18/12/2018

Mathematicians as protagonists are (still) very rare in fiction.

Livros analisados formam massa de 500 bilhões de palavras Foto: Furcifer Pardalis

How many books have you read that feature a mathematician as the protagonist? If you can't remember any titles, don't worry, it's not a memory problem. Considering the years from 1800 to 2008, the picture is discouraging. And it gets worse from 1952 onwards, if we analyze English fiction . The irony is that this is only possible to discover because of the contribution of mathematicians involved in the creation of the Google Books Ngram Viewer.

The online tool allows us to perform statistical analysis of the content of millions of books found in public libraries worldwide and digitized by Google to populate the Google Books search results.

The more than 15 million books already scanned produce an immeasurable amount of information, data that can help us, for example, to understand changes that have occurred in culture over time.

“What remains is a collection of 5 million books, 500 billion words, a sequence of characters a thousand times larger than the human genome – a text that, when written, would extend from here to the Moon and back more than ten times – a true fragment of our cultural genome,” details Erez Lieberman Aiden in a TED talk given with Jean-Baptiste Michel, both researchers at Harvard University in the United States. The pair, along with two other researchers, created the Bookworm prototype, which served as inspiration for Ngram.

It's possible to search using just one word, such as "mathematician," a phrase, or even letters – something important, for example, for philologists. The search can be done in different languages – unfortunately, Portuguese is not included – or corpora, such as those that include fiction books in English.

In a search for "mathematician" in the English Fiction corpus, we see that the peak of the word's appearance was in 1840. But don't get too excited: the percentage in that year reached 0.0010847217%. And, since 1952, it has only been declining. According to Aiden, programs like Ngram Viewer are tools that enable the use of what he calls culturomics.

“It’s like genomics. Except genomics is a lens that allows biology to see through the window of base sequences in the human genome. Culturomics is similar. It’s the application of the analysis of the enormous amount of information collected to the study of human culture. Instead of looking through the lens of a genome, we look through digitized pieces of the historical record. The good thing about culturomics is that everyone can participate,” says the expert.