Bumblebees Can Learn

Check out this cool new study published in SCIENCE Magazine. The study proves that bees can learn, and they can adapt what they’ve learned to new situations.

The researchers created some really clever tasks for the bees, and the descriptions of what the bees had to do are pretty complicated. First, the researchers showed the bees a small yellow ball at the center of a blue circle. The ball had a sugar solution inside, and the bees learned to go up to the ball, and get the sugar, pretty quickly (within 48 hours).

Next, they put the yellow ball outside the blue circle, and the bees could only get to the sugar after they pushed the ball into the center of the circle. The researchers started by showing the bee how to do it–they made a stick with a plastic bee at the end, and the manipulated the stick so that the plastic bee moved the ball into the center. At first, the bee could eat the sugar once the plastic bee had finished, but after a few times of this, the bee had to move it himself to get the sugar.

14 bees figured out how to move it themselves within 10 tries. The researchers got rid of the four dumber bees who couldn’t figure it out. Then, the researchers gave the bees a much larger blue circle, and all of the bees still could move the yellow ball to the center, ten tries in a row. The bees kept learning; on each of the ten trials they took less time to finish, and their path to the center become more direct.

Next the researchers put the bees through a complicated task that’s hard to describe briefly. In short, they showed that the bees learned best when they could watch another bee doing it, compared to another learning situation where they didn’t have another bee to watch. That’s social learning–learning from watching and imitating someone you recognize.

Then, with yet another complicated experiment, they showed that the bees aren’t just copying what they see another bee do, but that they learn to adapt what they’ve learned to new situations. For example, if they were shown another bee moving the farthest away ball, they knew to move the closest ball instead of that farther one. And second, if the ball color changed to black, they could still do it.

The researchers point out that these artificial tasks are way harder than anything bees have to do in the wild. Evolution didn’t require this adaptation (of being able to learn this way). This means that animals can end up smarter than they need to survive in the wild. (We’re talking about you, human beings!)

Here’s their conclusion:

Such unprecedented cognitive flexibility hints that entirely novel behaviors could emerge relatively swiftly in species whose lifestyle demands advanced learning abilities.

Are You Too Old to Be Brilliant?

When are brilliant scientists the most brilliant? What age are you likely to be when the Nobel committee comes calling? Pick one of the following answers:

  • You need a lot of expertise and wisdom to make a big breakthrough. You need professional connections, lots of research money, and big laboratories. Scientific breakthroughs come from people in middle age, or maybe even at the end of their careers.
  • It’s the young upstarts who have lots of energy and fresh ideas. After all, the old scientists are stuck in ideas from the past. They’re already past their prime. They’re tired and don’t have much energy any more. Am I talking about myself at the ripe old age of 56? I didn’t get much sleep last night, and my knees are kind of sore 🙂

A new study gives us the answer: None of the above. There’s no relationship between age and creative scientific contribution. The authors of the study analyzed 2,856 physicists, working from 1893 to the present. They found that the best predictor of exceptional creativity is productivity. It’s lots of hard work. The scientists who do the most experiments, and test the most hypotheses, are the ones with the big contributions. The researchers found that once they’d controlled for productivity, age doesn’t add any additional predictive power.

The researchers identified a second variable that’s related to scientific impact: They called it Q, and it includes intelligence, motivation, openness to ideas, ability to write well. Another surprise: The variable Q doesn’t change over your career. (Otherwise, you’d be back to the theory that age predicts creativity.)

It’s still true that younger scientists are more likely to make a significant contribution. But it’s not because a person has more brilliant insights in your 20s, and it’s not because their ideas are fresh and unbound by old-fashioned tradition. It’s because they work harder and that’s why they’re more productive. So if you’re older, there’s still hope.

Now if only I could get a good night’s sleep.

The Emergence of Creativity: Matt Ridley’s New Book

You’ve got to read the excerpt from Matt Ridley’s new book in today’s Wall Street Journal. Just released this week, his book is called The Evolution of Everything: How New Ideas Emerge. I have a lot of respect for his previous books, so I’m delighted to learn that his new book makes the same points as my 2007 book Group Genius.

Here are the key features of innovation, described in both of our books:

  • The stories we hear about genius inventors, like Thomas Edison inventing the light bulb, are always myths. Ridley and I both describe the real history of the light bulb, which involves lots of people way before Edison. (Group Genius, pages 110, 196)
  • “Innovation emerges from the bottom up,” I write in Group Genius  (page 16). I show that innovation emerges from self-organizing systems, and this is Ridley’s main point, too.
  • Ridley writes that innovation is “incremental” rather than “revolutionary.” That’s why I called one of my chapters “Small Sparks”: to emphasize that innovation doesn’t come from a big flash of insight. “Successful creators know how to keep their sparks coming in a process that unfolds over time” (Group Genius, page 97).
  • Ridley describes the historical research on multiple discovery, as I do on pages 192-193, with this example: “In the 1920s, numerous teams invented television in parallel.”
  • Ridley argues that patent protection is too broad and is based on the mythical view of the lone inventor. I make the same point on pages 176-224, especially pages 221-225: “Current policy favors linear, centralized innovation and blocks the natural rhythm of innovation”.
  • Ridley demolishes the idea that innovation comes from a linear process; this is the most important point of Group Genius  (for example, pages 158-159, “Beyond Linear Innovation”)

Ridley’s WSJ  excerpt is filled with great stories of real innovations. I come to the same conclusions, with some of the same historical examples, and also by drawing on the science of creativity. Inspired by my studies of jazz and improv theater, I think of creativity as improvisation. Group Genius argues that the most creative improvisations are non-linear, emergent, unpredictable, and inefficient. Ridley has a bit more to say about the political and economic implications of this new, more realistic, understanding of innovation (for example, he concludes that government doesn’t need to fund scientific research). I have a bit more to say about how you can use this research to become more successfully creative, both on your own and in teams. It’s cool that Ridley and I come to the same conclusions from really different directions. If you like Group Genius, you really should check out The Evolution of Everything. (I’ll post a review after I’ve read the whole book.)



The Lone Genius Loses to the Team

What’s your visual image of a brilliant scientist? A nerdy man in a lab coat, working late in some basement laboratory with beakers and test tubes? Someone typing at a computer in their office? Well, clear your mind of that image, because science today is all about collaboration and teamwork. This is the message of a truly impressive study published in SCIENCE magazine 18 May 2007. Three professors at Northwestern University, Stefan Wuchty, Benjamin F. Jones, and Brian Uzzi, analyzed huge databases–of 19.9 million scientific papers over 50 years, and 2.1 million patents–and found that collaboration is rapidly becoming the norm in science and in invention.

They focused on a few key numbers. First, the databases allowed them to determine which papers, and which patents, had one author, two authors, or more. Two or more authors means that the creation was collaboratively generated. In science, the average team size (number of co-authors) doubled over 45 years–from 1.9 to 3.5 authors per paper. Of course, science has become a lot more complex, and requires a lot more funding, and that might account for the larger team size. But the databases also had data about the social sciences and the arts and humanities; social science research hasn’t increased in scale and cost the same way particle physics and medicine have. And surprisingly, even in the social sciences, collaboration has become a lot more important. In 1955, only 17.5% of social science papers had two or more authors; in 2000, 51.5% of those papers did. And although papers in the arts and humanities still are mostly sole authored (over 90%), the trend over the last 50 years has also been toward more collaboration.

But what about quality and creativity? Can we find out if the collaboratively generated papers are any better? Fortunately, the databases allowed the researchers to determine the impact and influence of each paper, and of each patent, because those databases keep track of how many times the paper or patent was cited by a later publication. More citations means a more influential paper; and more citations have been shown to correlate with research quality. And guess what: over the 50 year period studied, teams generated more highly cited work in every research area, and in every time period. The implication is that teams generate better scientific research than solitary individuals.

One final interesting finding is that the creative advantage for teams has increased over the last 50 years. Although teams generated more highly cited work back in 1955, by 2000 the advantage of teams over sole individuals had become even greater. In 1955, team-authored papers received 1.7 times as many citations as sole authored papers; in 2000, they received 2.1 times as many.

In a later issue of SCIENCE magazine (14 September 2007) several letters challenging this research were published; the authors convincingly responded, by providing additional data. There’s no question that teams do better science than solitary individuals, and that the trend is working in teams’ favor.