Alun Salt
I’ve been sent a paper today to see if we can put out a press release for it. It’s fascinating and it touches on a famous philosophical problem. Does red really look red? This paper looks at plants from a bee’s eye view. That changes a few ways that you can see plants. They become bigger, relatively speaking, and connections between then differ because you can fly. Another way things change is that red flowers no longer look red. Bees see further into the UV spectrum than humans, but they can’t see the red part of the spectrum. So do they see black? Black is a good idea, and it’s the authors.
I’m not sure how helpful this is though. The paper isn’t musing on the nature of reality, its research into pollination by a specific type of bee. Does this observation explain why the research is novel, or does it distract from the interesting things the team has found. Fortunately this paper has a very accessible problem, so there won’t be a struggle to explain why the research is tackling a real problem.
We’ll also have some genomics papers coming up. They’re good too, but they’re difficult to talk about to a lay audience. For example when talking about genes how much information is being transferred? A common tactic is to say that N gigabytes of data are in a speck of DNA. It sounds good by realistically I don’t know how much information is in a Gigabyte. I know numerically, but I can’t usefully imagine it. I doesn’t help that I can look at two identical photos, one raw from the camera is 6MB and the other as a JPEG is 250K. Yet on my screen they look exactly the same.
An alternative is to consider information as the equivalent of books. I know what a typical novel is like so the information could be the equivalent of so many million books. The problem is can you really imagine a million books? A similar problem has been tackled by John Finnemore in Cabin Pressure where the characters attempt to imagine a hundred otters with the aid of their plane, GERTI.
Another problem is a matter of experience. Genetic Modification is often presented as a bit like copying and pasting a function from one computer program to another. This is true, but it overlooks our experience of computers. The aim of the analogy is to say there’s no danger, but we know computers crash on a regular basis. The function analogy is sensible, but it overlooks how we really use computers.
So today’s work is get excited about some new research (easy) and then try to explain why its exciting to someone who doesn’t care (more difficult). This has to be done in a way that shows quickly what the problem without any analogies distracting from the important part of the paper.
i know some researcher cringe at using analogies. Why can’t we let the research speak for itself? The answer is: how many members of the public have a background in plant sciences? It’s not about dumbing down. The people who are interested in this will be intelligent. It’s about giving these people a starting point to find out more about the research that’s going on.
AJ Cann
Analogies can be as misleading as they can be illuminating, but for better or for worse we cannot use language without them. (Already I have mixed an old metaphor about leading with one about light in my word choice.). Anything you can imagine is both similar to and different than something else, but when we compare a raven and a writing desk, they are more different than a raven and a crow. I can only guess at the difference, whereas I can easily see the difference between that RAW and jpeg file you see as identical on your monitor. So analogies are also strained by expertise, where familiarity sharpens perception, and experience points out the many differences overlooked in the analogy comparing things. This is why scientists hardly recognize their work in press releases, but also why they need to learn to communicate their research to a wider audience. It’s not, as you point out, a lack of intelligence, but a lack of expertise and experience that makes science communication hard. I see scientists and doctors struggling to communicate with colleagues, where neither assumes the other is dumb–but they have to compare their research to something the other understands in their own field to succeed. Analogy.
What interests me about “Bad Analogies” is the variety of cases you invoke. You must imagine another species’ P.O.V., which necessarily requires analogies to our own perception. Then move on to the problem of relating the information stored on DNA to computers, which is itself already the underlying analogy assumed by genetcists in all their hypotheses. its our best analogy, which is why its assumed, but the more that is understood about both DNA and computation, the more strained the analogy becomes. Then you move on to analogies of immense proportions, where all of us, especially mathematicians, know we are incapable of imagining them. Math provides a system for proceeding where we would otherwise be intuitively blind. Not just you and I, but Brian Greene can’t imagine 10 dimensions, and the symmetries and connections between subatomic particles are not strings entangled like iPhone headphones. Positing extra dimensions in string theory allows other variables in the equations of quantum mechanics and general relativity to be compared.
I think the main challenge in science communication is to convey this sense of proportion, so they can assess risk, and to teach the public to accept answers in the form of probabilities and distribution. But this means the fisherman needs to hear climate change in terms of a rougher sea and a seasonal catch out of sync with his license to use nets; and a comedian has to hear the way applause varies in different cities on different nights for different jokes. It’s very hard to find a universal analogy for a spread.
I myself am besides myself whenever information is counted in numbers of words or books. Computers store all the words and books already, but because they can barely read they don’t know a damn thing about them. DNA RNA and proteins, the cellular and organism’s environment are all storing information about each other in such complex ways, that our going analogies fail consistently. They’re neither encoded in binary bits, nor subject to the kind of word order grammar and declension by case we expect. In your short piece you auto corrected a them as “then”, dropped an “is” and an “s”–and it was still perfectly comfortable to read–yet this level of copyerror would be unacceptable in our own genome…maybe that fidelity is the high price necessary to encode our complexity with so few genes?