guttural for a palatal (dic=dik, for dice = diche) may be paralleled from Sanskrit as now pronounced, but the change of sound is marked by a change of letter, and the palatal letters are not dependent for their sound on one vowel rather than another. But in classical Latin the change supposed is not justified, so far as I know, by any analogy. Changes of consonantal sounds are frequent, but they are rarely caused by any change of the subsequent vowel: and the change of sound is frequently shown by a change of the spelling, e.g. in veh-ere, vec-tum, which is the nearest analogy that I know. 2. The letter c was used in early times in words which were afterwards spelt, some with c, others with g; and some instances of this use remain in early inscriptions (see §§ 56, 104). Whether these words were at the time pronounced with the flat guttural, or whether the sharp and flat guttural were not clearly distinguished (cf. App. A. vii.), it is not easy to say. But k was also in use, and is found in a few inscriptions, generally before a, but also before o, and (in one inscription regarded on this account by Mommsen as Græcising) before e; e.g. kalendas, korano, dekembres; and it was the regular abbreviation for the prænomen Kæso and for kalendas (§ 103). It is not likely that if c before e and i was pronounced otherwise than before a, o, and u, no attempt should have been made to retain k for the guttural. Yet such an idea does not appear to have occurred to any of the reformers of Latin orthography-neither to Accius nor to Lucilius nor to Claudius Cæsar, in the name of each of whom (see however § 946 n.) c occurs before one of these supposed influential vowels. Quintilian (1. 7, § 10) speaks of the desire on the part of some grammarians to write k before a, (not before o and u also,) but his remark on this seems clearly to imply that c had but one sound. "k quidem in nullis verbis utendum puto, nisi quæ significat, etiam ut sola ponatur. Hoc eo non omisi, quod quidam eam, quotiens a sequatur, necessariam credunt, cum sit c littera quæ ad omnes vocales vim suam proferat." 'k should not in my opinion be used in any word except in those for which it can stand by itself as an abbreviation. I mention this because of the opinion of some persons that must be used if the vowel a follow it, though c is a letter the sound of which is heard before all vowels.' 3. But with these facts must be considered, in order that their full force may be seen, the fact that there is no hint in any ancient writer whatever of c having more than one sound, since the early times mentioned in the last paragraph (Schneider, Lat. Gr. I. 244, 247; Corssen, Aussprache, 1. 48). And this is the more remarkable, because there are many parts of their writings in which such a variety of sound could hardly help being noticed, if it had existed. For instance Quintilian (1. 4, §§ 7—9) first refers to the discussion of the grammarians whether the Romans lacked some necessary letters, and then to the counter question whether some were superfluous, and speaks of k and q. In 7 § 28 he is speaking expressly of what is written one way and pronounced another, and instances this very letter c as used to denote Gnæus (cf. infr. § 104). Terentianus Maurus (who is generally thought to have lived at end of 3rd century p. Chr.) referring to the fact that the names of the three letters c, k, q contained each a different vowel (ce, ka, qu; comp. App. A. xxiii.), says expressly, as I understand him, that k and q are alike in sound and are both superfluous, because it matters not whether c, k, or q be used, whichever of the vowels follow (vv. 204-209)1. See also Diomed. pp. 423, 424, ed. Keil; Priscian Inst. I. 14. 17; pp. 12, 13, ed. Hertz; Servius, p. 422 ed. Keil; Pompeius, V. 110 ed. Keil; Max. Vict. p. 1945, Putsche; and others quoted in Schneider, Lat. Gr. I. p. 292 sqq. 4. c is invariably represented in Greek transliteration by k, be the vowel that follows what it may; and is invariably represented by Latin c2. Now Greek has never been, and is not either 1 The lines stand thus in Lachmann's edition, but the whole passage, beginning at v. 85, should be read: i. e. k perspicuum est littera quod vacare possit; et q similis, namque eadem vis in utraque est ; Whatever vowels you please to utter after forming the guttural contact for C, you must change accordingly the last part of the sound (i.e. the vowel part of the syllable ca, co, ce &c.), but it matters not whether the former part (i. e. the consonant) be k or q or c. 2 Except possibly in a few early words, the spelling of which may be accounted for from c being once the common sign of both the sharp and flat guttural. palatalised or assibilated before any vowel, but is the sharp guttural mute1. Against this argument it may be urged that as the Latin c coincided in sound with κ before a, o, u, it was only natural for the Greeks to use κ for c before e and i, unless the sound before e or i was clearly different from the sound of κ and was readily expressible by some other Greek letter2. Now the actual sounds given to c before e or i in words derived from Latin are (1) Engl. ch (=tsh) by the Italians and Wallachians. (2) Engl. th (sharp) by the Spaniards. (3) s (sharp) by the other Romance peoples (and the English). (4) The Germans pronounce it in Latin words as ts. Further it may be argued on physiological grounds that it may have been sounded as ky, or Germ. ch, or sh; these being possible mediating sounds between the sharp guttural mute and the various existing sounds of Latin c. (See v. Raumer, Gesam. Schriften, PP. 40—43, 90—95; Schuchardt, 1. 164; Ellis, p. 204, quoted in App. A. xxv.; Max Müller in Academy for Feb. 15, 1871.) Could these sounds have been represented in Greek? The sound of s could easily and accurately have been expressed by Greek σ. sh could be expressed by either σ, σo or σi (cf. Mullach, Gram. d. Griech. Vulgarsprache, p. 115). th (sharp) would be expressed far more nearly by σ than by κ. The sound of sharp th is now expressed in modern Greek by 0, but it is not clear when first obtained this sound. ts could easily be expressed by Tσ or τ (see below). I regard this value for Latin c, until at least some very late period, as utterly inadmissible. No combination was so thoroughly alien to the Romans, who never tolerated a dental mute before a sibilant in the 1 The Tzakonians say Te for kaι (see below, p. xk Mr D. Bikelas (in the Academy for 15 March, 1871) says, 'in many of the Greek 'islands & is pronounced like Italian c before the vowels e, ‹, v.' 2 Prof. Max Müller says: Unless we admit that c in Cicero was 'pronounced either exactly like for exactly like σ—and this nobody 'maintains-nothing remained to the Greeks but to use k as the nearest 'approach to the modified c.' Surely this is going too far. He himself explains the fact that the Germans wrote z or tz for c, as proving, not that z or tz was the exact pronunciation of c, but that they came nearer to c than did the Germ. k, or ch. (Academy, p. 146.) same word. (Etsi is of course two words.) Nor did the Greeks either. Χ Germ. ch is a sound which, so far as I know, has never yet been actually proposed as a value of Latin c before e and i. In modern Greek x expresses it exactly, but I do not think had this sound, at any rate till late Imperial times (cf. Curtius, Gr. Etym. p. 371, ed. 2). It is enough for the present to wait till some spark of evidence for such a sound is produced. It can never be a formidable claimant. Engl. ch=tsh was expressed in Greek by τ by Procopius in the sixth century p. Chr. (in the word τČovρovλóv, now Tchorlu, and others in Benseler's Lexicon), and probably in the Ravenna documents of the same time, e.g. aktČIO, DovatČloves, for actio, donationes (Corssen, I. 65 sq. Ellis, p. 529). So in modern Greek T is used to represent either ts, or sh, or tsh (Engl. ch) or zh, i.e. French j (Mullach, p. 115). Compare the Tzakonian dialect, Mullach, p. 94 sqq. M. Schmidt in Curtius Studien, III. 349. Prof. Max Muller objects to the supposition of having been possible, ‘because § was 'looked upon as a double consonant, and in the middle of a word 'would have made a preceding short vowel long.' This argument is no doubt good in reference to verse in the Augustan age: I am not sure of its being applicable to prose even then, if ci had really been sounded as chi, and I believe it has little or no weight as applied to transliteration in the 2nd or 3rd century, when yet к represented c. (See Prof. Munro's account of an Algerian inscription in Donaldson's Varronianus, p. 522, ed. 3; Mullach, p. 71; Luc. Müller's 2nd Appendix to his De re metrica.) But is not the prosodiacal argument as good against the supposition of ci being = tshi, as it is against its being represented by ¿? (cf. v. Raumer, p. 40); and is there any trace whatever of a tendency, at a time when quantity was felt, to make the first syllable in e. g. cecidi, long? There remains one theoretical sound for ce, viz. kye. Here it is necessary to discriminate. It is possible I believe to articulate ke at the same part of the mouth as ka, but neither English nor Germans nor, so far as I know, any other European people do so. ke is palatal and ka is guttural, but the difference is imperceptible. the real question is, had Latin ke either a full y sound or a slight y sound, such as is sometimes heard in Engl. kind, card? Mr Ellis But several times (e.g. p. 525, comp. 204) suggests that it had, but he nowhere defines the time to which he is referring, and he seems to think the distinction of ke and kye is too slight for us to rely upon its being noticed. I can only say that the distinction is one which seems to me obvious enough, far more obvious than many which I find noticed by Roman grammarians; and neither my ear nor tongue is acute enough to find or make any clear distinction between sounds which Ellis discriminates, viz, a palatalised k (as heard in the occasional pronunciation of kind, &c.) and a full ky. may, if the distinction was not obvious, surely we ourselves about it; if it was, then would not the Greek κ have been a tolerable representative? Yet no Greek gives us κinvσwp for censor, or Kukepov for Cicero. But be that as it need not trouble 5. Latin c was represented by Gothic k, and the early Latin words, received into High German, were all spelt with a k, whatever vowel followed; e.g. Cæsar, Kaiser; carcer, Goth, karkara, Germ. kerker. Later adoptions into German were spelt differently, e.g. census, Germ. zins; cancelli, Germ. chanzella, &c. (Prof. Max Müller accounts for this as due to the early poverty of the German alphabet, not to the identity or similarity of the sounds; and as regards Gothic, partly to this cause, partly to a (supposed) habit of taking letter for letter without regard to distinctions of sound, partly to the possibility of Ulfilas having received the words through the Greek.) But the argument most pressed, for c having sometimes a different sound from k, is the confusion which existed between ci before a vowel and ti before a vowel. Now first, whatever force there may be in this argument, it is one which cannot justify our attributing an altered sound of c to ce, ci, &c. when before a consonant. Secondly, it seems tolerably clear (Corssen, I. 50-67) that many instances of the miswriting are due to the confusion not of two sounds but of two distinct suffixes -cio, tio; and that there is no probable instance of ti for ci before at least the end of the 4th century p. Chr.; and only seven instances of ci for ti in inscriptions before the 7th century p. Chr. Further, of these seven instances, three 1 Corssen points out (II. p. 1003) that Mommsen speaks to the same purport (Liy. Cod. Veron. p. 175). Numquam in libro Vero |