More on dogs and self
Irradiated by LabRat
This is slightly reworked and crossposted material from a discussion elsewhere, because I wound up spending my evening on that instead of spending it on a long post about sexual strategies, mate selection, and human gender politics as I’d intended. So you get this instead.
One of the problems with judging if another animal has something like “a sense of self” is that their minds are highly likely to be RADICALLY different from ours, to the point where they may not even really be comprehensible to us, as our minds are heavily rooted in our own perceptions, instincts, and frames of references. Our intelligences have as much to do with being hominids, with all the special circumstances and influences that go along with that, as they do with being intelligent.
However, we can infer basic things about the minds of other creatures from the kinds of behaviors they’re capable of, what we know about their senses and therefore perceptions, and what we know about how minds in general work. For example, since abstract reasoning is the last ability to appear in the development of a human mind and for developmentally impaired individuals it may never appear at all, we can probably assume that abstraction is a uniquely human innovation. Since all animals need to be able to find necessary resources and avoid dangers, we can assume that even a pre-mind in a creature without a proper brain is capable of setting up some neurological rules that make certain kinds of tasks much simpler. It’s the areas in between that have the question marks on them, especially with respect to animals that are quite complex but not human.
What we really don’t know- and we try to infer when we do behavioral and psychological research- is which things are human innovations that go along with great intelligence, and which things are actually necessary tools for any animal that behaves in a complex fashion in a given context. If you’re going to swim efficiently, you need to have fins, and it doesn’t matter that much except to the details of what you’re doing whether you have a bony fish’s spined fin, a shark’s smooth one, a whale’s tail and flippers, a seal’s modified paw-flippers, or a human’s rubber fins; it’s just a tool that makes swimming much easier and you won’t be half as good at it without, which is why no animal that spends a lot of time free-swimming doesn’t have a structure like this. To borrow terminology from one of my favorite science writers, the exact kind of fin or flipper the animal has- the specific parts on a fish, shark, whale, seal, or human with special tools- are parochial features, unique to that species or general family of animals, but the existence of flippers themselves- a broad, flat surface attached to a limb and used to make moving in water more efficient- are universals, a ubiquitous or nearly-so solution to the same problem, innovated many times over. Lungs and gills are parochials; blood-rich areas of tissue with maximized surface area used for gas exchange from air or water to blood are universals. Chemotaxis, hunting, gathering, and filter-feeding are parochials; foraging is a universal.
My question is: what mental tools are similarly necessary to be a complex social animal? I think self- not our convoluted inner worlds that humans experience that let us do things like plot revenge or write poetry about how fall makes us feel, that’s a parochial, but a very basic sense of “I”- may well be a universal. “I” doesn’t need to be profoundly complicated just because human selves are; if that kind of logic held, then sharks would be unable to swim between continents because they haven’t got submarines. “I” need only function so far and be exactly as complex as the animal requires- and the very presence of abstraction in human thought, and its deep attachment to how we think about ourselves, makes even thinking about the prospect difficult for us. In order to start wrapping your head around the most simple I, strike up a conversation with your nearest available three-year-old. For now, back to dogs.
In training, we know that relationship matters- whether or not the dog trusts you to know what you mean and be consistent, whether the dog trusts you to protect him (and thus does not feel that he has to deal with a “threat”, like an approaching tall man with a weird umbrella, himself), whether or not the dog respects your ability to lead and to enforce your leadership. Those factors- the way the dog tracks your individual behavior and the way you behave in relation to him- influence how fast or well the dog will learn, as well as whether or not the dog will even bother to try.
In a broader context, sociobiologists are seeing in more and more and more and more varied different sorts of social animals (other than insects) that a hugely important factor, maybe THE important factor, to animals that can have complex socal groups is their ability to remember cheater/cooperator distinctions with individual members of their group. This is the most basic form of relationship- and also happens to require a certain advanced ability to distinguish individuals, remember them individually, and remember how they behaved toward you specifically… and if we’re going to get into this level of complexity of distinguishing individuals and keeping track of their behavior, a basic sense of self- if only to have an internal point of reference, the central individual self, to relate all those external individuals to- is a simple solution to an otherwise complex problem. In other words, likely to be a universal among diverse groups of animals facing the same fundamental problem.
Speaking of complex problems, we also have learning to rapidly cope with chaotic and novel situations, and planning- both of which dogs can do, as any shepherd knows. Human trainers use stepwise small learning events to create complex behaviors by chaining them together until the dog grasps the entire sequence as a whole, but wild animals don’t work like that. A wild dog learning to hunt simply cannot rely on learning to capture prey by small, digestible, simple sequences that then eventually link up into a complete behavior; the way a prey animal behaves is FAR too chaotic and unpredictable to rely on that kind of learning, because the sequence would never be repeated in the same way, would rarely even be begun in the same way. Thus, the animal must make simple plans based on the rapidly changing circumstances, and be able to think in a flexible enough way to try to solve problems as they arise rather than repeating stereotyped sequences of behavior and varying them slightly.
In order to plan, you have to briefly project yourself and the other object into the future- only a few seconds or minutes at a time, but even so- it seems to me that this relies on HAVING a sense of yourself as an individual whose actions are under your control, rather than having a series of instinctual or rote responses for every possibility, which would actually be a vastly more complicated and inefficient system. So now we have self as an elegant solution to TWO problems (or three, when you add trying different ways of solving novel problems in novel situations, without necessarily having to develop a plan), which makes it even more likely to be a universal. Human trainers have to use the much more simple and stepped approach to teaching because, unlike a group of dogs on a hunt, they share no common language or frame of reference with the creature they’re trying to teach- thus, reversion to the simplest and most common shared vertebrate mechanisms of learning is the best approach.
We used to assume no animal but humans had a sense of humor, until we thought to thoroughly test that assumption, and we found otherwise. We used to think no other animal used tools, until we looked and found multiple examples. We used to think no other animals had culture, until we looked and found multiple examples across many intelligent and social, but otherwise unrelated species. (Obviously not culture as in Shinto and opera, but behaviors, innovations, and mannerisms that varied with local groups.) If the argument I’ve made suggests that other animals than dogs, perhaps some that are actually less intelligent, have a sense of self… well, I’m really not all that convinced we’ve looked all that well.
October 29th, 2008 at 10:23 pm
Very well done - and, of course I agree with much of your position.
But… regarding the ability to keep track of cheater/cooperator relations - it might be interesting to look into game theory for perspective. An elegantly simple solution to the prisoner’s dilemma that basically involves giving back what you got (Tit for Tat), is incredibly successful - adaptively speaking. A brainless computer (after programming by a human) handles it quite well. [sidebar - somewhere here I have an excellent book reference to that, but my library is now small and I can’t remember it off the top of my head]
Kinda makes one wonder if it’s not that they’re (i.e. dogs, crows, porpoises) not smarter than we think - but maybe that we’re dumber than we’d like to believe.
With regard to your comparison of human back-chaining type training to the way wolves learn to hunt prey - I think the analogy your chose is not right for the situation. The very structured, step-by-step process of back-chaining is, IMO, an utterly human invention of somewhat limited use (tho I will admit that I find it very handy for some things). I am not aware of any parallels to this kind of training in the animal world. I think it would be more accurate to compare how wolves learn to hunt by, well, looking at how we learn to do it. Ferinstance: I started out as a young pup being first shown a gun and told what it was while, in the same time period, being also shown various kinds of game animals in the wild and on our table while having it explained to me that proper use of the gun brought that food to us. As I got older, I was shown how to use a gun (shoot, clean, carry, store) and how to find and observe game. In other lessons, I was shown how to butcher and cook it. The last bit was putting it all together to safely take a gun to field, find game, kill it, clean it, butcher it then cook and eat it.
This is probably a pretty good approximation (minus the cooking and guns parts, of course) of how a wolf pup is taught to hunt.
And when you look at the process *that* way - isn’t it a pretty close approximation to… culture?
October 29th, 2008 at 10:29 pm
*&$%# (llloooonnnggg day today….)
That should have been “my library is NOT small”
(husband jokes about “the books that ate our house”)
I just buy more.
If you want reference, just let me know.
October 30th, 2008 at 8:03 am
This reminds me of an article I read a while back. Zoo elephants were shown a mirror in which they could see their reflections, and they definitely had a sense of self. If some paint was daubed on their face or ear, where they couldn’t see it, when facing their reflection they’d touch the spot with their trunk, not in the mirror, but on their face/ear/whatever. Fascinating stuff! I agree that we humans have routinely overestimated ourselves and/or underestimated all the critters out there. But we’re learning more every day.
October 30th, 2008 at 8:11 am
You might want to re-think the supposed universality of “I”.
Taking a chapter out of Julian Jaynes’ work, try comparing the Iliad to the Odyssey … how the concept of “I” is handled is very different from the earlier Iliad.
The invention of narrative and writing has had some profound effects on “I”, and the ability to abstractly think about the thoughts of others.
October 30th, 2008 at 11:33 am
This is all good and I have to cogitate on it some.
Meanwhile, I wonder whether Ataika’s new “trick” of learning to hunt in a woodland environment with a kind of hawk that is unfamiliar, with whom she must work very closely, has any relevance. She did it for the first time this week and in half an hour and one rabbit chase had it down.
She previously worked with falcons in open country- utterly different- and with the gun on quail and rabbit within shotgun range. Larger point being that she is a Kazakh saluki- “tazi” of field background, and wasn’t “taught” anything. Her breeders in Kazakhstan alleged she would figure it all out by herself including gun, and she has.
Link here, with photos.
October 30th, 2008 at 1:07 pm
I do suspect that the fact that Ataika so quickly picked up this new skill is related to the fact that this work came naturally to her. In the time I’ve spent working with dogs I’ve seen that - even though they may seem to *us* to be more difficult/complicated - breed-related skills like herding sheep, hunting game, tracking and the like, make much more sense to dogs than ‘basic’ obedience skills like heeling or sit-stays. A stock dog can take beginning skills in balancing stock to the handler and very quickly translate that skill into driving stock with the handler. Most learn this new skill pretty effortlessly (if the handler doesn’t muck it up). OTOH - The dog that takes heeling skills and translates them into other work *on his own* is an incredibly rare beast.
October 30th, 2008 at 2:25 pm
Dogs are actually somewhat of a special case because they’ve been co-evolving with humans in a very intimate manner over a considerable period of time. One thing that is clear (or ought to be) is that complex systems can’t be properly understood solely by a reductionist, atomic approach. And the evolutionary development of humans and domesticated animals (most especially dogs) has had profound effects on both - to the extent they can’t properly be considered in isolation of each other in the larger analysis. See, for instance,
http://www.physorg.com/news92339486.html
Also, it appears to me that you imply something that may not in actual fact be true: that humans are superior in all cognitive abilities in all respects to dogs or even all other animals. This may not true. As an example of this, there is some evidence chimps have superior short term memory to humans.
Personally, I’m sure the cognitive abilities and inner personal world of a dog is vastly richer than we dare imagine - but, as you say, it’s essentially ineffable because we’re not dogs and for as much as we’re alike we’re also very different.
I reckon I’ve known dogs that are clearly alot smarter than alot of people I come across, for sure… is it just me or are people on average getting dumber and dumber as time wears on…
October 30th, 2008 at 2:51 pm
These are great comments- I may get an entire other post out of them, plus the discussion that it originally came out of. However, today I feel like I ripped off a layer of my cortex… totally fried! So it’ll have to wait a bit…
Smartdogs- The back-chaining thing is an artifact of the original context of the discussion… I was answering a trainer, and trying to show why that same style of learning didn’t fit with a wild hunting scenario. If you want, I can e-mail you a link to the original discussion.
Oh, and Steve- Ataika and the hawk were very much on my mind when I wrote this.
October 30th, 2008 at 6:23 pm
Yes, when your cortex recovers please send link. I love to dog training debates. You’ve got my addy.
October 31st, 2008 at 9:05 am
Does Zydeco have one of these yet?
November 3rd, 2008 at 6:48 pm
Smartdogs- OK, I’m back and a full weekend of mindless labor has allowed me to recover some…
First, with regard to game theory and simple tit-for-tat rules, I think this is more demonstrative of the principle that despite some who seem to believe there’s some sort of physical “conservation of complexity” law, you can get extremely complex behavior out of simple rules. Second, life in a social group is not like a game, with defined rules and a self-limiting nature for the interaction; you have to live with these individuals (and any others that arrive) and the task of recalling a list of whether individual X gave you tat or tit and how much of it eventually gets to be fairly gargantuan even if the rule remains as simple as that- that’s a LOT of interactions, with differing levels of cooperation and cheat. If you punch me in the face and take my lunch, you can’t make up for it by giving me a jellybean.
Also, I too have noticed the way dogs learn differently based on whether the behavior is a natural one- it makes a huge difference in teaching the Akita-monsters to do anything. Kodos has excellent house manners and required minimal training- “don’t damage my things, all things are mine unless I make it clear otherwise”, “don’t shove”, “don’t soil in the
den” were all things that made sense to him. When I asked him to grasp heeling and “finish” in class… well… that was a far, far different matter. (He still won’t do “finish”, but it’s not because he never eventually got it so much as he seems to think it’s a silly thing to do. Kang, who has no dignity that it would fall beneath, has a different opinion.)But, even natural behaviors have a very large degree of learned component- as you certainly know, since the hunting dog and the stock dog do go through learning curves, they just require far less granular human assistance to learn. Perhaps the fact that they can learn apparently nonsensical things with which they relate to another species is, in and of itself, evidence of a developed sense of self and recognition of another’s perspective?