Scientists revisit a monkey study gone wrong

• Watchdogs concluded ex-Harvard scientist Hauser committed research misconduct
• Others are checking his scandal-tainted research findings about a language instinct
• 'I needed to know if I believed (his) results, or not,' says researcher Julie Neiworth

When science goes wrong, who picks up the pieces?

Other scientists. And in at least one prominent case, their monkeys.

Consider the case of ex-Harvard scientist Marc Hauser, author of Moral Minds: How Nature Designed a Universal Sense of Right and Wrong, who had headed heralded research into behavior in monkeys, and its implication for how people evolved their abilities to think, feel and speak. On Sept. 5, the federal Office of Research Integrity concluded a long-running investigation, finding that Hauser had committed research misconduct in his work, faking data in a 2002 study on cotton-top tamarin monkeys, and fabricating results in later study descriptions. His university came to similar conclusions last year.

When the scandal erupted into public view in 2010, psychologists who had cited Hauser's work faced a quandary.

"I stopped teaching his articles, but there was a more fundamental problem," says cognitive psychologist Julie Neiworth of Carleton College in Northfield, Minn. "I needed to know if I believed (Hauser's) results, or not. And my lab was about the only one in the country that could find out if they were right or wrong."

Not many labs work with cotton-top tamarins. In a retracted 2002 study, Hauser had claimed these monkeys reacted to shifts in patterns of syllables in ways that pointed to the roots of language in humans. Playing syllable noises, or phonemes, to the monkeys in regular patterns (say a-a-b), the study claimed the monkeys reacted when the pattern shifted to other patterns (say a-b-b). You could substitute different noises into the pattern without reaction, but changing the pattern made the monkeys look up.

"One of the ways that humans are seen as different from other animals is in how we perceive language," Neiworth says. The 2002 study results made it seem possible that a language instinct resided in monkeys native to South America, ones separated from humans and other apes by tens of millions of years of evolution. "I felt compelled to test the idea when these experiments fell into doubt."

A lot of other researchers wanted the result checked as well. On Sept. 20, the National Institutes of Health awarded Neiworth and her lab another round of funding, bringing the total over the last year of work to more than $600,000 so far, to check out Hauser's results. Carleton College has a colony of 13 tamarins, and Neiworth and her students set out to re-do the Hauser lab experiments.

"Everything is videotaped and checked with outside observers. There will be no funny results from my lab," she says. So, what has she found? Was Hauser right or wrong? "He was right, but he was also wrong, and in an important way," Neiworth says.

In early results, the lab did observe the same reaction to pattern breaks in sounds of syllables played to tamarins.

"I actually used the exact same sounds as the 2002 study. Marc gave them to me," Neiworth says. "He was very helpful and wanted to see the research replicated."

But Neiworth added something more to the original study. She also tested patterns of computer beeps, not syllable noises, to see if the monkeys reacted to pattern breaks. And they did, in exactly the same way.

"I don't think this behavior is a marker for human language at all. It likely is something abstract and deeper tied to hearing and recognizing sounds," she says. How does she know that? Because she also tested them with patterns of tamarin call noises, barks, hoots and the like, split into their simplest forms and put into patterns.

"That's their language, those noises." Neiworth says. "We don't know what they all mean, but we do know a lot of them and those are the ones we used."

And how did the tamarins react to tamarin noises? "They reacted to the sounds like they would to language, not to pattern breaks," she says. "If the pattern was, here's food, I-hate-you, I-hate-you, they would get confused. If one sound was a snake warning, then they would run around in circles, which is what they do then."

So tamarins don't regard human language syllables as anything different than any other noises. One of the other areas that got Hauser in trouble was research suggesting that tamarins possessed an understanding of what others thought, a "theory of mind." He suggested that monkeys understood whether testers who had seen a treat being hidden really knew where it was, as against a tester who hadn't seen where it was being hidden. Hauser's research suggested that monkeys more often ran to treat hiding places pointed to by testers whom the monkey had observed were in the know (similar research has been done with dogs and other animals) as against clueless ones.

But in research still underway, that finding is looking like a bust, Neiworth says. "We think the effect is just one where they follow the pointing more often of the trainer with whom they are most familiar."

A lot of work is ahead to test just the most significant of the results from Hauser's lab. One moral of the story is that animal research is difficult even with the best intentions. "You can't test just one thing and conclude you have a really important result no matter what," Neiworth says.