Another outcome of the naïve optimistic view, as realizing the complexity of scientific specialization, is the doctrine that a simple, practical system may be kept abreast of scientific progress by the addition of new details. This elaboration of schedules is compatible with what we term "expansion." Expansibility is essential to the very life of a notation, but it may be overworked. Certain systems have, I fear, expanded beyond the capacity of their safety valves to save them from explosion. Thousands of the details of those inflated schedules are practically useless even in the largest library. Such abnormal distension of the bibliographical body, or hypertrophy of its special parts, is not now for the first time called a disease of the bibliothecal system. That the subjects and topics are innumerable and of intricate complexity has led to the misconception that a classification for libraries should embody an infinity of captions in infinite complication. An alphabetical subject-index is believed to be all that is requisite to operate this maze of entangled details. This view may be termed the subject-index illusion.
Classification for libraries is to be distinguished on the one hand from notation and on the other hand from an arrangement of bibliographical subjects indexed. Notation and index are but correlative to classification, and, however requisite to a practical system, are in truth of minor importance. They are the fingers and the feet of the body and brain that organize the materials of knowledge. Yet it is these fingers and feet that have chiefly occupied the attention of most classifiers.
In the theory of classification subjects are to be distinguished from classes as contents from containers. The subject is that which is denoted by its definition; the class is the aggregate of particular things—books, or other things—that are comprised by the definition. A class may be comprehensive of many subjects or aspects of subjects. Such need not appear in the schedules of the classification, but they should be in its subject-index. Thus, Botany is a subject, to which Botanical Books is the corresponding class; Plant Physiology, a less general subject, has a less comprehensive class of books. Geotropism is a specific subject in the physiology of plants. The question arises, is there a class of books and pamphlets treating especially of this subject, the tendency of plants to respond to gravitation, as a stimulus? "Have you in your library," I might ask individually of the majority, "have you an aggregation of books on this subject?" The A. L. A. List comes nearest in the sub-headings under Plants, where with Movements appears Heliotropism, a kindred subject. This caption Movements is for a veritable class of subjects, and it might indeed comprise Geotropism. That is just what the Library of Congress schedule does, subordinating under QK 771 "Movements, Irritability in plants, (general)", the caption of 776, "Miscellaneous induced movements: Geotropism, Heliotropism, etc." In my own classification, the mark GCM goes with the caption, "Movements, Heliotropism, Geotropism, etc." It seems well thus to provide for a future group of monographs. If I criticise the Library of Congress classification today, or elsewhere, be it remembered that I recognize its correct treatment of this and thousands of other subjects. But is the E. C. justified in reaching into the dim future for subdivisions of specialization such as its NESGD, Diatropism, and NESGL, "Lateral Geotropism?" That is where we must open the safety valve or burst.
The body of the D. C. is congested with thousands of names of persons, places, and events which may be subjects, but hardly for classes of books. Systematic schedules might provide for most of these, reduce the bulk of the system, and make for economy and convenience. The L. C. schedules suffer from similar but more astounding expansion. Class H, Sociology and Economics, is needlessly immense, having 551 p., of which but 51 are index. According to the principle laid down a moment ago, the number of subjects in the index should by much exceed those in the schedules.
The "Expansive" Seventh expansion expanded so much with its own specialistic tissue that it could afford to omit such bulk of proper and place names. For instance Aves (Birds), covers 8 pages of fine print; there are all the taxonomic terms, for example, PGSLPI is for Phalacrocoracidæ, some family related to the pelicans; but there appears besides only the single subject Oology (eggs), at the end as PGZ. No place under Birds for their structure, their habits, for the popular bird-books, and for such interesting subjects as their migration, flight, etc., about which there are books! However much there is to interest, to commend, and to admire in this great undertaking, it must be admitted that this is not practical classification for libraries. It is the province of the subject-catalog to bring together topics and titles which are too special for classification to bring into collocation.
But let us return to the main question of the feasibility of better classification. There are three answers, I said. Two we have considered, the naïve, and the pessimistic, also their offspring, the subject-index illusion, but we have not yet completely answered the pessimistic. This we may now proceed to do in connection with the third answer, which is optimistic and constructive, while at the same time critical. This affirms that better classification is feasible, that it may be sufficiently flexible and durable, that changes and adjustments may be provided for in alternative and reserved locations, that the notation may be quite simple, and that the index may be as full and specific as comports with convenience.
The purpose of library classification is to group books and to collocate groups for the convenience of readers and students in their average wants. It is not so much for those who want a book, whose author and subject are known, or any good book on a particular subject; for such, the author and subject-catalogs may suffice. But classification is for those who want books, in the plural, directly, without preliminary handling of cards. Three types of such wants are to be distinguished.
(1) To all libraries come (the prevalent type) those who wish a few good books on the subject, or a few facts to be found in the standard books. They do not care to fuss over the card-catalog. The reference librarian, the selective lists, may serve such wants, but close classification usually does so most economically and most satisfactorily. For very specific subjects, however, the subject-catalog in the large library may often best serve this type and may make it less dependent upon free access and close classification.
(2) The second type wants all the good books treating of the subject especially. From these the user himself is to make selection according to his purpose or point of view. Free access and classification are here requisite. A bibliography, if there be one, would be most likely an embarras de richesse.
(3) The third type is that of exhaustive research: all the available literature is wanted, not only the books and pamphlets treating especially of the subject, but also those on related subjects and those of broader scope. Subject-catalogs and bibliographies are needed preliminaries, but access, continued access to the books, is the desideratum. It is for this type that the most carefully guarded libraries give access to their precious collections. Classification, not merely any old kind of subject, or close classification, but good, scientific, close classification, based upon good, consistent, broad classification, is here of paramount importance. The test comes when the student turns from the special to the more general and the related subjects, which are mostly in related branches of science. The tendency to organization in science is rapidly and surely growing. The more consistent with the consensus, to which studies on the average are adapted, however original and divergent their aim, the more convenient will be the classification. It is in subordination of the specific to the broader subject or class and in collocation of related subjects and subdivisions of classes that most systems fail; and here that most classifiers fail to understand either the fault or the remedy.