If you could listen to - and deeply analyze - several lifetimes' worth of music, what would you learn?
You would probably find out whether certain elements - rhythms, note patterns - are consistent across cultures, regions or time periods. You may find out if certain elements of music resonate with all of humanity.
Until recently, such questions were hypothetical. Nobody can listen to several lifetimes' worth of music. But today researchers are harnessing the power of massive supercomputers to do the listening for them. In the process, they are asking questions about humanity that would have been impossible to answer just a few years ago.
And it's not just music. Researchers are for the first time finding out what millions of articles from long-forgotten literary journals can tell us about Victorian society; or what census data, old maps, voting records and other historical fragments can tell us about how Canada evolved as a nation.
The rise of high-performance computing - supercomputers able to perform billions of calculations in the blink of an eye - over the past decade has provided academics with one of the most useful tools in the history of organized research. Scientists have used them to measure the movement of atoms and to predict how the global climate will change over the next 100 years. But increasingly, some of the most innovative and important uses of new digital technology are coming from the humanities.
Academics are increasingly attempting to apply the power and scope of supercomputers to answer questions about human culture and psychology the same way scientists use such machines to try to guess the age of the universe. What can thousands of hours of recorded speeches tell us about the way people use pauses and intonation, for example? What can 53,000 letters from the 18th century tell us about the Enlightenment?
"If you think about the amount of evidence you can handle in a lifetime, you can only handle so much," says Geoffrey Rockwell, professor of philosophy and humanities at the University of Alberta. "That limits the types of questions we can ask."
The ongoing digitization of humankind's works, along with the increasing power of modern supercomputers, has given researchers in areas ranging from history to literature to music the ability to ask new questions. Indeed, they're using the world's brawniest computers to study the very notion of ideas.
For example, Prof. Rockwell notes, the conventional model of how ideas spread is mostly linear: Socrates passes an idea on to Plato, who passes it on to Aristotle, and so on. But is this actually the case, or is the transfer of ideas much more amorphous, spreading through communities in different ways?
"Now we can ask questions about very large collections of information," says Prof. Rockwell. "We can use this massive evidence to study patterns of how ideas are moving."
Canada has had a Society for Digital Humanities, specializing in just this kind of work, since 1986, but it's never been as active as it is now. A 2006 study of Canada's up-and-coming research sectors - areas where the country is punching above its weight and where research dollars should be going - listed humanities computing alongside more predictable areas such as petroleum engineering.
But it is only recently that academics have focused on using supercomputers to look at massive quantities of words, pictures and music, among other data. Late last year, the biggest research agencies in the UK, U.S. and Canada held a grant competition for the best proposals, called the Digging into Data Challenge.
Among the winning projects came one from a group of researchers, including some at McGill University, who plan to collect 23,000 hours of music across every style and region imaginable, and digitally analyze the data to find the underlying structures of global music.
Another project proposes taking digitized classic Greco-Roman text and automatically "enriching" it by having a computer run through the text and automatically tag it with links, much the same way as text on a website contains hyperlinks to other pages. For example, the word 'Athens' would be automatically tagged with a classical map of the city, or an entry to the city's Wikipedia page.
"We've digitized most of important stuff - now we can now start going in," Prof. Rockwell says. "If we have a gazette of all known locations in the classical world, we can run it against text - words can become links towards place-names in the atlas. That's a form of automatic enrichment."
Enrichment may not teach researchers anything new about the texts they're enriching, but it makes teaching others much, much simpler, by linking ancient texts to modern teaching resources.
Another one of the Digging into Data proposals involved having a computer look at almost 200,000 recently digitized records of trials from the Old Bailey, London's central criminal court and generate assertions based on what it read - for example, hypotheses about the lives of lower-class London residents, and punishments for certain crimes.
However, the researchers still needed a way to test these assertions, so they proposed "crowdsourcing" - essentially, presenting assertions to users who search the public Old Bailey database for related terms and asking if they want to help prove or disprove them, turning the public into a part of the researchers' arsenal.
(Likewise, the researchers working on the Victorian literature project decided to create a Twitter feed, called " conjecturator," to post their automatically generated conjectures to the world).
The Digging into Data projects are expected to run until March 31 of 2011, but researchers are already looking at others. Ultimately, the aim of such computer-powered humanities research is to model humanity. Imagine predicting humanity's social evolution on a massive scale the same way scientists model and predict climate patterns.
Such problems would have seemed impossible to tackle just a few years ago. Now, says Prof. Rockwell, "It's not the case that everybody is asking these questions, but nobody questions their importance, or looks at me like I don't belong here."
Omar El Akkad is The Globe and Mail's technology reporter.Report Typo/Error