Bracketing the target

2011-01-09T13:08:25Z

If you're firing a howitzer, get two shells on either side of the target, and then you can quickly home in on it using a simple (often trivial) binary chop. This may actually be better than two shells very very close but on the same side.

The same often applies to arguments: a totally crazy but totally different idea may be more useful than a cautious but unassailable refinement of the previous suggestion if the best answer is contained somewhere between. For instance:

(a) This is what people mean (or should mean) but "there are no stupid suggestions" -- even a ridiculous suggestion can contain a nugget of an undeveloped good idea or useful constraint. (Not always, but often.)

(b) If you're trying to get someone to grok something, you may say "it's a bit like [this film] and a bit like [this film]" or "it's a bit like a wave and a bit like a particle" or "it's a bit like a religion and a bit like a culture" then even if both examples are totally and utterly false, they may well give people a good intuitive idea of the domain of answers within which the correct one lies.

(c) if you come up with a long, complex philosophical argument, spend five minutes saying "would this be convincing to an intelligent person who doesn't know anything about philosophy, or would they say 'well, I can't tell you exactly where the flaw is, but I'm pretty sure it's false because here's a counterexample'"

comments

jack

Bracketing the target