"Digital Humanities" is a rather amorphous term that appears to cover just about any consideration of the "humanities"--in particular literature--as it is manifested in the existence of texts in digital or online form. Wikipedia defines it very vaguely as "an area of research, teaching, and creation concerned with the intersection of computing and the disciplines of the humanities."
Perhaps the most prominent area of "inquiry" in the digital humanities is "data mining," the exploitation of texts in digital form to examine them in purely mechanistic ways to uncover patterns or extract statistical, measurable "information." This is essentially the form of digital humanities that Stephen Marche attacks in a recent Los Angeles Review of Books essay. Quite simply, Marche tells us, "Literature cannot meaningfully be treated as data."
This is quite obviously wrong. "Literature" as the accumulation of written texts, like any other accumulation of texts regarded from a particular perspective, can certainly be "meaningfully" approached as a source of data. The real question, of course, is whether this is something one would want to do. To the extent this really the issue Marche is raising, his critique of data mining deserves to be taken seriously, but that many intelligent people do indeed clearly believe that data mining and statistical analysis of literature broadly defined is worth their time can't really be just abruptly dismissed.
Marche's contention that digital humanities is "yet another next big thing" at a time when literary studies needs a new big thing is a more cogent response to the rise of digital humanities, and in my opinion gets at the most significant limitation of data mining as a phenomenon in literary studies. Academic criticism has been for the past forty years certainly, and perhaps for the entire history of academic literary criticism, a series of new "things," new approaches to the "scholarly" study of literature. What all of these approaches, including digital humanities/data mining, have in common is that they take the emphasis away from "literature itself," from the inherent value of the reading experience itself, to other ways of "using" literature--for historical or political analysis, for illustrating theoretical positions, etc. Data mining is no more excessive in its abandonment of literature for other, more "cutting edge" pursuits than these earlier "advanced" agendas.
Most of the points Marche makes about the ability of literature to elude the reductiveness of something like data mining are perfectly well-taken, but in the context in which this activity occurs, the academic disciplines that ostensibly take literature as their subject, they are entirely irrelevant. It is indeed the reading experience afforded by the greatest works of literature that ultimately makes them valuable, which cannot be accounted for by this most recent development in the "literary" academy any more than the other non-literary approaches to literature that have dominated academic criticism for decades. But academic criticism is no longer concerned with reflection on or enhancement of the reading experience, is no longer concerned with literature as a subject of humanistic study at all. If Marche wants to convince the literary academy to return to the reading experience as its subject, good luck, but otherwise simply accusing digital humanists of neglecting the literary in literature is more or less redundant.