pose. A classification such as that contained in the following passage, for example, has little of the exactness which a strictly logical division would demand; yet for the purpose the writer had in view, it is exceedingly useful:

The human element in production, whether in the work of guidance or in obedience, varies as widely as human nature and capacity. Tot homines, tot capacitates. For services to production, laborers may be roughly classified by strata, as in the accompanying diagram:

The unskilled men in A, the slightly skilled in B, the highly skilled artisans in C (such as the locomotive engi

E

D

с

B

A

neers), the highly educated professional men in D (such as civil engineers, electrical experts, and the like), and finally the exceptionally capable managers in E. In any one industry some of each kind are required, but not with the same intensity of demand; nor are they wanted in the same relative numbers in different industries.

The unskilled man in A has no choice of occupations that he can enter; he can do only the work demanded of his class. And yet, as compared with the demand for them, the number in this strata is enormously large. Moreover, in the A class there is the least capacity to set the future gain above the present indulgence. Thus we find increasing numbers in the very group whose activity is restricted to a given kind of work. Among those least competent to add

to production, there is the greatest supply relatively to the demand for them. Their share is small, not only because their industrial efficiency is small, but because the supply of them is excessive.

As we go up in the scale of industrial efficiency, we find the numbers in the strata of the more highly skilled diminishing, while the intensity of the demand for them increases. Hence wages increase the higher we go. In the top strata, containing the most efficient managers, we find the highest wages paid throughout the whole industrial field. When a blundering or incompetent manager costs a company millions in losses, a fifty-thousand-dollar man, who adds millions in gains, is a cheap laborer. In this struggle up the scale from A to E we find the real social conflict. It is a contest between different kinds of laborers,-a contest of varying grades of industrial capacity with each other. It is a free-to-all race, in which the most competent win. The great industrial manager, being the most highly skilled laborer, obtains enormous wages for exceptional services to production. This exposition gives us, in brief, the economic reason why, in a country of phenomenal resources like the United States, men of exceptional industrial ability can acquire exceptionally large fortunes legitimately.1

Expository classification is thus, as we see, something more than mere logical division. Nevertheless, a good classification always conforms to the rules which govern logical division. These rules are: (1) There must be only one principle of division; (2) There must be no overlapping in the classes or species; (3) The sum of the species must equal the genus or class divided.

1J. Lawrence Laughlin, Atlantic Monthly, July, 1905.

The first of these rules is especially important, inasmuch as a change in the principle of division absolutely vitiates a classification. Thus if we were to classify the people of Asia as Mongolians, Malayans, Hindus, and Mohammedans, we should be proceeding first on the principle of racial character, and then on that of religion. There would therefore be no unity or meaning in our classification. It would be like filing letters at haphazard, now according to the writers' names, and again according to some other principle. The rule against overlapping or cross division, as it is sometimes called, is also important, since, if the classes are not distinct, the classification loses much, if not all, its value. To classify the writings which make up literature, for example, as poetry, fiction, history, biography, and the essay, would be objectionable, because the classes" poetry "and" fiction" would not be mutually exclusive; a romance such as Scott's Lady of the Lake, for example, might be put in either. As to the question of exhaustiveness, the rule is imperative only where exactness and precision are of first importance. In ordinary literary exposition it is seldom insisted on. A classification may have great practical value even if it be not exhaustive. In fact, no classification can be held to be absolutely exhaustive, logically speaking, unless it is made according to the method known as dichotomy, where there are only two divisions, one of which is expressly stated to contain all the individuals of the genus not included in the other. For example, the citizens of the United States may be classed as those who have the right to vote, and those

who have not. The classification is absolutely exhaustive, since every citizen must belong to either the one or the other of these two classes.

It should be noted that classification always implies definition. The two, indeed, very commonly go together. Whenever we make a classification, we do so on the supposition that the various classes can be defined. Exposition by classification is generally, in fact, first a division and then a definition of each member of the division. Notice how, in the following, Huxley first divides the science of zoology into its various subordinate sciences, and then defines each one of these in turn:

I shall use the term zoology as denoting the whole doctrine of animal life, in contradistinction to botany, which signifies the whole doctrine of vegetable life.

Employed in this sense, zoology, like botany, is divisible into three great but subordinate sciences,-morphology, physiology, and distribution, each of which may, to a very great extent, be studied ir dependently of the other.

Zoological morphology is the doctrine of animal form or structure. Anatomy is one of its branches; development is another; while classification is the expression of the relations which different animals bear to one another, in respect to their anatomy and their development.

Zoological distribution is the study of animals in relation to the terrestrial conditions which obtain now, or have obtained at any previous epoch of the earth's history.

Zoological physiology, lastly, is the doctrine of the functions or actions of animals. It regards animal bodies as machines impelled by certain forces, and performing an amount of work which can be expressed in terms of the ordinary forces of nature. The final object of physiology

is to deduct the facts of morphology on the one hand, and those of distribution on the other, from the laws of the molecular forces of matter.1

4. DESCRIPTIVE AND ILLUSTRATIVE EXPOSITION

As has been said, the typical moods of exposition are definition and classification. Exposition, however, is often used for various explanatory or illustrative processes not formal or exact enough to be brought under the category of definition or of classification. Among such uses may be mentioned, (1) the description of a type-form or the setting forth of the general characteristics of a class of things; (2) the explanation of a method or process; (3) the illustration of a general law or the application of it to particular cases.

In the first of these uses we have a kind of writing which suggests description, on the one hand, and definition, on the other. In some cases it verges so closely upon description that it can scarcely be distinguished from that form; it always deals with the type or with a class of things, however, whereas description properly deals with the individual rather than with the type. In trying to set forth the general characteristics of a type-form or of a class of things, descriptive exposition, to be sure, is trying to do very much the same thing as is done in definition; but its method is usually more suggestive of description than of definition.

An illustration or two will make the point clear,

1 See Huxley's Lay Sermons.