Saturday, June 8

Animal Behaviorist: We'll Soon Have Devices That Let Us Talk With Our Pets on earthquake prediction

We're fast approaching the point, says Con Slobodchikoff, when computers will help to mediate our communications with animals. We all try to talk with animals, but very few of us do so professionally. And even fewer are trying to build devices that could allow us to communicate with our pets and farm animals.
Meet one person who is trying to do just that: Con Slobodchikoff, a professor emeritus at Northern Arizona University, and a modern-day Dr. Doolittle. Slobodchikoff is an animal behavioralist and researcher who has devoted his career -- 30 years of it, at any rate -- to the decoding of animal communications. And though Slobodchikoff has studied those signals across different species, he has focused his original research on the communications of the prairie dog. The creatures, he says, talk to each other using "the most sophisticated animal language that has been decoded." The animals have word-like phonemes, combining those into sentence-like calls. They have social chatter. They can distinguish between types of predators that are nearby -- dogs, coyotes, humans -- and seem to have developed warnings that specify the predators' species and size and color. 
To arrive at those findings, Slobodchikoff relied on statistical analyses of the alarm calls produced by one particular species, the Gunnison's prairie dog. He cross-referenced the acoustic qualities of the animals' cries with the circumstances in which they were uttered, using the calls' natural contexts as clues to their meanings.
For a detailed (and totally delightful) guide to the prairie dogs' different alarm calls, see this Radiolab interactive. Meanwhile, the video below offers a summary of Slobodchikoff's research and findings. My conversation with him, lightly edited, is below.

Lots of animals have communications mechanisms. Whales have sonar; apes have signs; birds have chirps; bees have odor plumes ... and on and on. So why study prairie dogs, in particular? What do they offer that other animals don't?Slobodchikoff: I personally think that whales and dolphins and monkeys are going to be shown to have very sophisticated languages. But we still need to design some of the experiments that will get at that. That's a problem with studying things like whale and dolphin calls, because they occur below the ocean surface where we can't really see what's going on. So we can hear their vocalizations, but we aren't really sure what their context is. And you need to have the context in order to crack the code -- it's the context that allows you to decipher the meaning of the message.
As for prairie dog language, I stumbled into that work by accident. I started looking at the social system of prairie dogs -- and prairie dogs have a very complex social system. They have alarm calls, which they give when they see a predator. And the alarm calls turned out to be a Rosetta stone for me, in the sense that I could actually decode what information was contained within the calls.

And the prairie dogs were, conveniently, on the surface.
Slobodchikoff: Yes -- there they were, on the surface. And fortunately, they lived in colonies, and the colonies never moved, so I could come back to the same colony, day after day, and note who all the individuals are, and be able to do field experiments -- which a lot of people who work even with monkeys can't do very easily, because the monkeys move from place to place. You come back the next day, and they're not there -- and who knows where they are.

So what were the mechanics of those studies? What was your sample size? How did you actually keep track of the individuals you were observing?

Slobodchikoff: In the particular prairie dog town that we studied, there were about 100 individuals -- so we had a large sample size. And we kept track of the individuals by trapping them in live traps. We "paid" them with sunflower seeds, which they really love to eat. (They would compete with each other, actually, to jump into our traps and eat the sunflower seeds. I saw moms push their babies out of the way so that mom could jump into the trap and get the sunflower seeds herself.)
And then we marked them with fur dye, identifying all the individual animals. So when we observed them from a distance, we could tell, "Oh, this is animal A4," or M7, or something. And we knew something about their basic history, and so on.

And the colonies are actually called towns?

Slobodchikoff: They're actually called towns.

I'm interested in this idea of phonemes and sentences playing out with prairie dogs -- discrete sounds the animals seem to combine to make meaning. Do you have a sense of the potential iterations that could exist there, given what we know so far about the semantic value of those units? How many possible "sentences" might there be?

[optional image description]
Slobodchikoff: You know, I really don't have a sense of that, because each experiment we do brings new surprises. For example, the last experiment that we did was showing the prairie dogs abstract shapes, like circles and triangles. And I absolutely had no idea that they would be able to come up with words for "triangles" versus "circles." It's just amazing to me: the more we study them, the more sophisticated the system becomes. So I think that we're just plumbing the very surface of things, and we'll find that their language is far more sophisticated than even we know right now, today.Plus, the prairie dogs have all of those social chatters, which we can't crack yet because we don't have a context for them. One animal just goes, "chatter-chatter-chitter-chatter," and another animal in the colony goes, "chatter-chatter-chitter-chitter." We can show that the chatters and the chitters differ, but what it means, we don't have any clue. It could be just, "chatter-chatter-chitter," or it could be, "Do you know where Sam was last night?"

And the jump-yip! I'd love to know more about what the jump-yip actually means -- if it means anything.

Slobodchikoff: Yes, and we don't know. But the jump-yip is really interesting because it's used in so many different contexts. And one of the things that people have not looked at is whether the jump-yip's acoustic properties differ from one context to another. Because it sort of sounds the same to us, but then the alarm chirp that I started working with sounded the same, at first, too, for coyotes and for dogs and for humans. Now, I can take anybody out and I can point out in about half a day what the differences are. And if they have any sort of musical ability, they can say, "Oh, yes, that's a coyote," or "Yes, that's a dog," and so on.
But nobody has really done that stuff with the jump-yips. And some of the time, the prairie dogs just do the jump-yip when it looks like there's no other context except them feeling good.

So, besides the jump-yip research that obviously needs to be conducted, what other kind of work would you like to see being done in this area? With prairie dogs, or just with animal communication more generally?


Slobodchikoff: Well, one of the things we're starting to do with prairie dogs is something that I think can be extended to animal language in general. A computer science colleague of mine and I are using artificial intelligence techniques to keep a computer record of the call that the prairie dogs were making, analyze it with these AI techniques, and then spit back the answer to us, which potentially could be in English. So the prairie dogs could say something like "thin brown coyote approaching quickly." And then we could tell the computer something that we wanted to convey to the prairie dogs. And the computer could then synthesize the sounds and play it back to the prairie dogs.
So I think we have the technology now to be able to develop the devices that are, say, the size of a cellphone, that would allow us to talk to our dogs and cats. So the dog says "bark!" and the device analyzes it and says, "I want to eat chicken tonight." Or the cat can say "meow," and it can say, "You haven't cleaned my litterbox recently."
But if we're going to get to that technology, it's going to take some research. And it's probably five to 10 years out. But I think we can get to the point where we can actually communicate back and forth in basic animal languages to dogs, cats, maybe farm animals -- and, who knows, maybe lions and tigers.

It's fascinating, thought-experiment-wise, to consider what that might mean for the whole relationship between humans and animals. Paradigms would be shifted, for sure.

Slobodchikoff: Yeah. It would be world-changing. Consider that, for example, 40 percent of all households in America have dogs, 33 percent have cats -- at least one cat, at least one dog. And consider that something like 4 million dogs are euthanized every year because of behavioral problems. Well, most problems are because of the lack of communication between animal and human. The human can't get across to the animal what the human expects, and the animal can't get across to the human what it's experiencing. And if we had a chance to talk back and forth, the dog could say, "You're scaring me." And you could say, "Well, I'm sorry, I didn't realize that I was scaring you. I'll give you more space."
What I'm hoping, actually, is that down the road, we will be forming partnerships with animals, rather than exploiting animals. A lot of people either exploit animals, or they're afraid of animals, or they have nothing to do with animals because they don't think that animals have anything to contribute to their lives. And once people get to the point where they can start talking to animals, I think they'll realize that animals are living, breathing, thinking beings, and that they have a lot to contribute to people's lives.

And is it fair to assume that language will be a vehicle for that partnership -- in other words, that most of these animals do actually have language, in some capacity? Even if you narrow the pool down to, say, just domestic dogs and cats, do you feel pretty confident that, should such a device come to fruition, we'd actually have animal-language data to seed it with?Slobodchikoff: Yeah. Cats have something like 35 vocalizations. Plus, they have a variety of body language signals. Dogs also have body language signals. They have a variety of different vocalizations with barks. Both dogs and cats also use odor, which we're not very good at detecting -- we don't really know what odors mean. And I don't think that, in the near future, we'll be able to use odors in the computerized analysis that I was describing. But we certainly would be able to use sounds. And we probably would, now, with video-capture techniques and facial recognition technology and so on, be able to monitor body language as well. So I think that at least on two of those fronts, we'll be able to get somewhere.
As for other animals, a lot of them either have clear-cut language, or at least are pointing to the possibility that they have language. So at this point, it's premature to say that all animals have language, because we simply don't have that information. But I can say that a lot of animals have language.

Why, then, have we resisted that idea so strongly? Why do we talk about "animal communication," but not "animal language"?
Slobodchikoff: If you talk to most biologists, philosophers, and linguists, they will tell you that we humans are the only ones who are capable of language. And all the other animals are incapable of that -- all they can do is communicate. So we have that kind of bias, generally speaking. It's not an accepted thing to talk about in biological and philosophical and linguistic circles. It could be just, "chatter-chatter-chitter," or it could be, "Do you know where Sam was last night?"
So the bias comes from a kind of possessiveness when it comes to language -- the claim that language is a fundamental part of what makes humans, ultimately, human?
Slobodchikoff: Right. I think that, for the most part, there is the thought that we humans have to be really special -- and language is part of what makes us special. Back when I was a graduate student, people used to talk about (at that time, quaintly) "man" as a tool-user -- the only one who was capable of using tools. Well, then we found that lots of animals could make tools, as well. So then the story shifted: humans were the only ones with culture. And then we found that lots of other animals have culture. So then we had language as the only other thing that distinguishes us from other animals. And now we're finding out that lots of other animals have language. So the idea of animal language rocks the world of people who would like there to be a big gulf between humans and the rest of the animals, and who would like humans to be completely special.
But, you know, I tell people: We are special. You don't see whales and dolphins having conversations about nuclear physics.
Right. (That we know of, at least!) And these ideas -- talking to animals, and thinking of them as capable of talking in the first place -- would seem to be liberating from a scientific perspective, too. There's so much to learn: To what extent, for example, do animals experience emotion? Obviously they do ... but how can we actually understand the nuances of that from our human perspective?
Slobodchikoff: Absolutely. In my book, I present a new theory called the "discourse system theory" -- where I suggest that we've kind of been barking up the wrong tree. We've been looking at the signals that humans put out and that other animals put out, and think that that's language. But what we really should be looking at is the whole biological system that's involved in language production, language reception, language interpretation.
So in humans, for example, we have all of these specialized structures for language: we have vocal cords, we have a larynx, we have specialized structures in our brain, our lungs are adapted for manipulating air in certain ways. And when you look at other animals, they have similar kinds of structures that are adapted for producing these signals. And once we look at that, language makes more sense from a neurobiological and anatomical evolutionary standpoint. Once we start looking at the continuity of these systems, we can see the evolutionary continuity. And we can see that we're not alone there.
I think we can get to the point where we can actually communicate back and forth in basic animal languages to dogs, cats, maybe farm animals -- and who knows, maybe lions and tigers.
It does seem productive, science-wise and otherwise, to frame things in terms of a connective continuity between animals and humans. When it comes to communication, it's Darwin, all the way down.Slobodchikoff: Yeah. For example, one of the things that we humans do is use body language. And studies have shown that when spoken language and body language conflict, the listeners pay attention to the body language, not to what's actually being said. So there are a lot of parallels between what we do and what other animals do. We just, for the most part, have been ignoring that because we make the assumption that other animals can't have language. So we don't look for it.
But actually, it's there, in the scientific literature. The authors of the papers I refer to didn't call it language -- but when you look at it from a language perspective, it really is about language. And we have a lot of information about it.

No comments:

Post a Comment

You can write whatever you want, but technically, the precise mathematical-statistical forecast of earthquakes based on information about the behavior of wild and domestic animals, birds, fish, and individuals available from 1995, with the advent of social networking.

THE STRUCTURE OF INPUT BIG DATA: API applications to social networks