The dolphin rolled over to present his belly. Sagan rolled up his shirtsleeves and began scratching Elvar like a dog. The dolphin demanded more, then more again, until Sagan tired of the game. Elvar responded by leaping so high that “only high tail flukes were in contact with the water.” As he did so, he emitted a single shrill syllable. It sounded to Sagan as if he had shrieked, “More!”
Sagan “bounded out of the room” and excitedly announced the news to Lilly, who seemed entirely unsurprised. “Good, that’s one of the words he knows,” he replied.
Sagan was delighted by the trip. Afterward, however, he began to wonder about the quality of the research. Sagan and his close friend the astronomer Frank Drake were becoming experts in identifying signals amid seas of noise picked up by radio telescopes. They agreed that the institute’s work set off alarm bells. Lilly appeared to be cherry-picking his data, carefully selecting the snippets of “dolphinese” that resembled human language and excluding the rest.
The most unsettling feature of the institute, however, was not Lilly’s quixotic obsession with dolphin speech. It was the fact that his dolphins kept dying.
Among those who questioned Lilly’s actions at the lab was Ted Nelson, a young Harvard graduate who had first met Lilly on a train in New Jersey. Nelson was intrigued by the idea of studying animal communication with computers. And he adored Gregory Bateson, with whom he shared an office. But Lilly struck Nelson as needlessly callous. True, the institute’s director was “brilliant and persuasive, and a fantastic self-publicist,” Nelson admitted. Yet he was, at heart, “ruthless ... a con man.”
For a time, Nelson worked on a promotional video about the lab’s work. It was meant to be a puff piece, but Nelson found the job unexpectedly difficult. The trouble was, he wrote, that the tropical splendor of the lab’s setting kept swerving unexpectedly toward a more unsettling ambience:
Through the windows, it was just before twilight and beautiful: The palm trees were blowing, pelicans scudded by. Atmosphere. Then I panned the camera across the room, showing concern on everybody’s faces as they watched oscilloscope traces leaping. We were actually hearing what the scope showed: the shrieking and whistling of our dolphin pals in the tank on the floor below, coming out of different speakers as the sound was transposed from different audio ranges.
The truth was that Lilly’s research with animals relied on what scientists at the time called “operant conditioning” — a polite term for using pain to control animal behavior.