by Carl Sagan
If a chick is tied to a peg by one leg, it peeps loudly. This distress call makes the mother hen run immediately in the direction of the sound with ruffled plumage, even if the chick is invisible. As soon as she catches sight of the chick, she begins to peck furiously at an imaginary antagonist. But if the fettered chick is set before the mother hen’s eyes under a glass bell, so that she can see it but not hear its distress call, she is not in the least disturbed by the sight of him.
… The perceptual cue of peeping normally comes indirectly from an enemy who is attacking the chick. According to plan, this sensory cue is extinguished by the effector cue of beak thrusts, which chase the foe away. The struggling, but not-peeping chick is not a sensory cue that would release a specific activity.7
Male tropical fish show fighting readiness when they see the red markings of other males of their species. They also get agitated when they glimpse a red truck out the window. Humans find themselves sexually aroused by looking at certain arrangements of very small dots on paper or celluloid or magnetic tape. They pay money to look at these patterns.
So now where are we? Descartes was prepared to grant that fish and poultry are also subtle automatons, also soulless. But then what about humans?
Descartes was here treading on dangerous ground. He had before him the chastening example of the aged Galileo, threatened with torture by the self-styled “Holy Inquisition” for maintaining that the Earth turns once each day, rather than the view, clearly expressed in the Bible, that the Earth is stationary and the heavens race around us once each day. The Roman Catholic Church was quite prepared to coerce conformity—to intimidate, torture, and murder to force people to think as it did. At the very beginning of Descartes’s century, the Church had burned the philosopher Giordano Bruno alive because he thought for himself, spoke out, and would not recant. And here, the proposal that animals are clockwork automatons was a far riskier and theologically more sensitive matter than whether the Earth turns—touching not peripheral but central dogmas: free will, the existence of the soul. As on other issues, Descartes walked a fine line.
We “know” we are more than just a set of extremely complex computer programs. Introspection tells us that. That’s the way it feels. And so Descartes, who attempted a thorough, skeptical examination of why he should believe anything, who made famous the proposition Cogito, ergo sum (“I think, therefore I am”), granted immortal souls to humans, and to no one else on Earth.
But we, who live in a more enlightened time, when the penalties for disquieting ideas are less severe, not only may, but have an obligation to, inquire further—as many since Darwin have done. What, if anything, do the other animals think? What might they have to say if properly interrogated? When we examine some of them carefully, do we not find evidence of executive controls weighing alternatives, of branched contingency trees? When we consider the kinship of all life on Earth, is it plausible that humans have immortal souls and all other animals do not?
The moth doesn’t need to know how to fly around the pane of glass, or the goose to retrieve eggs but not beer bottles—again because glass windows and beer bottles have not been around long enough to have been a significant factor in the natural selection of insects and birds. The programs, circuits and behavioral repertoires are simple when no benefit accrues from their being complex. Complex mechanisms evolve when the simple ones will not do.
In Nature, the goose’s egg-retrieval program is adequate. But when the goslings hatch, and especially just before they’re ready to leave the nest, the mother is delicately attuned to the nuances of their sounds, looks, and (perhaps) smells. She has learned about her chicks. Now, she knows her own very well, and would not confuse them with someone else’s goslings, however similar they may seem to a human observer.
In species of birds where mix-ups are likely, where the young may fledge and mistakenly land in a neighboring nest, the machinery for maternal recognition and discrimination is even more elaborate. The goose’s behavior is flexible and complex when rigid and simple behavior is too dangerous, too likely to lead to error; otherwise it is rigid and simple. The programs are parsimonious, no more complex than they need be—if only the world does not produce too much novelty, too many windows and beer bottles.
Consider our prancing insect again. It can see, walk, run, smell, taste, fly, mate, eat, excrete, lay eggs, metamorphose. It has internal programs for accomplishing these functions—contained in a brain of mass, perhaps, only a milligram—and specialized, dedicated organs for carrying the programs out. But is that all? Is there anyone in charge, anyone inside, anyone controlling all these functions? What do we mean by “anyone”? Or is the insect just the sum of its functions, and nothing else, with no executive authority, no director of the organs, no insect soul?
You get down on your hands and knees, look at the insect closely, and you see it cock its head, triangulating you, trying to get a sense of this immense, looming, three-dimensional monster before it. The fly strides unconcernedly; you lift the rolled-up newspaper and it quickly buzzes off. You turn on the light and the cockroach stops dead in its tracks, regarding you keenly Move toward it and it scampers into the woodwork. We “know” such behavior is due to simple neuronal subroutines. Many scientists get nervous if you ask about the consciousness of a housefly or a roach. But sometimes you get an eerie feeling that the partitions separating programs from awareness may be not just thin, but porous.
We know the insect decides who to eat, who to run away from, who to find sexually attractive. On the inside, within its tiny brain, does it have no perception of making choices, no awareness of its own existence? Not a milligram’s worth of self-consciousness? Not a hint of a hope for the future? Not even a little satisfaction at a day’s work well done? If its brain is one millionth the mass of ours, shall we deny it one millionth of our feelings and our consciousness? And if, after carefully weighing such matters, we insist it is still “only” a robot, how sure are we that this judgment does not apply as well to us?
We can recognize the existence of such subroutines precisely because of their unbending simplicity. But if instead we had before us an animal brimming over with complex judgments, branched contingency trees, unpredictable decisions, and a strong executive program, would it seem to us that there is more here than just an elaborate, exquisitely miniaturized computer?
The honeybee scout returns to the hive from a foraging expedition and “dances,” rapidly crawling in a particular, fairly complex pattern over the honeycomb. Pollen or nectar may adhere to her body, and she may regurgitate some of her stomach contents for her eager sisters. All this is done in complete darkness, her motions monitored by the spectators through their sense of touch. Given only this information, a swarm of bees then flies out of the hive in the proper direction to the proper distance to a food supply they’ve never visited as effortlessly as if this was their daily, familiar commute from home to work. They partake of the meal described to them. All this occurs more often when food is scarce or the nectar especially sweet.8 How to encode the location of a field of flowers into the language of dance, and how to decode the choreography is knowledge present in the hereditary information stored inside the insect. Maybe they are “only” robots, but if so these robots have formidable capabilities.
When we characterize such beings as only robots, we are also in danger of losing sight of the possibilities in robotics and artificial intelligence over the next few decades. Already, there are robots that read sheet music and play it on a keyboard, robots that translate pretty well between two very different languages, robots that learn from their own experiences—codifying rules of thumb never taught to them by their programmers. (In chess, for example, they might learn that it is generally better to position bishops near the center than near the periphery of the board, and then teach themselves circumstances in which an exception to this rule is warranted.) Some open-loop chess-playing robots can defeat all but a handful of human chess masters. Their moves surprise their pro
grammers. Their completed games are routinely analyzed by experts who speculate about what the robot’s “strategy,” “goals,” and “intentions” must have been. If you have a large enough pre-programmed behavioral repertoire and if you are able to learn enough from experience, don’t you begin to appear to an outside observer as if you’re a conscious being making voluntary choices—whatever may or may not be going on inside your head (or wherever you keep your neurons)?9
And when you have a massive collection of mutually integrated programs, capability for learned behavior, data-processing prowess, and means of ranking competing programs, might it not start feeling, on the inside, a little bit like thinking? Might our penchant for imagining someone inside pulling the strings of the animal marionette be a peculiarly human way of viewing the world?* Could our sense of executive control over ourselves, of pulling our own strings, be likewise illusory—at least most of the time, for most of what we do? How much are we really in charge of ourselves? And how much of our actual everyday behavior is on automatic pilot?
Among the many human feelings that, although culturally mediated, may be fundamentally preprogrammed, we might list sexual attraction, falling in love, jealousy, hunger and thirst, horror at the sight of blood, fear of snakes and heights and “monsters,” shyness and suspicion of strangers, obedience to those in authority, hero worship, dominance of the meek, pain and weeping, laughter, the incest taboo, the infant’s smiling delight at seeing members of its family, separation anxiety, and maternal love. There is a complex of emotions attached to each, and thinking has very little to do with any of them. Surely, we can imagine a being whose internal life is nearly wholly composed of such feelings, and nearly devoid of thought.
——
The spider builds her web near our porch light. The fine, tough thread reels out from her spinneret. We first notice the web glistening with tiny droplets after a rainstorm, the proprietor repairing a damaged circumferential strut. The elegant, concentric, polygonal pattern is carefully stabilized with a single guy thread extending to the cowl of the lamp itself, and another to a nearby railing. She repairs the web even in darkness and foul weather. At night, when the light is on, she sits at the very center of her construction, awaiting the hapless insect who is attracted by the light and whose eyesight is so poor that the web is quite invisible. The moment one becomes entangled, news of this event travels to her in waves along the threads. She rushes down a radial strut, stings it, quickly wraps it in a white cocoon, packaging it for future use, and rushes back to her command center—composed, a marvel of efficiency, not even, as far as we can see, a little out of breath.
How does she know to design, construct, stabilize, repair, and utilize this elegant web? How does she know to build it near the lamp, to which the insects are attracted? Did she scamper all over the house tallying the abundance of insects in various potential campsites? How could her behavior be pre-programmed, since artificial lights have been invented much too recently to be taken account of in the evolution of spiders?
When spiders are given LSD or other consciousness-altering drugs, their webs become less symmetrical, more erratic, or, we might say, less obsessive, more freeform—but also less effective in catching insects. What has a tripping spider forgotten?
Maybe its behavior is entirely pre-programmed in its ACGT code. But then, couldn’t much more complex information be locked away in a much longer, much more elaborate code? Or maybe some of this information is learned from past adventures in spinning and repairing webs, immobilizing and eating prey. But then look how small that spider’s brain is. How much more sophisticated behavior might emerge out of the experience of a much larger brain?
The web is anchored opportunistically to a local geometry of lamp cowling, metal railing, and wood siding. That could not per se have been pre-programmed. There must have been some element of choice, of decision making, of connecting a hereditary predisposition to an environmental circumstance never before encountered.
Is she “only” an automaton, unquestioningly performing actions that seem to her the most natural thing in the world—and being rewarded, her behavior reinforced by an ample supply of food? Or might there be a component of learning, decision making, and self-consciousness?
Adopting high standards of engineering precision, she spins her web now. She reaps the reward later, maybe much later. She patiently waits. Does she know what she’s waiting for? Does she dream of succulent moths and foolish mayflies? Or does she wait with her mind a blank, idling, thinking of nothing at all—until the telltale tug sends her scurrying down one of the radial struts to sting the struggling insect before it frees itself and escapes? Are we really sure she doesn’t have even a faint and intermittent spark of consciousness?
We would guess that some rudimentary awareness flickers in the most humble creatures, and that with increasing neuronal architecture and brain complexity, consciousness grows. “When a dog runs,” said the naturalist Jakob von Uexküll, “the dog moves his legs; when a sea urchin runs, the legs move the sea urchin.”10 But even in humans, thinking is often a subsidiary state of consciousness.
If it were possible to peer into the psyche of a spider or a goose, we might detect a kaleidoscopic progression of inclinations—and maybe some premonitions of conscious choice, actions selected from a menu of possible alternatives. What individual nonhuman organisms may perceive as their motivations, what they feel is happening inside their bodies, is for us one of the nearly inaudible counterpoints to the music of life.
When an animal goes out to seek food, it often does so according to a definite pattern. A random search is inefficient, because the path would turn back on itself many times; the same places would then be examined again and again. Instead, while the animal may dart off to left and right, the general search pattern is almost always progressive forward motion. The animal finds itself on new ground. The search for food becomes an exercise in exploration. A passion for discovery is hardwired. It’s something we like to do for its own sake, but it brings rewards, aids survival, and increases the number of offspring.
Perhaps animals are almost pure automatons—with urges, instincts, hormonal rushes, driving them toward behavior which in turn is carefully honed and selected to aid the propagation of a particular genetic sequence. Perhaps states of consciousness, no matter how vivid, are as Huxley suggested, “immediately caused by molecular changes in the brain substance.” But from the point of view of the animal, it must seem—as it does with us—natural, passionate, and occasionally even thought out. Perhaps a flurry of impulses and intersecting subroutines at times feels something like the exercise of free will. Certainly the animal cannot much have an impression of being impelled against its will. It voluntarily chooses to behave in the manner dictated by its contending programs. Mainly, it’s just following orders.
So when the days become long enough, it feels an unfocused restlessness, something like spring fever. It hasn’t thought through conception, gestation, the optimum season for the birth of the young and the continuance of its genetic sequences; all that is far beyond its abilities. But from the inside it may well feel as though the weather is intoxicating, life is tempestuous, and moonlight becomes you.
——
We do not mean to be patronizing. The depth of understanding exhibited by our fellow creatures is of course limited. So is ours. We also are at the mercy of our feelings. We too are profoundly ignorant about what motivates us. Some of those beings have, as familiar aspects of their everyday lives, sensibilities wholly absent in humans. Other beings have different tastes and appreciations of the outside world—“To a worm in horseradish, the horseradish seems sweet,” as an old Yiddish folk adage has it. Beyond that, the horseradish worm lives in a world of smells, tastes, textures, and other sensations unknown to us.
Bumblebees detect the polarization of sunlight, invisible to uninstrumented humans; pit vipers sense infrared radiation and detect temperature differences of 0.01°C at a distance of half a meter;
many insects can see ultraviolet light; some African freshwater fish generate a static electric field around themselves and sense intruders by slight perturbations induced in the field; dogs, sharks, and cicadas detect sounds wholly inaudible to humans; ordinary scorpions have micro-seismometers on their legs so they can detect in pitch darkness the footsteps of a small insect a meter away; water scorpions sense their depth by measuring the hydrostatic pressure; a nubile female silkworm moth releases ten billionths of a gram of sex attractant per second, and draws to her every male for miles around; dolphins, whales, and bats use a kind of sonar for precision echo-location.
The direction, range, amplitude, and frequency of sounds reflected back to echo-locating bats are systematically mapped onto adjacent areas of the bat brain. How does the bat perceive its echo-world? Carp and catfish have taste buds distributed over most of their bodies, as well as in their mouths; the nerves from all these sensors converge on massive sensory processing lobes in their brains, lobes unknown in other animals. How does a catfish view the world? What does it feel like to be inside its brain? There are reported cases in which a dog wags its tail and greets with joy a man it has never met before; he turns out to be the long-lost identical twin of the dog’s “master,” recognizable by his odor. What is the smell-world of a dog like? Magnetotactic bacteria contain within them tiny crystals of magnetite—an iron mineral known to early sailing ship navigators as lodestone. The bacteria literally have internal compasses that align them along the Earth’s magnetic field. The great churning dynamo of molten iron in the Earth’s core—as far as we know, entirely unknown to uninstrumented humans—is a guiding reality for these microscopic beings. How does the Earth’s magnetism feel to them? All these creatures may be automatons, or nearly so, but what astounding special powers they have, never granted to humans, or even to comic book superheroes. How different their view of the world must be, perceiving so much that we miss.