Theory of Mind

This post reminded me of what happens in the backyard at home when the dogs are thinking about getting up to mischief they know is going to get them banned to inside if they get caught. You can stand in the kitchen and watch them through the kitchen window. The moment they can't see you through the glass door in the room adjoining the kitchen, they'll sneak off down the backyard with the intention of poking around the fences for a way to escape to go do something more fun. They'll get halfway down the yard in their sneaky poses when you'll yell from the house something along the lines of "Pyry! Come back here!". The dogs all halt and stare up at the house, but they can't see us. They look around them as if trying to decide whether we really can see them or if was a fluke and they can keep sneaking down the yard. Another shout will make up their minds that humans are magic and can see them possibly at all times and they come trotting sheepishly back to the house to gaze through the glass door looking for the people.

Now, obviously this is all a bit of a stretch scientifically and it's heavily interpreted and possibly incorrectly, but this scenario makes me suspect they might have a theory of mind because they seem to think if they can't see us we can't see them. As I understand it, if they didn't have a theory of mind they would assume we knew everything they knew, thus deception is impossible and all the sneaking around and waiting until they can't see any humans before heading off to make mischief wouldn't occur. But then, maybe they just learnt that if they sneak when no one is visible, they sometimes get away with it and it has nothing to do with whether they think they can be seen by someone else or not. Or in another interpretation, they don't have a theory of mind BECAUSE they don't think we can see them if they can't see us. The old "if I cover my eyes so I can't see you standing right in front of me then I'll be invisible to you, too" thing wee little kiddies love to indulge in.  

We get the same thing, Corvus. When Shadow is out in the yard and doing whatever, he will look at the kitchen window. When I first got him, he was all about the eyes. I could be mostly visible but if I was standing so that my face and eyes were hidden, he would run about until he could see my eyes, again. It was so important to him to see my eyes. As I have read and learned, that is part of the Sibe traits.

So, maybe it's a bit of both. The dog looks to us for cues, documented in more than one experiment, may understand or learn to understand what eyes closed means to our perception, but even that may not prove that dogs know we have a mind separate from them, i.e., dogs having their own theory of mind regarding humans. Which then, for me, leads to the question:

If dogs can have a theory of mind, how would we know, aside from behavior regarding us that could very well have been learned? I feel dogs are more intelligent and complex than amoeba. And it's not difficult to imagine they have theory of mind, or that may be hitherto unseen anthropromorphization peaking up, again. The other problem is that whatever we discuss has its own limits. Here, for example, the discussion is in the language of English with people mostly from english speaking countries. Any language has a particular limitation and a particular freedom and is not totally disaparate from the psychological and social values of the culture it is in. For example, in German, the word Gemuetlichkeit (old german spelling since I don't have a font with umlauts) doesn't have a direct English equivalent but is meant to convey a general sense of contentment and happiness.

What if it were possible to catalog the language of barks and growls, not totally disimilar to Chinese, a language of tones, inflection, and context. Would there not still be words without direct translation? Could such words as we could figure out lead to a better understanding of what a dog thinks we can think?

As I understand it, if they didn't have a theory of mind they would assume we knew everything they knew, thus deception is impossible

I've seen dogs engage in deception, both towards humans and towards other dogs. They can and do lie. Without a "theory of mind" how do you explain deception and lying?

You raise a good point, MP, and I think Corvus provides some evidence of that when one dog grabs and bogarts a toy or frog, not because she really wants it but because the other dog really wants it and she wants to create a play scenario based on the reaction of the other dog. Would that not imply a theory of mind in one dog regarding another? That is, yes, they have reactions to stimuli but they can also actively control and use stimuli for their own, sometimes "nefarious" purposes? Corvus' dogs - viva la Revolucion!

"...evidence of that when one dog grabs and bogarts a toy or frog, not because she really wants it but because the other dog really wants it and she wants to create a play scenario based on the reaction of the other dog."

 My two Siberians engage in this kind of play all of the time. If you give them each a toy, the gal will not be interested in her toy.  It is most fun if there is only one of a toy, then they can challenge run and make up the game. 

Not only do dogs have a real working mind, to devise and contemplate and obtain end results, They also have a sense of humor.  My two Siberians are incredibly full of humor.  You can tell when (Usually it is the male )that may not be not interested in play, and sister wants to play...She will work around him, rubbing her face in his front paws, all the while he may be baring teeth. But she knows just what his measure is and it is so funny.  Because she knows and can manage to put him into a play mode.

To me it is not just a theory, it is a fact!  And just because animals do not speak in human tongue, they and many other creatures have complex language skills.  The more you study a creature, and observe his natural ways the more you will realize that there is more there.  Much we may never understand. 

editing in:

In fact I am pretty sure some dogs have a more solid "theory of mind" than many humans. Just look at the news! But, humans get away with what dogs are "put down" for.  So, does that make the human more or less intelligent? hmmm. 

ron2

You raise a good point, MP, and I think Corvus provides some evidence of that when one dog grabs and bogarts a toy or frog, not because she really wants it but because the other dog really wants it and she wants to create a play scenario based on the reaction of the other dog. Would that not imply a theory of mind in one dog regarding another?

Jumping in to play devil's advocate, again... 

I think it is possible to explain the deception/tomfoolery without postulating a ToM- a dog figures out that if it runs over to the door barking, the other dog will follow (can just be by association), and so "infers" that if the second dog has a good sleeping spot that first dog wants, it can get it to move by running to the door...

The same can be said about the bogarting of toys (I love that word, bogart, by the way )- the dog knows the consequence of stealing a toy another dog is chasing, sniffing, looking at, it will get chased....

I think the main question when it comes to issues like these, is whether ToM or conditioning/association is the more parsimonious explanation, a la Occam's Razor.

Vinia

ron2

You raise a good point, MP, and I think Corvus provides some evidence of that when one dog grabs and bogarts a toy or frog, not because she really wants it but because the other dog really wants it and she wants to create a play scenario based on the reaction of the other dog. Would that not imply a theory of mind in one dog regarding another?

 

Jumping in to play devil's advocate, again... 

I think it is possible to explain the deception/tomfoolery without postulating a ToM- a dog figures out that if it runs over to the door barking, the other dog will follow (can just be by association), and so "infers" that if the second dog has a good sleeping spot that first dog wants, it can get it to move by running to the door...

Exactly right.

I was discussing this issue on another board, and was given half a dozen examples, all different,not as to whether dogs have a theory of mind, but whether they can think logically.

In one instance a dog owner posted a very similar scenario to yours: two dogs, one who was an inveterate barker at any noises coming from outside the front door. The 2nd dog would bark also, but wasn't as committed to it as the 1st. The 2nd dog also "coveted" a certain toy that the 1st dog was always bogarting. The poster asked me how the 2nd dog could have developed this game she had, where every time the 1st dog had the toy she wanted, she'd growl and bark and do all kinds of funny stunts to try to get the 1st dog to drop it. When that didn't work, she'd go to the front door and start barking at it, which caused the 1st dog to drop the toy and come running. Then 2nd dog would run back into the living room and scarf up* the toy.

I read this story, and puzzled about it for a while, then it hit me: I posed the following question (not verbatim) to the original poster:

"Could there have been a specific time, where when the 1st dog had the toy, and the 2nd dog wanted it really badly, that the 2nd dog heard a noise at the front door, and in her already energized state, ran to the door to begin barking as a way not necessarily of alerting to danger, but just downloading some of the excess tension she felt about not being able to connect to the toy? If she had," I conjectured, "and her barking brought the 1st dog running, and if the 1st dog's barking was always stronger and more insistent than her own, perhaps she stopped barking at the door, letting her "packmate" take over. And since the door situation was well in hand, she was almost magnetically pulled back to the spot where she'd originally been energized. She found and grabbed the toy, and a learned behavior was born. Could that have been the genesis of this behavior?

Of course I expected this dog owner to write back and tell me I was nuts. (I get that a lot.) But she didn't. In fact she vaguely remembered that what I'd described was pretty much exactly how this toy stealing strategy of the 2nd dog got started. (In fact,if I recall correctly there was nothing in her original post about how committed either dog was to barking at the door; that was part of my theoretical analysis that, by luck, happened to be true.) And here's a funny thing I've discovered over the years, one that fit my conjecture: when a dog really, really wants something very badly, and is blocked from getting it, but then finds a way to get past that obstacle and finally attain her desire, she can learn a new behavior once, just once, and the lesson will stick with her for the rest of her life.

So, yeah. You have to whittle things down using Ockham's razor (the spelling I prefer).

It also helps your analysis when you know big a part desire plays in how dogs learning new strategies for getting what they really, really want...)

LCK 

*(in this case scarf up is the opposite of scarf down) 

 This next example of mine doesn't really prove ToM either way, but I was just thinking about some of the interesting leaps a dog's brain can make and I remembered one time I was eating an apple. Penny is food obsessed and wanted some. Now, I knew Penny doesn't really like apple, but Kit the hare loves apple and he always gets my cores. However, Penny was dancing around eagerly, staring at the apple and finally I sighed and gave her a little bit. To no great surprise, she mouthed it, then spat it out and looked at me hopefully to see if I had anything better. I tried to coax her to eat it and she wouldn't. Finally, I said "I'll give it to Kit" in a light approximation of tone I might use with her when expressing my vague disapproval of something (which typically has zero effect on her) and the moment I said that she dived on the apple and ate it. She dived with such gusto and urgency that my housemate - a talented scientist himself - stared at me in disbelief and said "I swear she knew what you just said." I'd never said that to her before as far as I can remember, or anything like it. And people tell me dogs don't understand English!

Anyway, back on topic, I'm inclined to think that deceoption in dogs probably doesn't come from them directly understanding that the individual they wish to deceive can be deceived so much as from them discovering serendipitously that certain behaviour provokes another behaviour that they want (such as dropping the coveted item). However, the case of Jill stealing Pyry's frog purely because he wanted it badly enough to chase her for it (which, mind you, almost never happens) is a decent argument that Jill knew Pyry wanted something that she didn't want, and that he wanted it very badly. The way she taunted him with that frog was pretty hilarious. I don't know that it is still a good argument for ToM, though. I think it's possible for an animal to know that another individual wants something without understanding that they see the world from a different viewpoint.

Pyry is far and away the smartest dog I've ever met (although my mother thinks his shiba inu cross vallhund predecessor was even smarter). But even he shows me sometimes that he is a creature of superstition mostly. He obviously thinks I can pluck birds out of the sky because a couple of times I've gone into the bush and caught a bird where he couldn't see me. For the next few days, he'll be walking along beside me and quite suddenly get it into his head that I have a bird somewhere on my person. Even though I haven't been out of his sight for the whole walk, and we haven't seen any birds in distress, he quite suddenly starts jumping on me and wanting to look in my arms and my clothes and any bag I'm carrying and behind me. I reckon you can see the moment the thought of birds occurs to him if you're watching. His ears suddenly *** forward, he slows down, lifts his head up, and then I know it's a matter of seconds before he decides I must have a bird somewhere. It's pretty funny, but I think it shows that he can't imagine how I catch birds when he can't see for himself. Which would make it fairly strong evidence for no ToM. Or maybe he smells birds on the air and he thinks it must be on me somewhere because I had a bird once, but that seems unlikely to me considering how many perfectly healthy birds we might pass before he decides I have a bird.

ok, what about these ones: one of my dogs decides to bury his bone, and goes outside. Other dog sees him going, and she goes out and hides behind the truck and watches him. As soon as he finishes his task and goes back inside, she goes and digs up the bone and then buries it in a different location.

Or the infamous buster cube incident: UPS guy is dropping off packages, and truck is parked in the driveway. My dog is sitting there looking at the truck, then she suddenly gets up, goes to the buster cube, and carefully pushes it so it's behind the wheel of the truck. Truck leaves, smashes buster cube, she happily scarfs up the freed treats.

Or the sock stealing: dog waits until no one is looking, dog opens closet door, takes a sock out of the laundry basket, and then CLOSES closet door and goes and hides sock under her bed. Took a video camera to figure out why all of our socks were vanishing.

corvus

Finally, I said "I'll give it to Kit" in a light approximation of tone I might use with her when expressing my vague disapproval of something (which typically has zero effect on her) and the moment I said that she dived on the apple and ate it. She dived with such gusto and urgency that my housemate - a talented scientist himself - stared at me in disbelief and said "I swear she knew what you just said." I'd never said that to her before as far as I can remember, or anything like it. And people tell me dogs don't understand English!

That's a great story. And what it tells me (or reinforces, anyway) is the idea that dogs can pick up mental images from us. In other words, it wasn't your words that sent Penny diving after that apple core, it was the mental image of you giving it to Kit that motivated her "odd" behavior.

LCK

(Both William Campbell and Rupert Sheldrake have written about this phenomenon.)
 

Well, I don't know the circumstances of how these behaviours developed, but hey, I'll offer an alternative to ToM anyway.

It's possible the dog hiding behind the truck has learnt crouching behind the truck leads to greater likelihood of being able to watch the burying of the bone uninterrupted. It's also possible the dog knows from experience that the other dog walking out with a bone means an opportunity to take the bone.

UPS guy, well that's a pretty impressive leap of logic, but doesn't really prove or disprove ToM. Crows do the same thing with nuts. They even wait at a crosswalk for the traffic to stop before placing a nut under a car's tyre or collecting the nut meat afterwards. Of course, I wouldn't at all be surprised to discover crows have a ToM.

Sock incident is rather a lot like a lot of problems I've heard people having with captive octopus! Those invertebrates are capable of observational learning, as are dogs, I think, but OL in itself doesn't really prove a ToM, and nor does the sock incident, really. Again, very sneaky, but the problem with deception, IMHO, is that it's not necessarily what it appears to be. It's all together possible that an animal learns that another individual watching them means an end to fun or a fight over food or something like that without understanding the other individual sees them and wants what they have - although that's what signals are to communicate. Anyway, from the other side, it's possible the dog just knows if they can't see eyes on them they're fun will be uninterrupted rather than understanding that anyone who might interrupt their fun can't see and therefore don't know what they're doing and won't interrupt them as a result.

I wouldn't rule out ToM in dogs, but I'm not yet convinced.

LCK, I suspect Penny just knows the meaning of a lot of words. She knows Kit's name and she probably knows "give" and possibly has some vague notion of what it means from sentences such as "give the dogs their dinner" or "give it to the dogs" which usually heralds food for someone, but not always Penny if she isn't fast enough. It's a pretty impressive grasp of what words mean if that's the case, but Penny is a VERY savvy dog and she appears to have chosen human linguist as her lifelong career (alongside "human manipulation through Cute";).

 what really impressed me about the sock thefts was the closing of the door afterwards, not doing it only when unobserved. I guess however she could have watched us and learned that after you open that door, you close it again, that's just the way it is, and it wasn't some master plan of ongoing deception. Course she also opens the front door when she wants to come in and has never closed it even though we always do so and she sees us do so.

you know, I don't think it's possible to prove a human has a TOM, let alone a dog. If children don't start displaying TOM-traits until what, age three or four? it could all be conditioned behaviors not TOM.

mudpuppy

you know, I don't think it's possible to prove a human has a TOM, let alone a dog. If children don't start displaying TOM-traits until what, age three or four? it could all be conditioned behaviors not TOM.

I agree MP, and children have to be taught a ToM.

I was just thinking about the fact that one of our common techniques for teaching bite inhibition relies on puppies having a ToM.  When a puppy bites a human and we yelp to tell the puppy to stop, we are relying on the puppy's ToM.  And the assumption that a puppy learns bite inhibition from its littermates by the same method needs a ToM doesn't it?  Why would a puppy let go when the other puppy yelps or even care if the other puppy yelps unless there is a ToM.  Indeed, being born into a litter rather than as a single birth as most humans are, would argue for dogs being more likely to have a natural ToM or learning about it earlier than humans would.

Puppies care when we yelp because it means we don't want to play anymore. I think they learn the yelp means fun will stop. They learn it from their litter mates before they even come to us.

Children don't need to be taught a ToM. It comes naturally on its own with ongoing social interactions. I've seen it tested by someone putting a button in a piggy bank, then asking the child what a friend that wasn't present would think was in the money box. Kids without a ToM say "buttons" whereas kids with a ToM say "money" and laugh. Another example is kids talking on the phone and answering a question by nodding or shaking their heads because they don't realise the person on the other end doesn't know they are making those gestures.

Mudpuppy, the sock incident reminded me of octopus because they habitually replace lids on their aquariums after escaping, and then again when they come back home! Who knows why they do? But then, it could be adaptive because octopus taste so good and everything wants to eat them. Not leaving a gaping hole behind them could help them survive. I don't really know why the dog would close the door after stealing a sock, but sometimes animals do things because they can. I watched a small cow once spend several minutes trying to turn over its water container. When it finally succeeded it nosed the bottom, then looked at the water spilt everywhere quite forlornly. I don't know what it thought it was doing, but it worked pretty hard at manipulating its environment. I feel like closing a door to prevent later discovery is a bit of a stretch because I haven't met many animals that care about being discovered after they've had their fun and mischief, but it's possible the dog might understand that to be discovered would mean the end of future fun. Although I know that kind of reasoning is beyond our little canine Einstein. He gets close, though. You can't trick him very easily anymore because he knows what trickery is going to end his fun and is good at generalising it.