Friday, April 27, 2012

Categories and Wastebaskets

Nelson Goodman's Problems and Projects (Indianapolis: The Bobbs Merrill Comapany, Inc., 1972) contains a chapter entitles "Snowflakes and Wastebaskets". It is about the notion of "categories" in Immanuel Kant and C. I. Lewis. Since Goodman believes that Lewis went "as far beyond Kant" as Kant went beyond his predecessor, he prefers Lewis's understanding of categories. In this context, he relates the anecdote that Lewis once said to his class that anyone who really succeeds in deducing the categories will make a "lasting name" for himself only to add after a pause: "but it cannot be done".

Kant thought that "categories" are the fundamental synthetic functions of thinking, that is, the necessary conditions of the possibility of thinking at all. without categories, we could not think at all. In particular, we could not think in terms of objects. Lewis rejected this view. For him, categories were just tools we use to sort out stuff that "comes" to us. They represent the order we impose on things. They represent our "filing system" and any kind of filing will do.

Goodman illustrates this point by telling another story. "A poet once had a filing system consisting of four folders severally marked: 'Unpaid Bills', 'Rejection slips', 'Love Letters', and 'Miscellaneous.'" As he became successful he made a slight improvement, renaming the miscellaneous folder "Acceptances' and throwing everything else into the wastebasket. He claims that as long as we have either a miscellaneous category or a waste basket, we will never be at a loss in filing or categorizing things. "That whatever we encounter will fit our scheme depends upon no assumption about what we shall encounter but only upon reasonable care in devising our scheme—especially by providing a wastebasket." Such a system "guarantees limited variety" and "insures against unending novelty." We can also be sure some repetition.

By fiddling with the categories we can make some recurrences more probable. But the reliability of the system decreases with the number of categories. Maximum reliability is achieved by having the waste basket as the only category. But reliability alone is not enough, so our filing systems tend to betray a tension between safety and the "need for specificity" we may have.

The important point for him is, however, that the regularity of the world does depend upon the arbitrary choices we make in categorizing the world. "Reality must be regular because reality is distinguished by the very fact that it conforms to the requirements of the non-wastebasket compartments of our categorical scheme." Chaos is impossible as long as we have some categories.

But absence of chaos is not really what we are after. We clearly want more. Therefore the choice of our non-wastebasket categories matters very much. The trick is to find those that have the right sort of specificity that allows not only for novelty and regularity, but also for revision.

This is why all note-taking systems need categories—any categories and why a "miscellanea" can never be sufficient. As Goodman said himself in other contexts, the "right inductive categories" are of paramount importance.

None of this, pace Goodman and Lewis, has anything to do with Kant. His "categories" are not meant to be inductive (and are thus of an entirely different nature as those of Lewis and Goodman). Nor are they directly relevant for note-taking. But that is a story that does not belong here.


Michael Leddy said...

That's a great story about the poet.

I don't know this book but I've found Languages of Art and Ways of Worldmaking really thought-provoking and useful.

Just so you know: in Google Reader, this post came out as one long paragraph. Much easier to read here. :)

MK said...

The book is really a collection of different things, some of them rather technical.

It may be that I did not put the paragraph mark "< / p >" in the first version.

Don't know why blogger insists on this nowadays.