So far, then, the inner history of word-languages shows three stages. In the first there appears, within highly developed but wordless communication-languages, the first names--units in a new sort of understanding. The world awakens _as a secret_, and religious thought begins. In the second stage, a complete communication-speech is gradually transformed into grammatical values. The gesture becomes the sentence, and the sentence transforms the names into words. Further, the sentence becomes the great school of understanding _vis-à-vis_ sensation, and an increasingly subtle significance-feeling for abstract relations within the mechanism of the sentence evokes an immense profusion of inflexions, which attach themselves especially to the substantive and the verb, the space-word and the time-word. This is the blossoming time of grammar, the period of which we may probably (though under all reserves) take as the two millennia preceding the birth of the Egyptian and Babylonian Culture. The third stage is marked by a rapid decay of inflexions and a simultaneous replacement of grammar by syntax. The intellectualization of man’s waking-consciousness has now proceeded so far that he no longer needs the sense-props of inflexion and, discarding the old luxuriance of word-forms, communicates freely and surely by means of the faintest nuances of idiom (particles, position of words, rhythm). By dint of speaking in words, the understanding has attained supremacy over the waking-consciousness, and to-day it is in process of liberating itself from the restrictions of sensible-verbal machinery and working towards pure mechanics of the intellect. Minds and not senses are making the contact.

In this third stage of linguistic history, which as such takes place in the biological plane[194] and therefore belongs to _man as a type_, the history of the higher Cultures now intervenes with an entirely new speech, the speech of the distance--writing--an invention of such inward forcefulness that again there is a sudden decisive turn in the destinies of the word-languages.

The written language of Egypt is already by 3000 in a state of rapid grammatical decomposition; likewise the Sumerian literary languages called _eme-sal_ (women’s language). The written language of China--which _vis-à-vis_ the vernaculars of the Chinese world has long formed a language apart--is, even in the oldest known texts, so entirely inflexionless that only recent research has established that it ever had inflexions at all.[195] The Indogermanic system is known to us only in a state of complete break-down. Of the Case in Old Vedic (about 1500 B.C.) the Classical languages a thousand years later retained only fragments.[196] From Alexander the Great’s time the dual disappeared from the declension of ordinary Hellenistic Greek, and the passive vanished from the conjugation entirely. The Western languages, although of the most miscellaneous provenance imaginable--the Germanic from primitive and the Romanic from highly civilized stock--modify in the same direction, the Romanic cases having become reduced to one, and the English, after the Reformation, to zero. Ordinary German definitely shed the genitive at the beginning of the nineteenth century and is now in process of abolishing the dative. Only after trying to translate a piece of difficult and pregnant prose--say of Tacitus or Mommsen--“back”[197] into some very ancient language rich in inflexions does one realize how meantime the technique of signs has vaporized into a technique of thoughts, which now only needs to employ the signs--abbreviated, but replete with meaning--merely as the counters in a game that only the initiates of the particular speech-communion understand. This is why to a west-European, the sacred Chinese texts must always be in the fullest sense a sealed book; but the same holds good also for the primary words of every other Culture-language--the Greek λογός and ἀρχή, the Sanskrit _Atman_ and _Braman_--indications of the world-outlook of their respective Cultures that no one not bred in the Culture can comprehend.

The external history of languages is as good as lost to us in just its most important parts. Its springtime lies deep in the primitive era, in which (to repeat what has been said earlier), we have to imagine “humanity” in the form of scattered and quite small troops, lost in the wide spaces of the earth. A spiritual change came when reciprocal contacts became habitual (and eventually natural) to them, but correspondingly there can be no doubt that this contact was first sought for and then regulated, or fended off, by means of speech, and that it was the impression of an earth filled with men that first brought the waking-consciousness to the point of tense intelligent shrewdness, forcing verbal language under pressure to the surface. So that, perhaps, the birth of grammar is connected with the race hall-mark of the grand Number.

Since then, no other grammatical system has ever come into existence, but only novel derivatives of what was already there. Of these _authentic_ primitive languages and their structure and sound we know nothing. As far as our backward look takes us, we see only complete and developed linguistic systems, used by everyone, learned by every child, as something perfectly natural. And we find it more than difficult to imagine that once upon a time things may have been different, that perhaps a shudder of fear accompanied the hearing of such strange and enigmatic language--an awe like that which in historic times has been and still is excited by script. And yet we have to reckon with the possibility that at one time, in a world of wordless communication, verbal language constituted an aristocratic privilege, a jealously preserved class-secret. We have a thousand examples--the diplomats with their French, the scholars with their Latin, the priests with their Sanskrit--to suggest that there may have been such a tendency. It is part of the thoroughbred’s pride to be able to speak to one another in a way that outsiders cannot understand--a language for everybody is a vernacular. To be “on conversational terms with” someone is a privilege or a pretension. So, too, the use of literary language in talking with educated people, and contempt for dialect, mark the true bourgeois pride. It is only we who live in a Civilization wherein it is just as normal for children to learn to write as to learn to walk--in all earlier Cultures it was a rare accomplishment, to which few could aspire. And I am convinced that it was just so once with verbal language.

The tempo of linguistic history is immensely rapid; here a mere century signifies a great deal. I may refer again to the gesture-language of the North Indians,[198] which became necessary because the rapidity of changes in the tribal dialects made intertribal understanding impossible otherwise. Compare, too, the Latin of the recently discovered Forum inscription[199] (about 500) with the Latin of Plautus (about 200) and this again with the Latin of Cicero (about 50). If we assume that the oldest Vedic texts have preserved the linguistic state of 1200 B.C., then even that of 2000 may have differed from it far more completely than any Indogermanic philologists working by _a posteriori_ methods can even surmise.[200] But _allegro_ changes to _lento_ in the moment when script, the language of duration, intervenes and ties down and immobilizes the systems at entirely different age-levels. This is what makes this evolution so opaque to research; all that we possess is remains of written languages. Of the Egyptian and Babylonian linguistic world we do possess originals from as far back as 3000, but the oldest Indogermanic relics are _copies_, of which the linguistic state is much younger than the contents.

Very various, under all these determinants, have been the destinies of the different grammars and vocabularies. The first attaches to the intellect, the second to things and places. Only grammatical systems are subject to natural inward change. The use of words, on the contrary, psychologically presupposes that, although the expression may change, inner mechanical structure is maintained (and all the more firmly) as being the basis on which denomination essentially rests. _The great linguistic families are purely grammatical families._ The words in them are more or less homeless and wander from one to another. It is a fundamental error in philological (especially Indogermanic) research that grammar and vocabulary are treated as a unit. All specialist vocabularies--the jargon of hunter, soldier, sportsman, seaman, savant--are in reality _only stocks of words_, and can be used within any and every grammatical system. The semi-Classical vocabulary of chemistry, the French of diplomacy, and the English of the racecourse have become naturalized in all modern languages alike. We may talk of “alien” words, but the same could have been said at some time or other of most of the “roots,” so-called, in all the old languages. All names adhere to the things that they denote, and share their history. In Greek the names for metals are of alien provenance; words like ταῦρος, χιτῶν, οἶνος are Semitic. Indian numerals are found in the Hittite texts of Boghaz Keüi,[201] and the contexts in which they occur are technical expressions which came into the country with horse-breeding. Latin administrative terms invaded the Greek East,[202] German invaded Petrine Russia in multitudes, Arabic words permeate the vocabulary of Western mathematics, chemistry, and astronomy. The Normans, themselves Germanic, inundated English with French words. Banking, in German-speaking regions, is full of Italian expressions,[203] and similarly and to a far greater extent masses of designations relating to agriculture and cattle-breeding, to metals and weapons, and in general to all transactions of handicraft, barter, and intertribal law, must have migrated from one language to another, just as geographical nomenclature always passed into the proper vocabulary of the dominant language, with the result that Greek contains numerous Carian and German Celtic place-names. It is no exaggeration to say that the more widely an Indogermanic word is distributed, the _younger_ it is, the more likely it is to be an “alien” word. It is precisely the very oldest names that are hoarded as private possessions. Latin and Greek have only quite young words in common. Or do “telephone,” “gas,” “automobile,” belong to the word-stock of the “primitive” people? Suppose, for the sake of argument that three-fourths of the Aryan “primitive” words came from the Egyptian or the Babylonian vocabularies of the third millennium; we should not find a trace of the fact in Sanskrit after a thousand years of unwritten development, for even in German thousands of Latin loan-words have long ago become completely unrecognizable. The ending “-ette” in “Henriette” is Etruscan--how many genuine Aryan and genuine Semitic endings, notwithstanding their thoroughly alien origin, defy us to prove them intruders? What is the explanation of the astounding similarity of many words in the Australian and the Indogermanic languages?

The Indogermanic system is certainly the youngest, and therefore the most intellectual. The languages derived from it rule the earth to-day, but did it really exist at all in 2000 as a specific grammatical edifice? As is well known, a single initial form for Aryan, Semitic, and Hamitic is nowadays assumed as probable. The oldest Indian texts preserve the linguistic conditions of (probably) before 1200, the oldest Greek those of (probably) 700. But Indian personal and divine names occur in Syria and Palestine,[204] simultaneously with the horse, at a much later date, the bearers of these names being apparently first soldiers of fortune and afterwards potentates.[205] May it be that about 1600 these land-Vikings, these first _Reiter_--men grown up inseparable from their horses, the terrifying originals of the Centaur-legend--established themselves more or less everywhere in the Northern plains as adventurer-chiefs, bringing with them the speech and divinities of the Indian feudal age? And the same with the Aryan aristocratic ideals of breed and conduct. According to what has been said above on race, this would explain the race-ideal of Aryan-speaking regions without any necessity for “migrations” of a “primitive” folk. After all, it was in this way that the knightly Crusaders founded their states in the East--and in exactly the same locality as the heroes with Mitanni names had done so twenty-five hundred years before.

Or was this system of about 3000 merely an unimportant dialect of a language that is lost? The Romanic language-family about A.D. 1600 dominated all the seas. About 400 B.C. the “original” language on the Tiber possessed a domain of little more than a thousand square miles. It is certain that the geographical picture of the grammatical families at about 4000 was still very variegated. The Semitic-Hamitic-Aryan group (_if_ it ever did form a unit) can hardly have been of much importance at that time. We stumble at every turn upon the relics of old speech-families--Etruscan, Basque, Sumerian, Ligurian, the ancient tongues of Asia Minor, and others--that in their day must have belonged to very extensive systems. In the archives of Boghaz-Keüi eight new languages have so far been identified, all of them in use about the year 1000. With the then prevailing tempo of modification, Aryan may in 2000 have formed a unit with languages that we should never dream of associating with it.