Sometimes one and the same phonetic change is gradually ex tended to a greater and greater number of combinations. Such is the case with the dropping of the first element of the English "long u," phonetically written (ju), after consonants. This hap pened first after r, as in true; then after 1 when preceded by an other consonant (blue)—in these cases the sound-law is carried through consistently. Next came 1 with no consonant before it, as in Lucy, and s as in Susan, suit, where there is a good deal of vacillation; after t, d, n, as in tune, due, new, the tendency has prevailed only among vulgar British speakers and some Ameri cans. As regards these speakers the "law" must be said to affect all consonants produced by the point of the tongue, while it will have to be formulated in various less comprehensive terms for other speakers. It is probable that many of the comprehensive prehistoric sound changes, of which we see only the final results, may have similarly spread from more modest beginnings.
While some phonetic changes bring about a convergence of two or more forms of the same word, which were previously distinct, as when the different Old English (Anglo-Saxon) case-forms sunu and suna became Middle English sane and modern son, other sound changes lead to divergences between forms of the same word, as when keep and kept, which had formerly the same vowel, now have distinct vowels ; a still greater divergence is seen in seek, sought, which in prehistoric times had the same vowel (o).
When philologists speak, as they usually do, of sound-laws or phonetic laws, it must never be forgotten that the word "law" is not used exactly as in the physical sciences, but always refers to historical occurrences restricted to one particular language or dialect and to one definite period of its existence. But from all these particular "laws" we may hope some day to be able to find out some more general laws determining linguistic development everywhere and at all times, though the result may be only com paratively vague statements of universal "tendencies." Among these we must mention the tendency toward assimila tion of neighbouring sounds so as to save some particular articu lation. Thus, the s is voiced, i.e., has acquired the sound of z, before a voiced stop 'n husband (a compound of hus, house), gooseberry, etc. The place of articulation is shifted, when n be comes m before a labial, as in Stamford from Stanford, vamp (upper leather) from French avant-pied, and when n is pro nounced with the back of the tongue before a back consonant, e.g., in handkerchief ; inversely in has become n before a t, be cause that is pronounced with the tip of the tongue, in ant from cemette, count frc,m Latin, comit-; similar examples abound in many languages, though the extent to which such assimilatory forms are tolerated varies a good deal between one language and another.
Analogy.—Irregularities in forms, whether due to such pho netic changes as have just been mentioned, or to other causes, will always tend to disappear. This is brought about by the springing up of analogy-formations, which have played and still play an enormous role in the development of all languages. They have led to the English genitive being now always formed by means of s, while Old English had a variety of f ormations; fa der father's, manna men's, cwene queen's, haran hare's, etc. The sending of the plural has been also extended by analogy to nearly all words; many verbs which in earlier times had a preterite of the "strong" or irregular form now form their preterite with the regular "weak" ending -ed. The same thing happens in all departments of grammar. Irregular forms may indeed hold their own against all levelling influences, but this can only be if they are in very frequent use, so that a child hears them and has oc casion to utter them constantly from a very early age, e.g., the plural men, the case forms I, me, she, her, the comparative better, verb forms like see, saw, stand, stood, etc.
In these various ways form-systems constantly tend towards simplification ; we see how the languages belonging to the Indo European family have reduced the number of cases, etc., inde pendently of one another, though with varying rapidity, and in some languages, e.g., Celtic, the process has been countered by the development of new complications.