Why did the show fail? The problem, at root, is with the premise of the show — the essential joke that Big Bird doesn't want to be known as a big bird. That's the kind of wordplay that a preschooler simply doesn't understand. Preschoolers make a number of assumptions about words and their meaning as they acquire language, one of the most important of which is what the psychologist Ellen Markman calls the principle of mutual exclusivity. Simply put, this means that small children have difficulty believing that any one object can have two different names. The natural assumption of children, Markman argues, is that if an object or person is given a second label, then that label must refer to some secondary property or attribute of that object. You can see how useful this assumption is to a child faced with the extraordinary task of assigning a word to everything in the world. A child who learns the word elephant knows, with absolute certainty, that it is something different from a dog. Each new word makes the child's knowledge or the world more precise. Without mutual exclusivity, by contrast, if a child thought that elephant could simply be another label for dog, then each new word would make the world seem more complicated. Mutual exclusivity also helps the child think clearly. "Suppose," Markman writes, "a child who already knows 'apple' and 'red' hears someone refer to an apple as 'round.' By mutual exclusivity, the child can eliminate the object (apple) and its color (red) as the meaning of 'round' and can try to analyze the object for some other property to label."

What this means, though, is that children are going to have trouble with objects that have two names, or objects that change names. A child has difficulty with, say, the idea that an oak is both an oak and a tree; he or she may well assume that in that case "tree" is a word for collection of oaks.

The idea, then, that Big Bird no longer wants to be called Big Bird but instead wants to be called Roy is almost guaranteed to befuddle a preschooler. How can someone with one name decide to have another name? Big Bird is saying that Big Bird is merely a descriptive name of the type of animal he is, and that he wants a particular name. He doesn't want to be a tree. He wants to be an oak. But three- and four-year-olds don't understand that a tree can also be an oak. To the extent that they understand what is going on at all, they probably think that Big Bird is trying to change into something else — into some other kind of animal, or some other collection of animals. And how could he do that?

There's a deeper problem. Sesame Street is a magazine show. A typical show consists of at least forty distinct segments, none more than about three minutes — street scenes with the actors and Muppets, animation, and short films from outside the studio. With shows like "Roy," in the late 1990s, the writers of the show attempted, for the first time, to link some of these pieces together with a common theme. For most of the show's history, though, the segments were entirely autonomous; in fact new Sesame shows were constructed, for the most part, by mixing together fresh street scenes with animated bits and filmed sequences from the show's archives.

The show's creators had a reason for wanting to construct Sesame Street this way. They thought preschoolers did not have the attention span to handle anything other than very short, tightly focused segments. "We looked at the viewing patterns of young children, and we found that they were watching Laugh-In," says Lloyd Morrisett, who was one of the show's founders. "That had a very strong effect on the early Sesame Street. Zany, relatively quick one-liners. The kids seemed to love it." SesameStreet's creators were impressed even more by the power of television commercials. The sixties were the golden age of Madison Avenue, and at the time it seemed to make perfect sense that if a 60-second television spot could sell breakfast cereal to a four-year-old, then it could also sell that child the alphabet. Part of the appeal of Jim Henson and the Muppets to the show's creators, in fact, was that in the 1960s Henson had been running a highly successful advertising shop. Many of the most famous Muppets were created for ad campaigns: Big Bird is really a variation of a seven-foot dragon created by Henson for La Choy commercials; Cookie Monster was a pitchman for Frito-Lay; Grover was used in promotional films for IBM. (Henson's Muppet commercials from the 50s and 60s are hysterically funny but have a dark and edgy quality that understandably was absent from his SesameStreet work.)

"I think the most significant format feature in a commercial is that it's about one thing," said Sam Gibbon, one of the earliest Sesame Street producers. "It's about selling one idea. The notion of breaking down the production of Sesame Street into units small enough so they could address a single educational goal like an individual letter owed a lot to that technique of commercials."

But is the commercial theory of learning true? Daniel Anderson says that new research suggests that children actually don't like commercials as much as we thought they did because commercials "don't tell stories, and stories have a particular salience and importance to young people." The original Sesame Street was anti-narrative: it was, by design, an unconnected collection of sketches. "It wasn't just the ads that influenced the early SesameStreet," Anderson says. "There was also a theoretical perspective at the time, based in part on [the influential child psychologist] Piaget, that a preschool child couldn't follow an extended narrative." Since the late 1960s, however, this idea has been turned on its head. At three and four and five, children may not be able to follow complicated plots and subplots. But the narrative form, psychologists now believe, is absolutely central to them. "It's the only way they have of organizing the world, of organizing experience," Jerome Bruner, a psychologist at New York University, says. "They are not able to bring theories that organize things in terms of cause and effect and relationships, so they turn things into stories, and when they try to make sense of their life they use the storied version of their experience as the basis for further reflection. If they don't catch something in a narrative structure, it doesn't get remembered very well, and it doesn't seem to be accessible for further kinds of mulling over."

Bruner was involved, in the early 1980s, in a fascinating project — called "Narratives from the Crib" — that was critical in changing the views of many child experts. The project centered on a two-year-old girl from New Haven called Emily, whose parents — both university professors — began to notice that before their daughter went to sleep at night she talked to herself. Curious, they put a small microcassette recorder in her crib and, several nights a week, for the next fifteen months, recorded both the conversations they had with Emily as they put her to bed and the conversations she had with herself before she fell asleep. The transcripts — 122 in all — were then analyzed by a group of linguists and psychologists led by Katherine Nelson of Harvard University. What they found was that Emily's conversations with herself were more advanced than her conversations with her parents. In fact, they were significantly more advanced. One member of the team that met to discuss the Emily tapes, Carol Fleisher Feldman, later wrote:

In general, her speech to herself is so much richer and more complex [than her speech to adults] that it has made all of us, as students of language development, begin to wonder whether the picture of language acquisition offered in the literature to date does not underrepresent the actual patterns of the linguistic knowledge of the young child. For once the lights are out and her parents leave the room, Emily reveals a stunning mastery of language forms we would never have suspected from her [everyday] speech.