« PoprzedniaDalej »
consonant, I will also quote Mr. Bell's description of that consonant: it is "formed by placing the centre of the lower lip on the edges of the upper teeth, while the breath hisses through the interstices between the teeth, or between the teeth and the lip." The close connection of these two sounds is shown by the individual tendency to substitute the lip for the point consonant, as in fink, frough for think, through, and by many changes in various languages. In at least one English word this substitution has become so fixed as entirely to supersede the original form. This word is strife, or as a verb strive, in A.Saxon stríð, stríðan, Germ. streiten, etc., and there may be other instances. These consonants are closely allied to their corresponding stops or mutes, and are intermediate to them and the primary consonants, which are formed without any contact. Both of these consonants are sounded with breath and voice, and the distribution of the voiced and breathed varieties in the Teutonic languages is one of the most interesting points connected with their history. For the sake of brevity, I shall, in speaking of these sounds, call them respectively divided d and t, b and p, the last two corresponding to our v and f, while the letters used to denote these sounds will be called by their ordinary names.
To begin then with modern English, it may be stated as a general rule that the div. t is used initially and finally, the d only medially. The last rule is, however, only true with certain limitations. There are many positions in which it is a practical impossibility to pronounce a voiced consonant; thus in the word sixths the th is necessarily pronounced with breath instead of voice, being preceded and followed by breathed consonants. The most accurate statement perhaps is that the voiced pronunciation is necessary when the th is preceded and followed by vowels, one of which has, however, often been dropped in modern English pronunciation, although sometimes preserved in writing. Examples are: paths, A. Saxon paðas, where the second vowel is no longer written; and lithe, A. Saxon lide, where it is written, but no longer pronounced.
The exceptions to these rules are as follow:-1st. There is
a large number of pronouns and particles which are always pronounced with an initial div. d. 2nd. There is the isolated word with, also with exceptional vocal pronunciation.
These remarks apply only to words of Teutonic origin; the foreign th is always voiceless, as in myth, mythic.
The analogy of the divided lip-consonant is very close. Not only are the general rules for the distribution of the two varieties the same, but the exceptions can also be matched. These exceptions are the prep. of, analogous to with, and the two substantives vat and vixen, the latter contrasting remarkably with fox (Germ. fuchs, füchsinn). The chief dif ference lies in the graphic forms. The one series has but one sign, th; the other no less than three, f, v, and ph: the latter uniformly voiceless and used only in foreign words.1 only modern Teutonic language besides English which possesses the two sounds of the point series is Icelandic, where the distribution of the sounds is different from the English and perfectly regular, the div. t being invariably sounded in the beginning of words, the d medially and finally. These sounds are written with two different letters, which seem to have been taken from the A.Saxon alphabet. The div. t is expressed by the p or porn, the d by the 8 or stungið dé. The distribution of the corresponding labials is exactly parallel, both sounds, however, are expressed by the ƒ alone. There is still another Teutonic language which possesses the div. d, but not the t. The d in modern Danish has this sound when it stands between two vowels or finally, but there can be little doubt that this Danish sound is of comparatively late origin and has no direct connection with the English and Icelandic sounds. This is shown by such words as sad, mad (English sat, meat, Icel. sat, mat), where the t has first become voiced and then divided. In the same way the d of the preposition ved, although corresponding in pronunciation to the div. d of the English with, and Icelandic við, must be considered as a modification of an earlier d. In all cases the two divided consonants in Danish and Swedish are written t and d, and, with the exception of the Danish words just mentioned, are pronounced accordingly. The dis
tribution of these ts and ds corresponds to that of the Icelandic div. consonants which they represent; thus we find ting in Swedish and Danish for the Icelandic ping, ed, klæde, answering to eið, klæði. There is, however, an important class of exceptions, consisting of those pronominal words which we have already met with in English. In contrast to the above-mentioned ting, we find in Swedish du, den, detta, which do not correspond to the Icelandic forms þú, þann, petta, but postulate an original Jú, Jann, detta. The same is the case in Danish. The dialect of the Feroe islanders has also substituted the stopped for the div. consonant in the beginning of words, and shows its close affinity to Icelandic by the forms tú, tín, tá (Icel. þú, þin, þá). These remarkable forms are characteristic of the dialect and do not occur elsewhere in the Teutonic languages.
Notwithstanding these Feroic sounds and the mod. Icelandic usage, there is strong reason for believing that the oldest Icelandic of the thirteenth century made the same distinction in the initial thorns as in modern English, many of the oldest MSS. constantly using the in such words as dú, dann, dá.3 The sole difference therefore between the modern English and the old Icelandic usage lies in the final sounds, which in English are breathed, in Icelandic voiced.
It will be necessary now to turn our attention to the older languages and examine the evidence afforded by their orthographies. In the Gothic of Ulfilas we find the two sounds of the div. t and d in English and Icelandic expressed by one simple sign, a modification of the same runic letter which is used in A.Saxon and Icelandic. The fact of the sign being simple and uniform makes it probable that the sound was also simple and uniform, either a div. t or d. A strong argument in favour of the latter pronunciation is afforded by the frequent and, in many cases, apparently arbitrary change between þ and d in the middle and end of words. When we find baup and baud, nimip and nimid constantly varying, it is difficult to believe that the voiced d would at once change to a voiceless þ, or vice versa. Further proofs may be gathered from an examination of the other old languages. In the old
High German language, which is next in antiquity to Gothic, we find the Gothic thorn generally represented by a d, which has continued in use up to the present day. In some of the oldest documents which verge towards Low German the combination dh is written for d in all positions, initial, medial, and final.
A serious objection may, however, be brought against the original voiced pronunciation, grounded on the connection of the Teutonic languages with the old Aryan languages in general, where the thorn is represented by a t. It cannot be denied that the direct conversion of a voiceless stopped consonant into a voice divided is phonetically improbable, or even impossible; but there is an intermediate stage possible, which, once established, removes all difficulties. We have seen above that the Danish div. d appears in all cases as a modification of, an older d. To this I will now add that in modern Greek the letter delta is pronounced as a divided d in all positions, initial, medial, and final, so that the sound of d is almost unknown in that language, and that the div. d in English and Icelandic often represents an earlier d, the change of d into taking place regularly in Icelandic as in Danish. I think these facts are strong enough to justify the assumption of an earlier stage of the Teutonic languages, in which the original t was changed into d, thus adding a fourth consonantal stage to the three usually accepted. The probable relation of this earlier Teutonic stage to the others is indicated in the table below :
If this theory is correct, the d in those words which fluctuate between d and þ in Gothic is the original sound. It is also possible that the High German d, which corresponds to the Gothic thorn, is in reality a remnant of this earlier stage, just as the Old High German verb shows many forms older than the Gothic, and in comparatively late documents.
Such forms as hapêm, hapêmês compared with haba, habam, may well justify an assumption of similar chronological priority in the consonantal system.
Nor are these early d forms confined to High German alone, they appear also in Gothic and the Low German dialects generally, but of course only as isolated irregularities. Such words are the Gothic fidvor and fadar, corresponding to the Latin quatuor and pater, the O.H.G. equivalent of the latter is vatar, which does not correspond to a regular Gothic fapar. The genuineness of the form fadar is confirmed by the other old languages, Old Saxon showing fader and A.S. fœder, while the normal father, faðir first appear in modern English and Icelandic.
The analogy of the corresponding labial consonant is so close, that a brief statement of the older forms will be enough, ƒ and b vary in Gothic in the same way as þ and d. Some words, such as sibun (Latin septem), preserve the b invariably, which also appears occasionally in High German. But in the majority of these words, High German show a divided consonant, thus disturbing the parallel with the dental series. It is important to observe that this sound is generally expressed by v, which letter appears also in the oldest low German or old Saxon, alternating with ƒ and a peculiar letter consisting of a b with a stroke through the stem.
The connection of these theories with that important generalization known as Grimm's law, is so intimate that some remarks thereon will hardly be deemed out of place, especially when it is considered that the views set forth here, and that law, are to a certain extent opposed to one another. Grimm's law may be compendiously defined as a generalization of the consonantal differences in the old Aryan languages and Low and High German. The investigation of these differences is perfectly easy as long as it is confined to the graphic forms, but it is evidently impossible to generalize with any certainty without an accurate knowledge of the sounds represented by these forms. The want of this knowledge is strongly illustrated by the vague use of the term 'aspiration.' Even if it be correct to talk of a divided or primary as opposed to a