Tell us about your Google search for a stolen letter.

I searched for 'Descartes' and 'autograph letter', and got a hit at Haverford College in Pennsylvania on the first page. The listing immediately caught my attention. I had been using those search terms for a few years, so I knew the first 30 hits very well. From this letter, to French mathematician Marin Mersenne, we learn that Descartes had changed the introduction to his 1641 book Meditations on First Philosophy at the last minute. The letter had been stolen in the nineteenth century by Guglielmo Libri, a gifted historian of mathematics and an ardent bibliophile and collector. Eventually he not only bought manuscripts, but also stole them. He became an inspector of French public libraries, and looted them. He took about 80 autographed letters of Descartes from the Institut de France in Paris.

Dozens of René Descartes' letters have been missing since the mid-nineteenth century. Credit: M. ROURKE/PRESS ASSOCIATION

How did they end up in the United States?

After he had cherished them for several years, Libri decided to sell. That is how his collection got dispersed. The scandal became public in 1848 and he fled to England. The majority of the letters returned to France, but about 30 were untraceable. Some were completely unknown and never published, such as the one I discovered at Haverford. It was a very rare find.

Have any letters proven problematic?

The standard edition of Descartes' correspondence contains one letter that is a fake. It was copied from a novel of the late seventeenth century, which ridicules the belief that people can talk to spirits. The most notorious forger is Denis Vrain-Lucas, who sold thousands of counterfeited letters, supposedly by Descartes, Jesus, Mary Magdalene, Charlemagne, Aristotle, Lazarus before and after his resurrection, and so on. Once unmasked, Vrain-Lucas pleaded innocent, saying that the question of authenticity is trivial. The problem is that there is money to be made from forgery. A genuine letter from Descartes could sell for US$200,000 or more.

Why create a new edition of correspondence?

Erik-Jan Bos Credit: NICK HIGGINS

The standard edition is a century old and contains a lot of supplements and errata; the next edition contains even more. The corpus is a nightmare. If you want to find particular material, you will get lost. With a new edition we can check formulae and mathematical texts. In previous editions, complete lines were left out, making the maths incomprehensible at times. Claude Clerselier, the first editor of Descartes' correspondence in the seventeenth century, wrote in the preface to the first collection that many of the manuscripts were difficult to decipher, so he had to guess. We can't be sure exactly what he did because the manuscripts he edited have vanished. After he died they went to a French scholar who died, then on to the next scholar who died, and finally all the material went back to the mother of the first scholar. I don't know what she did — maybe put them on the fire during a cold winter?

How has the work of finding correspondence changed?

The amount of labour my predecessors had to go through is unimaginable. After I Googled the lost letter, the librarian at Haverford immediately took digital pictures of the manuscript, plugged in the camera and sent them to me. Twenty years ago, I would have had to wait for weeks. As for the search, half a century ago, if you went to a library in Paris, you could search for weeks and still miss important material. In the past decade, inventories of manuscripts have come online. But you have to be clever. If you just search for 'Descartes', you get millions and millions of hits.

Why study Descartes' letters?

He comes alive. He looked down on contemporary philosophers, scientists and mathematicians, including Pierre de Fermat, with utter disdain. He also thought that he was always right. And there are unexpectedly personal letters. We have one by Descartes written to the local bailiff pleading for leniency towards an accused murderer. He wrote wonderful letters to friends who had lost loved ones. Other highlights include a marriage contract drawn up in 1644, for which one of the witnesses was Descartes. The bride was the mother of Descartes' daughter. He recognized his fatherhood but never married the mother, although presumably he saw to it that she got married. That contract was a way to take care of her.

Has reading Descartes helped you with maths?

An encounter with maths in his letters can be like reading Greek. The seventeenth-century way of doing maths was different from today's: they were interested in other problems. We live in a post-Cartesian era, mathematically and philosophically.