Would it be possible to breed (not gene-engineer) a dog with near-human-level intelligence?

Yes, there is some evidence that some primates are capable of abstract thought.

Emotions are different but the abstraction resides in thinking back on an action and realizing you were wrong, ect. This is, of course, my own opinion.

An ant moving around trying to take different paths is not the same thing as a dog attempting to reach through a fence for a ball, then running around to the other side of the house, sneaking under a gap in the fence there, then running back around the achieve the ball. The former is blind variation, whereas the latter indicates some understanding of both non-local spatial relationships and objective-resultant reasoning. However, that doesn’t mean that we can expect a dog to compose, or even appreciate, a sonata.

We got here from single celled organisms not be selective breeding (which requires conscious decision) but by natural selection, a process of winnowing a vast number of variations by selective pressure. Our particular route to intelligence–something that is still highly debated–would not be the same as that of any other organism starting from a different part. In other words, we couldn’t breed canines to have the identical neural structures and cognitive processes as humans, even if we start with the goal of developing intelligence any more than one can make lamb stew from a pig. We may be able–with much directed breeding and applied evolutionary pressure–to develop something like human cognition and language skills in canines, but it would be at least subtly different (and more likely significantly so) from that in humans owing to dramatically different evolutionary pathways.


Agreed. But, I don’t think of this as abstract thought. Mice can figure out how to get through a maze in order to get a treat. This is more pattern recognition. Now, if the dog pulled a piece of wood over to the fence and used the wood as a ramp I would be more inclined to call it abstract thought.

My dog can do both of those, mind you. (Of course, she is a standard poodle, second only to border collies for intelligence.) She can abstract from ‘heron’ (which is outside and perched on the dwarf cherry tree) to ‘heron’ the stuffed animal, and he can abstract from ‘squid’ the red toy to ‘squid’ the canvas toy, made of a different material, and of a different size. (The math tricks are fairly rudimentary, but she can tell between two frisbees in play of the same type and three, and insists on finding the third before we go in, as well as knowing when someone has thrown two and when someone has thrown three. Various addition and subtraction tricks follow, and while she’s not great at math, she can generally do them.)

Funny thing is that the dog has come up with some idea of a ‘plot’ in the nature shows we like to watch. One notable incident was when we were watching a show on bears, and a man came out of a hidden blind, and she barked at him, apparently telling him to go back somewhere.

Her uncle once recognized his uncle on television, which I found fascinating.

And that doesn’t count the fact that the uncle figured out how to turn the TV on with a remote. I think that’s fairly advanced problem solving.

Yorick, as far as the ramp thing, you ever see that video of the beagle chimneying his way out of a shack to get out?

You simply can’t compare animal intelligence to human intelligence. Tool using is common in the animal kingdom. And yet, the propensity to use tools does not seem to be related to other forms of intelligence. Trap door spiders build complex tools. Birds can build amazing nests. Otters use rocks to break open shells of mollusks. Some animals use sticks to dig honey out of hives or dig termites out of mounds.

Whales go to ‘school’ when they are young, and are taught significant behaviors by their parents rather than knowing them instinctively. I’ve seen a video of a killer whale pod where the youngsters watched as an adult whale would knock a sea lion off an ice floe - and then leave it alone to climb back on, only to knock it off again. Then the kids would try it. It was clearly a training session.

There are crows in New Zealand that have learned to use traffic signals - they break open their shelled food by flying over roadways and dropping it from a height onto the concrete, then they land at the crosswalk, wait for the light to go red, and walk out into the street to collect their food. They don’t even fly out - they walk with the other humans. It’s the damnedest thing. But that doesn’t mean they are intelligent, or self aware, or capable of conceptualizing and planning complex tasks. In the case of crows, it probably means they have an advanced ability to learn through mimicry. A crow dropped a shell, it broke, other crows saw it and did the same thing.

An African Grey Parrot has shown the ability to use abstract concepts, to be able to count, to discern objects of different colors and materials, etc. The way it learned is through observational mimicry. Trying to teach it like a dog did nothing. Instead, two assistants would perform the behavior while the bird watched. It apparently built up some sort of complex decision-tree type mapping for distinguishing the various objects, and could put on a very good display of ‘intelligence’, but we have no way of knowing whether there was any real thinking going on of the type we would recognize in ourselves.
As another example, dogs have the ability to figure out from where your eyes are looking what you want them to do. There have been experiments where a person held perfectly still, then told the dog to fetch an item on the left or right merely by looking in that direction. The dog will figure it out. But is it really thinking, “Hey, that person is looking over at that thing over there. That must be what she wants me to get.” Or is there a more primitive simple trigger that merely says, “black dot moved that way, I go that way.”

As another example, A border collie has been able do show the ability to think in abstract concepts - an experiment was done in which the border collie was shown two pictures - one of a dog, and one of a landscape. If the border collie touched the picture of the dog with its nose, it got a treat. After learning to do this, it was shown completely different pictures of different dogs and different landscapes, and it was still able to figure out which picture to touch. This suggests that the border collie at least has the ability to generalize the concept of ‘dogness’ vs ‘landscapeness’ to some degree, even if the dog and landscape is a completely different color and shape.

And yet, dogs are utterly unable to recognize themselves in mirrors. They may see a ‘dog’, and even bark at it or paw it or whatever. But they don’t seem to understand that they are looking at images of themselves. Apes, however, can. Experiments have been done in which a colored marking was applied to an ape without its knowing. Then, when the ape looked in a mirror and saw the colored mark on the ape in the mirror, it actually reached up to touch the mark on itself, indicating that it had some concept of ‘self’, or at least the knowledge that it was the same kind of thing as the thing in the mirror, and that the colored mark didn’t belong.

But again, it’s easy to assume that the ape thinks about itself as an individual and understands what that means in a human-like way. In reality, it might be a completely alien thought process that it’s employing to make the same kind of connection.

I meant that it amounts to the same thing. Selective breeding and natural selection are still selection, and if intelligence is highly select for, the result is not that dissimilar, except that natural selection is selecting for more than one thing at a time.

I also never said that you would get identical neural structures; I said that, given enough time, you can even account for genes that aren’t present in the original population (due to mutations), but that what you got would probably be less dog-like over time if you selected only for intelligence. For example, if you’re crazy enough to try to make a dog that talks, you’d be hard-pressed to call it a dog at the end.

As I always say, dogs are probably the closest species to human. We evolved together.
The ability Sam mentioned, for dogs to recognize people indicating things, is really, really unique. Even chimpanzees have trouble with it.

Human brains are amazingly complex innate language machines and animal brains are not. It’s that simple. There are hard limits to what can be done via “breeding”. Unless a dog brain can be substantially mutated towards language you don’t have prayer of getting a dog with human like intelligence.

How do we know that they just don’t find the image interesting? Our Border Collie, who knows the names of many different animals, doesn’t care one bit about her own image in a mirror, but if she saw a cat reflected in one, I guarantee she would find it pretty fascinating. She just isn’t narcissistic, IMHO.

She has excellent vocabulary though. Mention a "squirrel’ or “raccoon” or “cat”, or any of a couple of dozen other animals that she might like to chase or herd around, and you get an immediate reaction. She also understands complete sentences. “Come over here” “Go to the back” “Leave that alone” “Do you want a treat?” etc., are all recognized by her. We speak to her pretty much as you would a child, and she seems to get it most of the time. Yes she is pretty darn smart!

But no, their brain structure is not the same as ours. There is a theory that we owe a lot of our own intelligence to dogs, in that when we first started partnering with them they specialized in areas of brain development related to scent and hearing, leaving us free to specialize in more abstract thought. Dogs have done us a lot of good over the last fifty thousand years or so, and may even be one of the reasons we are able to bandy it about on the internet. A great example of symbiosis.

It would certainly be handy if a sheep dog could count, so he wouldn’t leave strays behind.

That would be a hypothesis, and one I’ve never hear of. Got a link? I’m not trying to be snarky about this, but that hypothesis does not fit in with our consensus understanding of either human or dog evolution.

Although Border Collies do often have a handful of common ancestors, one thing that should be noted is that they also have an uncommonly low inbreeding coefficient. What they do (or should) have in common is dogs not only bred for high intelligence, but dogs bred to do very advanced sheepherding tasks out of sight of the shepherd. Dogs “on the hill” must navigate another species, with it’s own intelligence and defense mechanisms against predators in a hostile environment (Scottish highlands, American range, etc) often out of sight and hearing of the shepherd/rancher. Add to that a very competitive and difficult trial standard and you’ve got a breed that not only reads and understands two or three different species extremely well (human, sheep, cattle) but must also be intuitive problem solvers.

I could write all day on the intelligence of these dogs, but I got to go work sheep with them right now ;)!

I think you overestimate human thinking. All of those examples seem like EXACTLY the sort of things we do, for exactly the same sorts of reasons. Training, observation, mimicry—you think a child can intellectually grasp the concept of “red”? No, she simply mimics adults who point to a cherry and say, “red.” Then when she grows up, she teaches her own children how to mimic in the same way.

Also, I think everyone is coming at human intelligence from the wrong angle—I don’t believe our intellect is so fantastic in terms of coming to grips with our external environment, which is what everyone seems to focus on. We seem to be talking about breeding dogs that can add and subtract and build jet engines, or something.

With humans, where our intellect really excels is in modeling the thought-processes and potential behavior of other human beings. I think everything else—our ability to do caclulus and build skyscrapers—is just a side-effect of needing all that processing power to create, in essence, virtual simulations of dozens or hundreds of other human brains in our heads, so that we can figure out how to get along with them and how to get ahead.

My guess is that it is not possible, because you can’t force-breed one particular trait to arbitrary heights. At least, I’m pretty sure you can’t.

Border collies are smarter than a lot of people I know. And much more pleasant to be around than many people.

Yeah! We’ve done experiments such as teaching animals to talk, but we never went too much trouble finding really smart gorillas to teach to talk. Imagine how much more successful they’d be than the run-of-the-mill talking (ok, signing) gorilla.

To the OP: It’s not so much that we have many “special function” centers, it’s that evolution has perfect in us the ability to evolve the brain, and to make sophisticated connections in. Without such aids, a dog’s brain will have a much harder time to evolve high intelligence. But directed evolution (being far more powerful than the ordinary kind) might overcome that.

What evidence do you have you need complex grammar for abstract thought? Ok, I admit it helps a lot. But even a seemingly limited grammar can be turing complete, no? A language like Chinese has a much more limited grammar when compared to Russian, yet the Chinese are as capable of abstract thought as anyone.

Besides, most of my best ideas come in the moments between thinking.

Ok, for the love of god, music is not some ‘abstract’ high-order facet of preeminently intelligent beings. It is actually quite primitive and instinctual, and evolved in us for its own sake. I agree a dog wouldn’t be into it unless specifically bred to.

What do you think abstract thought IS if not pattern recognition? Everyone does it, just a matter of degree…

This is utter nonsense.