body-language readers, and adaptability is a survival mechanism. So it makes sense that dogs can adapt relatively easily to body language especially over time, and, to nitpick, it seems unlikely that his tested domestic dogs were raised with NO human contact whatsoever (it says "minimal", but it also doesn't mention how well-socialized the dogs' mothers and the dogs around them were - if the "role models" the puppies had in their earliest weeks were friendly toward and accepting of humans, even with only minimal exposure, it's entirely possible that it was enough for the puppies to develop some level of trust in people), therefore it's entirely possible that being able to read people is a learned trait, and since dogs do a lot of learning by example (i.e. from their mothers and other dogs around them), I don't know that the fact that the singing dogs raised from puppyhood by people are necessarily disproof of that (dogs of domestic breeds which are hand-raised from birth by people have a wide variety of problems, especially with socialization to people and other dogs, as do dogs raised ferally by adults suspicious of humans - I'd be interested in seeing how well one of these dogs does in this test). My take is that dogs need to learn how to read humans from other dogs, as well as from humans (as is amply shown by the problems of dogs under-socialized with people AND other dogs) - there's very likely a genetic component, of course, but I suspect the skill is primarily a learned one.
posted by biscotti at 10:14 AM on February 21, 2004

MzB, not sure I understand what you're asking. Are you curious as to whether my dog understands pointing and other body gestures? Definitely.

This is also why it's easier to train a dog with hand signals than with verbal commands. My dog will sit 100% of the time with the hand signal and 75% with just the word "sit". He's about the same with other tricks as well. He also can tell the difference between hand signals that are only slight different (rotating my index finger clockwise is a different command from counter-clockwise and results in a different trick, for instance).

I used to have a dog that mimicked my movements. I would lie on the floor, he would lie on the floor. I'd put my head down and so would he. I closed my eyes and went to "sleep" and so did he.

There's an old joke that goes: "I don't like dogs because when you point at something they smell your finger." It's a stupid joke because, in my experience, it's not true.

Most dogs know the difference between "pointing at something" and "offering a finger to smell". How? By watching the rest of your body. When you point at something, you look at it as well. When you offer your finger, you look at the dog.
posted by dobbs at 10:36 AM on February 21, 2004

Instead of a dog, could I use a homeless person?
posted by sharksandwich at 10:58 AM on February 21, 2004

Fascinating but unconvincing. Even assuming he's right about the "singing dogs" being formerly domesticated, I'm not convinced that a couple of thousand years would eliminate genetic changes. Assuming that changes so complicated would be purely genetic. Like biscotti, I suspect his methodology is flawed, and the kennel-raised dogs had more human contact than he thinks.

Another thing that occurs to me is that both humans and dogs evolved as pack hunters / scavengers on plains. Maybe pack hunters need to communicate to each other in similar ways. When a dog wants to take you a particular way, he finds it pretty easy to communicate that to you.

I suspect that humans actually may have a lot more in common with dogs than with chimps. The genetic differences aren't much greater, and chimps are highly specialised to a particular rainforest lifestyle, unlike us savanna-roaming dogs/humans/jackals.
posted by TheophileEscargot at 11:00 AM on February 21, 2004

-- My question: are you able to reproduce his results?

I would be, but since that unfortunate incident in the mid-90's they've limitted my access to free-range Chimpanzees...
posted by jkaczor at 11:08 AM on February 21, 2004

I've been told anecdotally that if you point at something, a dog will look at what you are pointing at, but a cat will look at your hand, this story seems to back that up.

I wonder if this behaviour evolved thousands of years ago or was bred recently into dogs, like Chimpanzees haven't been bred to the degree dogs have, maybe if you found some Chimpanzees that responded better than normal to hand signals and bred them with Chimpanzees that could do the same I bet you eventually could come up with Chimpanzees that responded as well as dogs.

I'd be curious if there was any difference in response between different breeds of dogs.
posted by bobo123 at 11:34 AM on February 21, 2004

Dogs are such wannabes....
posted by rushmc at 11:53 AM on February 21, 2004

I'd be curious if there was any difference in response between different breeds of dogs.

Good point, there probably are differences between different groups of breeds. Breeds developed to work closely with humans are probably more likely to "tune in" faster than those developed to work independently, because the traits you want in, say, a herding or working dog are different from the traits you want in a terrier or gundog, and you select breeding animals in large part according to these traits (which is the main reason people need to consider a breed's purpose before choosing one). That said, I suspect that overall, the main differences come down to socialization, both of the individual animal, and of the animals which raise it. I have a feeling that the differences between well-socialized canids of any type raised by well-socialized canids are smaller than the differences between well-socialized canids raised by well-socialized canids and completely unsocialized canids raised by unsocialized canids. In other words, a well-socialized wolf raised by well-socialized wolves is likely more similar (in terms of reading human body language) to a well-socialized dog raised by well-socialized dogs, than it is to an unsocialized dog raised by unsocialized dogs.
posted by biscotti at 12:02 PM on February 21, 2004

Speaking of dogs, behold this year's Best in Show, a Newfoudland named Josh. Lord Byron had a Newf he adored named Boatswain (judging by the painting, I'd guess Boatswain must have been a Landseer.)
posted by homunculus at 12:07 PM on February 21, 2004

Hmm I proved dogs could smell food reeaaallll well,
Did it with mine and the dog just sniffed both and went for the one with the food, DAMN that dog is smart.
posted by Elim at 12:32 PM on February 21, 2004

How dogs became man's best friend...

Kipling tells it differently...
posted by Shane at 12:37 PM on February 21, 2004

I was noticing just this morning how attuned my dog is to pleasing me, and he's pretty new to our family. Surely dogs have adapted to please us in order to share in the food, love, couch, etc. And we domesticated dogs because they have traits that suited our needs. Now, how do we get them to adapt to not chew shoes?

On preview, if you want confirmation of dogs' using hand signals, and non-verbal commands like whistles, check out herding dogs.
posted by theora55 at 12:42 PM on February 21, 2004

Dogs know when it's time to hit the road.
posted by cookie-k at 12:47 PM on February 21, 2004

Sure, dogs know all the tricks, and chimps are a little harder to get.
But what about the other pets: cats, parrots, hamsters? Are they able to identify a (conscious) hidden signal? (It is not about a trick that the pet was able to learn from us)
posted by MzB at 2:44 PM on February 21, 2004

But what about the other pets: cats, parrots, hamsters? Are they able to identify a (conscious) hidden signal?

I agree with bobo123 about cats. They usually don't get pointing. If you want to show something to a cat, you need to point very close to the item, because the cat will watch your hand instead of the invisible target of the pointing.

That said, I have been surprised by how well my cats have figured out some other pretty subtle signals. One of them figured out that going-for-the-toilet-paper means I'm about to get up off the toilet. She'll sit over by the bathroom door, and once she hears the sound of the roll, she'll get up. Surprised the hell out of me the first time she did it. My cats also seem to have a pretty good idea of what I'm looking at. If I look intently at the treat drawer or the lazer pointer toy, they get excited even before I make a move for it.
posted by vorfeed at 4:15 PM on February 21, 2004

Anyone who owns a dog could have told them that!

Domestic animals ARE different and that's been recognized by the scientific community for a long time. We are all pre-programmed to be "nice" to juveniles and most juvenile mammals share similar characteristics. Domesticated animals, for the most part, retain their juvenile characteristics far longer than they should. That includes an inquisitive, submissive personality but also physical things like floppy ears, playfulness, large eyes relative to head size, shortened muzzles, some coloration markers and generally other things we find "cute" as opposed to threatening. We've been conditioned too.

As an example: foxes don't really domesticate well. If you raise one from early cubhood or even birth it will stay with you happily for 1-2 years and once it matures will leave. On fox farms, foxes are quite aggressive. When they bred them for temperment, they found that within a few generations they got floppy eared, big eyed, docile animals who never attained their proper adult coat. No good for fur obviously!
posted by maggie at 6:18 PM on February 21, 2004

The answer is obvious, dogs have souls, cats don't :p
posted by Trik at 1:44 AM on February 22, 2004

The difference between cats and dogs:

The dog looks at his master and says: "He feeds me, he takes care of me, he shelters me... He must be a god!"

The cat says: "He feeds me, he takes care of me, he shelters me... I must be a god!"
posted by eas98 at 7:39 AM on February 23, 2004

I find that cats react well to audible clues rather than visual ones. For instance, our cats (one Siamese, one Maine Coon mix) will come when their names are called; know when we're done in the bathroom, as vorfeed's cat does; know that our suitcases appearing in the apartment means we'll be going away.

It's Pavlovian, but more sophisticated than simply bell=food.
posted by me3dia at 1:36 PM on February 26, 2004