In the preface to this special issue of FORUM on Readers and Writers, I wish to take the opportunity to think briefly about a question that preoccupies all historians of reading, to a greater or lesser extent, and that is the problem of evidence. Reading is an evanescent activity, which mostly goes unremarked, unrecorded, and very often, unnoticed. Under such circumstances, how can we retrieve its history? From Robert Darnton in 1986, outlining his ‘first steps toward a history of reading’, to those of us still working on the history of reading in 2016, we have been wrestling with precisely the same problem. As Simon Eliot put it in 1992, ‘any reading recorded in an historically recoverable way is, almost by definition, an exceptional recor...