No, not all snakes are poisonous. But if you see a snake move in the grass you better presume it might be poisonous and back off.

Now maybe it was just a King snake. But unless you know for sure  you better presume it might be poisonous. Of course the inference that this snake is poisonous because some snakes are poisonous is fallacious. But is that what is going on profiling? The inference is that it might be. And it is the 'might' that is the weasel factor. The degree of probability is all important as is the degree of danger. The profiling that is involved here is a kind of generalization: objects with characteristics XYX are presumed, to a certain degree of probability, to also have characteristics UVW. The presumption is based on what we know or think we know about the probability relationship.

Walter Williams, the noted economist and political analyst, characterizes profiling as follows, "We can think of profiling as a method to economize on information costs by using easily observed physical characteristics as a proxy or a guess for some other characteristic more difficult or costlier to observe."

Of course we can be done in by our profile invocations. After all, if you were standing next to Dzhokhar Tsarnaev, while watching the Boston Marathon, you  would probably profile him as not being dangerous since he seems well mannered and to fit in with the rest of the crowd.

In order to get things done in a reasonably timely manner we need to make snap decisions based on invoking what we consider to be the appropriate profile of the situation. If I am walking into Kroger out of the parking lot and a car slows down for me [XYZ], I invoke the profile of a considerate driver [UVW] who will let me pass. If  I am turning left onto a highway and see a car coming from the left with it's right  blinker on and seemingly slowing down, I start to pull out onto the highway if all else is clear. I have invoked the profile of competent driver who is giving me indication of his future actions.

We make thousands of such decisions every day. We don't have time to check out our presumptions. From a survival point of view it is better to err on the side of caution but even then, given time concerns, we continually take our chances. We have to profile. There is no need to council us to live dangerously, we have no alternative.

With the car example,  as indicated, I have, in effect, invoked the profile of a competent driver who is aware that his right blinker is on, presumed that the driver is not deliberately out to deceive me, that his car is functioning properly, that my car won't stall ... .

Holding on to one's profile of what to expect in an assessment of a given situation, which involves a host of presumptions about associated characteristics, in the face of increasing evidence that such associations are relatively mistaken, is irrational and possibly fatal. But making the initial decision in time critical situations is a Darwinian necessity.

If one is constantly being judged as probably having characteristics UVW because of having characteristics XYZ when in fact one does not, then one has a right to be irritated. But such is life. Dumb blonds? "Creepy assed crackers"? Not only may one be the exception to the general rule, the general rule invoked may be cockeyed. Certainly racial or ethnic profiling may be statistically wrong, as can any other profiling. But we all make snap decisions on the basis of what we know or think we know about associated characteristics, however unfair that may be in any individual case. Till we have perfect knowledge that means that Obama's black-man-in-the-elevator is going to hear the snap [sic] of your purse.

As Walter Williams concludes his article on the practical necessity of profiling, "God, unlike us mortals, wouldn't have to do any kind of profiling, because he knows everything. We mortals, with our imperfections, must find substitutes for his omniscience."