(pericise, ocio, prudencius) are not of early times, and are given by ¿ Corss. }7.04. collectors who lived at a time when the spelling ocio at least was usual; one (renunciationem) is from a notoriously bad collector: a fifth (disposicionem) is from a late Neapolitan inscription containing several misspelt words1; the remaining two (terminac[iones], defenciones) are from an inscription at Medjana in Africa of the time of Alexander Severus (222—235 p. Chr.). Even if these last be rightly copied, (which is not certain,) an inference from African spelling or pronunciation in the 3rd century to ordinary Roman spelling and pronunciation in (say) the Augustan age would be about as justifiable as an inference from the usage of words or constructions in Apuleius or Tertullian to that of Cicero or Quintilian. It is curious that the grammarian (Pompeius) whom Prof. Max Müller quotes as his authority for saying that' we 'know for certain that in the 5th century it was considered wrong 'not to assibilate ti before a vowel,' was also an African, from Mauretania, and as regards his age all that is tolerably certain is that he did not live before the 5th century, and not later than the end of the 7th century (Keil, Gram. Lat. v. p. 93. See also Teuffel, Gesch. d. Röm. Litt. p. 982). And again, another African, Commodianus, of the 3rd century, has in an acrostic the word cum for the initial word of the line which is to give the last letter but three of concupiscentiae2 (L. Müller, De re metr. p. 262, quoted by Corssen II. 1003). Thirdly, what does this confusion really prove as to the pronunciation of ci before a vowel, at the time, be it what it may, when the confusion existed? Prof. M. Müller says; 'The only 'point where these two letters (c and t) can possibly meet is the 'assibilation. Ti may go as far as tsi, but unless ki also went as 'far as tshi, the two could not have met, and no Roman whether in 'Italy or Africa could have attempted to write renuntiatio by 'renunciatio' (Academy, p. 146). I reply (1) by referring to Prof. nensi commutatas reperies litteras c et t, quod qui ante septimum d Müller's instructive Lectures, 11. p. 168, where, quoting Marsh, he says, 'We are told by careful observers that the lower classes in [French] Canada habitually confound t and k, and say mékier, moikié for métier, moitié. Quintilian (if the MSS are correct, I. II. 5, ed. Halm) speaks of that 'fault of pronunciation by which c and g are softened into t and d' (comp. Schuchardt, III. 81, sq.). (2) I refer to an authority whom Prof. Müller will respect-Mr Ellis (quoted in App. A. xxv.), who explains distinctly how the confusion of t with c arises, and in the stage of ky, ty, before either is assibilated; and v. Raumer (who seems to me to have inspired M. Müller in his argument generally) says the same (Gesam. Schriften, p. 92). (3) I venture to go still farther, and, while fully admitting the theoretical possibility of palatalised k and t (ky, ty) having been the mediator between ce, ci and the modern assibilated pronunciations, such as s, ts, or th, I hesitate as to its reality. For, as Corssen says (1. 49), there is not a spark of positive evidence for it: and, if c once became t, the change of t to s is far too common a phenomenon in Latin to necessitate an explanation, which applies only to t before i (cf. § 191 and infr. p. liv.). It must be remembered that the palatalisation of c into ch=sh in French is before the vowel a1. (Diez, I. 249, considers here the intermediate step to have been a guttural aspirate, Germ. ch.) To sum up; as there is not one particle of trustworthy evidence, before at least the fifth or sixth century, for any other pronunciation of c than that of the sharp guttural, except the few reminiscences of the sound of g, two African inscriptions, and the African acrostic of the 3rd century with the doubtful inferences deduced from them, I am unable to see how it can be any defence of so thoroughly confusing a pronunciation of the Latin of Cicero and Quintilian, as arises from sounding c as s, that it is theoretically possible for the Romans to have made a difference in ci compared with ca, which was yet so small that no grammarian noticed it, and no writer attempted to express it. 1 So in English the pronunciation of c as ky took place only (?) before a; e. g. card, kind (=kyaind), sky (= skyai). On g before æ, e and i. That g in Latin was not pronounced as English J (=dzh), and that it was always hard before all vowels, may be inferred from the following arguments. (Compare also the discussion of c before the like vowels.) 1. Closely connected forms exhibit perpetual alterations of the vowel following g, without any evidence of a desire to change g before e or i; e.g. malignus for maligenus; gigno for gigeno; tegmen for tegimen; tignum compared with tigillum, &c. Similarly rego, regis, regit becomes rectum (for regtum); reg- makes regis, regi, regum, regulus, and rex (for reg-s, rec-s); ager, agri; fuga, fuga, fugax, fugio, fugitivus. し 2. In Greek g is always represented by y; and y is represented by g. It is true in modern Greek y before and is Eng. y; but it is by no means certain when y first gained this sound. And moreover the sound of y is not that of Engl. J. 3. There is no trace to be found in the grammarians of any different sound of g before the several vowels. This is the more noticeable, because they speak of the effect of g and c, upon a preceding n, in converting the dental into the guttural nasal. But they make no allusion to any difference in the g. Yet the instances adduced contain the lingual as well as the labial vowels, e.g. angvis, ingenuus, anceps, Longinus, angulus, angens. It is no doubt not impossible that this change in the sound of n should be made before palatals such as Engl. ch and j; but we do not make it in English. I infer that the Latins had (in these cases at least, and if in these, why not in others?) c and g hard, whether e and i, or a, o, u followed. 4. There is no evidence of g having such a sound as Engl. j before the 4th or 5th century p. Chr., according to Schuchardt; before the 5th century, according to Corssen. Diez (1. 268) infers from the Anglo-Saxon alphabet that g was the guttural flat mute up to the 7th century. The omission of g before 1, in major for magior, does not appear to imply the assibilation of g. For it takes place before v as much as before i, e. g. nivis for nigvis, malo for magvolo; and g is too commonly omitted before consonants to make its omission before semiconsonants unnatural. There is evidence in the 4th and 5th centuries of its having the sound of Engl. y (=j), e. g. magestates for majestates, BELEVTL for viginti. Possibly this may have existed dialectically earlier. On dentals; especially ti before a vowel. On the pronunciation of ti we have a distinct statement by Isidore in the beginning of the 7th century p. Chr., viz. that before a vowel tia should be sounded as zia. And Pompeius (v. pp. 104, 286, ed. Keil) and Consentius (v. p. 395, ed. Keil) appear to say the same. But these are apparently not much, if at all, earlier witnesses; and accordingly donationem, donationes, are represented in Ravenna Greek of the 6th century by δωναζονεμ, δονατζίονες. And since the 6th century, according to Corssen, instances occur of a similar assibilation, in which the i was not preserved, e.g. constantso, constanzo are written for constantio. Schuchardt (1. 104. 150) thinks that assibilation began as early as the 2nd century p. Chr., but did not become general till a much later period. In Umbrian and Oscan it appeared before the first Punic War, and the origin of such forms as formonsus is probably to be found in formontios (see § 813). On di before a vowel see § 154. A final d was often pronounced as t (§ 150); and Quintilian's words (1. 7, § 5) imply, I think, that there was no difference in the pronunciation of ad and at, though the difference in spelling appears to have continued long. But d is rarely final (§ 155), and Velius Longus (beginning of 2nd cent. p. Chr.) speaks of apud and sed being pronounced with d (p. 2231, Putsche). Mr Munro calls attention to the fact that the continental t (and therefore of course d) is more dental than with us. Mr Ellis (Phil. Soc. Trans. 1867, Suppl. p. 12) describes the European dental as formed by pressing the tongue against the teeth, whereas in English the tongue scarcely reaches the gums. (See however Eng. Pron. P. 477, n.) But I do not think this can affect the question of the interchange of d and t. That interchange depended on the tendency to drop the sound of the voice at the end of the word, as the Germans do now, e.g. unt for und (Brücke, pp. 38. 46. See also below, App. A. vii.). On bs, x, bt, &c. That bs is=ps, not bz, follows from the general law of Latin, that the former of two consonants is made conformable to the latter, and from the fact that s was the sharp hiss. Some instances are found of araps, urps, pleps (Neue, I. p. 137). Compare also scribo, scrip-si, scriptum (cf. § 78). Plutarch writes iepòv övekovévτns for templum obseqventis (Fort. Rom. 10). Similarly is for ks, not gz. Compare rego, rexi, rectum. Reg-si first becomes rec-si, then is written rexi. So also obtulit was pronounced optulit: optimus is for ob-timus, (see Quint. I. 7. 7). And usually with the prepositions in composition, we shall be justified in thinking that, though MSS. and inscriptions vary much in their spelling, the assimilation, entire or partial, was expressed in pronouncing; the spelling, as is natural, oscillating between the claims of etymology and sound; e.g. accedo, attulit, affero, &c. On a before gutturals; gn. The pronunciation of n as ng before a guttural (c, g, qu) is clear from Nigidius Figulus, ap. Gell. XIX. 4. 7. No mention is made of the absorption of the g. And in the Greek to which it is compared the y is written twice, ἄγγελος. Gn is (or was) in Germany, I believe, pronounced like ng+n, i.e. dignus is sounded ding-nus. In Italian and French it is like ny in dīn-yus. There appears to be no allusion to such pronunciations in any of the Latin writers, although they frequently discuss ng. This seems decisive against the above-named pronunciations of gn, at least in the absence of any other evidence for them. (See Schneider, Lat. Gr. 1. 272; Corssen, II. 262, ed. 2.) |