House of Leaves came to me highly recommended by someone whose literary taste I trust, and I have to say that for the first 500 pages or so I was not disappointed. The genuinely creepy story at the center of the book is a masterpiece of suspense and terror, made all the more horrifying by the framing narrator's frequent delays, interruptions, and (what we assume is his) psychosis. By itself, the account the book gives of the fictional documentary The Navidson Record would have made an excellent novel. But Danielewski wanted it to be something more, and to some extent, he succeeds.

Using a conceit employed in Umberto Eco's The Name of the Rose and a great deal of Borges’ work, the central story in House of Leaves is based on a lost original that may never have existed in the first place. This emptiness at its center informs the book from start to finish, and despite the proliferation of information, the book, like the continually expanding house it describes, only increasingly reveals the impossibility of meaning and the futility of attempts to locate truth or significance. As a postmodern foray into the ultimate impenetrability of the sign, the book succeeds in a big way. I read a review that I can’t seem to locate again (lost original?) that called House of Leaves an “assault on the nature of the story,” a description I thought particularly apt. The book is also an immensely entertaining parody of academic criticism, mimicking the tendency of academics to pick apart at the insignificant details of a thing while comically (or tragically) failing to account for its entirety.

The book is less successful in its portrayal of its framing narrator. Johnny Truant is a poorly drawn caricature of a drifter/druggie whose obsession with the manuscript he assembles is immensely uncompelling. Perhaps Johnny is uninteresting because his insanity is rooted not in existential despair but in the most banal of sources: his difficult childhood. As such, the narrative's eventual obsession with Johnny's mother feels a little cheap. In all fairness, though, it's possible that an author as intelligent as Danielewski is poking fun at people's tendency to use personal psychology to explain problems more universal in significance. I'll have to read the book again to really be sure what I think of that.

The book's experimental design is at times hugely satisfying and at others annoyingly self-indulgent. The use of multilayered footnotes and voices is immensely effective, adding a dimensionality to an already complex text. And occasionally, when Danielewski only prints a few words per page or turns the text sideways or upside down to emphasize/reveal the architecture of the text, the effect enriches and extends the narrative. But the other half of the time, his use of experimental printing and layout to heighten dramatic moments adds little to the narrative. Like most attempts to treat text visually or cinematically, House of Leaves misunderstands the profound difference between the act of reading and the act of viewing, and rather than addressing this difference, attempts to elide it.

About 500 pages in, the book proper ends and the appendices and exhibits begin. And this is where the book lost me. Unlike some readers, who are fascinated with “decoding” the book and delving deeper into its various “secrets,” I just got bored when the narrative ended. While I can appreciate the denial of narrative as a technique, I still didn’t enjoy the last 200 pages of the book. But that’s not true of everyone. I also have to say that the saccharine ending of The Navidson Record was also disappointing. But, like much with this novel, I can’t take for granted that this letdown of an ending is not structured as such.

Despite its unevenness, I’d definitely recommend House of Leaves to others as a (mostly) superb feat of storytelling as well as an exceedingly clever meditation on the nature of narrative. But at the same time I would also remind people that testing the limits of narrative convention can be alienating and occasionally tedious.