THE ORIGIN OF THE OCEANIC LANGUAGES

THE ORIGIN OF THE OCEANIC LANGUAGES

BY P. A. LANYON-ORGILL

AT THE end of his chapter on “The Comparative Study of Melanesian” Ray (M.I.L. p. 25) cites an article by Professor D. van Hinloopen Labberton entitled “Preliminary Results of Researches into the Original Relationship between the Nipponese and the Malay-Polynesian Languages”. In the article Labberton says nothing of the connections of the Oceanic tongues and Japanese, for, as he himself writes (p. 280), “these preliminary researches have had, and could not but have, an analytical character”. He also remarks that sections 6 and 7, in which he will expound his theory, will appear later. It is apparent, however, that they did not ever appear, for the Journals of the Polynesian Society from 1924 to the present year are devoid of any such concluding article and the Presidential Remarks at the Annual Meeting of the Society held on August 21, 1933, state that “the matter has not been followed up” (J.P.S. 42 (1933) p. 235).

Although Labberton has not put forward a theory of his own with supporting evidence he has made a number of important points, which will provide us with a very sound basis for the present paper. Above all he casts aside the possibility of a “relationship between the Aryan or Indo-Germanic and the South Sea or Malay-Polynesian languages”. He notes the chief expounders of this theory, Karl Wilhelm von Humboldt and Franz Bopp, but omits to consider other writers. This question of the possible relationship between the Oceanic and Indo-European families cannot be dismissed as lightly as all that, however, for it has received the attention of many outstanding scholars.

An examination of the vocabulary of the chief Indonesian languages reveals a certain content of Sankrit, Sir Richard Winstedt (pp. 21–22) remarks that “the oldest foreign loan-words in Indonesian languages are Sanskrit, which are found pre-eminently in Javanese, and then in Malay, but also as far afield as Celebes, the Philippines, and Madagascar. The borrowing includes not only words of religious, moral, and intellectual ideas, but some astronomical, mathematical, and botanical terms, a court-vocabulary, and a large number of everyday words”. He also notes a few words borrowed from Persian and Hindustani (p. 23). With regard to Java we may note the words of Marsden (p. 21), written over a hundred years ago: “Of the other Polynesian dialects, the most highly cultivated and most distinguished for its literary compositions is that of Java or jawá; a superiority for which this fine island is indebted to its having, for several centuries, enjoyed the advantage of a regular government, as well as to the adventitious circumstance of its having, at an early period, become an asylum for expatriated Hindús, chiefly of the Buddha sect, but also, at a later, by those of the Brahminical or orthodox class. Of these strangers, the influence is shewn, not only by the existing ruins of temples dedicated to their religious worship, but by the communication of much of their language, and the translations made of their most popular works”. Labberton (p. 247) particularly stresses the import- - 26 ance of Javanese, for he notes that “we possess samples of Javanese in writing, covering a period of over a thousand years, not only in short stone and bronze inscriptions, but a rich prose and poetic literature has been handed down”. In the second volume of his great work K. W. von Humboldt makes frequent references to the Sankrit origins of the Kawi words, and Brandstetter's twelve Indonesian monographs contain numerous references to Sanskrit—in no. 6, for example, he quotes from the Mahabharata at least four times (pp. 17, 37) as well as from the Ramayana (p. 17).

Malay and Javanese are not, however, the only languages in which this Sanskrit element has been traced. Minguella, in his 31-page monograph published just over half-a-century ago, made a careful comparison of Sanskrit and Tagalog, the chief idiom of the Philippine island of Luzon. Both Kern and Blake have similarly discussed the Sanskrit element in Tagalog in articles, and the latter, in addition, notes in his detailed grammar of the language (p. 1) that “Tagálog has been influenced to some extent by the languages with which it has come in contact. First in some prehistoric period it was subject like most of its sister tongues to a strong Sanskrit influence, which resulted in the borrowing of a considerable body of Sanskrit words, e.g., mukhâ ‘face’ (San. mukha), bása ‘read’ (San. bhâsâ ‘language’), salitâ ‘story’ (San. carita ‘done’), ása ‘hope’ (San. ácâ)”. He also notes that there have been borrowings from Spanish, Chinese, and, more recently, American-English.

It will be seen, however, that neither Winstedt nor Marsden, Kern or Blake ever suggested that these were any more than ‘loan-words’. The importance of this fact cannot be over-stressed.

In 1896 John Fraser wrote an article entitled “The Malayo-Polynesian Theory” in which he compared a number of Maldivian (an Indonesian language) words taken from a vocabulary by François Pyrard (1611) with their equivalents in Modern Sinhalese, Pali, Sanskrit, and other Indo-European, as well as some Oceanic languages. Some of his comparisons are strange and, indeed, interesting. The Maldivian udu, meaning ‘heavens’, he compares with the Sinhalese udā, with the same meaning, the Pali uda-kam and Sanskrit uda-m, both meaning ‘water’ and the Oceanic words for ‘rain’ (p. 93). Of words in which he finds close parallels between the Maldivian and Sinhalese we may note the following, adding a summary of his Oceanic examples (which may be discovered in Codrington's M.L., etc):

	Maldivian	M. Sinhalese	Oceanic
sin, wickedness	papa	pāpa	Malay pāpa
death	mare	marana	mati
tree	gats	gas	N. Heb.	kasu
pig	ūr	ūra	Melanesian
blood	le	lē	Melanesian
foot	pae	paya	Samoan vae
house	gué	gé	Efate, N. Britain

Fraser gives us a new angle on the question of borrowing when he summarizes his arguments in these words (p. 99): “My discussion of the Malay-Polynesian theory has hitherto centred on the correspondence between Malayan and Polynesian words merely, for it was on this ground alone that K. Wm. von Humboldt first advanced that theory, and with the same arguments others have since maintained it. But it is clear to me, and I hope to my readers also, that these arguments can be proved to be fallacious. Francis Bopp and others - 27 were of that opinion long ago, and asserted that many of these resemblances came from India, but through the Malays. On the contrary I assert that these words did not come through the Malays, and that the Malays have nothing whatever to do with the formation either of the physical frame of the brown Polynesians or the structure of their language; that the Malays are the borrowers, and that, on their first landing in the Indonesian islands, they found the ancestors of these Polynesians there, and gradually adopted their language”.

S. E. Peal's comments on Fraser's theories are of no interest. He cites the unknown researches of J. R. Logan but gives no real philological evidence. The weaknesses of Fraser's paper are obvious. The very proximity of the Maldive islands to the Indian shore renders their language more or less useless for comparisons of this sort.

In 1913 Christian tried to re-awake interest in the Indo-European theory by quoting nine Sanskrit words and supposed Eastern Polynesian equivalents and adding that Maori scholars “will welcome with interest the probable [misprinted ‘probably’] ethnological and philological connection of Arahura, the name of the famous canoe of the migration with alfur, the name of an aboriginal tribe in the Moluccas, and Arafura the geographical name of the tract of sea that washes Timor and that portion of the Malay Archipelago that lies north of the Gulf of Carpentaria” (p. 79). This effort on the part of Christian called forth a piece of strong criticism from the pen of the great S. H. Ray. He rightly corrected Christian on the word alfuru, noting it to be “a Portuguese application of a native term meaning ‘bush men’ (in the Tidore dialect of Halmahera, hale land, furu wild)”. His mode of dealing with Christian's actual examples is, to say the least, sweeping. Christian (p. 78) had the following:

• “Sanskrit: bhek, toad, frog
• Persian: ghek, toad, frog
• Inca (Peru): bhek, loathsome, disgusting
• Inca (Peru): uekka, filth
• Java: beka, filth
• Micronesia: bek, filth
• Maori and Polynesian: wheke, heke, feke, cuttle-fish, originally ‘The Loathsome or Foul One’”.

Of this example Ray wrote: “Words like bhek, ghek given for ‘frog’ are plainly onomatopoeic, and show no more evidence of connection between Sanskrit bhek, frog, and Maori wheke, cuttle-fish, than do similar words in other languages. (Cf. Papuan (Toaripi pakeke), African (Lunyoro ekikere, Sechwana segwagwa), Turkish (kurbagha), Garo (bengbek), Khamti (khet), Siamese (khamkhok), all meaning ‘frog’, or the American (Abenaki moskeke) word for ‘toad’).” At the beginning of his short paper Christian refers to “an early Arab and Persian influence, which has tinctured Eastern Polynesian languages quite strongly, probably even to the extent of one-fifth of the words”. This fantastic remark leads us to consider the theory of the Semitic origin of the Oceanic languages, also successfully attacked and swept to the ground by Professor Hinloopen Labberton. 1

The most celebrated supporter, if not the only one, of this theory was Dr. Daniel Macdonald, who put forth his views on the subject in a series of articles: in 1896 when he compared the “Numerals, Personal Pronouns, Phonology and Grammar”, in 1901 the “Formative Suffixes”, - 28 and in 1904 “Triliteralism and Internal Vowel Change”. These, and other materials, were embodied in The Oceanic Languages in 1907. In that book he showed, to his own satisfaction at least, that “Arabia is the motherland of the primitive Oceanic, as it is of the Ethiopic, Amharic, and Tigre, and of the Assyrian, Phoenician, Hebrew, and Aramaic” (p. 94). He informs us that “the primitive Oceanic must be regarded, not as a descendant of, but as a sister to the Arabic, Himyaritic, Ethiopic, Assyrian, Phoenician, Hebrew, and Aramaic, and the Efate, Samoan, Malagasy, Malay, etc., as cousins to the Mahri, Amharic, Tigre, Mandaitic, Modern Syriac, and vulgar Arabic dialects, due allowance being made for the fact that these latter have been always more or less under the conserving influence of the surrounding Semitic literature and civilization, from which the Island dialects have been for ages completely cut off, as well as completely isolated from each other” (p. 94). Ray (MIL. p. 197) remarks that the Efate words in the dictionary have been “distorted into a Semitic mould”. In the second chapter of The Polynesian Wanderings Churchill critizes the work severely; for example, he writes (p. 8): “But when the definition begins to squint at something ulterior, when a word is added with the sly insinuation of a purpose to link the fact and the theory more tightly, when definition becomes polemic, such exhibitions I do not hesitate to stamp as scientific dishonesty; they prove an absence of conscience, without which speculation is mere trickery held in check only by consideration of the risk of discovery”. [In Ray's own copy the word ‘dishonesty’ is underlined and the comment ‘Right’ is in the margin.] In closing his criticism, however, Churchill (p. 11) writes: “Imperfect as it proves itself to be, Dr. Macdonalds dictionary is amply the most valuable contribution to our knowledge of any speech of Melanesia”.

Of Macdonald's theory little can be said. In the first chapter of his book he goes so far as to write of “the commercial fleets of Solomon” (p. 5) and quotes Josephus, “than whom there is no weightier historical authority on the subject”. Collingridge, in 1896, wrote a short paper in which he considered the question of whether the Arabs were acquainted with New Zealand or not. The subject of his argument is a map by the Arabian geographer El Edrisi (A.D. 1160). The map is described in these words (p. 124): “The circumfluent ocean surrounds the world as in the very earliest known representations of the world. The furthest country towards the east in the southern hemisphere is Ouak-Ouak, which, being in proximity to Sofala, would correspond to the South of Africa. Amongst the large islands in the Indian Ocean, Serendib corresponds with Ceylon, Ramy probably with Sumatra, and Malay with the Malay Peninsula, although represented as an island. The islands to the east of Malay must be intended for the various islands known to the Arabs, such as the East Indian Archipelago, and the Spice Islands”. The other side of the argument rests entirely on the supposed relationship of the names of two large birds, both now extinct, the Arabic Seêmoah (Persian Simurgh or Simorg) and the Maori Moa. On the whole Collingridge seems to think that the Arabs did not know of New Zealand.

Churchill criticized Macdonald's dishonesty in respect of the meanings which he gave his words. It would appear, however, as if the work should be criticized rather on the grounds of the spelling of both the Efate words and their supposed Semitic uncles and brothers. Captain Rason wrote a letter to Churchill (P.W., p. 11, fn.) which was apparently misread by Ray (MIL., p. 197). In this letter Rason - 29 writes: “When the missionaries established themselves on Efaté he [Dr. Macdonald] was in Havannah Harbor, and natives who first became Christians left their villages and came to the mission station for protection. Thus the language of the mission station became a medley of all the dialects around. This gradually coalesced into a special dialect which became a lingua franca with the natives and was partially understood by all. As the heathen natives died out or became Christian the mission language was claimed as the language of the island. Then the Bible was translated into this language and Dr. Macdonald wrote a dictionary of it as if the missionary language was the original language of the various villages before they were Christian”. [Ray's own copy has the note ‘Not Bible but NT’ in the margin.]

Thus the Efate words are of a polyglot nature. In addition to this, however, Macdonald has employed a system of representing both Semitic and Oceanic sounds by a personally constructed alphabet, and in so doing he is able to make dissimilar words resemble each other by implication. If we care to examine, for example, the Ethiopic alphabet as put forth by Macdonald (p. 9) with that given by authoritative Semitic scholar like Mercer (pp. 10–11) we are immediately struck by the forced nature of Macdonald's symbols. In carefully working out a comparison of the Arabic and Efate alphabets (pp. 10–31) we might refer Macdonald to that now-classic article by Prof. Seidel on Das arabische Element in Suaheli (Z.A.O.S., i, 1895, p. 100) in which the writer shows that every letter but two in the Oman dialect has a parallel in Swaheli. From this, no doubt, Macdonald would be prepared to deduce that the Bantu-Sudanese languages also had their motherland in Arabia.

Other writers have noted a number of Arabic loan-words in Oceanic tongues. Winstedt (p. 23) remarks that those in Malay “deal especially with the sphere of religion and law” and enumerates a dozen which are fairly common. Marsden (p. 23) quotes Crawfurd as saying that “the Arabic language has made but small progress in Java, although the religion of the Korán has been professed there for some centuries, and in the schools a smattering of it, with a religious view, is the only branch of instruction, Javanese literature itself being nowhere taught as a branch of education, but left to be picked up as occasion offers. The number of Arabic words introduced into the language is consequently extremely small, greatly smaller than into any other of the more cultivated languages of the archipelago”. The only truly Semitic feature about Malay is the alphabet, but if the readers care to turn to the two engraved plates in Marsden's book (pp. 116–117) he will find “Alphabets of Hither Polynesia” for the “Korinchi of Sumatra, Rejang, Lampung, Batta or Batak and Bugis or Ugi and Makassar” which bear not the slightest resemblance to any other alphabets. In any case we may note that the Malayo-Arabic alphabet contains thirty-three letters as compared with only twenty-eight in the ordinary Arabic. Of the five additional characters, two are borrowed from Persia (which has thirty-two letters) or from Hindustani (which has thirty-five letters) and three are Malay inventions. In addition Malay has a symbol which is in reality a combination of two letters (lam alif) and a special character to represent “a cerebral d found in Sanskrit, Javanese, and Madurese, though not elsewhere in Malayo-Polynesian languages; it is exotic and not employed now” (Winstedt, p. 36). The Arabic characters have been adopted only since the coming of Mohamedanism (Winstedt, - 30 p. 37). The same state of affairs apparently obtains in other parts of Indonesia, for Cowie (pp. ix–x) notes that the Sulus “have adopted the Arabic character with certain modifications. They, however, always employ bāris or signs to assist them in distinguishing their vowels, whereas the Malays generally leave them out”. In respect of Tagalog Blake has been able to write two pages on “Analogies between Semitic and Tagalog” but nothing more. Parker, in writing of Malagasy (p. iv), notes that “the influence of the Arabs is seen in the names of the days of the week, the Hova names for the months, and in many terms connected with dress, bed, money, musical instruments, etc.”

In considering the origin of the Oceanic languages it is certain that we may confine our attention to the continent of Asia. Macdonald, in one of his saner moments, writes (O.L., pp. 2–3): “Three great Continents, Asia, Africa, and America, or, counting Australia, four, border on the Island world. It may be held as certain that the relationship of the Oceanic, whatever it is, is not African, American, or Australian. Madagascar is near the African coast, but the Malagasy, which belongs to the Malayan or Tagalan branch of the Oceanic, is not related to the African languages. Easter Island approaches nearest, though not very near, to America; but its language, which belongs to the Polynesian branch, is not related to the American languages. And the Melanesian branch, which approaches Australia, is not related to the Australian languages”. Over a hundred years ago Marsden quoted (p. 92) two “Hottentot” vocabularies by N. Witsen and J. Barrow respectively to show that Malagasy is in no way related to the neighbouring African tongues. A cursory study of the chief languages of northern Mazambique and Nyasaland leads us to the conclusion that there is not the slightest possibility of relationship between Malagasy and, say, Makua, Yao, or Nyanja, or any other of the Bantu languages as shown by Sir H. H. Johnston.

Marsden also quotes (p. 104) three American vocabularies (Araucanian, Quichua, and Maya-Puctune) to show that they have no possible connection with the neighbouring Oceanic language of Easter island or any inter-relationship between themselves. It is indeed strange that the native languages of the American continent should present such a series of diverse forms. A glance at the numerals reveals a large number of different systems of counting. In some languages, for example, Cree (Hudson Bay) and Kaugian (N.E. Colombia) we find short, well-developed forms, but near them we find tongues like Tukudh (N. Canada) and Betoya-Tukano (N.W. Brazil) with complicated and unwieldy forms. In his article on Numeration in the Encyclopoedia Britannica (14th ed., vol. 16, p. 614) Padre Schmidt notes that the Quaternary system of counting (5=4+1, 8=4+4, etc.) is found only in the Salina and Chumash languages of California. Arawak uses a quinary system, Maya and Tukano an imperfect decimal; Cree, Kaugian, Quichua, and Araucanian a perfect decimal; and Camacan (Bahia, Brazil) a modified dual (like that commonly found in Australian languages). Why there should be this diversity is a problem which does not concern us, but we can definitely state that no American language has yet appeared which bore the slightest resemblance to any Oceanic tongue from the morphological point of view.

In his article which appeared in 1913 F. W. Christian repeated his remarks on “the purely Sanskrit origin of the word kumara, the original kumad, the Edible Lotus, and passing through the crucible of the Hindu-Malay of Sunda (N. Java) kumadjang, and of the Philippin- - 31 Malay kamote, and the Rúk (Central Caroline) kamal, into the Polynesian kumara and kumala, and leaving in the Quichua and Aymara dialects of Peru kumara, the choice white potato”. In 1932 the same author showed that the Chimu and Inca languages contained a number of common words, almost certainly of either Sanskrit or Oceanic origin. Mr. Christian shows that the following Chimu words can be compared with forms in Polynesia and Micronesia, as well as in Malay and Sanskrit (sic.) (pp. 144–146):

• faoa, smoke
• pong, stone
• ssi, ssu, water, rain, a lake
• nass, clean, pure
• pess, bad, evil
• cubun, a word
• fitillko, fisillko, bowels
• pelak, testiculus
• sllyang, syyang, sun
• si, moon
• ava, maize
• mach, mati, hand (Inca maki, Maya mash)

He even goes so far as to compare si with the Babylonian sin and ava with the Sanskrit yawa (‘barley’) and Japanese awa (‘millet’). He also finds almost identical forms in Malay for the following Inca words (p. 146):

• senka, nose
• simi, mouth
• siki, siqui, hinder parts
• tullu, bone
• puchu, tail

Mr. Christian also finds parallels in names of plants in Araucanian and in some unnamed tongues spoken in the basin of the mighty Amazon. It should be observed, however, that Mr. Christian merely suggests that South America is the goal of a great migration, not a source or origin.

Of the possibility of the Australian connection nothing need be said. The hopeless lack of form and generally debased aspect of the languages of Australia and non-Melanesian New Guinea is especially apparent in such works as the translation of the Gospel into Arunta (Ewangelia Taramatara, London: B.F.B.S., 1928). Seligmann summarizes this side of the problem when he writes (p. 4): “Many of the Papuo-Melanesians (including all the Massim) speak languages with a common Melanesian grammar. These languages are divisible into groups, the constituent languages of each of which contain numbers of common words, all related to the common stock language of Oceania. With the Papuan languages it is otherwise; with regard to vocabulary they present a number of apparently unrelated stock languages, while of their grammar it can only be said that though a number of them conform to certain rules, it is clear that none of them present Melanesian characters. A great difference is also seen in the system of enumeration of the two races. While the Papuo-Melanesian counts easily by the quinary or vigesimal system, the Papuan has only two numerals, one and two, and counts with certainty only up to five or six by combining these”. A study of the Papuan vocabularies in Ray's book on the British New Guinea dialects, of the comparative vocabulary in Hanke's fine study of Bongu, of Saville's notes on Mailu (which attempt to place a debased tongue in an elevated position, e.g., by quoting some fifteen tenses of the regular verb), of W. H. Bird's careful notes on Chowie, or, especially, of Padre Wilhelm Schmidt's article, Die Gliederung der australischen Sprachen, leads one to the absolute and final conclusion that the Australian is the most impossible of all possible origins of the Oceanic family of languages.

It is apparent, therefore, that the origin of the Oceanic languages is to be found in Asia—southern Asia, in fact. The south-western - 32 parts of that great continent have been shown to be merely sources of a few loan-words. The same applies to the Indo-European part of India. The motherland must, then, be either in southern (Dravidian) India or in south-eastern Asia. Both of these are Turanian or Mongolian regions, and most scholars believe that the solution to our problems lies in some part or other of that region.

In 1904 Edward Treagear stated (p. 105) that the “land of ultimate origin [of the Polynesians] was probably in South-Central Asia, but it may have been in Lithuania, or by the shores of the Caspian Sea; wherever it ‘may have been’ it was, as I believe, in that locality wherein those branches of the Indo-European family now occupying North-western Europe had their birth”. In 1885 the great Dr. Codrington in considering the suffixed personal pronouns of the Oceanic languages, wrote (p. 127): “From whatever region, by whatever routes, they have reached their present seats, it is evident that these Pronouns were among them before they parted. This is a point of very great interest, because of the Pronouns of Khamti, one of the Tai languages of the Asiatic continent, kau I, mau, thou, man, he. If it be supposed that ku, mu, na [the Malay forms], have come from the continent of Asia, from the valley of the Irrawady, into the languages in which they are now used as suffixes, these Pronouns must have come into them while still undivided”. We thus see that the one occasion on which Codrington made any comment on this subject he suggested a Taic origin.

To return for a moment to Labberton's paper, we should note that he opens by confirming the unity of the Oceanic family (pp. 246–248) as also does Ray in considering “The Common Origin of the Oceanic Languages”. The latter states (pp. 58–59) that “four propositions may be made as to the relationship of the proper Oceanic languages to one another:—

• 1. That the vocabulary shows evidence of a common origin.
• 2. That apparent differences in the grammar are modifications of the same method rather than actual differences of structure.
• 3. That the principal constructive particles are the same.
• 4. That the languages are in various stages, of which the Polynesian is the latest”.

In the course of noting the unity of the family Labberton makes this important and far-reaching statement: “The investigations of Padre Schmidt, published in Anthropos and elsewhere, have established the relationship of the Island tongues with the monosyllabic Mon-Khmer of Kambodge and the Santali and Munda-dialects beyond the Himalayas”. Apart from making a grave geographical blunder, Labberton has apparently thought it fit to decide that Padre Schmidt's conclusions were conclusive, and, at the same time, to suggest that the real solution to the problem lay many hundreds of miles to the east! In 1906 Schmidt wrote a monograph entitled “Die Mon-Khmer-Voelker, ein Bindeglied zwischen Voelkern Zentralasiens und Austronesiens”. This challenging title, together with the typically brilliant quality of the work, led to its translation into half-a-dozen or so languages. The work speaks for itself and we need not quote examples or other points.

We thus see a distinct tendency for scholars to turn their gaze to south-eastern Asia, notably to Burma and Cambodia. Dr. H. L. Shapiro, writing in Oceania (3 (1933) p. 367), states that “the general opinion of scholars and the dictates of common reason agree that the Polynesians embarked on their Oceanic odyssey from some point in south-eastern Asia”. But the Turanian area is a vast one, and - 33 includes many distinct types. In Siberia and northern Russia we find tongues like Samoiede, Ostiak, and Votiac which retain many of the features (e.g., a dual number and double forms of the pronouns) which we associate with primitive languages. Closely related to them are languages like Finnish, Estonian, and Magyar which are more advanced. In the ordinary Turanian language, Chinese or Burmese for example, the suffix remains a single syllabic form, but in Finnish that suffix has become an ending to be attached to a root. Finnish and Estonian therefore show a form of noun-declension, such as we also find in most of the Indo-European languages. In this way it might appear as if Finnish were the missing link between the Turkish and Samoiede on the one hand and the Lithuanian and Sanskrit on the other. This view was first put forward by Penka in 1883 in his spirited Origines Ariacae, and was developed by Weske in 1885 in his Uber die historische Entwickelung der finnischen Sprachen in Vergleich mit der Indo-germanischen, which is now assured a place among the great classics of philology.

In eastern Asia we find many and varied tongues, all obviously members of one family at the base, but differing widely in their grammatical structure and vocabulary. In attempting to decide the origin of the Oceanic languages we must keep the various possibilities in mind and attempt to decide which is the most practicable and the most probable. We cannot hope, initially, to formulate rules of development or draw up detailed comparative vocabularies, but must be satisfied with more general statements and considerations.

In 1906 the editor of the Korea Review, Dr. Homer Hulbert, wrote a short work in which he showed that Korean had certain definite affinities with the Dravidian languages of India. His views on this subject do not concern us here, but the vital importance of the book lies in the first four appendixes. In the first of these entitled “Korean and Efate”, the author shows definite relationships between about 130 words in the two languages with many additional words from about twenty other Oceanic tongues. Before giving this comparative vocabulary Hulbert remarks “that the phonetic systems of the two are very much alike. In each we have the continental vowel sounds of a, e, i, o and u. In each there is but one character for b and p. In each the k, l, m, n, r, s and t are sounded as in English. There are three differences. The g of Efate is pronounced ng as in certain parts of Japan, and except in one of its dialects the letter h is not found, its place being taken by s. But in Korean the letters h and s are very often confounded. . . In Efate we find the letter f” (p. 125). In the second appendix Hulbert quotes a “Comparative Vocabulary of Ten Savage Groups in Formosa”. Out of a vocabulary of fifty words he finds “fifteen in which a distinct similarity [to Korean] can be traced, and in not a few of the fifteen the similarity amounts to practical identity”. When the distances involved are considered we need no longer question at least the partial validity of Hulbert's theories. His other two appendixes are of an anthropological nature and do not directly concern us here.

An examination of Hulbert's Efate words shows them to be taken from Macdonald's dictionary (which had been published separately a few years before), as also are the parallel examples in other South Sea languages. We have stated previously that Macdonald's system of orthography in the Efate is unreliable, but it is not so unreliable as to make Hulbert's parallels any the less convincing. In any case, Macdonald's examples from other languages are in a more consistent - 34 form. From Hulbert's Korean vocabulary we may abstract the following words, the Oceanic equivalents of which reveal identical or practically identical forms:

• ab-i, ab-a-ji, father
• e, to, at, in
• ham-ham, smooth
• i, this
• iri, here
• kă, dog
• kariawa, to itch
• keu, that (near)
• k'o, nose
• komi, spider
• kue, crab
• kŭt, thing
• man, alone, only
• man, many
• mogi, mosquito
• na, I
• năamsă, a smell
• nal, sun
• nani, which has been born
• ne, yes
• nye, yes
• pi, to rain
• po, to see
• pu, to blow
• pu, to swell
• pui-ta, to be empty
• pul, fire
• sal, arrow
• ta, to pluck
• tang, earth, ground
• tokenui, axe
• toki, axe

We thus find for the first time a group of languages which shows some possibility of relationship with the Oceanic. But Korean is but one member of a large family of tongues spread over half a continent. We may indeed ask ourselves if the Oceanic languages are connected specifically with the Korean. Hulbert has shown us the family but Korea is unfortunately off the main track of the great migrations. It is rather a destination than a temporary halt. If we consider the probable movement of tribes from the upper Brahmaputra valley—the generally accepted cradle of the Turanian races—we are forced to think of a movement into Burma and south-eastern Asia, coupled with a dual development in the direction of the Pacific islands to the east and south-east and to Formosa and Japan to the north-east. The core of the whole system, the focal point, therefore lies in the islands around Formosa. The native dialects of Formosa as shown by Hulbert in his comparative vocabulary, and more especially the Favorlang dialect as put forth by Happart are obviously Indonesian, as also are the dialects found in the Botel Tobago (Koto-Sho) and Bashi or Batan islands lying to the south, whilst the tongue used in the Ryukyu or Lu-chu islands to the north is equally obviously a pure Turanian one. Luchuan possesses all the characteristic elements of Ainu, Japanese, and Korean, as well as those of Mandarin Chinese to a lesser extent—no distinction of gender or number, the Isolated Case, pronouns which are no more than simple nouns, a highly complicated verb-system, and the presence of long vowel sounds of phenomenal length. Luchuan and Japanese are linked by a set of regular phonetic laws, stated by Chamberlain (p. 49) to be as follows:

• Juchuan: I U CH J F Y N
• Japanese: E O K G T R M

Unfortunately our materials in Luchuan are very meagre, but we gain a very definite impression from Chamberlain's notes: that Luchuan is a much abbreviated tongue. Keeping the words given by Chamberlain in mind we may turn over the pages of Happart's Favorlang vocabulary and taking examples completely at random we find the Formosan forms ioa (p. 137), numma (p. 163), and tomma (p. 189) which are most certainly cognate with the Luchuan ya (‘you’), nū ('what?), and tā (‘who?’). The spelling of the Favorlang should be carefully considered, for the vocabulary was originally collected many years ago. The Luchuan word tammē, meaning ‘grandfather’, bears distinct resemblances to the Kiwarawa and Vonum of Formosa tama (Hulbert, p. 135) and the Melanesian tama, all meaning 'father).

A striking parallel, amounting to almost certain proof of relationship, between Luchuan and Formosan is contained in the translation - 35 of the Lord's Prayer into the Tokunoshima dialect of Luchuan made by Tanashiti Baru and printed in the form of a four-page pamphlet in 1899. The present writer has no copy of this but has transcribed the first sentence. It is as follows: “Nama tau tamasa-i biga te-a busu, ypata choa nan”. This single sentence of a Luchuan dialect is almost identical with the opening of the Favorlang translation as given in Campbell's book (p. 1): “Namoa tamau tamasea paga de boesum, ipádassa joa naan”. The difference of spelling (e.g., choa for joa) may be accounted for by the antiquity of the Favorlang version. Favorlang, or Vovorollang (Campbell, p. xvi), incidentally, is spoken around Chiang-hoa. Campbell's book also contains (p. 102) a translation of the Paternoster into the Sek-hoan dialect by the Rev. D. Ferguson, M.A. This dialect is spoken around Toa-sia, fourteen miles north of Chiang-hoa, and also “among the villages of the Pawsia Plain, some two days' journey to the east of Chiang-hoa”. This translation commences with the sentence: “Niam a A-bah kai-dih ba bau ka-wuss”. In this version we find the Korean ab-i, ab-a-ji meaning ‘father’ which is a parallel form to the Malay bāpa and Maori papa, just as in the Favorlang we find tamau, corresponding to the Luchuan tammē, already discussed.

In 1898 Christian (J.P.S. 8 (1899) p. 224 sqq.) presented a vocabulary of the Polynesian language of Nukuoro, spoken to the south of the Caroline islands. In this vocabulary the author connected some of the words with Japanese and Chinese. Ray, in writing his grammar notes on the same language in 1912, commented on these words and noted (p. 166) that “their affinities are obviously with Samoan”. The following are quoted:

Nukuoro	Japanese	Chinese	Samoan
tongo, mangrove	tangara	—	tongo
susui, wet	sui	shui, water	su
tokotoko, pole	—	chok, bambaa	to 'o
mune, ship	fune, junk	—	folau, voyage
sakasaka, dwarf	chiko	—	sa 'a

What affinities the Samoan folau has with the Nukuoro mune is far from clear, but the real value in these examples lies in the fact that they show a Turano-Polynesian relationship. It is not that the Nukuoro is connected with the Samoan and not with the Japanese; it is rather that it is connected more closely with the Samoan than with the Japanese. The importance of the link between Chinese shui, Japanese sui and Samoan su cannot be overstressed, and an additional link with Favorlang maspe, ‘wet’ (Campbell, p. 156) and the “Central Asiatic” su, ‘water, liquid, a lake’ (Christian (1932) p. 144) cannot be ignored. It will be remembered that Christian introduced the Japanese awa (‘millet’) as being cognate with the Sanskrit yawa (‘barley’) and Chumi ava (‘maize’) in the same paper (1932, p. 145). One is inclined to wonder whether the Motu ava (‘weed’), Lau 'aba, Sa'a and Ulawa apa, Bululaha (Mwala) apapaie (all meaning ‘leaf’) also are not connected with this remnant of “a very early tide of migration from Asia”.

Moving to the south-west we come to Indo-China, Thailand, and Burma. Schmidt's paper on the relationship to the Mon-Khmer of Cambodia has already been noted, as also has Codrington's suggestive reference to the Khamti pronouns. All over this region we find Austro-Asiatic languages, ranging from the Dafla, Lepcha, and Naga of Assam and Munda of Bengal to the Mon (Talaing), Annamite and Khmer (Cambodian) of the south. The group also includes Taic (Siamese), Nicobarese, and certain Burmese dialects as well as Sakai, Jakong, - 36 and Semang, three aboriginal tongues spoken in the more mountainous parts of the Malay peninsula. Jakong is apparently a language of a different type from both Sakai and Semang. It is probable that it is a dialect of Malay rather than a member of another family, as can be seen from the numerals as quoted by Marsden for Jakong (p. 87, collected by Raffles) and Semang (p. 133, by James Scott) with which we may compare the modern standard Malay:

Malay	Jakong	Semang
1. satu, sa	sah	sang
2. dua	dua	wang
3. tiga	tigu	wu-ip
4. ěmpat	amput	lemang
5. lima	lima	lepang
6. ěnam	anam	puit
7. tujoh	tuju	sunto
8. dělapan	delapan	lunto
9. sěmbilan	sambilan	lang
10. sa-puloh	sa-pulu	pu-un

Of these various Austro-Asiatic languages we need not consider tongues like Dafla which are spoken so far away from the Oceanic scene. Indeed, we may confine our attention to Nicobarese, Burmese and its associated dialects, Taic, and Khmer. We might, in passing, note that in Munda the word for ‘moon’ is bona and for ‘sun’ is sin—both common Oceanic forms by transposition (Sarat Chandra Roy, The Mundas and their Country, Calcutta, 1912).

In Nicobarese our knowledge is limited to three vocabularies, collected by the Danish Mission, N. Fontana, and G. Hamilton reprinted by Marsden (pp. 113–114), and an incomplete and unreliable translation of S. Matthew's Gospel made in 1780. In the vocabulary of the Danish Mission we find three possibly related words: sualafan (‘nine’), olmat (‘eyes’), and ka (‘fish’) with which we might venture to compare the Malay sembilan, mata, and ikan. Fontana's vocabulary has the equivalent forms encata (1), holmat, and cha or ka. Hamilton has ka. The translation does not yield any material. The “Rules of Pronunciation” at the front are of little use, for words like dewshe (‘God’) and rey (‘king’) are obviously loan-words, and not a single one of the other words shows any possibility of close relationship with Malay or any other Oceanic tongue.

Burmese is not, strictly speaking, a pure Austro-Asiatic language but is rather the link between the Austro-Asiatic and Indo-Chinese families. We may note the numerals in the chief Burmese and Indo-Chinese lanugages and compare them with the Oceanic. The forms are given in standard (Rangoon) Burmese, Barma (spoken to the north-east of the capital), Min-kia (S.W. Shina), Lolo (S. Szechwan), Song-kiang-houa (Kwei-chow and Hunan), Cantonese, Foochow dialect, and the standard Mandarin Chinese.

Burmese	Barma	Min-kia	Lolo
1. tit, ta	tit	a	t'i
2. hnit	hnit	go	n'
3. thone	sông	sa	so
4. lay	lé	shi	li
5. ngah	nga	ngo	ngo
6. hkyouk	khyowk	fou	tch'o
7. hkoo-hnit	khuhnich	ts'hi	cheu
8. shit	shyit	pya	hè
9. ko	ko	ddyou	keu
10. ta-hsai	ta-che	ts'eu	ts'eu

- 37

Song-kiang-houa	Cantonese	Foochow	Mandarin Chinese
1. i-keu	yat	it	i
2. leang-keu	y, ea	shi	eul
3. sè-keu	sam	sa	san
4. se-keu	si	si	seu
5. n'keu	keung	gou	ou
6. lôh-keu	loc	lug	lou
7. ts'it-keu	tsat	chit	thsi
8. péh-keu	pat	pe	pa
9. kieu	kau	kau	kiou
10. zéh	chab	chap	cheu

Of all these forms we can most certainly say that they show no probable relationship with the Oceanic, with the possible exception of the Burmese word for ‘one’. Bearing Jarnsen's maxim in mind, we need not pursue our investigations in this field.

In Taic we see a tongue which is a link between the Chinese dialects to the north and Khmer to the east. It has many affinities with Chinese, as may be seen from the numerals: 1. nung, 2. song, 3. sam, 4. si, 5. ha, 6. hug, 7. ket, 8. pet, 9. kau, 10. sihp. These forms are no more illuminating than the Burmese or Chinese, and we must therefore turn our gaze to the last of these regions, that of Tongking, Annam, and Cambodia.

The Mon-Khmer region was the solution suggested by Schmidt in 1906, but twenty-three years earlier a certain Mr. Keane contributed the following note to the preface of Parker's Malagasy grammar: “The infix syllable om (um, am, om) is a feature which Malagasy has in common with Khmêr (Cambojan), Javanese, Malay, Tagala (Philippine Archipelago), and, no doubt, other members of the Malayo-Polynesian family.

• Ex.: Khmêr: slap, dead; samlap, to kill.
• Javanese: hurub, flame; humurub, to inflame.
• Malay: pilih, to choose; pamilihan, choice.
• Tagala: basa, to read; bumasa, to make use of reading.

Originally a prefix, as it still is in Samoan (ex. moto, unripe; momoto, to die young), this particle seems to have worked its way into the body of the word by a process of metathesis analogous to the transposition common to most languages”. Mr. Keane apparently thought that Khmer was actually an Oceanic tongue.

Since we have taken the most stable of all forms, the numerals, in the other Turanian groups, let us now examine them in the languages of Indo-China. Since we have quoted from Marsden in other languages we may as well use his vocabularies in this instance, which are (1) by Alexander de Rhodes and (2) from the Leyden MS. Consultation of modern standard works like S. Tandart's Dictionnaire français-cambodgien in two volumes, and Pannetier's Lexique français-cambodgien confirms the reliability of these forms. An examination of the two vocabularies shows that in no less than twelve out of thirty-four of the words there are obvious connections with Malay and eight others show probable links with other Indonesian tongues. In the following list the relevant words have been transcribed from Marsden, the Malay from Cowie, and other languages from Wallace, etc.:

• two: Kh. (1) hai, dôi; Malay dua, Sula Islands gahu.
• three: Kh. (2) teng; Malay tiga.
• five: Kh. (1) lam, (2) lang; Malay, etc. lima.
• seven: Kh. (1) that; Mysol tit.
• nine: Kh. (1) (2) chin; Mysol sin.

- 38

• ten: Kh. (1) thap, muoi, (2) tap; Malay sa'puloh, Sula Is. poha, Gilolo megio.
• eyes: Kh. (1) (2) mat; Malay, Celebes dd. mata, Teluti (Ceram) matocolo.
• hand: Kh. (1) (2) tay; Malay tangan, Ahtiago (Ceram) taiimara.
• night: Kh. (1) dhem, (2) dêm; Sulu dūm, Camarian (Ceram) ameti.
• dead: Kh. (1) mât; Malay mati.
• black: Kh. (1) (2) tham; Malay hitam.
• fire: Kh. (1) hoa, lua; Gilolo lutan, Amboyana aow.
• water: Kh. (2) thuy; Rotti oëe.
• earth: Kh. (1) dhat, (2) 'dat; Malay tanah.
• swine: Kh. (1) heo, hoi; Gah (Ceram) boia, Amboyana hahu.
• egg: Kh. (1) tlung; Camarian teruni, Ahtiago tolin.
• fish: Kh. (2) ka; Malay ikan.
• sun: Kh. (1) mat-bloi; Malay mata-hari.
• moon: Kh. (1) mat-blang; Malay bulan.
• stars: Kh. (1) (2) tinh; Matabello toin, Wahai (Ceram) te'en.

We thus see, after examining the possible relationships with all the continents, that the origin of the Oceanic languages lies in one of two regions: Korea and Formosa or Cambodia. The most probable explanation of having two solutions lies in the theory that the origin is to be found in Cambodia and that the peculiar linguistic affinities to be found in Luchuan and Korean are offshoots from the northerly line of the great migration, the Tagalog-Formosan line.

Sufficient evidence has been produced to suggest where the truth lies. This whole theory can only be proved beyond all doubt by the production of a comparative grammar and dictionary of Khmer and the Oceanic languages. The compilation of such a work is already in hand and it will appear as a separate publication at some date in the not far-distant future. This paper will no doubt prove its own usefulness in clearing and surveying the ground prior to the construction of the actual final book.

ADDENDA.

SINCE writing the above a small number of additional points have come to my notice which I think are of sufficient interest to be included here.

In an article on the Melanesian Demonstratives (Bulletin of the School of Oriental Studies, vol. 9, part 2 (1938) p. 405) the late Dr. W. G. Ivens adds the Japanese forms to his list of those in thirty Oceanic languages because, in his own words (p. 397) “this seems to show decided resemblances to certain demonstratives in the other lists”. These forms were supplied to Ivens by the late S. H. Ray and are, unfortunately, attributed with the wrong meanings. On the authority of the best Japanese grammars we may state that ano, given by Ivens as meaning “this”, really means “that”; that are is a noun; that koko means “here” and not “there”; and that sono is an adjective and not a noun. Bearing these corrections in mind we can still see some degree of relationship between these forms and especially those of the Central Solomon Islands, some of which are also shown in Ivens' article. Incidentally, we may wonder why only four demonstratives are quoted in Malay and only two in Sangir (p. 404), yet Cowie enumerates half-a-dozen in Sulu (English-Sulu-Malay Vocab- - 39 ulary, p. xxvi) and H. C. von der Gabelentz (Grammatik der Dajak-Sprache, Leipzig (1852) p. 24) finds as many again in Dyak.

Following on the researches of Pelliot, Shirokogoroff, Kotwicz, Kanazawa, and Matsuoka, S. Yoshitake in his article on the “Japanese Names for the Four Cardinal Points” (B.S.O.S., vol. 7, part 1 (1933) pp. 91–103) has shown the relationship of Japanese to Turkish and many other Altaic languages (cf. Lanyon-Orgill: The Languages of the World (1940) pp. 106–113). He closes his remarks by saying that “a similar comparison may yet be made of Japanese with the Austronesian and the Finno-Ugrian words, when more convincing results may be obtained”. Dr. Jan Lobkowicki has followed up Yoshitake's suggestion in respect of Samoan, Tahitian, and Hawaiian (Die Himmelsgegenden, Krakow (1937) pp. 11–14), Motu (id., p. 25), and Tagalog (id., p. 39). Unfortunately, whereas Yoshitake's explanations were admittedly “50 per cent. doubtful”, Lobkowicki's were 100 per cent. so, not one of the four points producing any definite evidence of relationship.

Anatole Galsevski, in an eight-page pamphlet entitled “The Relationship of Dyak and Kanauri” (New York, 1921), has shown some interesting supposed connections, but unfortunately the Dyak forms quoted are nothing like those shown in Gabalentz's grammar noted above and the Kanauri forms find no confirmation in either the Rev. T. Grahame Bailey's vocabulary of Standard Kanauri (R. Asiatic Soc Monograph, no. 13, 1911) or in his grammar and vocabulary of Lower Kanauri (Linguistic Studies from the Himalayas, R.A.S.M., no. 17 (1915) pp. 46–77). The forms of the personal pronouns will show this difference—

	“Dyak” (Galsevski)	“Kanauri” (Galsevski)	Dyak (Gabalentz)	Standard Kanauri (Bailey)	Lower Kanauri (Bailey)
I	nau	na	aku, jaku	gö	gu
thou	karu	keran	ikau	ka, ki	ka'
he	ko	ko	ia	ju, du, nu	nau, no
we (incl).	ni	ni	ikei, ita	nina'	kishu
we (excl.)	nito	no	ikei, ita	kishona'	kishu
you	kitan	kintan	keton	kinā'	ki
they	kong	kon	ia, awen	jugo, dogo, nugo	nogau

Five of Galsevski's “Kanauri” forms are similar to the Purik (another Mongolo-Assamese language in some respects similar to Kanauri) forms noted by Bailey (Ling. Stud., p. 15). I cannot find any trace of the so-called “Dyak” forms anywhere and can only presume them to be an invention of Galsevski's own mind. Incidentally, Galsevski does not quote any dual forms although both Dyak and Kanauri possess complete sets in this field.

Although Galsevski's short study is in itself worthless it does show very clearly the ever-present problem which faces the student of Sino-Tibetan (or Mongolo-Assamese, as I prefer to call them) and Altaic languages. Sir George Grierson, the genius behind the Linguistic Survey of India, has remarked that non-literary Mongolo-Assamese languages change very rapidly. In the course of an article on the Lui and Kadu languages he notes that “there are cases on record in which a Naga [a race of eastern Assam] colony which migrated to an isolated hill, in two generations spoke a language that was unintelligible to members of the parent tribe” (B.S.O.S., vol. 2, part 1 (1921) p. 41.)

- 40

In an article on the words for “a hundred” in Mongolo-Assamese languages (B.S.O.S., vol. 6, part 3 (1931) pp. 667–668) J. Przyluski and G. H. Luce have given us a good example of this wide divergence of forms when they construct a Primitive Mongolo-Assamese word pargyak or parugyak from which are derived forms as varied as these—

Taic	pāk	Ladakhi	rgya
Nung	pāk	Burmese	rā
Dioi	pāk	S. Chin Dialect (Yawdwin)	pra
Chinese (6th century)	pak	C. Chin Dialect (Pankhu)	raja
Classical Tibetan	brgya	Kuki Dialect (Purum)	riyah

The modern Tibetan is gya or gya-tham-pa, incidentally (C. A. Bell: English-Tibetan Colloquial Dictionary, Calcutta (1920) p. 234).

Turning from the supporters of the Turanian origin to a solitary one of the Indian solution as such, we find Dr. W. J. Perry in his little book entitled The Growth of Civilization (1924; Pelican Book A. 15 (1937) pp. 110–118) actually suggesting, indeed stating, that the archaic megalithic culture, which has occasionally left its mark in Oceania, came from India (cf. Lanyon-Orgill: The Easter Island Script, J.P.S., vol. 51 (1942) p. 189). He also finds some affinities with ancient Egypt and sums up his general conclusions in these words: “it can be shown that in those parts of Oceania which are rich in pearls, gold or some analogous forms of wealth the culture of the peoples approximates, or formerly approximated with varying degrees of closeness, to that of Egypt in the Pyramid Age. We find mummification for rulers, the dual organization of society, the ruling class of the Children of the Sun, the sun-cult, irrigation, polished stone implements, megalithic monuments, and many other features of culture which, as has been urged in The Children of the Sun [another of Dr. Perry's more lengthy perpetrations], came into existence in Egypt. On the other hand, in those parts of Oceania where such valuable things are lacking, such as the Gilbert, Ellice, Tokelau and other groups, which have been colonized later, this archaic civilization is in ruins, and but little trace of it can be detected”.

He draws attention to the Ponape remains in the Caroline Islands and contends that their builders were “quite capable of civilizing America” (p. 114). He closes by stating that both the breadfruit and banana came direct from India—whether this is so or not I do not know but the contention of a supposed origin of the Oceanic culture in India cannot be supported by philological facts at all, as has been shown in the first part of the present paper.

In his excellent summary and criticism of Dr. Erland Norden-skiold's views on the Oceanic influence on American Indian culture (J.P.S., vol. 51 (1942) pp. 126–135) Mr. Kenneth Emory has concluded that “if ever it existed, the influence of the one upon the other was negligible”. I am inclined to think that in respect of the possibility of some connection between the Oceanic and American Mr. A. C. Bossom in his “Short Account of the ancient craftsmen of British Columbia and their Arts” (Guide to the Exhibition of Arts and Crafts of British Columbia Indians, London: Imperial Institute, 1929), has taken the correct line when he divides the American Indians into two main groups: those of British Columbia who resemble the Chinese in their culture and somatological make-up, and those of the rest of America who differ in both these aspects. The reader may take Mr. Bossom's remark (p. 5) that “the only natives who came near to their - 41 [i.e. the B.C. Indians] great ability, are those elaborate wood carvers of the islands in the Southern Pacific which were formerly known as the Bismarck Archipelago” in whatever way he likes. I can only remark that Haida, Tlingit, Kwakiutl, and Tsimshian (British Columbian languages) bear no possible resemblance to the tongues spoken in the Bismarck Archipelago.

Mr. Johannes C. Andersen has kindly brought to my notice a remark by the Rev. J. Stack (Transactions of the Australasian Association for the Advancement of Science, vol. 3 (1891) p. 369): “Mr. Abraham Fornander . . . believes it to be of Aryan origin, but Aryan of a pre-Vedic and pre-Iranian era, and that its identity with the pre - Malay dialects still existing in Malaysia is now fully established; and not only so, but that it is manifestly the older surviving form of a once common language. Thanks to its isolation for long ages, it has preserved the ancient simplicity of its structure. The rudimentary simplicity of its alphabet attests the early age at which it separated from its kindred in the Asiatic Archipelago”.

It seems to me as if Fornander's theory is the same—basically—as that which I put forward in my paper on “The Easter Island Script” (J.P.S., vol. 51 (1942) pp. 189–190) in which I briefly outlined the general movements of the World's peoples as suggested by the evidence of languages. It can be said that certain features are common to all races, whether they reside in Western Europe or the basin of the Amazon, New York or Timbuktoo.

Finally, in reply to possible critics of my method, I should like to cite two sentences by Edward Tregear (J.P.S., vol. 13 (1904) p. 107): “The narrower scholars among the ranks of philologists will certainly condemn me at the outset for daring to compare words in inflected with those in agglutinative or monosyllabic languages. My answer is that I am dealing with absolutely prehistoric words, or roots of words, having their origin in ages so remote that in common honesty they can no more be classed as Aryan than as Semitic or Turanian.” Kornelius Jarnsen contended that certain groups of words were never wholly borrowed, i.e. that at least some of them were always native. Such words may be grouped as follows:

• (a) Pronouns.
• (b) Numerals.
• (c) Basic relationship terms.
• (d) Names of external parts of the human body.
• (e) Names of natural forces and conditions, especially words dealing with the weather.
• (f) Names of basic colours (under certain conditions).

Rivers' remarks (The History of Melanesian Society, Cambridge (1914) vol. 2, p. 464) that “the comparison of Melanesian terms of relationship . . . led me to the conclusion that the islands of Melanesia were at one time inhabited by peoples possessing a great diversity of speech. The uniformity of the terms for certain relationships led me to the view that into this condition of linguistic diversity there had come a people through whose influence modifications of the social organisation were produced, and that with these modifications a need for new terms had arisen which had been supplied from the language of the immigrants” is a very adequate example of Jarnsen's third group in operation.

Hulbert (op. cit., pp. 114–115) in his chapter on the “Glossarial Affinities” of Korean and Dravidian discusses the whole question in a - 42 very concise and erudite fashion. His views on the affinities of Korean and Oceanic have been discussed above: all that now remains is the final piece of evidence in the form of a complete comparative grammar and dictionary.

BIBLIOGRAPHY AND REFERENCES.

Abbreviations used:

• Anth.—Anthropos, International Review of Ethnology and Linguistics, etc. (Vienna).
• B.N.I.—Bijdragen tot de Taal-Land-en Volkenkunde van Nederlandsch Indië (Den Haag).
• J.H.U.C.—Johns Hopkins University Circular.
• J.P.S.—Journal of the Polynesian Society.
• J.R.A.I.—Journal of the Royal Anthropological Institute of Great Britain and Ireland (London).

(1) OCEANIC LANGUAGES.

• BLAKE, DR. FRANK R.—A Grammar of the Tagálog Language, the chief native idiom of the Philippine Islands (New Haven, Conn: Amer. Oriental Soc., 1925, vol. 1).
• BLAKE, F. R.—“Sanskrit Loan-Words in Tagálog” (J.H.U.C., 163, pp. 63–65).
• BLAKE, F. R.—“Analogies between Semitic and Tagálog” (J.H.U.C., 163, pp. 65–66).
• BOPP, FRANZ—Uber die Verwandtschaft der Malayisch-Polynesischen Sprachen mit den Indisch-Europäischen (Berlin, 1841).
• BRANDSTETTER, DR. RENWARD—Monographien zur Indonesischen Sprachforschung (12 vols.) (Luzern: Verlag der Buchhandlung E. Haag)—

• 1. Die Beziehungen des Malagasy zum Malaiischen, 1893.
• 2. Die Geschichte von König Indjilai. Eine Bugische Erzählung ins Deutsche übersetzt, 1895.
• 3. Tagalen und Madagassen, 1902.
• 4. Ein Prodromus zu einem vergleichenden Wörterbuch der malaio-polynesischen Sprachen, 1906.
• 5. Mata-Hari, 1908.
• 6. Wurzel und Wort in den indonesischen Sprachen, 1910.
• 7. Sprachvergleichendes Charakterbild eines indonesischen Indiomes, 1911.
• 8. Gemeinindonesisch und Urindonesisch, 1911.
• 9. Das Verbum. Dargestellt auf Grund einer Analyse der besten Texte in vierundzwanzig indonesischen Sprachen.
• 10. Der Artikel des Indonesischen verglichen mit dem des Indo-germanischen, 1913.
• 11. Indonesisch und Indogermanisch in Satzbau, 1914.
• 12. Die Lauterscheinungen in den indonesischen Sprachen, 1915.

[Of the above, 6, 8, 9, and 12 have been translated into English by C. O. Blagden (Roy. Asiatic Soc. Mon., 15, 1916), 1 into English by R. Baron, 2 into Dutch by M. C. Poensen, and 3 into Spanish by P. L. Stangl.]

• CAMPBELL, REV. WILLIAM—The Articles of Christian Instruction in Favorlang-Formosan, Dutch and English, from Vertrecht's manuscript of 1650, with Psalmanazar's Dialogue between a Japanese and a Formosan, and Happart's Favorlang Vocabulary (London: Kegan Paul, Trench, Trübner and Co., Ltd., 1896).
• CHRISTIAN, F. W.—“Some Hindustani Cognates of the Maori” (J.P.S. 22 (1913) pp. 77–79).

- 43

• CHRISTIAN, F. W.—“Polynesian and Oceanic Elements in the Chimu and Inca Languages” (J.P.S. 41 (1932) pp. 144–156).
• CHURCHILL, WILLIAM—The Polynesian Wanderings: Tracks of the Migration Deduced from an Examination of the Proto-Samoan Content of Efaté and other Languages of Melanesia (Washington: Carnegie Institution, 134, 1911).
• [I have used the late S. H. Ray's own copy, including many additional MSS., at present in the Library of the School of Oriental Studies in London, no. 38579.]
• CODRINGTON, REV. ROBERT HENRY—The Melanesian Languages (Oxford, at the Clarendon Press, 1885).
• COLLINGRIDGE, GEORGE—“Supposed Early Mention of New Zealand” (J.P.S. 5 (1896) pp. 123–124).
• COWIE, ANDSON—English-Sulu-Malay Vocabulary, with Useful Sentences, Tables, &c. . .edited by William Clark Cowie. Grammatical Introduction by the Editor (London: Theo. May, for the Br. N. Borneo Co., 1893).
• FRASER, DR. JOHN—“The Malayo-Polynesian Theory” (J.P.S. 5 (1896) pp. 92–100).
• HAPPART—(see Campbell).
• HULBERT, HOMER B.—A Comparative Grammar of the Korean Language and the Dravidian Languages of India (Seoul: Methodist Publishing House, 1906).
• HUMBOLDT, KARL WILHELM VON—Uber die Kawi-Sprache auf der Insel-Java, nebst einer Einleitung über die Verschiedenheit des menschlichen Sprachbaues und ihren Einfluss auf die geistige Entwickelung des Menschengeschlechts (3 vols.) (Berlin: Gedrukt in der Druckerei der Königlichen Akademie der Wissenschaften, 1836–38–39).
• KERN, H.—“Over zogenaamde verbindingsklanken in het tagala en wat daarmea overeenkomt in't Kawi” (B.N.I. 25 (1876) pp. 138–157).
• KERN, H.—“Sanskritische woorden in het tagala” (B.N.I., 4, 4, pp. 535–564).
• LABBERTON, DR. D. VAN HINLOOPEN—“Preliminary Results of Researches into the Original Relationship between the Nipponese and the Malay-Polynesian Languages” (J.P.S. 33 (1924) pp. 244–280; originally read at the Pan-Pacific Science Congress, Melbourne, August, 1923).
• MACDONALD, REV. DANIEL—“The Asiatic or Semitic Origin of the Oceanic Numerals, Personal Pronouns, Phonology and Grammar” (J.P.S. 5 (1896) pp. 212–232).
• MACDONALD, D.—“The Asiatic (Semitic) Relationship of the Oceanic Family of Languages: Triliteralism and Internal Vowel Change” (J.P.S. 13 (1904) pp. 197–209).
• MACDONALD, D.—The Oceanic Languages, their Grammatical Structure, Vocabulary, and Origin (London: Henry Frowde, 1907).
• MARSDEN, WM.—Miscellaneous Works of William Marsden, F.R.S,, &c. &c. (London: Printed by J. L. Cox and Son. Published by Parbury, Allen, and Co., 1834).
• [1. On the Polynesian, or East-Insular Languages; 2. On a Conventional Roman Alphabet, applicable to Oriental Languages; 3. Thoughts on the Composition of a National English Dictionary.]
• MINGUELLA T.—Unidad de la especie humana probada par la filologia (Manila, 1889).
• PARKER, G. W.—A Concise Grammar of the Malagasy Language (London: Trübner & Co., 1883).
• PEAL, S. E.—“The Malayo-Polynesian Theory, comments on Dr. Fraser's Paper” (J.P.S. 5, (1896) pp. 101–108).

- 44

• RAY, SIDNEY HERBERT—A Comparative Vocabulary of the Dialects of British New Guinea . . . with Preface by Dr. Robert Needham Cust (London: S.P.C.K., 1895).
• RAY, S. H.—“The Common Origin of the Oceanic Languages” (J.P.S. 5 (1896) pp. 58–68).
• RAY, S. H.—“Notes on supposed Hindu Cognates of the Maori” (J.P.S. 22 (1913) p. 225).
• RAY, S. H.—“Polynesian Languages of the Micronesian Border” (J.P.S. 21 (1912) pp. 164–172; 24 (1915) pp. 62–64, 92–97).
• RAY, S. H.—A Comparative Study of the Melanesian Island Languages (Cambridge, at the University Press, 1926).
• SELIGMANN, DR. C. G.—The Melanesians of British New Guinea, with a chapter by F. R. Barton, C.M.G., and an Appendix by E. L. Giblin (Cambridge, at the University Press, 1910).
• SERRANO, DON ROSALIO—Diccionario de Termínos Comunes Tagalo-Castellano sacado de graves autores (Manila: Imprenta de Santos y Bernal, 1910).
• TREGEAR, EDWARD—“Polynesian Origins” (J.P.S. 13 (1904) pp. 105–121, 133–152).
• WALLACE, ALFRED RUSSEL—The Malay Archipelago; the land of the Orang-Utan, and the Bird of Paradise. A Narrative of Travel, with studies of Man and Nature (2 vols.) (London: Macmillan and Co., 1969).
• WINSTEDT, SIR RICHARD OLAF—Malay Grammar (Oxford, at the Clarendon Press, 1927).

(2) SEMITIC LANGUAGES.

• MERCER, SAMUEL A. B.—Ethiopic Grammar with Chrestomathy and Glossary (Oxford, at the Clarendon Press, 1920).

(3) BANTU-SUDANESE LANGUAGES.

• JOHNSTON, SIR H. H.—The Bantu and Semi-Bantu Languages (2 vols.) (Oxford, at the Clarendon Press, 1919–1922).
• [I have also consulted: Ch. Maples on Makua (London, 1879), M. W. Bulley on Nyanja (1925), Meredith Sanderson on Yao (1922), Murray-Jardine (1927), and several others for individual languages of the area.]

(4) AUSTRALIAN AND PAPUAN LANGUAGES.

• BIRD, W. H.—“Some remarks on the grammatical construction of the Chowie-Language, as spoken by the Buccaneer Islanders, N.W. Australia” (Anth. 5 (1910) pp. 454–456).
• HANKE, A.—Grammatik und Vokabularium der Bongu-Sprache, (Astrolabebai, Kaiser-Wilhelmsland), mit . . . einem Vokabularium der Sungumana-Sprache (Berlin: Archiv für das Studium deutscher Kolonialsprachen, bd. 8, 1909).
• KEYSSER, CHRISTIAN—Wörterbuch der Kâte-Sprache gesprochen in Neuguinea, Dictionary of the Kâte-Language as spoken in New Guinea. Siebentes Beiheft zur Zeitschrift für Eingeborenen-Sprachen (Berlin: Diètrich Reimer, Ernst Vohsen; Hamburg: C. Boysen, 1925).
• MARSDEN, op. cit.
• SAVILLE, REV. W. J. V.—“A Grammar of the Mailu Language, Papua” (J.R.A.I. 42 (1912) pp. 397–436).
• SCHMIDT, DR. PADRE WILHELM—“Die Gliederung der australischen Sprachen” (Anth. 7 (1912) etc.).

(5) AMERICAN LANGUAGES.

• HORDEN, RT. REV. J.—A Grammar of the Cree Language (London, S.P.C.K., 1934).

- 45

• KOCH-GRUNBERG, DR. THEODOR—“Betóya-Sprachen nordwestbrasiliens und der angrenzenden Gebiete” (Anth. 7 (1912) pp. 429–462).
• MARSDEN, op. cit.
• MCDONALD, VEN. ARCHDEACON—A Grammar of the Tukudh Language (London, S.P.C.K., 1911).
• PREUSS, PROF. DR. K. TH.—“Forschungsreise zu den Kágaba-Indianern der Sierra Nevada de Santa Marta in Kolombien” (Anth. 14–15 (1919–1920) pp. 314–404, 1040–1079).

(6) TURANIAN LANGUAGES.

• ASTON, W. G.—A Grammar of the Japanese Spoken Language (Yokohama: Kelly & Walsh; Tokyo: The Hakubunsha, 1888).
• BARU, TANASHITI—Kirisuto-kyo: Inori: Tokunoshima (Christianity: Prayer: Tokunoshima) (Tokyo: Kyu-du-shu, 1899).
• CHAMBERLAIN, PROF. BASIL HALL—“A Preliminary Notice of the Luchuan Language (J.R.A.I. 26 (1897) pp. 47–59).
• COMMERCIAL PRESS—Model English-Chinese Dictionary with Illustrative Examples (Commercial Press and Shanghai-Utopia University Press, 1929).
• DARROCH, DR. J.—Marlborough's Chinese Self-Taught and Grammar (London: Marlborough and Co., 1922).
• HAMILTON, R. C.—The Dafla Language (Shillong, 1900).
• HULBERT, op. cit.
• JUDSEN, REV. DR. A.—A Grammar of the Burmese Language (Rangoon: American Baptist Mission Press, 1888).
• LIETARD, LE PERE A.—“Au Yunnan, Min-kia et La-ma jen” (Anth. 7 (1912) pp. 677–705).
• MARSDEN, op. cit.
• MORAVIAN MISSIONARIES—The Gospel of Matthew in Nicobarese (Nancowry Dialect), edited and revised by F. A. de Rolpstorff . . . and his widow in 1884 (Calcutta: Auxiliary Bible Society, 1890).
• MULLIE, P. JOSEPH—“Phonetische Untersuchungen über die nord-pekinesischen Sprachlante” (Anth. 8 (1913) pp. 436–466).
• SCHMIDT, P. W.—“Die Mon-Khmer Völker, ein Bindeglied zwischen Völkern, Zentralasiens und Austronesiens” (Braunschweig, Archiv für Anthropologie, Neue Folge, Bd. 5, Heft 1 und 2, 1906).
• SLOAN, W. H.—A Practical Method with the Burmese Language (Rangoon: American Baptist Mission Press, 1911).
• SUMMERS, J.—Handbook of the Chinese Language (Oxford, at the Clarendon Press, 1863).
• TERRY, E. G.—Chinese Simplified; being a Practical Grammar of the Chinese Language (London and Durban: Luzac and Co., 1905).
• VERGIN, B. B.—English-Siamese Conversation (Penang, n.d.).
•  http://www.jps.auckland.ac.nz/document/Volume_52_1943/Volume_52%2C_No._2/The_origin_of_the_Oceanic_languages%2C_by__P._A._Lanyon-Orgill%2C_p_25-45/p1?page=0&action=searchresult&target=