Um...: Slips, Stumbles, and Verbal Blunders, and What They Mean by Michael Erard. (Amazon, B&N, Powell's) This was a neat one. It's got everything from Spoonerisms to Freudian slips to Bushisms, lots of other verbal flubs, the theories behind them, the Toastmasters, Cicero, and plenty of other reasons to make you pay more attention to your speech.

But note that it doesn't necessarily make you want to un-um (de-um?) your speech. Umlessness wasn't especially prized as the ideal for speakers until the advent of recorded sound, when we could finally hear our own voices, and how full our speech was of ums and ers and sentence repairs and other errors. The Toastmasters aside, to "er" is human. Erard makes the point that pause fillers and self-corrections can actually aid communication as much as it might hinder it.

It can get a little annoying if you start catching all your own blunders, and maybe even more if you start noticing everyone else's. But it can also be fun to turn that on and off if you're listening to someone boring.

I'm not saying any of these were boring, but just since I began Um..., I noticed several blog posts highlighting speech errors. Here's Prof. Volokh on "severe or persuasive harassment" (instead of "pervasive"), Think Progress on White House transcripts correcting President Bush's grammar (there's stuff in Um... about this), and my favorite, The Slog catching Barack Obama saying "The time for come has change."

I'd recommend Um... if you're a language geek. I thought it might help me be a better speaker, but I'm not sure. It's more about why we "um" than how to stop it," but that interesting in itself. Overall, I'd say, um, not bad.

(Previous 50 Book Challenge reviews)

I saw this interesting item from Wired. A scientist has developed a DNA test that will reveal the race of the donor. The technology has been used to help solve several crimes, or at least narrow the range of suspects. Yet lots of people, including police and prosecutors, feel sort of squeamish about it. A prosecutor from Louisiana who acknowledges that without this DNA test the police likely would not have caught a serial killer he tried still says, "If I could push a button and make this technology disappear, I would."

I don't get this at all. How is this different than an eyewitness telling police that the suspect is a member of a certain race? Oh yeah -- it's more reliable. I suppose that one day there might be enough genetic mixing among the races (a situation that might exist now in some places) for the margin of error to grow too large for use as a forensic device. But assuming this test is reliable (the developer got 20 out of 20 right in a blind test), I don't think it amounts to "profiling."

It's not like we're saying that members of a certain race are more likely to have committed a crime; we're saying that this particular person committed the crime, and among other characteristics we can tell from a DNA sample (sex, blood type, etc.), the suspect is probably a member of a certain race. Why isn't it a good idea to have that additional information? Or am I missing something?