Where rats and robots play

THINK of a university and what comes to mind may be the cloistered calm of Oxford, the architectural chaos of MIT or even, perhaps, the 1950s brutalism of Moscow. A Daliesque building on a subtropical island, with a view of the ocean and signs on campus warning of venomous snakes, is more unusual. But that is appropriate, for the Okinawa Institute of Science and Technology (OIST), inaugurated as a graduate university on November 19th, is intended to be unusual. It was built from scratch on a forested hilltop overlooking the East China Sea, and its approach to science starts from scratch too. It has no departments. Instead, its biologists, chemists, physicists, mathematicians and computer geeks intermingle, sharing laboratory equipment, teachers and money. After two centuries of science becoming more and more specialised, the idea is to bring back the generalist.

Surprisingly, this experiment is taking place in Japan—a country with one of the most rigid academic hierarchies in the world. Locating it in Okinawa, though, is a masterstroke. The island, which is closer to Shanghai and Taipei than to Tokyo, is as cut off as is possible to imagine from the mainstream of Japanese academia. The result is a skunk works: a place where novel and possibly controversial ideas can be tested without constant interference from institutional vested interests. If an idea from a skunk works fails, it can be buried quietly, and nobody gets hurt. If it succeeds, it can be launched fully formed, and bureaucratic resistance thus overcome.

The OIST certainly has the money to do this. The government has spent ¥77 billion (almost $1 billion, at current exchange rates) over the past six years to create it. It has the personnel: 212 researchers, and five Nobel laureates on its board of governors. Soon it will have the students, too. The first intake, next autumn, will only be 20. But that will rise to 100 in five years' time.

According to its president, Jonathan Dorfan (a physicist lured from Stanford University), the OIST seeks to address three shortcomings in Japanese science. First, it wants to nurture independence of thought in young researchers, encouraging them to work for themselves rather than as foot soldiers for professors. Second, it wants Japanese science to become more open to the outside world. Third, it wants to stimulate the emergence of technology clusters in a country where there is disturbingly little interaction between universities and industry—and few Silicon Valley-style start-ups as a consequence.

That graduate students and post-docs are fodder for the ambitions of departmental heads is not unique to Japan. But the phenomenon is generally thought to be exaggerated there: a toxic consequence of Confucian respect for authority and an academic system borrowed lock, stock and barrel from 19th-century Germany. Young Japanese researchers, though, are fed up with it. One result is that the number of graduate students in natural sciences has been falling in Japan since 2003.

Those youngsters are, nevertheless, oddly reluctant to deal with their disgruntlement by going abroad. Between 1996 and 2007, 28% of the science and engineering doctorates awarded in America went to Chinese; 11% to Indians; 9% to South Koreans; and 7% to Taiwanese. Japanese, by contrast, picked up just 2% of them. That stymies the exchange of ideas on which good science depends.

If young Japanese scientists cannot be persuaded to study abroad, though, perhaps abroad can be brought to Japan. At the moment, 85 of the OIST's researchers are gaijin. Ultimately, the organisation aims to recruit half of both its faculty and its students from outside the country.

And, as if that were not enough to stir the pot, these scientists have to mix with each other, too. The OIST's laboratories (designed by Kenneth Kornberg, the son of a Nobel prize-winning scientist and brother of another) abut one another and share microscopes, refrigerators and other equipment. Dr Dorfan says everything is designed to eliminate the barriers—physical and otherwise—between traditional fields of research.

That breakdown of barriers is even apparent in individual researchers. Kenji Doya, for example, is an engineer turned neuroscientist. He runs two laboratories at the OIST. One explores learning by robots. The other looks at how a brain chemical called serotonin regulates the patience of rats. An algorithm for regulating patience in the brain, he believes, could be used to improve robotics.

Which is just the sort of thing that, if brought to fruition in Dr Dorfan's alma mater, would have the researchers who had come up with it bolting out of the door and into the offices of the nearest venture capitalist, prior to hiring a suitable garage to build their prototype in.

Whether that will happen in Okinawa, remains to be seen. Its physical climate is certainly as pleasant as California's. Changing the intellectual climate to match may prove a challenge.