A couple of weeks ago, Niall Harrison published something of a lament about review venues that are more inclined to publishing favorable reviews than unfavorable ones. He says:
We’ve been here before with regard to the insidiousness of “mostly positive” reviews, but this seemed worth pulling out as an example where the harm caused by the policy is more obvious than usual. It does a disservice both to readers who might have seen the review and now will not, and to the field of sf reviewing and criticism as a whole, for which full and honest discussion must be a priority; I hope, though I accept it is likely in vain, that Prominent SF Magazine Editor feels a mite embarrassed by their reviews policy today. That the writer in question has subsequently decided not to review at all, at this stage in their career, also makes me sad — it impoverishes the dialogue, in more ways than one — but it is understandable.
For the past 21 years, we have run The New York Review of Science Fiction on the principle that what people like about a book is more interesting than what they dislike, and we aim to publish reviews on the strengths and weaknesses of good books. This means that we publish reviews that are for the most part positive; occasionally publishing a truly negative review if the review itself offers significant illumination.
There are all kinds of reasons one might react negatively to a book, many of them personal. One can be deaf to the virtues of a writer for several years before finally getting it. I would be quite embarrassed if I had been publishing my notes on stories I truly disliked over the years I've been doing Years' Best volumes. There are several really fine writers I really didn't get until the right story came along and things suddenly clicked into place for me.
We publish to promote the aesthetic advancement of the field and are not a buyers' guide. Some review venues that may think of themselves as buyers' guides may do things differently, which is fine.
Vast numbers of books are published each year. Many of them do not merit much discussion or review attention, in our opinion.
Ideally, book reviews are about books, not about reviewers and their feelings. A review should accurately characterize the project of the book and how it fits into the genre. Whether the reviewer is in sympathy with the project of the book is secondary to its accurate depiction.
The literary essay is a form more suited for exploration of rifts between critics feelings and the books they encounter.
UPDATE: David points out to me that he wrote a NYRSF editorial on this general topic in 2004, entitled "Blooming" which is about negative reviewing as performance.
FURTHER UPDATE: James Nicoll thinks I missed the Fail Fandom subtext of Niall's post. I didn't: I did understand what Niall was trying to say, and I disagree. An essay is a more appropriate form for a critic who disagrees with the basic project of a book than a book review.