While I share with Steven Shaviro the inclination to "only write about books I like, letting the ones I don’t be passed over in silence," I can't share his belief that, even while writing about admired books (perhaps especially when writing about such books), it's appropriate to "make no pretense of describing them accurately, as they are — but rather [to] use them (or appropriate them) to come to some understanding for myself, which means that the author might just as well be upset by my (admitted) misinterpretation, as he/she might be pleased by the fact that I liked their book."
I'd be less concerned that I might have misintepreted a particular work than that I made no attempt to "describe [it] accurately." I'd go so far as to say I would consider any writing about literature (in particular a book review) that didn't make such an attempt to be something other than literary criticism in the strict sense of the term. (Perhaps Shaviro would as well; perhaps he doesn't regard his own work as ""criticism" per se. If so, he is assuredly following the practice of most of academic criticism, which for decades now has mostly abandoned what was once understood as "literary criticism" for more the more ethereal pleasures of Theory and other philosophy-inflected modes of discourse.) Thus, while I might wildly misinterpret a work because of my own flawed reading, if I have nevertheless made a good effort to account for the work--to describe poems and fictions "as they are"--I am still engaging in a
recognizable if ultimately error-ridden act of literary criticism.
You might think that, given my assent to John Dewey's conception of art and literature as modes of experience--which in one of its subsequent reformulations became Stanley Fish's version of "reader-response" theory--I might find the notion that I could properly use art "to come to some understanding for myself" entirely acceptable, even an immediate consequence of Dewey's own philosophy of art. But Dewey did not license an unrestrained, free-for-all approach to criticism and interpretation. (And ulitmately neither did Fish: in his schema, interpretation is limited by an "interpretive community" that can accept some some readings as accurate or persuasive while rejecting others as misguided.) For an aesthetic experience of literature to be as complete as possible, the reader must undergo "an ordering of the elements of the whole that is in form. . .the same as the process of organization the creator of the work consciously experienced." If this sort of "ordering" doesn't occur at all, if the reader is free, in effect, to create a different work than the one the artist has presented us with, aesthetic experience has been impeded, cut off. If criticism is, in part, one reader's attempt to communicate his/her own "experience" of the work at hand, some effort to put aside what I want "for myself" and to represent (re-present) the work on its own terms is required.
Certainly the reader or critic is then free to re-process what Dewey calls the "detail" for him/herself. This reprocessing is what we usually regard as "interpretation." We don't have to stop with what "the creator of the work consciously experienced." Much of what we value in some works of literature goes beyond what the artist "consciously" intended. That we are able to reassemble the details into a new synthesis of text-writer-reader only underscores the reader's contribution to aesthetic experience, and such a synthesis only enriches the text. Sometimes we might even think that what the writer intended is at odds with what the text actually does on a close reading of it. (Paradise Lost, for example.) Such an intepretation might be found compelling by other readers and subsequently become an accepted version of what the text "means." In some ways, the creator is thus left behind, the text itself taking on a life of its own, so to speak.
Still, to entirely forfeit the obligation to "describe" the object of one's reading experience does seem to me a critical dereliction. The work is there, and we should begin by acknowledging it.