For as long as book reviews have been published, writers have argued that book reviewing itself is in a state of crisis – a pointless exercise, a waste of time. In 1846 Edgar Allan Poe called reviews nothing but a "tissue of flatteries". Virginia Woolf worried that the reader was none the wiser because "the clash of completely contradictory opinions cancel each other out". According to Elizabeth Hardwick in 1959, "sweet, bland commendations fall everywhere upon the scene; a universal, if somewhat lobotomised, accommodation reigns … For sheer information, a somewhat expanded publisher's list would do just as well as a good many of the reviews that appear weekly." And it's even more fashionable now to be "against reviews"..

Today, the crisis takes a different form: the challenge of the web; the decline of the critic – you know the deal. More narrowly, there's Amazon, and its anonymous, unmarshalled reviews. There have been numerous flare-ups about these – the self-reviewing, the hate-reviewing, the downright-unreadable-reviewing, and so on. The latest unholy behaviour to come to light is of authors paying for positive reviews. As the New York Times has recounted, and Salon has discussed, one Todd Rutherford set up a now-defunct operation called GettingBookReviews.com that would, for a fee, ensure the publication of dozens of five-star consumer responses to a submitted book. It filled his pockets with cash, and, in at least one case it seems to have worked, helping to create an ebook bestseller out of a self-published novelist.

So: are Amazon et al, with their bought-and-paid-for notices, killing off the book review? Or are they rather making the traditional, commissioned book review more important than ever?

The unsavoury Amazon stuff notwithstanding, no one is about to write off the whole business of reader-reviews. They are, in any case, unstoppable, and the sheer weight of numbers suggests that only a tiny fraction of them can be corrupted. Undeniably, they represent the latest stimulating chapter in the rather agonised history of book reviewing (read Orwell on the subject, and Edmund Wilson, and Cyril Connolly, and James Wood …) The ones most to be trusted, however, are perhaps more likely to be found on smaller, more specialised sites than Amazon – Goodreads and Librarything, for example (and hopefully among the Guardian's reader reviews too). Yes, online anonymity will always raise problems, and no one can ensure, with this kind of reviewing, that what the New York Times calls "the sacred arm's-length relationship between reviewer and author" is being preserved. But there are book communities and book communities, and it surely pays to choose carefully where you read and write your reviews.

But it also seems to me that the Amazon scandals reaffirm the importance of the much-maligned traditional book review. Reviews in, say, newspaper books sections (I'm biased) are vital in offering a properly critical (often negative) opinion of new books: a necessary accompaniment to (also important) articles in the same sections that simply showcase books, or report interviews with authors: these can all too easily become elegant exercises in PR. The book reviewers are chosen by commissioning editors, they don't choose themselves, and their judgments, if the editor is doing his job properly, must be properly backed up. Yes, there's only one wise voice rather than the wisdom of the crowd, but these critics are convincing, independent, entertaining and trustworthy enough that, time and again, they are paid to offer their opinion. And not in the way that Todd Rutherford was paid, by the authors of the books themselves.