This dissertation is about the psychology of imagination in the last two decades of the nineteenth century. In the critical account of this period, much has been written about the relation between literature and sight; it has hardly been noted, however, that the period was marked by the emergence of a field of research into a different kind of "vision" -- the images produced by words on a page. My dissertation addresses this gap in two ways: first, in an account of a major shift in the psychological understanding of the mind's eye in this period; second, in a series of readings which explore the ways in which writers and critics responded to this new science. Both accounts begin with Francis Galton's 1880 publication of "Statistics of Mental Imagery" -- the first study of its kind. His findings -- still cited by psychologists today -- disrupted the idea that words predictably or even reliably produced "pictures" in the mind, thus troubling more than a century of philosophic and literary debate over the nature of mental representation. As William James observed in 1890, Galton's study had "made an era in descriptive Psychology." After repeating Galton's investigation in his own classroom, James concluded that "There are imaginations, not `The Imagination,' and they must be studied in detail." My dissertation traces the work of a series of writers who drew upon this research. In chapters centered on Charlotte Perkins Gilman, Mark Twain, William James and Helen Keller -- all of whom were familiar with Galton's study -- I locate a literary tradition which found its value not in objective correspondence with the outside world, but rather, in the embodied feeling of the mind at work. These writers took from psychology the premise that mental vision, like physical vision, had limits -- limits defined by the body. While this limitation could be understood as a constriction, it also suggested the possibility that the imagination could take on the status of physical experience -- that the mechanical act of transforming shapes into signs could become a form of training for "real" life. In order to understand these texts, I argue, we must attend to what James described as the "half" of reading that is not present on the printed page -- the "half" provided by the reader him or herself. In pursuing this claim, I model a style of critical analysis that remains grounded in close reading, but that nevertheless seeks to account for the reader's imaginative experience. This style of reading critically re-orients our understanding of these texts, moving us away from "problem" plots and unresolved themes, towards larger structures of perception. These writers, I argue, do not seek to inform us about another person's experience; rather they provide us with a grammar of experience -- a technique for living intended to last well beyond the moment when the book is set aside.