Friday, April 11, 2008

A Good Argument (2): generalization

One thing that endlessly frustrates me in arguments is the use of quantifiers and appropriate generalization.

Generalization is necessary. We can't think or operate without it.
If we couldn't generalize, we couldn't function. We couldn't recognize food or people or friends. We generalize from specifics to future situations--this is something like the sense in which the pragmatists say that knowledge is what works--we eat one tomato, and we generalize that experience to other tomatoes, allowing us to eat them.

But generalizations can become problematic when used carelessly. Obviously this is true with the type of generalization known as a stereotype. But similar generalizations pervade scientific thought and are both necessary and perilous.
We want to be able to make generalizations that can predict (and thus guide) behavior; but situations are not always similar--and who is to say how a minor factor affects predictability?

We, therefore, ought to be careful with our generalizations and our use of quantifiers: there is a huge difference between claiming "All members of a class have property X" and claiming that "Members of the class are commonly observed to have property X."

Or, to use a well-used example: we may observe many white swans without observing any of other colors, but this does not mean that we know that all swans are white. Saying "All swans are white" is but waiting for one example to be shown wrong. Saying "many swans are white" cannot be disputed if you have observed these swans.

A good argument understands and utilizes the proper level of generalization.

No comments: