Friday, 20 December 2019

Review: Agency

Agency Agency by William Gibson
My rating: 3 of 5 stars

I never thought I would use the phrase "tedious William Gibson novel," but apparently this is the version of the world we now live in.

This tedious William Gibson novel is clearly a William Gibson novel: it has the effortless prose, the vivid (if occasionally inaccurate) imagery, the geek-culture namedrops, the characters who are outsiders to power and the mainstream. What it doesn't have much of is a plot, and what the characters don't have much of, by irony that may or may not be unconscious, is agency. They do almost nothing that has any impact on anything. In fact, they do almost nothing, and it's narrated at great length.

The author establishes a strict alternation between two viewpoint characters: Verity, in a version of 2017 California where the US election of 2016 and the Brexit vote went the other way, and Netherton, in a post-apocalyptic future descended, quite possibly, from our version of the timeline. This strict alternation regardless of what's going on and who has the most at stake at the time, and this choice of viewpoint characters, soon begin to work against the success of the book.

Both viewpoint characters are essentially passive. Verity spends most of the book as a passenger, being moved around to escape from a corrupt corporation who hired her at the start of the novel. It's never really clear to me why anyone involves Netherton in events, rather than just going direct; he's a go-between and a middleman and an observer, and the one effective thing he does (fighting off a random encounter that has no lead-up and no follow-through) is entirely by accident. Many of the chapters, particularly the Netherton ones, consist of someone, usually Netherton, repeating something we have just been shown in the previous chapter to someone else who wasn't observing at the time. (Not very far into the book, the two viewpoints connect, by a technological means of communication that's never explained in any depth, but looks to the users like VR.)

The beginning is promising. Verity is hired by that dodgy corporation because of her reputation as an "app whisperer" to do vague things with a new alpha build, an advanced AI called Eunice. Eunice is templated on a feisty, fiercely intelligent and capable African-American woman, and is by far the most interesting character in the book; after (view spoiler) (relatively early on), the novel immediately bogs down in exposition, pipe-laying, long descriptions of logistics (she sat here, she put her bag there, she looked at this), and people explaining things to other people that we've just been shown in the previous chapter. Once (view spoiler), the book wraps up rapidly, but without much involvement of the carefully-gathered group of people who are supposedly the protagonists; they have spend all their time while the world was threatened with nuclear disaster doing mostly mundane or evil-corporation-avoidance-related things, rather than working on anything to do with the threat, and (view spoiler).

I got the feeling partway through that the excessive number of secondary characters with backstories that didn't seem relevant to the current story were left over from a previous novel, and indeed it seems this is a sequel to The Peripheral. I was surprised to discover, looking at the front of the book where they are listed, that I'd missed three novels by Gibson since the last one I read, so I don't know if the mediocre dullness of this one is a new development or part of a trend. Since I got a pre-release version from Netgalley, I also don't know if the couple of glitches (such as placing jungle-dwelling orangutans on the savanna) will be fixed before publication; they may be. What I don't think can be fixed is the overnarration of mundane logistics that stands where a plot would normally go, or the limp and ineffectual puppets that are the viewpoint characters.

Accordingly, I'm awarding this my non-prize for Most Disappointing Novel Read in 2019.

View all my reviews

No comments: