A Thought, Part 2

As I said in the last post, a classification system (library or otherwise) can be thought of as a sort of map or guide to a certain field of knowledge or a certain group of items. It helps map out the territory, which then helps guide you to the specific item or place or bit of knowledge that you seek. So, if a library classificationist is the one who writes the guidebook to all human knowledge, this assumes that he knows where various fields are in relation to each other, which further assumes that he knows at least something about all of these fields of knowledge. But how can anyone really know how every subject relates to every other subject, if there even is an objective answer to such a question? I said last time that all classification systems are guesses. It might be more appropriate to say that all classification systems reflect the preferences (intellectual or aesthetic) and the practical needs of whoever designs the system. For example, the old edition of Roget's Thesaurus of Words and Phrases that I've been using divides all words (which is to say, ideas and/or things) into six classes:

I. Abstract Relations

II. Space

III. Matter

IV. Intellect

V. Volition

VI. Affections

The whole system descends in a hierarchy, from these six main classes to sub-classes and sub-sub-classes and so on, till you get down to the most specific ideas. This is all fine and dandy, and it gives the whole realm of knowledge a comforting sense of order and structure. However, in a more recent edition, called Roget's International Thesaurus, the categories are not nearly as hierarchical. It is divided into 15 main classes, but within each class, the order of sub-classes is what librarians call "enumerative" (i.e., a listing), rather than hierarchical (which can be thought of as a family tree). Is this an indication that Roget's (and, by extension, the rest of us) have given up on the idea of structuring human knowledge into some kind of logical order? The enumerative list is just that... a list. It doesn't attempt to understand or explain how ideas are related to each other, except in the simplest sense of grouping similar concepts together and putting similar groups under a general category. Worst of all, however, is the alphabetical style of thesaurus, which simply lists words by how they happen to be spelled. This can be very useful, of course, but that's the whole point... it is merely practical, without making any attempt whatsoever to classify our ideas and therefore to provide a sort of road map to all knowledge. Whereas an old-style thesaurus can be used in a similar way to a library classification system (except that it is a guide to words instead of books), the modern dictionary-style thesaurus serves no such purpose, leaving one just as lost amid the sea of words and ideas and things as without it.

One reason the classification project has held such allure for me is that, at some level, I have seen it as a way of locating things in the world. I mean "locating" both in the sense of "finding" and in the sense of "placing", as in "let's locate it here". And, although I mean "finding" in the sense of discovering something new or recovering something lost, I mean "finding" even more in the sense of discovering where something is in relation to everything else. Instead of something being lost in the flood of ideas, images, and experiences that we encounter each day in our 21st century lives, or floating adrift in the vast sea of knowledge, a classification system is a way of "locating" or "finding" its place in the world. Perhaps, ultimately, it's a way of making the world more like home, by organizing the random, chaotic clutter of our experiences and knowledge into a meaningful and beautiful order.

In my previous post I mentioned an uncertainty principle (specialization vs. generalization). What I have been talking about here leads me to what I might call an incompleteness theorem of classification. It goes something like this: In order to know how to organize knowledge and experience, one must already remember everything that needs to be organized. If you forget something, then that thing gets left out of the system, and the system is therefore incomplete. But in order to find everything that needs to be included in the system, we first need a system; without a system, we are left to the whims of memory, which is always incomplete and leaves things out. We can never be sure that we have included everything that needs to be included. In other words, we can't make a complete system without already having a complete system. Therefore, all classification systems are incomplete, and in fact it is impossible to create a complete classification system.