On some Doubtful or Intermediate Articulations : 
 
 An Exi'P^RiMENT in Phonetics. 
 
 liy Horatio Hale, Esq. 
 
 In many liinp;ua<,'e,s, as is well known, thcro are elementary 
 sonnds of an indeterminate character, which seem to float 
 between two, and sometimes even three or fonr, diverse articnla- 
 tions. Tiie American and tiie Polynesian lan<i;uayes afi'ord many 
 instances of this S(nt, which have much perplexed those who 
 have attempted to reduce them to writing. A striking example 
 is found in the Hidatsa (or Minnetaree) speech, a language of the 
 Dakota stock, of which we iiave an excellent account by Dr. 
 Washington Matthews. In this language, he informs us, "there 
 are two series of interchangeable consonants — a labial series, 
 consisting of vi, h, and w, and a dental, or linguo-dental series, 
 consisting of d, /, ?i, and r." Dr. IMatthews regards the m as the 
 "stantlard letter "of the labial series, and the d as the " standard 
 letter " of the dental series, and the other letters in each series 
 as mere variants of these. That is, the word viia, mother, may be 
 frequently heard with the pronunciations wid and Inn ; and the 
 word dopa, two, may be heard as nopa, lopa, and ropa. 
 
 In the Samoan and Hawaiian languages of Polynesia, spoken 
 in the Navigator group and the Sandwich Islands, tlie linguo- 
 dental series has almost as wide a range of variation. The / and 
 the r are constantly interchanged, and frequently pass into the 
 d sound. Hilo, the name of a district in Hawaii, has in past 
 times, before the orthography was settled by the missionaries, 
 been spelt Hiro and Hido. Fa/e-a/ili, the name of a place in 
 the Navigator Islands, has in like manner l»een written Frde- 
 andi. In the (knienga (or Mohawk) language the sounds of 
 / and r, of g (liard) and k, of o and ii, are constantly interchanged. 
 The word for man (or, rather, " he is a man") may be indiffe- 
 rently written roni/ive, rungwc, lomjwc, Imujivc, rotikwc, riinkwc, 
 lonkwe, or lunkwc. 
 
 In the Hawaiian a remarkable interchange occurs between 
 the sounds of t and /.'. Teiti and kdki for child, tamita and 
 kanaka for man, are heard, and were formerly written indiffe- 
 rently. The element is really the Polynesian t, as is shown by 
 comparison with other langimges of that stock. The Hawaiian, 
 so far as is known, is the only language of this family in which 
 this singular interchange of t and k occurs. The missionaries, 
 it is said, were perplexed in attempting to determine whether 
 to use the / and r, or the /• and /, in the al])hal)et of this spee<di. 
 
 a 
 
234 IT. Halk. — On some Doubtful or hitcrmcdudc Articulationn: 
 
 They finally concluded to submit the question to the kin;,', who 
 decided in i'avour of the /• and the /. So t'iir as tlie / wiis 
 concerned, the clioiee was a matter of indifference; l)nt the use 
 of the /.• has had the rather unfortunate effect of somewliat 
 dis^'uisinjf, in the written lan^niage, the close siinihirity which 
 exists between the Hawaiian and the other idioms of Tolynesia. 
 The IJev. ^Vnl. Ellis, the distin^nnshed missionary writer, autlior 
 of " Polynesian Itesearehes," and other valuable works, visited 
 the Sandwich Islands in 1823. Throu^'hout his narrative the 
 well-known names which are now written Kaniehameha and 
 Liholiho, are si)elt Tamehameha and IJihoriho. In this ortho- 
 graphy they correspond with the ibrms in the Tahitia'J language, 
 with which Mr. Ellis was familiar. 
 
 To a student of languages, in considering these interchangeable 
 sounds, there are three hypotheses which may occur. It be- 
 comes a point of considerable importance, in pursuing an inquiry 
 in regard to the origin of the variation of languages belonging 
 to the same stock, to determine which of these hypotheses is the 
 correct one. 
 
 1. We nught sup]iose that every member of a people speaking 
 one of these languages uses these interchangeable sounds 
 indifferently — that a Hidatsa Indian, for example, in uttering the 
 word for mother, says at one time mia, at {mother win, and at 
 another hia, as the fancy may strike him, or the euphony of the 
 sentence may seem to re([uire ; and so a Hawaiian may say 
 tannta or kanaka, lUhoriho or Lihuliho, according to his momen- 
 tary caprice or some casual notion of euphony. 
 
 2. Another view might be that some speakers prefeiTcd one 
 sound in the series, and others preferred one or other of its 
 variants. One Hidatsa might usually say viia, while another 
 more commonly pronounced the word hia, and a third was 
 more accustomed to say wia ; just as in English one speaker 
 may pronounce the vowel in the word " aunt " with the broad 
 sound of a in far, while another may give it the slender sound 
 of a in fat ; or as one person may omit and another pronounce 
 the aspirate in " humble." 
 
 3. A third supposition wouhl be that the difference of sound 
 was not in the speaker's utterance, but in the ear of the listener ; 
 that the sound as spoken was an indistinct articulation, inter- 
 mediate between the sounds represented by the two or more 
 letters of each series, and tliat the hearer, unaccustomed to 
 sounds of this peculiar character, involuntarily made distinctions 
 where none really existed.* 
 
 Of the three theories thus suggested, the last would, at first 
 
 • The same subject is treated in Prof. Max Muller's Lectures on the " Science 
 of Language," vol. ii, pp. 183-189. 
 
Ati Kiprriincnt in P/ionetie.s. 235 
 
 thou«j;lit, set'iii the Iwist likely to bu the covrect one. Those who 
 liave studied the lunguiiges iii which these umertuiu sounds 
 occur hiive generally lulopted one or other — or sonietin^es holh — 
 of the two lornier suppositions. This, I must admit, was the 
 case with myself, after considerable experience in this line of 
 study. The third view, which su^jpo.ses the discrimination of the 
 sounds to be due, not to the speaker, but to the listener, had not 
 occurred to me until it was forced upon niy attention by the 
 unexpected result ol' the experiment now to be recorded. 
 
 In .July, 1872, I had the pleasure of sjtending a few days at 
 the hospitable home of my friend, Professor Alexander Melville 
 liell, the distinguished author of " Visible S})eech," and other 
 esteemed philological works. Mr. Bell then resided near the 
 city of Jirantford, Ontario, at a short distance from the (Irand 
 Kiver Reserve, which is occupied by the (Canadian remnant of 
 the Irocjuois Confederacy. On one occasion we were joined by 
 an intelligent Indian friend, Chief George Johnson, the Warden 
 of the lleserve and Government Interpreter for the six nations. 
 Chief Johnson was a well-educated man, a Mohawk chief of the 
 highest rank, and spoke fluently the dialects of all the Invjuois 
 tribes. The idea occurred to nie of taking advantage of this 
 opportunity to clear up, with the aid of the practised ear of 
 Professor Jiell, some dou])tful ])oints in Iroquoie ])honology. 1 
 proposed that we should take down a list of words in the 
 Canienga (or Mohawk) dialect — Mr. ]>ell in the nicely discrimi- 
 nating ali)habet of his "Visible Speech," and I in tlie method 
 which 1 usually ado])ted in writing these languages. This was 
 accordingly done, and the dujtlicate list, in Mr. Bell's manuscript 
 and my own, was left with me for study and comparison. 
 
 The result was unexi)ected, and, as it seemed to me, instruc- 
 tive and valuable. In the languages of the Iroquois group, no 
 distinction is made between the r and /. In the Canienga 
 dailect the pronunciation seems to incline more to the sound of 
 r, while in the softer Oneida speech the / souu'l appears to 
 predominate. All the missionaries. Catholic, Anglican, and 
 Methodist, though dill'ering widely in some points of ortiiography, 
 unite in using the r to represent this .sound in the Canienga 
 idiom. In the list of words which we wr()te down this element 
 occurred twenty-one times. Of these, I fountl on e.Kamination 
 that I had written it ten times with /, ten times with r, and on 
 one occasion had, in doubt, repeated the word with both ortho- 
 graphies. Mr. liell had used the I riineteen times and the r 
 only twice. In two cases in which he had employed the / sound 
 he had adopted the character which represents the " non-sonant 
 /," a delicate modification of that liquid which he discerns in the 
 pronunciation of the French word (fmp/e and in the English /^/^ 
 
23G II. Halk. — On some DonUful or Inttrmi'diatr Aiiiciddtiovs: 
 
 From this stntenient it is evident that in ei<»ht words where I 
 heard the sound /•, Mr. Hell at the same moment heanl the 
 sound of /, cither sonant or non-sonant. The eonclusioti 
 appears inevital>le that the sound which we lieard was really 
 neither /• nor /, hut an utterance n»i<hvay Initweeii the two, and 
 of such a character that to one listener it secuned an r, and to the 
 otlu'ran /. One of the words, as has heen stated, I wrote at the 
 time in two forms ronika and ///'/( ///»<, meaninj^ " his father." 
 .1 was unahle to decide which orthoj^'raphy most accurately 
 represented the jnonunciation 1 desired to preserve. This word 
 was written hy Professor Hell lu'ni/in. 
 
 It will lie noticed that in writing,' this word I was uncertain 
 both as to the first consonant and as to the lirst vowel. What 
 may he called the "round vowel" sound (o or n, ])ronounce(l as 
 in Italian) occurred in our list thirty-one times. 1 v/rote it 
 eighteen times with o, twelve times with //, and once— in the word 
 just cited — with l)oth (*and ?^ Mr. Bell, with j^reaterre^ailarity, 
 and ju'ohahly a nicer ear, employed the n throughout. The 
 Catholic missionaries, on the other hand, use only tlij o. The 
 I'rotestant versions have o for the most ])art, hut emjtloy the w 
 in a few winds. 
 
 Many years a<io, in takinu; down some of the lan<,'uages of 
 Eastern Australia from the lijw of the natives, 1 ascertained the 
 curious fact that their lan«,niaji;es nnide no distinction between e 
 and i, or between o and v.' They had, in fact, but three vowel 
 s((unds, which mij.';ht be rei»resented either by <r,(', and (>, or hy a, 
 i, and n, at the pleasure of the writer. The Iroipiois make 
 a clear distinction ))etween the e and the i. which are not 
 more frequently confounded in their dialects than in the Indo- 
 European idioms. Ihit between o and u in Irocpiois no distinction 
 exists, and from the evidence of the exjieriment new detailed 
 it is clear that the sound is not a varying one, inclining at 
 one time to o and at another to u, but a sound so exactly 
 midway ]>etween the two as to ])erplex an English ear, anil to 
 lead two hearers to write the same utterance with ditt'erent 
 characters. 
 
 The Canienga language makes no distinction l)etween the k 
 and the //, or Ixitween the t and the d. The English missionaries 
 use all these letters ; the Freiuih missionaries employ only the k 
 and t. The evidence of our list shows tiiat the latter are most 
 nearly accurate, as it is clear that in the native pronunciation 
 the. sound approaches more ch)3ely to the vowel than to the 
 sonant utterance. Mr. Bell has written the /; twenty-four times 
 and the ff only six times ; he has the t twenty-six times, and the 
 d four times. I wrote, in the same words, k throughout, and t 
 
 ' All the vowels ure to be sounded ivs in ItaUau or German. 
 
All. KrperiiiniU in Phonetics. 237 
 
 in every iiistiiMce l)ut oiki— the word for "hunci," which was 
 writUni Uy me onuinhi, and l>y Mr. \WA\ itnvmizi. The same 
 word with ^. prelix ('iny heiid ") was written by Mr. Jiell 
 ai/enitnfziiin, and hy nio in two lonns, akmuiitsine and 
 akoiKudzitii'. 
 
 The lro((uois hiii;.,'ua<,'e has a strong <^'uttural aspirate, which 
 the Kn;^dish niissionaries express in some words hy h, in others 
 by hh, and in many instances hy ////. The early Jesuit mission- 
 aries, as appears from Ihuyas's well-known work {liadires 
 Vn-boridu /nx/meondii), had also a threefold notation for this 
 sound, employing sometimes the h, sometimes the (Jreek x, t^nJ 
 sometimes the (Jreek ,'if)iri/its aqwr ('). The modern French 
 missionaries, after long and careful study of the language, 
 have decided that all these sounds are but variations, real or 
 ai'parent, of a single element, which they represent hy h. The 
 experiment now recorded shows not merely that this view is the 
 correct one, l)ut also that the variations are only apparent, and 
 depend rather on the ear of the listener than on any actual 
 difference of enunciation. In our list I have written the h 
 twenty times and the stronger aspirate (here represented by q) 
 six times. Mr. Hell in the same words heard only the h ; but 
 he in three instances emidovs what he terms in hia svstem the 
 "breath-glide" (which 1 transcril)e by the ,<piritiis asptr ') when 
 I have used the h or the q. This occurs only in conjunction with 
 the " non-sonant /," {or l), as in e/hal!, dog, which 1 have written 
 elhnl/i,ixnd in hi'hife, tree, which 1 wrote /iCql/iitc. The aspirate 
 and the litpiid in these cases are so combined that it is difficult 
 to say which is first uttered. 
 
 The first impulse of many persons on reading of these 
 indeterminate vowcds will dvmlttless be to account for them by 
 the fact that the languages in which they occur are in the un- 
 * cultivated or barbarous stage, lurther consideration, however, 
 will show that this view cannot be maintained. We know from 
 the evidence of the Vedas, the Homeric poems, and the 
 Moallakat, what was the state of the Sanscrit, the C J reek, and 
 the Arabic, at a time when the people who spoke these languages 
 were nnle.ttered barbarians. We are aware that the speakers of 
 those tongues discriminated sounds with an accuracy and a 
 variety which their more civilised descendants have failed to 
 preserve. Further, we discover that nuuiy narbarous communi- 
 ties of the present day ex])re3S delicate shades of pronunciation, 
 which we can only with difficulty imitate. In the proper 
 Dakota (or Sioux) language, for example, we learn from the 
 excellent gi-ammar of the Kev. S. K. lliggs, that not only are the 
 surds k, p, and t, distinguished from the sonants (/, b, and d, as 
 in English, but tliat there is another distinction which does not 
 
238 H. IIalk. — On nome Donhf/iil or f/ilrrmallttfiArflinInliunn: 
 
 exist in oiir liiiij,Miii;;('. Kacli of tlio surds, as well as tlio 
 connHisite souml wliit-li roin'spoiuls to the Kii^lish ch in "chin," 
 has a ati'on^f exphtsivc or cinphiitii; utterance, \vhi(th makes of it 
 a distinct element. This dilference of sound is indicated in Mr. 
 Kig^'s's alphabet l»y a dot under the emithatic hotter. Imlelauli 
 of these characters we may emidoy " small ca])itals." The fol- 
 lowing list will show how accurately tlu'se barharous 8i)eakers 
 discriminated in their phoiutlo^'y : — 
 
 Hotiunt. 
 
 Surd. 
 
 EmpliBtic. 
 
 1)0, to hitfrh 
 du, to ask 
 
 1)0, nhiirp 
 
 KII, to CoiHf 
 
 111, moone 
 clitt, when 
 
 PC, flm 
 Kii, to ifive 
 Til, /() die 
 ciitt, to dig 
 
 No instance of the sonant <f is ^'iven, as that in the Dakota is 
 merely a dialectical variation. The lan<;ua<j;es of the Maya, or 
 Central American, family aliound in nice distinctions of elementary 
 sounds, which f(»rei<;ners have a dillicidty in acquirinj^. The 
 ]Iawaiian, which confounds the / with the r and the t with the 
 k, has preserved a jieculiarity so ajjparently sli«j;ht that the mis- 
 sionaries unfortunately have not deemed it worth indicating; 
 and yet it is, in fact, of the first im])ortance, both in science and 
 in practical use. It is a hiatus, or catching of the breath, which 
 shows where an element formerly in use lias disa]»i)eared from 
 the language. This element is the rolynesian /■, which is still 
 retained in the dialects of New Zealand, the Friendly Islantls 
 (Tonga), and some other groups, but has disajijieared from those 
 of Samoa, Tahiti, and Haw aii. Thus the original Polynesian ika, 
 fish, becomes in Hawaiian i'a ; aliki or <(riki, chief, becomes <il%i,' 
 kai, to eat, becomes '«?', and soon. The first missionaries to these 
 islands were, intelligent and well-educated men; but, accustomed 
 only to the Englisli pronunciation, they lailcd to notice this 
 delicate trace of utterance, or did not think it worth indicating. 
 They themselves never ac(|uired it, though their children, born 
 and reared in the islands, use it habitually, like their aboriginal 
 companions. In the native ]tronunciation, the words ao, daylight, 
 ao (for the Polynesian r?/.(>), to teach, and '(to (for the Polynesian 
 kao), to sprout, are ])laiuly distinguished ; but in the ordinary 
 orthogTa])hy of the language, all these v;ords are confounded in 
 one spelling ao. It is precisely as though the English language 
 were to be written for the first time by ])er.sons who could not 
 distinguish the aspirate. No difference would be made in their 
 orthography between heat and cnf, hair and ni)' ; and in reading 
 
All Kr/friinent in Phoml.icd, 239 
 
 of tlio " ciljro " of !i fu'ld, \v(« .should ii(»t kiiow, from the Hpclling, 
 whether th»^ writer referred to its hctrder or its hedjre. 
 
 The same iH-culinrity is found in the InM^iois diiduct.s, and 
 has been e(|Uidly lu'^^h-eted Ity the ndssionaries, exeej)! in oiio 
 instance. The Ilev. Asher " VVrij;ht, the late accomjilished 
 
 missionary an \i the Seneeas, who liad a turn for ithilolofjy, 
 
 and esjx'cially for distin;^'uishin,i,'souiuls, has employed a peculTar 
 charaeter, a modification of the h (which we may represent liy ii) 
 to indicate this hiatus, lie remarks of it: — "This letter never 
 j)recedes a vowel. Kollowin<,'one, it should bespoken by f;ivin" 
 the vowel an explosive forc(^ and breakinj,' it olf .suddenly, in 
 such a manner as for tJu! iii.stant to stoj) the breuth entirely, as 
 we often hear while |teoj»le in hastily pronouncinjj; the interjection 
 Oh : — especially when they repeat it several times in rapid 
 succession, in indicatinj^' to a child that it is doinj^ something 
 wront,'. This .sound is very abundant in Seneca, and, used in con- 
 junction with ctatain other modifications, the mode and time of 
 verbs, and various other circumstances, are denoted by it. Often, 
 also, it forms the chief distinction between words of very dis- 
 similar meaning,'. No one can read or write Seneca intelligibly 
 who does not pay the strictest attention to this character and 
 avoid confounding it with the rough as])irate of the common A." 
 
 Thus, among the examples, we find tiuit hawnih, "my lather" 
 (speaking of him), has for its vocative form linwniw, " my 
 father" (.speaking to him). Wa-n^ii, "he said," (litters only by 
 the absence of this element in the first syllable from vMii-a^h, 
 " she thought." In the Tuscarora an example given to me by an 
 intelligent school-teacher of that nation was o/i.sdukira, " finger," 
 which dilfers oidy in this element from ohsuhkica, " lip." In the 
 Canienga or Mohawk dialec-t, this hiatus was noted by both Mr. 
 Bell and myself, though, as was mitural in writing a strange 
 language, we did not always remark it, and in some instances 
 it was noticed by one ami omitted by the other. 1 have usually 
 represented it by an apostrophe, as in ihsha'a," child," I iriiih a, 
 " his father." Whether the hiatus indicates in the Iroquois, as 
 in the Hawaiian, the loss of an element, or is a mere trick of 
 utterance, is a question not yet determined. 
 
 Many languages which have been reduced to writing of late 
 years, in America, Oceania, and Africa, have undoubtedly 
 suffered a serious impoverishment in their phonology from the 
 fact that the persons by whom they were tirst written were 
 foreigners accustomed only to the European mode of utterance. 
 If the Sanscrit had been tir.st written by an Englishman, a 
 Frenchman, or a German, it is very doubtful if the distinction 
 between the lingual and dental elements would have been 
 preserved. The Arabic, under the like circumstances, would 
 
240 11. H.VI.K. — (hi Homr Diiiiht/ul ,ir I iitnnutHoti Arliiuhttlona : 
 
 proliiihly Imvc siilU'rcd a serious (l('t('riuniti(»ii in its (hnitiils nnd 
 its ^'tittunils. 'i'lic t'liilmmtfrirciiinsliinct! tliat Mr. Ili^';,'s was an 
 nc(!()iii|»lislie(l |tliiiiil<);j;ist lias lucsi'rvfd in tlic hakota lan!j;ua;,'»j 
 ilistinrtiniis that Wfaild pmlialtly olln'rwisc havi' Ikm'ii lost, liut I'or 
 this "luckv at'cidcnt," the readers ui that hmLjua^e would liavo 
 had no evidence of the dillereiice (tf pionniieiation whie.h exists 
 between ha, to conic, and K/', to jj;ive. l>ii, to ask, tn, a nioost;, 
 (iiiil '1V^ to (he, nii<^ht havu hecu cont'ounih'd in the hakota.as 
 ao, daylijU'ht, '"" tti teach, and n'o, ti'spriait.are in»w conloiuided in 
 the Hawaiian. Xo one, prohalily, lail a srhohir laniiliar with tho 
 Seiniti<! ton^'ues wonhl have ihstjunjnislu'd and represenfed in tho 
 Dakota, the form of <; which expres.ses "a ileep sonant ^Mittural 
 rcseinldin^j; the Arabic ghain (i) and the form (»f ii which 
 
 represents a strong; sunl },Mitlural resemblin<,' the Arabic kha (^). " 
 
 There are still many unwritten lan^'naj,'es in Westein Oceania 
 niid in Central Africa, for which aipl:.ibets will have to bo 
 provided. It will be fortunati! if the persons to whom this 
 imp(»rtant duty iw entrusted shall Ik; .scholars trained in tho 
 scientitio study of lan;4ua;4e. If this advantaj^o cannot bo 
 secured, care may at least lie taken that the work of settlinj^ 
 the alphabet shall not in any ca.se be entrusted to om; person, 
 however intellijj;ent nntl W(dl-iustructed. Tla^ result of tho 
 experiment now recorded will show how essential it is that, 
 to determine the real distincti(»ns in the ehunentary sounds of a 
 lun<i[ua<;e, its words should be taken (U)wn by two or more 
 persons, listeninj^ and writing' simidtaneously. In this way 
 alone it will be possil)le to avoid, on the ont^ hand, the dan<^er of 
 coidbundin;^ sounds which should be kept distinct, and, on Mm 
 other, that of tlndin<,' distinctions where none; really exist. 
 
 As rej,'ards the special results of this exi)eriment, it will not 
 be safe to infer that in all cases where such uncertainty jirevails 
 as is found to exist in tho Iro<|uois and the I'olynesian dialects 
 in discriminating between sounds which to us seem widely 
 ditterent, this uncertainty is due to a lack of clear ]»erception 
 in the listener. It is hi^ddy probable that in some cases tho 
 proimnciation of the dilferent natives, or of the same native at 
 different times, actually varies. Hut the experience now set 
 forth will at least serve to ])rove that this is not always, or 
 ])erhaps usually, the case. We must reco<^nise the fact, which 
 has heretofore been overlooked, that in certain languaj^e.s — and 
 possibly in all languages — there are elementary sounds which 
 affect so diflferently the ears of two listeners accustomed to a 
 foreign speech, that in noting them they will be likely to use 
 different characters to represent the same utterance. In other 
 words, there are elements, in some if not in all languages, which 
 
Jti K.ijMfiiiicnl in J'/wnetivs. 
 
 hold a mitlille plnco between two correa])ondinj? elements 
 of some other liuiKuiiK'e. And the sounds which these medial 
 elements thus represent may even be as widely diverse as, 
 according,' to our notions, the r is from the /, and the t tVoni 
 the k. 
 
 In our own lan^'uaf,'e a 8inj;;ular instance of this peculiarity 
 has been noted and well described in an article on Shakespeare 
 by the distin«,'uished philologist and Shakes|)earean connuentator, 
 Mr. Kichard Grant White, which appeared in the Atlantic 
 MoiUhlij for June, 1884. Mr. White brings to notice the 
 curious fact that in Shakespeare's time, and earlier, there were 
 many English words in which the dental sound was written 
 indiscriminately either with t and th, or with d and dh. 
 He gives numerous examples, such as stidwart and sttdivorth, 
 Jifth and Jift, hctfcr and ict/ur, hundred, hnndret, and Imndreth, 
 sivart/i7j and sioart)/, invrdcr ond murt/ier,hin'dtn and hurthm ; 
 and he continues — speaking of his own edition of Shakespeare's 
 works : — " It will be observed iliat in tlie lii ersid j Shakespeare, 
 murder has both its modern spelling and tiie form murt/wr. 
 The variation is that of tlie old copies, which was purposely 
 retained. The pronunciation was not viurlher, with the theta 
 sound, which is j)oorly indicated by th, nor exactly that of d, 
 but just that, I am sure, which has survived in the north of 
 Ireland (carried there by English invaders, and chiclly by 
 Cromwell's trooju-rs), and whicli we have all heard, viiirdher." 
 This acute suggestion of Mr. White gives us a case which is 
 exactly in point. Every one who has heard the Irish pro- 
 nunciation of the word in question is aware that it is sucli as 
 one hearer might represent by viurder, while another would 
 prefer to spell it mvrther. In other words, it is an intermediate 
 sound between the d and the th, exactly analogous to the inter- 
 mediate elements which are found in the Iroquois and other 
 languages. 
 
 The manner in wliich diflferences of dialect, and finally of 
 language, might glow out of this peculiarity of pronunciation is 
 evident. If an emigrating liorde, whose speech possessed these 
 indeterminate elements, were to conquer and absorb a tribe 
 accustomed to a more precise mode of utterance, these indistinct 
 articulations would, in the mingled race, tend to assume a fixed 
 and positive character in one direction. A like result, though 
 in a ditt'erent direction of change, might happen in a second or 
 a third migration, encountering and overcoming other tribes. 
 Thus theHidatsa people might send out three conquering colonies, 
 in one of which the word for " mother " might come to be always 
 pronounced hia, in the second mia, and in the third wia. If 
 the original Aryan speech possessed intermediate articulations 
 
 h 
 
242 H. Hale. — On sunie DoaUfnl or Inkrmediatc Artu'idatluns: 
 
 of the kind now described, it is easy to understand how in the 
 progress of the conciuering migrations of Aryan hordes wliich 
 absorbed the original populations of Northern India and of 
 Europe, the varieties of pronunciation signalised in " CJrimni's 
 law," as well as many other changes, consonant)^ji and vocalic, a/ 
 would gradually arise. These changes, with the accompanying 
 alterations in grannnar, and the inevitable acquisitions of new 
 words from the idioms of the conquered tribes, would linally 
 produce the various Indo-European languages. 
 
 The following is the list of words taken down by Mr. liell 
 and myself. It must be borne in mind that the words were 
 written hastily, with only one hearing, and with no opportunity 
 of revision. Under such circumstances some mistakes are 
 inevitable, and must be allowed for. In the orthography here 
 adopted, the consonants have in general their English sounds, 
 and the vowels their Italian or Crerman sounds. The (; is 
 sounded like the sk in " shine," and t^; represents the sound of 
 cA in "chest." The q represents the German vh ((Ireek 'x). 
 The d is the Urvocal, or tlie short English u in " but." The 
 Erencli nasal n is indicated by a small " above the line, and the 
 English nasal {ny in " song") by h. The apostrophe ( ' ) marks 
 the hiatus, or sudden catcliing of the breath, already referreil to. 
 
 I have added the corresponding words in the forms severally 
 adopted by the Anglican and the Iioman Catholic missionaries, 
 the former written for me by an educated ]\Ioliawk, and the 
 latter derived from the Iroquois Lexicon of the Ilev. J. A. Cuoq. 
 For the purpose of comparison all the words are transliterated 
 from the differing missionary orthographies into the alphabet 
 employed in my own list. A study of those various forms, 
 taken in conjunction with the facts of our experiment, will 
 probably be found sufficient to establish the existence of several 
 of those intermediate articulations whose part and influence in 
 the phonology of language have been generally overlooked. 
 
 Canienga (or Mohawk) Vocabulahy in Fouk Hendehings. 
 
 
 Bell. 
 
 Hale. 
 
 Englisli 
 Mission. 
 
 R.C. Mission. 
 
 
 
 
 
 one . . 
 
 sl'>skHh 
 
 u"ska . . 
 
 u"skalit, or 
 a"ska 
 
 tl"8kat, a°sha. 
 
 two .. 
 
 tekinih 
 
 tekenih 
 
 tekeni 
 
 tekoni. 
 
 three 
 
 aqsa" . . 
 
 ahsa" . . 
 
 ahs^" . . 
 
 aksa". 
 
 four , . 
 
 kfivelih 
 
 kajelili 
 
 kayerili 
 
 kaieri. 
 
 five . . 
 
 wisk . . 
 
 wisk . . 
 
 wisk . . 
 
 wisk. 
 
 six . . 
 
 yava'k* 
 
 yavak. . 
 
 yaynk . . 
 
 iaiak. 
 
 seven 
 
 tjatah 
 
 t^atii' . . 
 
 jadahk 
 
 tsiatak. 
 
 eight 
 
 satekii" 
 
 sa'teku" 
 
 9adeko"li 
 
 sateko". 
 
An Experiment in Phonvl.us. 243 
 
 Canienga (or Mohawk) Vocadulauy in four Renderings — continued. 
 
 
 Bell. 
 
 Halo. 
 
 English 
 Mission. 
 
 B.C. Mission. 
 
 
 
 
 
 nine . . 
 
 tyuhti" 
 
 tiohto" 
 
 tyohdo"h 
 
 tiohto". 
 
 ten . . 
 
 uyeli 
 
 oyeli . . 
 
 oyerie. . , . 
 
 oieri. 
 
 eleven 
 
 a"8kaeaw)V'li . . 
 
 a"ska yawjloli 
 
 a"8kaht 
 yawjl"re 
 
 il"skat iawk"ro. 
 
 twelve 
 
 tegeniawfk"li . . 
 
 tekeni ya\vil"li 
 
 tekeni yaw^"re 
 
 tekeni iaw&"ra. 
 
 twenty . . 
 
 towilishk" . . 
 
 tewficisi" 
 
 tewahsh^" 
 
 tewastl". 
 
 thirty 
 
 aliBiluewiihs^" 
 
 aqsilniwaqsil'' 
 
 alisi," 
 
 niwahsha" 
 
 ahsa" niwasi". 
 
 forty 
 
 kayeliniwalisik" 
 
 kayeliniwaqsi" 
 
 kayerih 
 niwahsh^" 
 
 kaieri niwasa". 
 
 fifty . . 
 
 wiskniwahsi," 
 
 wisk niwuqs^" 
 
 wiak 
 
 niwahshil» 
 
 wi8k niwasa". 
 
 one hundred 
 
 ^"skahtewil"- 
 
 a"ska tewi"- 
 
 a"8kaht dew&,°- 
 
 e"skat tewil"- 
 
 
 'niawi 
 
 'nifiwe 
 
 niawe 
 
 niawe. 
 
 two liundred 
 
 tekinidewil"- 
 
 tekeni tew^"- 
 
 tekeni dewi"- 
 
 tekeni tewil"- 
 
 
 'uifiwi 
 
 'niawe 
 
 iiiawo 
 
 niawe. 
 
 three hundred 
 
 ahstV'dew^"- 
 
 ahsi" tewil"- 
 
 alisi" dewa"- 
 
 ahsi\" tewS,- 
 
 
 'nifiwi 
 
 'niiiwe 
 
 niawe 
 
 niawe. 
 
 one thousand 
 
 uy elite wa"- 
 
 oyeli tewi"- 
 
 oyeri dewa"- 
 
 oieri tewii''- 
 
 
 'niawi 
 
 'niawe 
 
 niawi 
 
 niawe. 
 
 my father , . 
 
 rake'nlha 
 
 rake'niha 
 
 rakcniha 
 
 rakenihn. 
 
 thy father . . 
 
 ya'niha 
 
 ya'niha 
 
 yaniha 
 
 hianiha. 
 
 hi8 father . . 
 
 lu'niiia 
 
 ro'niha, lu'nTlia 
 
 roniha 
 
 roniha. 
 
 my mother . . 
 
 IstiV'a 
 
 i.sta'''a 
 
 isd^"ah 
 
 isti'-ha. 
 
 tliy mother . . 
 
 8a'ni8ti"ha .. 
 
 8iini8i^"ha . . 
 
 sanisdiV'ha . . 
 
 8anisti\"ha. 
 
 his mother . . 
 
 lu'nisti'Mia .. 
 
 ro'ni8ti"ha . . 
 
 roni8t(\"ha . . 
 
 ronisti"ha. 
 
 my head 
 
 a){entt"tilnt\ . . 
 
 akenu"tslne or 
 akeno"dzino 
 
 akeno"djih or 
 akeno"djineh 
 
 akeno"tsi. 
 
 thy head 
 
 zanuntslnii . . 
 
 srinu"t8lne . . 
 
 8ano"djih 
 
 fcano"t8i. 
 
 his head 
 
 laiintlntsTn^ . . 
 
 raonu°tsine . . 
 
 raono"djih . . 
 
 raono"t8i. 
 
 my liair 
 
 agenu"kwi8 .. 
 
 akenu"kwi8 . . 
 
 akeno"kwi8 . . 
 
 akeno"kwi8. 
 
 hair . . 
 
 iinu"kwi8 
 
 onu"kwi9 
 
 ono"kwi8 
 
 ono"kwi8. 
 
 head. . 
 
 flnundzih 
 
 omV'dzi 
 
 onoMjih 
 
 ono^tsi. 
 
 eye . . 
 
 tikaliV 
 
 okara . . 
 
 okara . . 
 
 okahra. 
 
 nose . . 
 
 u'nyu"'8a 
 
 o'nIu"'8a 
 
 onyo"3a 
 
 onio"9a. 
 
 teeth 
 
 unfiwi' 
 
 onawi 
 
 onawi.. 
 
 onawira. 
 
 ear . . 
 
 uhu"ta 
 
 ohu"ta 
 
 oho"da . . 
 
 oho°ta. 
 
 liand 
 
 U8nii"sa 
 
 08nu"8a 
 
 08no''sa 
 
 osno^sa. 
 
 tree . . . . 
 
 keL'hite 
 
 keqlhite 
 
 kerhide 
 
 kerhite. 
 
 dog .. 
 
 elhaL* 
 
 elhalh 
 
 erhar . . 
 
 erhar. 
 
 house 
 
 kanu"sa 
 
 kaniV'sa 
 
 kano°8a . . 
 
 kano^sa. 
 
 town 
 
 kanuta 
 
 kanuta 
 
 kanada 
 
 kanata. 
 
 large town . . 
 
 kanatuwuni\" 
 
 kanatowiink".. 
 
 kanadowantV" 
 
 kauatowan^". 
 
 man . . 
 
 lu"gwe 
 
 runkwe 
 
 ro"gwe 
 
 ro°kwe. 
 
 woman 
 
 ikslnVa 
 
 iksluVa 
 
 exhaah 
 
 eksaa. 
 
 husband 
 
 lune . . 
 
 lone . . 
 
 vone .. 
 
 rone. 
 
 wife . . 
 
 tiagenitelu" . . 
 
 tiakenltelu" .. 
 
 tpyagenidero" 
 
 teiakenitero". 
 
 white 
 
 k(\"lrik^" 
 
 ktt"raka" 
 
 ki\"i*aki'' 
 
 kJlrak^". 
 
 black 
 
 kahu"dzi 
 
 kahu'Hfi 
 
 kaho"jih 
 
 kaho"t8i. 
 
 yellow 
 red . . 
 
 utsTnakwaL* . . 
 
 otsTnekwahr . . 
 
 odjinekwar . . 
 
 otsinekwar. 
 
 unekw!l"kark 
 
 onekw&"tara . . 
 
 onekwk"dara.. 
 
 onekwi"tara. 
 
 green 
 
 uhu"'te 
 
 ohunte 
 
 oho"de 
 
 oho"te. 
 
ISeprinted from the Journal of the Anthropological Institute, February, 1885.] 
 
 Harrison and Sons, Printers in Ordinary to Her Majesty, St. Martin's Lane.