MICIUV 
 
 LIBRARY 
 
 UNIVERSITY OP 
 CALI#OtNIA 
 
/4 
 
 REPORT 
 
 PRESENT STATE OF OUR KNOWLEDGE 
 
 OP 
 
 LINGUISTIC ETHNOLOGY, 
 
 MADE TO THE 
 
 AMERICAN ASSOCIATION TOR THE ADVANCEMENT 
 OF SCIENCE, 
 
 AUGUST, 1856, 
 
 BY 
 
 PROFESSOR S. S. HALDEMAN. 
 
 PUBLISHED FOR THE ASSOCIATION, 
 BY JOSEPH LOVERINO, 
 
 PERMANENT SECRETARY. 
 
 1856. 
 
 MMMM 
 
 „s£\ 
 
[After the following Report was presented to the Association, the 
 author was commissioned to continue the subject in a further report, 
 4o-be~ f>rcsentod at Montreal in August, 1667, upon a system of alpha- 
 betic notation adapted to American and exotic languages. He would 
 therefore respectfully ask the advice and co-operation of observers 
 and scholars who take an interest in this intricate subject. Communi- 
 cations may be addressed to him directly, at Columbia, Pennsylvania, 
 or to the care of John Penington and Son, Booksellers, Philadelphia ; 
 or to the Smithsonian Institution, Washington City.] 
 
 CAMBRIDGE: 
 METCALP AND COMPANY, PRINTERS TO THE UNIVERSITY. 
 
f'zn 
 
 ,3 
 
 REPORT. 
 
 This Report will be restricted to the portion of the subject 
 pertaining to speech, — a portion which, although less extensive 
 than the grammatical, lexicographical, or etymological portions 
 of language, has not attracted proper attention until a recent 
 period. This neglect is due to the difficulty of the subject, of 
 which a constant example is at hand in the difficulty experi- 
 enced in pronouncing foreign languages properly, even when 
 they belong to the same stock, as Persian, German, Belgian, 
 and English. The difficulty of pronouncing, appreciating, 
 locating, explaining, and writing down the various phases of 
 speech is so great, and there are so many sources of error, that 
 we must be more cautious in accepting statements here, than 
 in other sciences of observation, few having as much educa- 
 tion in this branch as would be required to make a chemist or 
 a musician ; or to enable a singer to write down a song prop- 
 erly, even in a notation of his own invention. We cannot 
 even trust an observer who claims for himself a good ear. 
 The English lexicographer, Knowles, makes such a claim, and 
 by his analysis proves that he does not possess the power to 
 discriminate sounds ; as in the case of ye and ivoo, which he 
 considers equivalent to the repetitions e-e and 00-00. The re- 
 1 
 
 072 
 
^ ON THE PRESENT STATE OF OUR 
 
 porter is willing that the objections which he makes to the 
 results of others should be urged against his own ; and that 
 his assertions should be received with as much caution as 
 those of any observer, having at various times held views which 
 further research proved to be untenable. 
 
 Spanish grammarians emphatically deny that their b ever 
 partakes of the power of English and Spanish v. They claim 
 both b and v, and assert that v (like /) is made by the appli- 
 cation of the lower lip to the upper teeth, and that their b is 
 never made thus, the lips alone being concerned in its produc- 
 tion, so that it is impossible that it can be anything else than 
 a genuine b. The facts are as here given, but the inference is 
 false. The Spanish b between vowels, the German w, and, 
 according to E. A. Sophocles, the Ellenic* £, differ from Eng- 
 lish v in being formed with the lips alone. It is therefore an 
 aspirate of b ( C -B) as the Greek <f> is an aspirate of p^ which / 
 is not. This is an important point in ethnology, which few 
 attend to. Doctor Lepsius does not allude to it, nor to the 
 very distinct Russian vowel bl, in his recently published Stand- 
 ard Alphabet. Authors continually confound English labio- 
 dental v with German labial w, and in giving an account of 
 the languages they investigate, they cannot be trusted upon 
 this point; so that we have yet to learn which of the two 
 sounds is present in certain languages, the phonology of which 
 is apparently treated with great fulness. Judging from a par- 
 tial investigation, the Russian ' B ' has the power of English 
 and French labio-dental v. 
 
 Some writers (as Le Brethon and Marsden), in comparing 
 the French and English vowels, refer & to that m/all, and e to 
 
 * Modern Greek is an awkward expression, and Romaic is incorrect, and as the 
 language has lost h, Hellenic has become Ellenic. The word Anglo-Saxon is equally 
 awkward, and degenerates into " Saxon," — a name which should be restricted to 
 Plattdeutsch in its modern and ancient or old Saxon form. The change of languages 
 involves a change of pronunciation, as in the German klar, which closes to clair in 
 French, and still farther to clear in English ; so the language of Anglia and the 
 Angles was Anglish, and passed through English, with e in met, to the modern 
 Inglish, with the vowel in Jit. 
 
KNOWLEDGE OF LINGUISTIC ETHNOLOGY. 3 
 
 that in end ; whilst others (Picot, Bolmar, &c.) refer & to the 
 English vowel in arm, and e to that in fate. Duponceau, in 
 1817 (Am. Phil. Trans. I. 229), refers the English vowel in 
 fat to the French vowel in terre, pere, an error which Germans 
 commonly make. In the alphabet of Lepsius, fat would be 
 written fet, and the French mere mer. Duponceau seems to 
 have been the first to show that the initial vowel of the Eng- 
 lish diphthongs in aisle or isle, and owl, is not that in arm, but 
 the French a, which is made with a narrower aperture than a 
 in arm requires, and with a slight tendency towards awe. The 
 same author gives as the components of English u in usage, the 
 vowels of eel and ooze, — an error which is retained by most 
 English authors, very few of whom know what a diphthong is. 
 Every vowel added to a word forms an additional syllable, 
 and as English u is a monosyllable, one of its elements is a 
 consonant ; namely, the initial when it is pronounced you, and 
 the final when, as is sometimes the case, it is a diphthong pro- 
 nounced like the Welsh iw and Belgian iew, with the vowel of 
 it (German hitzig) and iv as a consonant in now, or German 
 u in haus. Similarly, the final element of clo-y is a consonant, 
 and of claw-y a vowel.* German and French writers seem 
 not to be aware of the nature of diphthongs, and in Latin- 
 English grammars they are described as vowels.f In most 
 ethnic alphabets, including that of Lepsius, the last element of 
 the diphthongs is represented by a vowel character, — a fun- 
 damental error founded upon the crude analysis of the an- 
 cients. % The term diphthong is itself almost useless, because, 
 
 * Hald. Latin Pronunciation, §§ 109, 111. Latham's English Language, 1841, p. 
 108, §68. 
 
 t Andrews and Stoddard consider the vowels of fall andyeeJ diphthongs, because 
 they pronounce the Latin LAVS and AETAS with them. They say that " two 
 vowels in immediate succession in the same syllable " (including UO, UA !) " are 
 called a diphthong. Yet if AETAS is read with but four elements instead of five 
 (the a in arm being omitted), the word does not contain " two vowels in immediate 
 succession." 
 
 \ Dr. Lepsius assigns (Standard Alphabet, p. 41) to the Latin diphthong oe (oi 
 in going, when pronounced as a monosyllable) the power of the German vowel o, and' 
 
4 ON THE PRESENT STATE OF OUR 
 
 as the first element is already a vowel, the peculiarity lies in 
 the second coalescing with the first, (an impossibility with a 
 second and subsequent vowel,) whence it may be called a 
 coalescent, meaning by this term those consonants that approx- 
 imate as nearly as possible to the vowels. 
 
 In a French work on Russian, the twenty-ninth Russian let- 
 ter, b, is explained by comparing it with the French " e mute," 
 whilst the twenty-seventh letter is said to have no sound, but 
 to indicate that the preceding consonant is to be pronounced 
 with force, and as if it were doubled. This gives a very in- 
 correct view of these letters. For example, the Russian word 
 for Jive is a monosyllable with the short a in mi^ which might 
 be represented pj^tj in Latin or German letters, and pyrfty in 
 English letters, the final y being the Russian (b), and the mod- 
 ified y-sound following I in the French "11 mouillee." This 
 addition to consonants is so common in Russian, that its ab- 
 sence is marked by the yerr, as in the word o-ke-an ocean, 
 which is written with the final yerr. The allusion to doubled 
 letters might cause Russian to be associated with Arabic, 
 Latin, and Italian, which are among the few languages which 
 have doubled or geminate elements, as in the Italian " Gio- 
 vanni," John, in which each n is as distinctly sounded as in the 
 English words one name. 
 
 Arabic has such doubled consonant sounds ; yet it would 
 be wrong to consider the sixteenth Arabic letter tta such a 
 gemination on the authority of Richardson, who describes it 
 (Grammar, p. 9) as "double t, or t with a slight aspiration," 
 a description which is void of meaning, the reader being un- 
 able to tell whether the conjunction is copulative or disjunc- 
 tive, and consequently whether the latter member of the sen- 
 tence is explanatory of the former. Brown (Journey to Dar 
 Fur) assigns to the Dar Runga, words like tta water, mmi wo- 
 
 ignoring the Latin nasal vowels, writes kplum for coelum. The dots are placed be- 
 low o to afford room above for accent marks, yet nasal vowels are indicated by ( " ) 
 above, which may call for vowel characters surmounted with the three signs (" •* ■*). 
 Thus the French word sans (saj is long, and cent (sa { ) is short. 
 
KNOWLEDGE OF LINGUISTIC ETHNOLOGY. O 
 
 man, ddSta mountain, wwi wind, ggo reprimanding ; but as he 
 gives no explanation of his notation, these are doubtful exam- 
 ples of geminate consonants. 
 
 Many English people fancy that they have double conso- 
 nants, because they spell certain words (e. g. all, well, off, lesson, 
 back, annex, allude) w T ith double characters. This practice is 
 in use to indicate a short preceding vowel, and when an Eng- 
 lishman writes a Latin Grammar, he is apt to believe and as- 
 sert that syllables which are long " by position " are really 
 short, but are "counted" or "considered" long, by a "me- 
 chanical rule " ; whereas, the doubled consonants heard in 
 Italian show that such syllables are really long, because it 
 requires more time to pronounce two elements than one.* 
 Similarly, Latin diphthongs are long, not by an arbitrary rule, 
 but because the two elements of AV, AE, OE, &c. require 
 more time than A and O alone. 
 
 The nasal vowels of but few foreign languages are properly 
 understood, and the ignorance of writers whose vernacular 
 does not contain them is frequently apparent. The error here 
 is sometimes so great, as to cause a confusion between vowel 
 and consonant, as in mistaking the nasal vowel of the French 
 fin for the English and German ng in fang. The French 
 word fin contains but two elements, a consonant followed by 
 a genuine vowel, whilst the English fang has three elements, 
 and ends with a consonant. The mistaking one for the other 
 would be exactly paralleled in the practice of a foreigner who 
 for pea would give peag as an English word. 
 
 This error in regard to the nasals appears in Riggs's valuable 
 Grammar and Dictionary of the Dacota Language, based up- 
 on the studies of a number of observers during a period of 
 eighteen years, and submitted to the inspection of a learned 
 committee previous to its publication by the Smithsonian In- 
 stitution. To a given character the power of 'n' in the French 
 
 * Most English grammarians do not distinguish between the length and the 
 quality of their own vowels, regarding a and o in fate and obey as long, and those of 
 fat and object as short, although, in these examples, the quantity does not differ. 
 1* 
 
D ON THE PRESENT STATE OF OUR 
 
 bon and English drink is assigned, so that the reader is unable 
 to pronounce with certainty the numerous words represented 
 by this character, — an l n ' the second line of which is pro- 
 duced and ends like 'j.' But it is probable that neither the 
 French nor the English sound occurs uniformly, for in the al- 
 lied Conzo (each o as in not, z in zeal) the reporter has heard 
 both, as (to use German characters) in hiing-ga leggins, with 
 u long and accented, and a short as in art. The French sound 
 occurs in the Conzo word for five, which is the English sylla- 
 ble saw accented and followed by t and the French un, as if 
 sawtu n . Using l n y for Mr. Biggs's letter, the Dacota word 
 for leg-gins is huwska, and for five, zaptaw. 
 
 Independently of the errors of observation, some writers 
 have a practice of referring the sounds they meet with to those 
 of other languages which they may know from description 
 alone ; and some proposers of general alphabets supply such 
 foreign sounds with characters, although they run the risk 
 of giving different characters to the same sound, or of con- 
 founding distinct sounds. An English alphabet-maker, upon 
 reading that the Lenape (lenape) aborigines use a whistle in 
 speech, might propose a character for it, although this sound 
 is nothing but English wh before a consonant, as in whte heart 
 (e in they), its occurrence in a new connection conveying an 
 impression analogous to that which a new sound would give. 
 In comparisons of sounds, the reader should be informed 
 whether the author has heard those he uses for comparison, 
 and whether his opportunities have been few or many. Some 
 sounds can be accurately described to those unfamiliar with 
 them, as the German w, Greek phi, English th in thin, then, 
 Welsh aspirate //, rh, which latter, together with the Oriental 
 ghain and its surd cognate (as they occur in Armenian) the 
 reporter was accustomed to pronounce before he heard them 
 from natives. 
 
 The alphabets used by various authors will give a good idea 
 of the state of our knowledge in this department, except that 
 they will not always enable us to establish a parallel between 
 
KNOWLEDGE OF LINGUISTIC ETHNOLOGY. 7 
 
 them. After Rapp's Physiologie der Sprache, the Essentials 
 of Phonetics of Mr. A. J. Ellis, A. B., London, 1848, may be 
 placed, as a conscientious and valuable contribution to the 
 general subject. Being printed in the author's alphabet, its 
 use is restricted to those who can speak English. The alpha- 
 betic portion of this treatise is so corrupt, that it ought not to 
 be used for any language ; but it has an important concession 
 to correct scholarship, in the use of Cay (and not Kah) as the 
 cognate of Gay. 
 
 Castren's Grammatik der Samojedischen Sprachen (St. Pe- 
 tersburg, 1854) contains a careful analysis of the sounds used. 
 The iotacized (mouillees) consonants, or those followed, and 
 in some cases modified, by the guttural cbalescent approach- 
 ing English y in million, are seven in number, /, r, n, t, d, s, 
 and English z, marked with a curved line through the stem 
 (on the right of n, d) of the characters, — an awkward nota- 
 tion requiring too many distinct characters. Ellis uses (j) de- 
 prived of its upper and lower dot (as in lj,) which is unexcep- 
 tionable. The peculiarity of these compounds is, that whilst 
 they commence with I, &c, the tongue passes to the iotacism 
 before the I is completed, million being milyyon when thus 
 iotacized, which, however, is not essential to its purity as an 
 English word. A soft lisped d is assigned to Lappish, which 
 is allied to a lisped r. The latter quality seems to remove it 
 from English sonant th in then. This curious sound should 
 be compared with the peculiar Irish Z, which the reporter first 
 noted as an 1-sound mixed with sonant th y but subsequently 
 determined: to be the sonant analogue of the Welsh surd aspi- 
 rate 11, to which it would bear the same relation that thy bears 
 to thigh. A sound between /and h is mentioned, — probably 
 Greek phi; and a consonant between I and r. Castren assigns 
 a peculiar u to Ostiak, and the Russian vowel bl to Samoiede. 
 The latter (which has been heard by the reporter) may be de- 
 scribed as a long and short vowel akin to English and German 
 i in still* but formed with a more open aperture, and the an- 
 
 * This is not the short quantity of the vowel in Jield, and cannot be correctly 
 represented by (?) of the Latin and Italian alphabets. 
 
8 ON THE PRESENT STATE OP OUR 
 
 gles of the lips drawn back. It has the pinched quality of 
 German o and U, but without the pursed lips used in forming 
 these well-known vowels. 
 
 Bohtlingk (Ueber die Sprache der Jakuten, St. Petersburg, 
 1851) mentions a nasal of the German J. The reporter* 
 mentions such a sound as present in Wyandot, and a close of 
 the glottis ( marked > ) which has since been observed in Chip- 
 peway. Judging from information received from a European 
 who had resided in Syria, this " close of the glottis " is the 
 Arabic effect termed " spiritus lenis" and marked (') by Lep- 
 sius. 
 
 S v unic (in German letters Schunjitsch, an Illyrian), De 
 Vera Orthographia', cum Necessariis Elementis Alphabeti 
 Universalis, (Vienna?, 1853,) admits twelve vowels, which, 
 with marks of accent and length, require seventy-two modi- 
 fications of vowel characters. He supposes that these twelve 
 correspond with the twelve semitones of the musical scale, a 
 view which is fundamentally erroneous. Of the consonants 
 he enumerates fifty, including the mouille kind, and a few like 
 ts y tsh, &c. He assigns to the German w the power of the 
 English w, and considers German b in haben different from the 
 ordinary b, his informant having probably been a provincial. 
 He omits English and German ng, or confounds it with the 
 French nasal vowels ; and his notation is over-crowded with 
 diacritcal marks. 
 
 Poklukar (probably an Illyrian) published a pamphlet at 
 Laibach in 1851, entitled, Ankiindigung eines nachst zu verof- 
 fentlichenden allgemeinen lateinisch-slavischen, zugleich deut- 
 schen, franzosischen, italienischen und eventuel auch eines Uni- 
 versal- oder Welt- Alphabetes, &c. He confounds German 
 ng in lang (Eng. long) with French n in loin; and French v 
 with German w. His notation is objectionable, although he 
 starts with the best possible rule to secure correctness and final 
 
 * On some Points of Linguistic Ethnology ; with Illustrations, chiefly from the 
 Aboriginal Languages of North America. Proceedings of the American Academy. 
 Cambridge and Boston. October 2, 1849. 8vo. 
 
KNOWLEDGE OF LINGUISTIC ETHNOLOGY. 9 
 
 uniformity, by preventing each author from being influenced 
 by the power which a letter may happen to have in the alpha- 
 bet he is best acquainted with. This rule requires that an al- 
 phabet should not contradict the Latin original, Latin being in 
 some sense the ( Weltsprache) universal language. The report- 
 er's Elements of Latin Pronunciation (Philadelphia, 1851) 
 grew out of a perception that, without such an investigation, 
 not a single step could be made in the right direction towards 
 a general alphabet, the construction of which should be based 
 rather upon scientific principles than upon the vagaries of each 
 individual who may be called upon to write a language for the 
 first time. Poklukar uses B, F, J, and other letters, correctly, 
 but by a false assumption he uses C as ts (although he had 
 already a t and an s in his alphabet!) and prefers x to the 
 Latin Cay or Greek Kappa. 
 
 Among the latest works upon the subject* is Professor Lep- 
 sius's Allgemeine linguistische Alphabet, (Berlin, 1855,) of 
 which there is an English version, entitled, " Standard Alpha- 
 bet for reducing Unwritten Languages and foreign graphic 
 Systems to a uniform Orthography in European Letters, &c." 
 (London, 1855.) The profound learning of the author, and 
 the use he has made of his alphabet in the languages of Nubia 
 and Dar Fur, render this a very important work. It has been 
 approved by the Royal Academy of Berlin, which has had the 
 necessary types cut to give the system publicity ; and many of 
 the missionary societies have adopted it, including the English 
 " Church Missionary Society," who have commenced using it 
 in the works of the Rev. S. W. Kolle on the languages of 
 West Africa. Professor Lepsius expresses a hope, that, in 
 cases where missionaries are disposed to make alterations in 
 his notation, "the Committees of Societies will require the 
 reasons of such deviations to be laid before them and dis- 
 cussed." 
 
 * Lauth's Vollstflndige Universal-Alphabet, (Manchen, 1855,) and Professor Max 
 
 Mailer's Languages of the Seat of War in the East with an Appendix on the 
 
 Missionary Alphabet, — have not yet been received. 
 
10 ON THE PRESENT STATE OF OUR 
 
 This system professes to have a physiological basis, and the 
 labors of the eminent physiologist, Joh. Miiller, are acknowl- 
 edged in this field. Mr. Ellis states that, in Miiller's account 
 of the elements, " the faults of a German ear are still con- 
 spicuous." Miiller (Elements of Physiology, English edition, 
 1848, p. 1051) does not understand the nature of his own J, 
 which he supposes a sonant German ch, as English z is a so- 
 nant s. With him, m is not a labial consonant, and he does 
 not know the distinction between p and b. He considers p as 
 having an aspirate quality, probably because an aspirate is 
 made after it, as in pronouncing tap', where a Chinese would 
 say tap\ If p in tap' and haphazard is to be named an aspi- 
 rate from the phase which follows it, the p in pay, play must 
 be a vowel, or the consonant I. If the t in boathook is an as- 
 pirate because h follows it, it is equally an aspirate when it 
 precedes, as in the Iroquois word a'hta. 
 
 Professor Lepsius says that in adna or anda we pronounce 
 " only half the n and half the d, whilst in ana and ada we 
 pronounce the whole of n and dP According to this reason- 
 ing, as n cuts off the first half of d in anda, and the last half 
 in adna, both halves of the consonant between consonants 
 must be lost in lend not, wends, endless, string, warps, and in 
 the German proper names Heindl, Jondl, Zarbl, Birkl, Schmolzl, 
 Dietzsch, &c. 
 
 This mode of regarding a consonant position and a vowel 
 position of the organs as in a manner constituting a unitary 
 element, has given rise to alphabets of a more or less syllabic 
 character, like the Cherokee, Ethiopic, Hebrew, and Sanscrit ; 
 and the system of Professor Lepsius is heterogeneous in ad- 
 mitting t and a for ta, whilst in the Hottentot dental* clack 
 he uses but one character for the consonant position of the or- 
 gans (a kind of t), and the reverberation which follows it in 
 the cavity of the mouth, set in a vowel position. 
 
 * Mentioned here because it is the only one heard from a native by the reporter, 
 who has, however, heard several others in American languages. A change of nota- 
 tion is required to distinguish the clacks formed by sucking in air from those in 
 which it is expelled. 
 
KNOWLEDGE OF LINGUISTIC ETHNOLOGY. 11 
 
 Professor Lepsius considers the vowel in worth as " inherent 
 in all soft fricative consonants," such as English v, in which 
 there is indeed sonancy, but no vowel power, and least of all 
 one requiring such open organs as that in worth. If anything, 
 the supposed inherent vowel in English v is German ii. 
 
 A consonant like /, r, and English z, may have the organs 
 so little closed as to approach the vowel quality, and in this 
 case the small circle placed beneath the character by Dr. Lep- 
 sius is a good mark. But he uses the mark with n and m 
 when they form syllables, although in these cases they do not 
 differ from ordinary m, n. Thus if z, in the Chinese word tsz 
 quoted by him, is English z following s, it does not want the 
 mark. In English, the second vowel of misses, horses, is often 
 omitted in hurried or careless speech, forming the dissyllables 
 mi-sz, hor-sz, as in su-dn (sudden), pri-sm (in which sm have 
 the same power as in pri-smat-ic), German v'r-la-ss'n, tfr-der-Vn. 
 In rare cases a mark of syllabication will be necessary, as in 
 prairie, often pronounced in English as a trisyllable, with or 
 without the vowel of utter in the first syllable, in the latter case 
 forming pr-ai-rie. 
 
 Professor Lepsius follows the English in admitting an " in- 
 distinct vowel sound" in nation, velvet, &c. This is the vow- 
 el of worth and urn, which stands on the throat side of the 
 vowel scale, opposite to awe on the labial side. It does not 
 yield in distinctness to any of the vowels, but as Latin U and 
 V (English w) and I and J are allied, so the vowel in urn ap- 
 proximates English smooth r, and coalesces with it. Hence, 
 if a person pronounces ramrod (ra-mrod), omitting -od, the lis- 
 tener accepts the remainder ra-mr as rammer. This so-called 
 " indistinct vowel " is doubtful as a German sound, being more 
 probably elided in lieWn, &c, as it sometimes is in the English 
 words nation, theatr\ &c. Its resonance may, we are told, be 
 lost " by partially contracting the mouth, or even closing it en- 
 tirely. In the latter case it is heard through the nose." This 
 supposed vowel is the consonant m. The English vowel awe 
 is given as an Italian sound, although this lies between awe 
 
12 ON THE PRESENT STATE OF OUR 
 
 and oiue (Ellis, p. 20). But six labial consonants are admitted, 
 p, b, m,f, and English v and w, no mention being made of </>, 
 Ellenic /3, nor English wh, although English examples are 
 freely cited, and the number of English consonants stated to 
 be twenty-two. English wh* was probably supposed to be 
 English w (Latin v in qvinqve, svavis, &c.) preceded by h, an 
 opinion in which Professor Lepsius was likely to be seconded 
 by the English committee which he met, and this view would 
 probably be sustained by the missionary committees to whose 
 decision he would have the results of original investigators re- 
 ferred, in case they should differ from the views laid down in 
 the Standard Alphabet. 
 
 If such grave errors can take place with the labials, the or- 
 gans of which can be seen and felt, in addition to the sounds 
 being heard, we may well doubt the analysis of sounds formed 
 out of sight, in the depths of the fauces ; and consequently, 
 the following observations are open to correction. Recalling 
 the admission that the reporter has never heard Arabic from a 
 native, yet he is vernacularly familiar with the German smooth 
 aspirate or spirant of gay in re'gen, which is the sonant of ch 
 in ich, is free from vibration, and belongs to the cay contact. 
 The Ellenic *gamma (judged by ear from native sources) be- 
 longs to the same contact, is made with a similar close of the 
 organs, but has the addition of a mild vibration, probably due 
 to the vibrant action of the edge, and not the body, of the pal- 
 atal veil. The French r grasseyee is probably formed by the 
 body of the palatal veil, with perhaps little or no contact of the 
 tongue and palate, wherein it would differ from ^gamma. 
 
 Some of the Oriental languages have a contact behind that 
 of cay, of which qof may be considered the characteristic. 
 Aspirating qof produces a faucal qh analogous to %, and when 
 this is made sonant the analogy is with aspirate gay. Profes- 
 sor Lepsius considers the German aspirate g, Ellenic 7, and 
 
 * In a former paper, the present reporter has affirmed that no orthoepist known 
 to him had been able to state correctly the elements which- occur in the English 
 word when. 
 
KNOWLEDGE OF LINGUISTIC ETHNOLOGY. 13 
 
 Armenian ghad identical, and of course represents them with 
 the same character ; and the surd form of ghad is considered 
 identical with ch in ich, — a greater error, apparently, than to 
 confound cay and qof or the Arabic " spiritus lenis n ( ' ) of 
 Lepsius, with Jain, as he marks it. 
 
 Richardson, in his Arabic Dictionary, says of ghain: " This 
 letter is articulated in the throat with a vibration producing a 
 sound like that given to r by the Northumbrians, or the noise 
 made in gargling. ... It seems to bear the same relation to 
 kh as b to p," — i. e. the relation of sonant to surd. The kh, 
 he states, " is generated by a gentle vibration in the throat " ; 
 consequently it is not the Greek nor German chi. Sir Wil- 
 liam Jones says that " the Persians and Arabs pronounce 
 their ghain with a bur in the throat and a tremulous motion of 
 the tongue, as in making the rough r " ; — and according to 
 S v unic its surd analogue is an aspirate " qui a graeco % chi in 
 eo difTert quod quasi gargarizando efferatur." 
 
 These two vibrant sounds appear in the following Arme- 
 nian words, premising that, as the character for the Armenian 
 ghad resembles an angular " 2," this will be used for it, whilst 
 £ q will represent the surd sound ; a the vowel in under, e that 
 in met ; a that in arm, but short in these examples ; r English 
 sh; c as k; and the objectionable character z as in English: — 
 
 dzndz2a', a cymbal; ^qclc, the mind; ^qatp, a crucifix; 
 ^Q^QaNTreL, a neigh. 
 
 The division of the consonants into contacts is natural, and 
 was appreciated by Aristotle, the Hebrew Grammarians, the 
 Abbe Sicard, &c. A consonant character indicates a closing 
 of the organs, as p, t, f, whether it precedes another effect, as 
 in fay, play; or follows, as in off, or does both, as in eft. 
 Each contact is subject to nearly the same phases, that is, if 
 the closed lips make p, the tongue will make t and its base k. 
 Adding sonancy to these gives b, d, gay; open the nasal pas- 
 sage, and these become m, n, ng. Professor Lepsius divides 
 the phases of the contacts into explosives or dividuce, as t, d, 
 2 
 
14 ON THE PRESENT STATE OF OUR 
 
 n ; fricativce or continues, as s, z ; and ancipites, as r, /. In 
 the labial contact, p, b, m are placed as explosives ; /, v, w, 
 (English) as fricatives, without any ancipites ; although Eng- 
 lish w is to b as I is to d, and English y to gay. Apparently 
 to accommodate the German nomenclature, voiceless conso- 
 nants, as p, f, are termed fortis instead of surd, whilst b and 
 English v are termed lenis instead of sonant. The fricatives 
 are the aspirates of other authors, whether sonant or surd, ex- 
 cept with those who think " sonant aspirate " contradictory, 
 and prefer " spirant " for such elements as English v, z, and th 
 in then. 
 
 English w and y (Latin V, J) are not fricative in the sense 
 of/, s, %, &c, but they become so when aspirated in the words 
 when and hue or hew, the initial of which is in neither case h, 
 as many suppose. Writing hue in Latin letters and marking 
 surd by (') it will stand "JJU or JhJU, English u being 
 normally Latin JU. 
 
 Instead of twenty-two u simple consonantal sounds " as- 
 signed to English by Professor Lepsius, the following may be 
 enumerated : — 
 
 Glottal. 
 
 iabial. 
 
 Dental. 
 
 
 Palatal 
 
 
 Guttural. 
 
 W 
 
 1 
 
 r 
 
 
 soldier 
 
 y 
 
 wh 
 
 ... 
 
 ... 
 
 
 nature 
 
 hue 
 
 m 
 
 n 
 
 ... 
 
 
 ... 
 
 n g 
 
 b 
 
 d 
 
 ... 
 
 
 ... 
 
 g 
 
 ... V 
 
 dh... 
 
 z 
 
 
 zh 
 
 ... 
 
 P 
 
 t 
 
 ... 
 
 
 ... 
 
 k 
 
 ... f 
 
 th... 
 
 s 
 
 
 sh 
 
 ... 
 
 To these twenty-six might be added an r (as some English 
 people use both a rough and a smooth one), and mh for the 
 English and German exclamation hm (really f/ mm), a surd as- 
 pirate followed by pure m. This aspirate is sometimes re- 
 placed by r/ n (found in Cherokee) and *'ng. See Rapp, Vol. 
 II., middle of p. 267, and Vol. I. p. 166, note. 
 
 It is here intended to assign to the English word nature a 
 
KNOWLEDGE OF LINGUISTIC ETHNOLOGY. 15 
 
 surd, and to soldier a sonant effect, allied to y in ye, you, but 
 made at the post-palatal point, and constituting the liquids of 
 which zh and sh are the aspirate mutes, and into which they 
 are apt to fall, just as r may fall into s or z, or w into v or /. 
 Hence nature is often natsh'r, as soldier is soldzh'r. In the 
 mouille effect, the modified y is not only drawn forward to the 
 palatal position, but when it follows a dental, this often recedes 
 to meet it, even when the double effect has become t-sh, that is 
 to say, in order to bring t nearer to y, or to sh, it is often drawn 
 back from the teeth, and the point placed against the base of 
 the lower teeth as a fulcrum.* 
 
 The notation of the " Standard Alphabet " is defective, 
 whilst tendencies towards uniformity are not fostered by its 
 third and fourth rules, which do not regard the purpose for 
 which a given character was invented. The four rules are : — 
 
 I. Every simple sound ought to be represented by a simple 
 sign. 
 
 II. Different sounds are not to be expressed by one and the 
 same sign. 
 
 III. Those European characters which have a different value 
 in the principal European alphabets, are not to be admitted 
 into a general alphabet. 
 
 IV. Explosive letters are not to be used to express fricative 
 sounds, and vice versa. 
 
 The first two rules are proper, the others are exceptionable. 
 The first is broken by its proposer in assigning a simple char- 
 acter for the contact and the subsequent resonance of the 
 Hottentot clacks. The second is broken by representing the 
 English combination t-sh partly by t and the character for sh 
 (p. 55), and partly (as in Arabic and Persian) by k surmount- 
 
 * When the iotacism follows a labial or guttural, it can scarcely be called a 
 mouille effect, although it is so considered in the Slavonic languages. The t and d 
 thus drawn back to a slight extent in the English t-sh and d-zh, would require little 
 to place them among the palatals, when the t would be the lenis of s, and the d of 
 z, or, if nasalized, it would form a sound between n and ng, perhaps the Sanscrit 
 palatal na. 
 
16 ON THE PRESENT STATE OF OUR 
 
 ed by an accentual. In Mpongwe (p. 57) this kf stands for 
 English ty, and on p. 42 if is assigned to ty in case it should 
 be required. Hence, in his alphabets of Kua and Herero (p. 
 57) we do not know whether k' means tsh, ty, or even ky. 
 
 By Rule III. c, ch 9 j, x, are excluded, and most of the char- 
 acters might have been got rid of by the same unphilosophical 
 process. If in the course of time the measures of the French 
 metre should become shortened, partly by the abrasion incident 
 to use, and partly by the file of avaricious dealers, whilst other 
 dealers, with a higher appreciation of strict accuracy, would 
 preserve their measures at the standard value, — if, under such 
 circumstances, the government were to enforce uniformity, those 
 who had allowed their standard to deteriorate would be clam- 
 orous for the retention of their own, as the best known. Strict 
 justice would require that the original metre should be restored, 
 although none but a few just traders might have it in use. 
 
 Notwithstanding the nations of Europe have faithfully pre- 
 served the vowel characters, even to the y (German ii) of the 
 Danes and Swedes,* there are several alphabets of English 
 origin, which (simulating the weights and measures of certain 
 dealers) fall so far short of the standard that every vowel char- 
 acter, even to that of O, has false powers assigned to it, the 
 opinions and practice of those being disregarded who had for 
 ages kept their standard pure.f No sophistry should induce a 
 Danish missionary to pervert a letter (Y) belonging to, and 
 made for, a labial vowel, to the power of a guttural consonant. 
 Philologically, it is worse than assigning to Latin, German, 
 Polish, &c. J, the power of English w. Let Latin ' V ' have 
 its vowel power in ooze, i P that in believe, and ' Y' that of 
 the French pinched u (which bears equal relations to this V 
 
 * Besides the correct use of Y, the orthography of the Danish word " havn," a 
 haven, (rhyming with town,) is strictly Latin. 
 
 t Some English authors have gone so far as to assign to Anglish cay the power 
 of tsh, as if to flatter superficial readers with a greater resemblance to English. 
 Were this view true, the English words broken, kernel, ache, and kin, would be 
 older than the Anglish brocen, cirnel, ece, and cynn, Irish cine. 
 
KNOWLEDGE OF LINGUISTIC ETHNOLOGY. 17 
 
 and I), when we may account for the form of ' Y ' by the fol- 
 lowing diagram of the affinities of the vowels : — 
 
 v 
 
 A 
 
 Y. 
 
 E L 
 
 The rule which rejects C should not retain its cognate G, 
 and that which assigns to the latter its original, standard pow- 
 er in get, give, should have retained C (or at least a character 
 like k deprived of its stem) with its Latin, Gaelic, Welsh, and 
 Anglish power, as Mr. Ellis has done. ' C (aided by ' Q,') 
 is the normal character for cay in the Romanic languages ; and 
 the Latin, German, &c. ch is a concession that, if ' ch ' repre- 
 sents the aspirate %, ' C ' without the aspirate mark h must 
 normally represent its lenis form cay. l C ' is rejected by Rule 
 III., on account of its many perversions, although still used 
 correctly in several languages, whilst ' Z,' with as many per- 
 versions, is improperly retained with a corrupt power. Its 
 powers are as follows: — 1. Ancient Greek, as English zd; 2. 
 Italian dz (and ts) ; 3. German ts ; 4. English in azure ; 5. 
 As s in Hungarian and Danish, and the German c tz ' ; 6. Its 
 French power ; 7. Its Spanish power. 
 
 The normal character for the sonant s in rose, misery, is s, 
 in German, French, Italian, and English ; and as the Latin 
 mode of distinguishing sonant from surd is seen in G, C, the 
 sonant ' s ' should have ended in some similar manner, as by 
 a comma point. In writing, this would degenerate into some- 
 thing like the numeral sign ' 3,' a form which is used in Rus- 
 sian for English z, constituting a very suitable letter. Never- 
 theless, the adaptability of a Z rounded into a reversed S 
 should be considered. 
 
 As English sh belongs to a different contact from s, it should 
 not be represented by a pointed f s,' nor French j by a point- 
 ed ' mf such a mode being as unphilosophical as to represent 
 % by a pointed sh, or th by a pointed /. The character r (but 
 not /"with its dot and curve below to be written with the stem 
 of script / and the tail of y, like the German script h) was 
 
18 ON THE PRESENT STATE OF OUR 
 
 proposed by Volney for sh, and has been used to some extent. 
 For French j, the Wallachian form is probably the best, being 
 somewhat like (j) inverted ' f,' with a curved line through the 
 stem, sloped in the direction of the acute accentual. Some 
 such characters are necessary, the paucity of aspirate conso- 
 nant characters in the Roman alphabet being admitted. 
 
 Rule IV. is probably based upon forms like <th' for the 
 Greek theta, — a false notation for this sound, because theta is 
 not an aspirate of t, but a member of a different contact, so 
 that to render theta lenis (0) it would be a kind of t formed 
 between the teeth; and truly to aspirate t ( e t) would be to 
 form a sound strictly at the t contact, with a quality between 
 the aspirates and s. Similarly, *s would be s deprived of as- 
 piration, forming a kind of t posterior to the normal t. 
 
 The use made by Dr. Lepsius of (0) for the surd consonant 
 of thigh is unexceptionable, but as he wants a character for 
 the sonant of thy* he wavers between $ and S, proposing at 
 the same time (? for ) thus establishing a complete confusion 
 in the use of the Greek spiritus asper and lenis marks. In & 
 an aspirate mark is added to that which is already aspirate, 
 that surd may be understood ; and in ff the lenis mark indi- 
 cates that vocality has been added, not to 0% but to an imagi- 
 nary 0. In h* the aspiration of the Arabic hha is enforced, 
 although the Ethiopic character (A) was present in an invert- 
 ed Greek ^-character, which would recall the European idea 
 of hh. The (%) with postposited (') is given for Arabic ghain 
 and Ellenic gamma, although (%') ought to mean k. In this 
 notation (f ) would mean /; (f) English v;-\ and (s') Eng- 
 lish z. 
 
 * That of Mr. Ellis formed on a (d) basis is probably the best, — or rather its 
 later form in the English phonetic journals. 
 
 t There should be a rule to the effect that, When a character is perverted from its 
 original power, its form should show the variation. Hence, if prejudice, or ignorance 
 of Latin speech, should prevent a missionary from using (V) with its Latin conso- 
 nant power, it should be marked to indicate its corrupt English power. Perhaps 
 (y) might answer, or a break towards the left, in the left branch, like that of italic 
 k. This would remove the confusion between r and v in writing. 
 
KNOWLEDGE OF LINGUISTIC ETHNOLOGY. 19 
 
 To accommodate Sanscrit writing, Professor Lepsius uses 
 (p) for p and h in uphold, although all that was necessary was 
 a statement of the fact that, in the Sanscrit alphabet, the se- 
 quents ph, bh, &c. (as well as a consonant and vowel in some 
 cases) are supplied by a single character ; a statement of this 
 kind being considered sufficient in regard to ps having a single 
 character in the Greek form of ellipsis. But whilst the spirit- 
 us asper and lenis marks are used in these heterogeneous and 
 unauthorized modes, the former is used as a separate charac- 
 ter for Arabic and Hebrew 'alef. This is at least a doubtful 
 view of the Greek spiritus lenis. As applied to a consonant, 
 we see it in apprjv, male (and apcryv, because the surd aspirates 
 rh and s* are allied), and Dr. Lepsius admits an (f) with a 
 Greek key-word, and he would probably write the Welsh rh 
 thus. As applied to a vowel, Chavee (Lexiologie Indo-euro- 
 peenne, p. 18), with great probability, explains the Greek spir- 
 itus lenis as the slight breath which precedes an initial vowel ; 
 for as the vibration of the vocal ligaments is due to the pass- 
 ing air, a little must necessarily pass before their quiescent 
 state can be changed. 
 
 Professor Lepsius uses the acute accentual (')f over (A;) for 
 tsh; over (%) to distinguish the German ch in ich from that 
 in ach; after (/), &c. to denote the mouille effect ; and in (7') 
 the Welsh surd aspirate /, for some unexplained reason. To 
 English sh and zh are assigned s and z surmounted by the ( w ) 
 mark for short quantity, — a perversion of a well-known mark, 
 required, with the mark of length (as in Dacota s and sh, 
 Riggs, pp. 184, 188), to distinguish long and short continuous 
 
 * But Dr. Latham, English Language, 1841, considers s the lenis of sh. He 
 omits A, wh, ng, rh, from his System of Consonants, p. 112, and he considers the 
 mutes specifically distinct to be no more than sixteen. To these add his " Semi- 
 vowels, to, y," and "Liquids, m, n, I, r," n being given as the liquid of t, d; and Z, 
 of h 9- 
 
 t As an accent mark, this should be thick above, and for the secondary accent 
 thicker below. If used for other purposes, it should be of equal thickness. Used 
 with ( " ) for sonant and surd, the middle part should be cut away to give the ap- 
 pearance of two dots in sloped directions. 
 
20 PRESENT STATE OF LINGUISTIC ETHNOLOGY. 
 
 consonants in some languages. S is thus prolonged in hissing. 
 " The people of Mallicollo use R in many words, two or three 
 being frequently joined together They express their ad- 
 miration by hissing like a goose." (Cook's Second Voyage.) 
 
 A dot is used over (n) by Lepsius, for English and German 
 ng, which is equivalent to representing n with a dotted m-char- 
 acter. Consistency should have required this dot to be ("), 
 because both indicate that the marked letter belongs to a pos- 
 terior contact. Ellis's character is much better, — an w-charao 
 ter with the second limb ending like (/). As a capital (and 
 capitals are of doubtful utility) that of Riggs is a good one, 
 being (N) with the diagonal shaped like a sloped (J.) As a 
 dot is the slightest of marks, it should be used (below the let- 
 ters) to indicate those slight evanescent consonants and vowels 
 which occur in some languages. 
 
 Rule IV., although it cannot be called unphilosophical, would 
 deprive us of a well-known and exceedingly definite mode of 
 notation with the aid of (' '), and obviate the necessity for 
 many new characters in the course of linguistic discovery. 
 By first assuming that I and r are " fricatives," Professor Lep- 
 sius admits (') with Z, r, but denies it to n, m, as " explosives," 
 without providing a means to indicate the same phenomenon in the 
 latter case. Sjogren (Ossetische Sprachlehre, St. Petersburg, 
 1844) has a good notation for the aspirates. He uses the 
 Russian alphabet for a basis, and instead of the ordinary A, 
 he curves the end towards the left below the line, in the shape 
 of (a), then uses this appendage as the indicator of aspiration 
 and spiration, by adding it to the stem of the Russian charac- 
 ters for p, t, g, and k, curving the last line of h backwards. 
 For English dz he uses J 3 , which is unnecessary, as the z 
 character should appear fully, if the sound exists. Lepsius 
 (p. 69) seems inclined to approve of this mode of writing t s , 
 &c. Such forms should be left for the concurrent emission of 
 consonants formed simultaneously (Rapp, I. 84). Thus a 
 child learning to speak formed I by applying the tongue to 
 the upper lip, and passing the voice over the lower lip, produ- 
 cing a sound having the quality of I and 7>. 
 
14 DAY USE 
 
 RBnmNTODBSKBROMWHXCHBORKOWBD 
 
 IOAN DEPT. 
 
 29Jul'65DP 
 
 — sec^-t^. 
 
 LD 2lA-60m-3 '6=5 
 (F2336sl0)476B 
 
 General Library 
 
 University of California 
 
 Berkeley 
 
U.C. BERKELEY LIBRARIES 
 ■lllllllllll 
 
 CD05513b70