The Study of Animal Languages

by Lindsay Stern

  Received wisdom holds that language is unique to human beings. In technical terms, this is another way of saying that syntax and displacement are unique to human language. With the exception of bee dances, which articulate the location of distant food, no other animal communication system has been shown to exhibit syntax, or to refer to objects outside the being’s immediate environment. Most linguists therefore consider their sounds—hissing, lowing, barking, and so forth—to be spontaneous responses to stimuli, no more significant than human laughter. This deficit helps explain why animals cannot think. It helps explain why they have not ascended, as we have, into the light of reason, but remain shackled, so to speak, to the walls of Plato’s cave.

  The study that brings me here today challenges that assumption. We found that zebra finches, like humans, can discriminate between different configurations of the same units of sound. Our method was simple. We began by recording a phrase of birdsong, and then divided it into its elementary units, known coincidentally as “notes,” which we labeled A, B, C, and D. We were able to isolate these units with the help of a spectrogram—a map of sound waves—on which the notes showed up as continuous traces along a sinusoidal curve. After that, we reordered the notes to build three new phrases—BCDA, DCBA, and CDAB, respectively—which we played back to the birds. Some of the birds listened to all three phrases in silence. But most chirped at the first, flew away at the second, and squawked—with great animation—at the third. That suggests that they are sensitive not only to individual notes, but also to their configuration. It therefore provides the first experimental evidence of syntax in a nonhuman vocal system.

  Our data do not prove, however, that finch songs qualify as language. To cross that Rubicon, my colleagues and I would have to discover unambiguous evidence of displacement in their melodies; to show, in other words, that—like human words—their notes can refer to objects outside their physical vicinity. There are a variety of ways we might test for this, though none are very promising. One option would be to present a group of finches with the same familiar object—a caterpillar, say—and record their vocal response. If we were to find the same note or note-sequence across all of their vocalizations, we could define this phrase, provisionally, as the finch term for “caterpillar.” We could then remove the caterpillars from their environment and analyze their subsequent melodies for that same phrase. If it recurred, we would have some evidence of their ability to “talk” about absent phenomena, and therefore of displacement in their vocal system.

  Yet for that evidence to carry any force, we would have to establish that the phrase of birdsong actually signified “caterpillar.” This is much easier said than done. Even if every finch uttered the same phrase in response to a given caterpillar, how would we know it wasn’t saying “food”? Well, we could confront it with another treat—millet, for example—and listen for that same phrase. If the phrase recurred, we would have some evidence that it signified “food.” But what if the finch repeated the phrase in response to a predator, pest, or other stimulus? How, moreover, could we be sure that it meant “food” and not “this looks good,” or “this looks gross,” or “thank you,” or “damn you,” or “let me out of here, you sick bastards”?

  We scientists therefore face a paradox hitherto confined to philosophy. How do you translate a word beyond reasonable doubt without speaking the language in question? Even if we establish a strong correlation between a given phrase of birdsong and a physical object or fellow animal—as Vincent Janik, Laela Sayigh, and other cetacean biologists have done in their work on dolphins’ signature whistles—we would have to assume that the phrase signified a physical object in the bird’s immediate vicinity. But what if it instead signified a thought? There would be no way to establish this, empirically, because—unlike caterpillars—thoughts are unobservable. We would have to learn their languages ourselves.

  That may sound like a pipe dream. But one day, the notion that the animal voices around us are talking gibberish may come to seem equally far-fetched. Whatever they may be saying, our study on syntax is hardly the first indication that it may be something more than “feed me” or “mate with me.”

  Consider, for example, the recent analyses of dolphin vocalizations using Zipf’s law, the method researchers have long used to search for extraterrestrial intelligence. The method was invented in the 1930s by George Zipf, a linguist who discovered a pattern shared by almost all human languages: in speech, the most common word—in English, “the”—occurs twice as often as the second most common word, which occurs three times as often as the third most common word, and so on. When he plotted these words logarithmically, Zipf found that the relation between them yields a line with a slope of -1. Today, researchers at the SETI Institute use Zipf’s method to test incoming signals from outer space, searching for gradients of -1 in the hope of identifying intelligent aliens.

  To model what an alien language might look like, astronomers have applied Zipf’s method to the sounds emitted by dolphins. To do so, they mapped each dolphin’s squeaks and whistles visually, using a spectrogram, and then looked for natural gaps in the resulting sound wave. These gaps enabled them to divide the dolphin’s voice into units, which they then mapped logarithmically. The resulting slope was -.95, virtually identical to that of human speech.

  There is a limit, of course, to what approaches like Zipf’s can teach us. Like us, the astronomers succeeded only in plotting animal voices mathematically. Were you to apply the same method to the words I am speaking now—to record and map them visually, along a sine curve—the image would reveal no hint of what I have been saying. It would expose every nuance of my voice, but it would not compel you to think my thoughts. Unlike my words, it would provide no evidence of a mind inside my body.

  Nor would it enable you to prove that I, too, could talk about the past and future, or God, or other abstractions. None of these things are empirically demonstrable. Nor is language, for that matter. I may be talking about it now, but given that my vocalizations have nothing to do with what is actually happening—nothing to do, for example, with the expressions on your faces, or the hunger in my belly, or the texture of my shirt—no scientist worth her salt would be able to show, using our methods, that I am speaking at all.

  I find myself in her position, with regard to my subjects—the small, quizzical faces I confront every morning, in my laboratory here on campus. I must cut a strange figure to them, scuttling around with my clipboard, blasting messages out of a small black box. Day after day, they watch my colleagues and me pore over these messages. They might find it amusing, if they experience such a thing, to know that we have no idea what they mean.

  Our position is curious. Some future species may share it one day, when they discover the ruins of our libraries. If the oceans haven’t wiped the volumes clean by then, they will find traces not only of our bibles and constitutions, but also of our symphonies. Perhaps they will take an interest in these primitive dots and lines—measuring their width, height, and so on—and believe that in doing so, they have discovered music, even as they work all the while in silence.

  If we succeed, one day, in translating birdsong into English, language will share the fate of that other fallen idol: tool use. It will no longer provide evidence of our superiority, but rather of our kinship with the beings that surround us. Will we find other ways to convince ourselves that we are the enlightened animals, when we are no longer the only animals that speak? What, if anything, will distinguish us?

  Perhaps the finches will tell us. They may have a ready answer, though I doubt, for my part, that we will like what they have to say.

  Six

  The applause begins cautiously at first, then swells. She collects her papers, blushing as someone whistles.

  “So that’s what she’s had up her sleeve,” Quinn whispers.

  I force my hands together a few times, my fingers numb, as the room floods with queas
y yellow light. Then I stare at my knees, the twilled cotton of my chinos. She has lost it. She has lost her mind.

  “We have time for a few questions,” Noboru calls out. “A microphone should be circulating—Lucas, can you—? Oh, terrific. Thanks.” He scans the room.

  People glance at one another. I will them not to speak, or laugh. Good God, let this be over, I think. Then I remember the party.

  “Right there, in the back.” Noboru gestures.

  A young woman in a toffee-colored blazer stands up, and the student hands her the mic.

  “That was super-interesting,” she says, “but I’m a little confused. How do you plan to decode the finches’ tweets, or, sorry, their cheeps”—scattered laughter—“without a Rosetta stone, or something? Also, if they really do have thoughts, how do you feel about keeping them in your laboratory?”

  Prue is a slight person, and in the glare of the spotlight she looks even thinner. I hold my breath.

  “Would you like my job?” she says. “Because you’ve just posed the two questions that haunt me every day. I may have shown that the finches’ vocal system has syntax, but by what stretch of the imagination have I convinced myself that I could ever find a way into its semantic register? And even if I could, what makes me think the birds would want to speak to me, their prison warden?”

  There is a round of uneasy laughter. But she continues, unabashed: “That second question is the reason I only keep any group of birds for three months, at most, before releasing them back into their home regions. As for the question of how we might translate their songs, I really don’t know. There are times when the attempt strikes me as a fool’s errand. But the notion that a vocal system would have syntax, but no semantics, seems even more absurd to me. We’ve found strong evidence that these birds are saying something. I want to find out what.”

  “How?” someone calls out. A young voice, probably a student’s. The dean turns and glares. There is a tittering.

  “Shall we move on?” Noboru suggests, but Prue lifts her hand.

  “It’s all right. I was just going to add that, with the help of our heroic librarian”—she gestures toward someone in the first row, who nods deeply—“Amina Singh, we have published an expanding online archive of their melodies. The data is public, available to any linguists, cryptographers, musicians, and other researchers interested in analyzing them.”

  She smiles brightly, as though she has not just accused her colleagues of torture. As Noboru calls on someone else, a familiar, unkind thought strikes me: As smart as Prue is, as graceful as she is, there is something ridiculous about her.

  “You mentioned Sue Savage-Rumbaugh’s research on language learning in apes,” the next questioner says. Twisting around, I find it is Clarice Hussein, a colleague of mine who works on Hume. “I’m curious: Why not go that route? Instead of treating birdsong as a text to decode, why not create an artificial lingua franca and try to communicate with your finches that way?”

  As she sits back down I catch sight of Walt a few seats away from her, texting. May, pen in hand, shoots me a furtive wave.

  “It’s a great question,” Prue says. “And I do admire that work. It’s profound for a number of reasons, especially in demonstrating that the qualities that seem to make us most human—the capacity for symbolic communication, for instance—are not uniquely human at all. But it doesn’t address nonhuman vocal systems on their own terms. That’s what I’m after. If we could translate these vocal systems, we could test the Sapir-Whorf hypothesis anew, with respect to animal languages. Does one’s language dictate one’s reality, as Sapir and Whorf claimed it does for human tongues? What would the world look like, through the lens of birdsong?”

  A man rises and edges toward the aisle, forcing half a row of people to stand. Heads turn. It is Jeremiah Wood, also a biologist, who championed Prue’s application to the College five years ago. Prue does not waver, though I see her see him clear the gallery and exit through the double doors.

  “The political ramifications would be Copernican,” she says. “And very uncomfortable, I would guess. Like the doctrine of manifest destiny, our dominance might begin to strike us as a vast, coordinated crime, rather than a logical expression of our superiority.”

  I begin to feel the pressure of other eyes. An administrator near the door glances gingerly in my direction. Another colleague, three rows ahead, meets my gaze before feigning interest in the wall clock. Her husband, I imagine him thinking. How must it feel to watch her bite the hand that feeds her?

  The next question comes from John Sawyer, professor emeritus of linguistics: “Yes, good afternoon and congratulations on your findings. May I suggest, with all due respect, that you are jumping the gun here? Syntax and displacement are only two of the many features a vocal system would have to exhibit in order to qualify as language. I have read your study. Your team showed that finches have a limited ability to construct phrases and identify differences in sound order, presumably in order to attract mates, repel rivals, and establish territory. You offer no evidence that birds can generate new sounds or new configurations of sounds, let alone in a recursive fashion. As for the research Clarice mentioned, let us not forget its fall from grace after a brief heyday in the seventies. The vast majority of ape language experiments were intellectually bankrupt, archetypes of the Clever Hans effect.”

  No tittering, this time. Only a shifting, the murmur of bodies. Noboru manages a look of vague bewilderment.

  “Has anyone heard of Clever Hans?” Prue says. A student in the front row nods maniacally. God knows what she is after now.

  “He was a horse who gained fame in the early 1900s for apparently solving arithmetical problems by stamping his foot. It turned out that what he was actually doing, unbeknownst to his trainer, was watching his human audience so closely that he inferred from their body language when to stop stamping. That the scientific community concluded that he was a fraud, rather than an emotional savant, is a perfect example of the kind of hubris I tried to bring out in my lecture.”

  “Are you defecting from science, then?” Sawyer calls out, sparking another bout of rustling. “Forgive me, but I don’t understand.”

  “I love science.” Prue draws a breath, finally ruffled. “That’s why I think it’s worthy of critique. Mapping birdsong mathematically will never prove that birds have language. As accurate as our spectrograms may be, they provide no information about whether the finches’ melodies meet the standards you mentioned, John.” She smiles to herself. “My husband works on this, actually—it’s a problem in epistemology: the gap between our data and the truth.”

  Quinn nudges me, and I shrink down in my seat. Please, my love. Don’t drag me into this.

  “That old fable, about a group of blind men touching different parts of an elephant?” she presses on. “The ones touching the tusks arguing that it’s hard and smooth, the ones touching the end of the tail saying it’s coarse, and so on. . . . Well, I think our species represents one of those blind men, confined by our language to one region of the elephant, one view of things. We have a philosophical imperative to change that.”

  A wet, ropy cough issues from behind me. Without looking, I can tell that it has come from Frank. Here is his chance to interpose himself, deflect attention from her, something I never thought I’d wish for.

  But he doesn’t, and now Noboru intervenes: “We have time for one last question. . . .” A hand shoots up ahead of me, and he brightens. “In the middle, yes.”

  A shuffling, as the microphone is passed.

  “I think it’s magnificent,” the questioner booms. “Your implication—that we’re not the wiser animals, but the demented ones. That insanity, not wisdom, explains our dominance.”

  Dalton. His legs are crossed, one arm draped across the back of the neighboring seat, probably occupied by someone he doesn’t even know.

  “Not Homo sapiens, bu
t Homo infirmum,” he adds. “Why else would we have vandalized the earth?”

  Give me a fucking break, I think. As I close my eyes Quinn whispers, “He’s great.”

  “You’ve raised the question of what other animals might say to us, if we would listen,” he continues. “What, I wonder, would you say to them?”

  “Ah . . .” Prue smiles mischievously. “Good question. I don’t know. Forgive me?”

  The reference to his book title prompts a ripple of warm laughter from the in-the-know.

  “Fair enough,” he says. I could break his knees.

  Noboru announces the conclusion of the lecture, triggering a final spasm of applause. I am out of my seat before it is finished, jostling past Quinn and up the aisle. May and Walt are putting on their coats, but Frank is still gazing ahead, beatific, hands steepled under his chin.

  “Let’s get out of here before the rush,” I hear myself say, handing May her backpack. To Frank I add: “Where did Prue park?”

  He ignores me. The woman beside him, trapped by his knees, taps his shoulder.

  “Uncle Ivan, guess what!” May brandishes her notepad. “I made—”

  “Do you?” I interrupt, turning to Walt.

  Walt shrugs. “Shouldn’t we . . . ?” He gestures toward the podium.

  “I’ll meet you outside,” I say, giving up on them. Dodging bodies, I charge through the side exit and out, into the whiteness.

  Part II

  Seven

  Dear Prue,

  Pardon my absence at tonight’s gathering. I had been looking forward to celebrating you. Noboru spoke for the Department in praising your team’s experiment, and we all know what a pivotal role you played in designing it. Your discussion of the results would have been the capstone of a distinguished tenure file.