Friday, November 19, 2010

The McGurk Effect

I was reading and watching some of Jonah Lehrer's oeuvre, and found him a bright, engaging speaker in the mold of Malcom Gladwell. He presents research anecdotes like the one I blogged on below, and then suggests this has some profound implication. In some sense it is the Freakonomics method: find a 'funky' result studied by genius professors (did you know Steve Levitt is widely considered the greatest young economist?), and note that it is totally outside the conventional paradigm.

For example: Most people think they are above average drivers. They can not all be. They still all think so (me too!). This is irrational. Models built on rational expectations are wrong. This disproves the Black-Scholes model that caused the mortgage crisis. QED.

I enjoy these stories as much as anyone, but notice that when they go into my domain they are 100x less interesting. Neural science, dark matter, animal phsychology all seem fine. When an expert or expertise seems really powerful everywhere but in the subject you know well, you should be skeptical.

For example, here is the McGurk effect, and it is rather fascinating. Basically, if you see someone saying "fa" but they pronounce "ba", you will hear "fa". Your sight affects your hearing.

Now, a skilled writer could probably present this as having some really important implication to finance or politics. It's interesting, true, slightly surprising because it's new to me, but not important. Important, new, and true are the hallmarks of Good Ideas. They are as rare as a good melody. One must watch out for authors who present new, important, and true ideas, but never at the same time applied to one idea. It's like the empty suit strategy of having strong opinions only on irrelevancies like sports or fashion, but then never taking a risk on real matters at work. People around you often won't think of you as indecisive, because you are such a strong Lakers fan, and you don't risk actually being wrong on something important to your business.


  1. Anonymous2:23 PM

    "When an expert or expertise seems really powerful everywhere but in the subject you know well, you should be skeptical."

    For some reason this made me think of Consumer Reports. Everyone seems to think that Consumer Reports is great except for X, where X is something they know about. Nobody reasons, "if they got X wrong, they're probably wrong about much else too".

  2. Anonymous8:09 AM

    Consumer Reports is a good starting point when you have no prior knowledge or expertise about a product line. If you're already an expert, of course it doesn't offer any marginal value.