Breaking the Spell

by Daniel C. Dennett

  Human memory is biased in favor of vital combinations, but so, presumably, is the memory found in the brains of all other animals. Animal memory has probably been relatively impervious to fantasy, however, for a simple reason: lacking language, animal brains have not had a way of inundating themselves with an explosion of combinations not found in the natural environment. How is an anxious ape going to concoct the counterintuitive combination of a walking tree or an invisible banana—ideas that might indeed captivate an ape mind if only they could be presented to it?

  Do we know that something like this fantasy-generation process has been taking place in our species (and our species alone) for thousands of years? No, but it is a serious possibility to investigate further. Using only materials that would have been put in place by evolution for other purposes, this hypothesis could explain the remarkably fertile imagination that must somehow be responsible for the world’s menagerie of mythical creatures and demons. Since the monsters themselves have never existed, they had to be “invented,” either deliberately or inadvertently (the way languages were invented). They are expensive creations, and the R & D required for the task had to be generated by something that could pay for itself. I’ve left the hypothesis quite unspecific for the moment, but more constrained forms of it are available, and they have the great advantage of having testable consequences. We can start scouring the world’s mythology for patterns that would be predicted by some versions of the hypothesis but not others.

  And we don’t have to restrict ourselves to the human species. Experiments along the lines of Skinner’s provocation of superstition in the pigeon might begin to uncover the biases and fault lines in ape memory mechanisms, in much the way Niko Tinbergen’s experiments with gulls (1948, 1959) famously showed their perceptual biases. The adult female gull has an orange spot on her beak, at which her chicks instinctually peck, to stimulate the female to regurgitate and feed them. Tinbergen showed that chicks would peck even more readily at exaggerated cardboard models of the orange spot, so-called supernormal stimuli. Pascal Boyer (2001) notes that, over the eons, human beings have discovered and exploited their own supernormal stimuli:

  There is no human society without some musical tradition. Although the traditions are very different, some principles can be found everywhere. For instance, musical sounds are always closer to pure sound than to noise…. To exaggerate a little, what you get from musical sounds are super-vowels (the pure frequencies as opposed to the mixed ones that define ordinary vowels) and pure consonants (produced by rhythmic instruments and the attack of most instruments). These properties make music an intensified form of sound-experience from which the cortex receives purified and therefore intense doses of what usually activates it…. This phenomenon is not unique to music. Humanseverywhere also fill their environments with artifacts that over-stimulate their visual cortex, for instance by providing pure saturated color instead of the dull browns and greens of their familiar environments…. In the same way, our visual system is sensitive to symmetries in objects. Bilateral symmetry in particular is quite important; when two sides of an animal or person look the same it means that they are facing you, a relevant feature of interaction with people but also with prey and predators. Again, you cannot find a human group where people do not produce visual gadgets with such symmetrical arrangements, from the simplest makeup or hairdressing techniques to textile patterns and interior decoration. [pp. 132–33]

  Why don’t other species have art? Once again, the answer that suggests itself—which does not mean that it is proven but only that it may well be provable—is that, lacking language, they lack the tools for creating surrogate stimulus combinations and hence they lack the perspective that permits exploration of the combinatorics of their own senses.1 Using acute observation and trial and error, Tinbergen cleverly devised the supernormal stimuli that enticed his birds (and other animals) into a host of bizarre behaviors. No doubt animals do on occasion trap themselves by inadvertently discovering a supernormal stimulus in nature and letting it do its thing on them, but what would they do next? Do it again if it felt good, but the generation of diversity on which true design exploration depends would probably not be possible for them.

  To sum up the story so far: The memorable nymphs and fairies and goblins and demons that crowd the mythologies of every people are the imaginative offspring of a hyperactive habit of finding agency wherever anything puzzles or frightens us. This mindlessly generates a vast overpopulation of agent-ideas, most of which are too stupid to hold our attention for an instant; only a well-designed few make it through the rehearsal tournament, mutating and improving as they go. The ones that get shared and remembered are the souped-up winners of billions of competitions for rehearsal time in the brains of our ancestors. This is not a new idea, of course, just a clarification and extension of an idea that has been around for generations. As Darwin himself surmised:

  …the belief in unseen or spiritual agencies…seems to be almost universal…nor is it difficult to comprehend how it arose. As soon as the important faculties of the imagination, wonder, and curiosity, together with some power of reasoning, had become partially developed, man would naturally have craved to understand what was passing around him, and have vaguely speculated on his own existence. [1886, p. 65]

  So far so good, but what we have accounted for is superstition, not religion. Hunting for elves in the garden or the bogeyman under your bed is not (yet) having a religion.

  What is missing? For one thing, belief! For, although Darwin speaks of belief in spiritual entities, we have not yet provided an account that secures anything so strong as that. Nothing has yet been said about having to believe the hobbyhorse idea that keeps recycling through your mind; it may be a hunch, or a wonder, or even an obsessively disbelieved little nugget of paranoia—or just a captivating morsel of story line. Nobody has ever believed in Cinderella or Little Red Riding Hood, but their fairy tales have been quite faithfully transmitted (with mutations) over many generations. Many fairy tales make up for not being true stories by having a moral, which gives them an apparent value—to the tellers and hearers—that makes up for their not being information about the wide world. Others conspicuously lack a moral—just what does “Goldilocks” teach our children: not to invade strangers’ houses?—and must persist in the transmission tournament for less obvious reasons. As is usual in evolutionary circumstances, a gradual ramp of intermediate states of mind is there to be traversed, from shuddering doubt are there really wicked witches in the woods?) and neutral fascination (a flying carpet—just imagine!) through nagging uncertainty (unicorns? well, I’ve never seen one) and on to robust conviction (Satan is as real as that horse over there). Fascination is enough to power rehearsal and replication. Almost everybody has a good strong copy of the idea of unicorns, though few people believe in them; but hardly anybody has the idea of pudus, which have the distinct advantage of being real (you can look it up). There is a lot more to religion than a fascination with counterintuitive agentlike entities.

  2 Gods as interested parties

  Why the gods above me

  Who must be in the know

  Think so little of me

  They allow you to go…

  —Cole Porter, “Every Time We Say Goodbye”

  Ancestor worship must be an appealing idea to those who are about to become ancestors.

  —Steven Pinker, How the Mind Works

  Whereas other species make limited use of the intentional stance—for anticipating the moves of predator and prey, plus a little bullying and bluffing—we human beings are obsessed about our personal relations with others: worrying about our reputations, our unfulfilled promises and obligations, and reviewing our affections and loyalties. Unlike other species, which have to worry all the time about lurking predators and dwindling food sources, we human beings have largely traded in these pressing concerns for others.
The price our species has paid for the security of living in large groups of interacting communicators with different agendas is having to keep track of these complex agendas and shifting relationships. Whom can I trust? Who trusts me? Who are my rivals and my friends? To whom do I owe debts, and whose debts to me should I forgive or collect? The human world is teeming with such strategic information, to use Pascal Boyer’s term, and what matters most about it (as in a card game) is this: “In social interaction, we presume that other people’s access to strategic information is neither perfect nor automatic” (2001, p. 155). Does she know that I know that she wants to leave her husband? Does anybody know that I stole that pig? All the plots of all the great sagas and tragedies and novels, but also all the situation comedies and comic books, hinge on the tensions and complexities that arise because agents in the world don’t all share the same strategic information.

  How do people deal with all this complexity?2 Sometimes when people are learning a new card game they are advised by their teacher to lay all their cards faceup on the table, so everybody can see what the others are holding. This is an excellent way of teaching the tactics of the game. It provides a temporary crutch for the imagination—you actually get to see what each person would normally be hiding, so you get to base your reasoning on the facts. You don’t have to keep track of them in your head, since you can just look down on the table whenever you need a reminder. This helps you build up skill in visualizing where the cards must be when they are hidden. What works at the card table can’t be done in real life. We can’t get people to divulge all their secrets during a practice session of life, but we can get practice “off line” by telling and listening to stories, narrated by an agent who sees all the cards of the fictional or historic characters.

  What if there really were agents who had access to all the strategic information! What an idea! It is easy enough to see that such a being—in Boyer’s terms, a “full-access agent”—would be an attention-grabbing concoction, but aside from that, what good would it be? Why would it be any more important to people than any other fantasy? Well, it might help people simplify the thinking that has to be done to figure out what to do next. A survey of the world’s religions shows that almost always the full-access agents turn out to be ancestors, gone but not at all forgotten. As the memory of Father is burnished and elaborated in many retellings to children and grandchildren and their grandchildren, his ghost may acquire many exotic properties, but at the heart of his image is his virtuosity in the strategic-information department. Remember how your mother and father often seemed to know just what you were thinking, just what mischief you were trying to hide? Ancestors are like that, only more so: you can’t hide from them, not even your secret thoughts, and nobody else can either. Now you can reframe your puzzlement about what to do next: what would my ancestors want me to do in my current situation? You may not be able to tell what these vividly imagined agents would want you to do, but, whatever it is, it’s what you should do.

  Why, though, do we human beings so consistently focus our fantasies on our ancestors? Nietzsche, Freud, and many other theorists of culture have articulated elaborate conjectures about the subliminal motivations and memories that arose from mythic struggles deep in our human past, and there may be substantial gold to be refined from this lode of speculation once we re-examine it with an eye to testable hypotheses of evolutionary psychology, but in the meantime, we can more confidently identify the basic mental disposition that sets up this bias, for it is considerably older than our species. Mammals and birds, unlike most other animals, often devote considerable parental attention to their young, but there is wide variation in this: precocial species are those in which the young hit the ground running, as the saying goes, whereas altricial species have young that require prolonged parental care and training. This training period provides a host of opportunities for information transmission from parent to offspring that bypasses the genes entirely.

  Biologists are often accused of gene centrism—thinking that everything in biology is explained by the action of genes. And some biologists do indeed go overboard in their infatuation with genes. They should be reminded that Mother Nature is not a gene centrist! That is, the process of natural selection itself doesn’t require that all valuable information move “through the germ line” (via the genes). On the contrary, if the burden can be reliably taken over by continuities in the external world, that is fine with Mother Nature—it takes a load off the genome. Consider the various continuities relied on by natural selection: those supplied by the fundamental laws of physics (gravity, etc.) and those supplied by the long-term stabilities of environment that can be safely “expected” to persevere (salinity of the ocean, composition of the atmosphere, colors of things that can be used as triggers, etc.). To say that natural selection relies on these regularities means just this: it generates mechanisms that are tuned to work well in environments that exhibit those regularities. The design of these mechanisms presupposes these regularities in the same way that the design of a Mars rover presupposes the planet’s gravity, the solidity and temperature range of its surface, and so forth. (It is not designed to operate in the Everglades, for instance.) Then there are the regularities that can be transmitted from generation to generation by social learning. These are a special case of reliable environmental regularities; they take on further importance since they are themselves subject to natural selection, directly and indirectly. Two information superhighways have been improved and enlarged over the eons. The genetic informational pathways have themselves been subject to incessant refinement over billions of years, with optimization of chromosome design and invention and improvement of proofreading enzymes and so forth, with the effect that high-fidelity, high-bandwidth transmission of genetic information has been achieved. The parent-child instructional pathway has also been optimized by a recursive or iterative process of enhancement. As Avital and Jablonka (2000) note, “The evolution of the transmission of mechanisms of transmission is of central importance for the evolution of learning and behaviour”(p. 132).

  Among the adaptations for improving the bandwidth and fidelity of parent-offspring transmission is imprinting, in which the newborn has a readily triggered and powerful instinctual urge to approach and stay close to and attend to the first large moving thing it sees. In mammals the urge to find and cling to the nipple is hardwired by the genes, and it has the side effect, opportunistically exploited by further adaptations, of keeping the young where they can watch Mother when they are not feeding. Human infants are no exception to the mammalian rules. Meanwhile, going in the other direction, parents have been genetically designed to attend to infants. Whereas gull chicks are irresistibly drawn to an orange spot, human beings are irresistibly captivated by the special proportions of a “baby face.” It brings out the “Aw, isn’t she cunning!” response in the steeliest curmudgeons. As Konrad Lorenz (1950) and others have argued, the correlation between an infant’s facial appearance and an adult’s nurturing response is no accident. It is not that baby faces are somehow intrinsically darling (what on earth could that mean?) but that evolution hit upon facial proportions as the signal to trigger parental responses, and this has been refined and intensified over the eons in many lineages. We don’t love babies and puppies because they’re cute. It’s the other way around: we see them as cute because evolution has designed us to love things that look like that. So strong is the correlation that measurements of fossils of newborn dinosaurs have been used to support the radical hypothesis that some dinosaur species were altricial (Hopson, 1977; Horner, 1984). Stephen Jay Gould’s classic analysis (1980) of the gradual juvenilization over the years of Mickey Mouse’s features provides an elegant demonstration of the way cultural evolution can parallel genetic evolution, homing in on what human beings instinctually prefer.

  But even more potent than the bias in adults to resp
ond parentally to baby-faced young is the bias in those young to respond with obedience to parental injunction—a trait observable in bear cubs as well as human babies. The free-floating rationale is not far to seek: it is in the genetic interests of parents (but not necessarily other conspecifics!) to inform—not misinform—their young, so it is efficient (and relatively safe) to trust one’s parents. (Sterelny, 2003, has particularly acute observations on the trade-offs between trust and suspicion in the evolutionary arms races of cognition.) Once the information superhighway between parent and child is established by genetic evolution, it is ready to be used—or abused—by any agents with agendas of their own, or by any memes that happen to have features that benefit from the biases built into the highway.3