Art. I.—Oceanic Comparatives.
Communicated by A. Hamilton.
[Read before the Wellington Philosophical Society, 5th May, 1909.]
Probably most students of Oceanic languages begin by merely comparing words of similar form and meaning in different languages. Later on, the testimony of words to the beliefs of the people, to their common origin, to the position and character of their home before they became scattered, becomes of great interest. But meanwhile certain principles of comparison emerge from the number of comparisons made, and gradually the comparisons become in consequence sounder. Some comparisons made at first without hesitation are seen to be wrong ones, while many comparisons formerly unthought-of can now be safely made. The following paper describes briefly a few principles of comparison in dealing with Oceanic words. They are principles which seem to be true when tested, but they were obtained by collecting a great many examples of comparatives, and then considering these examples to see if they had anything in common. They appeared to conform to certain rules. These rules were then tested by applying them to many fresh comparatives, with excellent results. In the hope that they may be of service to others, or, at least, may be suggestive, it seemed worth while to write them down.
One side of the subject, however, is not discussed—the question of phonetic changes; it deserves separate treatment, and is too large to be discussed here. But the principle may be at once laid down that no comparison should be made of words whose forms are different unless the difference can be explained according to well-known phonetic laws. One might suppose this caution hardly needed, but it is, in fact, generally ignored. One need only quote a sentence from “Maori and Polynesian”: “The Aryan ra (to rest, to be delighted, to love). … is not far off in origin from the root la (to yearn for or desire), which appears in Sanskrit lash (to desire). … It seems much the same as the Maori reka (pleasant), and, in the other sense of ‘calm,’ as the Maori whakaruru (sheltered from the wind).” Such comparisons may easily be a source of confusion.
Another principle to be insisted on is that all the letters of a word must be taken into consideration. If the Mota vula (moon) is compared with the Malay bulan, the n must be accounted for; if namu (mosquito) is compared with namok, the k must be accounted for. In these two cases the n and k suffixes, common in Malay nouns, are of great interest to the philologist. It is common to ignore single letters, especially at the end of a word. Much of great interest and value is thus passed by.
It is perhaps best to set aside onomatopœic words, interesting as they are. The author of “Maori and Polynesian” compares the Maori mumu (to hum) with the Latin murmurare (to murmur), while “koko (the tui) may be set beside the Sanskrit kakh (to laugh), English ‘quack’ and cackle.’”* Probably such words appear in all languages. There is a Malay word for “dog,” asu, and this is found in Melanesia, but not in Polynesia; so it has been supposed that the Maori au (to bark) represents the Malay asu (a dog)—in fact, that nothing but the bark is left in New Zealand. It is a comparison one is loth to forego when one remembers the classic case of the Cheshire cat; but one must allow that “au au” is very like the sound a dog makes everywhere.
It seems reasonable to suppose that the fuller form of a word is the root. We are familiar enough with shortened forms—cab, from cabriolet; bus, from omnibus; car, from carriage; phone, from telephone. It is, no doubt, a common process in all languages. Now, if we take any word of two or three letters (not a particle) in any Ocean language, it appears to be the case that a fuller form of the word can always be found in some other language. Of course, the fuller form may have been lost everywhere, but no such case has come under the writer's observation. These fuller forms consist of two consonants and two vowels, or, at least, of two syllables. Thus, though hua is “moon” in one language of Indonesia, hula is found in others; though niu is “cocoanut” in many Oceanic languages, niru is found in one (Vella Lavella). It also appears to be the case that no root-forms occur of more than two syllables. There are many longer words in the vocabularies, but they are either roots plus a prefix or suffix or they are compund words. Thus, in the Mota word nonom (think) the m is the common Oceanic verbal suffix; in the Mota malumlum (soft), ma is the common adjectival prefix, while lumlum is the reduplicated root (lumu). The Florida tidalo (a god) is a compound word, ti appearing also in tinoni (a living man), whatever the meaning may be of dalo—the same word, doubtless, as the San Cristoval ataro (a god); Mota tataro (a prayer)—from the first word with which all prayers or charms began; Tahitian tarotaro (a short prayer to the gods). The Florida word appears to be indirect evidence that the primitive religion of the people was ancestor-worship. The Mota geara (a fence) is really two words—ge and ara—a form of the latter being commonly used for “fence” in Oceanic languages.
[Footnote] * I cannot refrain from giving instances of the curious reasoning and inaccurate statements in the chapter of this book entitled “The Maori as seen in his Language”: “Malay, as the tongue of a Mongoloid people, is assumed to be agglutinative,. … But there is nothing agglutinative about [Maori]. … Now, the only inflectional languages are either Aryan or Semitic.” Therefore Maori is either Aryan or Semitic ! So much for the reasoning. Later on he speaks of the “coincidence of Maori ruma (an apartment), used all through the Pacific in the sense of house, with the English ‘room.’” Now, there is no such Maori word at all, and, though ruma does occur in Melanesia, neither the word itself nor any form of it is known to occur in Polynesia.
The following principle seems, then, to hold good: that all Oceanic roots consist of two syllables. The first point to determine about a word is this root. But the full form is often missing, and the next point is to find out how these root-forms are modified. In the first place, the first consonant may be dropped. The Florida aho (the sun) is a modified form of raho or laho. This modification is a common one. Still more common is the dropping of the second consonant, especially in Polynesia. The full form of the Maori kau (to swim or wade) is karu, the word really meaning “to move the legs and arms.”
In many cases both consonants are dropped, as in the Rarotongan ai (fire), the root of which is kapi; or the Maori ao (dawn), from raho.
Sometimes the second vowel is dropped, as in the Mota kor (dried), from koro; but the first vowel remains, and is rarely altered; it is, in fact, the most stable part of the word. On the other hand, the second vowel is liable to frequent change.
The commonest modification of all is the dropping of the second syllable. Thus, the Maori ra (sun) is from raho; po (night) from pongi; pa (a fence or stockade) from para. It frequently happens that when only the first syllable remains this is strengthened by reduplication. Thus Mota koko (enclose) is from koro; Mota rara (scorch), from raho; Mota roro (deep), from roto; Mota lolo (inner parts), from roto; Mota roroi (news), from rongo; Maori rara (a branch), from rana; Fiji baba (side of a hill), from bara. Such reduplications are very common.
Perhaps there are cases of the dropping of the first syllable of the root; but until they are shown to exist, such comparisons as that of the Mota gana (to eat) with the Malay makan (to eat) must be viewed with suspicion. Manga, which also appears as maka, means originally “a cleft or rent,” in some languages “a mouth”; in Mota “to keep opening and shutting, as the gills of a fish.” This is, no doubt, the Mota representative of the Malay makan (to eat), the n being the common verbal suffix.
There is yet another fairly common modification of the root—by metathesis. Thus, the Florida labu (strike) is the Malay palu; the Mota toro (deep), the Maori roto; the Mota ma-vinvin (thin), from nihi; Maori ngaro (fly), from rango; Maori raku (scratch), from karu; Aneityum mulmul (soft), from lumu; Tongan pelupelu (bent), from lupe; Maori tumu (cape), from mutu. It may be asked which is the root and which the transposed form. The answer is that the transposed form is usually local, and is shown to be later by the fact that modifications of it are absent or rare; whereas the real root is found all over the Pacific, and broken-down forms of it are very common.
Of course, many forms of the same root may, and generally do, exist in one language. An instance may be given from Mota. The root koro means “to encircle, enclose”; hence, “to contract, shrink,” &c. It is found in many Oceanic languages, as koro, kolo, koko, &c. Probably it is the source of the following words in Mota: kolo (contract), koloi (a hole), kolkoloi (a small contracted thing), kor (to shrink with heat or dryness, dried breadfruit), koko (keep close, contract, carry water in the hands), kokor (enclose, keep carefully), kokos (enclose, as fish in a net; fowls, by the people catching them), kokot (enclose, contract), kokota (narrow, contracted), kokorou (fold the arms or wings), gogo (shrink or shrivel), gogorag (gather together with the hands), goro-vag (throw the arms round), goro (to cut round, and so generally to cut), gorogoro (harvest), goro (to embrace).
golo (fade or wither). There are other more doubtful examples of it, such as oloolo (a sacrifice), where probably (as in Efate gorokoro) the idea was that of protection, coming into the enclosure of the god sought and being shut out from the power of other gods.
Some examples of the application of these rules may now be given. Three roots will perhaps be sufficient to illustrate the subject.
Taking the Maori word roto, it is evident that we have the root itself. We may expect, however, to find it modified in various ways, and may look for it in such forms as loto, oto, lo, ro, roro, lolo, toro. The meaning of roto is “inside, within, the midst, places inland, a lake” (ha-rotoroto, a pond). From this meaning of “within” (which is the root meaning) the word comes to mean in Samoa (in the form loto) “the heart, desire, or will; the understanding; the interior of a house; a deep pool in the lagoon”—loloto (deep, depth), lotoi (to be in the middle). The Tahitian roto and Hawaiian loko have much the same meanings. In Tonga loto means “the mind”; loloto “deep, the deep, ocean.” In Mangareva we get a further meaning, for here roto is “deep, depth, the inward parts,” and “to obscure or darken”; while rotoroto is “the shallow sea,” a curious meaning, to be explained presently. We may follow this full form of the root to Melanesia, where in Saa Malaita roroto is “a dark cloud”; in San Cristoval rodo is “dark,” rodomana “the abode of the dead, the dark land”; and in Florida rorodo is “blind.”
The form oto does not appear in Maori, but from the meanings presently to be given of the Maori roro it appears more than probable that oto has been modified into uto, which means in Tonga “the brains (the inner parts), the spongy substance of an old cocoanut”; the Mangareva uto (marrow, yolk of eggs); Mangaia uto (little kernel growing in a cocoanut); Fiji uto (heart, pith of trees, marrow of bones); Mota utoi (pith of trees); Malay utak; New Celebes utok. Uta (inland country), “forest” in many languages, is doubtless the same word. A final o is often represented in Malay by a, and in Malay otak means “brains” (k is a noun suffix in Malay).
We now come to ro, which appears in Maori, meaning “in, into”; and lo, which in Santa Cruz means “night or darkness.” Taking next the reduplicated form of this, we find the Maori roro, meaning “brains or marrow.” The Samoan lolo and Tahitian roro have similar meanings. The Hawaiian lolo means “the brain, marrow of bones, insane.” Tongan lolo (oil, oily fat) and Mangareva roro (soft) are perhaps meanings derived from “the brain.” In Santa Cruz lolo is “fat or oil”; in Mota loloi means “inside, the inner part, the affections,” and is used in very many compound words to express states of feeling; roro means “deep, to sink down, be deep,” also “to be sunk down,” and so “shallow”; the water is said to roro in the well, hence the well is roro (shallow), an explanation of the Mangarevan word. R frequently becomes n, and there can be little doubt that nonom (to think) and nonon (to smear)—the m and n being verbal suffixes—are both from roto. In Efate, New Hebrides, roro means “to think,” rorona “thought”; and Dr. McDonald mentions that these words are sometimes pronounced trotro and trotrona.
Taking now the final modification, the transposition of the letters, the Mota toro means “deep”; toron, “to desire or long for”; Florida dolo, “to love”; San Cristoval doodoo, “black”; doa, “blind.”
In Oba, New Hebrides, lologi is “inside,” and rorogi “deep”; in Malekula, roror is “deep,” and ror “the afternoon” (gi and r are the noun suffixes in these two language). The meaning “afternoon” is from the
going-down (roro) of the sun; the sun is said to roro in Mota when it sets. Many more examples of this root might be given, but enough has been written to show how it follows the rules given above.
The root which means “light and heat” is a difficult one to follow, because the two consonants it contains are interchangeable. It seems best to give raho as the root, though this form appears to have been lost, because the modified forms of raho (aho, ao, ra, rara) are the most widespread.
If we take raho as the root, with the meaning of “light and heat, the day, the sky, clouds,” &c., we may expect to find several forms of the full root, because r and h are interchangeable letters, and r may become l, while h becomes s or t, and in the case of such a root it is difficult to say which is the transposed form; however, we may be guided by the number of modified forms which appear.
Laso, the Kayan word for “heat,” is a full form of the root, and so too, apparently, are salo (a cloud) in Saa Malaita, perhaps also sato (the sun) in Saa, and warowaro (the moon), in Saa w representing h, as it commonly does. The three last may be considered transposed forms of the root, and thus be classed with the New Guinea (Kerepunu) haro (the sun) and the Maori Haronga (a mythical sky-power who begot the sun and moon).
The modified forms we might expect to find by dropping the first consonant are aho, aso, alo, aro, and ato. The first four are found. Aho means “the sun” in Florida, and Ysabel (Solomon Islands), and Oba (New Hebrides). In Tongan it means “day”; while ahoaho means “bright or shining,” as the moon on a clear night. The Maori aho means “radiant light,” and aho-roa is “the moon.” Aso in Rotuma and Guadalcanar is “the sun”; in Samoa, “the day”; in Efate, “burning.” Alo is “the sun” in New Celebes (mata-alo) and in Santa Maria. Aro is “the sun” in Hula and Bula'a (New Guinea), “the sky” in San Cristoval; while the Malagasy maso-adro (the sun) is almost identical.
We get two forms by dropping the last vowel—the Mota sar (to shine, give heat) and the Aneityum lah (light) and lav (to shine).
Dropping the second syllable gives ra and la. Ra (the sun) in Maori is said to be the son of Haronga and the sister of Marama (the moon), a very exact account of the words, except that marama, from the root rama, is quite distinct from raho, and occurs along with it in many forms in many Oceanic languages. Ra in Maori also means “the day” and “the east.” In Samoan la means “the sun, daylight; to be intensely hot.” The words, with these meanings, are common in Polynesia.
Reduplicating the first syllabe, we get rara and lala. In Maori rara means “to spread out on a stage to dry,” and hence the stage itself. In Tahiti the word means “to scorch over a fire”; rarararauri, “sunburnt” (uri=black); Fiji rara, “to warm one's-self by a fire”; Mota rara, “to dry or warm by the fire”; rarang, “to dazzle by reflection”; rarangiu, “hard dry ground”; ma-rang, “barren ground, lazy.” In New Guinea the Motu raraia is “to dry in the sun”; Wedau arara, “to shine”; arai, “to burn”; Celebes rarang, “to dry.” The Florida ma-rara means “light”; Mota ma-rarara, “transparent, translucent”; Malay arang means “charcoal.” Samoan ma-lala is “charcoal”; Tongan lala, “to broil”; ma-lala, is “charcoal”; Kayan lala, “to wither”; Wedau, New Guinea, lalai, “light”; Mahaga, Ysabel, lalaha, “light”; New Celebes lallav, “scorch”; Mota lalav, “scorch, burn”; lawa, “blaze or flame; to be red, brilliant in colour.”
By dropping both consonants we get ao, the Maori “day, dawn, become light, cloud”; Tahitian “day, bright clouds, the bright land of heaven.” In Mangareva we get a, meaning “sun,” where remarkably little remains of raho!
Unusual forms, such as elo (“sun” in the New Hebrides), loa (“sun” in Mota, “cloud” in San Cristoval), have not been referred to.
Perhaps it may be allowed to make two conjectures in conclusion. Maro means in Mota “famine”: may this be ma-aro, and mean “a time of sun without rain, when the ground becomes hard and dry” (Hawaiian malo, “to dry up as water, to wither as a tree”; Mangaia maro, “hard and dry”; Mangareva maro, “hard”)? Mr. Elsdon Best gives the Maori tau maro as “a year in which there are no crops.”
The word tataro in Mota means “a prayer.” Dr. Codrington writes thus of it: “The tataro of the Banks Islands, which may be called ‘a prayer,’ is strictly an invocation of the dead, and is, no doubt, so called because the form begins with the word tataro, which certainly is the ‘ataro of San Cristoval—that is, ‘a ghost of power.’ The Banks-Islanders are clear that tataro is properly made only to the dead, yet the spirits (vui, Qat and Marawa) are addressed in the same way.” Ataro means “a god” in San Cristoval (tidalo in Florida, tidatho Ysabel, tida'o Guadalcanar). “The soul” in Malaita is akalo. In Samoan tatalo is “to pray” (Tahitian, tarotaro; Hawaiian, kalokalo; Tongan, talo-monu, “solicit by actions the favour of the gods,” tolotalo, “cast lots”). Apparently the meaning of taro was “a spirit or ghost.” A man's spirit was perhaps thought to pass to the sky; he was no longer a ta-nun, a living man, but a ta-mate, a dead man, or a ta-taro, an inhabitant of the heavens; so in Florida he was no longer ti-noni, but ti-dalo. In the same way the soul was called akalo, just as ata in Samoan means “a spirit” or “the light”; Tahitian ata, “a cloud, a shadow, a certain prayer”; Tongan ata, “the air, free, the morning light, a shadow, reflection”; Maori ata, “early morning, the soul, a reflected image”; Motu (New Guinea) vata, “a ghost”; Mota atai, “soul.” If tataro meant originally “the spirits of the dead,” it might easily be made to include other inhabitants of the heavens; or, if it meant “the gods,” then, when ancestor-worship grew up, the term would come to include the ancestors of the tribe.
The root bara or para seems to mean “crosswise, lying across.” Other full forms of the root are pala, bala, vala, fala. From these we should expect to get ara or ala; pal, par, &c.; papa, baba, &c.; and transposed forms such as rapa and lapa. As a matter of fact, all these forms are found.
The way in which the meaning becomes modified is quite intelligible: from “crosswise” we get such ideas as—an angle, a ladder, tongs, a fence, a barricade, the shoulder, layers or strata in rocks, and hence flat slabs and anything flat, or, with a slightly different advance of meaning, the sloping side of a mountain, or an axe where the blade is not only flat, but is set sideways, crosswise, unlike the adze. With these preliminary remarks as to the meaning, we may follow the changes of form.
In Wedau, New Guinea, bara is “bent”; barabarana, “a bend or angle”; barabara, “a shelf, a wall-plate.” In San Cristoval “a fence” is bara; in Saa Malaita it is para; while para'a is “the side.” The Fiji baravi means “the sea-coast, the side of an island or mountain”; Maori para-hi, “a steep slope or acclivity.” The Mota parapara is “an axe,” a recent use of the word, according to Dr. Codrington. The Formosa parai
means “to fence”; the Malay parai “to go in zigzags, as a ship tacking.” An island in the Banks Group is called Ure-parapara, “the island of steep slopes”: the sea has entered the crater, the steep sides of which give the island its name. The Malay parang is “a chopper or chopping-knife,” or to cut with one; and this is the Maori para, “to fell trees or cut down bush.” In Mota, para also means “to go off at an angle”; sus para is “to crouch aside.”
Pala means in Mota “to set stick across stick, to set across, wattle”; pala ta vava, “to strangle a man by pressing a stick across his throat”; i-pala, “tongs”; palapala, “scissors” (a recent word). In Florida palapala is “a ladder.” In Malay palang is “a cross-bar or transverse beam.” In Santa Cruz pala-po is “lightning” (po meaning “downwards” and also “red”). In Tonga palaa means “a piece of a reed fence”; while in Samoa pala'au is “a wooden fence.”
In Mahaga Bugotu babala is “crosswise”; gai babala, “across.” Vala means in Mota “the fence of small stones round an oven.” In Madagascar vala is “a wooden fence or partition.” The Efate (New Hebrides) vala means “a ship's yards,” because set crosswise; while falafala is “a ladder,” which is made by fastening cross-sticks to a tree. Vala in Florida means “the shoulder”; varat in Mota “the purlin of a house.”
Leaving now the full root, of which many more examples might be given, we come to ara and ala. In Mota ara means “to keep off,” while ge-ara is “a fence.” The Saa Malaita ala means “the shoulder.” The Maori arai means “a veil, screen, or curtain; to block up”; Mangaian arai, “to ward off”; Tahitian arai, “to interpose, obstruct”; Hawaiian alai, “to obstruct, to block up a door or passage by sitting in it, to form a circle round a person for defence, to defend.” The Torres pi ala is “a fence round a garden.”
By dropping the final vowel we get par and pal. Par means in Mota “to slice, cut,” as in Par mal, the name given to a class of secret societies the members of which were wont to par a mal or young cocoanut and drink the milk in common, after which they were accounted brethren. Pal in New Britain means “a room” (just as niu in Mota means either “partition” or “room”); in Duke of York Island, “an outhouse”; in Raluana (New Guinea), “a house.”
Dropping the second syllable we get pa and ba, exceedingly common and important forms. A few examples must suffice. In Fiji bai means “to fence round a town or garden,” while ba is “a fish-fence.” In Maori pa means “to block up, obstruct; a fort or stockade, a weir for catching eels, a barricade; to protect”: Samoan, pa, “a wall”: Tahitian, pa, “a fence or hedge”: Hawaiian, pa, “hedge or fence in; the wall of a town”: Paumotu, pa, “a rampart or bulwark.”
The reduplication of this gives us papa and baba. In Malagasy baba is “a wall or fence in fortification”; Formosa, babas, “an earthen dam”; Tahitian, papani, “to block up”; Mota, paparis, “wall of a house”; Maori, papa, “to close up or fasten; the layers or strata of rocks.” It is from this last that the idea of a slab may perhaps be derived, and so papa or baba commonly means “a slab, board, anything flat.” In Wedau, New Guinea, baba means “slab, side of big canoe”; babai, “to build up with slabs”; babana, “canoe built with timbers”; Maori, papa, “anything broad or flat—a slab, board, door, or shutter”; Samoan, papa, “board, floor-mat”; Tahitian, papa, “a board, seat, the shoulder-blade”; Mangareva, papa, “foundation”; Motu, New Guinea, papapapa,
“flat rock”; Celebes, papang, “a board”; Malay, papan, “a plank or board”; San Cristoval, paparagan, “surf-board”; Malaita, baba, “surf-board.” Papa means also “the earth” in Maori, and no doubt the Maoris conceived of the earth as flat, as our own forefathers did; this word, however, may be from a root meaning “soil.”
Taking last the transposed forms rapa and lapa, we get the Maori rapa-rapa, meaning “the flat part of the foot”; Hawaiian lapa, “a ridge of land between two ravines, steep side of a ravine, having a flat or square side”; Samoan lapa, “flat”; Mangareva raparapa, “flat”; Paumotu rapa, “flat blade of paddle”; Motu, New Guinea, ilapa, “a sword”; Mota irav, “a board, slab of wood in canoe or house”; lapwai, “the flat of a blade, tail of eel.”
These roots are merely given as examples of the working-out of certain principles of comparison. Other roots might be easily given from which a still larger number of words are derived. But the following of such principles as those given above should make comparisons at once easier to obtain and more likely to be correct.
Confusion is especially likely to arise when the first syllable only remains, or where it has been reduplicated. For example, the Maori ta has many meanings, because it is derived from a number of different roots. Papa may be derived for patu (a stone or rock), para (across), para (sediment, dirt, &c.); rara from rana, raho, &c.; roro from roto, rongo, &c. If comparisons are to be sound, they should always have regard to the root-form and its meaning.
The study of the Oceanic languages should throw more light on the general problem of language than the study of the Aryan family or the Semitic, because the problems to be solved are less complicated, owing to the fact that the peoples have been living isolated for so long, and have been unaffected by civilisation. When the Oceanic family has been carefully studied, comparisons may be made with Aryan and Semitic languages. Apparently there were true Aryan and Semitic words in the original Oceanic language.