[ceceron]

> It's all a matter of how wrong you want to be, and what information you want to use for deduction. If you see a humanoid approaching you, do you start with "that is a human (99.99% certainty)" or "that is a man or woman (98% certainty)"? I can see the use in the latter, but only here when quantified and compared to more generic alternatives.

Generalization is one of the main mechanisms making the communication (language) work. Otherwise one should probably treat every human as an individual, similarly to the Duns Scotus and his haecceity concept.

It's somehow related to the concept of overfitting, you acknowledge existence of the outliers but at the same time you may want to generalize them for the sake of the better generalization...

> On personal attacks: I've only lost karma in this thread and I'm going to continue losing it; I'm not sure by what means you could imagine that I'm effectively winning.

Who cares about karma. What's important is that your comments may encourage some readers to think ;)