Thus, for instance, our present word day is found in Anglo-Saxon as— Nom. and Acc. Sing. dæg Plur. dagas Gen. ” dæges ” daga Dat. ” dæge ” dagum, where æ was pronounced as the a in man, hat, etc., and a as a in father: æ is therefore a ‘front-vowel,’ like the a in fate, ee in feet, etc., while a of dagas was a ‘back-vowel,’ as are o or u.
The phonetic development of final or medial g differs according to the vowel which preceded it. If this was a front-vowel the g became y (vowel),[88] if it was a back-vowel the g became w. Thus, e.g., A.S. hnægan, E. neigh; A.S. wegan, E. weigh; A.S. hálig, E. holy: but A.S. búgan, E. (to) bow; A.S. boga, E. bow; A.S. ágan, E. to own. Accordingly dæg, etc., in the singular became day, whilst in the plural we find in M.E. dawes, etc. As soon, however, as analogy had established the ‘regular’ s plural to the sing. day, plur. days, the verb (to) dawn, A.S., dagian was thereby isolated completely, and no speaker who is not more or less a student of the history of English, connects the verb with the noun.
Another instance maybe found in the word forlorn.
To understand the history of this word we must know what is meant by Verner’s law.
Among the first illustrations of the regular correspondence of the several consonants in Latin and in the Teutonic languages are such pairs as mater, mother; pater, father; frater, brother; tres, three; tu, thou: in all of which a th is found in English where the Latin shows a t. This and other similar regular interchanges were generalised by Grimm and formulated by him as a law, part of which stated that if the same word was found in Latin, Greek, and Sanscrit, as well as in Teutonic, a k, t, p, in the first three languages appeared as h, th, f in Low German, of which family English is a representative.
All our sets of examples seem to illustrate and confirm this law. If, however, we trace the English words back to older forms, we see that this absolute regularity is disturbed. In Middle-English almost invariably, and in Anglo-Saxon invariably, we find fader, moder, brother, A.S. fæder, módor, bróðor, in perfect agreement with O.S. fadar, môdar, brothar, and Goth. fadar, brothar (cf. Mod. Ger. vater, mutter, but bruder). It was Karl Verner who explained this irregularity, and proved that it was connected with the place of the accent in the Teutonic languages, not as we find it now, but as it can be proved to have existed in those languages, where it corresponded generally with the Greek accents, or more closely still with the accent in Vedic Sanscrit. There we find that in the corresponding forms pitar, mâtar, and bhratar, the accent or stress lay on the FIRST syllable in bhratar, but on the LAST in pitar and mâtar. Verner proved by numerous examples that only where an ACCENTED vowel preceded the p, t, k, Teutonic showed the corresponding f, th, h; but that, on the other hand, where the preceding vowel was UNACCENTED, instead of f we found b, and d instead of th, g instead of h. And also, instead of s, which was elsewhere found both in Latin or Sanscrit as well as in Teutonic, z was found, which z further changed into r in Anglo-Saxon.
Thus—to give one more instance—the suffix ian, used to form causatives in Teutonic, once bore the accent, which afterwards was placed on the root-syllable. Accordingly, the causative of the verb rís-an (to rise) was once rás-ian,[89] which, with z, and, later on, r, instead of s, changed into rǽr-an, Mod. Eng to rear.
The so-called Grammatical change in Anglo-Saxon (and other Teutonic languages) now becomes clear: The verb in past sing. plur. p. part. céosan (to choose) has caés curon coren sniðan (to cut; Scotch, sned) snáð snidon sniden téon (to drag) has téah tugon togen and all this series of regular sound-change depends upon the fact that in the past plural and in the past participle the accent fell ORIGINALLY on the termination. Similarly, (for) léosan,—léas,—luron,—loren, from which last form we have our word forlorn, meaning, therefore, ‘completely lost.’ Already, however, in Anglo-Saxon, in very many verbs all traces of this grammatical change have disappeared, and the history of the strong conjugation in Middle-English shows the gradual supersession of the consonants in the past plural and past participle by those found in the present and past singular. Hence those forms in which these older consonants remained were more and more isolated from the groups with which they are etymologically connected; and as little as in popular consciousness to rear is grouped with to rise, so little is the adjective forlorn thought of as a member of the group to lose, lost, etc.
We have had already more than one occasion to point out that not only words, but also syntactical combinations and phrases can and do form matter groups. Nay, even the various meanings of a syntactical relation are thus combined.
Such a relation, for instance, is that expressed by the genitive. Though we employ—and formerly employed more generally than now—this case with various meanings, all these meanings are more or less (rather less) consciously felt as one, or at least are closely related—and they continue to be so felt, i.e. the grouping remains a close one—as long as these various usages remain general and what we may call living. When, however, any one of these usages becomes obsolete, and the relation indicated finds another form of expression in some other syntactical arrangement, some few examples of the older mode of expression, strengthened as they are by, e.g., very frequent employment, remain, but cease to be felt as instances of that relation.