Would turne their owne perfection to abusee,
To seeme like him. So that in speech, in gate ...
He was the marke, and glasse, coppy, and booke,
That fashion’d others.
The spreading of a new pronunciation through imitation must necessarily take some time, though the process may in some instances be fairly rapid. In some historical instances we are able to see how a new sound, taking its rise in some particular part of a country, spreads gradually like a wave, until finally it has pervaded the whole of a linguistic area. It cannot become universal all at once; but it is evident that the more natural a new mode of pronunciation seems to members of a particular speech community, the more readily will it be accepted and the more rapid will be its diffusion. Very often, both when the new pronunciation is easier and when there are special psychological inducements operating in one definite direction, the new form may originate independently in different individuals, and that of course will facilitate its acceptation by others. But as a rule a new pronunciation does not become general except after many attempts: it may have arisen many times and have died out again, until finally it finds a fertile soil in which to take firm root. It may not be superfluous to utter a warning against a fallacy which is found now and then in linguistic works: when some Danish or English document, say, of the fifteenth century contains a spelling indicative of a pronunciation which we should call ‘modern,’ it is hastily concluded that people in those days spoke in that respect exactly as they do now, whatever the usual spelling and the testimony of much later grammarians may indicate to the contrary. But this is far from certain. The more isolated such a spelling is, the greater is the probability that it shows nothing but an individual or even momentary deviation from what was then the common pronunciation—the first swallow ‘who found with horror that he’d not brought spring.’
XV.—§ 12. Reaction.
Even those who have no linguistic training will have some apperception of sounds as such, and will notice regular correspondences, and even occasionally exaggerate them, thereby producing those ‘hypercorrect’ forms which are of specially frequent occurrence when dialect speakers try to use the ‘received standard’ of their country. The psychology of this process is well brought out by B. I. Wheeler, who relates (Transact. Am. Philol. Ass. 32. 14, 1901; I change his symbols into my own phonetic notation): “In my own native dialect I pronounced new as [nu·]. I have found myself in later years inclined to say [nju·], especially when speaking carefully and particularly in public; so also [tju·zdi] Tuesday. There has developed itself in connexion with these and other words a dual sound-image [u·:ju·] of such validity that whenever [u·] is to be formed after a dental [alveolar] explosive or nasal, the alternative [ju·] is likely to present itself and create the effect of momentary uncertainty. Less frequently than in new, Tuesday, the [j] intrudes itself in tune, duty, due, dew, tumour, tube, tutor, etc.; but under special provocation I am liable to use it in any of these, and have even caught myself, when in a mood of uttermost precision, passing beyond the bounds of the imitative adoption of the new sound into self-annexed territory, and creating [dju·] do and [tju·] two.” One more instance from America may be given: “In the dialect of Missouri and the neighbouring States, final a in such words as America, Arizona, Nevada becomes y—Americy, Arizony, Nevady. All educated people in that region carefully correct this vulgarism out of their speech; and many of them carry the correction too far and say Missoura, praira, etc.” (Sturtevant, LCH 79). Similarly, many Irish people, noticing that refined English has [i·] in many cases where they have [e·] (tea, sea, please, etc.) adopt [i·] in these words, and transfer it erroneously to words like great, pear, bear, etc. (MEG i. 11. 73); they may also, when correcting their own ar into er, in such words as learn, go too far and speak of derning a stocking (Joyce, English as we speak it in Ireland, 93). Cf. from England such forms as ruing, certing, for ruin, certain.
From Germany I may mention that Low German speakers desiring to talk High German are apt to say zeller instead of teller, because High German in many words has z for their t (zahl, zahm, etc.), and that those who in their native speech have j for g (Berlin, etc., eine jute jebratene jans ist eine jute jabe jottes) will sometimes, when trying to talk correctly, say getzt, gahr for jetzt, jahr.[72]
It will be easily seen that such hypercorrect forms are closely related to those ‘spelling pronunciations’ which become frequent when there is much reading of a language whose spelling is not accurately phonetic; the nineteenth century saw a great number of them, and their number is likely to increase in this century—especially among social upstarts, who are always fond of showing off their new-gained superiority in this and similar ways. But they need not detain us here, as being really foreign to our subject, the natural development of speech sounds. I only wish to point out that many forms which are apparently due to influence from spelling may not have their origin exclusively from that source, but may be genuine archaic forms that have been preserved through purely oral tradition by the side of more worn-down forms of the same word. For it must be admitted that two or three forms of the same word may coexist and be used according to the more or less solemn style of utterance employed. Even among savages, who are unacquainted with the art of writing, we are told that archaic forms of speech are often kept up and remembered as parts of old songs only, or as belonging to solemn rites, cults, etc.