These four classes of libraries, memory libraries, pictorial object libraries, “mnemonic” libraries, and picture book libraries, form thus the field. All of them existed before what may be called historical libraries; all are found among uncivilized peoples of all times; all have their faint remainders in popular custom among modern civilized nations, and suggestions of all may be found in child-study. Three of these classes, memory libraries, mnemonic libraries and picture book libraries, correspond to well recognized book forms. The term “mnemonic”, which is commonly used to include quipus, message sticks, wampum, and similar records, is itself not a very exact term, since all outward symbols, whether representative or conventional, are mnemonic. Moreover, what is generally meant by the term is the use as symbols of objects which do not represent or directly suggest their meaning—in short, of object signs with conventional rather than pictured meaning but as a matter of fact image signs with conventional meaning i.e. all ideograms or phonograms are equally “mnemonic” with conventional objects. A better distinction is therefore into the memory libraries, object libraries (including both representative, or pictorial, object sign collections and conventional object sign collections) and image libraries (including also both representative or pictorial images and arbitrary or conventional signs). For practical purposes, however, we may perhaps use the terms, memory, object, mnemonic, and picture, understanding by object, pictorial object, by mnemonic, mnemonic object and by picture, pictorial image, as distinguished from the mnemonic or conventionalized images known as ideograms and phonograms. To avoid confusion in this matter it must be kept clearly in mind that writing is not picture writing because its symbols are pictures, but because they picture something. If an ox’s head or its image (aleph or alpha) stands for an ox it is pictorial writing but if it stands for “divinity” it is ideographic and if, as it usually does, it stands for the sound “a” it is phonetic-alphabetic writing: It is pictorial writing only when it suggests its own meaning.

Again it must be said that pictorial writing is not confined to image writing as is usually implied by the phrase “picture writing” but applies just as well to objects. A real ox’s head and horns may mean “ox” or “divinity” or “a” just as well as a painting, drawing or sculpture of it.

Yet again it should be noted that the picture of an ox’s head is itself an object as truly as the head itself. The two kinds of objects might be called real or original objects and image objects but for short “objects” (originals) and “images” serve well enough. Again it should be remembered that an object is not a real object because it is in three dimensions or pictures necessarily drawings or paintings. A petroglyph is as suitable for “picture” writing as a painting (indeed most hieroglyphics are sculptured not drawn or painted). On the other hand a petroglyph is no more an “object” than a painting or drawing is.

With these distinctions in mind the following table of the kinds of symbols used in ancient records will make clear the kinds of primitive libraries.

For each of these kinds of “written” records there is a corresponding kind of library or record collection.

The question of the order of evolution among these various kinds of record collections is closely bound up with that of the evolution of language and handwriting, the very invention of handwriting probably implying a feeling of need for kept records.

The commonly recognized ways of human utterance are gesture and oral speech—the one appealing to the eye, the other to the ear, and each leaving its record probably at different points and in different molecular form in the brain. Hand gesture came in course of time to be the highest type of gesture language, evolving as it did into a highly complex and adaptable type of language, and modern hand writing is simply a form of hand gesture which, by means of ink or lead or chisel, or some other material or instrument, leaves a trail of the hand movement in permanent record.

The question whether gesture language preceded sound language may perhaps be settled by the answer to the question whether in the evolution of living beings the eye preceded the ear. If in the age of reptiles one saw the other glide or the grass move before he heard a swish or hiss, and if he himself first stayed still in order to escape being seen rather than heard, then doubtless gesture language began before sound language, and doubtless again also language began among men with simple gestures rather than simple cries. The biologists say in fact that reaction to light came earlier than reaction to sound, eye before ear, and if this is true, gesture language doubtless preceded oral speech. But, however it may be about simple utterance, when it comes to the matter of permanent external documentary record of utterance, it is clear enough that the records of gesture preceded the records of sound, and for some six thousand or eight thousand years, more or less up to yesterday, the only permanent records, or records in external material, were gesture records. Even phonetic writing, so called, is not sound record but a record of sounds translated into gestures; writing is a gesture sign which stands for a sound, not a record of sound. It is only within our own generation that, through the invention of the phonograph, oral or other sound utterance has been recorded in permanent material and libraries of sound records made possible.

The written recording of even signs for sounds did, however, in the evolution of record keeping mark a very decided advance over all previous methods. It was as great an advance perhaps as articulate speech itself is over gesture language or pantomime, and even greater than the next great step in human evolution, the invention of alphabetic writing. It was certainly a longer step in time from the very first beginnings up to this point than from here to the alphabet, perhaps longer than from 3400 B.C. to 1913 A.D., and the period of premnemonic record collections, therefore, it may be said in all seriousness, is perhaps longer than all later periods of library history put together.

The very first rudiments of record keeping were doubtless developed in the animal mind long before it learned expression to other animals and are to be found in the results recorded in its very structure, of its reactions to its environment. Certainly they began at the point where any experience, say of contact with an obstacle, left such record that on the next occasion action was taken in view of the previous experience.

The first attempt at expression or the effort of one individual to communicate an idea to another by signs may have been a mere movement to attract the attention of the other to the simple fact of its existence, and the first record of expression may have been the simple memory of this movement in the other’s mind.

However this may be, in the course of time and among human beings memory was the first record and as long as life was so simple that a man’s memory was sufficient for his own record uses and he felt no need of communicating to a distance, whether in space or time, the necessity of external records was not felt. As soon, however, as the number of a man’s cattle or cocoanut trees, or the contents of his hunting bag got beyond his count (perhaps beyond the number of his fingers and toes) or he felt the need of sending a message of defiance, peace, or ransom to a neighboring tribe, or from a hunting party back to the cave or wigwam, he began to make visible records—objects, specimens, images, and conventional signs of one sort or another. As the art progressed and became more and more complex, pictures of objects and pictures of gestures became the usual form of record until finally these pictures were recognized as standing for certain groups of sounds and phonetic writing had been invented.

Very soon after the introduction of phonetic writing documents began to abound and the chances of survival, therefore, to multiply. The Palermo stone seems to show that actual records by reign and by year of reign began in Egypt as early as the first king of the first dynasty. However that may be, within a few centuries of this time records and collections of records in Egypt had become abundant and varied, and these contained economic records, records of political and religious events, laws, censuses, etc., at least. In Babylonia too, long before 3200 B.C., there had been collections of laws, and a great variety of economic and religious documents.

In brief it may be said therefore that about 3400, or at least 3200 B.C., the vast number of documents, the firm establishment of phonetic record, the pains taken to insure permanence and the suggestions of methodical arrangement and custody point to the beginning of a strictly historic period.