In a review of Doris Lessing's The Four-Gated City, John Leonard created my favorite first line for a review: "On finishing this book, you want to go out and get drunk."
I could say the same for Light, which has been called M. John Harrison's triumphant return to down-and-dirty science fiction, which he seemed to abandon after The Centauri Device (a book I think I remember Harrison saying he feels is his worst). It is, indeed, a return, and it is certainly a triumph -- a triumph of vision, a triumph of prose, and a triumph of construction. It has garnered strong reviews from Ian Banks, Paul Di Filippo, Jeff VanderMeer, Cheryl Morgan, and others.
It is an easy book to admire, a difficult book to like. Perhaps that explains what Cheryl Morgan described as "the bored thumbs down it got from [David] Hartwell [of TOR, etc.], [Gardner] Dozois [of Asimov's] and [Charles N.] Brown [of Locus] at a panel at ConJose". Those boys should know better, though it may be that their critical senses have been dulled by reading too much mediocre SF and thinking it's what everyone should aspire toward. Light is worth the work of learning to move beyond admiration, and I will go out onto a shakey limb and say that calling it "boring" reflects on the reader the same way calling The Brothers Karamazov boring does -- yes, m'boy, you might find it to be so, but that is only because you're not willing to work hard enough.
Casting hyperbole aside for a moment, Light is not The Brothers Karamazov, though it could be as influential on SF writers as the messy monster of a Russian masterpiece was for 20th century writers, so long as the SF writers were willing to be influenced by a great book rather than market forces.
Before I start talking a lot about the difficulties Harrison's novel poses, let me offer a visceral response to the tone-deaf charge that the book is boring. I am a slow reader when I like a book, because I read every word, sometimes over and over. I read the first 100 pages of Light in my normal manner, 20 pages here, 20 pages there. Then, like avarice or the flu, it consumed my life. It took me a little over ten hours to read the last 200 pages of the book. Ten consecutive hours, with short, occasional breaks for food and other necessities. I last did that four years ago with the final two hundred or so pages of War and Peace. Another boring book. (Yes, all my touchstones are 19th century Russians, who in Dostoyevsky and Tolstoy showed us the potentials and limits of the novel, and in Chekhov the same for the short story and play.)
Light is not boring. Be aware, and it will infiltrate you, it will propel you through its mazes and architectonic wonders. Enter looking only for "entertainment" (governor of the vegetative state) and you will be disappointed, because though the novel is hugely entertaining, it is also disturbing, it is cold, it is cruel, and it offers no easy answers to any question you could conceive. Enter seeking escape, and you will find yourself mocked in the many scenes of masturbation.
What Harrison has done in Light is give us the alien love child of Cordwainer Smith and Louis-Ferdinand Celine, and like any child it is bound to its parents at the same time it renounces them. There is Smith's vision (and cyborgs and cats), a hint of Celine's bile (toward humanity and, perhaps, women), but there is also hope, a kind of yearning for redemption in characters you would never invite to dinner, but who are, in all their awfulness, compelling as a car crash. The only real hope, the only true possibility for redemption, is that the universe can be rebuilt -- a hope and a possibility that is alive on the last page.
What is most stunning about Light, though, is its construction. The prose has been rightly praised, the ideas and speculations have been lauded, but I haven't seen enough regard for Harrison's ability to build a book where it seems every sentence connects not only to the sentences around it, but to sentences 100 pages away. Jeff VanderMeer rightly invoked Nabokov, and this is a novel that could provide plenty of fodder for Ph.D. dissertations. After reading the book once, it pays to go back to just about any page and think about how it fits in to the grand scheme ... although the grand scheme itself can be difficult to discern in its details, because the details are scattered like the many motes of light in the book throughout its pages. It's fine to read this book the first time as any devoted fan of space opera might -- indeed, that's probably the best way to read it first -- but it deserves at least one more reading, read from the point of view of a quantum physicist looking at the behavior of radiation.
Light will be published in the U.S. in August, so if you live here and haven't been able to scrounge up a British import, you can pre-order it. Read it when you're ready for both challenge and delight.