l)V77 suiim Cornell University Library The original of tliis book is in tine Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31 924031 1 80783 ELEMENTS LATIN PRONUNCIATION, USE OF STUDENTS LANGUAGE, LAW, MEDICINE, ZOOLOGY, BOTANY, AND THE SCIENCES GENERALLY IN "WHICH LATIN WORDS ARE USED: Jf'MALI BY SrS^ALDEMAN, A. M., PSOFESSOR OF NATCSAL HISTOBT IN THE UHIVEKSITT OF PENNSTLVAKIA. PHILADELPHIA: LIPPINCOTT, GRAMBO AND CO., SUCCESSORS TO GRIGG, ELLIOT &. CO. 1851. s^ Entered according to act of Congress, in the year 1851, by LIPPINCOTT, GRAMBO AND CO., in the Cleric's Office of the District Court for the Eastern District of Penn- sylvania. PHILADELPHIA : T. K. AWD P. G. COLLINS, PKINTERS. PRELIMINARY REMARKS. HIO ENIM TJStfS EST LITERARTJM, UT CUSTODIANT VOCES ET VELTJT DEPOSITUM EEDDANT LEQENTIBUS : ITAQXJE ID EXPKI- MERE DEBENT, QUOD DICTURI SUMUS. — QXJIISrOT. INST. OR. I. VII. The use of letters is to preserve vocal sounds, and, as it were, return the deposit to the reader : therefore they should express what we have to say. In making some inquiries into the phonetic peculiarities of the aboriginal languages of North America, I found myself at a loss, from the want of an alphabet in which to record my results, those of Europe being more or less corrupt; and finding the statements respecting the Latin alphabet to a certain extent contradictory and unsatisfactory, I resolved to investigate it, with the intention of using it strictly according to its Latin signification, as far as this could be ascertained. This special inquiry being made, a view of the results is here presented. Pronunciation is the basis of philology, and without a know- ledge of it, in examining the various writings likely to be used for philological purposes, little progress can be made in this science. It is of little use to show a person unacquainted with Arabic and Greek characters, that Idmistry is derived from L;,t^ and not from jiDoj or to inform a pupil that the South English wordi plow, is derived from a North English word, written [plough] with six characters, if he does not know what words these characters are intended to represent. If the learner has better success in Latin words, it may arise from an acquaintance with many of the characters, as P, P, B, D, ii, T, if they happen to be used in writing his vernacular; but he 4 PRELIMINARY REMARKS. may be deceived if he fancies that similar characters must indi- cate similar words, as in the case of the Latin words mare, miles, and the English ones mare, miles. The materials upon which this work is founded, are as follows : 1. The ancient grammarians and their 2. Modern commentators. 3. Ancient false orthography. 4. Natural relation of the elements. 5. Interchange of the elements. 6. Ancient words transmitted pure. 7. Names of places transmitted pure. 8. Oriental etymologies. 9. Keltic etymologies. 10. The powers of the alphabet among those nations who adapted their spelling to the successive changes of their language. A comparison of such materials ought to produce trustworthy results, because, an error which might arise under an individual head, will be likely to be exposed under some of the others. Of these sources, not more than half are usually consulted by writers on the subject. Under the second head, Schneider's Elements of the Latin Language (Elementarlehre der lateinischen Spraohe, Berlin, 1819) is the most valuable, four-hundred pages being devoted to pro- nunciation, a subject to which about a page is often given, which prevents it from being acquired, except from such professors of the language as have studied it. Dr. Eapp devotes 56 pages to this subject, in his Versuch einer Physiologie der Sprache, Stuttgart, 1836. This work is useful upon the interchange of the vowels; and upon the nasal vowels. Justus Lipsius discusses the subject pretty fully, in his work DB recta pronvnciatione latin.® LINGVO, Antverpi.^!, 1586. The chief ancient authors to be consulted upon Latin pronun- ciation are Cicero, Quinctilianus, Marius Victorinus, Terentianus Maurus, Terentius Scaurus, Velius Longus, Priscianus, and Dona- tua. Schneider quotes fifty ancient authors, upon various points PEELIMINARY KEMAEKS. 5 of pronunciation; probably not one of whom was consulted by John Walker* My results usually agree with those of my predecessors, and when they do not, the adverse opinions are given, so that the reader may exercise his own judgment upon them. Deceived by the title, I have procured several English works upon Latin or Greek "pronunciation," which do not contain a word upon the subject. Among the new views, will be found the table of the alphabet, (§ 35, note 36a') — an explanation of the Grreek phi as the cognate of the digamma and Spanish B — the double nature of H in certain positions — the power of the vowel character pro- posed by Claudius — an additional argument (§ 224, note 58), enforcing Quintilian's view of the Greek Zeta — a refutation of the English J (§230) — the deductions generally from the natural relations of the elements, and from comparative philology; and many of the illustrations, both Latin and transmontane. The alphabet of no modern language corresponds exactly with that of the Latin; although there is a greater or less resemblance, where there has been an endeavor to preserve the characters with the powers they have always had, and should always retain in every language using the Roman alphabet. Latin is often read as if the logographs were Italian or German, and with some show of reason, because the German and Latin letters generally agree, and the full open vowels of the Italian have doubtless been transmitted pure, whilst the elision of sylla- bles in Latin poetry has its counterpart in Italian versification. (See the first note.) In the dipthongs and nasal vowels, the afSnities are greatest between Latin and Portuguese. To what extent that can be considered Latin, which a Eoman would not be able to comprehend, can be judged from the so-called French reading of a German, who would pronounce the French word written [poche] like the very dissimilar German word spelt [^poche]. This certainly would not be French. An Italian would * C. Kraitsir has published a useful little work on the " Significance of the Alphabet;" but it was found more difficult to get it from Boston, U. S., than the volumes of Lipsius, Cellarius, and Manutius from Europe. 1* 6 PEELIMINART REMARKS. be equally in fault, in pronouncing the French words qui est dif- ferently from his own chi e. An English boy might be inclined to smile at the Latin name SCiPio, because he fancies that it should agree with his dog's name Sipio; but the discrepancy is not due to the Latin, but to a defective education, which leads him to write the English name Sipio, with the Eoman cay, although he rejects it when writing slave and slander. Another difficulty arises from a queer association of ideas with what seem to be familiar words, as in the Latin word for praise, when the final consonant is preserved pure; but such cases must occur under every system of pronunciation. The existing materials upon Latin pronunciation are sufficiently explicit to teach it better than French can be taught by books alone without the aid of oral instruction; but they have been so effectually perverted and kept out of view by the authors of spu- rious grammars, that we may meet with respectable teachers of what is by courtesy called Latin, who are not aware of their ex- istence. Those who assert that the pronunciation of this language cannot be ascertained might be deemed honest in their opinions, were the proper length of syllables attended to, this being well ascertained, and the basis of Latin poetry; but spurious rules un- known to the Latin grammarians, have been foisted into poetry as well as prose. (See the second note.) Some are inclined to reject what is incorrect, but they find bad habits too firmly fixed, although they alter their use of the ver- nacular, or of a modern language, from day to day when they find themselves in error. If any who are already educated possess a false pronunciation in any language, this should not prevent those still to be educated from acquiring a correct one. In fact, this outline is intended for the learner, for professional students, for such as have occasion to quote sentences or words, and for the use of schools of both sexes where Latin is not taught, but where attention is paid to Etymology, Zoology, Botany, and the sciences generally in which Latin words are freely used. Although not intended for the proficient, who may be presumed PRELIMINARY REMARKS. 7 to be acquainted with the subject, some quotations are given in the original, to save the trouble of a further reference. Illustrations from the Greek have been sparingly introduced, because it is not usual to represent this language in Roman cha- racters, and G-reek characters would be of no use to many readers. On this account, when Greek words are quoted, they are generally written with the Roman alphabet, or such characters have been selected from the various forms left to us in Greek inscriptions, as most resemble their Roman analogues. This will account for the preference of the Greek characters, S to S, R to P, e to E, and o to a. When illustrations are taken from other languages, they are usually printed orthographically in italics, and phonetically in Latin characters, although in some cases the two do not differ. Italic characters are also used where the pronunciation was doubt- ful, or where it could not be represented by the Latin alphabet, as the Irisli word for silver (§ 258). For the same reason, illus- trative words containing the vowels in fall, not, and the Oriental cerebrals, &c., had to be avoided. In an elementary work on etymology, which the author is preparing, farther illustrations will be found upon, subjects touched upon in this volume. The Oriental etymologies are not expected to have weight with those who consider them "fanciful," or with those who prefer the superficial Richardson to the philosophic Webster, whose chief defect is his sparing use of "Oriental analogies." An apology ought perhaps to be made for employing etymology at all, since in a recent conversation in a railway car with a student who had just graduated with honor in an American college, he stated that he could perceive no etymological connection between words like the Greek GeRANos and the English CRANe, or between the same GeRanOS and the Latin GRUS. In the following pages the word diphthong is written so as to indicate the pronunciation approved by Mr. Smart; and /;; is used in writing '' Celtic," because the initial of the word intended to be used is not a sibilant. Columbia, Pa., Sept. 1850. INTRODUCTION.* § 1. When a student commences the study of a language with the aid of books, his first inquiry has reference to the power of the characters which represent the words. 2. A character is an arbitrary mark, meaning nothing until it has been assigned to a certain sound made use of in speech. For example, 3. The character H cannot be correctly referred to a sound until we know the alphabet of which it forms a part. In Greek it is a vowel identical with the Roman E ; in Russian it represents the N of the Roman alphabet ; and in Ethiopic it is equivalent to the French or English syllable za. 4. Before pronouncing a written word, therefore, we must know to what language it belongs, or wb may read a Greek trissyllable (APBTH) as a dissyllable in Roman characters, which would make apeth out of A-ee-te ; an error which has a strict parallel in the practice of reading Latin as if the letters were those of a transmontane vernacular. Hence a Russian cannot with pro- priety knowingly confound the H and N, a German the Z and C, or F and V; nor an Englishman the G and J, or C and S of the Latin alphabet. 5. Latin being spoken to a considerable extent among the learned, particularly between the residents of different countries ; grammars which profess to teach it, as they must be drawn from the same original source, should correspond in every particular, • The asterisks which precede the numbering of some of the paragraphs, refer to notes in the concluding pages bearing similar numbers. 10 INTEODTICTION. recommending a uniform mode of pronunciation, ■whether printed at CaHKmnemep6yprb, S^_ J^J' or J^ ^ . 6. Latin is called a dead language, and on this account(!) some pretend that they are at liberty to give it the sounds which hap- pen to he represented by similar characters in their own vernacu- lar ; a practice which would result in as many jargons as there are perversions of the Eoman alphabet. 7. By mispronunciation, much of the value of Latin is lost to etymology and general philology, both of which depend, to a great extent, upon the accuracy with which words can be recalled by the aid of appropriate characters. 8. The use of Latin for philological and conventional purposes renders a uniform conventional pronunciation necessary when it is brought to life in oral discourse. 9. This would prevent ambiguity between certain words; as STJROtTLtfs a twig, ciRCULtJS a circle; SEDO to calm, CEDO to yield; sosndo to climh, XSOENDO to mount; SCfiLESTI wicked, coELESTi heavenly; silicem a. flint, ciLiCEM of Cilicla; caelum a chisel, COELUM heaven; iNGESSi I have carried into, iNJEOi I have thrown; and many others. 10. Many languages, as the Russian, Armenian, Georgian, Arabic, Greek, &c., are not written in the nomau character, a fact overlooked by such grammarians as confine their superficial directions to those whose vernacular is supposed to be represented by the Latin alphabet; without giving a Greek or Arab any idea of the subject. 11. Having a rule before him which requires the imposition of vernacular barbarisms, and even forbids a uniform mode of pronunciation, the Russian cannot do otherwise than turn his Latin B into v or /, and Latin H into n, as if there were neither 6 nor h in the language; whilst 12. The Greek is forbidden to pronounce XERXES as he usually does, namely CSerCSes, although he is told that the Roman X is OS. Nor must he make Oh identical with his own Chi, although he knows this combination was made expressly for it. 13. Philological relations were not taken into consideration by those who, instead of endeavoring to ascertain the true power of INTRODUCTION. 11 tte alphabetic characters, fancied that they must be identical with such as resemble them in their own alphabet; confounding the Russian C and Armenian U (Roman S), or Coptic T (pa), or English J, or French J, or Cherokee J (gu, § 240), with the Roman characters of the same form. 14. Whilst some of the Latin characters have been corrupted, they are preserved pure as Greek letters, the initial of the logo- graph of OyDoN being pronounced correctly when it is considered Greek, but confounded with that of siDoN when spoken of in a Latin connection. 15. Probably no one acquainted with the subject pronounces the character C differently in the Anglosaxon (a dead language) from the power it still bears in Gaelic and Welsh, or the word hirk would cease to resemble its original cyrc. So the Welsh and Scotch word cist, the Irish cisde, the German and Danish Mste, Swedish hiMa and Arabic his, correspond with the Greek xii-tti and the Latin ciSTX. In English, these have been deve- loped into Ulmt, and its literary corruption tshest. 16. The English word poop (of a ship) is the first syllable of the Latin word puppis; the Swedish, English and Dutch word hinh is the first syllable of the Latin word cinctum; the English word croak is the root precisely of the Latin word CRoCiO; and the German term for emperor [kaiser] differs little from its original caesxb, which the Romans also wrote [oaisar]. See note to § 165. *17. It is an important law in the interchange of consonants of different contacts, that a guttural, as h, readily changes to a palatal, as s, but not the reverse. Hence *18. We falsify a fundamental law of philology, if we assert that a word which contains a guttural, has been derived from one whose corresponding letter was a palatal, as Jdst from sisla, canker from canser, the Greek form kiKeRon from the English name Cicero, or the English words Jceep, kin,, kitchen, from the Anglo- saxon cepan, cyn, cycene, if these contain a palatal. No one pre- tends that cover (from convert) in its local form (§ 77) civer, has an initial s; that the biblical Icgographs [Kedron, Eliakim] are 12 INTRODUCTION. not identical with [Gedron, Eliacim], or that the anatomical term sacciform is pronounced saxiform. 19. The reading of Latin should be the successive enunciation of the power of each letter, which would make it strictly phonetic, as it was among the Romans. 20. When in certain words the dipthong au (in hrovm, German braun) was replaced by the vowel o, (he orihograjihy was changed, as in o-LLA, which had been previously aulx, and in suffoco to suffocate, from faux the throat. So when de nouo or de novo became contracted, it was written denuo. *21. That this was a point of pronunciation is proved by Pestus; who also states that the rustics pronounced ORUM for aurum. In some cases the pronunciation was not uniform, as in lautum and LOTUM, caudex and codex, claudo and cludo. 22. Thus we find at the beginning of the 5th century a tend- ency towards a change which has been consummated in French ; but the French still write au instead of the proper character, o, contrary to the correct usage of the Latin alphabet. 23. That the vowel o replaced the dipthong au in a dialect of Italy, is no excuse for the French perversion of the latter, because the Romans did not usually follow the practice of the Greeks in writing their dialects. Cicero, however, wrote as he spoke, when his pronunciation difi^ered from the general standard. §§ 89, 96. Some English authors, as if to justify their Latin cacophony, imagine the English consonant combination dzh might have been known in Italy (§ 230); but whether or not, the fact should have no influence in reading Latin, for the reason just given. 24. Were such perversions proper, we might with the Sabines replace H with F in hIrcus, hordEum, or read an initial H in ADRiANus when it is omitted in the writing. 25. The following rules for reading Spanish (from Cubi's Grammar) are well adapted to Latin reading : " To sound every vowel fully and distinctly, leaving, as it were, the consonants to take care of themselves. Never to pass over the small words, but to pronounce them clearly and distinctly. Never to give a very strong emphasis to any particular word in the sentence ; for, as every word is fully pronounced, there is not much room for particularizing any one with uncommon vehemence." ELEMENTS LATIN PRONUNCIATION. 1. OF THE ALPHABET. *26. An alphabet is a collection of tlie characters representing the vocal elements in a language. The term alphabet is also used as a name for the aggregate of the vocal elements, which are termed letters by Priscian, who, about the year A. D. 525, wrote a voluminous Latin grammar extending to 900 pages. He says that a letter is a sound — the smallest portion of the voice ; and the written characters he considers representations of the letters. Other ancient authors use the word letter as synonymous with character. 27. The Latin alphabet has varied in extent at different times. Some modifications were introduced to assist in representing words taken from the Greek, and it is to be presumed that those who introduced them gave them the proper pronunciation ; because in most cases, when the words had become naturalized, the foreign characters gave way to the Koman ones. *28. The division of characters into capital and small letters is scarcely admissible in Latin typography, and Priscian makes no mention of them in writing. The small letter alphabet being used for its convenience in transcribing, lost its peculiar value when printed. 2 14 ALPHABET. 29. In the more ancient manuscripts, there was not so much dissimilarity as at present between the two kinds of characters, as in the case of the small b, f, g, l, n, r, t, i (without a dot); whilst the small d, h, m, q, u, y, did not differ materially from the modern form. 30. According to Priscian there are 23 characters in the Latin alphabet. These are ABCDEFGHIKLMNOPQK S T V X Y Z. To these the moderns have added J and U. 31. Strictly speaking, C, K, Q, are not distinct Latin letters. Some of the old grammarians do not consider X a letter; some improperly reject H as an aspirate, and others S as a hiss. They might as well have rejected Thi as a lisp, and F as a piiff. *32. The names of the characters are given by the ancient Latin grammarians, except those of Greek origin, and N when it represents its power in the English word anger. 33. The character K (and also Q) is a duplicate of C; X of CS and GS; Z of SD; and x of PS. Some of the characters, as U, K, iS, I, 3c, were rarely used; whilst Ch, Eh, Ph, Th, Y, Zi, are of Greek origin. 34. In the following version of the alphabet the duplicate characters are introduced, as well as the representatives of such Greek sounds as are represented in it, although foreign to the Latin language. The alphabetic order of the Greek equivalents is indicated by the numbers. AUHABET. 15 Iharacteis Name in Latin and English. Greek equivalents. ' A A Ah * A,» ' O ' B Be baj ' B,8,|3 * C Ce cay " K,* as ct Chi ... "^ X,z ' D Db day * A,S ' E B a = H,, long. ' F eF ayf] ' E, £ as in end. ' G Ge gay " r,y,r " H Ha hah 10 • 10 , I e "I,. " J JOTA yota I;' ^ \- .... ...] » » *' K Ca kah 13 K " L eL ail " A, A " M eM aim. « M,,* " N eN aia " N,v IS |\|Gt ayng f-y « o o[o " 0,o IS p Pe pay *» n,rt,ra- («> .... ph Phi ^ *,* *" Q Cu coo 21 O « Fi eE air ^ P,C,p («?..., . Rh Rho ^ «, s eS ace "^ S,5,cJ »a y Te tay i" T,* (-.... TH ThETA thayta " ®,S>,S =» u u 00 OT,8 23 V Ve way 8 « J DIGAMMA /3ou « f,^ X iCS eecs « S,l (»B ... . Y Y =» T,i; (» ... Z SDeta edayta ' z,?,? . oc ANTISIGMA ^ *,4- SADE, sav 6 so -Tf) 16 VOCAIi ELEMENTS. 85. Scheme of afBnities between the vocal elements in Latin. o - .o \H0 to ot> I tfs i ; " : «> IJ :2Q :K|- Od^ s> :20Q-nQ-ii. ■ Qi t^ O Q> t^ Oi f-* o o o -a 3 P4 >J P ^ EH ;3 I ^ *36. In the foregoing scheme the consonants of the respective contacts are represented in vertical columns, whilst those having certain qualities in common, are indicated in the transverse lines. It will be found of service in studying inflections and euphonic VOWELS. 17 changes, not only of the Latin, but of other languages when the letters belonging to them are properly distributed in it. 37. Elements or sounds of the same contact interchange most readily; and after them, those of adjoining ones. For example, B being the liquid of the palatal contact, is readily interchange- able with s, as in the double forms xrbor abbos; honor honos. So OS becomes oris; aes aerIs; haerEo haesi; but aestas makes aestxtis ; uro iissi become ustum, taking the next con- tact, T and s being made nearly at the same point. A similar law appears in the connection between tubEr a hnoh, &c., and tumor a swelling; m being a nasal b, as n is a nasal d. The dropping of M in rumpo to form Rupi, is paralleled in SOiNDO, SClDi. 2. OF THE VOWELS. 38. The Latin vowels are either long (marked ") or short (marked "), the former being double the length of the latter, according to the ancient grammarians. *39. The power of the Latin vowel characters is heard in pro- nouncing the following English logographs, or written words, in which they appear. A is loTug in Arm, s\ort in Art. B " thEy " Eight. I " marine " deceit. O « Own " Obey. U " fOOl " fUll. 40. "The sound of the long vowels was that of the short vowels doubled." — G. Walker, in Scheller's Latin Grammar. "The sound of the long and short vowels, though elementarily the same, were always distinguished in length." — Scheller. 41. In Latin it is rather the syllable than the vowel which ia long or short, or the subject of quantity. 42. Two consonant characters (excepting h) following a vowel character, usually make a syllable long " by position." Words like coNSuMO to' wasU, and consummo to accom;plish ; vita life, 2* 18 VOWELS, ACCENT. and viTTA a band, are distinguished by doubling the consonant where its character is doubled. 43. Dipthongs are long. A vowel preceding another is usually short (even when separated by h), as in ohdos chaos, cbviA. plenty, Mini to me. The quantity of vowels, which is long or short at the pleasure of the poets, is called common. 44. In old Latin, instead of the mark of length, the- vowel character was doubled, as in paacem for pacem. Afterwards the succeeding consonant character seems to have been sometimes doubled for the same purpose. *45. A doubled character indicated a long syllable, because every addition increases the time, as in the double forms tantu- Liis, TANTiLLus ; PXLATiNus, PALiiATiNus; although there was possibly but little if any difference in the pronunciation of certain words written with a single or double consonant character, as in the forms LIT ERA and litteea; XPulIx and appulia; balistx and BALLisTA; causa and caussa; numus and nummus; bacca and baca; belliiA and beltja. § 19. 46. In the Teutonic languages a doubled consonant character marks a short vowel, so that there is a tendency to pronounce the first syllable of words like pEnns, &c., as if it were the English syllable pen. This fault should be carefully guarded against. In Latinising Teutonic names, this peculiarity should be rejected. Accent. *47. Vowel characters often have the accent indicated by the grave Q) and acute (') accentuals in typography; when the former indicates a long and the latter a short syllable, as in mJilus an apphtree, and M&Lus had. To apply this mode to English, we would write tarry, st&rry ; nhughty, Tcndlty; slavish, Idvish; profdne,pro/dnity. The acute accentual only is preserved in inscriptions. *48. In dissyllables the penult syllable (or second from the end) is accented, as in cJissis a helmet; cAnis a dog. 49. In polysyllables the antepenult (or third from the end) is accented, unless the penult contains a long vowel, when that is accented. VOWELS I, E. 19 I 50. The most contracted of the primary vowels, that formed with the -shortest tube when the vowels are made mechanically, is I. It is heard in the English words marine, field, and is repre- sented throughout Europe, and in general alphabets, by the Roman character. Its short quantity is heard in the English words deceit, feet, equal. " The English &eat retains the Roman pronunciation of situs, that is, seeius." — Webster's Dictionary. 51. Some give the secondary English and German vowel in Jin, Jit, pity, as the "short" quantity of I, a sound which has a different quality, and is unknown in French. Moreover, " every letter retained an invariable sound." — G. Walker. 52. Victorinus describes the vowel I as being made with the mouth nearly closed. It was identical with the Greek I, which "was sounded like the e in mste. The modern Greeks so pro- nounce it : and here again the English, in differing from the modern Greeks, differ from all the nations of Europe." — Pen- nington on the Pronunciation of Greek, p. 36. 53. A vowel being a simple sound, the Roman I would not be one if its power were that of English i {ai, in aisle) this being a dipthong or double sound. 54. The second Latin vowel is heard when long in obedio, obey ; vEnx, vein; VERBENX, vervain; and when short, in the English words freight, hate, eight, weight. It is long in SECius less, and short in sEcitjs otherwise. 55. The natural position of E being between I and A, it shows its affinities to each, as in tEnEo, eEtinEo; lego, BiLlGOj BABBX, IMBERBIS; XPTUS, INEPTUS; CiRPO, DiSO-ERPO; and in the double forms sivE, sev; alxmani, alemani; Xlex- ANDRlx, alexandrea; herculius upon coins, herculEus in stone. In English, a similar relation appears in dbsy, obidient. 56. Varro considers E the vowel in the cry of the sheep (bee), so that it cannot be the English e. The character E is recognized with the Latin power in Europe, and when it is dropped for I, 20 VOWELS E, A. the character changes with the sound. Thus the Latin SEOuRus has become sicuro in Italian, as creatura and sllEvo have become criatura and alliviar in Spanish. 57. As the short and long sound of E differ only in length, they readily flow into each other, as in prEhendo, vEhEmens, which, by dropping H, take the form prendo, vemens. *68. The power of I and A being determined, there is no character left for the vowel in vein except E. 59. The natural order of the primary vowels, as determined mechanically, is I E A U or U A E I, as heard in the English words field, vein, far, owe, ooze. This order should be so well impressed upon the memory that the vowels may be repeated fluently in either direction, as it will be found useful in studying the inflections of words ; and on this account the ele- ments are here treated according to their affinities. 60. The fundamental vowels are the guttural I (in field), the palatal. A (in far), and the labial U (in ooze). The closeness of aperture in I and U approximates them to the nearest consonants, into which they are apt to fall, the first into the liquid or semi- vowel of the guttural, and the last to that of the labial contact. E and exhibit in a less degree the tendency to become conso- nants ; whilst A, from its openness, and its want of relation to the extreme vowels, is farthest removed from the consonants, and is consequently the type and most noble of the vowelsj A 61. As with all the vowels, the power of A is its name. 62. The almost universal power of the first character of the Roman alphabet is heard when long in ear a kind of wheat; FAS right, and in the Latin and Italian word amo. It is short in XRAR the river Saone, in the final of armx, arX, and in the English word art. 63. The French §, (as in S.me) approaches A, but is not so open, that is, the mouth is less open in its formation. Standing in the middle of the vowel series, A is the most open of the whole, and its use gives great power to Italian vocal music. It is equi- valent to the Grreek A, which Dionysius of Halicarnassus properly VOWELS A, 0. 21 calls the most agreeable of the long vowels, and which is made, as he informs us, with " the mouth as much opened as possible." — Pennington, p. 28. "It seems clear from the description of Dionysius, that this letter wa& pronounced as we sound the A in father. The modern Greeks so sound it, as do most, if not all, the other nations of Europe. Our English mode of pronouncing the Greek A is peculiarly unfortunate, excluding the very sound which Dionysius thought the most agreeable." — Penningtcm, p. 34. 64. The descriptions of the Latin authors agree with the above, as in the expression EiCTtr paxulo of Victorianus Afer and Te- rantianus Maurus; and in the hiatu oris of Marcianus Capella. 65. There is no evidence to show that in becoming short, the quality of any vowel varied, as in that case it would be a different vowel. §§40,51. Theauthor of "Living Latin" (London, 1847) says, p. 11 — "That the Latin vowels have only one sound each, long or short, is clear from Priscian, who, when he would enu- merate the varieties of sound which they admit, mentions only those of accent and aspiration, which are merely varieties of its accidents, not of the sound itself." Yet this author, instead of perceiving that the a in the English words art, Cairo, kite (Jeait') is short when compared with that in aye, ah, car, gives a vowel scarcely known except in English j namely, that in faf, which, so far from having a quantitative relation to the vowel in arm, has both a long and short quantity of its own, as in Welsh, where bach [ch as in German) means little, and l&ch a hooh. o 66. The Latin is heard when long in the English words owe, moan, lo; and when short, in ohey, ocean, note, invoice (voco). It is found in the Latin words 5!, oMfiN, sol (the sun, preserved in Swedish and pronounced like the English soul or sole) ooEXNiiSj OCOllBO, OCTO, OTioSUS, ORO, ORNO, OS, OBNOXiiiS, QUOMODO, HODlE, JOViS, JOCOR, OLOR, ODOR, OBORioR, OB, ODIUM, KOTX, DEMOPHoSn son of THESEUS. It is long in morari to he foolish, and short in m8rari to delay; long in otcl5ps and short in cecrops. It is found in the Spanish words seiior, Colorado, and in the German jpol, lob, los. 22 VOWELS 0, U. 67. The round form of the character (0) was intended to picture the lips in forming the sound ; a form which does not accompany the vowel in nor, not, which is less common. "Our Towel in fond occurs but seldom, if ever, in Arabian, Indian, or Persian words." — Jones, Asiatic Researclies, 1, 15. 68. When long, the Latin agrees with the Greek g, in form- ing which, according to Dionysius of Halicarnassus, "the mouth is rounded and the lips disposed in a circle, and the breath strikes upon the extremity of the lips." — Pennington, p. 29. *69. and U being labial vowels, if the organs commence closing, or assuming their quiescent state before the vooality ceases, a labial dipthong will be formed with each, as guttural ones are formed with I and E under similar circumstances. This has induced the author of "Living Latin" incorrectly to consider the ordinary a dipthong, and as the Latin requires it to be a vowel, he replaces it with the power in all, nor. The expression of Victorinus answers better to the ordinary 0. 70. The interchange of and U shows a greater afSnity be- tween the two than would be the case with awe, which has a greater affinity with A. Thus njvibus, consul, effugiunt, stand in earlier inscriptions navebos, cosol, exfociont. So we have the two forms oPiLio, uPiLio; PyLos, pyLus; volsella, VULSELLA, tweezers. The Greek proper name HEKABE became HECOBA in old Latin, and finally hEcubx, passing through A, o, u, of the natural vowel scale. 71. In some parts of Italy did not exist, its place being supplied by U; in other parts U was wanting and replaced by O. — Priscian. The of tolosa has become U in its modern name Toulouse ; and mutina has become Modena. u 72. The Latin U is long in the English words jiooZ, cool, room; arfd in the Latin words RtiMOR, Kfts, CKus, LflNA, lux, sfis, Mus, suppLEX. It is short in the English words pull, full, root; and in the Latin words RuiNA, Liipus, supebbus, TuMuLiis, Rubicon. *73. The Latin U is recognized with its proper sound and character throughout Europe ; and the position of the organs in VOWELS U, Y. 23 forming it is well described by Oapella. It is preserved pure in the following geographical names ;— Anamour from ANEMtranrM. Tersoos " tarsus. Courtenay " cuRTiNrACtrM. 74. The character (U) is angular in inscriptions (V) ; and in old printing the two forms are used indiscriminately, as in laetjus, vsuRPATUR, vsTJM, VT, ACVTVS, Diuisio, QviNQVE. The more common, but not universal practice of the present day, is to limit the rounded character to the vowel power. 75. U and O being nearly allied, are interchangeable, as in the old form EPISTTJLA, ADUIiBSOENS ; of EPiSTOLA, ADOLESCENS; and in humu, used by Varro for humo. The same law appears in the changes to which the English words gold, move, Home, door, floor, have been subject. 76. U and I (completing the circle of the primary vowels) are interchangeable, as in famulus, eXmilia; simul, similis; ex- suLO, EXSiLiiiM; CONSULO, CONSILIUM; in the old Latin of Scipio's tomb; PLOIRUME for plurimi; in the double forms Hie, Hue Aere; iLLio, iLLue there; libEt, lubet; libentEr, ltibentEr; and in tEgumEn, tegimen, which became tegmen. 77. The relation between U and I being organic, their inter- change is common, as in the English words food feed; brood breed; blood (formerly) bleed; flew flee; you ye; thou (where it has not become obsolete) thee; foot feet; rood reed; leward lee. 78. The English corruptions of you for U, and eye for I, whilst they are disproved by this law, tend greatly to mystify the student who wishes to understand the genius of the Latin language. 79. There was a tendency to elide V, as in the change from audivisti, mavelim, nevolo ; to audisti, malim, nolo. Y *80. Dionysiua describes this Greek vowel as pinched or com- pressed, and that it is a labial appears from Capella'S' description^ which assimilates it to the French v. or German u, with which the scholars of these nations consider it identical. Pennington is of 24 VOWEL Y. the same opinion, and cites Mr. E. P. Knight as follows: "Per- haps the nearest letter to it in modern alphabets is the French accented U, the sound of which is indeed poor and slender ; but such as Dionysius informs us that the Greek T was." It is both long and short, as in the na»e of HypsiPyLE, a queen of lemnos. 81. From the form of the character T, it appears to have been intended to represent a compressed V (U), and it is correctly preserved for such a sound in Polish, Danish, and Swedish. In modern Greek, as in some dialects of German and French, it has degenerated into I, except in a few words. The sound is unknown to the Italian, Spanish, and Portuguese. 82. To form Y, the organs must take the position for I, the jaw must then be dropped to enlarge the cavity within, and the lips pursed and projected, and made narrower than for U. The resulting sound must resemble U rather more than I. "The un- pleasing sound and the ungraceful position of the lips agree with the description of Dionysius." — Pennington. *83. Terentianus Maurus states that the Latin language wants the Greek T ; and according to Victorinus the Greeks represent U by ». The Y was therefore used by the learned who under- stood Greek, but was replaced by U and I as the words became naturalized. In old Latin it never appeared, until introduced by pedantry — a cause which has had an improper influence with modern transcribers, so that it is often diflGicuIt or impossible to determine the orthography of the ancients. 84. Standing between U and I in the natural alphabet, Y readily falls into one or the other (but chiefly into U), as in the forms STLLA and sullAj amymone and amimone; alcyone and alcione; stmboLa and sumbola; CHYtra and chutra; TyRO, TURO, and tiro a novice; cyMA, CUMA, cima. 85. The change from Y to U is most frequent, as in ToRTiiS, TuNDO, Mus, sus, DUO, cuPRESSus, which are from the Greek j and in fact, in the dipthongs, and in the JEolic dialect, (Y) had the power of (TJ), as in some of the Sclavonic alphabets. On this ac- count, when iY cannot he pronounced, it is best replaced with U, and the preceding consonant would be more likely to be preserved VOWELS Y, I. 25 pure, as in the proper name CyRiis, which, as the name of an Asiatic river, has become Cur or Koor. 86. In fewer cases the Greek r has become I, as in stipEs, SXTIRS, STILUS. Sometimes Y has been improperly placed in regular Latin words, as in siLVX, pensilvaniX, hiEMS, tibo, CLIPEuS, LACRIMX (§ 283), PXPiRlUS, PIRUM. 87. Instead of perceiving that the interchange of allied vowels is organic, Scheller propounds the erroneous opinion that U was pronounced like O (§§ 71, 75) and the Greek T, citing stjlla from SYLLAS as an example of the latter. — Lat. Gfram,., 1, 16. Now it is evident, that being unable to pronounce the Greek sound, it was naturalized by the use of U, as in the case of duo, &c. 88. Scheller, on the strength of the two forms vertex and VORTEX, states that B was pronounced as 0; a view which a foreigner might take of the allied English words vertex and vortex. Cellarius ( Orthogr. Latino) considers the word hiems winter as not of Greek origin; whilst Scheller not only asserts it, but in- sists that it should be spelt [hyems], as if Y could not have changed in so common a word (§ 86). In ancient inscriptions it is spelt with I. I- 89. This character, according to Priscian, Donatus, and Velius Longus, was proposed by the Emperor Claudius for a vowel repre- sented by the characters U and I, but in which they had not their true power. Among the words cited by the ancients as contain- ing it, are maxxjmtjs (preferred by Cicero) or maximus; proxtj- Mtrs or PRoxiMus (and doubtless all superlatives), possmmiis, VOLWMUS, NOLMMTJS; ARTIBUS; MXNIBUS; AURTJFEX or AURl- PEX; MANUBLAE Or MANIBIAE, VIR, AUCUPIUM or AUOIPIUM. 90. The vowel f- was probably that in the English words it, fit, in, pin; a distinct vowel which is long in Sclavonic and Turkish, and whilst it is allied to I, approaches U by being formed with a more open aperture. 91. Priscian states that short I followed by D, T, a, m, x, seems to have the power of Greek T, as in vidEo, vMum, vim, viRTus, vix; but we do notinow the power of T in his day, its 3 26 TOWELS I, u, £, 0. true power being expressly stated not to have occurred in Latin; and it was already provided with a character. 92. Victorin, whilst he mentions those who consider this vowel "thicker" than I and "thinner" than U, recommends the exam- ples in which it occurs to be written and pronounced I, a mode which has prevailed. It is probable that many pronounced these words with pure U or pure I (§§ 76, 77), as Velius Longus states that AVBiTEX sounds better with I and avcupare with U. o 93. This character occurs in an ancient inscription, replacing final A in the word dicata. Its power may have been that of the vowel in further, or the final one in comma, altar. In the volume on Philology of the U. S. Exploring Expedition, Mr. Hale uses this character for this sound. [£] 94. It is not likely that the Greek epsilon occurred in Latin, where it was always replaced by E, as well as the eta (§ 135). This vowel is heard in the English words fen met, and is long and a little more open in the French word mime. 95. The final vowel in hErE or hEri yesterday, according to Quinctilian, was neither pure E nor I. It was probably the dip- thongal sound ej (§§ 133-4), condemned by Velius Longus as a mispronunciation in tJbi (tibej) ; a sound which arose in the Gothic forms AKE JT, avrke js ; borrowed from the Latin words aCeTuM vinegar, and uRCEuS a pitcher. *96. This form is not rare in inscriptions, as in the words Si, uBi, iBi, SINE; which stand SEI, dbei, ibei, sinei; and the termination is occurs as Eis. Cicero is said to have written CIVETS for civis, omneis for omnis, &c., this being probably a dialectic variation. [OJ 97. There is no evidence that the vowel in the English words on, not, and French mol, noce, was found in Latin, and no author asserts that (0) had a second power, even in the dipthong oi (§ 67). NASAL VOWELS. 27 3. OF THE NASAL VOWELS. 98. An examination of the Indo-European languages from a period long anterior to the Greek might induce us to suspect the occurrence of nasal vowels in Latin. Thus we find the Sanscrit originals of the Latin words donum, Sanscrit dana (in Roman characters), antrum (antra) to have a final nasal vowel. 99. The ancient Latin grammarians are sufficiently explicit on the subject of nasal vowels, which they associated with m, as in the Portuguese of the present day, where (hom,) is equivalent to the French (hon). In English, nasal sounds are often associated with ng, as in (hong) for the French (bon). In Polish, a mark somewhat like a comma forms an appendage beneath the character. 100. A nasal vowel, like a nasal consonant, is made by pro- nouncing the letter with the nasal passage open. *101. Prisc-ian makes a distinction between m final, initial, and medial. In the first it is obscure (that is, nasal), in the second with its ordinary power; and when medial, as in umbrx, it probably had its ordinary power, in addition to nasalizing the preceding vowel. 102. Verrius Flaccus indicated the nasality by writing but half the character [m], thus [i J, and it retained its place with so little permanency,' that TIRO, antioco,* have been found in in- scriptions (the final to be probably understood as nasal) for viRTJM, ANTIOOUM or ANTIOChtTM. So VENEo is from VENU"" Eo, and ANiMADVERTO from XNiMu™ ADVERTO. The m in ciEOuM dis- appears in cJROuiTus, and QVAMSI (through qvasi) becomes QVXSI. 103. Manutius copies an inscription (p. 143) in which a small curved line ( ~ ) is used (at least by him) to represent M, N and N (ng), as in the logographs poenv, ivicti, ovctarvm, for POBNTJM, INVICTI, CUNOTAETJM ; SO that there is antique authority for this mode of graphic representation. 104. In Latin manuscripts and printed books, M (and also n) is frequently indicated by a straight or curved line over the pre- * This final has become pure in Italian, wliich is without nasal vowels. 28 NASAL VOWELS. — DIPTHONGS. ceding vowel character;* but this is inconvenient, as such a mark interferes with the placing of the accentuals and marks of quantity. On this account, when nasal vowels are to be illustrated, I adopt as near a modification of the Polish mode of indicating them as ordinary typography affords, a mode which is no novelty in Latin typography.! 105. The Latin nasal vowels are i E A o u, as in ENi"", dKce"", XA", FLovio" (§§ 71, 75), Tti™. 106. The nasality may have been also associated with (n), as in QVOTiES from qvotiens, and the inscriptive forms costjl or COS. for CONSUL, and cojux for conjux. The letter in question is omitted in the first syllable of the Greek form of const ^.NTiNiis and in hortensius. The Latin word conspirare appears in Italian under the two forms conspirare and cospirare. 4. DIPTHONGS. 107. As vowels are distinguished from consonants by the amount of interruption, it may happen that this may be so small as to leave a doubt as to whether the resulting sound is a vowel or consonant, and this really takes place. 108. The extremes of the vowel scale, i and u, have a great aflSnity to the allied consonants or semivowels J, v (English y and w), and readily interchange with them, and their little difference respectively has deceived good grammarians, as in the case of the English word well, which has been asserted to be merely oo-ell, as yard has been considered as e-ard. The syllables woo and ye disprove such views, as they are not repetitions of a single vowel. 109. But a still closer approximation exists in a pair of coa- lescents intermediate to the semivowels and extreme vowels ; and * As in — DAKonu begu hebouthongs are A J, bj, OJ; and perhaps ij, UJ, YJ, by poetical license, ae and ai, by the concurrent testi- mony of the ancient grammarians, had the same power, but [Ai], the older and more correct orthography was allowed to fall into disuse; so that the words ajmilius, qvajstor, ajternus, &c., became aemilius, qvaestor, abternus, perhaps from a jealousy of the Greeks and their literature. (§ 118.) 129. [ai] is used instead of [ae] when poetry requires two syllables, as in teerai frugifErai; affording an argument in favor of the double nature of AE, and proving the inconsistency of the later orthography, [ae] is not employed in GRAIus, Greek, GRAJORU. *130. According to Terentianus [ae] had the power of the Greek Ai; and according to Varro, in the rural word hedus, a was inserted in the towns, making haedus a kid, whence it is 32 DIPTHONOS. evident that ae was a double sound, as in praebEo, aenISus, abbreviations of prakhibeo, aheneus. 131. In Portuguese the Latin orthography is preserved, as in the word pae or pai (father). This dipthong is represented by [ai] in most of the languages of Europe. In French (as in faire), in English (as in fair), and in modern Greek the charac- ters [ai] represent a vowel sound. There was a tendency towards this change in the time of Varro, who asserts that whilst some said FAENUS, others said pentjs. Dialectically [Ai] has its Latin and universal power in French. 132. The few modern authors who consider \_m\ a vowel cha- racter, make it the French t, which is the English vowel in fen lengthened, without becoming E, as in fairy. This word (when properly pronounced) is a lengthened form of ferry. 133. EJ differs from the dipthong in aisle by having the vowel E as an initial. It may be learned by omitting the final vowel from the English syllables lay-ye, which will give the Portuguese monosyllable lei, law. In old Latin the forms navejs, olasejs, were used for naves, clxssis. § 96. 134. EJ is the Latin modification of the classic Greek dipthong £t, which finally became the vowel i. It is found in the poetical forms DEJNDE, AVREJS, PERSE JS, NEREJ. 135. Those literary Komans who pronounced the Greek CM, may have used ej instead of ej in words from the Greek; and as the character [j] is merely a rounded form of [e], and not un- known to Latin typography, there is no objection to it in repre- senting unnaturalized Greek words, by those who believe its power was used by the Romans. 136. OJ has two forms [oe, oi], of which the latter is the more ancient and correct. Both forms are used in Portuguese, as in \^foe, fox\ Tie has been. OJ was interchangeable with u in old Latin, as in coiravit, coeravit, curavit; ojno (ojno) for TJNUM — and with I, as in liberum from loebestjm. 137. The vowel in or being further removed from ir than o is, the change between oj and u indicates that the of the dipthong was pure, and the ancient grammarians say nothing to the con- DIPTH0N6S. 33 trary. It is pure and short in the LsNlPt word (r as sh) rlHXMoJs a crevish or crayfiih. 138. The trissyllable PRo-vi-dens provident, became pro-i- DENS, PROJDENS, PRODENS, PRUDENS, prudent. PROELIUM a skirmish, is a contraction of pro iliijm; and from the compound word co-£o to meet, we have coetus a crowd. 139. Having from a false theory (§ 69) given to o the power in on, the author of "Living Latin'' makes oe rhyme with hoy^ instead of heau-y, or the first syllable of co-egual, if this word is pronounced in two syllables. 140. vi is a dipthong of which the initial is u {00"), and the Portuguese are more consistent than the Romans in writing it Me and ui. The Portuguese monosyllable /itt, I have been, nearly resembles the Latin dissyllable pui. uj is heard in the German word pfui, with which the English word buoy (booy) rhymes. 141. UJ occurs in the interjection htij; and in the poetical forms cuj (not cvi, according to the analogy of CUJ, cujus) HUJO, ETJJT. Terentianus Maurus and Julius Scaliger regard the final of cui to be J ; and Priscian considers the final of the vocative form CAJ in the same light, 142. TJ may occur in shortening words or inflections like POIjYIDUS, IMITYIS, COTYIS, CAPYIS, ITTI. 143. When ti is followed by a vowel, the coalescent is apt to become a consonant, as in har-py-ia (HAR-py-JX), or the French words essuyer, noyau. 144. ij may occur like TJ, as in Livij, used by Ausonius. 145. As the Romans used but a single character for the vowel and consonant power of I and v respectively, errors in pronuncia- tion may occur from inability to determine when the consonantal • " Rightly to find the Latin diphthong ce The sound of o and e you must enfiploy. No Roman ever sounded it as we Who make the M and oe like English e ; For so our doctors taught us docile boys, Not to distinguish between ju's and cb's. And thus a hundred errors find their way, By this confounding iE and ie with E." — P. 40. 34 DIPTHONGS. — LABIAL CONSONANTS. powers should be employed, as in Huic, which might be made to rhyme with the English word wirl-, destroying the dipthong by turning u into a consonant. A parallel change would " be made if lu dipthong were changed into the English syllable you, that is, from iv to J v. 146. The union into a single character of [^M, oe] is improper, because it breaks the uniformity of notation, no other vowel cha- racters being so represented. Characters are sometimes united to economize space upon Roman coins, of which the syllables AV, ME, NE, ET, VE, TV, VAL, MAR, and others, afford examples. In " forming [val], the second line of [a] would be applied to the top of [v], whilst it would form the stem of [l]. 5. LABIAL CONSONANTS. V 147. If the lips be gradually closed upon the vocal current, the liquid of the labial contact will be formed. Its quality approaches that of the vowel u so closely, that in Roman inscrip- tions the two were represented by the single character v, as in PVBLicVs public, LiQVOR a liquid. *148. The ancient grammarians include b, p, f, m, in the labials; generally confounding v with u (§ 108); but Cicero adds v when it has its consonant power. They do not hint that the consonant is formed differently from the vowel V, so that this must be the English toe and not the English ve. 149. Those Latin authors who, in treating of the alphabet, describe P as being formed with the lower lip and upper teeth, say nothing about v ever being formed in this manner ; and the G-reeks, in representing Roman names, make no distinction com- patible with such a difference, as in v0LTURNUS [Oi;ovx*oi;gvoj, 8«7ii'»gvo{ (§ 169)], the initial syllable being the English wool. *150. According to Pennington, "The Roman v was more probably our w," an opinion with which Webster, Donaldson, Rapp, and the author of " Living Latin" agree. 151. The English syllabic ^wai/ occurs in the second syllable LABIAL CONSONANTS. 35 of MANSVBTTJS tame, and in the first syllable of SvlVUS, but not when it is the trissyllable suEvtis relating to the Suevi. tenuis thin, has three vowels, but when contracted, into TENvis it has but two. The reverse takes place in siLVX, which Horatius uses in three syllables. It is not probable that the same word was intended to be as dissimilar as the use of the English letters u and V would make it. 152. The affinity between u and v is proved by the derivation of NAVTA and oavtio from navixa and cavitio ; and by the poetical use of dissoluo, etoltiam, peevoluent ; instead of the trissyllables dissolvo, evolvam, pervolvent. 153. Cicero and Plini relate that M. Crassus, hearing a crier of a kind of figs cry CAVnSas (usually printed 'cauneas'), took it for a bad omen, understanding the cry to be CXvE ne hAS, beware how thou goest, the first E being probably indistinctly enunciated, and the two others lengthened and confluent (§57). The English mode of pronouncing the first syllable of 'cauneas' caw and CAVE with English v, destroys the analogy between the two forms. 154. In the • following examples the Latin way has been pre- served in English : — VXLEo to be well. vermis a worm. VALLum a wall. VESPa a wasp. VADO to walk, wade. ViDiiX a widow. VAcillo to wag, be fickle. viGiLO to wake, watch. ViSTo to waste. VINCA a winkle (shell). VASTATTJS laid waste. VOLo to will. VANNO to winnow. ViCus a village or wic. VERRuOX a wart. BERVicium Berwick. vix a way or road. ViNum wine. VEHA a wagon. ViTrum woad, glass. VEHO to carry, whence weigh. voLVO to roll or wallow.* ViLtiis wool. ^ VuLNus a wound. VELLus a fleece. vEgSo to be strong, to wax. VENTus wind. ViscuM glue, whence wax. 155. Although the use of the rounded or angular character • Wheel is from the same root, namely, the Sanscrit taji to turn. 36 LABIAL CONSONANTS. [u, v] is almost a matter of indifference in Latin typography, it is better that one form should be invariably used for the vowel and the other for the consonant (§ 112). 156. tJ, "when followed by another vowel in the same syllable, becomes a consonant and should be written v; as XQVX, SANQViS, &o."—G. Walker. 157. In a few instances V was dropped after 0, as in the word SECuTUS, which had been seqvutus; in the double form sEqviits, and sEcius; and in colliqviae, used by Columella where Plini uses COLLICIAE. 158. If there was no v after Q (as in the French and Spanish g'ui), C0R could not be derived from QVaRe, conCUTIO from QVaTIO, CUJuS from QVIS, the Spanish cmtro, the Gothic fidur and English four, &c., from QVATUoa; nor the Spanish agua from XQVX. The Greeks represented the Eoman name qvinttjs by [kyintos]; and many rejected Q altogether, writing CYIS for QVIS, &c. See under Q, §§ 290, 292. M 159. When the lips are entirely closed and the voice is allowed to pass through the nose, the labial nasal will be the result. 160. M is a nasal b, which accounts for the derivation of GLUMA from GLtJBO ; and the two forms proboscis and promus- OIS. Its relation to v is shown in promulgo to publish or pro- claim, from PRovuLGo, having the same meaning. B 161. When the lips are closed without opening the nasal passages or stopping the vocality, the sound of B will be given. 162. P had, as far as we know, a uniform power; and as B replaced it, as in the forms repo, rebo; soribo, scripo; PyR- RHuS, BURRUs; poPLicus {itujfKixoi in Greek], publicus; popli- coLX; puBLicoLX; it is certain that the power of the French and English b existed in Latin. d *163. When the vocal barrier is broken by forcing b through LABIAL CONSONANTS. 37 the lips, a sound will result somewhat resembling the English, French, and Spanish v, which latter is a labio-dental, and Will he here represented by v for the sake of illustration. 164. The pure labial consonant J is represented by the German [to], and by the Spanish [6], in certain cases, as in [^Cdrdoba], from the Latin cqrduba. It is probably the Hebrew heith; and it occurred in Greek, where its character was termed pav (pro- perly JAv), or aeolic digamma, because its form [F] resembles a union of two gamma characters [r]. *165. To prevent it from being confounded with their own character for F, the Romans inverted it, as in writing viR, VIRTUS, sERVus [JiR, jiRTUS, SERjus ;] but its Use was soon relinquished. 166. The power of v (English w) is usually attributed to s, probably because the German and Spanish sound alluded to (§ 164) is scarcely appreciated as distinct from vi or b. 167. The small Greek letters being more easily written than the capitals, the digamma would be written [yjr], which would pass into [w], and the origin of the latter character being mis- understood, it was confounded with [u, v]. From this it appears that [w] has its proper value in German. It has no place in the Romish languages, and when it appears in French, in foreign names, its power is that of v. 168. The Greek and Latin [b], and Hebrew [3, or 3 when aspirate], must have had precisely the unstable power of the Spanish [6], sometimes lene or pure, and sometimes aspirate, and when aspirate, forming s. and allied to V. This seems borne out by the Hebrew arnabb^A, and Arabic arneb a hare; the Greek phRAToR, GoMPhoS, and the Macedonian variation brator, GoMBos; and the old inscriptive forms berum, probaberit; for VERUM, probaverit. § 174. 169. The views taken here elucidate the use of the Greek digraph m or a (Latin u) and p, in representing allied or identical Latin words, as in Aasfns and Aaj3ivi,a'tu>ii, used by Dionysius in allusion to LAViNiu"". So we find the name of varro given as Ba^puv and Hap^uv. See note 83. 170. It is a curious fact in connection with these discrepancies, 4 88 LABIAL CONSONANTS. that there are Spaniards who, knowing that their [6] is never v, fancy it is always lene or pure ; and I have known it to be in- sisted upon that the German [w] is identical with the English M- 171. As J was used in some words written with [v], this character, as well as [b], probably represented it to some extent; as in the old inscriptive forms DANwrvs, acervvm ; for danu- BIUS, ACERBTJM. 172. The Greek [B, (3] aspirate? became in some cases the Latin v, as in the words yicix, volo, vivo. 173. The character a being rejected, its power probably re- mained associated with [b] as in Spanish; as we find the double forms bEn£, vene; BXsis, vASis; lsbor, lavor; it being extremely easy for the aspirate b to fall into the allied v, of which an example is furnished by the German words tvein, will, when compared with the English words wine, will. 174. The B was probably at first aspirated in words from the Greek having phi, as BJLAENa, NEBuLa, albus, orbus, ambo. 175. V became B or b, as in bellum from dvellum, bis from DVis ; and in besica for vEsica, larba for LARva, berna for VERNA, &c. 176. B sometimes replaces v, as in absens, absolutum, from the older forms apsens, apsolutum. So we find conlabsum for COLLAPSUM. 177. B must have been pure in the numerous cases where it interchanged with p. P 178. "When the vocality of b is stopped, p is the result; it is therefore a surd b, as b is a sonant P. It was not subject to aspiration, which formed a sound foreign to the language. 179. P sometimes replaces b, as in caNoPus [Kovweos]; and in the inscriptive form optinebit for obtinebit, &c. § 162. 180. The character for P arose from the Greek form [n], the right side of which was frequently made but half the length of the left, forming a character [r] to be seen upon old Koman liABIAIi CONSONANTS. 39 Ph P-\ 181. When the breath is forced through the lips (as in blowing a small object), the Greek phi [*, ^], (a labial/) is the result. It is therefore the corresponding surd of the sonant digamma. It is heard in Swedish, and is represented by [/] in the German word kop/weh. According to Pennington the modem Greeks pronounce [*] soft and full, "more like a sigh, thongh it is not easy to express the diflference in writing." P. 71. 182. Phi occurs in certain words of Greek origin, and there- is sufiScient evidence of its distinctness from F, as in the case of the Greek witness ridiculed by Cicero for pronouncing a proper name phTJNDANiDS instead of fundanius. If the Greek digraph [mi] was not AV but af, as the modern Greeks maintain, Cicero's witness would have had no difficulty with the Eoman r. 183. In naturalized words of Greek origin phi became F, as in F-iLius a son; FJMa fame; fugS flight; fur a thief; fKro to hear; FSLLO fo deceive; FJGus a ieech; FESTiSR a brother. In inscriptions we find faselus a skiff; falErae trappings; SIFO a siphon; elefas (and ELEPhAs) an elephant; delfintjs a dolphin, which the moderns write with [ph]. The carpensian Virgil and that of the Vatican have sulpur for sulfur sulphwr. The ancient phAiuM is written [Fdwum] or [Faiooni] in books. 184. In some cases phi became p, as in purpureus [yto^^^toi], PALANGA or phaLANX, PROSEEPiNaj in the double form tru- PERA, from TRYPhERA; and in the change from phoEnicius Pheiiician, to PoiiNicus, and finally to PuNicus punic. *185. Some authors suppose that the Greek phi (also Rh, Th, and Ch) was P followed hy an aspirate, as in the English word havnazard, because the ancient grammarians regard it as P and an aspirate, as it is in fact; for if an aspirate is made and the lips be gently closed toward the P position, phi will be formed. Hence this sound is not a post-aspirate, but a co-aspirate P, or this element modified by a synchronous aspiration. 186. Quintilian admired this sound as a pleasant breathing, which shows its nature ; whilst Terentiau wished it to be prQ< 40 LABIAL CONSONANTS. nounced in introduced Greek words, although this could not be done by the Romans without special instruction. *187. The power of [f] is known (§ 190), and Priscian de- scribes it as composed of P and an aspirate, so that it is related to phi, which is not the fact with P followed by H. Moreover 188. The Greeks, -who could not pronounce f, represented it by [*] phi in Latin names, as in *a6ioj, ^avstv\os; for fabius, PAVSTULUS. 189. In words borrowed from the Greeks, phi is sometimes represented in Roman inscriptions by a character formed of a union of [p] and the right hand poriion of H. See Manutius' Orihogr., Venetiis, 1566, pp. 215, 271, in the words nioepiob, HILEMON. The former inscription has [th] united in the word XMARSNThus, by adding the horizontal line of [t] to the left hand line of [h] lengthened upwards. F *190. The Roman f is correctly pronounced by the moderns; as the ancient grammarians describe it as being made with the aid of the lower lip and upper teeth. There is no evidence of its corresponding sonant y existing in Latin; and it is also wanting in German. pH,etc. 191. Modern writers on Latin grammar have falsely assumed that if a Latia word is derived from the Greek it must follow a certain orthography; and if not derived from this language, it cannot have PH in it. It might as well be said that the English words haPHazard and uPHold are incorrectly spelt, because PH should be placed only in words of Greek origin. 192. The orthography of bosphortjs is said to be incorrect, because the etymology requires it to be BOSPORtrs, an assertion which virtually denies that H can be acquired in words from the Greek where it is absent, although hibisou and helops are examples to the contrary. 193. The word bosphorus would not be [B(.>s^>ogo{] in Greek LABIAL CONSONANTS. — ^DENTAL CONSONANTS. 41 characters, but [Bujrtogos], as isthuo (ist-huc) would be [ij*'sas] and not [ij^s*]. 194. H followed p, &c., (when written) in triumphus, inchoo, COCHLEA, BACCHUS, and other words. 195. Cicero thought H should be rejected from TRiUMPHtis, PULCHER, CARTHAGO, and CETHEGUS, probably because he did not pronounce it, and his authority is sufficient for its rejection. This fact is sufficient evidence that ph in triumphtjs are not equivalent to p ; and the inscriptive forms triumphs, pulcer, are sufficient authority for its rejection. 196. If this view of the double natvtre of [h] is correct, there is no means of readily determining when it is to have its inde- pendent power. On this account [h] is used when the pure sound is supposed to be represented; and [h], which is an ancient form, when it is merely a diacritical mark of co-aspiration. 6. DENTAL CONSONANTS. L *197. L is the liquid or half interruption of the dental con- tact, and the descriptions of Victorinus and M. CAPELLA corre- spond with the ordinary modern power. 198. L interchanges with the liquid of the next contact R, as in PAVLUS and parvus litUe; parilia and palilia a kind of coat. This change is extremely common in the languages of Polynesia, and is observed in the Spanish esclavo and the Por- tuguese escravo. 199. L interchanges with D, as in dacrima and lacrima a tear; odor an odour, oleo to scent. N 200. N bears the same relation to d that m bears to b. D 201. D is the sonant, fully interrupted member of the dental 4* 42 DENTAL CONSONANTS. contact. It is interchangeable with t, as in Xput for apud, set for s£D. This change corresponds with that of P to B. § 179. T 202. t is a surd d, or d deprived of its vooality. Th *203. Th is an aspirate, usually, but not necessarily formed a little in advance of the ordinary position of the dental contact, like the Irish d in certain cases. It is the equivalent of the Greek theta, which the' modern G-reeks pronounce as in the English word thin. 204. When Th replaces s it forms a lisp, and this interchange indicates the co-aspirate nature of theta [®, ^] as in Thjos god, in the Doric dialect Stos, and in the Laconian sioE. § 220. 205. The Germans, French, and Italians are not familiar with the sound of Th, and they accordingly replace it with T, which the author of "Living Latin" justifies. 206. The English sound of [Th] is common in the Oriental languages, and so is h following t, as in foffsnold. 207. The Sanscrit t remains t in Greek and Latin, and D and T followed by H become xh, and also T in Greek; which is in favor of the post-aspirate theory that theta represented the t and H in poTUook. But 208. The Sanscrit d followed by H becomes the Greek delta [a, 5] to which the modern Greeks give the power of Dh in this. Moreover, the Sanscrit pure T also becomes the Greek theta, which is against the post-aspirate theory. 209. The post-aspirate theory would remove an anomaly from the Latin alphabet, namely, the representation of a single sound by two characters, but language must not be sacrificed to writing. 210. By taking [Th] as the representative of two sounds, as in penthouse, Beethoven; or of T alone, as in isthmus, Thomas, Anthony, Luther, Rothschild, Othello, we establish a rule which must be followed with [ph] and [ch]. *211. H occurs after T in isthuc, isthic, anthac (when ANTEHAo is condensed), posthjc, posthumus posthumous, and DENTAL CONSONANTS. — PALATAL CONSONANTS. 43 probably in penthemimEris (from the Greek PsNTs hemi MsRIS). 212. The modern corruption of reading the English sh instead of T in words like lectio a reading, rxtio reason [and of Cay in ocfiANus ocean) is improper; and a rejection of sh implies that of the Italian corruption tsh. *213. Saint Hjerom (who died A. D. 420), after stating that the Hebrew Samech is S, finds himself unable to give his Latin readers an idea of the Hebrew ty (or V)) shin (or sin as he was compelled to write it), because the sound does not occur in Latin. 214. According to the same authority the Hebrew Sadde (usually but improperly read as ts) is disagreeable to Roman ears. It is a peculiar aspirate consonant of a quality between English sh and ch (jx), and equivalent to the Arabic sad, and Greek jafirtt or 5AN, whose place in the Greek alphabet is next after tt. 7. PALATAL CONSONANTS. R 215. R, the liquid of the palatal contact, must be trilled or vibrated to make it agree with the descriptions of the ancients. It cannot therefore be replaced with the English smooth r. It sometimes interchanged with L, as in pst£rx and pxtellx. Rh 216. Rh, the aspirate of R, is surd or whispered. It is used almost exclusively in words taken from the Greek, which gets the sound from the Oriental languages. In the Eomic languages it is preserved in the French terminations pre tre ere. 217. The Greek aspirate R (g") was not always preserved in Latin derivatives, as in roSX a rose, RESiNX resin; being, like other foreign somids^rejeeted from properly- naturalized words. The old Latin word BtfRRUS, which was legitimately developed 44 PALATAL CONSONANTS. from the Greek, was afterwards considered a Greek name and replaced with PYaahus. s 218. B, in French and English, when it occurs between two vowels, is apt to be affected by their vocality, and to become sonant, as in miserable; and we might hence incorrectly conclude that the rule is universal. 219. In Spanish, s preserves its pure hissing power, and as the ancient grammarians do not mention a sonant power, the same sound must be preserved, as in the English syllables say (SE himself) ace (Es tJum art). The Latin word tees three, is preserved in Spanish, and pronounced like the English word trace, which is a little shorter than the Latin word. 220. s was interchangeable with its liquid, as in the old forms PLisiMA for plurima; qvaesumus for qvaerimus, § 76; PAPisius for PAPiRius; lases for lares; asas for aras; MELios for melior; rusius for purius; and fasena by the Sabines, for arena. 221. z, which (like v) was not used in old Latin, is a double letter according to the -ancient grammarians, and it accordingly lengthens a syllable by "position." 222. z is composed of s and d, but we are not informed whe- ther the hissing (surd) or buzzing (sonant) sound was used. In the vocal scheme (§ 35) the latter is assumed. *223. According to Dionysius, z is composed of s followed by D, "and that this is done advisedly appears from a passage in Herodian." (See the note.) — Pennington, p. 70. According to Maximus Victorinus z is sd, the proper name mezentius being MESDENTius; and in consonance with this view, the proper name EZRA or EZRAS is given as tsSgoj (esdras) by Origen. The ancients named d a mute and s a semirvowd, and Yerius Flaccus (if the text is pure) says that without doubt z ends with a mute. The Dorians wrote [sd] at length, instead of [zj, as in [MeUSDo] PALATAL CONSONANTS. 45 for [McLizg], SO that SD is correct, at least as far as the Doric dialect is concerned. *224. Surd consonants being less difficult to form than sonant ones, they may be expected where the latter occur. The Italian dz and ts are not Greek combinations, and were the former in- cluded in z, we would still want ts, which should be at least as common as ds. But compound forms like Esdechomai show that SD is a Greek combination, although usually represented by [z]; we may therefore naturally expect its corresponding surd ST, which we find so common that it has been provided with a character [y], as in aygoi/ (astron) a star; and words like S^aigigt'^^ioi/ from s^aifiju, saArfijT'jjf from saJ.rtifu, x^vft^tj from xiM^a {x%vsSui), and the double form mastos and mazos (masdos) are conclusive. Characteristics of a language should be faithfully recorded, and none should be relinquished because the develop- ment of certain modem languages has taken a different direction. *225. z, in naturalized Latin words, took one or the other of its constituents, as in medentius and messentius, instead of MEZENTius; MTisso, spisso; and in old Latin we find the Latin- ized forms SONA, SEVXis; of zoNX, ZEVxis. Replacements like these are also found in Greek, as SS (lisped?) for z in Doric; and KYPRiSMos from kyprizo. Compare ^^^fo, t?<*5") e^^aSov. 226. In modern Greek [z] is a vocal s, as in rose; and this seems to have been its ancient power (or a dialectic variation) in a few words, as in zmirnx S'myrna.^(^Pennington, p. 69.) This view is confirmed by the variations in inscriptions, as lezbia for LESBlS; PHILOGENEZ for PHILOGENES; *ao»A^^ for *aoxX^S. — (^Schneider, p. 382), although, according to Priscian, no true Greek word ends with z, and Ssij for Ztvs, and bySANTION as well as BTZANTION upon ancient coins. So the name of the Spanish town SXGUNTTJS was written with an initial [z] by the Greeks. The English z was probably found in the Eolian ZA for dia, z and d being allied sounds. 227. In Italian, z has become both dz, as in lazaro ; and ts, as in oalza; and in this language (and in German) s has crept in between t and lo, as in nazione, from NSTio a people. 228. The corrupt mode of reading Latin by inserting s after T 45 PALATAL CONSONANTS. in words like nxtio and gratiX, is supposed to date as far back as the beginning of the seventh century ; as we leam from Saint Isidor,* of Spain, that justitia was at that time pronounced JUSTIZIA, but we are still in doubt whether his z had its ancient power, its English one, or the Spanish lisp. Lipsius quotes a passage (p. 74, which Schneider does not consider ancient] in which TZ is assigned as the power of this t. 229. Latin ceased to be the vernacular language of Italy towards the end of the sixth century. (dzh, zh) ^230. A few English authors have endeavored to justify the English dzh in Latin from the word jupitEr, which they think may have had an initial D, being, as they say, derived from Diu PITER ; although others derive it from jovxpatee, and diespitee. But DIUPITER would merely have produced the first syllable of the English word due (not Jew), as in aDJUnctus, aDJUto- RIUM, which have become acfyuntivo, arfyutorio, in Spanish. In a similar manner, the gay of the Irish word cuig seven (the first syllable like coo, the second as in igmiie) changes dialectically to the Latin dj, forming cuiDJ, not coo-idzh, which would be an Unglish development. 231. Moreover, a consonant preceding another was frequently dropped, as in the proper name bellius from dvellius ; in BIS from dots; the English when from the Latin qvando; the Dacro-romanic ava for aqva ; and the Oscan peltora when com- pared with qvatuoe. 232. As the sonant phase of a consonant is more difficult to form than its surd equivalent, it cannot be expected that the Eomans, who could not pronounce shin (§ 213), would be able to form its sonant, the French j as in azure, which is included in dzh. 233. If [t] is another form of [s, sh, or TS] in oriosiis at ease; [d] must be read as English z or zh, or be followed by z, in 5l>iosus hateful. * Schneider, p. 356. Grotefend, 2, 272. GUTTURAL CONSONANTS. 47 8. GUTTURAL CONSONANTS. 234. The guttural contact is formed with the base of the tongue and palate. J *235. J (yota) is the liquid, or half interruption of this con- tact. It is heard as the initial of the English words you, yoJce, and in the last syllable of hallelujah, which is spelt in English with the proper character, as recognized in Italian, German, Polish, and most of the languages of northern Europe. 236. The natives of jIfa (written [Y af a] by a recent English traveler), and "Yebna," the ancient joppa, and jamnia, pre- serve the initial of these names pure; and in the Levant the initial of the names John, Jacob, and Joseph, corresponds with the German sound, and no one pretends that the Hebrew originals should be pronounced diflferently. The river of India named JOMANES by the Romans, is now called jamuna (Yamoona) by the natives, and corrupted into Jumna by the English. 237. The English pronoun you occurs in Swedish, where it is spelt [ju] as yule is spelt [JUl], with the proper characters. The English word young differs only in the vowel from the German Jung, which is the Gothic jimgs, the Latin juvenis and the Sanscrit juvan, giving the last in Eoman characters, the first and third letters being the English [y] and [w]. From the Sanscrit juga resulted the Latin jugu, the Gothic and Dutch JUK, the German joch, the Spanish yiigo and the English yoke. 238. The consonant j has been retained in many Spanish words of Latin origin, as in ayunav, jEjunare; yacer, jxc£re; yactura, J XCTimTi. ; yamhigo (Ital. Samhico') lAMBiCUS; yugular (Italian Jugulare) JUGULSRIS; yuxtaposicion, JUXTAPOsiTIO ; conyugal (Italian conjugali) CONJUGIALIS. 239. J has to some extent become dzh in Italian, as in giacere from JACERE ; and also G (in the same contact), as in coiighieitura (Spanish conyetlura), although conjettura is also used. A some- what similar change has taken place, where J has become ch in 48 QDTTtTRAL CONSONANTS. Spanish, j, g, and ch being members of the same contact, and. therefore interchangeable. 240. The Spanish use of the character j to represent ch, is therefore less of an error than to make it the representative of zh or dzh; just as the character G is less perverted when read J (as in Bohemian and Gothic), than when it is read zh as in French, or dzh as in English. When [g] represents the Koman [j] it is surmounted by a dot in Irish, as in writing saigitteoir, Latin SXGiTTARius an archer. 241. Double forms like iosEph and JosEph; xbiegnus and ABJEGNUS made of fir ; and the use by Plautus in the dative case singular of eji for Ei show the close relation of j and i. NG *242. The guttural nasal is heard as a final in the English and German word sing. It is represented by [N adulterinum] in Latin, where it occurs before the gutturals c (q, x) g, and ch, as in SNCoRX an anchor, ANCIlla a maid servant, iNQViRO to in- quire, Longus lonxj, ANGViLLX an eel, ANXiilS troubled, ANChiSES Anchises, angElus an angel, ingEnuus ingejitcous, longinus Longinus. 243. The Greeks and Romans neglected to provide this element with a peculiar character. In old Latin it was more correctly represented by [g] in the Greek manner, as in XGGuLiis an angle, AGGVILLA (aqgulxjs, aggvilla), the sound having a nearer relation to G than to N, and to represent it by the character of the latter, conveys a false idea of its affinities, as if the character for M were allowed to represent n, this being a parallel case. The change was probably made to prevent words like aggEr a mound from being pronounced like anger. 244. g has a proper character in Sanscrit, as in the original of the Latin word coNchx a conk shell, and its Greek form, in which it is represented by the gamma. It is also found in the original of ANGUSTTJS narrow, from the Sanscrit verb AGG to contract. 245. G was in some cases derived from m by assimilation, as in PRiNCEPS (pRiMiCEPs), siNciPiiT (semicaptjt), hobunce (horumoe) unqvam (umqvam). GUTTURAL CONSONANTS. 49 246. In transcribing Eoman words, the Greeks used their gamma instead of [n] when this character represented ng, as in the name of the British tribe cangiani, which became kafkanoi in Greek, as Plutarch uses nRirKiniA for PEiNOiPiA ; affording additional proof of the guttural nature of the Latin sound. G 247. G (gay) is heard in the English words gear, give, gay, get, go. It is the sonant of Gay, and was represented by the same character until the little mark was added which distinguishes [g] from [c]. This mark was introduced by Carvilius, after whom it may be named a carvilium. 248. At a remote period, the celtae or Kelts entered Spain from France by crossing the Pyrenees, and having become per- manently established, they formed with the Iberians the CELTI- BERI. 249. This accounts for the Keltic names in Spain which the Romans adopted; and for the occurrence of the same names in Spain and England, as Ahono in the former and Avon in the latter; or that of Jura in the Hebrides, and in Switzerland. Asturias is derived from AS a torrent and tir land, meaning the land of torrents; Sardinia from sard the larger, and INIS island; and Lusitania from TANA the country, o of, luis flowers.* 250. The languages of the Keltic stock, having preserved the Roman gay and cay pure to the present day, supply us with the pronunciation of ancient Latinized names, as tulingi from tul a flaod; ungead, leaping; cingetorix from cingead valiant, and EIG (Latin regs) a king ; cogidunus from COGAC war, and DUN a Mil; vergesilaV'nus from fear (Latin vie) a man; gals expert, saelan (with a) spear. The name of a British king, VORTIGERN, is from for, chief, and tigerna, Imd; and that of the British tribe durotriges is from dur, water, and trig, an inhahitant. 251. In French, Guines is derived from gisna; and Bigorre * The Gael and Cymbri, by Sir W. Betham, Dublin, 1834. The works of Johnes and of Priohard may be consulted on the same subject. 5 50 GUTTURAL CONSONANTS. from BIGERRENSIS AGER; whilst the German name Bregentz is from BREGENTIUM. In German, Spanish, and some other lan- guages, the character [g] never becomes the representative of a palatal articulation. c 252. The Latin [c] cay has the power of [k], and no other ; as in CALCO to tread, CALOitro to kick, CALOEus a sandal. The character [<:] was used at an early period as the representative both of G and o, whilst [c] has the third place in the Eoman alphabet, which is occupied by [r] gamma in Greek. 253. In the name gajus or cajus the sonant or surd form is used indifferently, and the Sanscrit jug {yoog) to throw, produced the Latin jACio. In the double form TRiOEsiMus thirtieth and trigesimus, if [c] represents s, [g] represents English z; or if the latter represents dzhi, the former represents tshi, as in Italian. 254. "The uniformity of pronunciation in c, G, &c., when fol- lowed by a vowel, is strikingly confirmed by the silence of all the ancient critics and grammarians, who, though treating ex- pressly of pronunciation, never indicate any variety," — G. Walker in Scheller. This argument is strengthened by the fact that the ancients are sufficiently explicit upon the varying power of [n] and [m]. 255. In the following lists the words written in the first column are derived from those of the second, and it will be remembered (§ 17) that palatals do not change to gutturals. sanctum sancitum doctum docitum lectum lecitum vinctum vincitum decuhia decem CEPi CAPiO PARCI PAROUS SECIUS SECUS AUDACTER AXJDACITER AXLEOTO AXtlCIO OCELLUS OCiiLUS GUTTURAL CONSONANTS. 51 256. The organic change from C to s was indicated in Latin orthography, as in censEo and sentio to be of opinion; census the censor's valuation, SBNSus sense (from the same root)j ful- crum a prop, EULCIO I prop, EULSI I have propped. We find also RA^cio, eaVsi; and sarcio, sarsi; -which, as variations respectively of the same word, would have been written similarly if so pronounced. 257. [k] is used in writing the Greek form of CAEDO to strike or kill, which is the Gothic scathia and English scath. Yet Caedo is frequently read as if it were siDO to perch, or cedo to grant. Moreover, the same root gives rise to cedo and cado to fall (§ 55), whence OCOASUS a fall, death, and occiDO to perish. LtiCERNA a lantern became lukarn in Gothic, and leuchte in German; and the Gothic easkja is from the Latin eSscia a hand. 258. The guttural contact in ac£r or acris sharp, hold, is preserved in acrid and eager; and in macSSr and the English meagre, the French maigre, and the German mager. It is also preserved in the Irish cer, in Latin cera, wax; airgiott, Latin ARGENTuM, silver; 0£L, Latin Celo, to conceal. 259. The Greek ktknos a swan, is the Latin cyCNus or cyGNus, and the modern Persian quqnus; a cherry is kiras in Arabic, cErAsus in Latin, and Mrsche in German; and the Greek mekao to hleat, became the Latin miceo, the German meckeren, and the Lithuanian mikenu. 260. If the initial of the Latin word oiihARA (spelt with k in Greek) was a palatal, the Italian chitarra (ch as k), French guitare, and English guitar would have a doubtful etymology, and the same doubt would exist in the case of the German words keller a cellar, kerker & prison, hicher a chickpea, which are from the Latin CELLARIUM, CSROEr, ciOEr ; or with the English words dk from ALCES, and skink from SOINCUS, a kind of lizard. 261. The Sanscrit schal (h, pure after k) to swerve, gave rise to the Greek skolos, whence the Latin scelus guilt, and subse- quently the Gothic skula, and modern Westphalian s'chuLD ; so that [so] in Latin is not a double character for s, any more 52 GUTTURAL CONSONANTS. than septic, sin are identical in English with sceptic and tkin (properly scin, from the Anglo-saxon). 262. The following Keltic etymolo^es of Latin names from Betham (§ 249), show the guttural nature of Latin Cay (§ 15). CALLAICI, from CAOILEACH, narrow (the narrow slip) ; aVlerci, from all great, and learq plain; CERONES, from CAOR sheep; ICENI, from gan hounds, and OICE the sea. 263. The consonant cay has been preserved in the modern names following : — Kaiserlautem from oaesarea ad lutram. Kaiserswerd " caesaris VERBA. Kyllourg " OELBIS BURGUS. Qtierguinez (qu=k) " cercina. Draguinan " dracenum. JExilles " OCELUM. Selefkeh " seleucia. The following Syrian names are from the maps of the Society for the Diffusion of Useful Knowledge : — LadiMyeh from LAODICEA. Kaisdriyeh " CAESAREA. Killis " OILIZA. Antakia " antiochia. Kerak " CHARAX. The following words are from a vocabulary of Albanian given in Diefenbach's work.* Ch has the power of k. Fachie from PAOIES. Pache " PACE. Pische " PISCIS Zepa " CEPAE. Sckanduem " scintilla. 264. When a prefix is added to a Latin word whose initial consonant is different from the final of the, prefix, an adapting change called assimilation usually takes place in the prefix. 265. Thus AD to becomes an when a prefix to necto to bind, * Ueber die jetzigen romanlschen Schriftsprachen, u. s. w. Leipzig, 1831. GUTTURAL CONSONANTS. 53 forming jnnecto to connect. The same prefix is modified in the first syllaMe of the compounds aliatbo to bark at, attribuo to attribute, APPELLO to drive towards, AFPLiiO to Jlow towards, AEROGO to claim, assideo to sit at; whilst in abduco it remains pure. In Pomoerium a limit, post is reduced to PO. 266. AD becomes AC by assimilation before cay, as in accolo to dwell near, acceptus accepted, aooingo to gird; and AG before ga^, as in aggravo to aggravate; aggEro to heap to- gether. Priscian cites qvicqvam (from qvidqvam) and accidit, as instances of this change from dai/ to cai/. AH these examples prove that the conjoined letters had the same sound, and to pro- nounce them diflferently reverses the Latin practice. 267. When a consonant character is doubled to represent assimilations, both should be pronounced. 268. The operation of assimilation is seen in the change of n to ng (§§ 242, 246) in words used as examples of the change bi/ the ancients tJiemselves, as ANGiSLiiS, inGeuus, longinus; and no exceptions are adduced to this change before cay and Gay preceding I and e, so that they must each have been uniform in all cases ; for had they been palatalized in ancTlla and longi- Niis, the N preceding them would necessarily have remained pure. 269. Every one pronounces the second c in Eccu™ behold him like the first, yet this word being a contraction of ECCE Jllum, if a sibilant is placed in eccb, it should have a place in the deriva- tive. So from Hie here are formed hicce this, and hiccInE he. *270. The Greek character [k] was never properly naturalized in Latin, nor was it used as in the modern Teutonic languages where cay would be likely to become s ; but (according to the author of " Living Latin") to prevent cay from becoming gai/; and as this distinction took place only before the single vowel a, and was obviated by a difierence of character, the [k] became useless, and was so considered by Priscian, who says it has no ■ other value than [c]. Varro and Nigidius Figulus (§ 292) re- jected the character K, whilst Quintilian, Terentian and others considered it useless. 271. According to Gregorius Placentinus [k] at one time had 5* 54 GUTTURAL CONSONANTS. a syllabic power equivalent to CA, when oSRus was written [KRUs] but read carus. 272. For the sake of brevity, the Komans introduced into their numerous inscriptions a multitude of abbreviations, in which a feip initial characters, or a single one, represent an entire word; as [c. CL. R.J for cavsa clari regi. 273. Manutius gives a list of fifty words for which the charac- ter [c] stands; thirty for [q], and twelve for [k]. Among the last are cxlenpae, carus, caput, cxlumnIx, words sometimes spelt with [k], a practice which the inscriptions fostered, although they were intended merely to prevent confusion by assisting the memory in reading them. 274. In reading Latin words as if they were represented by English characters, there is a singular discrepancy in the case of [c, g], which have preserved their pure power before A when the character is insulated, but not before [i] when incorrectly read AJ, nor before ae, as in caenepolis (the modern Qen6 or Keneh) although in ancient inscriptions [k] is sometimes found in the latter case, as in kaesonia, kaesius. 275. If C and g had not their pure power before I E, as in gear, gay, hey, cane, the alphabet could not represent these syl- lables, and Latin would be a more corrupt language than any of those derived from it, and, in fact, the most anomalous known. *276. The ready interchange of c and G, &c., seems to prove that the surds P, t, C, were what Rapp terms "indifferent," or pronounced with a greater surface of the organs in contact, as in some modern languages ; a phenomenon which causes surd and sonant to be confounded, as in German, where dinfe or iinte (ink), hrod or hrot (bread), are used according to the inclination of the speaker. 277. and g must be fully pronounced before N (§ 19), as in the German word gneiss (GNAJs) a kind of rock; knie (the old English knee); and in the Irish cndib, Latin cannabis hemp; gne, Latin GENus KiNc?. The same remark applies to T in TMoLus, the name of a mountain ; and to P in the proper name PTOLEMAEiis, and in psalmus (psalm in German and Flemish, and SALM in Danish). The Greek combination kt is found in GUTTURAL CONSONANTS. 55 Latin in borrowed proper names like CTESias, and although it is somewhat difficult, it must be pronounced if accuracy is desired, as in some of the aboriginal languages of America. X 278. According to Maximus Victorinus and Diomedes, "the ancients," before the invention of [x], wrote as and OS, the former in words inftected with G, like rex, rexi, maximus, anxius, and the latter in such as are inflected with 0, as Pix, LUX, felix, Dixi. In cases of doubt, as in nix (compare ningo), connixi, ALEXANDER, [x] may be read OS, especially as Varro (a cotem- porary of Cicero) asserts that no one can distinguish any difference between its two powers. Ch 279. Chi is the aspirate of c, a sound which is retained in modern Greek, Scotch, German, Spanish, and many other Euro- pean languages. 280. In Latin, chi occurs in words ta:ken directly from the Greek, and when it cannot be pronounced, it may be replaced with 0, as in the Latin double forms chARiTAS charity, coohLEA a snail, and (more correctly) carItas, coolEA. The pronuncia- tion should, however, follow the orthography. § 19. 281. The Greek ohi sometimes became g in Latin, as in GAL- BANUM, ANGO, LINGO, CULIGNAJ and H, aS in HUMOR, HERES, HiEMS, Hio, HIRUNDO; but more frequently c as in oroa, lan- CEA, SCINDO. 282. In old Latin the Greek chi was replaced by cay, as in the inscriptive forms braoio, bacas (BAOchAs), baoanalibus, ANTioco (antioco), Subsequently pronounced ANTioohu by the Greek scholars, who were numerous in Rome. So the proper names chLOE and chARMOSYNE were Latinized at an early period into CLOE and carmosne. CH *283. followed by h (§ 194) is found in the Latin words OHORS (from cohors), pulohee (also sculptured pulcer), an- 56 GUTTURAL CONSONANTS. — GLOTTAL CONSONANTS. CHORA (but Marius Victorinus considered ancora the more cor- rect word) and lachrima, which seems to be less proper than LACRIMA. The following forms are taken from three different inscriptions, a line from each : MISER . QVTD . GEMIS . ET . LACRIMAS IGITVR . LE'CTOR . LACHRIMES TRIE . LACRVMAS . POS, 284. Plutarch introduces chi into the Greek form (nii\xf?) of the name pulcher, following a Greek rule (note 211), or mis- taking every union of c and h for the Greek chi. The same thing occurs in the Greek form of gracchus or graccus, from which [h] is rejected by Varro and some inscriptions. 285. The power of c and h is doubtful in Punic proper names like BARCHA, BOCCHUS. 9. GLOTTAL CONSONANTS. Q 286. The Oriental Qof and Greek qoppa (9) is a very ancient character; its form in Egyptian hieroglyphic writing is the figure of a man's head and neck, the latter being for some time repre- sented by a vertical line (as in the Greek cppoTQN for kroton; 9OPIN0OS, &c., upon old coins), which degenerated into the ordinary appendage. In Hebrew and Samaritan, which are writ- ten from right to left, the tail is placed upon the left side. 287. The sound, a glottal k, was found in Hebrew, Phenician, and Zend ; and exists in Hindoostlini, Arabic, Persian, Coptic, Armenian, and Gurgistanic (Georgian). It was not a Latin sound, although it was probably found in Italy, judging from Etruscan monuments. Modern scholars use [q] to represent this sound. 288. If Qof had been a Roman consonant, it would be repre- sented in certain words of Eastern origin, as coRNti a Jiorn (He- brew and Arabic qarn); qorban a, gift, a Hebrew word introduced by Saint Mark vii. 11; csMiNiis a. furnace (Arabic qamin. Per- GLOTTAL CONSONANTS. 57 sian QUMIN a chimney); CSEXBUS a crab; (Arabic xqrab a scorpion) ; CYMSTIUM wavy carved work, from the Greek kyma a wave, and Hebrew QUM to rise. To these probably belong CXTus, Arabic qitih a cat; and vicix a vetch; Greek bikion, Arabic BAQL. §§168,172,175. 289. In the following proper names, qof and not cay occurs in the originals : bamasqus, isaaq, jaqob, qain, q£muEl, XMA,l6q (a as in fall), aqeLD^ma, Acts i. 19. *290. The use of [q] in Latin was to indicate that the follow- ing [v] represented a consonant, but it became as useless as K when the two forms [v, u,J were introduced, so that its use should have been relinquished, as in the ancient examples acvae for AQVAE, cvo for Qvo, and cvando for qvando. The last word is retained in the Spanish cuando. 291. The character coo was of little account in indicating the consonant v, because the difficulty was not obviated in cases like uNGVENTu™ an unguent, sXNGVis blood, SVXDEo to persuade, so that regularity in orthography as well as ancient authority would allow it to be rejected. 292. On account of its identity with o in Latin, some ancient authors did not consider coo a letter, and Vellius Longus states that QVis might be written cvis. The celebrated orator and friend of Cicero, Lioinius Calvus, avoided this character, and Nigidius Figulus made no use of k, q, nor x. 293. In a few ancient examples [u] followed [q] as the repre- sentative of a vowel, as in laqus, q«ea, qtjrvus, pequnia; for lacus, etc. It would have been etymological in XBXQiis and SYQXMOBtrS, H 294. H is the liquid of the glottal contact, in which the requi- site amount of interruption is secured by emitting the unvocalized breath with a certain velocity, and not by reducing the vocal . passage. It is heard in the English and German syllables hut, hat, held. 295. This sound has become almost entirely obsolete in the Bomish languages, which might lead us to believe that it did not 58 GLOTTAL CONSONANTS. exist in Latin. The ancient authors, however, speak of it as an aspiration, and no exception is made for words like honor or HONOS honor, HoRS an liour. 296. In XRioLATio or hariolatio, and some other words, h was indifferently used or rejected; whilst AVE hail! was esteemed more correct than have. 297. The following are a few of those words likely to be con- founded by the Fr-ench and Italians, whose vernacular has lost the typical aspirate : XBiTus a departure. HABiTus condition. AC and. HAO this way. AMATOR a loven: HAMATOR a deceiver, ARA an altar. HARA a hogsty. ELiCES gutters. HfiLiCES spirals. EV ! well done. HEV, HAV ! alas. OSTIA doors. HoSTiA a victim. *298. "There seems no good ground for supposing that the sound of H was ever suppressed by accurate speakers." — O. Walker. The rule (§ 19) would require h to be pronounced after a vowel in the interjections OH, PROH ; as in the Bohemian word lehhy ; the Konza word for nose, which is the English syl- lable paw followed by H, as if pa'h; and in the Hebrew words GOMORRAH, MILOAH, MAChpELAH, REBEQAH, LOTH Lot, NINE- VEH, EPhAH. 299. H occurs rarely after N, as in ANHELO to gasp ; and in the proper name panormus, which stands panhormus upon some coins, as synhodus stands for stnodus in an inscription. 300. Catullus condemns the affectation of saying hinsidiab for INSIDIAS, and ohommoda, that is o'ho-mmo-da, for commoda. NOTES. Page 5. As examples of the elision of syllables in poetry may be cited mtjltum ille et teeris, which is read mult' ill' et TERRiSj ULTRO ASiAM, read tjltr' asiam. So we find scj'- ABSURDE for SCIO ABSURDE; xVaRBITRATTJ for TTJO ARBITRATU; m'ercle for MEO herctjle ; and mj'ajc for mea haec. So in Italian poetry we find hewe for heeette; cap6i for cwpilli; cor for cdgliere; dicestu for dicesti tu, &c. JPage 17. According to Adam's so-called "Latin G-rammar," Prosody " teaches ihe^proper accent and quantity/ of syllables, the right pronurudation of words, and the measures of verse A long syllable in pronouncing requires double the time of a short one. . . In most Latin words of one or two syllables, according to our manner of pronouncing, we can hardly distinguish by the ear a long syllable from a short one." That is to say, if this author were to hear the word ajax or ajax, he could not tell whether the speaker used a long or a short initial. It is well that Prosody teaches the proper quantity of syllables, because the false gram- mars do not; the rules just quoted, and all others in the same book, being contradicted, superseded, and rendered'worthless by the first paragraph, in which the learner is informed that "Latin should be accented and pronounced by us according to the pre- vailing analogies of our own language, without regard to the prosodial accent and quantity of the ancients." Hence if "our own language" is Italian, sine die contains four syllables and as many vowels; if Grerman, three syllables and vowels; and if English, two syllables (like in tine) with a dipthong in each, such being the prevailing analogies; and as these do not exist 60 NOTES. between Latin and Armenian writing, this rule will prevent Armenians from reading Latin at all, although they learn it. English analogies will allow a pyrrhyc to be read as a tribrachys (as in haX-tal-ion or bat-ta^-i-on, in the annexed foot note*), and a fully pronounced dissyllable to be used as a monosyllable, as flour, flower, bower, heav'n, &c. The quotation will show whe- ther it is correct to say that "In English heroic verse, every line consists of ten syllables, five short and five long." Quantity being a matter of the voice, the varieties of Latin poetic feet must be judged by the ear (so at least thought Hora- tius and Cicero), that being a long syllable which is long in pronunciation. Hence in English reading, the feet in hominibus and FoRTissiMus are identical, because both are pronounced in the same time, instead of the two first syllables of the latter occupying the entire time of the former word. It is chiefly inat- tention to quantity which annoys the student when he is learning to distinguish a proceleusmaticus from a dispondens or a dijam- bus; or to determine whether the nature of armament JRiuM is molossidactylic or dispondentipyrrhic, to be of no use to him when known, if he makes it antibacchiodactylic, by reading ac- cording to false quantity; especially if his Hudibrastic teacher is satisfied with tragododidascalicological names of the feet, rather than with the feet themselves.f * In the following example to is marked short, whilst too or two would be long ; and throng is short when compared with wrong. The succession of short syllables in the first line, and of long ones in the second, conveys the idea of a rush followed by greater composure when the battle begins. The first is rapid, and the second deliberate, so that and and men might have been marked as long. S3 I tS thS I fight the | thick bSt- I taliSns I throng, Shields I urg'd on | shields and | mgn drove | men a- | long. Pope's Homer's Iliad, iv. 485. t H' had bird words r&dy to show wh^ And t611 what rules he did it b^; For k\\ a rhetorician's rules Teach n6tbing bilt to name bis tdols. Hudibras, 1. 85 — 90, NOTES. 61 English poetry is written and read appreoiatingly without a knowledge of the Latin feet, with which some have attempted to cripple it. As Latin feet depend upon quantity, and English feet upon accent, the two cannot have the same names. In English there is no difference in the use of what would be a molossus in Latin, as loathesomeness ; an antibacchius, as flamingly; a dactyl, 5iS harmonj^j and a tribrachys, as pitiful; one being capable of replacing the other if the proper accent be preserved, as pUif'Bl enemies instead of the two first feet in the following approxima- tion to an English scansional hexameter, compared with one of Virgil's : but although this may be done, the iambus billow can- not replace the iambus bel5w.* "SEMX Vi- Ru"QVi6 ox- NO TROJ 5; QVi PRIMUS XB ORIS." arms and the hero i sing who first from ilion's borders. Literary people fancy that — " From the low | pleasfires Sf | this fallen | nature ..." is an example of dactylic verse, although low is as long as the next syllable; this is shorter than fall; and the first syllable of nature is as short as the second, or equivalent to fate^ which is as short as fat, and shorter than fane. This example, like my own, has the natural or prose accent at the beginning of each foot, which is not the case in Latin. The following line will therefore give the unclassical reader a better idea of Latin versifi- cation, in which it may happen that a foot (like the fourth) is without a natural accent. This accent is marked in the example, the long with a grave accentual, and the short with an acute one. Boreas is given as an English word. Stdrms and billlows and h6r|rors thr6e|fold thTolBdreas'lwMlings. This is prose, and if Latin verse was recited with the prose accent (and the Italians recite it in this manner), the listener could not distinguish it from prose, except by the quantity, and * The I billows ] float in I order I to the I shore, The I wave be- 1 hind rolls | on thS I wave be- I fore. Pope^ Homer's Iliad, iv. 480. 62 NOTES. for this accent Bentley contends ; ■whilst Mekerchus was in favor of a mode like scanning. The author of "Living Latin" would have dactyls, anapests, trochees, and iambics accented on the long syllable; tribrachs, spondees, andpyrrhics to take the accent upon the first syllable among dactyls, &c., and upon the last in iambic or anapestio verse. The ancient grammarians leave this question in doubt, and amidst such conflicting opinions the least objecti^^r able mode seems to be to pay strict attention fe) quantity, and to avoid the use of accent. § 25. Dr. Gaily, a literary person of the last century, who has many followers at the present day, asserts incorrectly that "No man can read prose or verse according to both accent and quantity. For every accent, if it is anything, must give some stress to the syllable upon which it is placed; and every stress that is laid upon a syllable must give some extent to it, for every elevation of the voice implieth time, and time is quantity." An unlettered person could not have fallen into such an error. In the Latin Grammar of the Rev. P. Bullion, D. D., N. York, 1843, it is stated that "In English every accented syllable is long." If these views were correct, there would be no diflference in quantity between tarry to remain, and tarry from tar. See paragraph 47. § 5. The names at the end of this paragraph are given in the original character to enforce the views of the preceding ones. The first name is St.- Petersburg, the second el Medina, and the third Canton. Via. The tongue being in a manner wedged into the throat, its base has not so free a motion as its apex, so that palatal letters are more easily made than gutturals. I do not assert that excep- tions to the rule of change from guttural to palatal do not occur, although I believe that some of the cases that may be cited are more apparent than real. 6. In accordance with a notion that English orthography must be etymological, I have avoided the use of words which seemed to indicate a change from palatal to guttural, as in "nuiSanCe" from noCeRe (§ 220.) I have consequently been compelled to use the original form of horatius, and to avoid the English form of PRAEFATiOj^DEFENsio, &c. The Corrupt use of [o] in English NOTES. 63 has caused authors whose classical knowledge could not be doubted, to use this character in writing insessores (perhaps as an English word) and sujpercede. Noah Webster says that to spell sigdr with [s] is a mistake of the grocers. It is evident that the mistake in writing "defence," &c., is grosser. 18a. The Sanscrit SAiCAs irrigation (whence the Latin suCOiis juice) is said to be from the Sanscrit (in English letters) seefsh, but this is probably rather modern, and it is more probable that both are from a lost form with a final cay, because the English palatals dzh and tsh are so common in Sanscrit as to induce the belief that cay and giai/ must have disappeared from many words before they were written. b. The Sanscrit SNU to sprinMe is supposed to have given rise to the Gothic snaiws and old English snow, &c., yet the Russian snieg, the Gaelic sneachd, and Irish SNSchio, have a guttural which must be looked for in some ancient collateral dialect. The Sanscrit root of the Greek AOo has the double form AG and (in English letters) adzh, of which the former must be the older, and therefore the true root. The initial of the Sanscrit analogue of cool, GELO, GELIDUS is English dzTi, although it was probably GAlilTAs at an earlier period. The Latin malignus is evidently older than the Sanscrit malinas, which probably lost a gay. The arguments used in this note tend to disprove views like the fol- lowing: "Perhaps f was retained because the original dialectic sound dsi passed over, among the Greeks, into Jcsi." — Buttman's larger Greek Grammar, by E. Robinson. Andover, 1839. c. The English words quacle, cuckoo; with the Latin COAXO, CUCTJLO, and their Greek forms, cannot be derived from the" Sanscrit OATsh (to which they are referred), because this must be a later form of KAkh to cry and to laugh, which gave rise to chucMe, giggle, gaggle, cackle, chicken (Angl. CICEn), cheek (Dutch hek), cough, hiccup; and probably cake, coke, cook, ciichen, if the original idea is connected with the noise of cooking. d. Although the predominance of palatals in Sanscrit where the Greek preserves the gutturals, leads to the conclusion that in certain points the latter is the older form, this need not prevent us from considering AGO as a derivative of AG, bejjause the latter 64 NOTES. probably occurred in languages which preceded the Greek. On the other hand, the absence of Sanscrit words like GALITas is a strong argument against its great antiquity under its present form. 21. "AVLAS ANTIQVI DICEBANT, QVAS NOS DIC-IMUS OLLAS." — Festiis. 26. The following extracts are given to show how imperfectly the rudiments of grammar are defined: "Sentences consist of words; words of one or more syllables; syllables of one or more letters. A letter is the mark of a sound. Letters [marks of sound] are divided into vowels and consonants. A vowel [mark of a sound] makes a full sound by itself. A consonant [mark of sound] cannot make a perfect sound without a vowel; as b. d." — Adam's Latin Grrammar, with Improvements. It appears from this that vowels, consonants, and syllables, instead of being parts of human speech, whether written or unwritten, are merely marks, also called letters; and that the consonauts b, d, 1, m, &c., cannot make a perfect sound, probably because a mark can make no sound, although a human being can, particularly a Sclavonian, who uses entire words without a vowel, as smrt, srp, Term, drhl. Sis speech, however, does not contain "marks of a sound," but the sounds themselves. The next sentence to the last quoted informs us that "A vowel is properly called a simple sound; and the sounds formed by the concourse of vowels and consonants articulate sounds." Hence, a sound and the mark of a sound are identical. The sections in the same Grammar devoted to the dipthongs and consonants are equally confused and inaccurate, which is unfortunate in a work which defines Latin Grammar to be " the art of speaking and writing the Latin language correctly !' 28a. For the forms of the Koman script letters, the Foreign Quarterly Review for October, 1841, may be consulted. h. The first character in Roman inscriptions is not larger than the rest, although a large letter was occasionally used, as I for ii in dIs. The use of characters of two sizes (unknown in most alphabets except modern Greek and Koman) is seen in the fol- lowing copies of parts of ancient inscriptions from Manutius. e. These examples show that the hyphen was not used, and that the only point was a dot separating the words, but which NOTES. 65 was not used at the end of a line, or where the modern period point would be used. d. In print, variations in the form of the characters cannot well be represented. The accentual is probably placed after the character to which it belongs, for the convenience of the printer. It was probably placed above in the original inscriptions. 1 MATV'EA . PER . sTtGIA . MOETe . SEQVAR 2 conIvgi . svo KAKISSIMO . ET . SIBI 3 COnIvgI . OARISSI MAE . B . M . EEC 4 EEC . E . CARISSIMO . PlIsSI MO . ET . SIBI . ET . SViS 5 SIT . TIBI . TERRA . LEVIS . MVLIER . DIGNI SSIMA . vItA . QVAEQVE . TVIS o'lim . perervere're . bonIs 6 D . M D . IVNIO . PRIMIGENIO QVl ViX . ANN . XXXV rVNIA . PALLAS . EECIT COnIvgI . KARISSIMO ET . PIENTISSIMO DE . SE . BENEMERENTI CVM . QVO . VIXIT . ANNIS XV . MENSES . VI DVLCITER . SINE . QVERELLA 32. In Bullion's Latin Grrammar, it is stated that "The Latin alphabet consists of 25 letters, the same in name and form as the English, but without the w." The same author discusses the letters under one head and the vowels under another, and under 6* 66 NOTES. the latter we are told that "A vowd is a letter, which represents a simple sound." 36o. Being founded upon organic laws, this table may be made a useful element in the construction of Grammars ; and it will be found a more important aid in etymology than any system of false orthography. This will appear in tracing the following words: — Greek D A K E Y M A Latin porous Latin L A C R I M A Germ. F e R K e t Gothic T A G R Welsh p £ R ch i ih French L A R M e Dutch B i Q Anglosaxon T e A E Eng. p i G English T I R French P R c "Welsh D a I ayR Irish M u • c « D U R, water. Gaelic m u ch The German p i che n Ui fojsple (or topt), and p e g eln U) dip into (dive), give the English p i ck e I a steeping liquid, the Scotch p i gg i n a rf^ping vessel, the German 6 e ck e n whence basin {pitcher?), and the English m u g.* Compare Margaret and Peggy, h. The Roman v (English w) is aspirated in the English word when (wh-w-e-n), l and R in Welsh, j (English y) in Cherokee, and in the English syllable B.ugh, hue or hew (yh-y-u)._ The aspirate of D is heard in this, and of Q in Dutch, and sometimes in German. 39. Prof. Anthon, in his edition of Zumpt's Latin Grammar, considers the short e in mete long, and a Professor of Latin and Greek has expressed to me his doubts as to whether the last syl- lable of deceit is any shorter than that of marine, redeem, &c. The former, as I pronounce it, is one fourth of a second long, and the latter is not less than half a second. Any musician who is an accurate timist may decide between us. This view of a short vowel was published by me in a review in 1846. 45. The Italians, whilst they give the same quantity to viTTA and viTA, give a distinct pronunciation to each consonant indi- * Mug — "I know not whence derived." — Webtter. NOTES. 67 cated in the former, so that the two are perfectly distinguishable in pronunciation. 47. The French (and some other nations) have introduced a corruption in using the accent marks to indicate distinction in sounds, as between de and dL 48. These rules for the accent of prose are those of Quintilian, who says besides that a final syllable is not accented. Subsequent grammarians, however, cite exceptions, which Scaliger thought unworthy of attention. Nevertheless, when an author or editor places an accentual in writing rect6, MAL^, pennS,, it is to be understood that he wishes these words to be accented accordingly. In my work on the Freshwater univalve moUusca of the United States, Philad., 1842, and Monograph of the genus Leptoxis (in Chenu's Illustrations Conchy ologigues), Paris, 1847, I have used accentuals in the Latin descriptions; but in the later crtpto- OEPHALINARtJM BOREAM-AMERICAE DIAGNOSES CUM SPECIEBUS NOVis, etc., I have made no use of them, nor of the combined [m, (e], and I have conformed to the European practice of writing adjectives like PENSiiiVANious with a small initial. In deference to the journal in which the latter was published, it is printed in the European character, a Latin character being used at the begin- ning of a sentence, although in Latin typography in the European character, a small initial may be used after a full point, as prac- ticed by Lipsius, and to some extent in German. 58a. Although this is not the place to treat of Grreek pronun- ciation, I may be allowed to give a few words upon the eta (h), which I believe to have been the pure Roman e in vein. From the formation of this vowel it is more closely allied to A than I is (the latter of which the modern G-reeks consider eta to have been), and in allied dialects it would be more likely to change with A than if it were identical with I. We accordingly find eta in the Ionic words hElios, AThEwAi, ThEsEUS, sophiE, ThoREcs, &c,, and A in hAlios, AThANAi, ihAsEUS, sophiA, xhoRAcs, in Doric. The Attic in the three first agrees with the Ionic, and in the two remaining ones with the Doric. So we find ».o|f s9ot and ^ijltyflot, from %ayxava. Compare ogo, agij; aif, viil Xiftiiv, Xijiivosi rtoiea, rtOHjjB; xr}^ for xtag, asxtj^tis Or asxs^s, to show 68 NOTES. that »j liad a strong relation to i and o, so that it could not be identical with i. Cratinus says the cry of the sheep is B^. b. The old Doric form ga (the earth) became GE in normal Greek, and this has become Gi (or Ji) in the modern pronuncia- tion, by the closing of the organs, just as "the Latin CLARtrs (G-erman klar) became daire in French and dear in English. 69. LONGTJM AVTEM PRODrOTIS LABIIS, RIOTU TERETI, eto. — Victorinus. 73. U ORE CONSTEICTO LABRISQVE PEOMINULIS EXHIBETER. — Oapella. 80. Y APPRESSIS LABRIS BPIRITUQVE PROOEDIT. — Capella. 83. That the G-reeks represented u by m or s is proved by their orthography of Roman names, as tibtir Ti)3aga, eegulus, Ptiys^os; ALBULA, AjL)38>.a; NOVUMOOMUM, No(38^*o/»«/it J VALE- RIA, SaJitgias. The following are Greek versions of Latin names in Britain: Stjiyouat selgovae; Novaviai, novantae; Aijbowo DEVANA; Stsei'ofta VICTORIA; ^axofiayov VACOMAGIJ KaffOinot OARNABir; Hevita VENTA; 'Pstsrttoi RhUTUPIS ; Angotgiyig DU- EOTRIGES. The English often pronounce u as i/ou and « as in rownd ! whilst an English scholar, as if to impress this barbarism permanently upon the Greek, uses s for the dipthong in roimd, in his phonetic English alphabet. 96. HEJO occurs in the following vertical inscription (Manutius, p. 113. Aldus, Venetiis, 1566) : — B s I s T A A H 8 E TJ I N T 101. "M OBSOURUM IN EXTREMITATE DICTIONUM SONAT, VT TEMPLUM; apertum in principio, tit MAGNUS; mediocre IN MEDiis, UT UMBRA." — Priscian. 110a. "DIPHTHONGI AVTEM DICUNTUR, QVOD BINOS PHTHON- NOTES. 09 GOS, HOC EST, VOCES COMPREHENDTJNT. NAM SINGULAE VOOAIES SUAS VOCES HABENT.'.' — Priscianus, lib. i. h. An inverse dipthoDg is where the eoalescent precedes the vowel, as in the French words oie, trois. This peculiarity is con- fined almost entirely to the French language, which wants the ordinary or direct dipthongs. 111. A dipthong is etymologically and practically a double sound, and has nothing to do with the number of characters used in representing it. Yet the French continually speak of words liie their au, eu, being dipthongs. The English word aisle is composed of one dipthong and one consonant, and ail of a vowel and a consonant. The Abb6 Sicard is in error in saying that eau is a word composed of vowels, because eau (au or 6) comprehends but a single vowel or continuous voice. In the Spelling-book of Wm. D. Swan, the word heat is said to contain a diphthong. 117a. I have noticed the peculiarity in the nXdaoo language of Texas, of the Latin dipthong AV being co-existent with the dissyllabic AU. 6. This distinction is rarely recognized by grammarians. Since the text was printed, I have heard the dipthong oj in Portuguese, as in oJTo (in Latin characters) eight. 130. "Alpha semper atqve Iota qvem paeant graecis SONI, A ET E NOBIS MINISTRANT." — Terentianus Maurus. "In LATINO RTJRE HEDUS, QVOD IN URBE, TJT IN MULTIS, A ADDITO HAEDUS."— FaiTo. 148a. V, LITTERAM QVOTIES ENUNCIAMTJS, PEODTJCTIS ET COEUNTIBUS LABRIS EFFERIMtTS. — Yictorinus Afer. h. "The Umbrians and Oscans distinguished between u and V. The latter was a consonant, and was pronounced like our w." "y must have corresponded to our English w." — Donald- son's Varronianus. 150a. The following strange argument has been adduced to prove that English and French v existed in Latin. " The Latin ear was certainly too delicate ever to have suffered the pronuncia- tion Wox Wentus instead of ventiis, which it seems to me would have been as strange to them as Woice for Voice, Went for Vent, 70 NOTES. Winegar for Vinegar, do to a well-bred person now in England." — H. Bonnycastle, Classical Museum, No. 23. b. It seems from this that v is more of a well-bred sound than w, so that of the two words from the same root, wine and vinegar, the former would be the more vulgar; and that the Germans who acquire English v sooner than to, throw a "well-bred" and classical air around their English in saying vind and varm instead of wind and worm. § 154. c. Supposing the Romans to possess English v, and the power of English w to be doubtful, the force of this argument may be tested by paraphrasing it for a Roman grammarian in this manner : "The English ear is certainly too delicate to suffer the pronuncia- tion worm instead of verm (from vermis). We know, moreover, that in French the power of the characters w and v is identical, and that in German (whence the English probably borrowed it) the character w does not represent the semi-vowel contended for in English, so that the verb went was probably identical in sound with the noun vent." d. When an unusual sound, or a sound used in a mode to which we have not been accustomed, offends our prejudices, we are apt to persuade ourselves that our taste alone has been offended. e. If English w is less pleasant than English v, r must be less pleasant than English z, the relation being about the same (§ 220) ; and whilst it accounts for the two English forms hurrah and huzza, it shows that to prefer the former is like preferring wine to vine; whilst to prefer the latter is like preferring vine to wine. 163. The proper character for English v in the Roman alphabet would be that of F with the middle line crossing the stem, so as to form a Carvilium. § 247. 165. s. appears in the following inscription, from Manutius ; — OOTAalAE C L A V D I I OAISARIS AVGVSTI.P.P P I I, I A B 185a. " With regard to the Greek