## August 31, 2013

### Books to Read While the Algae Grow in Your Fur, August 2013

Chelsea Cain, Kill You Twice and Let Me Go
Mind candy. The series is steadily ratcheting up its sheer gothic weirdness, and I'm not sure I can recommend these for new-comers, but they were fun.
Erin Morgenstern, The Night Circus
Mind candy: lovely literary fantasy about an enchanted traveling carnival, and the wizard's-duel/love-affair of two apprentice magicians. (Thanks to RCS for a copy.)
Daniel Schwartz, Travelling through the Eye of History
Beautiful contemporary photographs of Central Asia; the text strives for profoundity but only achieves pretentiousness.
Jessica Hagy, How to Be Interesting (In 10 Simple Steps)
Inspirational life advice. I think it's good advice for people who are in a position to follow it; too many people aren't.
Linda Nagata, The Red: First Light
Mind candy: cynical, intelligent and gripping near-future military hard SF. Stefan Raets's review at tor.com pre-empts most of what I wanted to say.
Tracy Thompson, The New Mind of the South
Journalistic impressions from a native daughter; despite the title, not really an attempt to update Cash. It's well-written (for journalism) and seems very plausible, but I lack the knowledge to say if it's right. (She doesn't seem to do enough comparison to the rest of the US, though, like the Midwest or the mountain West, when discussing the hollowing-out of rural areas.)
Nilanjana Roy, The Wildings
Valuable, but I did want to argue with it a lot. In particular, I don't understand his objection to the causal Markov condition, and so to what he calls the approach of "Glymour et al." Saying it won't hold for some of the official economic statistics because they're measured too slowly confuses, by his own principles, the causal structure of the macroeconomy with the econometric issue of discovering that structure. Hoover's own positive principle for determining the direction of causation is that if $X$ causes $Y$, interventions which change the distribution of $X$ shouldn't alter the conditional distribution of $Y$ given $X$. If the causality goes the other way, however, interventions which alter the distribution of $X$ will also change $Y|X$. This is sound, but also easily explicated within the graphical-model tradition. (It's how we calculate effects of interventions from "surgery" on graphs.)