THE HISTORY OF LANGUAGES While in the case of the vast majority of languages we know only their present stage or quite recent stages, as these languages have only been noted down in recent periods by missionaries and philologists, there are fortunately other languages whose history we can trace back for centuries, some even for a few thousand years. But we must remember that even with regard to those languages in which writing goes furthest back, the period thus known to us is extremely short in comparison with those thou sands of years which have left no record. Besides, writing is at best a very imperfect medium for language-study, and its de ficiencies are especially glaring in the oldest systems of writing. The Egyptian hieroglyphs and the early Semitic alphabets leave us almost entirely in the dark as to the vowels of these languages. Chinese writing is even worse, and the nature of the sounds of early Chinese is only now slowly revealing itself to the patient researches of B. Karlgren and other scholars. We are of course much better informed with regard to such old languages as Sanskrit, Greek, Latin, Old Turkish, etc., where not only con sonants, but also vowels are indicated. But even the best of these alphabets are incapable of denoting a great many nuances—in the qualities of the sounds, in quantity, stress, tone, etc.—which are of importance in the spoken language and which may exercise a far-reaching influence on the development of speech. Speaking and writing are really incompatibles because they address them selves to two distinct senses, and even the neatest phonetic script is nothing but a makeshift contrivance for the study of the real, i.e., the spoken language. Our descendants will be better off in that respect as they will have gramophone records for the study of our speech.
A great difficulty is created for the historian by the conserva tism of people who persist in writing words in a customary spell ing, even where the sounds have changed greatly from what they were (see above "Standard Languages"). Sometimes this disad
vantage is remedied by the express statements of early phonetic observers who describe the pronunciation of their own times and call attention to the discrepancies between spelling and pro nunciation ; in other cases we must draw our conclusions as to earlier pronunciation in a more indirect way, through loan words to and from the language concerned, through rhymes and rhythms in poetical works, and through mistakes in spelling. When we find, for instance in the Shakespeare folio, spellings like solembe, we are justified in inferring, both that n in solemn and that b in words like comb, had at that time disappeared from actual pronunciation. Mistakes made by illiterate spellers have proved a very fruitful source in the hands of recent investigators (H. C. Wyld and others).
In writing people will tend to be more conservative than in speaking, not only in their spelling, but also in their grammar and in their choice of words. The verbal forms in -th (goeth, bath) were kept in writing long after the forms in s (goes, has) had become universal in speech; thus also forms like mine own, etc. The subjunctive in "If he come" is much more frequent in writing than in actual speech. Some literary words have prac tically disappeared from the spoken language, e.g., abode, wax ( =grow), anent, belike. Such things are found in all languages that have a long literary tradition, though their extent and char acter may vary considerably between one language and another. In Swedish and Dutch the old word-genders survive only in the written language. In Modern Greek the tendency to pre serve as much as possible of the old glorious language has led to an extreme artificiality in the written language, which creates infinite difficulties for all its users, and the same is the case in some of the Southern Indian languages, such as Telugu. The first requirement of a linguistic historian must therefore be that he is constantly on his guard against false impressions due to the way in which languages are written.