SMITHSONIAN INSTITUTIONBureau of American EthnologyBuUetin 173
Anthropological Papers, No. 60A CAROLINE ISLANDS SCRIPT
By SAUL H. RIESENBERG and SHIGERU KANESHIRO
269
471762—60 19

CONTENTS PAGEIntroduction 273Type 1 script 273Type 2 script 277Present-day knowledge and distribution of the writing 279Number and sequence of characters 282Origin of the writing 282Distinctness of the two types 282Similarity of Type 2 script to the alphabet 283Syllabic values 283Type 2 history reconstructed 284Possible sources 284Linguistic affiliations 287The alphabet of Truk 287Alfred Snelling and the Trukese castaways 288Adoption of the Trukese characters 290Spread of Type 2 writing 293Type 1 history 294Invention at Faraulep 295Foreign influence 296Date of invention 297Derivation of characters 297Analysis of the writing 299Phonemes, phonemic combinations, and characters 299Length of vowel not distinguished. 300Vowel characters 302Lack of exact correspondence of characters and syllables 303Effect of dialects 304Representation of final consonants 305Comparison of native texts 307Summary 309Literature cited 311Tables 313ILLUSTRATIONSPLATES FOLLOWINGPAGE42. Tattooing in native script on arm of Maralatuy, a woman of FaraulepIsland, Faraulep atoll. (Photographed by E. Quackenbush.) 33443. Tattooing in native scrip on leg of Letaweribul, a woman of FalalapIsland, Woleai atoll. (Photographed by E. Quackenbush.) 33444. Native script and Japanese katakana on canoe-house beams, FaraulepIsland, Faraulep atoll, (a, b, Photographed by S. Kaneshiro;
c, photographed by E. Quackenbush.) 334
271
272 BUREAU OF AMERICAN ETHNOLOGY [BuU.173TEXT FIGURES PAGE25. Type 1 symbols and values (facing) 27426. Type 2 symbols and values (facing) 27827. Paragraph from Smith's roman text (1951, pp. 3-4) as written in nativescript by R. of Ifaluk 30828. A page of text from a song written by R., a man of Ifaluk 310MAP
1. Area of the occurrence of the script. _. 281
A CAROLINE ISLANDS SCRIPT
By Saul H. Riesenberg and Shigeru Kaneshiro *
INTRODUCTIONThe existence of a peculiar system of writing in the Woleai Islands ^has received almost no attention from ethnologists or paleographers,and none at all from linguists. It is a script which, the present writersare convinced, has developed in consequence of stimulus diffusion, touse Kroeber's term, and as such is to be reckoned among the smallnumber of scripts so originating, such as the Bamun ideographicscript and the Vai and Cherokee syllabaries. It is the purpose of thispaper to explain its origin and development and to analyze its form,content, and use. TYPE 1 SCRIPTThe first published notice of a Caroline writing was made by J.Macmillan Brown (1914, pp. 89-91; 1927, pp. 117-120). In 1913 hepaid a brief visit to Woleai atoll, where a chief, Egilimar by name,wrote for him a sentence in the native script; later he received a list of51 characters and their phonetic values. Brown did not inquireinto or was not told anything of the origin of the writing. The listis reproduced here in column B of figure 25. It is the type of writingwhich we will refer to hereafter as Type 1 . As Brown points out, thesjnubols do not resemble those of any system known from elsewhere.The script is now known only to five men on the islet and to some in Faraulep,an islet a hundred miles distant. But it is probably a relic of a wide usage in thearchipelago. There is no possibility of any one of the five having inventedit. . . . This Oleai script is manifestly the product of long ages for the use of the
' The authors are indebted to Drs. A. L. Kroeber and William Sturtevant for advice and criticism, to Dr.y. Uyehara for assistance with Japanese characters, to Dr. Ward Qoodenough for advice on Trukese lin-guistics, to Dr. S. H. Elbert for critical reading of the text and for assistance with tape recordings, and toMr. Frank Mahony who obtained considerable information incorporated in the text from Truk and sur-rounding islands. Several students from Micronesia at the University of Hawaii were also of much help,among them Tosiwo Nakayama of Truk, Bethwel Henry and Bailey Olter from Ponape, Nicholas Leon yGuerrero from Saipan, Edmund Qilmar from Yap, and David Ramanii from Palau.
' The term "Woleai" as generally used has two meanings: Woleai atoll itself; and the whole Woleai groupin the Central Carolines, consisting of Woleai atoll, Eauripik, Ifaluk, Faraulep, Qaferut, Olimarao, Elato,Lamotrek, West Fayu, Satawal, and Pikelot. Of these islands, Gaferut, Olimarao, West Fayu, and Pikelotare uninhabited. The group is often referred to also as simply "the Woleai." Just to the east of the grouplie Puluwat, Pulusuk, and Pulap, known today at Truk as "the Western islands;" we will be concerned inthis paper with Puluwat. 273
Explanatory notes for figure 25ColumnA. Numbers assigned to characters and used in text. Numbering after Brown'ssequence, 1 to 51, and arbitrarily thereafter.B. Brown (1927, p. 118). Brown's sequence, cliaracters, and attributed values.Woleai, 1913.C. Damm (1938, fig. 279). Sequence and characters from wooden board.Faraulep, 1909.D. Damm (1938, fig. 280). Characters from beams of men's house. Faraulep,1909.E. Kramer (1937, pi. 15 and fig. 109). Characters from canoe and bamboobox. Woleai, 1909.F. Damm and Sarfert (1935, figs. 128, 272). Characters from tinderbox andflute. Satawal and Puluwat, 1909.G. Someki (1936, fig. 5, p. 178; 1945, figs. 189, 230). Characters and attributedvalues. Presumably Faraulep, Ifaluk, and Elato, 1934.H. M. of Pigue, Faraulep, 1954. Sequence, characters, and attributed valuesin Smith's orthography.II
- C. of Pigue, Faraulep: 1955A, 1955B, and 1956. Sequence, characters,„ andattributedvalues; attributed values of column J in Smith's orthography.
L. L. of Pigue, Faraulep, 1957. Characters and attributed values.M. N. of Pigue, Faraulep, 1957. Characters and attributed values.N.l R. of Ifaluk; 1955 and 1957A. Sequence, characters, and attributed valuesO.J in Smith's orthography,P.IT. of Lamotrek; 1955 and 1957. Sequence, characters, and attributedQ.J values in Smith's orthography.R. B. of Falalap, Woleai, 1957. Sequence, characters, and attributed values.S. A. of Falalap, Woleai, 1957. Sequence, characters, and attributed values.T. S. of Ifaluk, 1955. Characters. REMARKS
1. Positions in sequence of column A assigned to characters in columns D, E, F,G, L, M, and T by comparison of form of character, and of value if given,with form and value of characters of the other columns.2. Bracketed characters are those not given in informants' lists but occurring inword samples; or are variant forms from word samples.3. Values in columns B and G are as given by Brown and Someki respectively.Attributed values in columns H, J, N, O, P, and Q are based on hearing oftape recordings and are in Smith's orthography, except those in parentheses,which were heard under hurried field conditions. Attributed values incolumns I, R, and S, in parentheses, not taped; are given as roughly recordedin field. Attributed values in columns K, L, and M, in parentheses, nottaped; values were written by these informants in Japanese katakana, and arehere shown in the orthography used by Prof. Y. Uyehara of the Universityof Hawaii, who has transcribed the katakana characters for us.*4. Underlining of numbers in various columns indicates the place up to whichinformants' sequences match the sequence of Brown (column B).
•The katakana used by Woleai natives Is a quite unorthodox one, and some of the combinations used areapparently Intended to form sounds Impossible In Japanese. Thus, the device known as nigori or chon-chon, used In Japanese writing to transform a character representing a surd consonant into a sonant con-sonant, is used by C. and N. (columns K and M) for various other purposes: e. g., it is used by both C. andN. as part of values of characters 21, 22, and 34, and seems to be intended in these instances to convert mto mw or r to rw. in other cases it is apparently meant to achieve vowel values intermediate between twoJapanese ones (e. g., characters 9, 10, 31, 67). Various combinations of katakana characters are transliteratedby us with hyphenated values, and seem also to be intended to form vowels absent in Japanese.
KO.
i.Al :f' 1 r '"."'v
1»

M
1

No*6o7'^''^' CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 275. 60]
Iorganisers of a highly-organised community of considerable size. In other wordsit must have belonged to the ruling class of an empire of some extent, that neededconstant record of the facts of intercourse and organisation.'There are some scattered references to Brown's discovery, but fewwriters have commented on it even briefly. Mason (1920, p. 152)accepts Brown's opinion, quoted above, uncritically. Diringer (1948,p. 448), agreeing in part with Brown, states thatthe origin of the Woleai script is perhaps in some way connected with the FurtherIndian branch of scripts, although this connection does not appear evident, eitherfrom the graphic or from the phonetic points of view. There is, however, thepossibility of the mixed process of invention and borrowing, called "idea diffusion."And Imbelloni (1951, p. 164 and fig. 25), in an attempt to link thescript to Easter Island writing and other scripts of his "Indo-Pacific"graphic system, refers to it as being based on Semitic syllabaries andhaving diffused to the Carolines via India and Malaya; he too regardsit as the remains of a formerly more developed system. A moresober judgment is that of Metraux (1957, p. 199), who says that thescript may very well have been "invented on the spot under the in-fluence of Malay, Indian, or even European writing."These theories derive entirely from Brown's report. Earlier thanBrown, in 1909, the Hamburg Siidsee Expedition had visited Woleaiand neighboring atolls, but the reports of ethnographic work on theseislands were not published until 1935 and later. These publicationsdo not discuss or even refer to the Type 1 writing, but they containillustrations of objects which bear the same characters. One of theauthors, Damm (1938, fig. 279), copied from a wooden board whichhe found at Faraulep a series of 29 symbols.* This series is reproducedhere as column C of figure 25. As will be seen by comparing theseFaraulep symbols with those from Woleai in column B, there is virtualidentity in graphic form and in sequence, as far as they go. Brown'scharacters Nos. 24, 25, and 28 do not appear, and his character 31does not seem very like the one in corresponding position in theFaraulep series, but otherwise they are alike. Apparently the boardrepresents the effort of someone to set down in proper order an alreadydefined set of symbols, perhaps for instructional purposes.There is also a line of characters copied by Damm from a beam in amen's house at Faraulep (1938, fig. 280). These are not a series, but
' Brown, 1927, p. 119. This is perhaps not so fantastic an idea as might at first appear, for the Yap
"empire," a religio-political hegemony which once may have stretched beyond Truk to the east, still includesin its domata all of the Woleais. But the Yapese themselves, who control the "empire," do not possess anynative script.
* This is the only instance of writing In which the symbols run from right to left, as reproduced by Dsmm.But it Is very likely that the board from which the figure was taken was held upside down by the copier.It is impossible to tell from the characters themselves if this was so, since, as will be seen from examination offigures 25 and 26, orientation of the characters is of no significance; but one Informant who was shown a copyof Damm's figure inverted the paper in order to read it.
276 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
constitute a sentence of actual writing.* Occurring in these lines are11 of the above 29 characters, 8 more characters which are not amongthe 29 but appear on Brown's list, including the missing No. 28, andtwo others which are only on lists furnished us by our present-dayinformants (fig. 25, col. D). In addition, the Siidsee Expeditionvolumes contain illustrations of a bamboo tinderbox and two canoesfrom Woleai, a Puluwat nose flute, and a tinderbox labeled "Satowalbzw. Polowat," all of them incised or painted with characters of thesame type (fig. 25, cols. E and F).It is evident, therefore, that the characters found by Brown in 1913at Woleai were known there and elsewhere in the Central Carolinesin 1909; and further, that Brown did not collect the complete set.^In the Japanese literature on Micronesia available to us we havefound only one reference to the writing, in spite of the long period ofJapanese occupation, 1914 to 1945. This is by Someki (1936, p.178, fig. 5; 1945, pp. 405, 476-477, and figs. 189, 230) who illustrates38 characters of the same type (27 of them occurring among Brown's51, 7 others which appear on lists furnished by our informants, and 4which only Someki gives). They are not presented in any formalsequence; we have located them in various positions in column G offigure 25 by means of comparison of their graphic forms with charactersin the other columns and by means of their attributed phonetic values,which, however, often deviate considerably from the values in theother columns. Someki states that the characters, which he ap-parently collected at Faraulep in 1934, occur only at Ifaluk, Elato,and Faraulep, and he illustrates a wooden bowl from Elato whichbears a few of the characters. He derives some of the symbols, whichhe identifies as of Roman alphabetical origin, from an early Europeaninfluence, and, like Imbelloni, speculates that the others are linkedto Easter Island writing.
• Two informants have read tbis line for us, as follows (the numbers are those of the characters in figs. 25and 26):39/X/4/ni/46/30/28/III/23/25/IV/III/ 11 /53/16/11/I/17Wolipwc/ masturbate / he /and/ FoimeyatIt will be noted that, as in the samples of writing collected by ourselves, there is no separation of wordsor phrases and no pimctuation. One of our informants, C, occasionally uses a tiny triangle or diamondbetween words at the level of the uppermost portion of the characters; its use is inconsistent, but it seemsto be intended to terminate phrases and sentences.
« There is also to be noted the existence In 1909 of a set of numeral signs at Faraulep (Damm, 1938, pp. 213-216). Damm attributes their Invention to a chief Saueru; he states that the symbols originated from tattoodesigns and from signs taken from Japanese newspapers, and that their recency of invention was evidentfrom their limitation to use in copra transactions. Since no other authority mentions them, and since nonative ofwhom we inquired during 1954-57 knew anything of them, we will not discuss them further.
nS*6o1^' ^''^" CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 277TYPE 2 SCRIPTThe Siidsee Expedition volumes, while they fail to refer directlyto the type of writing we have called Type 1, present a second set ofcharacters of a very different type. From Ifaluk, Damm (1938,fig. 180) gives a list of 18 characters belonging to this second type andshows their phonetic values. Damm and Sarfert (1935, fig. 278)give almost the identical script from Satawal; it contains 19 charactersof the same graphic form with nearly the same values as the Ifalukcharacters, in slightly different sequence. These two lists are pre-sented by the German anthropologists without comment or analysis,except that Damm attributes the introduction of this writing atIfaluk to a castaway missionary from Truk. The symbols arereproduced here in columns B and C of figure 26, and are of the tj^peof writing which the present authors will call Type 2. Not only arethe symbols and their values different from Type 1 writing; they areclearly derived, as is evident upon simple inspection, from Romancharacters, while the symbols of Type 1 in nearly all cases show noresemblence to Roman alphabetical characters.These are the only two series of Type 2 that we have found in thepublished materials. But in the lines of writing from the Faraulepmen's house, previously mentioned, there are also some symbols ofType 2. And in the three words which Brown appends to his Woleailist there are three characters which mystify him, since they do notoccur in his list of 51 Type 1 characters, but which can be identifiedfrom the Ifaluk and Satawal series as belonging to Type 2. Besidesthis, all of the illustrated objects previously mentioned bear char-acters of this type in addition to the ones of Type 1, and there areadditional illustrations of a Lamotrek house and a Puluwat canoedecorated in Type 2 characters only (fig. 26, cols. D-H). And inSomeki's list there are 11 characters of this type (fig. 26, col. I).We have, then, evidence that in 1909 both types of S3anbols wereknown at Woleai, Faraulep, Puluwat, and probably Satawal, if notelsewhere, and that at least Type 2 characters were known at Ifalukand Lamotrek. In 1934, the date of Someki's visit, both systemswere known at Ifaluk and Elato.
Explanatory notes for figure 26ColumnA. Numbers assigned to characters. Numbering after M.'s sequence (col. J).B. Damm (1938, fig. 180). Sequence, characters, and attributed values.Ifaluk, 1909.C. Damm and Sarfert (1935, fig. 278). Sequence, characters, and attributedvalues. Satawal, 1909.D. Damm (1938, fig. 280). Characters from beams of men's house. Faraulep,1909.E. Damm and Sarfert (1935, fig. 128). Characters from tinderbox. Satawalor Puluwat, 1909.F. Damm and Sarfert (1935, figs. 212, 272). Characters from canoe and flute.Puluwat, 1909.G. Kramer (1937, pi. 9c). Characters from beams of men's house. Lamotrek,1909.H. Kramer (1937, pi. 15 and fig. 109). Characters from canoes and bamboobox. Woleai, 1909.I. Someki (1936, fig. 5, p. 178; 1945, figs. 189, 230). Characters and attributedvalues. Presumably Faraulep, Ifaluk, and Elato, 1934.J. M. of Pigue, Faraulep, 1954. Sequence, characters, and attributed valuesin Smith's orthography.C. of Pigue, Faraulep: 1955A, 1955B, and 1956. Sequence, charactersand attributed values; attributed values of column L in Smith's orthog-raphy.R. of Ifaluk: 1955, 1957A, and 1957B. Sequence, characters and attributedvalues; attributed values of columns N and O in Smith's orthography.Q. A. of Falalap, Woleai, 1957. Sequence, characters, and attributed values.R. L. of Pigue, Faraulep, 1957. Characters and attributed values.S. N. of Pigue, Faraulep, 1957. Characters and attributed values.T. O. of Falalus, Woleai, 1957. Characters and attributed values.U. P. of Eauripik, 1957. Sequence and characters.V. S. of Ifaluk, 1955. Characters.REMAEKS1. Positions in sequence of column A assigned to characters in columns D-I, R,S, T, and V by comparison of form of character, and of value if given, withform and value of characters of the other columns.2. Bracketed characters are those not given in informants' lists but occurring inword samples; or are variant forms from word samples.3. Values in columns B, C, and I are as given by Damm, Damm and Sarfert,and Someki respectively. Attributed values in columns J, L, N, and Oare based on hearing of tape recordings and are in Smith's orthography,except those in parentheses, which were heard under hurried field conditions.Those in column N were obtained from informant R. reading not his ownlist of characters but the list by informant M. (given in column J). At-tributed values in columns K, P, Q, and T, in parentheses, not taped; aregiven as roughly recorded in the field. Attributed values in columns M,R, and S, in parentheses, not taped; values were written by these inform-ants in Japanese katakana, and are here shown in the orthography usedby Prof. Y. Uyehara of the University of Hawaii, who has transcribed thekatakana characters for us.*4. Numbers preceding characters in columns B, C, J, P, and U, form separateseries, independent of any series of numbers in figure 25. Numbers preced-ing characters in columns K-0 are in the same series as those precedingcorresponding lists of characters in figure 25 by the same informants, sincethese informants gave lists containing characters of both types.
•See footnote following notes to figure 25.
]
t
.< H
O.,'
J
I
oa)
ovD
Io
r~

B CDEFGHIJKLM N P Q R S TU V
I

^°tj5°P- P'^P- CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 279PKESENT-DAY KNOWLEDGE AND DISTRIBUTION OF THEWRITINGWe were able, in 1954 to 1957, to obtain lists of symbols of one orboth types from various living informants of Woleai, Faraulep,Lamotrek, Ifaluk, and Eauripik. These sjmnibols are included infigures 25 and 26 under informants' names. In addition we havesamples of the writing, though not lists of characters, from severalother persons of all these atolls; and additional persons were able toread or at least to identify many of the characters. We also havesamples from tattooing and from canoe-house beams (pis. 42-44).The number of people who today know the writing is not certain.At Faraulep two men and two women provided us with lists whichincluded both types of characters (fig. 25, cols. H-M, and fig. 26,
cols, J-M, R, S), and the two men wrote sample words and texts;another two men and two women (two of them of Woleai origin) wereable to recognize from 14 to 38 Type 1 characters and 15 to 17 ofthose of Type 2; still another woman is said to know the wi'iting; inthis list of Faraulepese familiar with the system we should also reckona tenth person, a man who died in 1955, but who the previous yearhad recognized 17 symbols of Type 2 and is said to have known theother type. As for the other atolls in the Woleais
:
Woleai atoll: Two women of Falalap Island gave us Type 1 listsand wrote Type 2 characters in sample words (fig. 25, cols. R, S;fig. 26, col, Q) ; a Falalus man wrote a Type 2 list (fig. 26, col. T) ; aWottagai woman and a Siliap man wrote some Type 2 charactersand words; two Wottagai men could read symbols of Type 2; and wecopied examples of writing in tattoo and on house beams and tinder-boxes at Falalus and Wottagai (as we did also at Faraulep) . Eauripik
:
One man wrote a Type 2 list (fig. 26, col, U), and another man recog-nized 16 of the Type 2 characters and wrote sample words; a thirdman, the last at this atoll who knew Type 1 writing, died recently.Ifaluk: One man wrote three lists containing both types (fig. 25,cols. N, O; fig. 26, cols. N-P) and a long text in characters of bothtypes, as well as sample words; another man wrote a similar text(his characters are shown in fig. 25, col. T, and fig. 26, col. V) ; and athird man, who claimed to have once known the whole system,could read a large number of the characters. Lamotrek: A man ofIfaluk origin who learned the wi-iting at Ifaluk by means of lettersfrom Faraulep, gave us two Type 1 lists (fig. 25, cols. P, Q) and wasable to read additional characters of both types; he also wrote samplewords; a second man recited the lists orally; and a third could readmany of the characters; and again there was writing on house beams.Elate: No one knew the system, although several persons bore
280 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
tattooing in Type 2 characters done by a woman now dead, and wecopied characters from house beams. Satawal: One man who diedin 1955 liad the previous year recognized some Type 2 characters;no one else literate in the writing could be found, although a Satawalwoman who did tattooing in both types is still living at Palau, andat least four women (one of them from Woleai) were tattooed withthe symbols; and waiting on house beams, most of it dating fromJapanese times, was once more recorded. As for Puluwat, east ofthe Woleais, though three informants knew the writing for what itwas, they could read none of it and stated that the few people whohad known it were dead. Altogether we have samples of writingand/or lists of characters from 27 living or recently deceased persons,as well as characters copied from tattooing and from beams and otherobjects. No doubt some of the people whom we tested only forreading ability can also write in the script.It would appear, then, that the writing has or once had a geo-graphical distribution from Eauripik in the west to Puluwat, 300miles to the east, and was known on all the inhabited islands between.(See map 1.) Specific inquiry elsewhere in the Carolines establishedthat it had not existed beyond these limits,^ but it was often recog-nized for what it was; people on Pulusuk, for example, have heard ofit as "writing of Faraulep." Within the area where it exists, notmany persons seem ever to have known it, and knowledge of it isdeclining. While formerly there was some interest among youngerpeople in learning the writing, many today use an adaptation ofJapanese katakana writing instead, and the children are being taughtto write in the English alphabet in Government schools. All of ourinformants were past their youth. Previously, Avhen more peopleknew the system, it was used for writing letters to one another, oftento request supplies of native and European commodities, but nowa-days, with travel made easier and with stores available, this functionof writing has lapsed. The few people who know the script todayuse it primarily to record chants and magical and medicinal formulae.One man says he learned the writing specifically in order to be able torecord songs, medicines, and magic, which he keeps in a notebook.A recent convert to Catholicism keeps a notebook of catechism lessonsin the writing. An Ifaluk man who, in late Japanese times, becamelost at sea, states that during his misadventure he kept an account inthe native writing which included the birds he saw "and their mean-ing." Lt. Kevin Carroll (tragically killed in Iran in 1957), who wasan administrator in the military government at Yap in 1946, told usthat he sometimes transmitted orders to the Central Carolines,through an Ifaluk amanuensis, in the native script.
' We have no Information for Pulap, just east of the Woleals.
Anthrop. Pap. CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 281No. 60]
I?,
10 E
H a
3 0.3o c
-i?
i
s
4-1a,
(L)
1)UcVu!-•uuOu
7^ ^ —r-o
282 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173NUMBER AND SEQUENCE OF CHARACTERSFrom all the sources previously mentioned and from the lists ofcharacters and samples of writing we ourselves have collected, we havea total of at least 78 characters of Type 1, to most of which we canassign phonetic values, and 19 of Type 2. We also know that there isa definite sequence. For Type 1, M.^ of Faraulep gives a list of 42characters identical in sequence with the first 42 of Brown's list fromWoleai. Of the three lists obtained from C, one is identical insequence as far as Brown's No. 21, one as far as No. 22, and the thirdto No. 50 (except for character 35, which is given as the 80th in C.'slist). Similarly, of R.'s two lists containing Type 1 characters, onecoincides with Brown's as far as character 43 (with some omissions^and the other up to character 20 (with one omission) and again from29 to 43. From T., one list (with two omissions) runs in the sameorder up to No. 18, and another list (with some omissions) to 47.B. gives a list which duplicates Brown's from 1 to 12, omits 13 to 38,but resumes at 39 and runs to Brown's 49. And the list by A. runsin Brown's sequence to No. 44 (with four omissions and with No. 28out of order). The list from the Faraulep wooden board of 1909,with three omissions, also runs in the same sequence up to No. 32.As for Type 2, only five lists are given in sequence separate fromType 1 (fig. 26, cols. B, C, J, P, U). It will be seen that four of themagree as far as No. VI. The Ifaluk list of 1909 and the Faraulep listof 1955 agree completely in sequence except for the omission of onecharacter in the former. The Satawal list of 1909 and the Eauripiklist of 1957 likewise are in agreement (not considering omissions) asfar as No. XI, and both have No. XIX in seventh position.It is evident, then, that we have here a system of writing which wasweU-defined some time before 1909.ORIGIN OF THE WRITINGDISTINCTNESS OF THE TWO TYPESWhat is the origin of the Carolinian writing? We may disregardthe speculations of Brown, Diringer, Imbelloni, and Someki, sincethere is no evidence to support them and they border on the fantastic.In answering this question, it is important to note that of the five listsof characters we have obtained from published sources, two (Brown'sWoleai list and the list from the wooden board found at Faraulep byDamm) contain only one type, the non-alphabetical type which we
' Designations of informants referred to in this paper are as follows: A.: Marutang of Falalap, Woleai.B.: Nachomai of Falalap, Woleai. C: Chiyemal of Pigue, Faraulep. L.: Laichib of Pigue, Faraulep.M.: Magilo of Pigue, Faraulep. N.: Nesawen of Pigue, Faraulep. O.: Maluchorang of Falalus, Woleai.P.: Fagolifek of Eauripik. R.: Maroligar of Ifaluk. S.: Tarof of Ifaluk. T.: Tachep of Lamotrek.
No*60]^' ^'^^' CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 283have called Type 1; two (the Ifaluk and Satawal lists published byDamm and by Damm and Sarfert) contain symbols which are all ofType 2; only one (Someki's), which appears to be in no formal se-quence, has both types. In other words, the natives from whom thelists were obtained themselves consider the symbols to be of twotypes; we have not sorted them out on any logical grounds. Also,some of our informants gave us the two types in two separate sets.Other informants gave us mixed lists, but nevertheless distinguishthe characters as belonging to two types of writing.SIMILARITY OF TYPE 2 SCRIPT TO THE ALPHABETWhen we first examined the symbols it was immediately apparentthat those of Type 2 were taken without great alteration from Romanalphabetical symbols ; they all appear to be modified forms of our ownupper-case letters; whereas most of those of Type 1 bore little resem-blance to the alphabet or, it seems evident from examination of theexhaustive compilation of other forms of writing illustrated by Dirin-ger, to any other known system of writing. It therefore suggesteditself to us that Type 2 was first introduced into these islands fromsome European source, and that, perhaps because it fitted poorlyinto the native phonetic patterns, another system, Type 1, was thendevised in order to fill a need for more adequate representation.SYLLABIC VALUESIt was apparent also that neither type of symbol was used alpha-betically, except for symbols representing vowel sounds alone. Thethree words that Brown gives us indicate that both types were beingused in 1913 to represent syllables, not single phones. The wordsand phrases we later obtained from our own informants verified ourguess that this was in fact a syllabary, and suggested what the processof development had been. AU the symbols, of both types, have nameswhich are also their attributed phonetic values (although, as weshall see, in actual writing values are often only approximate). Ex-cept for characters representing vowels alone, nearly all of whichbelong to Type 2, they represent open syllables composed of aninitial consonant or semivowel followed by a vowel. Fm'ther, everysymbol of Type 2, excluding those representing vowels alone, hasan attributed value whose vowel portion is a long i, while all symbolsof Type 1, with two exceptions (Nos. 7 and 67) have as their vowelportions attributed values other than i. Writing is accomplishedby a mixed alphabet-syllabary system; when a syllable consists ofa vowel alone, the character for that vowel is used, as in alphabeticalwriting; when it is formed by a consonant-plus-vowel or semivowel-plus-vowel combination, the appropriate syllabic character, of eithertype, is used. (See table 2.)
284 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173TYPE 2 HISTORY RECONSTRUCTEDThis evidence caused us to guess that a European alphabet or amodified form thereof had been introduced to these islands, but withnames attached to the letters different from those we know them by;that the names for the consonantal letters consisted of the phoneticvalue of the consonant followed by an i suffix; that the natives didnot understand the acrophonic principle upon which the names werebased, hence did not attempt to write alphabetically but took thesenames as having syllabic value and tried to write their language withthem ; ^ and that they devised the other system of writing, Type 1
,
when they found the first system of syllabic representation, Type 2,inadequate to reproduce all the sounds of their language. Thisreconstruction of history seemed consistent with the consistentlyopen form of the syllable in this language, syllable-final consonantscommonly occurring only at the ends of words. Binary geminatesequences occur, but dissimilar consonants are almost always sepa-rated by at least an excrescent vowel. Among the 301 Woleai wordsthat we have assembled from the text material in Smith (1951),written in his orthography, only three combinations seem to be excep-tions, those italicized in the place names So/w/g, Ysi/n/rw/ipii/g, and Ya/nga/ZA/ge/ra/i/lh. Thus, since the spoken language in large part iscomposed of open monosyllables, such a system of wi'iting serves itwell. When a spoken word has a terminal consonant, only the con-sonantal portion of the final character used in writing the word re-tains phonetic value, and the vowel portion which follows it becomesvalueless. POSSIBLE SOURCESWhere could the natives of these islands have obtained theRoman characters? Dates of discovery by Europeans range from1686 for Faraulep to 1828 for Eauripik, but the natives were in inti-mate contact with other islands which had earlier contact with theWest; Yap, for example, was discovered in 1526, and Fais in 1543.In the late 18th and early 19th centuries, visits by explorers and trad-ers became fairly frequent, and a number of ship's deserters andcastaways have left accounts of their stay in various of the Carolines.But missionaries failed to establish permanent stations in these islandsuntil the end of the 19th century, and the natives remained in vir-tuallyaboriginal condition. The natives themselves were probably moreactive agents in culture dispersal in those days than were explorers,traders, or missionaries. They are skilled mariners and navigators,and possess a remarkable geographical knowledge. Much has beenwritten of the navigational skill and exploratory zeal of the Polyne-
• For examples of writing In 1909 with exclusively Type 2 characters, see Damm and Sar/ert (1935, p. 277).
No*6oT"^*^' CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 285
sians, but the similar qualities of the Micronesians have remained inobscurity. According to Hornell (1936, p. 438), "In all the Microne-sian groups of islands the design of the outrigger canoe reached ahigher level of development than in any part of Polynesia, as did alsothe knowledge of the science of navigation possessed by certain ofthe islanders." Meinicke (1876, p. 374) likewise refers to the nativesof the Carolines and Marshalls as the foremost mariners of the Pacific,far surpassing the Polynesians in this respect, and Kramer comparesMicronesian and Polynesian geographical Imowledge with similaradvantage to the former. In earlier days a flotilla of canoes from theCentral Carolines assembled each April at Gaferut^° and made the300-mile trip thence to Guam in the Marianas in 8 days; canoes fromWoleai, Faraulep, Lamotrek, Elato, Satawal, Puluwat, and possiblyPulusuk and Namonuito participated in this expedition; they tradedshells, mats, cordage, and canoes for iron knives, beads, and cloth,and made the return voyage in May or June. Guam, be it noted,had been missionized by the Spanish beginning in 1668. This com-merce, apparently interrupted by the Spanish conquest of the Mari-anas, was resumed in 1788 and persisted until 1873. Kittlitz (1858)encountered Caroline natives in 1827, e. g., at Faraulep, who alreadyspoke fluent Spanish, an ability no doubt acquired on visits to Guam.Objects of Marianas origin were traded farther east than the islandsdirectly involved in the commerce, e. g., to Truk and Namoluk (Finsch,1900, p. 48; Gu-schner, 1912-13, p. 180), which got theh- iron toolsand tobacco from Puluwat. Lutke (1835, p. 295) found cats on Luku-nor (in the Nomoi group), known by the Spanish name "gato," which,no doubt, were obtained from the Marianas via such native traders.A permanent colony of Carolinians grew up in the Marianas after1815, with settlements at Guam and Saipan, and later at Tinian.These colonists, who numbered many hundreds, came, and continueduntil recent years to come, from the very islands we are here concernedwith, as well as from others which lack the writing; visits back andforth to their home islands were frequent; much trading involvingEuropean manufactures occurred.There was also contact with the Spanish in the Philippines. TheSpaniards at Guam employed Caroline crews to take them as far asthe Philippines. The journal of the Salem ship Clay, Capt. W. R.Driver, reports finding in the Fijis in 1827 two natives of the Carolinesleft there by a Manila brig 5 years before; these men signed on theClay as crew members and returned to Manila. Traditions of sea-farers cast away in the Philippines and successfully returned homeare known as far east as Puluwat. The first knowledge of the Woleaiscomes from Spanish accounts of 30 canoeloads of people from these
" Not West Fayu, as Is often stated In the older literature on the subject.471762—60 20
286 BUREAU OF AMERICAN ETHNOLOGY [BuU.173
islands driven ashore at Mindanao and elsewhere in 1664, and theliterature contains many more such reports in later years, includingthe most recent case in 1954. Many of these castaways made theirway home, either by themselves or aboard foreign vessels.Another point of contact with the outside world was by way ofYap, whose dominion over all the Central Carolines has been describedby several writers, ^^ and continues in diminished degree until today;formerly it extended farther east than the islands we are concernedwith here, and contacts with Europeans would have been possible atboth ends of the area involved. Mission activities at Yap, to be sure,did not commence until 1886, but traders were active there earlierin the century.Finally, maps drawn by natives of these islands and sailing directionswhich they are capable of giving reveal knowledge of places in thePhilippines, parts of Indonesia, the northern fringe of Melanesia,and some of the islands of the Gilberts and Polynesia.Nevertheless, none of the foregoing gives us any clear indicationthat some early contact with the West or with literate natives fromother regions resulted in familiarity with writing among CentralCarolinians. We have only one report of writing from this areabefore the German expedition's visit in 1909: Arago (1822, p. 35)reproduces a letter from a Satawal chief written in response to theorder of a trader at Rota in the Marianas; the writing used in thisletter is purely pictographic, the chief having made drawings of theobjects he desired in return for the shells which he had for barter, andthere is not the slightest resemblance to the system of writing we arehere concerned with. Nor is this system reported by Chamisso,Choris, Dumont d'Urville, Freycinet, Kittlitz, Kotzebue, Lesson,or Liitke, all astute observers, in the early 19th century, and it is notlikely that it would have escaped the attention of Christian, Finsch,or Kubary in the latter part of that century if it had existed then.It seemed to us therefore that the writing must be of more recentorigin. Also it appeared that the most likely place to look for itssource was in the islands to the east of the area concerned. To thewest and north are Palau and the Marianas, whose inhabitants speakIndonesian languages, and Yap, whose language, although it isusually classified as Micronesian, is very different from the languagesof the Central Carolines. The borrowing of Roman characters fromthese islands, even though some of the natives of the Woleais speakYapese, would for these reasons have been difl&cult. But moreimportant, alphabetical writing was introducted to the Marianas,Palau, and Yap by the Spanish, and the letters of the alphabet, as
11 See Lessa, 1950.
No*6o1^'^^^" CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 287given orally by the natives of those islands today, all have modifiedSpanish names, very different from the names in the Central Carolinesof the Type 2 characters which we have considered to be of alphabeticalderivation. LINGUISTIC AFFILIATIONSThe language of the Central Carolines belongs to a larger linguisticgroup which has a geographical range stretching from Ulithi in thewest nearly to Truk and Nomoi (or the Mortlocks) in the east, andincludes also the islands south of Palau. This area has four sub-divisions: in the west, Ulithi, Fais, and Sorol; to the southwest,Sonsorol, Pulo Anna, Merir, Tobi, and probably Mapia; to the east,Pulap, Pulusuk, and Puluwat; and in the center, all the inhabitedislands of the group commonly referred to as the Woleais: Faraulep,Eauripik, Lamotrek, Elato, Ifaluk, Satawal, as well as Woleai atollitself. The Woleais, together with Pnluwat, constitute the areawhere the writing is known. The languages of these four subdivisionsare mutually intelligible, differing only in some phonemic shifts andin some minor vocabulary changes. Farther east, the inhabitants ofTruk and of the islands to the north and south of Truk, including theNomoi group, speak dialects of another language, whose relationshipto the first language group is close but not sufficiently so to permitmutual intelligibility. However the two vocabularies offer a greatnumber of cognates and the phonemic patterns are not remarkablydifferent (Smith, 1951).THE ALPHABET OF TRUKIn the first language area there has not been, until the time ofthe American administration, beginning after World War II, anyliterature in the native tongue. But in the second area there hasbeen, since 1878, a series of Bibles, hymn books, catechisms, arith-metic instruction books, and so on. These are all in the Nomoidialect, and were all written by the missionary Logan, who was as-signed this task by the American Board of Commissioners for ForeignMissions.^^ The books are used not only in the Nomoi group but in
" Available to us are the following, all by the missionary Logan:Maku en Mark. 1880 and 1882 editions. Honolulu.Kapas fel, puk eu, kapas en lorn kana, Mortlok. 1880. Honolulu.Nor an lamalam kana; kapas an Mortlok (Mortlock Island Hymns). 1881. Clnciimatl.Puk an afalafal, kapas an Mortlok. 1881. Honolulu.Testament Sefa an amam Samol o Ran amanau JIsos Kralst. 1883 (1st edition) and 1944 (7th edition).New York.Kapas fel, puk eu: kapas en Kot, Mortlok. 1884. New York.Aritmatik. 1887. Honolulu.Puk an als fel, me ais an lamalam kana (Mortlock catechism). 1888. Honolulu.Puk an k61, me k61 an lamalam kana, Ian kapas an Ruk me Mortlok. 1888 (2d ed.), Honolulu; andn. d. (3d ed.), San Francisco.
288 BUREAU OF AMERICAN ETHNOLOGY tBuU.178
all of the Truk area without accommodation to dialect difference.They are printed in ordinary roman characters, some of them alsousing a number of diacritical devices. '^Trukese today, who write alphabetically and who still use the LoganBible unaltered from its 19th century form, when asked to recite theiralphabet nearly all respond with the following sequence : A E I O U F SKLMNATTRT^T.^* These are the very same characters that areused in three of Logan's books, except that Logan also uses a J. (Jis used on Truk by the older generation, almost solely for writingbiblical names.) Also one of the Logan books {Puk an afalafal, 1881,p. 1) contains, apparently for instructional purposes, almost the samesequence:AEIOUFJKLMNA^PRi?ST. In oral recitation,the vowels have approximately Spanish values, and the consonantswhich follow are given as though suffixed by i, thus: fi, si, ki, li, mi,
ni, ngi, pi, ri, chi, ti. That is, the names of the Trukese consonantsare the same as the names and phonetic values of the Type 2 syllabiccharacters of the Woleais.^^ There are only five vowels as againstthe eight in the Type 2 lists, and the sequence of characters is shghtlydifferent—we will shortly attempt to explain these differences—butthe relationship would seem to be obvious. The Truk area wouldseem, then, to be the source of the Type 2 wiiting of the CentralCarolines. Moreover, inasmuch as Logan began his work in 1878,Type 2 writing must have come into use since that date.ALFRED SNELLING AND THE TRUKESE CASTAWAYSHow did the Trukese alphabet get to the Woleais? No AmericanBoard missionary was ever stationed there, but the Ifaluk informantsof Damm told him that a missionary from Truk, who was en route fromUlul (in Namonuito atoll) to Truk in a sailing canoe, together with anumber of Trukese, was driven ashore at Eauripik at some unspecifiedtime; there he remained until a Woleai chief arrived and invited himto accompany him back to Woleai ; the Trukese already knew how to
'3 Logan's books all contain the following roman characters; A, E, I, 0, IT, F, J, K, L, M, N, P, K, S, T.Three have in addition AT and R. Three others have N, R, A, and O. The 1888 catechism has all of thesecharacters, plus an umlaut E. The second edition of the hymn book published in 1888 has all of thesecharacters again, as well as diacritical devices—accent, two forms of circumflex, and umlaut—for A, E, I,and O, which also have italicized forms. The 3d edition of the same hymn book (n. d., but published before1899) converts all of these forms to Italicizatlon. After 1900 the only nonstandard orthographic devices areItalicized N, R, A, and O.1* The Spanish, before 1900, Introduced their own alphabet, which has never caught on, one reason beingthat two of the consonants, as pronounced by the Spanish, form words with vulgar meaning.
" The alphabets of Ponape and Kusaie, in the Eastern Carolines, and of the Mai shall Islands are similar;that is, they are given with the vowels, pronounced with Spanish values, fii-st, then the consonants whichagain all have names with i endings. Apparentl y this regularization of consonantal names was the standardpedagogical device of the American Board of Commissioners for Foreign Missions missionaries. Theseislands, however, have languages with phonemic systems different from those with which we are concernedin this paper, a fact reflected in the local alphabets; thus the missionaries did not introduce F at Ponape,where there is no corresponding phoneme, but did introduce both D and T. These alphabets, therefore,could not be the source of the Type 2 writing.
Na6oT"^^^" CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 289
write and taught the art to their hosts at Eauripik and Woleai; theaccount goes on to say that the Trukese went back to Truk via Ifalukand Lamotrek and taught writing to the inhabitants of these atolls too.To this account is to be added that of Burrows and Spiro (1953, p.201), who say that writing "came to Ifaluk from Woleai, where it wastaught by a missionary." (Burrows gives no further informationabout the writing.)One of our informants, a man of Eauripik, confirmed the foregoingin the following words: "An American Protestant missionary fromTruk got lost on a boat during German times (1900-1914). Hestayed there (Eauripik) and taught the people (how to write) ....His name was Misinining. He was there only for three or four monthsand left for Falalap (in Woleai) where he soon died." AnotherEauripik man refers to this missionary as Misililing and remembersthat he and the Trukese all gave instruction in writing.All of these accounts, of course, must refer to the writing we havecalled Type 2, since the castaway party from Truk must have usedfor instruction the alphabet known to them, and we have seen that thealphabet is the inspiration of Type 2 writing.The mention of a missionary called "Misinining" sent us to therecords of the American Board of Commissioners for Foreign Missions,where we found that a Rev. Alfred Snelling was missionary to Trukbeginning in 1888, and that he was lost at sea in a native boat in 1905(Bliss, 1906, pp. 129, 162; American Board of Commissioners forForeign Missions, Annual Reports, 1886 to 1906). The equation ofMr. SneUing and "Misinining" is obvious. Damm (1938, p. 133)and Kramer (1937, p. 203), in context unrelated to anything bearingon the script, also refer to Snelling. Damm relates only that Snellingsailed in a native canoe for 93 days until he reached Eauripik, whencehe went to Woleai. Ki'amer gives the story in greater detail, tellinghow Snelling, en route from Puluwat and Ulul to Truk, went astrayand drifted for 3 months until he reached Eauripik, whence a boatcrew from Woleai fetched him to their island, where he died.None of these published accounts which mention Snelling say any-thing about his role in the introduction of writing. However, Mr.Frank Mahony, presently District Anthropologist at Truk, has estab-lished the connection in the following interview with a Trukese namedAiras. The story of Snelling is well known at Truk, but Airas, whowas a student at Snelling's school in 1900, and who made the ill-fatedvoyage with him, is apparently the last survivor of that journey.Mahony relates:
Snelling left Totiw [Tarik] Island [in Truk atoll] intending to go to Ulul andPiserach [both in Namonuito atoll]. He picked up the chiefs of Puluwat and Ululon Etten [in Truk atoll], then west to Tol [in Truk atoll] to get Sirom and Kinion
290 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173[two men]. Snelling took 400 baskets of preserved breadfruit with him to giveto . . . Piserach and Ulul. They left Truk and went to Piserach, thence toUJul, then back to Piserach. They left Piserach around 2:00 PM to make thereturn journey to Truk. . . . Ten hours later . . . they were still not outof sight of Piserach. . . . At 3 :00 AM a light breeze blew up . . . While Snellingslept some passengers they had picked up on Ulul . . . talked the crew intochanging course. . . . From then on they searched in vain for Truk. Theyhad a little copra and some drinking coconuts aboard. ... At the end of seven-teen days this was all gone. They managed to catch a few fish. They were atsea about ninety days until finally they got to Eauripik. Meanwhile . . . fourof them had . , . died of starvation. The people of Eauripik took very goodcare of them. . . . When a big chief of Woleai named Okupeniar [Kramer giveshis name as Agupelior] came they went with him to Woleai . . . [The three Jap-anese living there] took Snelling into their house. . . . Snelling knew he wasgoing to die and asked to be moved ... to the men's house , . . and in twenty-four hours he was dead. . . . After a few months Airas and the others leftWoleai and went to Lamotrek. . . . Then they went on to Satawal and on toPuluwat . . . then after a couple of weeks . . . back to Truk.Maliony:Did you hold school on the islands?Airas:No, we 1* just taught them to write. . . . They didn't know how before. . i .We taught (the Trukese alphabet) to the people of Eauripik and Woleai . . .but not to the people of Satawal, Puluwat, and Lamotrek. . . . We didn't staylong enough on these islands, and besides they already knew how to write. OnLamotrek the Guamanian wife of the white trader had taught the people how towrite. . . . The people of Satawal probably learned from Lamotrek.ADOPTION OF THE TRUKESE CHARACTERSAiras recalls the following alphabet as used by himself and theother Trukese instructors: AEIOOUFSKLMNNPRRJT.It differs from the modern Truljese alphabet only in having two extraletters, O and J. Of these letters, all six vowels occur in the Type 2writing of today, where the vowels A, E, I, and O are to be found inpositions I, III, IV, and V, and where U occurs in different positionson different islands (see fig. 26). The fifth character in Airas' series,O, is no doubt the sLxth in the various versions of Type 2 writingand is to be equated with the italic O of some of the Nomoi boolisof Logan. As for the other two vowels in Type 2 ^vriting, one of them,the second in all the series, is very likely the italic A of most of theNomoi books, and the E-like character. No. XVI, is possibly derivedfrom the italic E which we have found in three of those books (thoughwe later suggest a possible Japanese katakana derivation.) Atras'failure to include these two letters in his series may very well be due
"0 For the record, we should mention the names of the other Trukese teachers, who, with Airas, carriedon their Instruction, using pencil and paper as -well as writing in the sand. They are Kinion, Sirom, TJne-itor, and Resapechik.
NaeSr"^^^' CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 291
to a fault of memory; 50 years have passed since his adventm-e, duringwhich time the Trukese alphabet has become standardized in theversion given on page 288, so that Airas may be influenced by modernwriting, which tends to omit italics.All the consonantal names in the alphabet given by Airas are pro-nounced by him with i endings. Identification of most of them withthe Type 2 characters is readily apparent upon inspection and neednot be discussed. A few Type 2 characters have been altered in ori-entation (as Nos. IV and XI) or in minor detail (as Nos. VIII, X,XIII, and XIV) .'^ Only the Type 2 symbol representing the soundwi (No. XII) is at all puzzling. Possibly it is the letter J, which oc-curs in aU the Nomoi printed books, where it is used interchangeablywith S; it occurs also in Airas' alphabet, where, like S, it has the name
si. J is disappearing in modern Trukese writing because of thisequivalence with S, but its former position is shown by its replace-ment by S between F and K. Now if J was indeed among the letterstaught to the people of Eauripik and Woleai, as Airas states, itsobvious superfluity may have resulted in its being used in Type 2writing to represent a sound for which no symbol was available butwhich was felt to be required, namely wi; the position of characterXII at the very end of the Satawal list (see fig. 26), as though it wastacked on after this transformation was effected, suggests this explana-tion.^*The only letter in both the Trukese alphabet and in Airas' list whichdoes not appear in Type 2 writing is L. In the Nomois, where Loganworked out the alphabet used in all the printed materials of the Trukarea, I and n are both phonemes, as they are also, apparently, in all ofthe low islands around Truk. But at Truk itself there is no I phoneme,only n, and L is used in writing when necessity is felt to defer to tradi-tional spellings based on the Nomoi Bible, the Bible used at Truk.Now at Woleai atoll there are four corresponding phonemes, whichSmith (1951) writes I, Ih, n, and nh (the h indicating a release). Thesephonemic distinctions are probably made also in the speech of most ofthe other Woleais. Yet in the writing no necessity is apparently feltto make the same distinctions; we shall see that this holds true for thecharacters of Type 1 also, that is, those characters whose syllabic
" Thanks to Mr. Mahony, an old man of Murilo (an atoll north of Truk) has furnished us with examplesof the alphabetical characters used In the Truk area about 50 years ago, apparently as the missionaries taughtthem. There are some differences in form of the characters when they are compared with the Type 2writing of today in the Woleais, but at least his E, N, and T bear the same peculiar embellishments asthe corresponding Woleai characters (see fig. 26), so it would appear that the alteration from the graphicform of Roman upper-case characters did not occur in the Woleais but in the Truk area.
" Airas would seem to have transposed J and S in position in his series. However, a Nomoi man whowent to Snelllng's school at Truk from 1897 to 1901 gives the final letter of the alphabet as he learned it asJ, so possibly Snelllng's party introduced It to the Woleais in this position. Modern samples of tattooingcollected by Mr. Mahony at Puluwat contain a J-llke character which Is read as wi, not «».
292 BUREAU OF AMERICAN ETHNOLOGY [BuU. 178
values contain these four consonants, suffixed with vowel sounds otherthan i. If L was among the letters taught by Airas and his com-panions, it had dropped out of use as a Type 2 character between1905, the date of Snelling's drift voyage, and 1909, when the SiidseeExpedition collected Type 2 lists at Satawal and Ifaluk. However, inthe various Type 1 series there is an L-like character at position 33,whose value Brown gives as la and the consensus of our informantsmakes to be ne or nae. We may guess that the Trukese castaways in-troduced L and N as equivalent in name, because of the Trukese pho-nemic pattern, and that L was therefore converted into a Type 1symbol whose value was rather similar; one of our informants, C,actually states that character 33 was indeed so derived.Airas' account is that the Trukese castaways taught the alphabetas they knew it only to the natives of Eaiu-ipik and Woleai; that theGuamanian woman at Lamotrek had already taught the people of thatisland to write; that Satawal probably got its writing from Lamotrek;and that by then the people of Puluwat also knew how to write. How-ever, it is hardly possible that the Satawal writing of Type 2, recordedin 1909, would have had the consonant-plus-^ value pattern, as it did,if it were of Guamanian origin. Guam uses a standard Roman alpha-bet, whose letters have Spanish names, while the Satawal charactersof Type 2 are in form, name, and value completely in the Woleai pat-tern. Moreover, the form, name, and value of the Lamotrek char-acters obtained in 1955, as well as the form of the Lamotrek and Pulu-wat characters of 1909 painted and incised on the various objects, pre-viously mentioned, are all of the same pattern. As for the Guamanianwoman, a Lamotrek informant who in 1905 would have been about 18years old denies that she taught anyone to write. Either Airas' mem-ory is at fault or whatever Guamanian influences may have been pres-ent in 1905 had disappeared by 1909. If missionaries had previouslyvisited Puluwat, as Snelling did the islands north of Truk in the courseof his missionary work, or if natives of Puluwat attended the mis-sionary school at Truk, it is possible that there may have been a sepa-rate introduction of the alphabet there; there do seem to have beennative missionaries from Truk at Puluwat from time to time. Thismight explain Airas' statement that the people of Puluwat alreadyknew how to write. Examples of relatively recent tattooing collectedby Mr. Mahony from islands all around Truk, including Puluwat, aredone in upper-case alphabetical characters, many of which show thepeculiarities of the Type 2 characters we have recorded from theWoleais (e. g., the triangular embellishments on the cross strokes ofthe E, T, and other letters) ; these may stem from a separate influencefrom Truk.
No*60]^'^^^" CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 293
In this connection, it will be noted (see fig. 26) that the two Type 2lists of 1909 are in rather different sequence. The Ifaluk list (col. B)of 1909, however, is identical to the Faraulep list (col. J) of 1954, ex-cept for omission of No. XV, and R.'s partial list (col. P) of 1957 agreesin having No. VI out of order and Nos. XIV and XV reversed. Onthe other hand, the Satawal list (col. C) of 1909 is in fairly close agree-ment with the Eauripik list (col. U) of 1957, as well as with the present-day Trukese alphabet and with the alphabet Airas says he taught atEauripik and Woleai m 1905; all of them have U (No. XIX) followingitalic O (No. VI), then F (VIII), S (XV), and, except for one of theselists, K (VIII) ; and P (XIV) is followed by R (XVIII) and italic R(XIII) except on the Eauripik list. This would possibly suggest thatthere were two separate introductions of the alphabet, each in adifferent sequence; but then it must be assumed that in each case thesame development followed, that is, the names of the letters were bothtimes taken as their phonetic values and the letters converted into asyllabary. More likely the alphabet introduced in 1905 at Eauripikand Woleai was simOar in sequence to Airas' list as he gave it 50 yearslater, and as it diffused among the Central Carolinians a second tradi-tion developed which became fixed by 1909, as seen by the persistenceof the Ifaluk sequence of 1909 until 1954 at Faraulep.It is also of interest at this point to note a similar development inType 1 writing. Characters 24, 25, and 28, occurring in identical po-sitions in the lists of Brown (Woleai) and of M. and C. (both of Pigue,Faraulep) , are either missing entirely or are out of sequence in both ofthe lists of R. (Ifaluk) and in those of T. (Lamotrek) and A. (Woleai),as well as on the Faraulep wooden board found by Damm in 1909.Also, T.'s hst and R.'s fists place No. 52 directly after No. 47. Itwould therefore seem that at least two slightly different sequences inType 1 writing already existed in 1913 (the date of Brown's visit), andthat this difference has persisted until today. This is probably whatR. has in mind when he states that there are two different systems, onewhich developed at Faraulep Island, the other at Pigue Island (bothin Faraulep atoll), and that further changes have occurred in thecourse of teaching the writing to other people. Though we have nospecific information as to inventors of new characters, it seems obviousthat there have been many since the first invention was made; thevariability in the different lists after the first 50 characters (see fig. 25)attests to such development.SPREAD OF TYPE 2 WRITINGWe do not have a great deal of information about the spread ofType 2 writing after the introduction of the alphabet by Snelling andhis party, and what we have is sometimes conflicting. According to
294 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173Damm the Trukese castaways taught writing to the people of Eauri-pik, Woleai, Ifaluk, and Lamotrek. A Eauripik man, Maninifek, statesthat SneUing himself gave the instruction to the people of Eauripik.P. of Eauripik, who attended some of the lessons, recalls that bothSnelling and the Trukese were the teachers. Informants at Woleai,Ifaluk, Lamotrek, and Satawal state that the castaway party stoppedfor brief periods of time at those islands on their return journey toTruk, but gave instruction only at Woleai. Lamotrek informants alsosay that they learned from one Reghipol, who returned to Lamotrekfrom Woleai about 2 months after a typhoon which we date as havingoccurred in March, 1907; he had learned to wi'ite at Woleai, un-doubtedly in Type 2. But two women of Woleai, A. and B., sayType 2 writing came to Woleai from Lamotrek; this can hardly becorrect, since we know that the alphabet, from which Type 2 writingsprang, was taught at Woleai by Snelling. C. and Tereso of Pigue,Faraulep, say that their atoll got its Type 2 writing from Lamotrekalso. And Burrows (1953) reports that the Ifaluk (Type 2?) writingcame from Woleai. These are the only statements we have as to thedispersal of the Type 2 script from its presumable Eauripik-Woleaicenter. TYPE 1 HISTORYWe have already stated our conclusions that the Type 2 writingcomprises, apart from its eight vowels, a syllabary consisting of sym-bols which are all of consonant-plus-'i phonetic value, that these valuesare the same as the names of the alphabetical characters taught by acastaway Trukese party in 1905, those characters having becomeconverted into a syllabary as the result of their names being taken ashaving syllabic value. The Type 1 writing was developed throughstimulus diffusion after a period of trial with Type 2 alone, when theinadequacy of the latter was recognized (a Woleai woman makes thisstatement in vu-tually the same words); it consists, with a very fewexceptions, of characters whose values are all consonants suffixed byvowels other than i.There are some examples of attempts at an early stage in thedevelopment of the writing to set down words with the use of Type 2characters exclusively (Damm and Sarfert, 1935, p. 277). Thus u/li/ge/tis written with characters XIX/X/VIII/XVII, bo/da/u with XIV/XVII/XIX, ja/li/gi/o/m with I/X/VIII/III/XI, etc. We also haveexamples in tattooing and from modern informants who know onlyType 2 writing. B., who knows both types and who writes her nameas 1/20/21/16, gives as an alternative form X/XIII/XI/IV, all inType 2. A., writing the island name Sa/taa/wa/1, WTites it as 37/25/14/33, but gives alternatively XV/XVII/XII/X. Obviously the modern
Anthrop. Pap.No. 60] CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 295
method of writing, which uses characters of both types, is much moreprecise than Type 2 alone.INVENTION AT FARAULEPInformants from all the islands agree that Type 1 writing wasinvented at Faraulep, and the script is generally called ishilh Foeshavlap(wi-iting of Faraulep). Even people who cannot read the writing, asfar east as Puluwat, at once identify it by this name. Three informants(two Ifaluk, one Lamotrek) state that they learned Type 1 writing atIfaluk from Faraulep visitors; another Ifaluk man says he learned itat Ifaluk from a man of Woleai origin who had long been a resident ofFaraulep. Several Woleai people also give Faraulep as the place fromwhich Woleai got its writing; three Woleai women learned the systemat Faraulep, two of them shortly after the 1907 typhoon. ThreePuluwat people, none of whom can read the script, say that Puluwatobtained it from a canoeload of Ifaluk voyagers. At Satawal a Farau-lep man is said to have taught the Type 1 writing, which no one atthis atoll now can read. There also appears to have been some instruc-tion among various Central Carolinians when they worked togetherduring Japanese times at the phosphate mines at Angaur, in thePalau group. We have several statements to the effect that the systemwas learned through exchange of letters between various of theislands.^* We have no other clues as to dispersal of the Type 1writing.Several informants give the names of the inventors, all of whom wereresidents of Faraulep.^*^ Though the Ksts of names differ, there isconsiderable agreement among them. The claim by Faraulep natives
Tairu
296 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
that the Type 1 writing was invented at Faraulep, the support forthis claim by natives of other islands who state that they learned thewriting at Faraulep or from Faraulep visitors, and the universalappellation of the writing, even among people who cannot read it, asisMlh Foeshavlap, leave us satisfied that the invention was made atFaraulep and was largely, if not entirely, the work of a group ofFaraulep natives. FOREIGN INFLUENCEWe are less satisfied about the possibility of alien influence. TheSudsee Expedition reports state that at various times between 1900and 1910 there was a copra station at Faraulep to which severalJapanese seem to have been attached. One informant from Eauripikstates that the Type 1 script was made at Faraulep by a Japanesenamed Soshaki or Soshiki, and that a Filipino named Serifino orSerbino may have helped. A Faraulep man states that the Farauleppeople themselves invented the writing but were later helped by aJapanese and a Filipino. At Woleai a woman of that atoll, wholearned the writing at Faraulep and who gives the names of fourFaraulepese as the inventors, states that the Japanese helped bycontributing two characters, Nos. XVI of Type 2 and 61 of Type 1,from Japanese katakana; these two characters do actually nearlycoincide with two Japanese characters in both graphic form andphonetic value. A Woleai man living at Faraulep since shortlyafter the invention insists that the Japanese Soshiki definitely did nothelp, and a Woleai woman who has lived at Faraulep since beforethe invention (and who is named by others as one of the inventors)denies that the Filipino was involved. All other informants statesimply that the Faraulep people whose names ihey give were theinventors. We examined the possibility of Filipino influence, remotethough it might be; but none of the symbols of any of the Filipinoscripts can be related to the Woleai symbols (see Gardner, 1943;Diringer, 1948; Conldin, 1953). We also point out that two charactersin addition to Nos. XVI and 61 are similar in appearance to Japanesecharacters; these are No. 26, which resembles the Japanese kanjiform for "sun," and No. 34, which is like the Japanese kanji form for
"wood" or "tree"; together, with the addition of one stroke to char-acter 34, they would stand for "Nippon," and undoubtedly Japanesegoods labeled thus were available to or seen by these islanders.However, in neither case is the phonetic value of the symbol similarto the Japanese value, so no more than the graphic form could havebeen borrowed. Also, in the case of character 26, informants haveidentified it as a representation of a canoe outrigger platform, as willbe seen. The evidence for Japanese influence goes no further, although
No.*60*]^' ^^^" CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 297
it is possible that the facts that katakana is a syllabary and thatSoshiki may have been consulted by the inventors may have reinforcedthe idea of creating a syllabary, first stimulated through the names ofletters of the Trukese alphabet having been taken as being theirphonetic values. DATE OF INVENTIONAs for the date of the invention of Type 1 writing, native informantsstate that it occurred "after the big typhoon," when the Germanadministration had to evacuate many distressed people to islands inthe same area less hard hit, as well as to Yap, Palau, and Saipan.This typhoon can be no other than the one that struck these islandsMarch 27-30, 1907; other typhoons of which there is record areeither too early or too late. Now the Siidsee Expedition ethnologistsworked in all the islands we are concerned with during Novemberand December of 1909, and found the writing as far east of Faraulepas Puluwat. The invention must therefore have occurred betweenthese dates, and would probably have been closer to 1907 than to 1909to have had time to spread so far by 1909.
DERIVATION OF CHARACTERSThe form of the characters and their values suggest several pos-
sibilities concerning then- derivation, apart from the four which maybe linked with Japanese characters. Some of the Type 1 symbolsappear to be modified forms of the alphabetical signs of Type 2.Thus character 3 is apparently an altered T, with the value changedfrom ti to ta. Using the same criteria of resemblance in form andvalue, character 8 would be derived from R, 11 from M, 12 from N,33 from L, 40 from S, 48 from A^, and 66 from F.Other characters, as their graphic forms show, are attempts torepresent natural or artificial objects, and the values of these char-acters are also the same as or close to the names of such objects.These constitute a kind of rebus writing. Among such characterswe may list the following identifications made by informants:
5. Sprouting coconut9. Bird's wing13. Ulcer, boil14. Canoe (represented under sail)16. Forked branch used for hand net and flying-fish net19. Portion of bonito26. Canoe outrigger platform28. Fish backbone29. Trigger fish30. Perfume bottle31. Woman's breast
298 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
32. Midrib of coconut palm leaf (showing leaflets to either side)35. Saw36. Coconut palm tree41. Porpoise43. Lure of bonito hook45. Leaf53. Leaf of Hibiscus tiliaceus56. Fishhook (modern type)60. Canoe seat64. A plant bearing this nameNos. 29 and 41 have the same graphic form and the same names asconventionaHzed tattooing elements, and may have been takendirectly from tattoo design rather than from the animals they repre-sent. No. 24 also may be a tattooing design.Some other proposed identifications, made not by informants butby the authors, using the same criteria of similarity in form andname, are:
2. Tattooing-rake handle10. Composite bonito hook15, Mast38. Cock's tail feather42. Ear ornament consisting of two interlocking rings44. FingernailBrown also suggests that No. 12 is "bamboo" and No. 37 is "knife,"but the resemblances in these cases strike us as elusive.The fact that the characters are so frequently of rebus type mayindicate that the inventors became preoccupied for a time with thisprinciple of representation during their development of the system,and that some of the gaps in our identifications might be filled byfurther research in the vocabulary of material culture. It is of interestalso that the phonetic value of the Type 2 character A^ (No. IX) isalso the native word for "tooth," which the form of the characterresembles, especially in the variant form given b}'- C. Perhaps it wasthis coincidence which first suggested the rebus principle.Of the other characters, some may well have been borrowed fromdecorative design elements, but we have not recognized any exceptthose already described. Most of the others are very likely the prod-uct of pure imagination. Indeed, this must needs be the case, forCarolinian dialects in these islands are extremely deficient in wordsconsisting of open monosyllables, upon which the syllabary is based,and even more so in such words which can be concretely represented.We should note that Brown's list, but none of our lists, includesseveral symbols (Nos. 14, 31, and 43) whose attributed values areshown as closed monosyllables. It is possible that he misheard thesounds: certainly the word for canoe (No. 14) is wa, not warr as Brown
No*60*]^" ^^^" CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 299
writes it, in the Central Carolines as far east as Truk; it becomes waronly beyond Truk, at Ponape. Similarly the name of the bonitohook lure (No. 43) has no final consonant as Brown gives it. Thefemale breast (No. 31) is indeed tut, as Brown has it {tuut in Smith'sorthography), just as the trigger fish (No. 29) is properly vwp, but inthe syllabary No. 31 becomes tu (tuu) and No. 29 becomes pu, in ourlists; the words for these objects seem to be compounded forms ofolder roots (Goodenough, 1953, p. 16, and personal communication),and these compounds are apparently sufficiently transparent to thenatives to allow isolation of the open syllables for the purpose of in-cluding them in the syllabary. Possibly this was done because ofpaucity of open monosyllabic words in the language. A similar proc-ess may have resulted in the discarding of final consonants of stillother words which may have been utilized for the syllabary, butwhich we have not been able to identify in such altered form asbelonging to the rebus type.ANALYSIS OF THE WRITINGPHONEMES, PHONEMIC COMBINATIONS, AND CHARACTERSWe wish now to examine the actual writing and its adequacy torepresent the language.According to Smith (1951) the Woleai language has 50 phonemes.There are 11 vowels which occur both long and short, 2 vowels oc-curring only short, the semivowels w and y, and 24 consonants. Forthese phonemes Smith has developed an orthography in roman letterswhose official adoption has been proposed to the Trust Territoryadministration; we will use it in the remainder of this study, as wehave done in the figures and tables. ^^
21 Smith's published work is nontechnical In language, and his phonemes were established by means ofminimal pairs. The values indicated below are, in Smith's words, "only a meager approximiation indeed."For this reason we are not satisfied that our transcription In his orthography of the attributed values of thevarious characters in figs. 25 and 26 is always accurate. His description of the orthography, which Is adaptedfor use without diacritical marks, is as follows:a fatherah fa-theraa fataah fa-tBe fedaeh fe-de fatheree safeeeh sa-fei seaih sea-11 sito oakoh oa-koa offcab o-ff06 hors d'oevresoeh hors d'oe-vresu bootuh boo-tuu Numberguuh NiZ-mbergV (like "oe" but with tip of tonguevh curved up and back)w woody pes
b
300 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173From Brown, Damm, Sarfert, and Someki, and from our variousinformants we have a total of 97 characters (78 of Type 1 and 19 ofType 2) which are used to write the Woleai language. All of thesecharacters, except those which represent vowels alone, have syllabicvalues of the consonant-plus-vowel or semivowel-plus-vowel pattern.Smith's phonemic analysis would indicate that there is a theoreticalpossibility of 624 such syllables (24 consonants and 2 semivowelscombined with 24 vowels) ; with the addition of the 24 vowels whenthese form independent syllables, there might be altogether 648symbols.Of course, it is most unlikely that all 624 combinations actuallyoccur in the language. We have assembled from Smith's text m.ate-
rials a list of 301 Woleai words written in his orthography, and intable 1 the frequency of occurrence of all syllables among these301 words is shown. There actually occur only 162 syllables of theconsonant-plus-vowel and semivowel-plus-vowel types, as well as13 syllables composed of independent vowels, a total of 175. It isstriking that the phonemes d and ngh do not appear to occur insyllable-initial position at all, but in these 301 words occm* only ter-minally. Some of the columns headed by other consonants (e. g.,ch, j, n, rw, ill, z) have only one or two entries under them. Morethan 175 different syllables must, of course, occur in the language.Our informants have written for us 222 of the 301 words (see table 2),have failed to use 21 of the symbols (VI, 7, 9, 24, 30, 44, 54, 55, 57,61, 63, 64, 69, 70, 71, 72, 73, 75, 76, 77, 78), and use 3 of them onlyas terminal consonants (5, 47, and 52). Presumably the attributedvalues of these symbols are not the same as those of any other symbols,so failm-e to use them would mean that the appropriate syllables donot occm- in these particular words but may very well occur in others.Even so, it is unlikely that more than, say, 250 or 300 syllables ac-tually occm* in Woleai speech. In the first 50 words (not countingrepetitions) of Smith's text there occur 62 different syllables; in thenext 50 words 29 more syllables occur; in the next 50, 18 others;in the 222 words written for us by our informants, there are alto-gether 157 different syllables; and in the total of 301 words takenfrom Smith, only 18 more or a total of 175. (These 18, which we donot have written in the native script, are: hvh, je, kii, loe, loeh, Ihii,me, mwah, rah, ree, soe, shii, waah, waeh, wo, yaa, yaah, yoe, in Smith'sorthography.) If a curve is projected on the basis of these figures,it appears to flatten out at something under 250 syllables in 750 or800 words. LENGTH OF VOWEL NOT DISTINGUISHEDWe may reduce still further the possible number of syllables in thewriting. The distinction between long and short vowels does not
Naeo?"^*^' CAROLINE SCRIPl^RIESENBERG AND KANESHIRO 301
seem to be utilized. We are not certain of the phonetic values ofsome of the symbols, but when we analyze words which informantshave written for us we are seldom confronted by any two characterswhich seem to be distinguishable only on the basis of difference invowel length; rather, when a native writes the characters for twosyllables which contain the same consonant and whose vowel suffixesare identical except for being short in the one case and long in theother, he almost always uses the same character to represent both.The following examples (see table 4), in Smith's orthography, showthis to be true (in each case h following the vowel indicates length):Syllable Character used {and number of times so used)i IV (44), III (1)ih IV (6)u XIX (11), XIX+39 (2)uh XIX (2)uu 16 (4), IV (4)uuh 16 (2)V 16 (6), IV (1)vh 16 (1), XIX+14 (1)bu 29 (5), 59 (3), 13 (1)buh 29 (1)fa 32 (40), 66 (1)fah 32 (2)ga 46 (29), 50 (4), 74 (3)gah 46 (2)gi VIII (26), VIII+III (1)gih VIII (2)la 1 (24), 6 (1), 33 (1)lah 1 (3)li X (5)Uh X (2), 36 (1)Iha 1 (3), 33 (3)Ihah 1 (4)ma 23 (9), 21 (1), 11 (2)mah 23 (7), 21 (1)maa 23 (12), 11 (1)maah 23 (7)mae 11 (24), 19 (2), 23 (1), 34 (2)maeh 11 (2), 21 (1), 23 (1)pa 43 (29), 26 (1)pah 43 (1)pi XIV (5), XIV+III (2)pih XIV (7), XIV+III (1)roe 35 (2)roeh 35 (1)sa 37 (7)sah 37 (1)see 40 (2), 40+IV (3)seeh 40+IV (3)Shi XIII (15), XIII+ III (1), XV (1)471762—60 21
302 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
Syllable Character used (and number of times so used)shih XIII (2)shv 28 (1), 28+16 (1)shvh 28 (1), 40 (1)ta 3 (18), 18 (1), 25 (2)tab 3 (13)to 17 (22)toh 17 (2)tuu 31 (5), 31+ 16 (1)tuuh 31 (1), 31+ 16 (1)ya I (50), III (1), IV+III (1)yah I (3)yoa XVI (1)yoah XVI (2), I (1), II (1)
It will be noted that, by and large, the frequency of occurrence oflong vowels is much lower than that of short vowels. Perhaps forthis reason the natives of the Woleais have found it unnecessary toinvent characters which make the distinction.Instead of 648 theoretically possible combinations, we have thenonly 338 (24 consonants and 2 semivowels combined with only 13vowels), and 13 vowels which can occm' not in combination.VOWEL CHARACTERSOf the 97 characters that we have found, a number have only vowelvalues; just how many is not certain. In table 1 it will be seen thatof the 24 (both long and short) vowel phonemes in the Woleai language,only 13 actually occur as independent syllables in the sample of 301words. Possibly if we had more than 301 words, more of the 24vowels would appear. The other 11 vowel phonemes occur only insyllabic combinations. Of this 13, 4 are long vowels and in all wordsthat informants have written for us they are represented by the samecharacters that stand for their short forms. This would leave 9vowel phonemes for which characters would seem to be needed.Now among our total of 97 characters there are some 13 which mayor may not be intended to stand for vowels: I, II, III, IV, V, VI,XII, XVI, XIX, 14, 16, 39, and 72. (Possibly No. 74 should beincluded among them.) In the 222 words wi'itten by our informants(see table 2), only 3 of these 13 characters never represent semivowel-plus-vowel syllables (although they may have semivowel values whenused terminally). Thus character IV usually represents the phonemei, and also occurs as ee, ih, ih-\-a, ih-\-y, uu, v, and -y; it never appearsin a written word in the form semivowel-plus-vowel. Similarly XIXis always u, uh, or -w; and 16 is always uu, uuh, v, vh, i-\--w, or -y.Four others of the 13 characters may occur either as vowels alone orwith semivowel-plus-vowel values: thus character I can be a, but also
No.*6o']^" ^'^^' CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 303
ya, yah, yae, or yoah; III is ae, ee, or i, but also ya, yae, ye and -?/; XVIis oa, but also yoa, yoah, and yoeh; and 39 is o once, otherwise wa,waa, wae, we, woe, -w. Four other characters of these 13 appear onlywith semivowel-plus-vowel values: II is ijoah in its lone occurrence;V is we, woa, or yoeh; XII occurs in our examples only as wih; and 14is wa, wae, or woa. We have no examples of words containing char-acters VI or 72. (See tables 2 and 3.)It is apparent that there is considerable variation in vowel valuein these characters, as there must needs be if they are to represent all13 (or 24, if we consider length) vowel phonemes when these phonemesdo not enter into syllabic combinations with consonants and semi-vowels. If we eliminate the 4 characters among the 13 that seem tooccur only with prefixed semivowel values and if we disregard char-acters VI and 72, which do not occur in oiu- 222 wi'itten words, we haveleft just 7 characters which may be used to represent the 9 (or 13, ifwe consider length) vowel phonemes which actually occur as inde-pendent syllables in the 301 words taken from Smith (table 1).It is likely that additional samples of writing from more informantswould clarify matters; it is possible that all 13 of these characters maybe read indifferently as simple vowels or with initial w- or y-. But fromthe evidence at hand, if we do not consider the three characters (IV,XIX, and 16) which do not seem ever to have the value of semivowel-plus-vowel, and ignoring for the time being variation in vowel valueof the others among the 13 characters, there remain 94 charactersout of the total of 97 to represent all of the possible syllabic combina-tions. It is obvious that the system is inadequate for truly phonemicrepresentation, and that many of the 94 characters must serve forseveral combinations.LACK OF EXACT CORRESPONDENCE OF CHARACTERSAND SYLLABLESThis conclusion is further borne out when we examine the range invalue of the various characters as they are actually used in the writing.We have seen how Avide this range is for the 13 characters we have justdiscussed. Table 3 lists all the syllables occurring among the 222words which we have been able to obtain in the native script, ar-ranged according to the conventional numerical sequence of figures25 and 26. We see that character 4, for example, is used for at least6 different syllables
—
ba, baah, bae, be, bee, and boe; character 33 isused for 11
—
la, lae, le, lee. Hi, Iha, Ihae, Ihe, Ihi, nhae, nhe; character50 serves for the syllables ga, gae, ge, ka, ke, and kee; etc.But not only must one symbol serve for several different syllables.It is apparent that the same syllable may be represented by different
304 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
symbols. Table 4, which lists in alphabetical order all the syllableswhich occur in the 222 words, illustrates this point; for example, thesyllable Ihi is written by one native with character 33 three times,character X 25 times, X+33 once, and X+IV once, and by a secondnative with X three times and 33+X once; similarly pu is writtenwith characters 10, 13, or 59; sha with 18, 20, or 37; etc.These two phenomena—the use of one symbol for several syllables,and the representation of one syllable by several symbols—explainin part the facts that the name or attributed value of a symbol, asseen in the various lists in figures 25 and 2G, sometimes differs, andthat in the writing of words the attributed value occurs only part ofthe time as the actual phonemic shape of the syllable the symbol isintended to represent. It would appear that when it is necessary torepresent a sound which differs from the attributed value of anycharacter, a choice may be made among two or more characters whoseattributed values are close phonologically. For example, we havenot recorded tv as the attributed value of any one character (it ispossible, of course, that we have misheard some of the tape-recordedvalues), but in writing the word f&ltvfih. one native uses 32/31/33and a second uses 32/XVI1/33; symbol 31 is otherwise used by thefirst man for tu and tuu, and XVII by the second man for ti and tii.It would seem that each man has chosen, to represent the syllable tv,for which neither seems to know a character which has this as itsattributed value, a character with an attributed vowel value on eitherside of it in articulation.EFFECT OF DIALECTSSome of the differences in choice of character by different nativesmay be due to dialect difference, but we have no way of assessing itseffect, since such differences as may exist have not been determined.The only statements that Smith (1951) makes on this point are asfoUows (p. 40): "Islanders . . . may have difficulty in deciding on 'a'or *aa' depending on their island of origin. These are separate pho-nemes on Weleeya (Woleai) and Foeshavlap (Faraulep) but perhapsnot on Yaurwpiig (Eauripik). This is of course independent of thefact that some words are pronounced slightly differently on differentislands"; and (p. 27): "On Eauripik 'tafeey' is pronounced 'tafee' and
'maaht' is pronounced 'maeht.' " (R. of Ifaluk apparently hearsta/fee/y, not ta/fee, for he writes the word with three characters:3/66/III; and he distinguishes maah/t, which he writes 23/25 or 23/17,from maeh/t, 21/40.) The occasional use by various informants of two,or even three, characters for what, in Smith's orthography, should bea single syllable, would also suggest dialect differences not recordedby Smith; thus, in the following examples of syllables written by
No.*6o1^" ^^^' CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 305Smith as consonant-plus-ee or consonant-plus-eeA, the informantswrite an extra character, namely IV, apparently for an additionalvowel
:
6ee/yae R.: 4+ IV/Ulkee/i R.: SO+IV/llIli/mee/g R.: X/ll+ IV/50rwee/g R.: 47-\-S9+IV/50see/g R., C, and M.: 40+IV/50seeh/gi R., C, and M.: 40+/F/VIIIREPRESENTATION OF FINAL CONSONANTSThe choice of a character to represent a final consonant seems tobe much more capricious than the choice of a character to represent afull syllable. We have already pointed out that the attributed valueof a character loses it vowel portion when the character is used for theterminal consonant of a word. In table 4 it will be seen how wide achoice of character seems to be permissible for such use. Neverthe-less, some regularities are discernible. For example, 18 words termi-nating in -g are listed in that table, some of them written severaltimes and by as many as three natives ; for the -g the informants havewithout exception used character VIII in four words, in four otherwords they agree on character 2, in another jfive words they all usecharacter 41, and they invariably write character 50 in five more.How are we to explain such consistencies in usage, sporadic thoughthey are? In some Malayo-Polynesian languages there occurs a finalvowel that is so weakly pronounced as to seem inaudible. While thisphenomenon does not appear to be a feature of Woleai speech today,it may well have been so in the past; and if we may credit the in-ventors of the Type 1 script with ability to reconstruct the ancientpu from modern pup and tuu from tuut, as previously described, it ispossible that in these instances the character chosen to represent -gis the one whose full syllabic value includes the vowel sound whichwas formerly terminal.In the use of final characters in some words there is a hint that per-haps the informants are anticipating a vowel which would appear onlywhen additional syllables were suffixed to the same words. Thus wehave character 33 for -Ih in the word mwae/ZA., and also for Ihe in thecompound mwae/ZAe/we; 33 is used again in mi/si/lh and in the com-pound mi/si/lhae/lh] but in the word rae/lh, informants use for -Iheither 33 or X, as though this word might be compounded in differentways, and in the one compound form which we have in native script,Tae/lhi, the syllable Ihi is written with the character X.Occasionally the vowel preceding the final consonant seems to in-fluence the choice of character for that consonant. Thus -I is almostalways written with character 33, occasionally with X; but character
306 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
36, which is used for the syllable luu in the word F^/la/luu/s and forIv in the word fae/Zv/w, also represents -I in the words vh/Z (as writtenby one informant out of two) and uuh/Z (by both of two informants)
;
in other words, 36 seems to be the choice in these instances becausethe vowels of the preceding syllables are similar to the vowel portionof that character's full syllabic value. Again, character 2, whichusually has the value go (see table 3), is used terminally as -g only inthe words yae/lo/^, So/w/g, to/g, and toh/g (but not in sho/g) , that is,in words where the preceding vowel is o. And the only times charac-ter 31 is used terminally are in the words tuu/t and tuuh/^, character31 otherwise having the value tu, tuu, tuuh, or tv; in practically allother cases of -t informants agree in the use of character 25.However, the principle which seems to develop from the foregoingdoes not seem to apply to many of the other terminal consonants. Insome cases it would seem that it is the consonant, not the vowel, ofthe preceding syllable that determines which character is to be usedfor the final consonant of a word. Thus all informants use character6 for the -n of bu/ri and -nh of hu/nh, but use 33 or 36 for the -nh ofpu/nA. Again, both C. and R. agree on character 12 for -ngh intah/ngh, but 52 for -ngh in Ihah/ngh; all informants use IX for -ng in
ch.Q,/ng, \ah/ng, and tah/ng, and all of them use 52 for -ng in Ihah/ngand u/ng; they agree again on XIII for the -c in wih/c and on 5 for the
-c in mwoa/c-mwoa/c ; and there are further cases of such unanimityin choice of character to be used for final consonants. Perhaps thereis some component of the phoneme which determines such seeminglyarbitrary choice.There is also some evidence, although not so strong, that sometimesthe reverse occurs; that is, that the final consonant may determinewhich character is to be used for the preceding syllable. For example,R. and C. use character 47 consistently for final r or ?^w. In the wordssorw, raesorw, galiisorw, and josoar the syllable preceding -r and -rw iswritten with character 68. The only other time 68 is used is for thesyllable sho in the word jo/sho/s; whereas for sho in the word sho/gcharacter 45 is used. How is this to be explained? It would seemthat R. has misunderstood the word joshos and taken the final con-sonant, -s, as -r or -rw, since he writes 47 for it, instead of XV or 62as he does for all other cases of -s. It therefore suggests itself thatthere is some kind of relationship between these two characters, 47and 68, and that the use of the former requires the use of the latter.Whether such relationships are of a functional nature, whether thesystem of spelling has in the course of years of use become arbitrarilyconventionalized, or whether some aesthetic judgment is applied whichprevents or encourages the juxtaposition of certain characters, is amatter that we are unable to settle. In many cases, perhaps in most
No.*6o7'^^^" CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 307
cases, there seems to be no clear rule; rather it may be whim whichoften dictates which of several final characters is to be used. Never-theless, examination of table 4 gives the impression that some rulesare operative. COMPARISON OF NATIVE TEXTSThe foregoing analysis is, of course, based entirely on examinationof the 222 words in Smith's orthography that we have been able toget natives to write in the Caroline script. Smith's text, which isaddressed to the natives and which consists of a guide to his proposedsystem of spelling, is naturally limited by his vocabulary and mode ofexpression. (See fig. 27.) But we have available to us two othertexts, of native composition, by R. and S.,^^ both of Ifaluk (fig. 28).They have not been tape-recorded or put into Smith's orthography,but they will enable us to discuss frequency of occurrence of thesymbols. S. uses 68 characters out of the 97, a total of 1,178 times;R. uses 65, a total of 483 times. Of the symbols (fig. 25, cols.N and O; fig. 26, cols. N-P) given by R. separately from the text, 10 donot occur in the text; and occm-ring in the text but not in the lists are6 symbols. In other words, he knows 75 of the 97 characters, but inwriting he makes use of only 65, of which 3 are the apparently simplevowels (IV, XIX, 16), so that he limits himself to 62 charactersrepresenting syllabic combinations for the 624 possible combinations.Of these, 11 are used only once, 8 only twice, and 10 only three times.We have no comparable hst of symbols and attributed values fromS. with which to compare his text. He does not use in the text 29 ofthe symbols known to us; 22 of these 29 are also absent from R.'stext, and 10 are among those apparently unknown to R. Of the 65which he uses, 10 occur only once, 4 twice, and 7 three times.Percentagewise, some comparisons between the two men are ofinterest, assuming that we have adequate samples of this universe,speaking statistically. Character 7, not used at all by R., occui-swith a frequency of 4.4 percent of the 1,178 characters written by S.Character 46 occurs in R.'s text with a frequency of 0.2 percent, butS. uses it 28 times as often, with a frequency of 5.6 percent. Therespective figures for R. and S. for character 35 are 2.1 percent and6.0 percent; for 50, 6.0 percent and 3.4 percent; for 39, 3.9 percentand 1.9 percent. There is less significance in the difference in occur-rence of other characters.Evidently, then, the system is flexible enough to permit of consid-erable freedom in expression of preference through employment ofone character or another. In order of descending frequency, the
« S. Is the man described by Spiro (1950) as a psychopath. He Is kept confined at Ifaluk in a coconut loghut. Whether this fact diminishes the value of the comparison between his writing and that of R. we arenot prepared to state.
308 BUREAU OF AMERICAN ETHNOLOGY [BuU. 178
^ 'v ^ i>^ ^ a /^ X c? X
Figure 27.—Paragraph from Smith's roman text (1951, pp. 3-4) as written in native scriptby R. of Ifaluk. (Words are shown separated by spaces for purposes of clarity; in theoriginal they run continuously.)(For explanation, see opposite page.)
preferences of the two men for their most frequently used charactersare: R. S.Percent PercentX 7.0 IV 8.050 6.0 X 6.6III 5.8 35 6.0IV 5.4 46 5.625 4.7 7 4.433 4.7 33 4 2In a smgle person's writing, however, there is considerable con-sistency. Part of R.'s text is written twice. The duplicated portioncontains 187 characters in both versions, which are identical exceptfor the following changes in spelling in the second version: substitutionof IV for 16 in one word; 37 for 40 in another word; 38 for 17 in atliird ; 6 for 33 ; the sequence III-50 for IV-46 ; and the sequence 6-52for 1-2 in still other words. We are probably dealing here again withcombinations of phonemes whose values lie within a range of overlapbetween values associated with two different characters, so that onechoice is as good as the other.
nS*W"^''^' CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 309EXPLANATION OP FIGURE 27
Characters used in figure 27 shown by numbers assigned to them (refer to figures 25and 26). Text and translation as given by Smith {1951, pp. S-/f, S3).IV 33 VIII 33 46 26 3 33 XV 4 32 17 VIII 45 41i /Iha/ gi/ 1 ka/pe/ta/ 1 si /bae fa/ to/ gi sho/ g4 III 4 39 39 46 26 3 33 VIII 20bae yae/be wae-wae ka/pe/ta/
1
gi /sb39 39 ? 2 42 X 39 33 I 35 X 11 33 3 IVwae-wae/lh go/zv/lhi we/lee/ya rae/lh mae/lae ta/ i46 26 3 X 39 33 I 40 23 36 III 4 17 XIX 1 26ka/pe/ta/
1
we/lee/ya si /mii/ Ih yae/be to/ u /la/ p41 1 I 20 4 III 4 23 20 15 VIII VIII 20gv/la ya/sh bae yae/be ma/sha/ra/ g gi /sh35 X 50 35 VIII 35 VIII X 46 26 3 X IV IX 33 XVrae/lh ge/ra/ gi /roe/ gi /Ihi ka/pe/ta/
1
i /nge/lii/ s4 XV 4 41 1 4 III 14 IV 35 XVbae si /bae/gv/la bae yae/wa/ v /rii/ sIV III 11 33 IV XIII 33 46 26 3 X IV IX 33 XVih / y mae/lhae i / shi / Ih ka/pe/ta/ 1 i /nge/ lii/ sIV 39 12 IV III 11 33 2 42 ? 39 33 I 12i /woe/ngoe ih/ y mae/lhae go/zv/lh we/lee/ya ngeIII 23 20 15 VIII 35 20yae ma/sha/ra/ g roeh/sh
Ilhagil kapetal sibae fatogi, shog bae yaebe wae-wae kapetal. Gish wae-waelhAll words written in this way are WTitten just as they sound. We spell Woleaigozvlhi Weleeya, raelh maelae tai kapetal Weleeya. Simiilh yaebe toulap gvlasounds, not Woleai words. Therefore Woleai has ayash, bae yaebe masharag gish, raelh geragiroegilhi kapetal Ingeliis bae sibaegvla.much better and easier way of spelling than English has.Bae yaewavriis ihy maelhae ishilh kapetal Ingeliis, iwoengoe ihy maelhae gozvlhIt is difficult to spell the words of English but it is easy to spell the soundsWeleeya, nge yae masharag roehsh.of Woleai. SUMMARY
In 1905 an American missionary from Trulc, Alfred Snelling, and aparty of Trukese were cast ashore at Eauripik. Here they gaveinstruction in writing with the alphabet which had been developed inthe Nomoi Islands. The natives of Eauripik took the names of theletters as being their syllabic values and converted the alphabet intoa syllabary. These letters constitute the symbols we have called
310 BUREAU OF AMERICAN ETHNOLOGY [Bull. 178
A^mo^f ^^h.tr k cf ^ g f ic^ EC r isu V' 7!/' A/ \Q/ vp r c r r-ari r/3),3^ g
Figure 28.—A page of text from a song written by R., a man of Ifaluk.
Type 2. The syllabary diffused to Faraulep where the deficiencies ofthe writing became apparent, all consonant signs of the originalalphabet now having syllabic values consisting only of consonants-plus-i. The Faraulepese, between 1907 and 1909, invented a wholenew set of symbols, Type 1, taking some of the signs from theirenvironment and their material culture and giving them as theirvalues the names of these objects; other signs were made by alteringthe form of Type 2 symbols; a few may be of Japanese derivation;and some are the products of imagination. By 1909 the writing, ofboth types, had spread to eight atoUs of the Central Carolines and itis still known on five of them today.
NaeoT''^'*^' CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 311
Previous authors have speculated that the writing represented theremains of a formerl}^ more developed system, that it was related toscripts of the Asiatic mainland, that it was linked to Easter Islandwriting, etc. But it has been demonstrated that the Woleai syllabaryrepresents a case of recent stimulus diffusion, like the Vai and Cherokeesyllabaries.The writing, v/hich is still being added to by new inventors fromtime to time, represents only crudely the language it is used for. Asymbol may be used for more than one syllable, and a syllable may berepresented by more than one symbol. In time, more exact corres-pondence might develop. However, the writing will probably dieout before this occurs.LITERATURE CITEDAmerican Board of Commissioners for Foreign Missions.1886-1906. Annual reports. Boston.Arago, J.1822. Promenade autour du monde pendant les annfees 1817/1820 sur lesCorvettes du Roi I'Uranie et la Physicienne commandees par M.Freycinet. Vol. 2. Paris.Bliss, Theodora Crosby.1906. Micronesia. Fifty years in the island world. Boston.Brown, J. Macmillan.1914. A new Pacific Ocean script. Man, vol. 14, No. 43, pp. 89-91.1927. Peoples and problems of the Pacific. Vol. 1. London.Burrows, Edwin G., and Spiro, Melford E.1953. An atoll culture. New Haven.CONKLIN, H. C.1953. Hanun6o-English vocabulary. Univ. California Publ. Ling., vol. 9,pp. 1-290.Damm, H.1938. Zentralkarolinen. G. Thilenius, ed. Ergebnisse der Slidsee-Expedi-tion, 2 B, vol. 10, pt. 2.Damm, H., and Sarfert, E.1935. Inseln um Truk. G. Thilenius, ed., Ergebnisse der Siidsee-Expedi-tion, 2 B, vol. 6, pt. 2.DiRiNGER, David.1948. The alphabet. New York.FiNSCH, Otto.1900. Carolinen und Marianen. Sammlung gemeinverstandlicher wissen-schaftlicher Vortrager, vol. 14, pp. 651-710.Gardner, Fletcher.1943. Philippine Indie studies. Indie Bull. No. 1, Ser. of 1943. San An-tonio.GiRscHNER, Max.1912-13. Die Karolineninsel und ihre Bewohner. Baessler-Archiv, vol. 2,pp. 123-215; vol. 3, pp. 165-190.GOODENOCGH, WaRD H.1953. Native astronomy in the Central Carolines. University Museum,Univ. Pennsylvania, Philadelphia.
312 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
HoRNELL, James.1936. The canoes of Polynesia, Fiji, and Micronesia. Bernice P.. BishopMuseum, Spec. Publ. 27, vol. 2.Imbelloni, J.1951. Las "tabletes parlantes" de Pascua, monumentos de un sistemagrdfico indo-ocednico. Runa, vol. 4, Nos. 1-2, pp. 89-177.I^TTLITZ, F. H. VON.1858. Denkwiirdigkeiten einer Reise nach dem russischen Amerika, nachMikronesien und durch Kamtschatka. 2 vols. Gotha.Kramer, A.1937. Zentralkarolinen. G. Thilenius, ed. Ergebnisse der Siidsee-Expedi-tion, 2 B. vol. 10, pt. 1.Lessa, William A.1950. Ulithi and the outer native world. Amer. Anthrop., vol. 52, pp.27-52.LtJTKE, F.1835. Voyage autour du monde ex6cut6 par ordre de sa Majeste I'EmpereurNicolas ler. Paris.Mason, W. A.1920. A history of the art of writing. New York.Meinicke, Carl E.1876. Die Inseln des Stillen Oceans. Leipzig.M^TRAtrx, Alfred.1957. Easter Island. New York.Sarfert, E., see Damm, H., and Sarfert, E.Smith, Alfred G.1951. Gamwoelhaelhi ishilh Weleeya (guide to Woleai spelling). HighCommissioner, Trust Territory of the Pacific Islands. Honolulu.(Mimeograph.)Someki, Atsushi.1936. Karorin shoto no ki-sara ni tsuite (the wooden dishes used in theCaroline Islands). Japanese Journ. Ethnol., vol. 2, No. 2, pp.172-190.1945. Micuronejiya no fudo to mingu (topography and implements ofMicronesia). Tokyo.Spiro, Melford E.1950. A psychotic personality in the South Seas. Psychiatry, vol. 13, pp.189-204.Spiro, Melford E., see Burrows, Edwin G., and Spiro, Melfobd E.
Anthrop. Pap.No, 60] CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 313
CO
^ o
CO
S
O
oS
CO
a —
S5s
3goEh
^
314 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173Table 2.
—
Words written by Caroline Island natives in Caroline script[On left, words from Smith, In his orthography; on right, spelling of those words in native characters.Characters are shown by numbers assigned them in figures 25 and 26. Numbers in parentheses indicatenumber of times word is thus written, if more than once, by each informant. Double diagonals followingthe number of a character indicate that the character is apparently used for two successive syllables. In-formant designations: R.= Maroligar; 0.=ChiyemaI; M.=Magilo; T.=Tachep; A.=Marutang.]ae/ta - R.: 111/3ae/tae/i R.: III/3/IVae/ta/1 R.: III/S/IVae/ta/i/fiiAh — - R.: III/25/IV/VII/33ae/to/u/la/p R.: III/17/XIX/l/26aAe/fah/pae/t - R.: 1/33/32/66/25 (2)baah/sh R.: 4/20; O.: 4/20; M.: 4/20ba/bl/o/rw R.: 4/29/XVI/22bae R.: 4 (21); C: 4bae/s -- --- R.: 4/XV; C: 4/XVbae/sh R.: 4/XIII, 4/40; 0.: 4/XIII; M.: 4/XIIIbae/y R.: 4/111bee/yae R.: 4+IV/niboa/d R.: 13/XVII; 0.: 13/XVII; M.: 13/XVIIboe/ga/re R.: 4/46/35buh/k — - R.: 29/2bu/n -- R.: 29/6; O.: 29/6; M.: 29/6bu/nh - R.: 29/6; C: 29/6ca/ng R.: lR/12, I8/IX; O.: 18/12; M.: 18/12ca/ngh.... R.: 37/52cha/ng R.: 18/IX; O.: 18/IX; M.: 18/IXee/bae R.: III/4ee/tae R.: 111/25ee/tae/mwaAh R.: III/3/19/33ee/ta/1 R.: IV/3/33faa/t-faa/t R.: 32/25-32/25; C: 32/25-32/25Fae/la/la/p.. R.: 66/1/1/26fae/lv/w R.: 32/36/39, 66/36/39Fa/go/sa/p R.: 32/56-I-IV/37/26Fa/la/luu/s R.: 66/1/36/62faAh - R.: 32/X, 32/33 (2)Fa/ra/gi/ye —
-
R.: 32/35/VIII/IIIfa/th-fa/th R.: 32/17-32/17 (2); O.: 32/17-32/17fa/th-fa/thvAM- R.: 32/17-32/17/X f4); C: 32/17-32/17/Xfa/tilAh C: 32/XVII/.33; M.: 32/XVII/33fa/ti/lh R.: 32/XVII/X; C: 32/XVn/X; M.: 32/XVII/Xfa/to/gl R.: 32/17/VIII (7); C: 32/17/VIIIfa/tvAh R.: 32/31/33: C: 32/XVII/33fl/to/u R.: VII/17/XIX+39 (2)foa/ri R.: 27/XVIII (3); C: 53/XVIII; M.: 27/XVIIIFoe/sha/v/la/p R.: 32/20/16/1/26, 66/18/16/1/26; O.: 66/18/16/1/26; M.: 66/18/16/1/28; T.:66/18/16/1/26 (2)ga/bu/ngh --- R.: 46/13/IX, 46/59/58ga/ch R.: 46/28ga/fl/tae/g R.: 46/VII/25/VIII (6); O.: 74/VII/25/VIIIgah/th R.: 46/17; C: 46/17ga/i/ga/bu/ngv - R.: 46/IV/50/59/58 (2)pa/1/ma/ngI/y R.: 46/IV/ll/IX/III, 2//23/IX/IIIQa/la/ya/li/ma/ng R.: 46/33/I/X/ll/IXgaAii/so/rw R.: 46/33/68/47 (2)ga/mwoe/lha — - R.: 46/19/33ga/mwoeAhaeAhi R.: 46/19/33/X (3), 46/19/3.V33, 46/19/33/X+33; O.: 74/19/33/Xca/re R.: 46/35, 50/35; C: 74/omlUedga/wae/waa/y R.: 46/39/39/III (2); C: 50/14-|-IV/39/IVge/pa/t R.: 50/43/25; C: 50/4.3/25; M.: 50/43/25ge/ra/gi/roe/giAhi R.: 50/35/VIII/35/VIII/X|i/sh -— - R.: VIII/20(5);C.: VIII/20goa/mw R.: 2/19 O.: 2/19go/mw - R.: 2/34; O.: 2/49go/sh R.: 2/5; O.: 2/5; M.: 2/5go/zvA — - R-: 2/42/33go/zvAh R.: 2/42/omittedgo/zvAhi. R.: 2/42/X (5), 2/42/33; C: VIII/omltted/33+XgvAa.— - — - R.: 41/1; C: 74/1I/faeAhuu/g R.: IV/66/36/41; C: IV/66/36/41i/gaeAha — R.: lV/50/1 (3)ih/a/j R.: IV//46Ih/mw - R.: IV/19; O.r IV/19Ih/r R.: IV/35ih/sh... - R.: IV/XVIh/y R.: IV/III (2), TV// (3)lAe/gih/r -. R.: IV/33/VIII/35; C: IV/3.3/VITI/35lAha/giA R.: IV/33/VIII/33; C: IV/33/Vin/XI/ngeAii/s R.: IV/IX/33/XV (2)1/shl.-.. R.: IV/Xni (4), IV/XIIH-III; C: IV/XIII1/shlAh - R.: IV/XIII/33 (6), IV/XV/33, IV/XIII/1; 0.: IV/XIII/X1/shiAhl R.: IV/XIII/33, IV/XIII/X+IV1/taeAh - R.: IV/3/331/taeAhl. R.: IV/3/X1/ti/paeAM - — R.: IV/XVII/26/X
Anthrop.Pap. CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 315JNO. blljTable 2.
—
Words written by Caroline Island natives in Caroline script—Continuedl/woe/ngoe R.: IV/39/12I/ya/ng R.: IV/I/omittedjaAhae — R.: 46/33jo-jo R.: 2-2 (3); C: 2-2jo/sha R.: 2/37; C: 2/18jo/sho/s R.: 2/68/47jo/soa/r - — - R.: 2/68/47ka - R.: 46(5)ka/lh - R.: 50/1ka/pa/tee/yae - - R.: 50/43/25/IIIka/pe/ta/i R.: 46/26/3/33 (6), 46/26/3/X (3); C: 46/26/3/33, 46/26/3/X, 50/43/25/33ka/pe/ta/lhi R.: 46/26/3/Xkee/i R.: 50+IV/III; C: 46//ke/pa/t R.: 50/43/25 (14); C: 50/43/25; M.: 50/43/25ke/pa/te/ka/y R.: 46/26/3/46/IV-l-IIIlae - R.: 33la/go/sha/g R.: 1/2/18/41; C: 1/2/18/41lah/ng. R.; 1/lX; O.: 1/IX; M.: 1/IXIhae/n R.: 1/XIhae/nh R.: 1/X (2)IhaiJ/ng R.: 1/52; O.: 1/52Ihab/ngh R.: 1/52; C: 1/52Ihl... R.: X (2)lih/mw -- R.: X/19; C: X/11lih/wa/nhae/y R.: 36/39/33/IVU/mee/g —. R.: X/ll+IV/50maah/lh - R.: 23/36; C: 23/36; M.: 23/36maah/t R.r 23/25,23/17; C: 23/25; M.: 23/25maa/lbl R.: 23/Xmaa/sv/r R.: 23/62/47; C: 23/62/47maa/t — R.: 23/25; C: 23/25; M.: 23/25maa/zv/r R.: 23/62/47; C: 23/62/47mae -- R.: 11 (9), 34 (2); C: 11; M.: 11maeh R.: 11; C: 11; M.: 23.maeh/t — R.: 21/40mae/lae. - R.: 11/33 (7); 0.: 11/1-|-1IImae/lae/1 - R.: 11/33/33mae/lhae. R.: 11/33 (2)mae/lhae/1 R.: 19/33/33(2)mae/taa/lh R.: 11/3/33(2)mah.. - - -- R.: 23; C: 23; M.: 23mah/ch R.: 23/20; M.: 21/28mah/lh... R.: 23/36; C: 23/36; M.: 23/36ma/ngi/y - R.: 23/IX/IIIma/sha/ra/g R.: 23/20/15/VIlI (3); C: 23/20/15/VIIIma/tae-mae/tae/1 R.: 23/25-23/25/Xma/th R.: 23/17; C: 23/17mi/lh.. R.: XI/33; O.: XI/33mi/si/lh R.: XI/XV/33mi/si/lhae/lh R.: XI/XV/33/33mwae/lh... R.: 19/33mwae/lhe/we R.: 11/33/39mwa/lhi.. R.: 21/Xmwl/lh R.: 67/33; C: XI/33mwoa/c-mwoa/c R.: 34/.5-34/5; C: 49/5-49/5rnwoe/lae/lh - R.: 19/33/33Na/mwoa/chU/g R.: 1/34/20/VIII; M.: 1/49/20/VIIIngaAhi R.: 48/Xnga/li R.: 48/X (2)nge R.: 12(7)nbe/pa/1 R.: 33/43/33 (4), 33/43/Xpaah/l_. R.: 43/IV; C: 43/IV; M.: 43/rVpah/nga/sh R.: 43/48/20pa/nga/lh — . R.: 43/48/33(3)Plh/g C: XIV/50;M.: XIV/50plh/pih - R.: XlV/XIV-t-III; C: XIV/XIV; M.: XIV/XIVpi/pL R.: XIV/XIV+III; O.: XIV/XIV; M.: XIV/XIV+IIIPu/lo/wa/th R.: 13/6/14/25pu/nh --. R.: 10/33; 0.: 59/36raa/sh R.: 35/20rae/lh R.: 35/X (5), 35/33 (7), 35/33+X; 0.: 35/omittedrae/lhi - R.: 35/X (2); 0.: 35/Xrae/so/rw... R.: 35/68/47ra/ra/lh R.: 15/15/33re— -- R.; 35(5)re/bae -- R.: 35/4 (4); C: 35/4 (4)roeh/sb R.: 35/20roe/sa/fa/to/gi R.: 35/37/32/17/Vinro-ro. - R.: 22-22; C: 22-22ro/w-ro/w - - R.: 15/XIX-15/XIXrwee/g R.: 47+39+IV/50 (2)Sae/pa/lh. R.: 37/43/33sah/g -- R.r 37/41
316 BUREAU OF AMERICAN ETHNOLOGY [BnU. 173Table 2.
—
Words written by Caroline Island natives in Caroline script—ContinuedSa/taa/wa/l R.: 37/25/14/33 (2); O.: 37/25/14/33; M.: 37/25/14/33; A.: 37/25/14/33see/a/w R.: 40//39 (6)see/g. - - R.: 40+IV/50; C: 40+1V/50; M.: 40+IV/50seeh/gi R.: 40+IV/VIII; C: 40+IV/VIII; M.: 40+lV/VIIIsee/mwoe/y R.: 40/34/16(2)see/o/w R.: 40//39se/mwoe/y - - R.: 40/34/16(2)sha/g --- R.: 18/41 (3), 37/41shib/m R.: XIII/19; C: XIII/19sho/g - R.: 45/41 (5); C: XV/omittedshv - R.: 28; C: 28+16shvh/w R.: 40/XIX; C: 28/39si R.: XV (2)si/a R.: XV/I (3)sl/bae R.: XV/4 (8); O.: XV/4si/t)a«/gv/la R.: XV/4/41/1 (3)si/boe R.: XV/4si/fa/to/gi R.: XV/32/17/VIII (2)si/ga/1 R.: XV/46/33si/mii/lh R.: 40/23/36; C: XV/XI/33si/tl/pae/lhi. R.: XV/XVII/26/Xsi/ya R.: XV/I (3)so/rw R.: 68/47 (2)So/w/g R.: 42/XIX/2ta/fee R.: 3/66+IIIta/fee/y — R.: 3/66/IIItah/d -— R.: 3/XVII; C: 3/XVII; M.: 3/XVIItah/1 R.: 3/33; C: 3/33tah/lh R.: 3/X; C: 3/Xtah/ng R.: 3/IX; C: 3/IXtah/ngh - R.: 3/12; C: 3/12tah/t R.: 3/25; C: 3/25ta/i R.: 3/IV; C: 3//ta/pi/lh R.: 18/XIV/omlttedTe/o/ma/L. R.: XVII/39/21/33ti/11/gi/ae/lh R.: XVII/X/VIII/III/Xti/rl/gi R.: XVII/X/VIII+IIItoe/toe... R.: 38/38; C: 38/38to/g R.: 17/2; C: 17/2toh/g R.: 17/2; C: 17/2to/u/la/p .- R.: 17/XIX/1/26 (3); C: 17/XIX/1/26, 60//1/26tu/tu R.: 31/31; C: 51/51tuuh/t R.: 31+16/31; C: 31/31tuu/t R.: 31/31; C: 31/31tuu/tuu... R.: 31/31; C: 31+16/31uh/ng R.: XIX/52; C: XIX/12u/lu/nhv/lhi R.: XIX/65/65/Xu/ng... R.: XrX/S2; C: XIX/52uuh/1 R.: 16/36; C: 16/36uu/r R.: 16/35; C: 16/35uu/rw R.: 16/47; C: 16/35uu/we.... R.: IV/39 (4)vh/1 R.: XIX+14/33; C: 16/36wae-wae... R.: 39-39; C: 14-omittedwae-wae/lh R.: 39-39/omltted; C: 39-39/XWe/lee/ya R.: 39/33/1 (14); C: 39/33/1 (6); M.: V/33+I/IV+IIIwih/c R.: XII/XIII; C: XII/XIIIwoa/m R.: V/X (2)woa/lo/w R.: V/6+XIX/39, 14/65/39Woe/tte/ga/i/w R.: 39/25/46/16//yae R.: Ill (2)yae/be R.: III/4 (10), 1/4; C: III/4 (2)yae/be/ga/ch R.: III/4/46/28yae/fa/th R.: III/32/17yae/ga/fl/tae/g R.: III/46/VU/25/VIIIyae/1 R.: III/33yae/lo/g R.: III/6/2yae/mwoe/lh R.: III/19/33yae/tae/tae/I R.: III/25/3/IVyae/to/u/la/p R.: I/17/XIX/1/26yae/wa/1 R.: III/14/33yae/wa/v/rii/s R.: III/14/IV/35/XVya/f.-- R.: I/VII; C: I/VII; M.: I/VIIyah/f R.: 1/66; C: 1/66; M.: 1/66ya/r. R.: 1/35ya/re/maa/t R.: 1/35/23/25; C: 1/35/23/25 (2)ya/re/maa/taa/I.. R.: I/35/23/3/X; C: I/35/11/3/Xya/sh R.: 1/20 (12), 1/40ya/th R.: 1/17ya/to/we R.: 1/17/39Ya/u/rw/pll/g.... R.: I/XIX/8/XIV/50, III/XIX/8/XIV/50yoah... R.: XVI; C: IIyoah/la/gU/lh R.: XVI/1/VIII/X; 0.: I/6/omitted/Xyoa/rw R.: XVI/22yoeh R.: XVI; O.: V
Anthrop. Pap.No. 60] CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 317Table 3.
—
Characters and their actual syllabic values, arranged according to sequenceof characters in figures 25 and 26[Numbers in parentheses following a word indicate number of times, when more than once, the Italicizedsyllable within that word is written by each informant with the designated character. Plus signs: in thefirst column plus signs indicate when two characters are used for a single syllable; in the second columnthey indicate when two syllables are written with one character. Informant designations: R. = Marollgar;C.=Chlyemal; M. = Magllo; T. = Tachep; A. = Marutang.]
Character No. Syllablesrepre-sented bycharacter Inform-ant Words In which character Is used
n..ni.
IV.
IV-I-IIIV
VIVII..-
vin.
Vlll-flllIX
aya
yah
yaeyoahyoahaeeeiyayae
ye
-yeei
ih
ih-l-aih+yuuV
-yya
-ywewoayoehnonefi
fll
-f
gi
gihgUgo
-g
gingengi
-ng
-ngh
R.R.C.M.R.C.M.R.O.C.R.R.R.R.R.
C.R.R.R.R.
O.M.R.O.R.R.R.R.R.C.M.R.M.R.C.R.C.R.R.O.M.R.O.M.R.O.R.O.R.C.M.R.R.R.R.O.M.R.
sio (3), alcfahpaet (2)jraremaataal, Weleeya (14), posh (13), siya (3), yotowe, iyang,Oalayallmang, yaremaat, yourwpiig, yar, j/oth, yaiyaremaataal, Weleej/a (6), yai, yaremaat (2)yaiyahiyahiyahtyaehe, yoetoulapyoaftlagiilhyoahtiligiaelh, aetoulap, aetal, oeta, oetalflilh, aetaei6<tae, eetaemwalh, eebaekeejFaurwpiigyaehe (10), j/aewavrils, i/oetaetaei, yaehegach, kapateej/a<,yoemwoelh, yaelog, yael, 2/oefath, j/aewal, j^aegafitaeg, yae (2),beeyaeyaehe (2)Faragij/eihy (2), baej^, mangly, galmangfj? (2), gawaewaay (2), tafeey
flegihr, ishl (5), flhagil, taj, /ngeliis (2), (shUh (8), ishilhi (2),ftaelhl, fyang, itaelh, iwoengoe, fgaelha (3), yaetaetaej,aetaj, aetajfiilh, gaimangiy, itipaelhi, gajgabungv (2),aetaei, paahi, /faelhuugflegihr, ishi, flhagil, ishilh, paahi, Zfaelhuugpaahiihy (2), iftsh, iftr, jAmwihmwihaiihy (3)MMwe (4)yaewanriislihwanhaeygawaewaaj/Weleej^akepateka^/M-'deeyawoalow, woa\h (2)yoehfitou (2), ga/5taeg (6), yaega^taegga^taegaetai^ilhya/ya/ya/ilhaffil, fatoffi (7), gerajiroeffilhl, tilifftaelh, roesafato^i,Faragjye, seehgi, ^ish (5), slfatogi (2)ilha(?il, fatojj, seeh^i, gishseehffiilcgihrilegihryoahlajfiilhgozvlhigafitaes (6), masharaj (3), yaegafitaep, Namwoachilj?gafltaei?, masharaj?Namwoachiii?tirigiInffdlls (2)manffiy, gaimanjiy (2)galayalimawj, chang, lahng, taiing, cangchan^, lahnff, tahngChang, \a,hnggahnngh
471762—60- -22
318 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173Table 3.
—
Characters and their actual syllabic values, arranged according to sequenceof characters in figures 25 and 26—Continued
Character No. Syllablesrepre-sented bycharacter Inform-ant Words in which character is used
X-f33.X+IVXI....
XII...XIII..
XIII+III.xrv
XIV-1-III-XV
XVI-
XVII.
XVIII.
XIX.-.
XIX-f-14.XIX-f39.
11lib
Ihl
ri
-1
-Ih
-n
-nhIhiIhimi
miimwlwlh
-c
shl
shlh
-sh
Shipi
plh
piipipihSi
shlshe
-shoayoayoahyoehteti
tiltv
-d
ri
uh
w-
-wvhu
R.R.O.R.
C.R.R.C.R.C.M.R.R.R.R.R.C.C.C.R.C.R.O.R.C.R.O.R.O.M.R.R.C.M.R.C.M.R.R.M.R.R.C.R.C.R.C.R.R.R.R.R.R.R.O.M.C.M.C.R.C.M.R.O.M.R.O.R.O.R.R.R.R.
tiZfgiaelh, Hmeeg, ngaH (2), QalayaHmanglihurvrlihmwraeZA) (2), fath-fathviftf (4), gozylhi (5), geragiroegi^Aj, mwalM,ngaZW, ulunhvMi, Ihi (2), kapeta/ftj, gamwoelhaettt (3),sitipaeZAi, maattj, itipae/Ai, itaeZftiTaelhi, fath-fathvZfti, gamwoelhaeMitirigi (an error?)yaremaataaZ, kapetaZ (3), matae-maetae?, nhepaiyaremaataai, Uhagii, kapetaZyoahlagiiift, raeZA (5), tlligiaett, faZft, woaZ/i (2), fatilh, tahlhyoalilaguZA, wae-waett, ishiZft, fatiift, tahlhSatmIhaewIhaenA (2)gamwoelhaeZftiishilhiwisilhaelh, wisUh, milhTnjlhsimnlhmwilhwihcwihcwlhcwihcUhi (4), ishilh (7), laftjlhi (2)Uhi, UhilhshihvashihmbaesAbaesfebae«AUhitapjlh, pipipipipihpmPihg, pihpihPihs, pihpihYam'wpitg (2)pipipipipinpiftsibae (8), aibaegvla (3), »iboe, aigal, siya, (3), ml«ilh, mi«i-Ibaelh, si (2), sia (3), sifatogi (2), sitipaelhi«ibae, s/mulhIngelii* (2), yaewavru«, bae«bae«l«Ailh«AogihsAbabioarwyoarvryoah, yooAlagiilhyoehTeomal<iligiaelh, iirigi, si/ipaelhi, Ujpaelhi, fa^flhtatilhta/.i\hfaliilhiaJiilhfa<»lhboad, tahdboad, tahdboad, tahdfoari (3)foarifoarjYaMrwpilg (2), towlap (3), ulunhvlhi, yaetowlap, aetowlap,wngMng, toMlapuhnguhngsowgTOW-TOW, shvhwvh\&tou (2)
No?6o7'^^'^' CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 319Table 3.
—
Characters andltheir^actuaVsyllabic values, arranged according to sequenceof charactersJn figuresl25.^and 26—Continued
320 BUREAU OF AMERICAN ETHNOLOGY [BuU. 17STable 3.
—
Characters and their actual syllabic values, arranged according to sequenceof characters in figures 25 and 26—Continued
^n*Rn*i^' ^^^' CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 321No. 60]Table 3.
—
Characters and their actual syllabic values, arranged according to sequenceof characters in figures 25 and 26—Continued
322 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173Table 3.
—
Characters and their actual syllabic values, arranged according to sequenceof characters in figures 26 and 26—Continued
NaGoT"'^"^" CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 323Table 3.
—
Characters and their actual syllabic values, arranged according to sequenceof characters in figures 25 and 26—Continued
324 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173Table 3.
—
Characters and their actual syllabic values, arranged according to sequenceof characters in figures 25 and 26—Continued
Character No.
59.
60.61.62.
636465.
66.
Syllablesrepre-sented bycharacter
66+III6768
69707172737475767778
buputo+unonesv
-szv
nonenonelolunlivfafae
feefoe
-f
paefeemwlsosoashononenonenonenonenonegagvnonenonenonenone
Inform-ant
E.C.C.R.O.R.R.O.
R.R.R.R.R.C.R.R.0.M.T.R.C.M.R.R.R.R.R.R.
O.O.
Words in which character is used
gaigaftungv (2), ga^wnghpunh(tmlap
maasprmaasprFalaluusmaazi^rmaazpr
woaZowuZanhvlhiulunftplhii^olaluus/aelvw, Faelalap, I/oelhuugI/aelhuugta/eeyi-'oeshavlapfoeshavlapFoeshavlapi^oeshavlap (2)yah/yah/yah/alefahjjaet (2) (an error?)ta/eemwiVasorw (2), galii«orw (2), raesorwjosoorjos/ios
pare, jofltaeg, ffomwoelhaelhi
Anthrop. Pap.No. 60] CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 325Table 4.
—
Syllables and the characters used for them, arranged according to thephonemic sequence of table 1[Numbers In parentheses following a character designation Indicate number of times, when more than once,that character is used, by each informant, to write the italicized syllable of the word preceding in thesecond column. Plus signs: in the first column plus signs indicate when two syllables are written witha single character; in the columns under informant designations they indicate when two characters areused to represent a single syllable. Informant designations: R.=Marollgar; C.=Chlyemal; M.=Magilo;T. = Tachep; A.= Marutang.]
Syllables
326Table 4. BUREAU OF AMERICAN ETHNOLOGY [Bull. 173-Syllables and the characters used for them, arranged according to thephonemic sequence of table 1—Continued
Anthrop. Pap.No. 60] CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 327Table 4.
—
Syllables and the characters used for them, arranged according to thephonemic sequence of table 1—Continued
328Table 4. BUREAU OF AMERICAN ETHNOLOGY [BuU. 1T3-Syllables and the characters used for them, arranged according to thephonemic sequence of table 1—Continued
Anthrop. Pap.No. 60] CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 329Table 4.
—
Syllables and the characters used for them, arranged according to thephonemic sequence of table 1—Continued
330 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173Table 4.
—
Syllables and the characters used for them, arranged according to thephonemic sequence of table 1—Continued
Syllables
-ng.
-ngh.
pa-
pah.,paahpae..
pe...pi—
.
pih..pll-.pu...
-p....
ra.
raa.
re.
ri...
rii..ro...
roe..
roeh
-r...
rw-..rwee
-rw..
Words in whichsyllables occur
ca.nggalayalimanjch&ng\ahngtohngIhahnffnngnhngtahnffftcdLTighIhahn^/igabumgftkepatgepotnhepalpangalhSaepalhkapateeyaekepatekaypnftngashpaaftisitipaelhiitipadhialefahpnetkapetalkapctalhitaptlhpipiPihgpihpihYawcwpiigPwlowathpwnhtoulapaetoulapyaetoulapFagosapFaelalapFoeshavlapmasharagraralhgeragiroegilbiFaragiyeraffshraesorwroelhraelhirerebaegareboegar«yarcmaatyarcinaataalfoar;tiri'giyaewavrjisro-rorow-rowgeragiroegilbiroesafatogiroeAshyarihrilegihruurjosoarmaasvrmaazvrYaurwpiigrwecgbab!oar««yoarwuurwsorwgaliisorjtfraesorw
Character used by—
R.
12, IXIXIXIXIX52526212525258, IX43 (14;4343(5^43(3)4343264343262666 (2) (error?)26 (9)26XIVXIV/XIV+IIIXIV/XIV-MIIXIV (2)131026(3)2626262626(2)15(3)15/153535353535 (13)35(2)35(5)35(4)35(2)353535XVIII (3)X (error?)3522-2215-15353535353535354747478(2)47+39+IV (2)22224747(2)47(2)47
C.
12IXIXIX52521212
524343
43
26 (2), 43
XIV/XIVXIVXIV/XIV
5926(2)
2615
3535
M.
12IXIX
4343
43
35(4)
35(2)35XVIII
22-22
35354747
35
XlV/XIV-flllXIVXIV/XIV
26
XVIII
26(2)
Anthrop. Pap.No. 60] CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 331Table 4.
—
Syllables and the characters used for them, arranged according to thephonemic sequence of table 1—Continued
332Table 4. BUREAU OF AMERICAN ETHNOLOGY [BuU. 173-Syllables and the characters used for them, arranged according to thephonemic sequence of table 1—Continued
Anthrop. Pap.No. 60] CAROLINE SCRIPT—RIESENBERG AND KANESHIRO 333Table 4.
—
Syllables and the characters used for them, arranged according to thephonemic sequence of table 1—Continued

BUREAU OF AMERICAN ETHNOLOGY BULLETIN 173 PLATE 42
rrfv
Tattooing in native script on arm of Maralatuy, a woman of Faraulep Island, Faraulep
atoll. (Photographed by E. Quackenbush.)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 173 PLATE 43
Tattooing in native script on leg of Letaweribul, a woman of Falalap Island, Woleai atoll.(Photographed by E. Quackenbush.)
BUREAU OF AMERICAN ETHNOLOGY' BULLETIN 173 PLATE 44
Native script and Japanese katakana on canoe-housc beams, Faraulep Island, Faraulep
atoll, {a, b. Photographed by S. Kaneshiro; c, photographed by E. Quackcnbush.)