On the Origin of Species: The Preservation of Favoured Traces. Click to launch »
In November 1859, Charles Darwin published his seminal work, On the Origin of Species, in which he introduced the world to his theory of evolution. Since its first edition, the book has undergone considerable edits and revisions.
Last week, Ben Fry, director of Seed Visualization and its research arm, the Phyllotaxis Lab, released a new tool, called The Preservation of Favoured Traces, which allows users to watch the book evolve across six editions as Darwin reconsidered his arguments and responded to criticisms. Fry’s visualization shows how scientific theories themselves are not static truths, but living ideas that are constantly evolving as new evidence comes to light. Seed spoke with Fry to learn more about his new project and what future he sees for visualization tools.
Seed: Why visualize the evolution of On the Origin of Species? What do you hope to accomplish?
Ben Fry: I spoke to a Darwin scholar about this project and she asked me the same question. “Why do this? We already know what all this stuff looks like,” she said. But by “we,” she meant the community of Darwin scholars that have access to all of this fascinating stuff. We wanted to get it out to a larger audience. People are curious about Darwin’s ideas and what his theory meant.
Furthermore, people are frequently under the impression that scientists know everything and that scientific theories are fixed. Within science, however, the opposite is true. Scientists know that the more that they find out, the more questions they will have. By showing how the text of The Origin… changed over time, we are visualizing how scientific ideas evolve.
Seed: What were the main challenges you faced when you first began the project?
BF: Finding the data set was certainly a big issue. In fact, when the idea came about a couple years ago, digitally transcribed versions of all of the editions of the book weren’t yet available. The Darwin Online folks actually did all the work of transcribing the texts and creating a very nice archive.
Also, developing visualizations with text and language is very different than working with numerical data. It’s more about sentence construction and the function of language. There is a shocking amount of work that goes into corralling that sort of data into something useful. The book is 150 years old, so I can’t just use a current dictionary.
Seed: Will this tool enable further research?
BF: One interesting thing you can see from the visualization is that over the course of Darwin’s edits he included more and more responses to people’s criticisms of his theory. Finally, in the sixth edition, published in 1872, he actually just took all of those scattered references and made them into their own chapter where he addressed them directly. This shows a shift in his approach to presenting his theory. Prior to the sixth edition, he had tried to stay out of that fray and rely on his network of friends to defend his ideas.
This particular project was really about being able to open up the domain of intense scholarship and render it accessible to a general audience. Visualizations engage people in research by giving them something they can actually play with and imagine. People connect to it. I’m working on a bigger piece right now that will serve as more of a tool for researchers.
Seed: How will the bigger piece be different?
BF: Users will be better able to compare subtle edits made to each edition—sort of Darwin with “track changes” turned on. For instance, there are places where Darwin increased or decreased stress on a particular idea, or backed away from certain points. We’ll add annotations, and it’ll become something historians and students can use.
Seed: This seems like a very scientific approach to historical research.
BF: It is. And to that end, I think we are about to see a whole lot more work like this. In the next few years we are going to have a lot more people doing work around text-based visualizations. We have really turned a corner with regards to the amount of text that is actually available for us to work with. One of the things that caught the eye of the creators of IBM’s Many Eyes—a website that provides users with data visualization tools—was that far more people were interested in creating text-based visualizations than numerical-based ones. An astonishing number of people really relate to them.
Seed: Looking forward, what other text-based visualizations are you planning to develop?
BF: I’m curious about how language changes over the course of a century or even in just five or ten years. There are a lot of things that get inserted into the language, like the way that people talk about things and frame issues. I don’t want to get into particular project ideas though, because I think it sort of jinxes me when I do.
Originally published September 10, 2009