TROUBLE WITH CONSTRUCTIONS

The previous section, starting with the supposition that the evidence was fairly well in hand, was concerned with the problem of knowing what could reasonably be asserted about it. There was no question as to what the situation involved, and it was assumed that the classifications and general notions were serviceable. All cows are brown -- the problem was the handling of all, rather than cows or brown. Does the evidence warrant saying all or only some? The notions cows and brown were assumed to be clear and intelligible. In this section the interest centers on the formation of the notions themselves, rather than how they are related together in propositions. The most difficult problems in managing the materials of argument arise in the forming of notions and classifications and the like. We shall use the term constructions to cover all notions and abstractions. All abstract thinking proceeds by selecting aspects of the situation and making constructions out of them. Some constructions are relatively easy to understand -- cows, for instance -- a three-year old can form the notion of a cow. The names such simple constructions bear are well known terms in the common speech. Every user of the language can learn the proper construction and hence the correct use of the name. The class name "child" represents a construction made up from certain things in the world, children, markedly distinguished from other things: babies, adults, and, for that matter, elephants. The construction reflects the situation in the world in a rather obvious way. A speaker, if pressed, can give at least rough conditions for a proper use of the term.

	------->		----------->
class name	means	construction	stands for	for situation

As the class name is often a familiar term in the common language and the situation is the state of affairs in whatever aspect of the world is under discussion, the only subjective part is the construction, though, perhaps, since the speakers communicate the construction, this part should be called "inter-subjective."

All constructions involve at least implicit classification -- babies are grouped under the heading "babies" and children under "children." Now speakers are constantly making new classifications. For this reason some constructions do not have a traditional name in the language, and a name must be made up for them. It may be a new word or a whole new phrase made up of old words, but whatever it is, it often names a construction that stands for characteristics of the situation not previously remarked. Since all writing and speaking of point proceeds by drawing attention to fresh groupings of things, it is evident that anybody who wants to make a contribution in discussion must understand the elements of classifying if his constructions are to be intelligible.

Suppose a boy gets from his uncle a bag containing a hundred marbles. And suppose this boy has a little sister who wants to play with the marbles. The little girl begins to sort them. First she arranges them by color, green with green, red with red. Then she sorts them within the colors, clear glass with clear, cloudy with cloudy. There are some hard to classify: opaque agates, little dull clay balls, marbles of mixed colors. She has disappointments. Though she is very fond of purple, there are no purple marbles. And some of the prettiest blues have little chips and flaws.

The boy says his sister is silly. This is the way marbles should be sorted. He first pretends to shoot with each one, then sets aside one of the "aggies" and two of the chipped "migs" as "shooters." The rest he arranges in several orders, regardless of color or marbling. The least mig, he tells his sister, is worth ten of the "dough babies," but the best migs are worth ten of the worst, and the aggies are worth at least ten of the best migs. A good shooter is worth the lot; in fact, it is beyond price, but he won't know if there are any good shooters until he has tried some out in competition. His sister doesn't understand anything he is talking about and decides it is really boys who are silly.

Here are only a hundred marbles. Yet it is at once evident that there are a great number of ways to classify them. If they are sorted as to color, there will be empty classes, purple for instance. If "cloudiness" (marbling) is the criterion, then the clay marbles present difficulty: they are not marbled, but they are not transparent, either. If cost is to be the consideration, then there are two obvious ways to put them in order: monetary value, for agates are relatively expensive, or trade value, which seems to depend on the fashion or taste of certain boys in the neighborhood -- one marble "costs" ten others. And so on, indefinitely.

The situation is present. It would be a truly "silly" error to assume, because the marbles are classifiable as to color, that there must be somewhere a purple marble. And maybe there will not turn out to be a first-class shooter in the bag. So the situation puts a limit to the number and kinds of constructions. And the names ("shooter," "pretty ones") or the descriptions ("worth ten of") must mean the constructions. This is all we intend by these terms, and there is no metaphysical commitment in the notion of the "situation."

Even in the simple cases of classification, there can be trouble. It will be seen that difficulty can arise over where to put marginal cases. When this happens, the classification may carelessly omit them altogether -- they get left out. Sometimes the classification seems to require putting the same object under two coordinate headings. In this event, the classification does not function in respect to that object, for the general function of a classification is to separate.

From these considerations, since the time of the early Greeks certain criteria of classification have been enforced. They can be expressed simply, and we shall try to state them in a set of rules. The fallacy of violating any of the rules falls under the general head of "faulty classification."

Rule 1. The classification must be exhaustive. (The dough babies must not be left out, simply because they don't fit the division of clear or cloudy.)

Rule 2. The classification must be exclusive. (Cross-classifications are permissible -- cloudy-red marbles -- but there must be what is sometimes called a fundamentum divisionis, which prevents putting something under two coordinate headings.)

Rule 3. The classification must be adequate to the purpose for which it is designed. (The little girl's classification of the marbles into colors, etc., though perhaps suitable to the purpose of the little girl, is not adequate to the game of migs.)

Rule 4. The divisions of the classification must be precise enough to avoid serious marginal cases. (If boys don't usually agree on, say, which marbles are shooters, then this construction is vague.)