Art. XXXVI.—Some Notes on the Language of the Chatham Islands.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 51, 1919, Page 415

Art. XXXVI.—Some Notes on the Language of the Chatham Islands. By Archdeacon H. W. Williams. [Read before the New Zealand Institute, at Christchurch, 4th-8th February, 1919; received by Editor, 11th March, 1919; issued separately, 19th August, 1919.] Chatham Island, called by the Maori Wharekauri, and by the earlier inhabitants Rēkohū, was peopled by a branch of the Polynesian race known variously as Maioriori, Mouriuri, and Mooriori (generally spelled “Moriori”).*This is probably not a tribal or race name, but, like the word “Maori,” an adjective meaning “native.” Bishop Selwyn, who visited the island in 1848, says that they called themselves “tangata maoriori.” Accounts of the early history of the island and its inhabitants are meagre in the extreme. Toi (c. 1125), making for New Zealand in search of his grandson Whatonga, is reported to have sighted an island “like a cloud in the distance.” It does not appear that he touched there, and no further information has been preserved. On reaching New Zealand Toi found portions of the North Island† Recent investigations indicate that this previous occupation probably covered both the Islands of New Zealand. occupied by a people now referred to as “Maruiwi,” the name of a former chief. Trouble soon arose between Toi's party and the Maruiwi, and a war of extermination was raged against such of the latter as had not intermarried with the new-comers, and a small remnant of them escaped in six canoes, sailing towards the south from Cook Strait. It is said that some of these reached Chatham Island, but actual proof of this fact is wanting. Moriori genealogies certainly contain the names of Toi, his son and grandson, among the demigods, but a knowledge of them may conceivably have come through another channel. Some little time later (c. 1200) one Kāhu, in a spirit of adventure, fitted out a canoe at the mouth of the Rangitikei River and set out to find the island reported by Toi. He reached Wharekauri, where he found inhabitants who are supposed to have been the Maruiwi refugees from New Zealand. But it must be borne in mind that the Moriori traditions claim that their ancestors had resided on the island for many generations before Kahu's arrival, a statement supported by notes to two Moriori genealogies (Shand, pp. 53, 55); though it is difficult to assess accurately the chronological value of these genealogies, seeing that according to them 1,750 years intervened between Kahu's arrival and the coming of the Rangimata canoe, a period which upon other data is estimated to be about 150 years. The Moriori claim that his ancestors antedated Kahu by a considerable period receives, perhaps, more support from the statement of Matorohanga that four canoes are known to have reached Wharekauri from Rarotonga—one of them, Te Ririno, “long before the visit of Kahu.” The names of two of the occupants of this canoe are given. One of them, Kapohau, had a son, Te Ao-marama, a name which occurs in both the Moriori genealogies as that of the father of Rongomai-whenua, who is stated to have been the first ancestor who resided on the island. But this was, according to the genealogies, some 128 generations before the beginning of this century—say, about 1300 B.C., or fifty-five years before the siege of Troy. Reverting to Kahu, it is said that he, with some of his companions, set out to return to New Zealand, but nothing is known of their fate. Others of the crew married on the island, and some of their descendants have since returned here.

The Moriori traditions also mention the arrival (c. 1350) of three canoes—Rangimata, Rangihoua, and Oropuke—from “Hawaiki,” which we may fairly assume to be Rarotonga. Lastly, in 1835 a body of Taranaki Maoris occupied the island, completely dominating the Moriori inhabitants, many of whom they murdered, while the survivors were reduced to a condition of slavery. Under this treatment the Moriori dwindled rapidly. It is believed that at that date there were about two thousand; Dieffenbach estimated their numbers at ninety in 1840,* E. Dieffenbach, Journ. Roy. Geog. Soc., vol. 11, p. 207, 1841. Bishop Selwyn took a careful census, and found 268, including children. and Mr. Shand gave the number as twenty-five in 1898. So far, then, as history is concerned, the inhabitants of the Chathams would appear to have come from “Hawaiki” (possibly Rarotonga); some may have been Maruiwi from New Zealand; those who came with Kahu were probably of mixed Maruiwi descent; while all would be much influenced by the later-arrived Maori tyrants. Matorohanga has preserved some descriptive notes of the Maruiwi, which conform in many respects with what is known of the Moriori. At the same time, many of the characteristics recorded appear also in the descriptions of the mixed dwellers on the island of Rangiatea, on which Whatonga was cast away. But this fact is not very helpful, for though we may identify Rangiatea with Raiatea, near Tahiti, we know nothing of the inhabitants of that island eight hundred years ago, or of whence they came. We may now turn to the language. For our knowledge of this we are indebted almost entirely to the patient and sympathetic investigations of the late Mr. Alexander Shand, who resided on the island for many years. A vocabulary by Mr. Deighton (Appendix to the Journals of the House of Representatives, G.-5, 1889) contains equivalents for nearly nine hundred English words. About sixty of his Moriori words are not recorded by Shand. Mr. Shand collected a number of traditions and legends during the years 1868–1869, when, as he says, “the old people could speak their own language.” The information thus obtained formed the basis of a series of articles in the Journal of the Polynesian Society (vols. 3–7),† These articles have been republished, and form vol. 2 of the Memoirs of the Polynesian Society. References herein are to the pages of this volume. in which Mr. Shand included original Moriori matter enough to fill about forty-four pages of the Journal. Moriori has been described as “a corrupt subdialect of New Zealand Maori”; and when the writer began work upon the recent edition of the Maori Dictionary it was suggested that Mr. Shand should be invited to contribute a list of Moriori words with a view to their incorporation in the dictionary. To this he readily agreed, and, though much hampered by ill-health, he compiled a vocabulary of some 2,500 words, many of which are illustrated by examples. He completed the work in 1910, the MS. reaching New Zealand by the mail immediately preceding that which brought the news of his tragic death. With him, unfortunately, perished a considerable amount of additional matter which he had hoped to publish; and we were deprived of his help in elucidating questions arising out of the vocabulary he had compiled. An examination of the vocabulary forced the conclusion that the opinion making Moriori merely a subdialect of Maori requires revision, and it was decided to hold it back for independent publication‡ The copy is now almost ready for the press. As a matter of fact, Moriori appears to be farther removed from Maori than the dialects of many of the islands of the Pacific. Peculiarities of

both grammar and vocabulary make the language more difficult for one conversant with Maori to read than Rarotongan, and not less so than that of Tahiti, Uvea, or Niuē. Structurally the dialect is exceedingly plastic, and affords examples of letter change in bewildering variety. In some cases the metamorphosed word has wholly displaced the form with which we are familar in Maori; in others the Maori form is used concurrently with one or more variant forms. This appearance of Maori forms may in some cases be a survival, while in others it may be due to adoption from the speech of the more recently arrived Maori overlords. In the matter of letter change we may, in comparing with Maori,* The Maori form is given in brackets. note the occurrence of— A for E:† Cases of this occur in some dialects of Maori.‡ This occurs in Marquisan. ta (te, art.), a (e, prep.), tikamata (tukemata). A for EI: ka (kei). A for I: ka (ki, prep.). A for O† takato (takoto), kotau (koutou). A for OA: mana (moana). A for U: oata (hoatu), mari (muri). AI for I: hokai (hoki). E for A:§ This occurs also in Rarotongan and Mangarevan. heu (hau), kei (kai), etu (atu), eriki (ariki), tinane (tinana). E for O: nape (napi). E for O: reimata (roimata), tike (tiko), mohewao (mohowao), heme (hemo), konehi (kanohi). E for U: ange (aku). E frequently inserted before H: ehau (hau), ehē (hae), maehanga (māhanga), maehara (mahara), aeho (aho), poeho (poho); and sometimes (as printed) after H: heanga (hanga)—but this last may be only a case of the modified H, which will be dealt with below. EI for A: eitū (atua). I for A: puti (puta), eringi (erangi), iki (ika), mehori (mahora), wihine (wahine). I for E: mi (me, conj.), ti (te, art.). I for O: rari (raro), korihiti (korohiti), tiri (tiro). I for U:† matī (matū), tikamata (tukemata). I inserted between two vowels: kaiore (kaore). I inserted before or after H: puiha (puha), hiaka (haka)—but the latter may be a case of modified H. O for A:† mahoro and mohoro (mahora), tokoto (takoto), mōū (māoa). O for E:† to (te, art.). O for I: toki (tiki), tango (tangi), pororo (porori). O for U:‡tongo (toku). O for OU: kotau (koutou). O for UA:† akonei (akuanei), aoreka (ahuareka). O for WA: rao (rawa), manao (manawa). O for WE: wao (wawe), kao (kawe), makao (makawe). U for A: ur′ (ara), hunau (whanau). U for E: tu (te, art.), hunū (whenua). U for I:† ku (ki, prep.), muru (muri). U for O: kunei (konei). U for OA: mōū (māoa).

It will be noticed that in the case of many of these changes the vowel appears to have been attracted, so to speak, to the vowel in the preceding or following syllable (though in not a few instances the vowel change has disturbed a previously existing symmetry), and the same rule may be observed operating loosely in the choice of alternative forms in a sentence: thus ku muru will generally be used for ki muri, ko ru kupu and ki ri kupu for ko ra kupu and ki ra kupu respectively (ru and ri representing the article te), no ro mē for no te mea. A good illustration occurs on page 77: Kitē ko Tu i rari i ri papa o ro waka, Tu was found beneath the flooring of the canoe. (In Maori, I kitea ko Tu i raro i te papa o te waka.) In addition to these changes a vowel is sometimes dropped (i) before a consonant, as na (ena), ha (aha); (ii) after a consonant, leaving a closed syllable, as rangat (rangata), nawen (nawene), hok (hoki), or (oro), mot (motu)* Ka is thus sometimes clipped to k′, and ta to t′.; (iii) after a vowel, in which case the preceding vowel is lengthened by way of compensation, as ahē (ahea), mē (mea), ī (ia), and the passive terminations hī, rī, tī (hia, ria, tia), ingō (ingoa), nō (noa), ahū, (ahua), tā (tae), marī (marie), kō (koe), ēnē (ēnei), nū (nui), rē (reo); and (iv) before a vowel, the remaining vowel, as in the last case, being lengthened, as pē (pae), tē (tae), wēwē (waewae), hēre (haere), and, apparently, kā (koa). Examples similar to the group (iv) occur in Tikopia, where we find fē, kē, mē for foe, koe, moe. In a number of words a vowel has become long where in Maori we have a short vowel, as ihē, tihē, īka, tāki, tīrā, apaāpā, tapōko. But I speak with some diffidence on this point, as I find that Mr. Shand has, in the MS. in my hands, corrected the quantity in several passages quoted from the articles which were printed some ten years previously. It is possible that in reciting the legends a rhythmical diction may have been adopted which placed on certain vowels a stress which did not accord with that of the normal pronunciation. The consonants are not quite so pliable as the vowels, but still offer no inconsiderable variety. In the cases of H, K, and T a peculiar method of pronunciation is sometimes adopted, in which the tongue appears to be somewhat arched into the palate and the letter uttered with a slight emission of breath which not infrequently produces the effect of a suppressed I, or sometimes E, sound before the proper vowel of the word. This stressed pronunciation is used with K only when followed by A, possibly only in the case of the particle ka; with H only when followed by A or O; but with T it occurs much more frequently, and with any vowel except O. Mr. Shand represented this peculiarity of utterance variously in the case of each of the letters named, using the combinations HHI, HI, or HE for H; KH, KHI, or KKHI for K; and TCH, TCHI, TCHE, or TC for T. He states in a note that it is difficult to represent the pronunciation in writing, and mentions that it appears to be used in some cases for the sake of emphasis. It is of interest to note that a somewhat similar usage with respect to H was observed in the north of New Zealand in the early days, and led the missionaries to write Shunghie and Shauraki for Hongi and Hauraki. In Tonga, too, the method of pronouncing T when followed by I leads to its being represented variously by J or S.

As examples of change in consonants we may note— H for W H:* Cases of this occur sometimes in Maori. huti (whati), hi (whi), hunū (whenua). H dropped: † H is completely discarded by Whanganui and neighbouring tribes. eke (heke), oeha (hoha), iki (hiki). K for NG: ka (nga, art.). (No other cases noticed.) K for T: ka and ki (te, art.), kihorei (tihore). M for W: iwi (iwi), pakihimi (pakihiwi), mahine (wahine). M for K: matahi (katahi). N for R: ‡ This occurs in Marquisan. manino (marino). (Uvean has naua, natou; raua, ratou.) NG for K: ngonei (konei). P for K: harapepe (harakeke), tupe (tuke). P for M: ‡ tupuaki (tumuaki). R for N: ‡ irokonei (inakuanei), hiringaro (hinengaro). R for T: rika (tika), rangat (tangata), rupuna (tupuna). W inserted: warero (arero), arowaro (aroaro). This sometimes indicates a U lost by substitution: hiwaki and hiwiki (huaki). Awainei for aianei involves more than the mere insertion of W. A further step in the transformation of words is taken in the dropping of a complete syllable at the end of a word, as hi for hine or for hinga, ma for mate, mo for motu. The effect of these letter changes is to introduce a certain amount of ambiguity in the interpretation of our texts; for, on the one hand, a single form may take on the functions of a number of others, while, on the other, the same word may assume a multiplicity of shapes. Thus a, which has seven distinct functions of its own, acquires in addition those of ae, āna, ana, and e; e has three functions of its own and those also of ae, he, hei, nge, and is said further to be inserted sometimes into a sentence with no other purpose than that of euphony—a very refuge of the destitute for a grammarian; i has two uses of its own and represents also ae, ai, e, he (sing. and pl,), hei,. ia, ki, te (sing. and pl.), and wai, and also perhaps a euphonic use; hi, besides its own duties, has those of he, hia, hine, hinga, and whiwhi; ka does duty for kahore, kei, koa, ki, nga, and te, besides retaining two previous meanings. Conversely, the particle atu appears also in the forms ati, etu, eti, whatu, wha, at, and et; whano may become whani, hano, hana, or hane; for the singular definite article we have the choice of ta, t'a, te, t'e, t'ei, ti, t'i, to, tu, t′, ne, re, ri, ro, ru, i, h'a, ka, ki, and ko, and three of these forms, i, ka, and ro, may also be used in the plural.§ The forms t′ and h′ are here used for the modified t and h referred to above. Metathesis is by no means uncommon in Maori and cognate languages, but some examples in Moriori are interesting: euwha for ehua, pass. of ehu, leads to eueuwha for ehuehua; huanui appears as uwhanui, and hiwiki (huaki) as hikui; whine for wahine would certainly seem to connect closely with the fine in Samoa, Tahiti, Tonga, &c., were it not paralleled as a case of modified metathesis by the occurrence of kawhai for kahawai. The last structural peculiarity remaining to be noticed is connected with reduplication. It was noted in dealing with the interchange of vowels that a principle of attraction seemed occasionally to be operating. In not a few cases, however, of reduplicated words, a vowel in one element of the word has been altered. Thus we have hinihina for hinahina, ateata for ataata, okihakehaka for hokohakahaka, and iarehara for harahara. This variation of vowels occurs very frequently in reduplicated words in Tongan.

It is not very surprising to find that many words have in Moriori meaning differing in greater or less degree from that in Maori. We may notice of few of these* The meaning of the word in Maori is given in brackets.: Anga, front, intestines (aspect); hua, keel of a canoe (hand-spike, lever); kamo, beckon, wave to (wink); kopa, veer, turn (bent, folded); mapuna, ripple, v. (well up); moteatea, shine (fearful, apprehensive); ngoi, weak (strength); oti, death, as well as (in Maori) finished; pao, open, a. (strike); ro, roto has extended its meaning beyond merely “the inside,” and we have such expressions as ko ro ta pari, “over the cliff”; ko ro to wai, “above the water”; tango, work (take, attempt); toke, small (out of sight, gone). Of all the words recorded by Shand about 10 per cent. appear to be from roots no longer preserved in Maori. It is possible that some of these may, on further investigation, prove to be familiar roots disguised by some of the letter changes discussed above. Mr. Shand remarks that “The language … retains many words more peculiar to the Rarotongan dialect”; but he unfortunately left no list of these,† One, which he mentions in a note, is open to doubt. and I have only succeeded in noticing one. A study of these non-Maori words reveals in the case of about 12 per cent. identity or close approximation with words in one or more of the leading Polynesian languages. Thus we have ha, sacred (sa, Sa., Uv.); maramara, nausea (malamala, Sa., Ta.); puni, many (Marq.); tau, reef (Ta.); topa, fall (E.I., Mang., Marq., Ta.); tohua, public place (Mang., Marq.); whariu, turn (fariu, Ta.); pehe, sing (Ta.; pese, Sa., Tik.). Among the approximations the more interesting are hakana, coffin (Marq., hide); murumuru, singe (Fut., warm oneself); nono, taut (Niue, bind; no, bind, Uv.); po, troubled (Sa., have war); poi, leap (Niue, run; Tik., go); ro, go (Tik., oro, go); t'iei, not (Marq., tie; Tik., sīe; Nuguria, teai; Sikiana, seai); toro, shoal, ridge (Sa., tolo, promontory); oko, many (Mang., Marq., strong; and apparently used to give a plural signification almost equivalent to “many”); nuno, of no account (E.I., nunu, thin). This would seem to be the place to notice the causative prefix. The usual form is hoko, but this varies to hoka, hoke, hoki, hok, ho, oko, oka, oke, oki, ok, and o. In some cases in the vocabulary Shand uses whaka; but this does not, I think, occur anywhere in the published narrative matter, and appears to be an adoption due to Maori influence. This inference is supported by the fact that under the head hoko Shand mentions whaka only as the Maori equivalent. The normal form hoko suggests comparison with ho'o, ho, in Hawaiian, which seems to be the only other language in which the vowel is normally o, though in a few Tongan words foko replaces faka. In the affinities thus disclosed the Marquisan dialect figures most frequently; then follow those of Tahiti, Samoa, Hawaii, and Tikopia; Easter Island, Mangarewa, Tonga, Uvea, and Futuna falling a little behind. Rarotonga falls in the last group; but it must be pointed out that the opportunities for a comparison with this dialect were somewhat restricted; and, in any case, it is a mistake to stress heavily numerical values in a comparison of this nature, particularly when the numbers are small and the available vocabularies far from complete.‡ Mr. Best gives (Trans. N.Z. Inst., vol. 48, pp. 435-36) thirteen Maruiwi words, of which one, kohi, appears in Moriori with a slight modification in sense, hasten for come.

It now remains to notice a few of the grammatical peculiarities of this language. With the limited material in our hands it is not possible to formulate a complete scheme of the grammar; but we are much helped by the fact that Mr. Shand has given a careful, and often literal, translation of what he published. Where he was unable to translate he had to content himself with the sense of the passage as obtained from the narrators. There is a tendency to clip sentences by the omission of— (i.) The definite article, as—No/taenga ki/tuatoru k” akina i tao o Whakatau, When it came to the third time Whakatau's spear was thrust forward (80); Ka oti/tarei, The adzing was finished (35); Pena mai i titike o/tupapaku, ti te hope to hohonu o/toto tangat, Such was the heap of the corpses that the depth of human blood reached to the loins (98); Ko roto i ri pu hamama o/totara, In the gaping trunk of a totara (97). (ii.) A verbal particle, as—/Ru mānga a Kura,/toru mānga a Kura, Kura took two mouthfuls, Kura took three mouthfuls (41); E kore/tē, He will not get there; Ko maru/horo, Maru swallowed it (63). (iii.) A preposition, as—Ka ra waihe ī/wahine mana, She became a wife for him (65). (iv.) The nominal particle, a, as—Ka whai mai/Nunuku, Nunuku replied (132). The distinction between the active and passive of verbs is not always clearly preserved, and the preposition e, which in Maori indicates the agent after a passive verb, is in Moriori used with either an active or a passive: Potehi etu e Maui, It was discovered by Maui (29); Tē potehitī mai kō e o hunau, Lest you be discovered by your relatives (35). Similarly, what may be called the transitive preposition i is inserted also after intransitive or passive verbs: Ka tak i te rau, The feather fell (63); Ka hapa i te hei, There was one necklet over (77); Ka pau i te hunu o tana kei i a Rohe, Part of his food was consumed by Rohe (29); Ti reira pea i te kiato o ta imi, Perhaps the bulk of the people were there (132); A te huna e ratau i te rangat o ratau, They were hiding one of their number (77). This usage may perhaps explain away the supposed euphonic use of i mentioned by Shand. The use of the particle e after personal pronouns and some adverbs is more frequent than that of its Maori equivalent nge, and may reduce slightly the number of cases of the euphonic use of the letter. In the narratives as edited the definitive particle, ko, is used with the subject to a verb more freely than in Maori: Ka puta ko Kura i t” ata, Kura went out in the morning (42). But in many apparent cases of this use it is preferable to regard the ko as another form of the definite article, though it is not so recorded by Shand: Ka ki mai ko tupuna ki a i, His grandfather said to him (129). The pronoun ena, those (near the person addressed), is used adjectively with the signification other: Ki te hunga ena ko Tamahine matua, With other people (she was known as) the senior daughter (76); K′ huihui mai ko Te Wheteina me na ka imi, There assembled the Wheteina and the other tribes (94). An additional indefinite pronoun, a, is used to obviate the repetition of a noun, somewhat as nga mea would be used in Maori: Kimi mei ki ka rakau, i a tika, Fetch some poles, let them be straight ones (40). The expression “hunaunga no ko,” a relative of thine, which occurs more than once, is peculiar; but we are unfortunately not in a position

to say whether this use of the personal pronoun in place of the pronominal suffix had a wider currency. Another anomalous use occurs in the phrase “ta imi na ratou e kai,” the people who ate him (132). The Maori idiom would be “nana i kai.” A curious inverted construction is found occasionally in which the direct object of the sentence is treated as if it were the indirect object: Ko ta imi t'iei haramai i tangat, ka pang etu ki ri ngakau, As for the tribes from which no men went with him, he threw them the entrails (80); Ri oro mai au nei ki t′ opeope, He threw me a scrap. The phrase “no ro mē,” because (Maori, “no te mea”), takes also the peculiar forms “ka ro a mē” and “ka ra wa mē,” which it is difficult to explain grammatically. A similar difficulty is raised by the sentence “E mē wa mē meheki naku” (132). Shand's translation, They are things belonging to me, no doubt gives the sense, but fails to explain the syntax. The points reviewed in this paper do not, in the opinion of the writer, make for any special theory as to the identity or origin of the Moriori race. In fact, it is well that we were pledged to no theory, for it seems that the only conclusion we have succeeded in establishing is the entirely negative one that the Moriori tongue is not correctly described as “a subdialect of New Zealand Maori.” So far from that being the case, it has as much right to be considered independent as any of the known dialects of the Polynesian language.