Nov 24, 2011
Since the 1990s, scientists have recruited idle home computers to help with data analysis. I remember that our family’s clunky old PC used to run a screensaver that used its spare computing power to analyse the structure of cancer molecules and send the results back to a distant lab. We felt pretty good about ourselves when the diagrams of cells floated across the screen but ultimately it was the processor that was doing all the heavy lifting.
These days a much broader range of research questions can be tackled through crowdsourcing techniques that harness genuine human intellect. People, after all, are still far better at identifying patterns than machines. Active communities of astronomers, medical researchers and physicists are now aided by teams of web-connected volunteers who make simple pattern judgments to classify images. Collectively this information contributes to discoveries that could never have been made by lone individuals.
Until recently it never occurred to me that crowdsourcing could be used to solve language problems. But then I listened to this seminar from Steve Bethard and Vicki Lai. Using Amazon’s Mechanical Turk the pair enlisted hundreds of individuals to make judgments about English sentences. This information equipped them to devise a targeted laboratory experiment for testing how the human brain processes metaphors.
And a recent article from Boston Globe, ‘How crowdsourcing is changing science‘, describes yet another linguistic application of the technique. A cache of 2000-year-old ancient Greek manuscripts, unearthed in Egypt over a century ago, are finally getting transcribed with assistance from the public.
The manuscripts are rich, fascinating, and varied. The texts include lost comedies by the great Athenian playwright Menander, and the controversial Gospel of Thomas, along with glimpses of daily life — personal notes, receipts for the purchase of donkeys and dates — and the occasional scrap of sex magic.
The pace, however, has been glacial. After a hundred-plus years, scholars have been able to work through only about 15 percent of the collection. The finish line appeared to lie centuries in the future.
But a few months ago, the papyrologists tried something bold. They put up a website, called Ancient Lives, with a game that allowed members of the public to help transcribe the ancient Greek at home by identifying images from the papyrus. Help began pouring in. In the short time the site has been running, people have contributed 4 million transcriptions.
Try it yourself. You don’t need to know a skerrick of Ancient Greek. Many of the fragments look completely impenetrable at first, but it’s a matter of simply clicking on the most recognisable letters, assigning a value and moving on to the next fragment. Persevere and you’ll gradually recognise more and more of the letters. In turn, this incremental work will eventually allow scholars to search for snippets of intelligible phrases and resurrect hitherto unseen examples of Greek literature.
In this case, it’s the small, random acts of human intellect that are slowly filling out the bigger picture. Or as a famous Aristotlean once put it ἡ φύσις οὐδὲν ποιεῖ ἅλματα – ‘Nature does not make sudden jumps’.